Sports Economics Theory, Evidence and Policy

Sports Economics
Books in the Sport Management Series

Sport Governance
Russell Hoye and Graham Cuskelly
Sport and the Media

Matthew Nicholson
Sport Funding and Finance

Bob Stewart
Managing People in Sport Organizations

Tracy Taylor, Alison J. Doherty and Peter McGraw
Introduction to Sport Marketing

Aaron Smith
More information on the series can be found online by visiting

www.elsevierdirect.com
Sport Management Series

Series editor
Russell Hoye
Sports Economics
Theory, Evidence and Policy
Paul Downward
Alistair Dawson
Trudo Dejonghe
AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD

• PARIS • SAN DIEGO • SAN FRANCISCO • SYDNEY • TOKYO
Butterworth-Heinemann is an imprint of Elsevier
Butterworth-Heinemann is an imprint of Elsevier
Linacre House, Jordan Hill, Oxford OX2 8DP, UK
30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
Copyright © 2009, Elsevier Ltd. All rights reserved
No part of this publication may be reproduced, stored in a retrieval system

or transmitted in any form or by any means electronic, mechanical, photocopying,
recording or otherwise without the prior written permission of the publisher
Permissions may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333;
email: [email protected]. Alternatively visit the Science & Technology Books
web site at www.elsevierdirect.com/rights for further information
Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons
or property as a matter of products liability, negligence or otherwise, or from any use
or operation of any methods, products, instructions or ideas contained in the material
herein.
British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
ISBN: 978-0-7506-8354-8
For information on all Butterworth–Heinemann publications

visit our web site at www.elsevierdirect.com
Printed and bound in Great Britain
09 10 11 12 13 10 9 8 7 6 5 4 3 2 1
Contents
Series editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

About the authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv
Sports Management series preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Preface to Sports Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
List of figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
List of tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
Chapter 1 The Economics of Sport

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 What is Economics? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Economic Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 A Core Economic Model: Perfect Competition and Efficient
Resource Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4.1 Positive economics . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4.2 Normative economics . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Market Failure and the Rationale for Policy Intervention . . . 14
1.5.1 Monopoly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5.2 Equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.3 Externalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.4 Public goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.5 Imperfect information . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6. Limitations to Policy Maker Intervention. . . . . . . . . . . . . . . . 24
1.6.1 The Coase theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6.2 Government failure . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.7 Expenditure-Based Rationale for Policy . . . . . . . . . . . . . . . . 25
1.8 An Empirical Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Appendix 1.1 The relationship between average and marginal
revenue in monopoly . . . . . . . . . . . . . . . . . . . . . . . 31
Appendix 1.2 The derivation of the Keynesian Expenditure
Multiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
v
vi Contents
Chapter 2 The Nature, Organization and Economic Significance of Sport

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 What is Sport? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3 Economic Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4 Some International Evidence on the Sports Environment . . . 40
2.4.1 Public policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4.2 Sports policy bodies . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4.3 Governing bodies and professional team
sports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.4.4 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.5 Events and Mass Participation . . . . . . . . . . . . . . . . . . . . . . . 53
2.5.1 Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.5.2 Mass Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.6 Historic and Current Policy in Sport: Changing Emphases
and Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Chapter 3 The Economics of Sports Participation

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2 A General Economic Model of Sports Consumption . . . . . . 65
3.2.1 The dual decision hypothesis . . . . . . . . . . . . . . . . . . 66
3.2.2. The income–leisure trade-off . . . . . . . . . . . . . . . . . . 69
3.2.3 The derived demand for sports . . . . . . . . . . . . . . . . 76
3.3 An Integrating Framework for Consumer Choice . . . . . . . . 82
3.4 Policy Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Appendix 3.1 Utility-maximizing behaviour . . . . . . . . . . . . . . . 89
Appendix 3.2 Durable goods, the time value of money
and investment demand . . . . . . . . . . . . . . . . . . . . 92
Chapter 4 The Economics of Sports Participation: Evidence

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Empirical Evidence on Participation and Expenditure in
Sports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2.1 Mass sports participation: Sources of official and
unofficial data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2.2 Mass sports participation: Descriptive
evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2.3 Expenditure on mass sports participation . . . . . . . . . 104
Contents vii
4.3 Mass Sports Participation: Theoretical and Empirical

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3.1 Leisure time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3.2 Participation in sport. . . . . . . . . . . . . . . . . . . . . . . . . 110
4.3.3 Expenditure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.3.4 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Chapter 5 The Supply of Participant Sport: The Public and Private Sector
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2 The Structure of Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.3 The Public Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.3.1 UK sports policy and provision . . . . . . . . . . . . . . . . 120
5.3.2 Other sports policy and provision . . . . . . . . . . . . . . . 121
5.4 The Economic Rationale for Public Policy . . . . . . . . . . . . . 123
5.4.1 Economic efficiency . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.4.2 Equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.4.3 UK policy revisited: Mechanisms to promote
participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.5 The Private Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.5.1 Market structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.5.2 Pricing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.6 Informal Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.7 Sports Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Appendix 5.1 Some elements of game theory . . . . . . . . . . . . . . 145
Chapter 6 The Supply of Participant Sport: Volunteers and Sports Clubs

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.2 The Sports Club Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.2.1 European hierarchical sports organization . . . . . . . . 148
6.3 The Economic Foundation of Sports Club Systems . . . . . . . 151
6.3.1 Club goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.3.2 Provision of club goods . . . . . . . . . . . . . . . . . . . . . . 155
6.3.3 Hierarchical form . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.3.4 Economic evaluation . . . . . . . . . . . . . . . . . . . . . . . . 160
6.4 Volunteering in Sport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.4.1 Defining volunteering . . . . . . . . . . . . . . . . . . . . . . . . 161
viii Contents
6.4.2 Motivation and constraints on volunteering . . . . . . . 163

6.4.3 The profile and scale of volunteering . . . . . . . . . . . . 165
6.5 An Economic Analysis of Volunteering . . . . . . . . . . . . . . . . 167
6.6 Policy Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
6.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Appendix 6.1 The general theory of second-best . . . . . . . . . . . . 174
Chapter 7 The Market for Professional Sports: General Themes

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
7.2 The Transition to Professionalism . . . . . . . . . . . . . . . . . . . . 178
7.2.1 Sports clubs: Key concepts revisited . . . . . . . . . . . . 178
7.2.2 Formalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2.3 Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
7.3 Sports Competitions as Economic Contests . . . . . . . . . . . . . 184
7.3.1 Contest/tournament theory . . . . . . . . . . . . . . . . . . . . 184
7.4 Foundations of the Analysis of Professional Team Sports . . . 192
7.4.1 The production of professional team sports . . . . . . . 193
7.4.2 The peculiar economics of sport. . . . . . . . . . . . . . . . 193
7.4.3 The league as a natural monopoly or as a cartel . . . . 194
7.5 US and European Leagues . . . . . . . . . . . . . . . . . . . . . . . . . . 196
7.5.1 Club objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
7.5.2 League characteristics . . . . . . . . . . . . . . . . . . . . . . . 197
7.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Appendix 7.1 Some properties of tournament discrimination . . 201
Appendix 7.2 The derivation of optimal tournament effort . . . . 202
Chapter 8 Uncertainty of Outcome, Competitive Balance and Bias

in Sports Leagues
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
8.2 The Time-Dependent Nature of Uncertainty
of Outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
8.3 Short-Run (Match) Uncertainty of Outcome . . . . . . . . . . . . 207
8.3.1 Relative league standings . . . . . . . . . . . . . . . . . . . . . 207
8.3.2 Betting odds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
8.3.3 Direct estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
8.3.4 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
8.4 Medium-Term Uncertainty of Outcome . . . . . . . . . . . . . . . . 213
8.4.1 Within-season uncertainty of outcome . . . . . . . . . . . 213
Contents ix
8.4.2 Seasonal (team-specific) uncertainty of outcome . . . 214

8.4.3 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
8.5 Long-Run Uncertainty of Outcome . . . . . . . . . . . . . . . . . . . 219
8.5.1 Competitive balance . . . . . . . . . . . . . . . . . . . . . . . . 219
8.5.2 Evolution of competitive balance . . . . . . . . . . . . . . . 224
8.5.3 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
8.6 Home Advantage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
8.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Appendix 8.1 Reconciliation of Jennett’s and Borland and
Lye’s models of medium-term uncertainty
of outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Appendix 8.2 Calculating the dispersion of sporting success
in sports leagues using standard deviations . . . . . . 231
Chapter 9 Cross-Subsidization in Professional Sports Leagues

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
9.2 The Rationale for Cross-Subsidization . . . . . . . . . . . . . . . . . 234
9.3 Methods of Cross-Subsidization. . . . . . . . . . . . . . . . . . . . . . 235
9.3.1 Sporting labour markets . . . . . . . . . . . . . . . . . . . . . . 235
9.3.2 Revenue redistribution . . . . . . . . . . . . . . . . . . . . . . . 237
9.4 An Economic Framework for Understanding
Cross-Subsidization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
9.4.1 Profit-maximizing competitive equilibrium . . . . . . 239
9.4.2 Cross-subsidization in a profit-maximizing
league . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
9.4.3 Equilibrium in a win-maximizing league . . . . . . . . . 248
9.4.4 Cross-subsidization in a win-maximizing league . . . 249
9.5 Relaxation of Assumptions. . . . . . . . . . . . . . . . . . . . . . . . . . 249
9.5.1 Introducing the effect of spectator preference for
absolute quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
9.5.2 Market size diseconomies. . . . . . . . . . . . . . . . . . . . . 251
9.5.3 A variable supply of talent . . . . . . . . . . . . . . . . . . . . 253
9.6 When Does the “Invariance Proposition” Hold? . . . . . . . . . 254
9.7 How Far Does Competitive Balance Matter?
The European Champions’ and other Leagues . . . . . . . . . . . 254
9.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Appendix 9.1 Why win percent and talent distribution
are not formally equivalent . . . . . . . . . . . . . . . . . 258
Appendix 9.2 The quadratic model under
profit maximization . . . . . . . . . . . . . . . . . . . . . . 259
x Contents
Chapter 10 The Demand for Professional Team Sports: Attendance and

Broadcasting
10.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
10.2. The Demand for Professional Team Sports . . . . . . . . . . . . . 262
10.3 Theoretical and Empirical Issues . . . . . . . . . . . . . . . . . . . . . 262
10.3.1 Measuring quantity . . . . . . . . . . . . . . . . . . . . . . . . . 262
10.3.2 Measuring price and income . . . . . . . . . . . . . . . . . . 263
10.3.3 Consumer preferences . . . . . . . . . . . . . . . . . . . . . . . 264
10.3.4 Estimation issues . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
10.4 Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
10.4.1 Measuring quantity . . . . . . . . . . . . . . . . . . . . . . . . . 268
10.4.2 Measuring price and income . . . . . . . . . . . . . . . . . . 283
10.4.3 Consumer preferences . . . . . . . . . . . . . . . . . . . . . . . 288
10.4.4 Estimation issues . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.5 Broadcast Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
10.5.1 A brief outline of sports and broadcasting history . . . 292
10.5.2 The economics of broadcasting and sport . . . . . . . . 295
10.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Appendix 10.1 Problems of measuring habit persistence . . . . . 299
Chapter 11 The Labour Market in Professional Team Sports

11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
11.2 Perfect Competition as a Benchmark Labour Market . . . . . 302
11.3 Historic Evolution of the Labour Market in Sport . . . . . . . 303
11.3.1 The United States . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
11.3.2 The UK and Europe . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.3.3 Player agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
11.3.4 Contract duration . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
11.4 Theorizing the Players’ Labour Market . . . . . . . . . . . . . . . . 310
11.4.1 Monopsony . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
11.4.2 Monopoly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
11.4.3 Bargaining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
11.4.4 Non-profit-maximizing behaviour . . . . . . . . . . . . . . 314
11.5 The Players’ Labour Market: Some Econometric
Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
11.5.1 The United States . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
11.5.2 Europe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
11.6 Managerial Efficiency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
11.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Contents xi
Appendix 11.1 General conditions for the demand for factors

of production . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Appendix 11.2 Profit-maximizing monopsony . . . . . . . . . . . . . 331
Appendix 11.3 The Nash bargaining model . . . . . . . . . . . . . . . . 332
Appendix 11.4 A win-maximizing labour market . . . . . . . . . . . 332
Chapter 12 The Economics of Sports Events and Infrastructure

12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
12.2 Sports Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
12.3 Investment Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
12.3.1 Theoretical issues . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
12.3.2 Public sector rationale . . . . . . . . . . . . . . . . . . . . . . . 345
12.3.3 The multiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
12.4 Types of Sports Event and Impacts . . . . . . . . . . . . . . . . . . . 354
12.4.1 Duration and scale of impacts . . . . . . . . . . . . . . . . . 354
12.4.2 Theorizing economic activity as an impact . . . . . . . 357
12.4.3 The practical measurement of economic activity as
an impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
12.5 Empirical Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
12.5.1 Ex ante evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
12.5.2 Explaining the variance . . . . . . . . . . . . . . . . . . . . . . 365
12.5.3 Ex post evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
12.5.4 Non-market valued effects . . . . . . . . . . . . . . . . . . . . 376
12.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
Appendix 12.1 NPV, wealth and profit maximization . . . . . . . . 380
Appendix 12.2 A simple input–output analysis of
multiplier effects given a simple
economic system . . . . . . . . . . . . . . . . . . . . . . . . 381
Appendix 12.3 Calculating welfare weights . . . . . . . . . . . . . . . 382
Appendix 12.4 The determinants of medal success . . . . . . . . . . 383
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
List of abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
Dedication
To Cath, Sarah-Louise and Joe who know that I am marginally less irritable playing sport than
writing about it, but without whom none of this would have been possible (PD).
To the friends whose affection sustained me during the writing of this book, and especially to the
memberships of the January the Second Society, and of the Last of the Summer Walking Club (AD).
To Ann, Arno and Ferre who have seen their father a lot of hours working at his computer and the dogs
Zeus and Dreyfus who accompanied me in the walks between the working hours (TD).
xii
Series Editor
Dr. Russell Hoye is an Associate Professor in the School of Sport, Tourism and
Hospitality Management, La Trobe University, Victoria, Australia. Russell
has been involved in sport management education since 1993, working in
Australia at La Trobe University, Griffith University, and Victoria University
and in China with The University of Hong Kong and Tsinghua University.
He is a board member of the Sport Management Association of Australia and
New Zealand (SMAANZ). He was the Guest Editor for the inaugural special
issue of Sport Management Review on professional sport in Australia and
New Zealand published in 2005.
Russell’s areas of expertise include corporate governance, organizational
behaviour, volunteer management and public sector reform within the sport
industry. He has acted as a consultant for the Australian Sports Commission,
Sport and Recreation Victoria and a number of local government and
non-profit organizations. His research interests focus on examining how
governance is enacted with sport organizations and how volunteers engage
with and are managed by sport organizations. He has published papers on
these topics in journals such as Nonprofit Management and Leadership,
Sport Management Review, European Sport Management Quarterly, Society
and Leisure, International Gambling Studies, Third Sector Review, Sporting
Traditions, Managing Leisure, Football Studies, Annals of Leisure Research,
and the Australian Journal on Volunteering.
xiii
About the authors
Dr Paul Downward is Director of the Institute of Sport and Leisure Policy,

University of Loughborough. Along with Alistair Dawson he is the author of
the textbook The Economics of Professional Team Sports (Routledge, 2000). He
is the editor of a book for the Office for National Statistics on the use of official
data in understanding sports, leisure and tourism markets, and the author of
numerous articles on the economics of sport. He has recently undertaken
consultancy work for Sport England on participation, volunteering in sports
clubs, and performance management, and for UK Sport on volunteering at
the Manchester Commonwealth Games. He is a member of the International
Association of Sports Economics, a founding member of the editorial board
of the Journal of Sports Economics, and member of the British Philosophy of
Sport Association. Outside of these areas Dr Downward has also published
on the philosophy of economics, and pricing theory. As well as refereed papers
he also has a book in each of these areas. He currently teaches a final year
undergraduate module in Sports Economics, and contributes sports economics
teaching to modules in Sports and the Leisure Industries, Managing Sports
Organizations and Olympic Studies, at Loughborough University. Paul
is currently playing veteran’s rugby (and can be spotted on the cover) and
coaches a junior side at his local club.
Alistair Dawson retired from Staffordshire University in 2000, but
maintains an active research interest, particularly in the theoretical problems
surrounding the measurement of uncertainty of outcome. Apart from The
Economics of Professional Team Sports, with Paul Downward, Alistair has
published articles in sports economics. Other fields in which Alistair has
published include the econometrics of wage inflation, and the application
of computer business and macroeconomic simulations to undergraduate
learning and assessment. Alistair hopes to devote his remaining years to hill
walking, photography, real ale and socializing with friends and family.
Dr Trudo Dejonghe is docent (professor) at the Lessius Business University
Antwerpen and guest lecturer in VUB Brussels, Hogeschool Brugge and
Copenhagen Business School. He is the author of Sport en economie: een
noodzaak tot symbiose (sport and economics: a symbiotic necessity) (Arko
Sports Media, 2004), Sport en economie: een aftrap (sport and economics: a
kick off) (Arko Sports Media, 2007) and Sport in de wereld: ontstaan, evolutie
en verspreiding (sport in the global space: evolution and diffusion) (Academia
Press, 2001; 2004; 2007). He has researched the location-allocation problems
of football clubs in Belgium (2001) and the Netherlands (2005).
xiv
Sport Management Series Preface
Many millions of people around the globe are employed in sport organizations
in areas as diverse as event management, broadcasting, venue management,
marketing, professional sport, and coaching as well as in allied industries
such as sporting equipment manufacturing, sporting footwear and apparel,
and retail. At the elite level, sport has moved from being an amateur pastime
to a significant industry. The growth and professionalization of sport has
driven changes in the consumption and production of sport and in the
management of sporting organizations at all levels of sport. Managing sport
organizations at the start of the twenty-first century involves the application
of techniques and strategies evident in the majority of modern business,
government and nonprofit organizations.
The Sport Management Series provides a superb range of texts for
the common subjects in sport business and management courses. They
provide essential resources for academics, students and managers and are
international in scope. Supported by excellent case studies, useful study
questions, further reading lists, lists of websites, and supplementary online
materials such as case study questions and PowerPoint slides, the series
represents a consistent, planned and targeted approach which:
• provides a high quality, accessible and affordable portfolio of titles
which match management development needs through various
stages;
• prioritises the publication of texts where there are current gaps in the
market, or where current provision is unsatisfactory;
• develops a portfolio of both practical and stimulating texts in all areas
of sport management.
The Sport Management Series is the first of its kind, and as such is
recognised as being of consistent high quality and will quickly become the
series of first choice for academics, students and managers.
xv
Preface to Sports Economics
With the help of Trudo Dejonghe, this book radically extends and revises
The Economics of Professional Team Sports by Paul Downward and Alistair
Dawson. Not only has the coverage of professional team sports been radically
updated, but the book also encompasses mass participation sport as well as
aspects of sports events.
This has provided challenges. On the one hand mass participation sport
has received relatively little research and textbook treatment in economics,
although some notable contributions, cited within, stand out. In this respect,
it is hoped that the book provides some innovative discussion. In contrast,
professional sport and the economics of major events comprise burgeoning
literatures, which provide challenges in summarizing its content. In this
respect it is hoped that the book provides a good introduction to the main
issues discussed in the literature.
A central reason for the combined treatment and coverage is that the
different contexts of sport are now often integrated as economic activities
in public policy discussion. As no previous economic literature exists that
examines all of these contexts and their connections, we hope that the major
contribution of the book is to show how this is possible, and that economics
can contribute greatly to our understanding of sport.
In order to achieve this goal, core principles of economic theory coupled
with specific conceptual innovations from economics in the context of sport are
employed throughout the book, whilst recognizing that each aspect of sport has
some distinctive characteristics. Relevant descriptive and inferential empirical
evidence is cited, as well as the implications of the theoretical and empirical
insights for policy. It should be emphasized that at no point are specific policy
recommendations offered. Rather, the aim is to use theory and evidence to
explore the underpinnings of alternative policy positions. Broadly speaking
these include a choice of either allowing market forces to allocate resources in
sport, or for policy makers to intervene in resource allocation to provide sport.
The economic logic and trade-offs of each position is essentially compared.
A distinctive feature of the book is that its pedagogy aims to make sports
economics accessible to those for whom economics is a relatively new area of
study, or is not necessarily a major component of their studies, for example,
for sports management students. In this respect, rather than using sports to
exemplify a traditional economics curriculum, this book makes the contexts
of sport the focus of analysis, and uses economics as and when it is needed
to illuminate these areas.
xvi
Preface to Sports Economics xvii
Moreover, because the pedagogy of economics can be technical, drawing

on mathematical and statistical analysis, most of each chapter, where
possible, explores the issues at stake in verbal terms, providing historical
and contextual commentary. Economic principles are presented in graphical
format with the statement of, and commentary on, key equations. Derivations
are kept to a minimum and developed most in appendices, for the more
technically-minded reader. Reflection questions and boxes are also used in
chapters to prompt the reader to think about specific points, as well as to
provide the context for specific theoretical or empirical contributions that
have been used to analyze sport.
The aim of the book was to be international in scope. However, partly as
a result of language and partly as a result of the emphasis of the literature,
it focuses primarily on comparisons between the US, the UK and mainland
Europe. In this respect it is hoped that the book prompts research into sports
economics in other contexts.
Chapters 1 and 2 provide overviews of key economic principles and
methodology, and the definition and economic and political context of sport,
respectively. Chapters 3 and 4 examine the demand for mass participation
sport, introducing key theoretical features of consumption behaviour, as well
as how economists analyze evidence. Evidence on participation is reviewed.
Chapters 5 and 6 then examine the supply of mass participation sport,
focusing on the policy rationale for public or private sector supply. The role
of clubs and volunteers in sport are also discussed, as well as the growth of
the private and informal sectors. In these chapters, the consumer–producer
nature of the economic agent, and the activity of agents within “club goods”
is emphasized.
Chapter 7 draws on this discussion to explain the origin and nature of
professional sports competitions, as an evolution of club goods in which
consumers and producers become specialized. Key aspects of the economics
of competition – as detailed in contest theory – are discussed in connection
with sport in general and then professional team sports. Themes in the
economics of professional team sports are reviewed connected with the nature
and objectives of the firm and the market structure of leagues. It is argued
that sports leagues are cartels whose members’ mutual interdependence
needs to be managed.
Chapter 8 provides a detailed digression on the nature and measurement
of “Uncertainty of Outcome” as a characteristic of this interdependence,
as highlighted in contest theory. This is because uncertainty of outcome
has assumed a central place in the literature examining the management
and evolution of sports leagues. Chapter 9 then reviews the efficacy of
policies that have been used by sports leagues with the purported intention
of managing uncertainty of outcome. It is shown that, broadly speaking,
the impact of policies will either be ineffective, or possible but difficult to
implement, depending on the assumptions made.
xviii Preface to Sports Economics
Chapters 10 and 11 then examine the key sources of revenues and costs
for professional sports; attendance and broadcast demand, and the labour
market. In Chapter 10 it is identified that uncertainty of outcome is more
likely to be important for broadcast audiences, and that market sizes and team
qualities and possibly loyalty are key drivers of attendance demand, which is
price inelastic. Chapter 11 argues that economic liberalization has produced
rising player costs for teams, but the specific nature of the market structure
is unclear. In this respect it is implied that league management policies have
traditionally exploited players rather than affected uncertainty of outcome.
The role of player–agents and the coach–manager are also discussed.
The book closes with a discussion of the economics of sports events
and investment in infrastructure in Chapter 12. Here a number of themes
corresponding to participation in, and provision of, sports events as distinct
economic entities to sports leagues are discussed, drawing on contest theory.
The chapter also examines in detail the economic case for investment in
sports infrastructure that could apply to events or professional team sports.
It is argued that the claim that benefits to society can be leveraged from
public sector investment in professional team sports or sports events lacks
a strong evidence base. Consequently, while a case can be made for such
investment, considerable caution and planning is required to harness any
spillover effects from the investments.
In concluding, it should be noted that sports economics is a growing area of
study and a number of individuals and contributions have added to a vibrant
development of ideas and research interests. It should become apparent
on reading the book where the driving force of these ideas originates. This
said it is clear that a book cannot do justice to all of the contributions that
have been made. It is also clear that much remains to be done in research
terms, and it is hoped that this book helps to stimulate some of this work.
Readers should note that there is now an International Association of Sports
Economists that hosts an annual conference dedicated to sports economics
(see http://www.iasecon.net/), and also a dedicated journal, The Journal of
Sports Economics, edited by Leo Kahane (see http://jse.sagepub.com/), that
provide key forums for current work.
List of Figures
Chapter 1
Figure 1.1 A perfectly competitive sports league
Figure 1.2 Increased demand for sports fixtures
Figure 1.3 Monopoly supply of sports fixtures
Figure 1.4 Monopoly versus perfectly competitive leagues
Figure 1.5 Positive and negative externalities
Figure 1.6 The intercept and slope of a linear function
Chapter 2
Figure 2.1 The sports economy
Chapter 3
Figure 3.1 Sports consumption
Figure 3.2 A simple circular flow of income
Figure 3.3 The demand for leisure and income
Figure 3.4 Income and substitution effects in leisure demand
Figure 3.5 Equilibrium demands for vests and training shoes
Figure 3.6 The effects of higher income on consumer demand
Figure 3.7 The effects of a price change on consumer demand
Figure 3.8 A theoretical demand curve for training shoes
Figure 3.9 A change in the relative price of good- and time-intensive
commodities
Figure A3.1.1 Consumption of training shoes over time
Chapter 5
Figure 5.1 Sports policy trade-off
Figure 5.2 Types of economic good
Figure 5.3 UK sports participation policy options
Figure 5.4 BV and policy outcomes
Figure 5.5 Payments versus subsidies
Chapter 6
Figure 6.1 The traditional (English) model of European mass sports
organization
Figure 6.2 The demand and supply of club goods
Figure 6.3 Heterogeneous demands and government failure
xix
xx List of Figures
Figure 6.4 Organizational efficiency

Figure 6.5 The volunteer decision
Figure A6.1.1 The second-best theorem
Chpter 8
Figure 8.1 The Lorenz curve and Gini coefficient
Chapter 9
Figure 9.1 Two-team league with Team 1 as the large market team
Figure 9.2 Equilibrium win percent and no incentive to trade players
Figure 9.3 Effect of revenue-sharing
Figure 9.4 Equilibrium in profit-maximizing (Ep) and win-maximizing
(Ew) leagues
Figure 9.5 Market size diseconomies (Team 1 maximizes profit at A,
wages Cp, Team 2 maximizes profit at B, wages Cd)
Chapter 10
Figure 10.1 A theoretical demand curve for match tickets
Figure 10.2 Actual and latent spectator demand
Chapter 11
Figure 11.1 Labour market structures
Figure 11.2 Wages and employment under monopsony (W2, E2) and
competition (W0, E0)
Figure A11.1.1 Equilibrium demands for labour and capital
Chapter 12
Figure 12.1 The impact of injections on prices and resource employment
Figure 12.2 The components of total economic value
Figure 12.3 The “all-or-nothing” demand curve
Figure A12.1.1 NPV and wealth maximization
List of Tables
Chapter 1
Table 1.1 Regression options
Chapter 2
Table 2.1 The origin and characteristics of other professional team sports
Table 2.2 Examples of official data participation sports
Table 2.3 Scale and distribution of sources of European football club
revenues
Table 2.4 US sports leagues
Table 2.5 Size of sports economy
Chapter 4
Table 4.1 Some official sports participation data
Table 4.2 Top ten and team sport participation rates
Table 4.3 Overall participation rates
Table 4.4 Socio-economic determinants of participation
Table 4.5 Consumer spending on sport
Table 4.6 The determinants of leisure time
Table 4.7 Results of a survey of studies using official data into the
choice to participate in various activities
Chapter 5
Table 5.1 Location of sports participation
Table 5.2 International comparisons of sport funding
Table 5.3 Externalities and participation in sport
Table 5.4 Sports participation 1987–2002 (%)
Table 5.5 Location of sport provision
Table 5.6 Private sector health and fitness clubs 2007
Table 5.7 Evolving marketing strategies
Table 5.8 Sports club participation
Table 5.9 Sports good sales
Chapter 6
Table 6.1 European sports structure
Table 6.2 Volunteer motivations and benefits
Table 6.3 Volunteer constraints
Table 6.4 Volunteer profiles
Table 6.5 The value of volunteering
xxi
xxii List of Tables
Chapter 7
Table 7.1 Comparison of voluntary and emergent professional sports
Table 7.2 Typical forms of tournament in sport
Table 7.3 Assumptions and implications of tournament theory
Table 7.4 Evolution of sports leagues
Table A7.1.1 Success probabilities as γ varies, for given levels of team effort
Chapter 8
Table 8.1 Some findings on the effects of match uncertainty
Table 8.2 Some findings on the effects of medium-term uncertainty
Table 8.3 Mean SDWPs in baseball and football by sub-periods,
1901–1990
Table 8.4 Some findings on the effects of competitive balance on
attendance
Table 8.5 Some findings on home advantage
Table A8.1.1 Significance of Team A’s last three games under various ass-
umptions about how many wins are needed (encompassing
model)
Chapter 9
Table 9.1 Revenue sharing in leagues with profit-maximizing teams
Chapter 10
Table 10.1 A selection of recent econometric findings on team sport
demand
Chapter 11
Table 11.1 US sport salaries ($)
Table 11.2 Premiership transfer spending
Table 11.3 A selection of econometric evidence on US labour markets in
sport
Table 11.4 A selection of econometric evidence on transfers and wages
in European sport
Chapter 12
Table 12.1 Types of sports events
Table 12.2 Forms of investment
Table 12.3 Summary of impacts of investment
Table 12.4 The economic impact of major events
Table 12.5 Flaws with economic impact studies
Table 12.6 Econometric evidence on economic impacts
Table 12.7 Contingent valuation analyses
Table A12.2.1 A simple input–output account € million
Table A12.2.2 Input–output coefficients
Table A12.4.1 Econometric evidence on performance (medals) in Olympic
Games
CHAPTER 1
The Economics of Sport
OBJECTIVES
& To understand some main features of economics

& To understand the relationship between sport and economics
& To appreciate the main features of economic methodology
& To understand how the perfectly competitive model of resource allocation
could apply to sport
& To appreciate why market failures are important in sport
1.1 INTRODUCTION
To understand how economics can be used to analyze sport requires an initial
appreciation of some key tenets of economics on which to base subsequent,
more detailed, discussion. To meet this need, this chapter begins by examining
definitions of economics and shows, in the subsequent section, how these are
related to the methodological emphasis of economic analysis. Section 1.4
illustrates the main emphasis of economics, and draws out the main policy
thrust of economics, by outlining the key theory of the perfectly competitive
model of resource allocation. A distinction is drawn between the production of
predictions from theory, or “positive” economics, and the evaluation of the
outcomes predicted, or normative economics. Section 1.5 presents elements
of market failure that are invoked to provide a rationale for policy intervention.
Each of the theoretical concepts, empirical approaches and policy implications
discussed in this chapter are then referred to or elaborated on in subsequent
chapters. Section 1.6 discusses the limits to policy action in sport, and a brief
introduction to the empirical approach employed in economics to test predic-
tions from theory is presented to conclude the chapter. 1
2 CHAPTER 1 The Economics of Sport
BOX 1.1 DEFINITIONS OF ECONOMICS

1. “Writers on political economy profess to teach, or to individual and social action which is most closely con-
investigate, the nature of wealth, and the laws of its nected with the attainment and with the use of the ma-
production and distribution, including, directly or re- terial requisites of well-being. Thus, it is on the one side a
motely, the operation of all the causes by which the study of wealth; and on the other, and more important
condition of mankind, or of any society of human beings, side, a part of the study of man.”
in respect of this universal object of human desire, is Marshall, A. (1952) Principles of Economics: An Intro-
made prosperous or the reverse. Note that any treatize ductory Volume, 8th edn. London: Macmillan and Co.,
on political economy can discuss or even enumerate all p. 1.
these causes; but it undertakes to set forth as much as is 3. “Economics is the science which studies human behav-
known of the laws and principles according to which iour as a relationship between ends and scarce means
they operate.” which have alternative uses”
Mill, J.S. (1900) Principles of Political Economy with Robbins, L. (1940) An Essay on the Nature and Signif-
Some of Their Applications to Social Philosophy. icance of Economic Science. London: Macmillan and
London: George Routledge and Sons, p. 13. Co., p. 16.
2. “Political economy or economics is a study of mankind
in the ordinary business of life; it examines that part of
1.2 WHAT IS ECONOMICS?

Box 1.1 presents three historical definitions of economics that Lawson (2003)
argues are widely acknowledged as foundations for the emphasis of economics.
They are abstract and general, and have similarities and differences. Both Mill
and Marshall, for example, emphasize the connection between individual
human activity and society, particularly through the production and distribu-
tion of material that contributes to well-being. There is some implication that
wealth is connected to measurable or tangible material that is, for example,
exchanged on markets. It is this that tends to link the traditional study of
economics with, for example, financial and industrial subject matter.
Clearly this could apply to professional team sports. In this context money
changes hands in the production, distribution and consumption of sport.
Money, of course, is the mechanism by which key sporting resources such
as players are obtained and allocated between the various teams, to use in
competition against their opponents on the field.1 The purchase and sale of
players, as well as payment to them to perform, requires financing decisions.
1
In this book the words “player,” “labour,” “talent” and “athlete” are used interchangeably to
refer to the competitors who, through their practise, produce sport. Note that this use of the
term is different to “athlete” in common language, which is often identified specifically for
those engaged in the disciplines of athletics.
Economic Methodology 3
Consequently, gate and television revenues need to be earned to pay players'

salaries. Clubs and their amalgamated organizations, such as leagues and
governing bodies, must coordinate match schedules since they cannot be
produced in isolation, and potential spectators must be informed where and
when matches are to occur. In turn, spectators need accommodation and a
means by which payment can be made, while restricting access to the sport to
non-payers. Historically, stadia were necessary features of commercial sports
supply. Spectators could also pay to watch broadcast sport, although payment
has not always been necessary.
Characteristics such as these indicate that, rather like the production,
distribution and consumption of other goods and services, professional sport
can be viewed as an economic process. Inputs, or factors of production, such as
labour (the athletes and manager/coach) are combined with capital (the sport-
ing field, equipment and so on) to produce, along with another team in the
league, a product (the fixture) that is sold to consumers (spectators and sup-
porters) typically in a stadium, or via broadcast media.
Yet Robbins' definition in Box 1.1, which has come to be embraced by
economics, is broader in concept and consistent with recognizing that eco-
nomic activity is not, of necessity, connected with the creation and redistri-
bution of material wealth per se. In this regard there are similarities between
professional team sports, as well as amateur sports and sport in general, as
economic phenomena, if choices have to be made over the allocation of
resources to supply and consume these activities. These issues are discussed
in some detail, for example, in Chapters 3 and 6. If nothing else, participating
or volunteering in sports activities “costs” time, which is a resource that is
scarce to everybody. This is in addition to the resource allocation issues
associated with the commercial provision of non-professional sports activi-
ties, as discussed in Chapter 5. Robbins has been one cornerstone of the
development of economic methodology, which is now discussed.
1.3 ECONOMIC METHODOLOGY

Robbins' legacies for economics are two-fold. The first is that scientific pro-
positions on economizing are associated with attempts to produce generalized
understanding or “laws” of behaviour. The second is that economics becomes
a tool for assessing rational choices between courses of action.
As far as testing economic theories is concerned, Friedman (1953) has been
particularly important and emphasizes testing the predictions of theory re-
gardless of the realism of its assumptions. Friedman's argument provides a
link between theories that can be constructed on idealized conceptions of
rational economic behaviour, that are literally false, for example motivated
by the definition of Robbins, and the empirical relevance of theory as an aid to

policy.
This broad method of analysis, in which rational economic behaviour is
postulated for economic agents which is used to deduce predictions that are
tested against data, is applied to all economic problems, including those
associated with sport. Economics purports to offer “covering law” explana-
tions of phenomena. The specific subject matter analyzed is understood in
terms of a general theory of behaviour. In its core assumptions connected with
rational behaviour, economic analysis does not make allowance for the anal-
ysis of different sports taking place in different countries or in different time
periods. However, it does attempt to accommodate different institutional
contexts. It is important to bear this issue in mind, as this approach is different
to those of disciplines such as sociology, history and politics, which may be
more familiar to, say, sports management students, and in which the specific
context and character of sports is explored.
The economic approach can be understood more clearly, as well as
highlighting some important economic concepts, by considering the core
economic model of perfectly competitive market allocation of resources. Both
its “positive” and “normative” characters are discussed. Although philosoph-
ically contentious these terms reflect, as distinguished by Friedman (1953),
that statements about economic activity can have the aspiration of being
value-free and concerned with testable theoretical propositions, or be con-
cerned with prescription, for example policy, based on value judgements.
1.4 A CORE ECONOMIC MODEL: PERFECT COMPETITION

AND EFFICIENT RESOURCE ALLOCATION
1.4.1 Positive economics
Although discussed in a more appropriate way in Chapter 9, the representa-
tion of a sports league is used to illustrate the model of perfect competition.
The sports league can be viewed as the industry, or synonymously the market,
producing the output of sporting contests, i.e., fixtures, with individual teams
or clubs within the league being viewed as firms within the industry. The fans
who pay to watch games are the consumers who demand sports fixtures. As
implied earlier, it is assumed that economic agents in markets are rational.
They have perfect information about the implications of their decisions, and
they pursue clearly defined goals. It is assumed that clubs or firms seek to
maximize their profit, while the fans seek to maximize their utility, which is
enjoyment, from viewing fixtures from the league, made possible through the
purchase of tickets.
A Core Economic Model: Perfect Competition and Efficient Resource Allocation 5
While this might seem to be a reasonably plausible scenario, notwithstand-

ing the assumption about the information possessed by clubs and spectators,
the model also assumes that each club and fan is “atomistic,” i.e., they are
infinitesimally small relative to the total number of clubs or fans. In this
regard, their individual decisions to buy and sell tickets to watch sports fix-
tures cannot affect industry-level activity as a whole. It is also assumed that
there is freedom of entry and exit to the industry. This implies that unprofit-
able clubs can leave the market or league, while profitable clubs may attract the
entry of other clubs looking to compete for their profits. Clubs may also offer
more or less fixtures to the league as required. In this respect, resources in this
sports market are free to come and go. It is also assumed that the product
supplied by any set of teams – a sports fixture – is identical to all others,2 and
there are no differences assumed between the quality of sports fixtures or the
identity of fans to particular teams. The only impetus to watching one partic-
ular fixture rather than another is the ticket price. However, because clubs
cannot affect ticket prices, they cannot brand their fixture, and cannot draw on
particular allegiances from fans. Everyone has to accept the ticket price as
established in the market as a whole. Clubs are “price takers.” These are clearly
unrealistic assumptions.
To explore the implications of these assumptions in more detail requires
exploring the demand side and supply side of the market in more detail.
Figure 1.1 provides a diagrammatic representation of the league. On the
right-hand side is a representation of the demand and supply of the league's
fixtures as a whole. On the left-hand side is a representation of the costs and
revenues facing an individual club within the league. As clubs are identical,
this club is representative of all clubs. Prices and costs are measured on the
vertical axis, and the number of club fixtures and the number of fixtures in the
league as a whole are measured on the separate horizontal axes (Figure 1.1).
The full underpinnings of the demand side of the market are discussed in
Chapter 3, consequently it is practical at this juncture simply to note that, for
the market as a whole, i.e., the sports league, it is assumed that a rise or fall in
ticket prices, will, in the absence of changes in other factors, reduce or increase
the demand for sports matches. The demand curve represents planned pur-
chases of tickets on the market for various prices. The inverse relationship
implied by the curve is often referred to as the “law” of demand. The supply
curve in a perfectly competitive market is the sum of the marginal costs (MC)
of the many, identical, clubs that supply the market. Marginal costs are the
2
The astute reader will note that, unlike manufactured goods, or say, gymnasium services
offered by a sports facility, sports events need sporting competitors. This unique element of
competitive sports is discussed at great length in Chapter 7.
FIGURE 1.1 A perfect competitive sports league.

Key: P = price; MR = marginal revenue; AR = average revenue; MC = marginal cost;
AC = average cost; q* = number of fixtures supplied by a club; Q* = number of fixtures in
the league.
extra costs incurred from producing an extra item of output. In this context
these will be the costs incurred from producing an extra fixture.
In this price-taking market, as discussed above, the coincidence of market
demand and supply determines the market ticket price (P) and quantity (Q*).
Here the intersection of these curves shows the mutually consistent set of
plans for consumers to buy tickets to view fixtures and of clubs to supply
fixtures in the aggregate. If demand and supply in the market set the price of
the product, individual firms have to supply the product at this price. Conse-
quently, the club's average ticket price for fixtures (average revenue, AR) and
ticket price for any extra fixture supplied, marginal revenue (MR) are equal to
the market ticket price (P). The intersection of the demand and supply curves
also determines the number of fixtures in the league Q*.
Reflection Question1.1
How many fixtures should each club provide to the league total?
Hint: Think about the assumed objective of clubs.
The theory of perfect competition assumes that clubs wish to maximize

profits. The maximum contribution to the firm's profit occurs where MC is
equal to price or MR.3 If the firm were to increase its output beyond q*, then
3
This is a necessary condition for profit maximization. It is not a sufficient condition
because one has to consider what happens to fixed costs.
the added cost, as indicated by the MC curve, exceeds the revenue received
from the additional sale at market price P. Likewise, if the firm reduced its
fixtures below q*, then contribution to profit is missed, as P = MR > MC,
which implies that profits are not maximized. Because the demand curve is
given to the firm by the market, it should be clear that the MC curve is
essential to understanding how much output competitive firms, and hence
the market, supply.
What, then, are the extra costs that a firm incurs as it increases output by
one unit? In economic terms, costs are broadly classified into fixed and vari-
able costs. Examples of fixed costs would be the debt payments on investments
for a new stadium. These debts have to be paid regardless of whether or not the
team actively competed in their sport. Variable costs measure the costs that
can be varied as the club commits resources to produce fixtures; consequently,
they only need to be paid when varying the level of output of the firm, which is
supplying sports fixtures.
Consider a professional sports club. Which of its costs are likely to be viewed as variable, i.e.,
would vary directly with the number of fixtures offered?
Hint: Consider who actually produces the sports fixture on the playing field!
In economic theory labour costs are usually assumed to be the variable

input to a firm's production in the short run, and clearly this is applicable to
sports, if one considers who is actually required to produce a sports fixture and
who would need to be paid to do so. The implication is that the firm can more
easily adjust labour than capital. Labour markets are extremely important to
the economics of professional sports, and are discussed more fully in Chapter
11. Intuitively, however, it is common to observe players being transfer-listed
or sold when clubs face financial problems. In contrast, clubs tend to relin-
quish their stadium only under extreme financial pressure – when they effec-
tively close. Of course, clubs also often build new stadia, to accommodate new
capacity or to relocate to new cities, which is common in the United States
(US). This issue is discussed more fully in Chapter 12. MC for clubs can be
defined as in Equation 1.1:
Wage rate of players

Marginal cost ð1:1Þ
Marginal product of players
This relationship implies that the only reason that MC can rise for a given
money wage rate is because the marginal product of labour (MPL) falls, as
BOX 1.2 WHAT IS MARGINAL COST?

In a competitive market, output such as fixtures can be sold fixture is e10 and the player performs over the whole fixture,
at a specific price, this is marginal revenue. Each fixture also then the marginal cost is e10. If the player only played for
has a cost, its marginal cost. Equation 1.1 defines marginal one-half of the fixture, perhaps because they were viewed
cost as the money cost of producing the fixture, the wage as underperforming, then the marginal cost of the player
rate, divided by the marginal productivity of the player pro- would be e20. This higher marginal cost reflects the lower
ducing the fixture. For example, if the wage rate that has to productivity of the player.
be paid to players to encourage them to be available for a
illustrated in Box 1.2. In the theory of competitive markets this is assumed to

apply, reflecting a “law” of “diminishing marginal productivity.” This implies
that, in putting on more and more fixtures, clubs end up having to make use of
less and less productive, i.e., skilful, players as the more skilful players are
already being used, or need to be rested at times. This assumption is needed in
the theory because without it there is no guarantee that MC would rise to equal
price, which is constant, and hence the profit maximizing level of output for
the firm identified where MR = MC at q*.
Does this imply that teams could be of any size?
Hint: What is assumed about resource availability?
In principle, this means that the size of the team can vary according to
profits. One could think of this as connected with variable squad sizes rather
than actual players on the field. Better ways of viewing the employment of
players in producing fixtures are discussed more fully in Chapter 9.
An important point to note from the above analysis is that actual profit for
the club per fixture is indicated by the difference between the AR and AC
curves. AC refers to both fixed and variable costs, and therefore describes
the full, total cost per fixture. Note that, unlike with revenues, MC and AC
are not the same. The former is the cost of producing the last fixture. The latter
is the cost of producing each fixture on average. Consequently, as implied on
the diagram, when MC is less that AC, AC falls. Conversely, when MC is
greater than AC, AC rises. It follows that MC = AC at the lowest point of the
AC curve. In Figure 1.1 the horizontal MR = AR curve was tangent to the AC
curve at the point where MR = MC. This suggests that zero economic profit is
earned. Technically this means that no supernormal profit is earned. In con-
trast, only the profit implied in the cost of capital, as incorporated in the AC
curve, is received by the club. An important distinction to note here is that

accountants refer to this as “profit,” in economics it is viewed as the cost of
keeping funds in the current line of economic activity. If this cost was not met,
i.e., payment to the owner or investor in the club, they would invest their
capital elsewhere. As discussed further in Chapter 12, economic costs reflect
the opportunity costs of alternative uses of resources. It is clear from the above
diagram that profit maximization in perfect competition ensures that costs are
at their lowest possible level.
Predictions can be derived from the theory. Assume that interest in the
sport rises from current levels. This can be illustrated on Figure 1.2 by drawing
a new demand curve to the right of the existing demand curve, to show that
spectators plan to view more fixtures for any given ticket price. This could
occur if current spectators all wanted to see more fixtures or new spectators
entered the market, or both. There is an important methodological point that
needs to be noted at this juncture. In economics the effect of this change is
considered ceteris paribus, that is holding other factors constant. This is not to
suggest that economists view this as a description of the world as it operates.
However, the focus of economic analysis is to examine the logical conse-
quences of a change in one variable on another variable in isolation. This is
a direct consequence of trying to establish causal links between variables as
specified by various theories. It is these conditional or partial predictions that
form the focus of testing in economics and why, in this particular case, a new
demand curve is drawn as the only initial change taking place.
FIGURE 1.2 Increased demand for sports fixtures.

In the current context, Figure 1.2 illustrates that following the increased
demand the consequent prediction is that the market initially establishes a
new higher ticket price, P**. For individual clubs this implies supernormal
profits per fixture, as indicated by the fact that P** = AR > AC where profit
maximizing behaviour predicts an increased number of fixtures at q**, where
MR = MC in the short run. However, the assumption of free entry and exit to
the market implies that new clubs will enter the league in the long run. In
other words, capital can now vary and gets invested in this profitable industry.
This suggests that the supply curve will now also move to the right, indicating
that more fixtures can be supplied at any price. The consequent prediction is
that ticket prices will fall.
To what level will ticket prices fall?
Hint: Remember each club is assumed to be identical.
Because clubs are considered to be identical, the fall in prices will imply that
supernormal profits will disappear and individual clubs will supply their orig-
inal number of fixtures. The difference now is that more clubs will be in the
league, which has expanded to Q** fixtures overall. The “dynamic” process
implied in the model is that prices act as a signal of profit opportunities for
clubs to meet the utility or demands of fans. The league or market thus adjusts
resources to ensure that more clubs emerge to meet this need.
1.4.2 Normative economics

Perfect competition also contains value judgements that are linked to eco-
nomic policy prescription. The central normative proposition of the perfectly
competitive model is that it represents the efficient allocation of resources.4
Efficiency is concerned with the direction of scarce resources towards their best
use, which produce maximum economic value or welfare. The current ap-
proach towards understanding this concept was introduced to economic dis-
cussion by Pareto (1906). The first and second fundamental theorems of
welfare economics state that not only will a perfectly competitive equilibrium
yield an optimal allocation of resources, for any given distribution of income,
4
Other normative propositions are that decisions should be understood as deriving from
individuals, and that the general value or welfare experienced by society is, in essence, a
simple aggregation of the utility of individuals. This reflects the utilitarian ethical tradition
of economics.
but also that every optimum allocation of resources is a perfectly competitive

equilibrium. The precise demonstration of these theorems is beyond the scope
of this book but, intuitively, it can be seen that the model discussed above
involved scarce resources, because costs were assumed in production. More-
over, in the analysis of Figure 1.2, resources were attracted into the sports
leagues in response to consumer demand. The consumer demand acted like a
vote that was effective, because of the exercise of spending money on sports
fixtures. How is this efficient though? To answer this question we can review
two types of efficiency; productive and allocative efficiency. Each of these,
although integrally connected, is best understood by considering the two sides
of the market: supply and demand.
Production efficiency is connected to the idea of a production function,
which conceptually describes how resources are employed in production or
supply. Equation 1.2 describes a production function in which output “Y” is
postulated to flow from the use of three factors of production: land (Ld); labour
(L); and capital (K), terms that were referred to earlier. In economics all
resources are associated with these generic categories, which are used to
describe resources as inputs to economic activity.
Y ¼ YðLd; L; KÞ ð1:2Þ
A function is a shorthand description that indicates that a dependent

variable, an object of analysis whose value can vary, is related to a set of
independent variables, objects whose values can also vary. Equation 1.2 is a
general function because no specific relationship is identified. In the case of
sports fixtures above, we can say that the number of fixtures produced by a club
will result from the employment of players as labour, a stadium as land and
equipment, such as strips, and the relevant ball, etc., as capital.5 But, of course,
in this general sense nothing is said about in what way and by how much. The
role of such functions is central in economic analysis and they are referred to in
the appendix to this chapter, and throughout the book. Production efficiency
then implies generating the maximum number of fixtures from the minimum
amount of each resource.
Box 1.2 alluded to the concept of productivity. We can now understand this
term more thoroughly by noting that marginal productivity is the additional
output following the employment of an additional unit of a variable factor of
production, or resource, combined with other factors of production, or
5
At this point the astute reader should recall footnote 1, in which it was indicated that the
presence of an opposing team is also required to produce a competitive sports fixture. For the
moment, accept the proposition that all players fall under the remit of labour.
resources, being fixed. Average productivity relates the total level of output
produced to the total level of resource inputs. In the context of the example of
sports fixtures above, the marginal productivity of a player is essentially the
amount of the fixture that they participate in every time that they are included
in the team to produce the fixture (to play in the particular stadium, etc.).
Average productivity is thus the total number of fixtures divided by the total
number of players employed to produce them. Maximum productivity would
thus imply that all players participated fully in every match and that players
were selected in order of their abilities, and that in perfect competition the size
of the team can be variable.6
The reason for this is connected to the second aspect of productive effi-
ciency, and this is associated with “economy.” The theory of perfect competi-
tion implies that clubs are price takers in the market for fixtures. Implicit in
this discussion is that the purchase and sale of players and other resources is
also facilitated on a perfectly competitive market. The wages of players are also
given to the firm, as are the costs of land and capital, from their respective
factor markets. As a result of this, the costs facing firms and the productivity of
resources used by the firm are directly related. Indeed, costs are determined by
the behaviour of productivity, as illustrated in Box 1.2. Maximum productiv-
ity, i.e., productive efficiency, must imply minimum cost, i.e., economy, as
it is clear that the pressure of competition in the market will ensure that
profit maximizing firms seek to minimize their costs by being as efficient as
possible in production. As Figures 1.1 and 1.2 illustrate, under perfect com-
petition the output of the club is determined at the lowest point on the AC
curve. It can now be appreciated that this lowest point is determined by
productivity being at the highest, so there are no better options for the em-
ployment of resources.
6
As the astute reader should now be beginning to appreciate, there is a degree of difficulty of
interpretation of these basic economic concepts in sports. These issues are discussed much
more thoroughly in Chapters 7 and 11. This is because the nature of output is rather elusive.
In the discussion above it is assumed that a fixture of appropriate duration is the relevant
output of a sports club. Casual observation tells us that many professional athletes
participate in most fixtures, and that clubs are not interested in producing fixtures per se,
as these have been agreed in setting the scale of the league and number of times each
team play. In this respect, teams may be interested in making profit and/or maximizing
their relative performance in the league, recognizing that player performances and
qualities can vary. Identifying and measuring this performance is difficult. In the case of
the physical production of goods, however, one might more easily observe the same
hours' work producing more or less units of output depending on effort, monitoring and
incentives, etc.
An important corollary of the above discussion concerns the markets for

factors of production, such as labour. Equation 1.1 indicated that profit max-
imizing output for a perfectly competitive sports club would be determined
where marginal costs were equal to the wage rate divided by the marginal
product of labour. Recalling that in a profit maximizing club, in perfect com-
petition, MR equals MC, and that MR also equals P, then Equation 1.1 can be
rewritten as Equation 1.3.
Wage rate of players

Price ð1:3Þ
Marginal product of players
Multiplying both sides of this relationship by the marginal product of

players per fixture gives Equation 1.4:
Price Marginal product of players Wage rate of players ð1:4Þ
or:
Marginal revenue product of players Wage rate of players
The marginal revenue product (MRP) of players is their contribution, in

producing fixtures, to the club's revenues, as the fixtures are sold at their ticket
price. Perfect competition suggests that player's wage rates should be equal to
this contribution, which could then be viewed as a “just wage.” This relation-
ship is very important to the discussions in Chapter 11.
In contrast to the above conditions that suggest necessary, if not sufficient,
conditions for the best use of scarce resources, because the focus is on the
supply decision only, allocative efficiency is directly concerned with both
demand and supply conditions. It occurs when prices reflect the true costs
of production and benefits to consumers. The above discussion implies the
former. So, also, while Chapter 3 covers the detail of the consumer decision, a
utility maximizing rational consumer might be expected only to pay a price
that represents their true individual subjective valuation, i.e., utility, placed on
the service or product consumed, that is the fixture whose ticket is bought.
However, care should be taken to note that it is logically possible that the prices
paid by consumers, or received by clubs, do not reflect the true costs and
benefits involved to society. This is discussed further below under the heading
of market failure. Consequently, additional assumptions have to be met before
one can guarantee that perfect competition is, indeed, an optimal allocation of
resources.
To summarize this section, however, the main point is that a perfectly
competitive allocation of resources in sports, and generally, is typically
identified by economists as the most desirable situation, as it maximizes

economic welfare. The core of conventional economic policy thinking, there-
fore, despite many conceptual concerns with the theory of perfect competition,
is that policies that move towards the allocation of resources by a competitive
market system are appropriate.
1.5 MARKET FAILURE AND THE RATIONALE FOR POLICY

INTERVENTION
The desirability of a market allocation of resources in sport versus an alterna-
tive perspective in which active intervention is promoted in sports-related
policy can be understood more clearly by reviewing what are traditionally
referred to as market failures; these are defined as such because they stand
in counterpoint to perfect competition. These failures, to a greater or lesser
extent, relax the assumptions of the perfectly competitive model and suggest
more realism. As a consequence, different predictions of behaviour become
apparent.7 This discussion also reveals that it may be conceivable that the
economic logic of a competitive allocation of resources lies in direct conflict
with the objectives of particular stakeholders in sport.
1.5.1 Monopoly
To begin discussion, perfect competition in markets has a polar opposite of
monopoly, which, in the extreme theoretical sense, is when only one firm
supplies the market. A league could be viewed as a monopoly. Theoretically
this would mean, in comparison with perfect competition, that the supply
curve of the market would represent the MC of the monopoly, and the equiv-
alent sum of ACs would represent the monopoly league's AC. Likewise, the
market demand curve would represent the monopolist's AR. Unlike the de-
mand curve facing a perfectly competitive firm, this is not now horizontal. In
contrast to the perfectly competitive club, in which AR and MR are equal,
because the club cannot influence the ticket price on the market, the monop-
oly league can adjust its sales of tickets for fixtures by altering ticket prices.
It faces the total market demand curve and is, therefore, a price maker.
7
This generates something of a methodological conundrum in economics. By producing
more realistic assumptions on a case-by-case basis, economic predictions become fully
contingent. Consequently, tests of predictions do not necessarily refute the model, but could
also reflect the inappropriate application of the model. Of course, analysis outside of
experimental conditions and any use of statistical analysis cannot provide definitive tests.
Market Failure and the Rationale for Policy Intervention 15
To increase ticket sales the monopoly league can reduce ticket prices and
vice versa.
What will the monopoly league's marginal revenue curve look like?
Hint: Remember the discussion of average and marginal cost.
Note that if, in order to increase ticket sales, additional tickets are offered at
a lower price than before, the MR earned by the league on these additional
tickets will be lower than the AR earned on previous ticket sales. Consequent-
ly, in general, MR will be less than AR for a monopoly or, indeed, any other
form of market that is not perfectly competitive and where firms can influence
the market price. Appendix 1.1 illustrates that, with a straight line AR “curve,”
MR will be twice as steep. Figure 1.3 illustrates the situation of a monopoly
league, using the same notation as previously.
Once again ticket prices and costs of production are measured on the
vertical axis and the number of fixtures in the league on the horizontal axis.
Profit maximizing behaviour implies that the number of fixtures supplied will
be Q* where MR is equal to MC. Because the monopoly firm faces the market
demand curve, a ticket price of P* can be set for each ticket, as this represents
what consumers are prepared to pay to see Q* fixtures. An important impli-
cation of monopoly supply is that even in the longer run, because of the lack of
competition and inability of clubs to enter the league or rival leagues to be set
up to supply the sport, supernormal profits can be earned. This is because, on
each ticket sold, there is a mark-up above average costs (which, remember,
includes normal profit) of P* AC*.
From the normative perspective discussed earlier, this is viewed as prob-
lematic for economic welfare. Productive efficiency is not maximized, because
the number of fixtures produced does not reflect the lowest AC. Even if de-
mand fell, because of a reduction of spectator interest for the sport which could
be indicated by a movement in the demand curve to the left such that ticket
price mark-ups were squeezed to zero, and just normal profits were earned by
the league, P = AR = AC would have to take place, in the diagram, to the left
of the lowest possible value of AC, where the demand curve was at a tangent to
the AC curve.8
8
It has also been argued that because monopolies do not face competition they become
“slack” organizationally, i.e., exhibit x-inefficiency, consequently productive efficiency falls
and costs rise (Leibenstein, 1966). The term “x-inefficiency” suggests that the inefficiency is
neither due to productive nor allocative inefficiencies.
Moreover, this suggests that the number of

tickets sold will always be less than would have
been the case under a perfectly competitive league.
This produces an allocative inefficiency, which
represents a deadweight loss of economic welfare.
This can be illustrated with reference to
Figure 1.4. Here, the profit maximizing ticket
price and number of fixtures of the monopoly lea-
gue, given by Q* and P*, are reproduced. However,
these are compared to the profit maximizing tick-
et price and number of fixtures of a perfectly com-
petitive organized league, given by Pc and Qc,
FIGURE 1.3 Monopoly supply of sports fixtures.
respectively.
What will be the profit maximizing price and output of a perfectly competitive sports league?
Hint: Remember that the sports league represents the market as a whole for perfectly
competitive sports clubs.
Recall from Figure 1.1 that, in a price-taking market, demand and supply
set the market level ticket price and number of fixtures, and this is consistent
with the profit maximizing behaviour of individual clubs. If it is assumed that
the monopoly and perfectly competitive leagues
face the same demand curve, and that the pro-
duction functions are the same in the aggregate,
then from the monopoly diagram the perfectly
competitive equilibrium would be where de-
mand equals supply (the MC curve).
Figure 1.4 illustrates that the monopoly
price of tickets, P*, is higher than the perfectly
competitive price of tickets, Pc, and the number
of fixtures supplied on the monopoly league,
Q*, is less than would be supplied in a perfectly
competitive league, Qc. The welfare cost that
results from having a monopoly league is indi-
cated by the triangular area xyz. What is this
welfare cost?
FIGURE 1.4 Monopoly versus perfectly competitive The total area xyz comprises two rectangles,
leagues. wxy and wyz. The first of these refers to the lost
consumer surplus resulting from monopoly organization of the league. The

second triangle shows the lost producer surplus. Conceptually speaking, the
first of these is defined as the amount that consumers are willing to pay for a
ticket above what they have to pay in a competitive market. Remembering that
the demand curve represents what consumers are planning to purchase, any
point on the demand curve represents willingness to pay. In a competitive
market, consumers would actually pay Pc for tickets. Consequently, triangle
wxy represents how much value is lost to consumers as a result of not having
up to Qc number of tickets to purchase at a price of Pc, and which are valued
more highly than Pc.
Likewise, producer surplus represents the difference between the price of
tickets per fixture received by clubs under a competitive market, and the cost
that clubs would incur but be willing to supply fixtures at. The supply curve
represents the cost of producing each fixture. Consequently, by restricting the
number of fixtures that could have been sold for ticket price Pc, a monopoly
league loses the economic value of wyz, a series of fixtures for which Pc is
greater than the cost of producing them.
It is important to indicate that these losses are viewed in terms of aggregate
economic welfare. This is implied in comparing the alternative league out-
comes of different economic organizations, or market structures. By construc-
tion, the perfectly competitive model represents the sum of values received by
individual consumers and individual clubs. Yet each clubs share of, say, the
forgone producer surplus, is by assumption miniscule. In contrast, the specific
monopolist can make substantial extra profits by restricting output. A com-
mercial organization may have objectives that do not correspond to those
of maximizing the welfare of the broader economy more generally. This
implies that there is always an economic incentive for organizations to look
to establish monopoly power. It is for this reason that economic regulatory
agencies often intervene in markets to reduce the monopoly power of specific
organizations.
For example, they might look to tax away the supernormal profits in a lump
sum form of the value (P* AC*) Q* in Figure 1.3. Such a policy would not
affect the costs of production or productivity, and would thus be efficient. This
policy would not, however, eradicate the deadweight loss. Alternatively, they
might seek to force a change in the organization of the market, to encourage
competition, or to enforce a greater supply of output, which will force price to
fall. There are also other arguments for intervention in markets. However,
these do not, of themselves, provide a rationale for trying to move towards a
market allocation of resource, but in contrast, to rectify the problems of
market allocation.
1.5.2 Equity
While the model of perfect competition offers an efficient use of resources,
which can be identified with maximum social welfare, an implicit assumption
in the model is access to the income required by consumers to express a
demand for sports league tickets through purchases. The logic of the model
is that economic agents can only be consumers because they have supplied
their labour on the labour market to earn the requisite income to spend. The
income earned, moreover, reflects their productivity. Likewise, those who own
capital or land will only invest it in particular economic activities if the
productivity of the investment is sufficiently high. Consequently, economic
efficiency does not recognize the need to explore issues of the inequitable
distribution of resources and the fact that those with higher incomes can
demand more goods and services than others. A moment's reflection suggests
examples where these issues may matter as indicated in the questions raised
in Box 1.3.
Essentially answers to these questions hinge on the view that is taken of an
unequal distribution of income, and if this biases outcomes in the market
allocation of resources in favour of particular groups. In general, equity can be
both understood, and addressed in policy, through vertical means, i.e., the
“unequal treatment of unequals,” or horizontally, through the “equal treat-
ment of equals.”
In the former case a sports club or league might have lower ticket prices for
the unemployed, elderly or student fans, but charge a premium price for
BOX 1.3 DOES EQUITY MATTER IN SPORT?

Participation sport 6. Are ticket prices too high for premiership soccer
1. Should local authorities subsidize/provide access to matches?
sports and leisure facilities? 7. Should clubs share their gate revenues with smaller
2. Should students pay the same gym fees as university clubs?
professors? 8. Should local authorities subsidize professional team
3. Should local authorities invest in a youth development sports?
funds? 9. Should the government subsidize facilities for profes-
4. Should the national government invest in sport facilities sional team sports?
and participation programs in the short-term to reduce
Mega events
social security costs in the long-term?
10. Should all tax payers contribute towards the costs of
hosting the Olympic Games?
Professional team sports 11. Should lottery funds only be directed towards athlete's
5. Are professional soccer players paid too much relative development for those with a chance of winning gold
to nurses? medals?
executive boxes. This is an example of a policy of addressing vertical equity.

One can argue that all are given opportunities to watch fixtures and are not
discriminated against by income in this policy. Moreover, this (more) equal
treatment of fans of a similar nature is an example of policy seeking horizontal
equity. Consequently all students would pay the same ticket price according to
the policy, as would all unemployed or elderly fans.
These policies may make economic sense for a profit oriented sports club,
as discussed in Chapter 5, if these sets of fans represent distinct market
segments. A policy dilemma only arises if all seats in a stadium could be sold
at the executive price, then clearly equity and efficiency come into conflict as
policies if one holds the value judgement that professional sport should be
accessible to all fans. Rationing the allocation of tickets by price only would
necessarily favour wealthy fans.
1.5.3 Externalities
A key assumption of the perfectly competitive model is that consumers and
producers have property rights over the purchases and sales that they make.
Clearly defined property rights, which are the legal rules that describe what
people can do with their property, are essential for markets to be allocatively
efficient as a form of voluntary exchange that takes place in the interests of
both consumers and suppliers.
However, when property rights are not clearly defined, then markets can
fail to produce an efficient allocation of resources. Externalities arise and can
mean that the private benefits and costs received by or paid to a consumer or a
supplier does not correspond to those of society. There will be spillover effects
to third parties not formally engaged in the transaction. Positive externalities
arise when the social benefits of consumption exceed those of the private
individual, or the social costs of production are less than the private costs of
production. Negative externalities occur in the opposite case. Here, the private
benefits of consumption exceed the social benefits, or the social costs of
production are greater than the private benefits. Box 1.4 gives some examples
of externalities in sport.
What are the consequences of externalities? Positive externalities imply
that the market undersupplies the amount of sport. Consequently, as with a
monopoly, society experiences a welfare loss in the form of an opportunity
cost. In contrast, negative externalities imply that the market oversupplies
sport. Again, there is a welfare loss.
These two cases are illustrated below in Figure 1.5 for demand-based
externalities associated with a sports league. Here, a competitive allocation
of tickets is given by price Pc, and number of tickets, Qc, determined by the
BOX 1.4 EXTERNALITIES ASSOCIATED WITH SPORT

Positive externalities: Negative externalities:
& Demand: fans that pay to watch professional sports fix- & Demand: fans that pay to watch professional sports fix-
tures are subsequently enticed to volunteer to provide tures act as hooligans and vandalize localities.
sports opportunities for others. & Supply: the growth of televised sport encourages
& Supply: successful sports teams or events raise morale “armchair” spectating, and reduces participation and
and encourage productivity in work efforts elsewhere. health benefits.
intersection of the marginal private benefit (MPB) of fans and the marginal
social cost (MSC) of their production. The presence of a demand externality
means there is a true marginal social benefit (MSB), i.e., value, of the tickets
that differs from the value associated with the actual purchases of tickets. This
is because of the spillover effects of demand that are not actually accounted for
in the market price. With supply-based externalities marginal social costs
(MSC) would differ from marginal private costs (MPC).
In the case of positive demand externalities, the diagram on the left illus-
trates that the market under-provides tickets, as Qc < Qs, the competitive
allocation of tickets is less than the socially optimal allocation of tickets. This
is reflected in the market expressing a lower value than should be the case, as
indicated by Pc < Ps, i.e., the actual price of tickets is lower than the socially
optimal price. In the case of negative externalities, the opposite is true and the
market over-provides tickets, as Qc > Qs and consequently Pc > Ps.
The presence of externalities provides a further direct rationale for active
policy in the allocation of resources in sport. In the examples above this
FIGURE 1.5 Positive and negative externalities.

implies providing a means of either increasing or decreasing demand to its

socially optimal level. This reallocation of resources would involve internal-
izing the externality.
Increasing demand theoretically could involve providing direct benefits,
such as grants to consumers to increase their demand. This is not an area of
policy activity, however, in sport. Marketing and information campaigns may
attempt to achieve these objectives, and recruitment campaigns for volun-
teers, for example, appeal to an underlying or latent demand. In the latter case
restricting access to fixtures for hooligans at stadiums or restrictions on trav-
elling to fixtures would reduce demand appropriately.
More generally, policy agencies can use taxes or subsidies to control
externalities that emerge from the supply-side of the sports economy. Taxes
act as an additional cost of production and thus shift the supply curve to the
left, indicating that for any given price less matches, i.e., tickets, can be
produced or equivalently, for any given level of tickets, the price will be
higher. In contrast, subsidies act as a reduction in costs and shift the supply
curve to the right. Unlike the case of monopoly such taxes or subsidies will
not be lump sum but, rather, reflect either a fixed amount added or sub-
tracted from each sale, known as a specific tax, or a fixed percentage of the
value being added or subtracted to the proposed selling price, known as an ad
valorem tax. Qualitatively the effects are the same. Of course, the actual
costs or revenues that are imposed or generated from subsidies or taxes will
vary, depending on the fiscal regime and the specific behaviour characteriz-
ing the sports market.
1.5.4 Public goods

If externalities are extremely severe, this implies that property rights cannot be
allocated between those engaging in market transactions and those that do
not. This leads to the free-rider problem, and means that markets fail
completely in the presence of public goods. Public goods are both non-rival
and non-exclusive. It is possible that goods and services may have one or both
of these characteristics. The former characteristic implies that MC = 0. This
might apply in sports contexts when an event or fixture takes place in a
stadium in which there are spare seats. The costs of providing the event or
fixture are not affected by the presence of additional consumers. In the exam-
ples of the competitive market above, it was assumed that spectators were
rival, i.e., resources such as tickets were scarce. If one fan bought a ticket for a
game, another could not see the same game. This is implied in the equaliza-
tion of the demand and supply schedules in which all consumer and supplier
plans were harmonized. Consequently, the competitive model assumes that
fixtures are always sold out, a corollary being that clubs can easily adjust their
stadium capacity to meet demand.
Non-exclusivity applies if consumers cannot be prevented from consuming
the good, which they can be in the example above. One can speculate that early
historic examples of sport, such as the early codes of football, were played on
land which was open access. Under such circumstances, in principle, all fans
that could physically fit onto the land could enjoy the spectacle. Of course,
stadiums and dedicated sports facilities make sports potentially exclusive to
avoid this scenario and to extract economic value from spectators. In the
absence of such possibilities a direct rationale for policy makers to provide
goods and services exists.
1.5.5 Imperfect information

Another form of market failure is connected with economic agents not
possessing perfect information. There are degrees of lack of information
discussed in economic analysis. Chapter 3 reviews some approaches that
maintain that economic agents simply cannot optimize. Consequently, for
example, they reject the idea that consumers maximize their utility and
argue instead that agents exhibit “bounded rationality” and, as a conse-
quence adopt rules of thumb, habits or decision heuristics to make decisions.
There is a clear desire to conceptualize the processes by which decisions are
made. This approach has its roots in psychology which is represented as
relevant for understanding sports participation, for example from Sport Eng-
land, in the UK.
A less radical approach that is often used to analyze employment or finan-
cial arrangements in economics, and has been used to understand aspects of
professional sports, assumes that economic agents possess asymmetric infor-
mation.
Information asymmetry occurs in transactions when one party possesses more or
better information than the other party.
In what contexts might there be information asymmetry in sports transactions?
Hint: Think about the relationship between various stakeholders.
In the absence of being able to correct the information deficiency, for

example, by monitoring, there are two main problems that can arise in
economic transactions. Moral hazard might occur when agents behave
differently than expected after a transaction has taken place. In contrast

adverse selection may occur prior to a transaction taking place. The impli-
cation of these effects would be to promote market failure, because the
quality of transactions falls.
Typically, and in professional sports, information asymmetry is impor-
tant in a principal–agent context, where the principal may represent an
employer, and the agent an employee or someone contracted to perform a
function on behalf of the principal. Professional soccer players are thus
“agents” to clubs, which are “principals.” The club signs the player on a
contract prior to their performance for the team. The player knows their
own ability best and likely effort levels best, and thus the club is potentially
at a disadvantage and, contrary to the result implied in Equation 1.4, pay the
player more than they become worth in terms of their productivity. Likewise
team managers are also agents, while club owners are principals. Managers
might seek to target successes, while owners target profits, particularly if the
team is publicly owned. While these may not be mutually exclusive objec-
tives, there are still differences, as discussed in Chapter 7, which has impli-
cations for understanding the impact of policies on league management.
Moreover, league and tournament organizers essentially contract teams and
competitors to deliver certain outputs, a certain level of quality competition
that the paying public wishes or expects to see. Such organizers need to
ensure that the appropriate levels of competitive effort are elicited. Chapters
7 and 12 discuss these issues in some detail.
A variety of solutions to the problems of asymmetric information have
been offered. Significantly, none of these, of necessity, requires the use of
replacing the market with public policy agencies. For example, warranties
act as signalling devices declaring the supply of appropriate quality goods
and services. Screening devices allow the agent to choose between options in
making a transaction. Product bundling and packages of tariffs, for example,
in television or telecommunications are examples of screening. The nature
of the customer is revealed and an appropriate “discriminatory” price is
offered.
Screening is particularly prevalent in employment contracts. Instead
of just receiving a fixed salary, incentive contracts, rewarding higher effort
and performance, can be designed and will be attractive to those confident
of their talents. Most professional sports have offered forms of incentive
contracts to players, with win bonuses being an obvious example. The
remuneration of tournaments, moreover, is predicated on providing
enough high prizes to attract the best talent, while not undermining the
incentives to participate by confining rewards for only the very best or
winner.
Tournaments, as discussed in Chapter 7, are a form of incentive contract that is

hierarchically structured, and in which performance is relative rather than absolute.
Naturally, this coincides with a conceptual description of sporting contests, whether
organized on a league or knockout basis.
However, there must be no incentive to focus only on the performance mea-

sure being directly rewarded. One can easily hypothesize that an incentive
scheme rewarding the number of goals scored by a particular soccer player may
promote undue selfish efforts, rather than concern with team performance,
which might have been enhanced by other players' better scoring opportunities
being continually overlooked. The general economic missive is that complex
jobs should not be explicitly rewarded through such contracts. In team sports,
then, team-based bonuses may also be appropriate.
1.6 LIMITATIONS TO POLICY MAKER INTERVENTION

With the exception of imperfect information, the above discussion of market
failures arising from monopoly, equity, externalities and public goods are often
presented as providing a direct rationale for policy maker intervention in the
allocation of resources. However, there are also theoretical limitations to this
possibility. In this regard there is no necessary presumption in favour of policy
intervention.
1.6.1 The Coase theorem

As well as policy maker intervention, it is possible that when an externality
affects few parties and if property rights are well-specified, economic efficiency
can actually arise from the market through bargaining. In other words, gains
from trade by internalizing the externality force a solution through the emer-
gence of market trade. This proposition is the Coase theorem, which was
developed by Coase (1960). There are varieties of the theorem in existence,
but essentially it suggests that in the above circumstances the output and
input configurations of resources will be independent of the allocation of
property rights in the market. The Coase theorem has been particularly influ-
ential in the analysis of professional sports leagues in which, for example, the
distribution of performances by teams, driven of course by the teams accruing
playing talent resources i.e. labour, is independent of the contractual regime in
the players' labour market.
1.6.2 Government failure

There are many reasons why government, or more broadly speaking, policy
making intervention in sports may also fail. Naturally, these circumstances
Expenditure-Based Rationale for Policy 25
undermine assumptions that non-market policy solutions to resource alloca-

tion in sport are automatically relevant. Philosophically speaking, one can
argue that elected government representatives may well have been chosen
initially to reflect the preferences and desires of the electorate or consumers.
However, over time electoral systems and political representation are not
flexible enough relative to markets to adjust to competing demands being
made on resources. Policy makers may also begin to pursue their own inter-
ests, as opposed to those of the electorate, which is known as regulatory
capture. This could shift focus onto the short-term benefits of policies, as
opposed to their longer-term legacies.
It is also possible that, over time, taxes and subsidies fundamentally distort
resource allocation and may impose their own externalities on society. For
example, taxes reduce consumer incomes which may reduce or crowd out the
social optimum levels of consumption of other goods and services, despite
apparently correcting a particular externality. Taxes may also provide disin-
centives to effort, which reduce productivity and raise costs. Of particular
relevance to sport is the role of the voluntary sector, as discussed in Chapter
6. Inasmuch as both markets and governments do not meet the needs of
resource allocation in sport that is provided by voluntary organization, this
is an example of government, as well as market, failure.
1.7 EXPENDITURE-BASED RATIONALE FOR POLICY

The above discussion has hinged on the choice of having either active policy
or market allocations of resources in sport, in the context of a typical
concern for efficiency, i.e., an optimal allocation of resources. There has
been, however, another historical reason why policy agencies have inter-
vened in economic decision-making. Unemployment can occur in situa-
tions when traditional industries are in decline, reflecting long-term
structural economic changes stemming from technological change, funda-
mental shifts in consumer tastes and changing global or political economic
conditions. Sport has been increasingly viewed as a means of offsetting the
implications of these changes in two major ways. The first is that the
advanced economies, such as Europe and the United States, have evolved
from being manufacturing providers to service-based economies. Sport is
part of this sector, as is discussed more in Chapters 2 and 12. Secondly,
investment in sport by policy agencies has been viewed as a means of
promoting this development. Importantly, the value of initial investment
need not be the full amount that is required to offset the structural problems
of long-term falls in income and unemployment, because of multiplier
effects. These are discussed further in Chapter 12 and Appendix 1.2.
1.8 AN EMPIRICAL FRAMEWORK

To close this chapter it is informative to outline, in broad terms, the empirical
framework employed by economists that has been applied to the economic
analysis of sports. The empirical approach is directly motivated by the meth-
odological principle that economic analysis builds theories that yield predic-
tions that can be tested against data. This section outlines the main elements
of this process.
In the above discussion, economic relationships have been presented in
equations and diagrams. The main reason for this is that they capture the
logical relationships, and hence predictions, implied in economic theory.
Diagrammatic analysis however only identifies qualitative predictions be-
tween two economic phenomena. Equations are central for the general quan-
titative analysis of these phenomena. To recap in the context of production
efficiency, Equation 1.2 examined how output flows from the employment of
resources described as land, labour and capital. There is a dependent variable
and a set of independent variables in a general function. What lies behind this
application to a particular context is a structure of functions. A specific form of
functional relationship can be written as:
Y ¼ b1 þ b2 X2 þ b3 X3 þ . . . þ bk Xk ð1:5Þ
Here, “Y” is the dependent variable and there are up to “k” independent
variables, labelled as “X” with a subscript “1” to “k” that are viewed as con-
tributing to the determination of “Y.” Theory will specify the variables. In this
general format the addition sign between each variable indicates that the
independent variables affect “Y” separately or independently of the other
variables. This is consistent with the desire to examine changes to “Y” follow-
ing a change in a particular “X” ceteris paribus, or other things being equal, as
discussed above. Convention specifies the presentation as positive, but in
practice the sign could turn out to be positive or negative, depending on the
actual behaviour being examined. The specification of independent positive or
negative relationships between the variables means that the behaviour exam-
ined is hypothesized to be of a linear nature. This linearity is implied in a very
specific sense. This is made clearer in considering the other terms in the
equation.
While Y and each of the Xs are referred to as variables, precisely because
their values can change as measurements of economic behaviour, the other “b”
terms in the equation are referred to as parameters. These are viewed as fixed
or unchanging, as they represent the structure or behavioural foundations that
govern how the variables are related to each other, and consequently how
An Empirical Framework 27
changes in one variable affect another. Each indepen-

dent variable has its own parameter associated with it,
as indicated by the common subscripts 1 to k with the
variables, although clearly there is one major differ-
ence. The term b1 does not appear to have a variable
X1 associated with it, yet each of the other bs, 1 to k,
does. There is good reason for this: b1 represents the
constant or intercept of the equation and is the value
taken by Y when each of the independent variables is
equal to zero. It provides the level of an economic
variable that is independent of the influence of other FIGURE 1.6 The intercept and slope of a linear
variables. In practice, constant or intercept parameters function.
often have little empirical meaning. This is because,
while they may refer to a scenario which is logically possible, in practice this
logical possibility is a huge extrapolation away from typical relationships
observed in data. Figure 1.6 illustrates the constant or intercept, indicating
why the latter name is pertinent. Graphically it shows where the function
intercepts the Y-axis at a value where X2 is zero.
Naturally, in a graphical analysis only two variables are examined, Y and
X2. Conveniently, however, the graph can also presents the logical nature of all
of the other parameters in the model simply by examining the parameter b2
associated with X2. b2 is described as a partial slope coefficient or parameter. It
is partial in that it examines the relationship solely between Y and X2. It is a
slope because it indicates how Y changes following a unit change in the value
of X2. As drawn, it is implied that b2 is positive. Specifically, the parameter
indicates that following a unit change of X2, Y increases by b2, ceteris paribus.
Each of the other bs would indicate the change in Y following a unit change in
the respective independent variable. Note that the graph is drawn in a straight
line. This reflects the fact that b2 is fixed. Regardless of the values of X2 from
which a change occurs, therefore, the subsequent change in Y will always be
the same. There is a hypothesized fixed relationship between changes in the
variables. It is important to note that this linear relationship applies regardless
of the units in which the variables are measured. Differences in these may
affect the economic interpretation of the relationship, but not the fact that
there is a constant degree of change implied in the relationship between the
variables.
It is the partial slope coefficients that are of direct interest to economists
because they describe how variables are related to one another. Assumptions or
hypotheses about these parameters thus form the basis of predictions that can
be tested against data. In testing these relationships, however, Equation 1.5
needs to be modified. As written, Equation 1.5 specifies a precise
mathematical relationship represented, for example, in Figure 1.6. This sug-

gests that the behaviour of Y is fully understood. In the real world this cannot
be. Theories might suggest which variables might be important in affecting Y,
but not definitively. Economic research does not operate through experimen-
tation.
An experiment engineers the isolation of intervening independent vari-
ables on a dependent variable. In a true laboratory experiment, the impact
of independent variables is, essentially, fully understood. Translating the
effects of experimental results to the real world outside of the experiment,
however, becomes reminiscent of economic research as the impact of different
variables needs to be isolated statistically. Because economic measurement
and analysis takes place in the context of actual human activity, such as sport,
assumptions have to be made about how factors other than those specified in a
particular theory affect the dependent variable Y. In economic research statis-
tical assumptions are made. To facilitate this, Equation 1.5 is rewritten as
Equation 1.6. This is referred to as an econometric model.
Yi ¼ b1 þ b2 X2i þ b3 X3i þ . . . þ bk Xki þ i ð1:6Þ
Two differences are noted. The first is that a subscript “i” is now attached to
each variable. This is referred to as the index of observation, and will represent
a particular set of observations or cases of the values of each economic variable.
Thus, i = 1,2 . . . 2000 would indicate that 2000 observations on each variable
were available to the researcher.The second difference is the final term in the
equation, “y i.” This is referred to as the stochastic or random error term of the
equation. It represents the non-systematic effects on the dependent variable Y.
Because it is a random variable, it makes Y, but not the independent variables,
a random variable. Consequently, Equation 1.6 decomposes the behaviour of Y
into a systematic component indicated by the independent variables and the
random component, which cannot be predicted. The aim of economic re-
search then becomes to use data on the variables to estimate the parameters
using regression analysis, and to compare these with theoretical hypotheses
about the sign and size of the parameters. Regression analysis helps to identify
the effects of each independent variable on the dependent variable separately.
Because the parameters are estimated from data on Y, which is random, and
the independent variables, the parameters are also random variables. In this
way statistical tests of the parameters can also be undertaken to make infer-
ences about the applicability of the parameter values beyond the sample of data
used in the analysis. Traditionally ordinary least squares (OLS) analysis has
been used. Briefly, this approach estimates the parameters so as to minimize
the sum of squared residuals in a regression, and thereby maximizes the
An Empirical Framework 29
BOX 1.5 ASSUMPTIONS ABOUT THE RANDOM ERROR TERM IN OLS

1. The mean value of yi is zero. 4. Each yi has a normal distribution.
2. The variance of y i is the same for all observations.
3. Each y i is independent of the others (i.e. their probability
distributions are independent).
“goodness of fit” of the regression. Residuals can be loosely viewed as an

estimate of the values of y i from the sample of data at the disposal of the
researcher. A statistic, R2, that lies between “0” and “1” indicates the propor-
tion of explanation achieved, with “1” implying a perfect fit with the data,
which is never achieved.
OLS assumes that y i follows a normal distribution, and as a result so do
both Yi and the parameters. Because samples are used, tests of the parameters
such as H0:bi = 0 against one of the following alternative hypotheses,
H1:bi „ 0 or H1:bi > 0 or H1:bi < 0 can be undertaken using the t-distribution,
with a t-statistic being calculated by dividing the estimated parameter by the
estimated standard error of the parameter. If the absolute value of the t-ratio is
greater than 2, i.e., |t| > 2, then H0 can be rejected at approximately a 5%
significance level. This implies that a non-zero estimated coefficient can only
have been a chance occurrence at most in 5% of occasions. OLS produces
“best-linear-unbiased estimates,” if the assumptions detailed in Box 1.5 are
met.
This means that the estimated parameters are the best possible given the
data available. The errors will be the lowest, and on average the parameter
estimates will be correct. Much economic research thus involves correcting
models to make these assumptions fit. In particular, assumptions 2 and 3 are
addressed under the headings of heteroscedasticity and serial correlation.
Solutions to these problems either involve adjusting the model (often referred
to as generalized least squares) or the standard errors used to make the statis-
tical claims.
There are an increasing number of regression estimators used in economic
research resulting from the growth in computational power available to
researchers, and increased sensitivity of researchers to more adequately cap-
ture the characteristics of the dependent variable being investigated. Examples
of other techniques noted in the literature are given in Table 1.1
Regressions can be estimated for cross-section, time series or pooled/panel
samples of data. These reflect observations across a sample of different eco-
nomic units at a point in time, the same economic unit over a set of time
periods, or a combination of these.
30
CHAPTER 1 The Economics of Sport
TABLE 1.1 Regression options
Type of estimator Measurement of dependent Typical application Coefficient interpretation

variable
1. Discrete choice
Logit/probit Binary values e.g., 1 or 0 Participation or not in an The change in the odds of the
activity by an individual possession of the characteristic
scored 1 following a unit
change in the independent
variable
Poisson/negative binomial Count data e.g., the number of The frequency of participation The change in the conditional
occurrences of an event in an activity or the number of mean number of counts
medals won by a country at the following a unit change in the
Olympic Games independent variable
2. Limited dependent variable (censored)
The Tobit model Some values of the dependent Attendance at sell out fixtures As with ordinary least squares
variable are collapsed to one used to measure the demand (OLS)
value for the fixture
The Heckman model Some values of the dependent The frequency of participation Two equations are estimated: a
variable are recorded only for in a sports activity depends on logit/probit to estimate the
sub-samples of the population the prior choice to undertake likelihood of inclusion in the
the activity sample, with coefficients
interpreted as above; and a
regression that models the
subsequent values of the
dependent variable, with
coefficients interpreted as with
OLS
The hazard/survival
model The dependent variable The length of career of The proportional change in the
measures a discrete and professional team sport duration of the hazard
varying duration managers
Conclusion 31
1.9 CONCLUSION
This chapter has sought to revisit some basic economic tenets that are relevant
to the study of sports economics. Consequently, it has been shown that the
conventional economic scope of analysis is concerned with an optimal allo-
cation of resources. The theory of perfect competition acts as a policy bench-
mark, which provides impetus to recommending the market-based provision
of sports. It has also been indicated, however, that active policy intervention in
sport can find a rationale through seeking to alleviate market failures, but also
to promote economic development. A basic sketch of the empirical approach
of economic research has also been given. Subsequent chapters employ these
concepts in a more detailed analysis of participant sport, professional sports
and sports events. The next chapter provides more detail on the definition of
sport, its measurement and significance, and the alternative policy emphases
of sports provision.
Appendix 1.1 The Relationship Between Average and Marginal Revenue

inMonopoly
A straight line AR curve can be given by Equation A1.1.1. As described in the text, this is an
example of a function in which the dependent variable, F (the number of fixtures), is related to
the independent variable, P (the average revenue or ticket price). It is a specific rather than a
general function, because the nature of the relationship between the two variables is
specified. In this case, a straight line relationship is implied in which the number of fixtures
equals “a” if the ticket price is zero, and reduces from that value as ticket prices increase
according to the value of “b” (a and b are referred to as parameters of the function). The
former is the intercept, or constant term, the latter is the slope coefficient.
F ¼ a bP ðA1:1:1Þ
Where:
P is average revenue or ticket price;
F is the number of fixtures;
a is the constant or intercept and gives the number of fixtures when price is zero. This is the
limit of demand.
b is the slope of the demand curve, and gives the number of the increase or decrease in
fixtures following a unit decrease or increase in ticket prices. Mathematically, b is the
derivative of the function F with respect to P, i.e., b = dF/dP.
MR is the change in total revenue following a change in the number of fixtures supplied. To
identify MR requires first finding total revenue (R). This can be obtained by multiplying
Equation A1.1.1 by ticket price P to give Equation A1.1.2:
F P ¼ R ¼ aP bP2 ðA1:1:2Þ
MR is then given as qR/qP, which gives Equation A1.1.3:
dR=dP ¼ MR ¼ a 2bP ðA1:1:3Þ
Note that the slope of this marginal revenue relationship is given by Equation A1.1.4:
dðdR=dPÞdP ¼ dMR=dP ¼ 2b ðA1:1:4Þ
The slope is twice as steep as the demand curve.
Appendix 1.2 The Derivation of the Keynesian ExpenditureMultiplier

Equation A1.2.1 describes consumption as a linear function of income, known as a
consumption function, and Equation A1.2.2 shows that income is determined by
consumption. In practice more complicated functions would be involved
C ¼ a þ bY ðA1:2:1Þ
Y¼C ðA1:2:2Þ
Where:
Y is income
C is consumption
b is the marginal propensity to consume (mpc, i.e., 0.8)
a is the hypothetical minimum consumption required if income was zero.
To determine the overall level of income these equations are solved simultaneously. Thus:
Y ¼ a þ bY ðA:1:2:3Þ
implying:
Yð1 bÞ ¼ a ðA1:2:4Þ
implying:
Y ¼ a=ð1 bÞ ðA1:2:5Þ
Therefore, if the public authorities increased spending by e100m this would be equivalent to
increasing a by this value, i.e. consumption that is not funded by current income. The
multiplier is calculated as:
dY=da ¼ 1=ð1 bÞ ðA1:2:6Þ

Conclusion 33
or:
dY=da ¼ 1=ð1 0:8Þ ¼ 5 ðA1:2:7Þ
Consequently, multiplying the increase in expenditure “da” by the multiplier indicates the
consequent rise in income:
ðdY=daÞ da ¼ e100=ð1 0:8Þ ¼ e500 ðA1:2:8Þ

This page intentionally left blank
CHAPTER 2
The Nature, Organization and

Economic Significance of
Sport
OBJECTIVES
& To appreciate the difficulties in defining sport as a distinct economic activity

& To understand the main features of the policy and economic environment of
sport
& To appreciate the main research themes of mass participation sport, sports
events and professional team sports
2.1 INTRODUCTION
The previous chapter provided an overview of economic analysis relevant to
the study of sports. In this chapter an overview of the sports economy is
attempted, so that the remainder of the book can provide a more detailed
theoretical and empirical analysis of specific features and issues in sports
economics.
This chapter addresses two main interrelated issues. The first is to explore
what is meant by “sport” as an object of enquiry. In the next section some brief
philosophical comments on the definition of sport are provided, before a
concrete understanding of the contingent nature of sport is illustrated by
addressing the second main theme of the chapter, which is to outline some
key economic and policy characteristics of the sports environment. This dis-
cussion makes it clear why it is important to focus on the main subjects of this
book; mass sports participation, professional team sports and sports events.
Although they share common origins and elements of organization, there are
also unique policy concerns and economic developments in each case that
need to be understood. It is through a discussion of these segments of the 35
36 CHAPTER 2 The Nature, Organization and Economic Significance of Sport
sports environment that the main sports involved in each case are presented.
The chapter closes with a brief review of the changing emphases of their policy
and economic environment to set the scene for subsequent more detailed
discussion.
2.2 WHAT IS SPORT?

In Chapter 1 it was argued that the most prevalent definition of economics was
derived from Lionel Robbins as “the science which studies human behaviour
as a relationship between ends and scarce means which have alternative uses.”
Robbins' definition is broad in conception and consistent with recognizing
that economic activity is not, of necessity, connected with the creation and
redistribution of material wealth. Consequently, it can apply directly to all
activities that could be classified as sport, such as professional team sports,
elite individual sports and, indeed, formal or casual mass participation in
other activities. This is because individuals who participate in these activities
need to make choices over their allocation of resources to supply and consume
these activities. But this raises an extremely important question which is:
what is sport? A moment's serious reflection reveals that, like many appar-
ently tractable objects, sport appears easy to define at first hand. However, the
boundaries that distinguish it from other activities are contentious and, in-
deed, can appear quite arbitrary. From a purely logical point of view, there is no
simple list of activities that can be called “sport.”
For example, at the outset of policy concerns with sports in Europe, the
Council of Europe (1980) and Rodgers (1977, 1978) developed overlapping
classifications of leisure, recreation and sport which have become accepted
policy categories, and are still implied in the 1992 Council of Europe, Euro-
pean Sports Charter (Gratton and Taylor, 1985, 2000). The 1992 European
Sports Charter argues that:
“Sport embraces much more than traditional team games and competition.
Sport means all forms of physical activity which, through casual or
organized participation, aim at expressing or improving physical fitness and
mental well-being, forming social relationships or obtaining results in
competition at all levels.”1
And this definition is accepted by the policy makers of European countries.
Within this definition, classifications essentially rely on differentiating activi-
ties according to their characteristics and organization, of which rules and
governance are a central part. For example, Rodgers argues that sports have
1
The charter was revised in 2001 but not the definition.
What is Sport? 37
four essential elements present. Physical activity is undertaken, for a recrea-

tional (that is non-obligated) purpose and this takes place within a framework
of organized competition that is regulated in an institutional setting.
However, despite the practical context, the boundaries between activities
remains blurred. For example, many sports are undertaken as a professional
activity, which implies that they are not recreational. Activities other than team
sports, such as swimming and cycling, may take place under similar compet-
itive and organized circumstances to team sports. However, these activities,
and indeed many others such as team sports, may also take place in a non-
competitive and non-institutional environment. In this regard they could be
viewed as “recreational” sports, inasmuch as formal rules of competition are
not followed. Finally, leisure activities may embrace reading, watching the
television, visiting a city, or indeed spectating at professional sports encounters.
They are neither competitive, rule bound, nor physical activities. Walking and
gardening are both physical activities, often undertaken for recreational pur-
poses. Walking, as illustrated below, is often included in sports participation
surveys by government and other agencies, but gardening is not. Sport England
has recently classified darts as a sport, but not chess. However, the Interna-
tional Olympic Committee recognizes that chess is a sport. In this regard its
governing body must ensure that its statutes, practice and activities conform to
the Olympic Charter, as with the other recognized activities discussed below. It
is clear then that, in practice, sport does not have a predetermined definition. It
follows that one should always bear in mind the context in which the terms
“sport,” “recreation,” “leisure” and “physical activity” are used.
This discussion illustrates the important point to grasp that the subject matter
of this book essentially reflects definitions of sport that have come to be used
within particular lines of enquiry, and advocated and discussed by various policy
making bodies and, ultimately, to have come to be accepted by the wider public.
As Gratton and Taylor (2000) note, definitions of sport involve “the cri-
terion of general acceptance that an activity is sporting, e.g., by the media and
sports agencies” (p. 7). This implies that this book will inevitably show a
degree of “bias” toward discussing sports and issues that are dominant in the
academic literature, as a subset of popular and policy discussion. None of this
is meant to imply that other issues or sports are of less significance per se.2 To
2
Consequently, for example, women's professional sports are emergent and reflect women's
participation in sports traditionally associated with males. Historically, Negro leagues
thrived in baseball in the US, and in the UK and in general, ethnicity and gender are
important in consideration of patterns of participation in specific sports. Issues such as these
are much more likely to be discussed in sociological literature. In economics, as indicated in
the definitions discussed in Chapter 1, attention tends to focus on “agents” making optimal
choices. These issues are discussed further in Chapter 4.
indicate what activities are defined as sport for policy and economic discus-
sion, it is necessary first to address the economic organization of sports. This
also indicates the various stakeholders that exist in the sports economy.
2.3 ECONOMIC ORGANIZATION

Figure 2.1 highlights a stylized sports environment which identifies the eco-
nomic and policy relationships involved in sport, based on a modification of a
schema presented in Gratton and Taylor (2000).
In Figure 2.1 the central triangular shape represents sports activity. Broadly
speaking the horizontal dimensions measure the number of people involved in
sport, the consumers and suppliers of sport. As one moves towards the apex
the professional nature of sports participation and supply increases, and
FIGURE 2.1 The sports economy.

Economic Organization 39
consequently the numbers of people involved fall. At the apex, thus, is “elite”
professional team sports (PTS). As one moves down the triangle one enters the
domain of formal mass participation, facilitated through sports clubs or
schools to informal and casual mass participation, in which consumers of
sport and suppliers become more and more distant in their relationship. For
example, as discussed in Chapter 5, traditionally sports can be viewed as being
consumed and supplied through the formation of voluntary sports clubs;
outside this domain participation can take place either through informal
use of municipal or commercial facilities. The boundaries of this activity
are, of course, difficult to establish, as reflected in the open-ended base of
the pyramid. Sports events bridge the elite and mass participation boundaries
because, as discussed further below, these are simply an episode of sports
competition. Yet at one extreme, say the Olympic Games, professional ath-
letes compete and are organized by economically powerful organizations. At
the other, many formal amateur sports compete in events as voluntary clubs.
The diagram does allow for the possibility of participation at lower levels,
feeding into higher level activity. A number of talent redistribution mechan-
isms might facilitate this. As discussed in Chapters 7 and 11 there are notable
differences in the North American model of professional sports and some of
the competitions in international sports, such as rugby union, where franchise
systems operate as opposed to a system of vertical selection of clubs, which is
common in European sports in general and soccer in particular. There are also
sports policies that are put in place that seek to extract talent from lower levels
of mass participation and to move this talent vertically.
The main commercial inputs to sport feed into the elite apex of the sports
economy through merchandising, sponsorship, television rights deals and
paying spectators, as discussed in Chapter 7. However, some sponsorship
exists in amateur formal sports activity, and fee paying spectators are also
possible. Non-commercial inputs are also very important. As discussed fur-
ther in Chapter 6, volunteering is integral to all organized sport. The diagram
also illustrates that flowing out of the direct involvement in sport are related
derived demands, for equipment, facilities, etc.
Derived demands express how interrelated economic activity is. For example, according
to Alfred Marshall: “The demand for raw materials and other means of production is
indirect and is derived from the direct demand for those directly serviceably products
which they help to produce” (Marshall, 1952, p. 316, italics in original).
Once again extremes are possible. One may be the clothing and equipment
required for an individual to participate in a casual activity. The other may
reflect the need to create an Olympic village. The important point to note here
is that sports are part of a set of interrelated markets.
All of this economic activity takes place within a regulatory framework. In

sports there are three main tiers to this framework. The first is the public
policy framework of national and supra-national policy-making agencies,
which is illustrated by the rectangular shape which borders the diagram. In
European countries, as implied earlier, the EU influences sports policy and
markets, as well as national governments. However, most national govern-
ments, in connection with a form of local government, also have sports policy
bodies which direct and implement policies. Often these have remits which
embrace both mass participation and elite sports. This is implied in the open
curly parentheses in the diagram. Finally, organized sport often takes place
under the remit of governing bodies. This is discussed further in Chapter 6.
However, and as discussed further in Chapter 7, governing bodies often face
powerful counter-organizational pressures from professional sports. More-
over, almost by design they have no control over informal activity. This is
why the curved brackets start at the juncture of formal and informal mass
participation sport, but do not reach the apex of the pyramid.
Of course, the emphases of these relationships can change. Moreover, the
comparability of the quantitative character of the sports environment between
cases can be questionable, because of the potential differences in definitions of
sport, but the qualitative structure of the sports environment, so defined, does
have some reliability, as the following discussion reveals. This discussion
draws on a wide variety of sources, but particularly useful overviews are
provided in Carter (2005), Green and Houlihan (2005), and Houlihan (1997).
2.4 SOME INTERNATIONAL EVIDENCE ON THE SPORTS

ENVIRONMENT
2.4.1 Public policy
Public policy involvement in sport has been steadily increasing. In part this is
because of the growth in the globalization and commercialization of sport. As
Green and Houlihan (2005) note, in Australia before World War II, federal
government involvement in sport was ad hoc, and the same was true of
Canada and the UK until the 1960s. As discussed in more detail later, the
emphasis ebbed and flowed between developing mass participation sport and
elite sport. Likewise, as Downward and Dawson (2000) argue, from an eco-
nomic perspective regulation of professional sports fell within the broad remit
of the Sherman Acts in the US and the Monopolies and Restrictive Practices
Acts in the UK. Currently, within Europe, the Treaty of Rome provides the
economic competition regulation framework. Box 2.1 illustrates issues that
the European Commission has recently examined.
Some International Evidence on the Sports Environment 41
BOX 2.1 RECENT EU RULINGS ON SPORT

1. Media rights for the UEFA Champions League. 8. The Mouscron case in relation to the UEFA home and
2. Multiple ownership of football clubs. away rule.
3. Football players' agents. 9. Doping case filed against the IOC by two swimmers
4. UEFA broadcasting rights. who were banned for doping. The commission ruled
5. Formula One – International Automobile Federation reg- that this fell outside the scope of community competition
ulations. rules.
6. Rules on the transfer of football players.
7. Grants to French professional football clubs. Source: http://www.euractiv.com
As can be seen, football provides a central focus for these regulatory initia-
tives, as it is the largest professional sport in Europe. However, more generally,
sport is viewed as subject to the subsidiarity principle, which means that it is
essentially a matter for member states. Box 2.2 indicates some important
developments in the consideration of sport by the EU.
In addition to such supra-national policy making, the organization of sport
in Europe has some common traits. Countries such as the UK and France have
BOX 2.2 SPORT AND THE EU

Sport is not dealt with separately as an EU policy area, of sport within the community framework. A sports unit
neither are there permanent groups for addressing sports was set up in the DG Education and Culture, to admin-
issues. These are dealt with in different committees, ister the European Year of Education through Sport
depending on the precise nature of the issue. The Commit- 2004.
tee on Culture, Youth, Education, the Media and Sport is & In Nice in 2000, a council declaration again supported
one main forum for debate. There is also a Parliament the need to preserve and promote the social functions of
Sports “Intergroup” which meets informally. Indirectly, sport. It also noted the specific characteristics of sport, to
sport is influenced by policy in a variety of areas, as detailed intensify European cooperation in the area of doping and
in the initiatives below. the UN Millenium Declaration on the promotion of peace
and mutual comprehension by means of sport and the
& 1974: the Walgrave ruling established sport as an eco-
Olympic Truce.
nomic activity, and subject to the Treaty of Rome.
& 2003: the council and parliament approved the estab-
& 1992: the Maastricht Treaty referred to the social signif-
lishment of the European Year of Education through
icance of sport.
Sport 2004.
& 1997: the Amsterdam Treaty stressed the importance of
& Article III–182 of the draft constitution refers to the ed-
amateur sport, and the need for dialogue with sports
ucational and social role of sport. Sport is also listed in
associations, and emphasized the social significance
Article 16 as one of the areas where the EU may take
of sport.
supporting, coordinating or complementary action.
& 1999: the Helsinki Report stressed the need to safe-
guard sport structures and maintain the social function Source: http://www.euractiv.com
more generalized ministries as the first layer of state involvement, but some of
these do have a sport remit. In the UK the relevant ministries are the Depart-
ment for Culture, Media and Sport and the Department for Education and
Science. In 2006, a new central government minister for Public Health was
established to work in partnership with the Department for Culture, Media
and Sport, the Department for Communities and Local Government, the
Department for Transport, the Department for Education and Skills and
sports delivery bodies to raise participation. Likewise, in France the Ministry
for Youth and Sports, together with the Ministry of Education are involved in
sports delivery. In Finland the Ministry of Education subsumes sports, where-
as in Germany sport is devolved to its 16 constituent federal L€ander (European
Commission, 1999).
Some similar patterns exist elsewhere. In Japan sport lies within the remit
of the Ministry of Education, Culture, Sport, Science and Technology, and in
Canada, sport resides within the Federal Department of Canadian Heritage. In
Australia there is much stronger direct federal support for sports agencies, in
contrast to the US in which there is none. This is suggestive of a distinction
between the US and other sports economies, a theme which is developed more
later on in the book.
As well as central government, direct government involvement in sport is
also mediated at local government levels. For example, in Germany and
Finland, local authorities are the principle providers of funds for facilities
and sports clubs. In Canada mass participation has tended to fall within the
remit of the provinces (Green and Houlihan, 2005). As Houlihan (1997)
notes, moreover, in the UK the distribution of responsibilities is complex
at a local level. Historically in the US, by contrast, sport was supplied by
the local, municipal and state authorities. However, there is a growing pri-
vatization of sport in the US, with facility redevelopment often being under-
taken by private investment. In addition, the college system is uniquely
placed in US sport in terms of bridging the gap between mass and elite
participation.
2.4.2 Sports policy bodies

Lying below central government agencies, and in addition to local govern-
ment, most countries have sports policy bodies which act as vehicles for
delivering central and local government initiatives or allocating finances to
sport. In the UK, UK Sport and four sports councils, Sport England, Sport
Scotland and the Sports Councils for Wales and Northern Ireland perform
these functions. However, as the Carter report (2005) indicates, there is a
blurred distribution of functions and funding between the agencies.
In Germany, the German Sports Confederation (Deutscher Sportbund)

and its regional offices perform the functions of public policy delivery and
funding, and in Australia the Australian Sports Commission (ASC) embraces
the delivery of elite sport via the Australian Institute of Sport. Policy associated
with mass participation is administered via Active Australia. In addition,
these agencies are partnered by the Australian Sports Foundation, which is a
company governed by the ASC, whose role is to raise funds. Finally, in Canada,
Sport Canada is directly responsible to the federal government and has three
main divisions: Sports Programs; Sport Policy; and Major Games and Hosting
Events.
2.4.3 Governing bodies and professional team sports

Once one moves away from the legislative and policy framework of sport,
attention turns to the main organizational stakeholders in the provision of
sport, the main elements of which are the participants and spectators, the
clubs that organize sports competition and the governance of sport by clubs,
along with other collective bodies. These stakeholders are related in a complex
way that reflects historically and socially contingent formation, not least of
which concerns the development of specific sports. It is significant to note that
it is out of this complex evolution that professional sports leagues emerged,
that large-scale events could develop, and that the structures that provide
much mass provision of sport developed.
Downward and Jackson, for example, discuss the complex evolution of
codes of football in the UK, its definition and governance. In brief, the impor-
tant developments reflected the emergence from public schools of organized
games of football. These were distinct from the informal traditional games
played between villages, for example, versions of “Shrove Tuesday Football”
such as are still played today in Ashbourne in Derbyshire, where two teams of
unlimited size struggle to carry a ball to either ends of the village. The descrip-
tion as “football” reflected the fact not that a ball is kicked, but simply that it
was distinct from pastimes such as hunting and riding which, being under-
taken on horseback, were confined to the aristocracy (Houlihan, 1997).
Commensurate with a rise in voluntary associations, the Football Associ-
ation (FA) was established in 1863 as the relevant governing body of the game
of soccer. Rule changes led to association football encouraging a passing game
using the feet (Williams, 1994, pp. 29–30). In contrast, Rugby School devel-
oped its own distinctive set of rules in 1862 to allow “old Rugbeians” to
introduce an alternative to those interested in soccer. The Rugby Football
Union (RFU) was established in 1871. At the outset the “Corinthian ideal”
of amateur sport was embraced by both codes of football (Houlihan, 1997;
Downward and Jackson, 2003). In this regard, voluntary participation and

organization was emphasized and remains today in mass participation sports
clubs.
There was also, however, a growing concern to manage the increased
popularity of both of the games. In large part this had been prompted by the
development of cup competitions which provided incentives for clubs to codify
their play according to particular sets of rules (Williams, 1994, p. 44); other
than cup matches, fixtures tended to involve traditional or local rivalries.3
Rapid expansion also carried with it the inevitable consequence of vertical
segmentation and sporting differentiation. Alternative strategies evolved to
cope with the expansion, particularly at the elite level. By 1888 soccer em-
braced professionalism at its highest level and the English Football League
(EFL) was established to work with the FA to manage the competition between
the top soccer clubs, while the FA was responsible for the management of the
game as a whole, including its huge amateur base.
Leaving behind their amateur status, professional clubs began to incorpo-
rate themselves as limited companies. The rapidly-growing potential for earn-
ing money meant that enclosed stadiums were built, entrance fees for
spectators extracted, and players paid and bought and sold. The need to pay
to attract the best players existed even before the formation of the EFL, with
evidence that paid professionals existed in 1876 (Vamplew, 1988).
As Dobson and Goddard (2001) note, the original 12 teams expanded to 16
in the top division. A further 12 clubs facilitated the creation of a second
division before World War I. Movement between the divisions was based on
merit, i.e., vertical selection. After the war, further expansion included a seg-
mented third division for teams constituted on geographical lines for the north
and south of England. This was achieved by incorporating elements of a pre-
vious southern league, which included some clubs from south Wales. By 1951,
92 league teams existed in the three divisions, and in 1959 the geographical
basis of division three segmentation was replaced by a division three and four
constituted on merit. During the 1970s and 1980s various changes took place.
The most important of these included increasing the numbers of teams
3
Such elements exist today in rugby union, the last sport to turn professional in the UK.
Here the varsity match, between Oxford and Cambridge Universities, retains its social
significance with the governing body and is played at Twickenham, which is the national
stadium owned and managed by the Rugby Football Union. Outside of the County cup
competitions also, amateur rugby union at a junior level is not organized by competitive
leagues, unlike soccer. Likewise, invited “barbarian” teams are often put together to play
international matches against national sides. The allegiance is to the game of rugby, with an
implicit ethos being to play an expansive passing and attacking game for public enjoyment.
relegated and promoted, and the introduction of a play-off system to determine

one of the promotion places. In 1993, the Premier League (PL) was formed by a
breakaway of the division one sides, an issue discussed further in Chapter 7.
Coupled with domestic expansion in soccer, growth in international com-
petitions promoted the need to develop supra-national governance. Procrasti-
nation by the English FA implied that in 1904, the Federation Internationale
de Football Association (FIFA) was formed at a meeting of seven other Euro-
pean countries in Paris on 21 May, although the English FA soon joined. Initial
competitions between countries took place at the Olympic Games before and
immediately after World War I. Other countries, such as South Africa, Argen-
tina and the US, joined FIFA, although England and Scotland refused to
participate in competitions involving former wartime enemies. By 1930, un-
der Jules Rimet's Presidency, the first world cup took place in Uruguay. With a
break for World War II, the world cup has been contested every four years since.
As can be noted below, a four-year cycle of international competition became a
common basis of international competition, reflecting the time and travel
costs associated with the expansion of competition. Competitions were of a
knockout variety.
Reductions in travel costs generally meant that by the 1950s more regular
international knockout competitions were held for both national sides and
clubs. The Union of European Football Associations (UEFA) was established
in June 1954 in Basel to oversee these developments in European football.
Other continental associations have since formed. UEFA oversees the Euro-
pean Football Championship which started in 1958 between national teams,
and a series of club-based international competitions. The European cup, for
national league club champions was first contested in 1955, and this has
subsequently evolved into the Champions league (CL), as discussed in Chap-
ter 7. The Inter Cities Fairs Cup was established in 1955 to facilitate compe-
tition between national major knockout cup champions. This evolved into the
UEFA cup in 1971. The UEFA cup also absorbed the Cup Winners Cup, which
was started in 1961, in which lesser national knockout trophy winners com-
peted.4
4
FIFA recently launched an international World Club Championship, but it has met with
mixed success. On one occasion Manchester United's participation in this cup, while
withdrawing from the national FA Cup competition met with much protestations. The
2002 event was cancelled, but in 2005 a much reduced tournament took place. UEFA also
oversees a number of other smaller competitions, including the Super Cup and the Intertoto
Cup. In the former competition, Champions League winners play winners of the UEFA cup.
The latter was originally a summer competition operated by some central European
associations, but has expanded to become a qualifying tournament for the UEFA cup.
Over the same period, in which soccer developed, the rugby code was
characterized by schism which was based on alternative approaches to
“rejecting” professionalism. The divide occurred along broadly geographical,
as well as class, lines. Many amateur rugby clubs were restricted to the
professional classes. This was particularly the case in the south of England,
but also in Liverpool, Sale and Manchester in the north of Britain. In
contrast, in the north of England and south Wales the game was opened
up to working men and artisans. Because of the length of the working week –
including Saturday mornings until 1 pm – it was common for northern
rugby union clubs to offer “broken time” payments to compensate for loss
of earnings. The RFU in London were vehemently against this development
and matters came to a head when, in the annual general meeting of the RFU
in 1893, two coalitions of clubs – the Yorkshire branch and the remainder –
debated the issue. The former argued that broken time would be the best
means of avoiding professionalism, the latter not. The result was cast
against broken time and the imposition of stringent rules against such
payments instigated.
However, in 1895, 21 clubs of the Northern Rugby Union, based in York-
shire and Lancashire, broke away to form their own league – the Northern
Rugby Football League. Initially “professional players” receiving broken time
payments were allowed, providing they were in full-time employment, and
players could transfer between clubs if allowed by the Northern Union (Moor-
house, 1995, p. 61). By 1922 the game was relabelled as the Rugby Football
League (RFL), to distinguish it from the RFU which managed the amateur
game for the remaining teams. As with soccer, professionalism was embraced
with the subsequent development of stadiums and incorporation of clubs.
Convoluted changes to both the structure of competition and rule
changes followed the breakaway.5 In the former case, traditionally struc-
tured county championships were contested for five years following the first
season. By 1902 two divisions were introduced with 18 sides each, and
promotion and relegation again depending on merit. By 1905 the second
division had contracted and the two divisions were scrapped again. A unique
system was adopted within which approximately 31 teams competed for a
single championship decided on the percentage of wins rather than the total
number of points. From 1907 the top four teams contested the champion-
ship by play-off. The next major changes were to come in 1962 in which
eastern and western divisions were installed, only to be scrapped again in
1964. In the new single division a play-off of the top 16 clubs decided the
5
See Downward and Jackson (2003) and any Rothman's Rugby League Yearbook for a
summary of key historical developments.
championship. This lasted until 1974, when two divisions were reinstated
with play-off finals. The next major change was the creation of three divi-
sions in 1995 making way for the creation of the super league (SL) in 1996
for the top sides, which will be discussed further in Chapter 13. From 1998
both the SL and the first division championships were to be decided from
play-offs.
In the case of rule changes, attempts were made to speed up the game and
encourage further handling of the ball. The main ones established a distinctive
code of the northern rugby sides as opposed to the remainder of rugby union.
From 1897 lineouts were abolished, because time-consuming scrummages
tended to follow. In 1906 two of the forwards were removed, reducing team
sizes to 13 men. A key rule change was the “play-of-the-ball” rule, which
removed the need for a scrum following a tackle and an unplayable ball.
Rewards for tries were increased – earning three points compared to the two
from goals. This was subsequently increased to four points in 1983. In 1966
the number of tackles was limited to four, which was then extended up to six in
1972.
From an economic perspective, it is important to note that the above
rule changes were not just directed at differentiation from rugby union. As
professional sports, soccer and rugby league were in direct competition for
both fans and playing talent. Consequently, teams such as Preston North
End experimented with playing both codes. In general, being part of a well-
organized football league with strongly professional aspirations, soccer
teams flourished, as evidenced by the growth in scale of the EFL as opposed
to the RFL (Williams, 1994, p. 82). Part of this expansion was at the
expense of rugby league teams, such as Radcliffe and Walkden in Lanca-
shire, which became unsustainable facing the competition for fans provid-
ed by the Bolton Wanderers and Bury FCs who were successful sides.
Others, the most notable being Manningham, who had won the inaugral
northern rugby union championship, converted to soccer in 1903 and re-
emerged as Bradford City FC, who compete in the EFL today. Players also
switched between codes. Part of the expansion of the EFL was, however,
because it could monopolize the “vacant” midlands lying between the
disputed geographical territories of rugby football. It was in direct response
to the economic threat from soccer, thus, that rugby league struggled to
adjust its form of competition and the rules of the game – to establish
barriers to entry. It was also important that Yorkshire and Lancashire clubs
played one another, because the initial competition with soccer was in the
latter rather than the former county (Williams, 1994). Indeed the consol-
idation of talent in the former ensured that Yorkshire sides were dominant
in the early years of rugby league. They won the county championship
against their rivals in seven of the first eight seasons from 1889, and hence
it was by no accident that they were the most vocal elements in the
breakaway.
International rugby league, as with association football, developed out
of early national competitions between members of the British Isles, how-
ever, competition with then colonial or near neighbours such as Australia,
New Zealand and France soon took place. Its formal organization only
developed with the formation of the Rugby League International Federation
in France in 1948, as an attempt to preserve the game in France. France
has struggled to recover from the political bans imposed on rugby league by
the Vichy government. Other countries, particularly the pacific islands,
subsequently played rugby league, as well as Russia and South Africa. In
general, the strongest sides come from Great Britain, Australia and New
Zealand, and currently these three countries compete in an annual tri-
nations series.
In contrast to rugby league, rugby union remained staunchly amateur,
although pressures towards becoming professional began in the 1960s. Es-
sentially this was driven by gate-taking clubs wanting to increase resources
through more regular competition. A national knockout cup was thus in-
stigated in 1971 and national leagues finally established in 1976. In addi-
tion, the RFU and major clubs became increasingly dependent on gate
money, sponsorship and media income. This was particularly fermented
with the growth of interest in national team competition, ostensibly orga-
nized around the traditional annual competition including the home
nations of the British Isles and France, the “five nations,” which has recently
been augmented to include Italy as the “six nations” in 2000. This interna-
tional interest was galvanized around the instigation of a world cup, which
was inaugurated in 1987 to promote more regular competition with south-
ern hemisphere teams by separate national sides, in addition to traditional
“British Lions” tours, on a four-year cycle, in which composite teams from
the British Isles played other rugby union playing nations. Between 1900
and 1924 rugby union had been played at the Olympic Games, but with
small numbers of participants.
By the 1990s English international players began using trust funds to get
around professionalism rules, implying that earnings from sponsorship and
appearances were not paid directly for playing, and there came public recog-
nition that professionalism was more overt in other countries, particularly
New Zealand and France. There was also concern that, coupled with the rise of
the PL and the SL, media companies were seeking to develop a professional
game, organized distinctly from traditional governing bodies. Consequently,
by 1995 rugby union was declared an open game by the International Rugby
Board (IRB).6
The IRB had been established in 1886 as the International Rugby Football
Board (IRFB) by Scotland, Ireland and Wales, and subsequently England, fol-
lowing the resolution of concerns over representation and law-making powers.
The current IRB comprises representatives from Scotland, Ireland, Wales,
England, Australia, New Zealand, South Africa, France, Argentina, Canada,
Italy and Japan, and actively seeks to develop the game internationally. It also
controls international competition through the world cup and international
seven-a-side tournaments.
Concern with the international games' development has also taken place
against the backdrop of growing media incomes for sport and the desire not to
lose players to rugby league deals. As with the development of the Premier
league in association football, the Super league in rugby league and the
Premiership in UK rugby union, the three main southern hemisphere
rugby unions formed SANZAR (South African, New Zealand and Australian
Rugby) which developed “Super” rugby in 1996. Growing out of earlier com-
petitions between South Sea islands, this evolved into a league of originally 12
now 14 regional sides from these unions. A tri-nations series between the
national sides of Australia, New Zealand and South Africa was also instigated
in 1996.
A clear implication of the above discussion is that codes of football have
evolved, reflecting both organizational and economic pressures, within an
international setting. It is notable, thus, that alternative forms of football have
also developed, for example in Ireland, Australia and North America. Links
between these codes of football and those above are implicit, if not explicit, and
reflect historical links between countries. Table 2.1 describes some character-
istics of other professional sports.
2.4.4 Summary
These issues clearly demonstrate that team sports are organized in a complex
way, with historically derived governing bodies managing mass participation
of a formalized activity through clubs. Some of these clubs became profession-
al, and leagues and other forms of competition emerged in both national and
international contexts. Despite specific differences between sports, however,
6
These pressures were also experienced in the southern hemisphere, with the
establishiment of a Super League in Australian rugby league. Fear over the potential loss of
players led to the Super 12 competition being developed, as discussed below, to raise
broadcasting revenue and to enable higher salaries to be paid.
50
CHAPTER 2 The Nature, Organization and Economic Significance of Sport
TABLE 2.1 The origin and characteristics of other professional team sports
Sport First club and Countries played Example of national International Examples of forms of
rules governance governance competition
Cricket Marylebone England, then In England and Wales The International International test cricket of
Cricket Club initially Australia all competitions take Cricket Council five days duration plus various
1788 and South Africa, place under the (ICC), founded in “limited over” one day games
then the West jurisdiction of the 1989 from earlier for full members of the ICC
Indies, New English and Wales bodies
Zealand and India. cricket board (ECB)
Subsequently
Pakistan, Sri Lanka
and Zimbabwe
One day form world cup
competition
National games for county
sides of four days duration plus
various “limited over,” one day
games Amateur games of
limited overs
Ice hockey 1877 McGill Canada, US, Czech In the US (24 teams) International Ice The Stanley Cup is awarded to
University Republic, Finland, and Canada (6 teams) Hockey Federation NHL league champions in a
students in Russia, Slovakia, the National Hockey (IIHF) play-off system
Canada Sweden, UK League (NHL) prescribes
the rules and governs
the professional sport.
Amateur sports in Hockey The annual world
Canada and US Hockey championships are competed
compete with rules drawn for internationally by nations
from the NHL and the
IIHF
Ice hockey has been played at
the Winter Olympic Games
since 1924
Basketball 1891 Springfield The sport has global Most early competition The International US college varsity
College popularity. The US, in the US was between Basketball Federation tournaments culminate in
Massachusetts South American, colleges and this led to (IBF) was formed in “March Madness” play-offs
European, Asian the formation of the 1932 for amateur
and Australasian Intercollegiate Athletic players. Amateur and
countries play Association (IAA) and professional
basketball, as well subsequently the NCAA. distinctions were
as Russia, Competition between dropped in 1989
Scandinavia and numerous professional
the Baltic States teams also took place.
The National Basketball
Association (NBA) was
formed in 1946. The
American Basketball
Association (ABA) began
in 1967 but merged with
the NBA in 1976.
The Womens National The NBA has a league
Some International Evidence on the Sports Environment

Basketball Association structure
began in 1997
Basketball has been
played at the Olympic
Games since 1936, with
women's basketball since
1976.
The basketball world
championships were
first held in 1950
51
52
CHAPTER 2 The Nature, Organization and Economic Significance of Sport
TABLE 2.1 (Continued )
Sport First club and Countries played Example of national International Examples of forms of
rules governance governance competition
Baseball The New York The US, Canada, In the US the National The International Each US league has a pennant
Knickerbockers Mexico, Caribbean League (NL) has existed Baseball Federation race for the champion
1845 countries, South since 1875, despite rival (IBAF) emerged in
America, Japan, leagues. It has 16 teams 2000 from earlier
Korea, Taiwan and versions beginning in
European countries 1938
such as Italy, the
Netherlands and
some African
countries, Australia
and New Zealand.
The American League (AL) The World Series set up in
has existed since 1900. It has 1903 is played between the
14 teams winners of the two main
leagues
The National Association of World series were also played
Professional Baseball among Negro leagues
Leagues was set up in 1901
and has become the basis of
the minor leagues
Many amateur black players The first official Olympic
were forced to play in tournament took place in
separate clubs and various 1992, but the game has been
“negro leagues” were formed, removed from the 2012 games
culminating in the Negro
American League in 1937.
This folded in 1960 as a
result of the removal of race
restrictions in 1947
In 2006 a world baseball
classic will take place replacing
the previous irregular world
cup that began in 1938. Mostly
amateur players participate
Events and Mass Participation 53
as Arnout (2006) also confirms, it is argued that the above does illustrate that a
pyramid structure of sports as represented in Figure 2.1 exists. There are
distinctions in the US, where professional sports and amateur sports have
developed more independently, but they are, of course, still inextricably linked
in the sense that college sports and the minor leagues feed talent vertically into
professional sports.
2.5 EVENTS AND MASS PARTICIPATION

Clearly, as described in Figure 2.1, mass participation activities provide the
basis through which individuals or teams come forward to compete. Such
competition then takes place according to an “event.” However, this event
can be organized by a league or knockout tournament, or a combination of
these, as discussed further in Chapter 7, for a separate sport or for a collection
of sports. In this regard what is meant by an “event” in this book has a specific
meaning. It is essentially connected with a less-regular sports competition, in
both the frequency of competition and the location in which it takes place
compared to a league.
2.5.1 Events
The Olympic Games is the largest sporting event and, under the stewardship
of Pierre Fredy (Baron de Coubertin), began formally in 1896 what we now
understand as the summer games. The Winter Olympic Games commenced
in 1924.7 The International Olympic Committee (IOC) was established in
1894 to organize the first modern games, held in Athens in 1896. The Games
have run up to the present day only being suspended in 1916, 1940 and 1944,
because of the World Wars.
The organization of the games is controlled by the IOC in conjunction with
partner agencies. These include international federations for specific sports,
National Olympic Committees (NOC) within each country, who regulate the
Olympic movement in that country, and specifically constituted organizing
committees (OCOG) to stage the games. The latter are disbanded after each
games. There are currently 202 NOCs and 35 international federations in-
volved in the Olympic Games, and they are planned to supervise 26 sports
scheduled for the 2012 London games. In the original games in 1896, nine
sports were involved: athletics; cycling; fencing; gymnastics; weight lifting;
shooting; swimming; rowing and wrestling. Current rules permit up to 28
sports. Individual sports require a two-thirds vote from the IOC to be
7
In England versions of national “Olympic” competition preceded these dates.
BOX 2.3 2012 OLYMPIC SPORTS

Aquatics Hockey Cycling Taekwondo
Archery Judo Equestrian Tennis
Athletics Modern pentathlon Fencing Triathlon
Badminton Rowing Football Volleyball
Basketball Sailing Gymnastics Weight lifting
Boxing Shooting Handball Wrestling
Canoe/kayak Table tennis
considered as Olympic sports, and thereby become eligible for inclusion in a

games. These include activities as wide in scope as climbing, bridge, chess (as
noted earlier) golf, roller skating and surfing, and of course team sports such as
cricket, baseball and rugby, which have been included in the past. Box 2.3
indicates the sports likely for the 2012 games, as baseball and softball have
been dropped (http://www.olympic.org/uk).
For the Winter Olympic Games, cross-country skiing, figure skating, ice
hockey, Nordic biathlon, ski jumping and speed skating have always been
involved. In addition, the bobsleigh, curling, ice hockey and luge are very often
included. Since 1960 in Rome, the Paralympic games for athletes with dis-
abilities have been a part of the Olympic programme, with events taking place,
for example, after the close of the summer games.8 It is important to recognize,
therefore, that the activities identified by the IOC as sport are also generated
out of particular historical emphases and policy decision-making.
In this regard the other major events such as the Commonwealth Games
(CWG), which is the second oldest multi-sport event, although lagging behind
the scale of the Olympic Games, involves a different list of sports, as illustrated
in Box 2.4.9
Since the Manchester Commonwealth Games in 2002, there have been
athletics, swimming, table tennis and power lifting events for athletes with a
disability.
8
It should be noted that the “Special Olympic Games” and “Paralympic Games” are distinct
organizations recognized by the IOC. Essentially the former is connected with the provision
of year-round sports training and competition for individuals (age 8 and older) with
intellectual disabilities. More than one million athletes in over 150 countries train and
compete in more than 26 Olympic-type sports at local, state, national and World Games.
The International Paralympic Committee, (IPC), in contrast, focuses on producing a games
parallel to the Olympic Games, and other multi-disability competitions for elite athletes.
9
For example, it is estimated that the 2006 games in Melbourne was about a third of the
scale of the 2000 Sydney Olympic Games. http://www.minister.dcita.gov.au/kemp/media/
speeches/ Accessed on 21 June 2006.
BOX 2.4 COMMONWEALTH GAMES EVENTS

Current Previous
Aquatics (swimming and diving) Lawn bowls Archery
Athletics Netball Cricket
Badminton Rugby (sevens) Fencing
Basketball Shooting Wrestling
Boxing Squash Judo
Cycling (road and track) Table tennis Rowing
Gymnastics Triathlon Ten-pin bowling
Field hockey Weight lifting
The Olympic Games and the Commonwealth Games are not the only
major multi-sport events. There are also continental championships such as
the European Games, the Asian, Pan American and African Games. As noted
earlier, moreover, there are, of course, world championships in various specific
sports, such as athletics. In this regard, only the world cup in soccer rivals the
Olympic Games in terms of scale. The economics of these events is discussed
more fully in Chapter 12. What is clear, however, is that these “events,” as
defined earlier, currently operate at the elite level, with participants that are
now all professional in orientation. Significantly too, staging these events has
become a competitive economic and political activity. Countries and cities
now actively compete, through committing resources to bids, to host them. It
is such developments which clearly merit economic analysis.
2.5.2 Mass Participation

Below the professional and elite level of sports is the mass participation sector.
Care should be taken not to treat this sector as homogenous. One segment
comprises formal club-based activities, for example operating under governing
bodies that also govern elite team and individual sports and competition in
leagues or events. In formal participation it is not uncommon practice to have
elements of paid coaching and facilitation of sport, combined with huge levels
of voluntary activity. In contrast, much participation now takes place by
individual consumers who pay to participate in classes at health and fitness
clubs or who participate in a highly informal and self-organized manner. The
Carter Report (2005) estimates that up to 43% of sports participation in the
UK operates on an informal basis.
Informal sport describes an activity that displays many of the characteristics of
“formal sport,” but does not involve structured competition governed by rules.
Therefore, two people playing basketball on their local “Outdoor Basketball
Initiative” site may be said to be engaging in informal sport, because the activity is
very close to being what we understand basketball to be, but some of the defining
characteristics of that sport have been compromised. Informal sports facilities can
include skateparks, BMX tracks, basketball courts, kickabout areas and multi-use
games areas. Associated facilities may include youth shelters or equipped play areas
for younger children (Sport England, 2006).
Coalter (1999) charts the rise of this activity in the UK, and argues that it
reflects more individualistic activity, flexibly organized on a non-competitive
basis and showing more concern for fitness and health. In this regard, keep
fit and yoga activities have recently increased very significantly for female
participants, a trend also identified by the Australian Sports Commission
(2004).
A detailed exploration of the reasons for participation in sports takes place
in Chapters 3 and 4, along with a discussion of the supply structure of mass
participation sports in Chapters 5 and 6. However, to indicate the types of
activities that are conceptualized as sport outside the elite level of team and
event sports, Table 2.2 presents examples on which different national public
authorities collect data. It should be noted that these are not the only data
sources for these particular countries, neither are they, in the light of informal
sport, definitive. This said three issues are of particular interest concerning the
definition of sport and its official measurement. The first is that there is some
broad commonality over the activities investigated. Indeed, in Europe, the
COMPASS project has been seeking to promote harmonization in the collec-
tion and reporting on official sport participation statistics.10 The second issue
is that some of the activities are unique to the country involved. This, of
course, reflects the historical development of sports within the various coun-
tries, as discussed earlier. The final point to note is that the terminology varies
according to context. Thus, soccer is used in North America and Australia for
association football, which is referred to as football in the UK. Conversely, the
term “football” is used for American football in the US, but qualified as such in
the UK.
It is also important to note that the data collected on these activities does
vary, reflecting differences in conception of participation. In Australia the
Exercise, Recreation and Sport Survey (ERASS) commenced in 2001 and is
conducted annually, commissioned by the ASC and state/territory depart-
ments of sport and recreation (see Australian Sports Commission, 2004).
Participation is defined as active involvement, rather than coaching, officiat-
ing or spectating, and data is collected for those aged 15 years and older. Data
10
See, for example, http://w3.uniroma1.it/compass/index.htm Accessed 13 June 2006.
TABLE 2.2 Examples of official data participation sports1
Australia Canada UK USA
Aerobics/fitness Badminton Athletics Aerobics/aerobic dance

Aquarobics Baseball American Baseball/softball
football
Athletics/track and field Basketball Angling Basketball
Australian Rules football Bowling Badminton Bicycle/exercise bike
Badminton Curling Basketball Bowling
Baseball Cycling Bowls Cross country skiing
Basketball Ice hockey Canoeing Football
Billiards/snooker/pool Rugby Climbing Gardening/yardwork
Boxing Running Cricket Golf
Canoeing/kayaking Skiing Curling Handball/
racquetball/squash
Carpet bowls Soccer Cycling Jogging/running
Cricket Swimming Football Skiing
Cycling Tennis Gaelic sports Soccer
Dancing Track and Golf Stair climbing
field
Darts Gymnastics Stretching exercises
Fishing Hockey Swimming
Golf Horse riding Tennis
Gymnastics Jogging/running Volleyball
Hockey Keepfit/aerobics Walking
Horseriding/equestrian Martial arts Water skiing
Ice/snow sports Motor sports Weight
lifting/strength
Lawn bowls Netball
Martial arts Rugby
Motor sports Sailing
Netball Skiing
Orienteering Shooting
Rock climbing Snooker
Roller sports Swimming
Rowing Squash
Rugby league Table tennis
Rugby union Tennis
Running Tenpin bowling
Sailing Volleyball
Scuba diving Walking
Shooting sports Weight lifting
Soccer Weight training
Softball Windsurfing
Squash/racquetball
Surf sports
Swimming
Table tennis
Australia Canada UK USA
Tennis
Tenpin bowling
Touch football
Triathlon
Volleyball
Walking (bush)
Walking (other)
Water polo
Water skiing/powerboating
Weight training
are collected on participation during the 12 months prior to the interview. The
sample size in, for example, 2004 was 13 662 persons.
In Canada, the activities above refer to those measured as part of the
National Household Survey on Participation in Sport (NHSPS) conducted in
2004, of 2408 households, and commissioned by the Conference Board of
Canada (see The Conference Board of Canada, 2005). Participation data are
collected according to both active and non-active variants, such as volunteer-
ing and attendance at games or events, over a 12 month period, for those aged
16 years and above. Motorized sports are explicitly excluded, as well as activ-
ities that do not seek to improve personal sporting performance, such as
jogging or cycling to work. In contrast in the US, the activities noted above
are investigated in the National Health Interview Survey, which is a household
survey conducted in 1985, 1990 and 1991 for individuals aged 18 years or
older (see US Department of Health and Human Services, 1996). Participation
is recorded for activity over the previous two weeks. In 1991, 43 732 house-
holds were surveyed.
In the UK, the activities refer to the General Household Survey (GHS) (see
National Statistics, 2004). While the GHS is an annual survey of UK house-
holds, it only reports at various intervals on sports and leisure participation.
The last two occasions, prior to the 2002 survey, on which the activities were
derived were 1993 and 1996. The survey investigates participation primarily
for those aged 16 or above in the four weeks before the interview with respon-
dents took place, and for the 12 months before the interview took place. The
activities noted above are categorized as “sports.” Other activities that are
categorized as “leisure” include watching television, listening to or playing
music, reading, acting, dancing, attending leisure classes and other activities.
It is clear that any empirical claims that are offered with respect to trends in
Historic and Current Policy in Sport: Changing Emphases and Values 59
mass participation sport should pay due regard to these definitional differ-
ences. More recently the Active People and Taking Part Surveys have been
commissioned for England and the UK by Sport England and the DCMS
respectively with more activites investigated. Chapter 4 gives some details
of these surveys.
2.6 HISTORIC AND CURRENT POLICY IN SPORT:

CHANGING EMPHASES AND VALUES
The above discussion has attempted to map out the broad structure of the
meaning of sport as comprising a wide variety of activities being consumed
and produced under different circumstances. The aim has been to provide an
overview of the sport environment. Before closing this chapter, however, it is
worth summarizing some of the main changes that have taken place in eco-
nomic and policy emphasis within the sports environment as described above,
as it is ultimately these changes that motivate issues of direct economic
interest in sports.
To begin with, a general traditional tension identified in sports policy arises
out of concerns for fostering mass participation and also elite sport success in
the non-professional sports context, and particularly major events such as the
Olympic Games. Houlihan (1997), Green (2004), and Green and Houlihan
(2005) note, for example, that the traditional emphasis of sports policy in say
Australia, Canada and the UK were aimed at raising general physical fitness
and enhancing social welfare, but such initiatives have gradually been eroded
in favour of the pursuit of elite success.
This suggests that one policy tension facing sport is the provision of public
funds for the support of elite versus mass participation activities. As discussed
in Chapter 1, there is an opportunity cost implied in supporting one rather
than the other. From an economic perspective, however, this also raises the
important prior questions of whether or not public funds should be used at all
to promote mass participation and, further, to underpin elite sport activity,
such as hosting and participating in sporting events which do not fund them-
selves. It has to be noted that economists are sceptical about the value of such
initiatives.
Issues such as these have become important, because the scale of economic
activity associated with elite sports at events has increased significantly. Blake
(2005) estimates that the operating costs of staging the 2012 Olympic Games
in London alone will be £1.010 billion, with further infrastructural invest-
ment costs of at least a further £642 million. Moreover, the 1972 Munich
Olympic Games and 1976 Montreal Olympic Games made losses of £178
million and £692 million respectively, while the 1984 Los Angeles Olympic
Games and the 1992 Barcelona Olympic Games made surpluses of £215
million and £2 million. Even as part of the process of bidding to host major
events, cities now invest huge sums of money. For example, the London bid is
estimated to have cost £17 million, while the 2010 winter games bid for
Vancouver cost $35 million.11 This competitive marketing of cities to hold
events, as discussed in Chapter 12, is now often used to capture sports fran-
chises, and has been described as a process of “glocalization” in which local-
ities compete on global markets for sources of tourism and related
expenditures (see, for example, Robertson, 1995). The economics underlying
these issues need to be clearly understood.
The above discussion also hints at the internationalization, globalization
and commercialization of sports generally.12 Of course, in the case of profes-
sional team sports, as noted earlier, it has been argued that the spread and
development of sports has occurred along increasingly commercial lines, de-
veloping and adjusting previous patterns of development. What is particularly
significant in this regard, however, is the relatively recent but rapid growth of
funds flowing between stakeholders in such sports.
To illustrate this point Dobson and Goddard (1995) note that in 1967
individual football clubs in England and Wales, in the most popular team
sport, received £1300 from television sources for the showing of highlights.
In 1978 the amount had only risen to £5800. In contrast, BSkyB paid over
£600 million for the exclusive rights to broadcast live Premier league associ-
ation football matches between 1997 and 2001, and £1.024 billion from
2002–2007.13 Since 2006, as discussed in Chapter 10, exclusive rights for
televising the Premiership do not exist. Table 2.3 also indicates the scale and
variety of income for three top European football clubs for the 2003–2004
season. As well as the scale of funds involved, the table also reveals that
revenue from television and commercial activities such as sponsorship and
merchandising are now as important as attendances, the historical source of
revenues, for the top clubs.
11
See http://www.londontown.com/London/Timeline_of_the_Bid Accessed 14 June 2006
and http://thetyee.ca/Views/2004/05/24/Blame_It_on_Olympic_Fever/ Accessed 14 June
2006.
12
Debates between theoretical notions of globalization and internationalization are beyond
the scope of this book. Contrasting perspectives could be reviewed in Maguire (1999) and
Houlihan (2005).
13
See, for example, http://news.bbc.co.uk/1/hi/business/4949606.stm Accessed 14th June
2006.
Historic and Current Policy in Sport: Changing Emphases and Values 61
TABLE 2.3 Scale and distribution of sources of European football club revenues
Club Matchday Broadcasting Commercial % Champions

League
£ e £ e £ e
Manchester United 92.4 61.2 94.5 62.5 72.1 47.8 15

Real Madrid 62.0 41.1 88.1 58.3 85.9 56.9 13
AC Milan 27.9 18.5 134.4 89.0 60.0 39.7 10
Source: Deloitte (2005a).
Indeed, Deloitte (2005b) note that across Europe the major professional
football leagues of the UK, France, Germany, Spain and Italy generated e5.8
billion in revenue in 2003–2004, of which the PL generated e2 billion.These
revenues reflect growth from e1.95 billion in the period from 1995 to 1996.
Coupled with this growth in revenue, wages and player salaries have also
escalated from a total of approximately e1 billion in 1995 to approximately
e3.6 billion in 2003–2004.
Similar developments have, of course, occurred in the US. Whereas
$65 000 captured the rights to broadcast the World Series in baseball to fewer
than 12% of US households in 1947, as Table 2.4 reveals, US professional
sports are now million dollar industries employing million dollar athletes.
Understanding the main determinants of revenue and cost flows for pro-
fessional sports, their changes and how, now justifiably viewed as large-scale
industries, such sports are subject to the scrutiny of economic regulators in the
light of claims made about the management of leagues is of direct interest to
economists. Of course economic activity directly associated with sports is not
just confined to the flows of revenues and costs within elite and professional
team sports. Figure 2.1 shows that, flowing out of activity associated with
sports are derived demands for related travel, equipment, facilities and
TABLE 2.4 US sports leagues
Sport League revenue $ Average player salary $

billion (2005) million (2005)
National Football League 4.8 1.25

National Basketball Association 3.1 4.92
Major League Baseball 4.1 2.5
National Hockey league 2.0 1.81
Major League Soccer 2.4 Not available
Source: Plunkett Research.1
1
See http://www.plunkettresearch.com/Industries/Sports/SportsStatistics/tabid/273/Default.aspx Accessed 15 June 2006.
TABLE 2.5 Size of sports economy
Country Spending e billion % of GDP
France 25.4 1.7

Finland 1.7 0.9
Canada** Not available 1.1–1.22
UK 20 2
US* 77.97 0.78
Source: Carter Report (2005).
*
Plunkett research: only includes equipment, clothing and footwear $ values converted at an average 2005 rate $1.244:
e1, US GDP in 2005 $12.36 trillion.
**
Second figure from The Conference Board of Canada (2005).
clothing. In other words, economic activity directly associated with sports

consumption and production “spills over” to the wider economy. Such effects
are essential to understanding the full economic impact of sports, which are
discussed in various forms throughout the book. Key to this are the multiplier
effects discussed in Chapter 1. For now, however, it is worth noting that the
increasing importance of sport to the wider economy has led to recent
attempts to formalize the economic magnitude of sports through charting
the value of spending as a share of the economy. While there are problems of
measuring such values, Table 2.5 provides some examples of the relative
importance of sport, as an indicator of the relative sizes of the sports economy,
deriving from the sports environment as defined in Figure 2.1. A salutary point
to note from the table is that while sport looms large in people's minds, in
economics terms it is still a small proportion of overall economic activity, at
least as measured. In this regard, it should be noted that it is the rapid changes
in the economic profile of sports as discussed above, coupled with general
interest in sport that, in essence, combine to make the study of the economics
of sport both important and compelling.
2.7 CONCLUSION
This chapter has provided an overview of what is meant by sport, and how
sport is organized internationally. It has been shown that “sport” is not a
uniquely defined activity but, rather, reflects the historical development of
particular activities that have come to be understood, provided and monitored
in various guises. In this regard what is considered to be sport and, more
importantly, how it is organized, can change. It has also been shown that there
are three main segments in the sports economy, although they are integrally
related, and these segments are the main subjects of this book: mass sports
participation; professional team sports; and sports events. It has been shown
Conclusion 63
that in many respects the latter two elite activities have grown out of mass
participation activity and, corresponding with such evolution policy, tensions
and shifts in the economics of the organization of the segments have occurred.
In particular, this is connected with a growth in the commercial provision, or
at least its logic, of sport and competition for public funds to supply sport.
Having outlined the main tenets of economic theory relevant to examining
sports, and having provided an overview of the sports economic environment,
the remainder of this book probes each of these sections in more detail. The
aim is to explore the relevant theory and evidence that yield a fuller under-
standing of these segments and their development.
CHAPTER 3
The Economics of Sports

Participation1
OBJECTIVES
& To understand economic theories of sports participation

& To appreciate the policy implications of economic theories of sports
participation
& To appreciate why market failures could be important in sport
3.1 INTRODUCTION
This chapter begins an examination of mass sports participation as part of a
broader process of sports consumption. Section 3.2 presents an initial “core”
sports consumption model, by drawing on the hierarchical nature of the sports
environment presented in Figure 2.1. Section 3.3 then presents how this
model can be used to understand sport as a consumption of time, and as a
consumption of goods. Section 3.4 presents an integrating framework for
these theories, after which Sections 3.5 and 3.6 review alternative approaches
to analyzing sports consumption, as well as considering the policy implica-
tions of the approaches.
3.2 A GENERAL ECONOMIC MODEL OF SPORTS

CONSUMPTION
Figure 3.1 presents a core economic model of sports consumption. The figure
shows that choice emerges from the interaction between the motives of con-
sumers and the resources they have at their disposal. It is postulated that the
sports consumer's motives will be influenced by their tastes or preferences, as
1
Elements of this chapter draw on Downward (2005). 65
66 CHAPTER 3 The Economics of Sports Participation
well as their objectives, which will imply a set of possible

courses of action for the consumer. The resources at the
consumer's disposal, i.e., their time and income, then
identify feasible courses of action from which consumer
behaviour emerges. Drawing on Figure 2.1, it is assumed
that following choices about the frequency and intensity
of participation, derived demands and expenditure on
travel, equipment, facilities and clothing will follow.
The next three sections outline these decision choices
in more detail, by first of all identifying some basic fea-
tures of economic decisions and then exploring situa-
tions in which the consumer allocates time to
participate in sports, and then allocates financial
resources to consume sports.
3.2.1 The dual decision hypothesis

A basic feature of economic analysis is that aggregate
economic activity constitutes a flow of resources be-
tween economic agents. This is known as the “circular
flow of income.” In the simplest example, the flow of
FIGURE 3.1 Sports consumption.
resources lies between “households” and “firms.”
Households comprise individuals who want to con-
sume products and services, but who also provide work effort for firms
to produce goods. In order to consume, the household needs to work to
earn income. The household demands goods and services, but supplies
labour effort to firms. Likewise, recalling Chapter 1, firms supply house-
holds with goods and services, but demand labour effort as a factor of
production, in this case from households, to facilitate this. The expendi-
ture of households will correspond to the income earned by firms and by
households as employees of firms. Figure 3.2 illustrates the basic principle
and the consequence that aggregate economic expenditure will correspond
to aggregate economic income, or national income, in an economy.2 The
clockwise direction of the flows indicates the transfer of money payments.
The anticlockwise flows indicate the real resources that correspond to the
monetary transactions.
In practice, the circular flow of income is more complicated. There are
leakages, in the form of savings by households, taxation by public author-
ities and imports of goods and services. Each of these, other things being
2
An examination of Appendix 1.2 illustrates this proposition.
A General Economic Model of Sports Consumption 67
equal, will reduce the circular flow of income and the volume of national
income.
Reflection Question 3.1 The circular flow ofincome

Why do imports, e.g., the purchase of overseas players by a football club or an overseas
golfing holiday, reduce the circular flow of income?
Hint: Think about the destination of the expenditures.
The reason why savings and taxation reduce national income is because
“expenditure” is being channelled away from firms, who cannot then use that
income to employ labour from households. In the case of imports it reflects the
fact that expenditures are being directed towards a different economy. Corre-
sponding to these leakages is a series of injections to the circular flow. These
comprise investment by firms, expenditure from public authorities, and
exports. One of the key sources of investment funds for firms is savings,
channelled to firms through the financial system. Thus, the banking system
takes savings from households and lends this to firms to fund their investment
decisions. Among other factors, therefore, the interest rate will make these
activities interdependent.
The interest rate indicates the rate at which additional funds are received by savers
for choosing not to consume their income now, but at some period in the future. It
measures the time value of money. Clearly, as more individuals save it becomes
difficult for financial businesses to pay the interest charges to savers, so they reduce
the rate of interest they will pay. In turn, this will make the cost of borrowing money
to invest less, and thus encourage less saving and more investment based on using
borrowed money and vice versa.
FIGURE 3.2 A simple circular flow of income.

Likewise, taxation is a source of income for public authorities to spend. In this

regard injections to and leakages from the economy are closely related. Exports
and imports, of course, are not directly related as decisions, but nonetheless
their behaviour is related through, among other things, the exchange rate.
The exchange rate is the price of one country's currency in terms of another country's
currency. Importing goods and services requires buying foreign currency and selling
domestic currency to pay for the goods and services. Exporting goods and services
implies receiving revenues in foreign currency which then needs to be sold and
domestic currency bought. Consequently, the exchange rate will change according
to the overall volume of importing and exporting that occurs. This affects the cost of
the transactions, making imports or exports appear cheaper or more expensive
depending on the circumstances, which, in turn, will affect importing and exporting
behaviour.
There has always been debate in economics about whether or not the
circular flow of income automatically balances to ensure full employment
and an efficient use of resources, as defined in Chapter 1. If all economic
activity behaved according to the predictions of the model of “perfect
competition” described there, then the economy as a whole would allocate
resources efficiently. Key to this argument is the efficacy of prices, and the
effectiveness of the market system, to channel resources to their appropriate
uses. In this respect, the interest rate and exchange rate are considered to be
important prices that coordinate the leakages and injections of savings–in-
vestment and imports–exports respectively. In turn, for many economists, it is
recommended that governments should strive to ensure that the public
finances balance and/or impose the smallest tax burden on the economy as
possible. The latter point is justified by arguing that taxation reduces incen-
tives to work. Such issues are of less direct interest for this book, although they
are important in examining the impact of investment in sports infrastructure
and events, as discussed in Chapter 12.
What is important to gain from this discussion is that it highlights that the
sports consumer faces a “dual decision” dilemma. As a member of a house-
hold, if the consumer wants to purchase products and services, including
sports goods and equipment, etc., then the individual will need to work to
earn sufficient income to do so. However, work provides an obligation on the
use of the consumer's time. Leisure, which is desirable, requires undertaking
activities that do not place obligations on the consumer's time. However,
because there is a time limit on the individual's capacity to work or to enjoy
leisure (and indeed consumption generally) this implies a trade-off. The next
section indicates how economists explore this trade-off in more detail, by
examining a theory of rational individual choice.
3.2.2 The income–leisure trade-off

Economists have devised a relatively simple model to explore the dilemma
faced by individuals between working to earn income, and not working and
consuming time as leisure. It is called the income–leisure trade-off model. The
main assumptions of the model follow.
3.2.2.1 Motives
1. Individuals seek to maximize their utility by consuming goods and
services or enjoying leisure time. Utility, as discussed in Chapter 1, is
simply defined as “satisfaction.” In this respect, it describes the
individual's preferences and tastes. It is considered to be a personal
phenomenon and to be ordinal in nature. This means that individuals
can compare two situations only for themselves, and express prefer-
ences over them, i.e., indicate which is associated with more utility,
without being able to say anything about the extent of the difference in
preferences or utility. Formally, this implies Equation 3.1, which indi-
cates that utility, “U,” depends on consumption through income “I,”
and the consumption of leisure time, “L.”
U ¼ UðI; LÞ ð3:1Þ
2. The individual always prefers more of both income and leisure to less of
both. This implies that they are both “normal goods” and that the
individual is willing, if necessary, to substitute more of one for less of
the other. It follows that, if the individual gives up some leisure or
income, then to leave the individual feeling as well off as previously,
more income or leisure is required. It is also usually assumed that if, for
example, the individual was currently consuming a lot of leisure, then it
would take a smaller relative increment in income to persuade the
individual to decrease their leisure by a certain amount, and vice versa.
This is known as diminishing marginal utility,3 and implies that rela-
tively abundant goods or services are of less value to the individual
compared with relatively scarce goods or services.
3.2.2.2 Constraints
1. Individuals are constrained in their activities by the rate of pay, “w” (per
unit of time, e.g., hour, day or month, etc.) that can be earned from work,
“W” and the finite amount of time available to them, “T.” Formally, Equa-
tion 3.2 expresses the relationship that total time is equal to work plus
leisure, and Equation 3.3 that income, “I,” and hence consumption, is
3
This is directly analogous to diminishing marginal productivity, discussed in Chapter 1.
work times the wage rate.
T¼WþL ð3:2Þ
I ¼ wW ð3:3Þ
2. Thus, if the rate of pay, w, is e10 per hour (and noting that there are
24 hours in a day, T), then this means that, in extreme cases, the individual
could:
(a) not work at all, W = 0, and hence receive no income, I = 0, for con-
sumption of goods, but consume 24 hours of leisure, L = 24.
(b) work for 24 hours, W = 24, and receive an income for consumption of
I = e240, but consume no leisure, L = 0.
3. These are, of course, extreme cases and are referred to as corner solutions
by economists. It follows that attention is focused on points in between
these extremes. Thus, the individual could also consume 12 hours of lei-
sure and receive e120 of income to consume goods, or consume 6 hours
leisure and receive e180 income and so on. The time constraint acts to
produce an income constraint, given a particular rate of pay. Equation 3.4
shows that the rate at which leisure can be substituted for income will
reflect the wage rate. The equation indicates a negative relationship to
imply the substitution of income by leisure.
Total income e240

¼ ¼ e10 ð3:4Þ
Total leisure e24
3.2.2.3 Leisure choice

To identify the choice of leisure that an individual makes requires exam-
ination of how the desire to consume more of both income and leisure
becomes constrained by the time constraint. Formally, this implies that the
consumer maximizes utility, as given by Equation 3.1, subject to the con-
straints given by Equations 3.2 and 3.3. Equation 3.5 illustrates the choice
problem formally:
Max UðI; LÞ subject to T ¼ W þ L and I ¼ wW ð3:5Þ

or:
I ¼ wðT LÞ
The choice decision is examined formally in Appendix 3.1. However, the

intuition of the decision is demonstrated in Figure 3.3, which illustrates the
time–income constraint facing consumers and their preferences, as well as the
derivation of demands for leisure time and income.
FIGURE 3.3 The demand for leisure and income.
In this figure, the vertical axis measures income. The horizontal axis
measures the time period available for leisure, for example, 24 hours. The
income constraint is represented by a straight line from 24 hours' leisure
and zero income, to a maximum income of e240 and zero leisure at a wage
rate of e10 per hour. The slope of this line is negative, as implied in Equation
3.1. Combinations of income and leisure along and below this line are feasible
choices for the individual.
Preferences or tastes are represented by the indifference curve. Only
one is drawn, but this should be thought of as one of a set which, just as
contour lines represent sets of coordinates associated with a given altitude
on a hill, represent combinations of income and leisure which yield
the same level of utility. It was assumed above that individuals prefer
more of both income and leisure, therefore indifference curves that are
drawn out and to the right of the origin represent higher and higher levels
of utility.
Why does the indifference curve slope from left to right convex to the origin?
Hint: Examine assumption 2 of the motives for choice.
The indifference curves are drawn sloping from left to right, with
a slope that is convex to the origin. In the first case this is because
both leisure and income are normal goods. This means that one can be
substituted for the other to leave the individual feeling the same level of
satisfaction – as implied by the concept of an indifference curve. In the
second case the slope of the indifference curve is known as the marginal
rate of substitution (MRS) of leisure for income (or vice versa). The
changing value of the slope is connected to the diminishing marginal
utility received from consuming more and more of one good or service
and less and less of another. Consequently, beginning from a situation in
which the individual was towards the top of the indifference curve, higher
income and lower leisure are demanded. As leisure is relatively scarce
with respect to income, the marginal utility of leisure is relatively higher
than the marginal utility of income. Thus, if the individual moved down
the indifference curve, i.e., reallocated their demand, relatively more in-
come could be given up than leisure time increased, but the individual
would feel equally satisfied. As the individual moves further and further
down the indifference curve, indicating greater and greater demand for
leisure, the marginal utility of leisure time falls coupled with a rise in the
marginal utility of income. Consequently, less and less income will be
given up to compensate for increased leisure. More specifically, as indi-
cated in Equation 3.6, this implies that the marginal rate of substitution
of leisure for income is equivalent to the ratio of the marginal utility of
income and the marginal utility of leisure. In general, the indifference
curve summarizes the rational planned consumption possibilities of the
individual.
Marginal rate of substitution

¼ Marginal utility of income ð3:6Þ
Marginal utility of leisure
3.2.2.4 Equilibrium choices of income and leisure

Combining the indifference curves and the income constraint enables us to
identify conceptually, although qualitatively, the individual's rationally de-
fined demands for income and leisure time. The logical point of utility
maximization is where an indifference curve is tangent to the income–
time constraint, as indicated on the figure. As indifference curves cannot
cross – because they represent particular levels of utility – moving to the
right or left of this point in reallocating leisure and income demands must
force the individual onto a lower level of utility. Technically, this tangency
implies that the consumer will be willing to substitute marginal incre-
ments in income against leisure at a rate equal to the wage rate. This is
an important result, as it indicates that the wage rate measures how much
value is placed on an hour of time by the individual. The wage rate is the
opportunity cost of leisure. Equation 3.7 summarizes this relationship and

shows that by equating Equations 3.6 and 3.3 an extra hour of leisure will
reduce income by e10.
Marginal utility of income
¼ e10 ð3:7Þ
Marginal utility of leisure
3.2.2.5 Predictions from the model

The above analysis is clearly abstract and, perhaps, appears to have no
obvious relevance to understanding the demand for leisure. This is
certainly true of the model's description of decision making. However,
recalling the discussion of economic methodology from Chapter 1, the
model should be understood as a basis from which to make predictions.
This theoretical objective lies at the core of much economic analysis. Let's
assume that the wage rate increases. Wage rates have, for example, in-
creased in most advanced economies over time. What are the implications
for the demand for leisure? The income–leisure trade-off predicts that two
effects will be set in motion.
3.2.2.6 The substitution effect

As wage rates increase, thinking about the idea of equilibrium, this suggests
that the valuation placed on an hour of time has increased. This means that
the price of leisure has increased. As a normal good, this must mean that
the demand for leisure will fall and the demand for income through work
rise correspondingly. Of course, the opposite will happen with a fall in the
wage rate.
3.2.2.7 The income effect

On the other hand, as wage rates rise, it is now possible that undertaking the
same level of work is still associated with a higher income. Consequently,
more of both leisure and income can be achieved. It follows that, as a normal
good, at least some more leisure time might be demanded. Once again, the
opposite would apply with a fall in the wage rate. It is clear, therefore, that the
net effect of these tendencies will produce the new demands for leisure and
income. If the substitution effect dominates, less time for leisure will be
demanded. If the income effect reflects a normal good then more leisure will
be demanded. This can lead to the supply of labour becoming “backward
bending,” as wage increases lead to less and less labour being supplied. Ulti-
mately what happens depends on the individual's preferences, and is an em-
pirical question. The empirical support for such effects is discussed in the next
chapter.
FIGURE 3.4
Income and substitution
effects in leisure
demand.
In terms of the graph, one way to illustrate the theoretical nature of the two
effects is by establishing what the new equilibrium point would have been for a
particular set of indifference curves, i.e. preferences. Figure 3.4 illustrates the
changes in demand that occur, assuming that the income effect partially off-
sets the substitution effect.
An increase in the wage rate rotates the income constraint clockwise
around its intercept with the horizontal axis. This is because, while the
numbers of hours in the day remain constant, each hour's work now
produces more income. This implies that the consumer's income–
time constraint has relaxed, and higher utility can be achieved from real-
locating their choices to new leisure and income demands. These will be
identified by a new tangency on the highest indifference curve that is now
possible. It is clear that in this case; overall more leisure is demanded,
as well as some extra income. Arriving at the new leisure demand is,
however, determined theoretically by the net outcome of the substitution
and income effects.
Constructing a hypothetical income constraint tangent to the original
indifference curve, but at a slope corresponding to the new wage rate, will
establish a further hypothetical point of equilibrium on the original indiffer-
ence curve. The movement from the original equilibrium to the hypothetical
equilibrium identifies the substitution effect. The movement from the hypo-
thetical equilibrium to the new actual equilibrium identifies the income effect.
This makes it clear that the substitution effect is essentially a reallocation of
demand purely as a result of a change in the relative price of leisure, whereas

the income effect identifies the change in demand following the enhanced
income from a price change.4 The substitution effect measures the “Hicksian”
demand, named after the economist John Hicks. The combined substitution
and income effect measures the “Marshallian” demand, named after Alfred
Marshall. Overall, the above analysis describes a simple general demand func-
tion for leisure time as implied in Equation 3.8:
DL ¼ DL ðw; PÞ ð3:8Þ
where leisure demand, (DL), will depend on the opportunity cost of leisure, or
wage rate (w) and the tastes and preferences of the individual (P). Theory
suggests the broad prediction that a rise in the wage rate will normally lead
to a fall in the demand for leisure, following the substitution effect, but
produces an ambiguous overall effect when one allows for the income effect,
assuming that leisure is a normal good. Figure 3.4 above implies that overall
demand for leisure rises. It is possible that overall demand falls, if the income
effect does not increase demand enough to offset the substitution effect or
because the income effect actually reduces demand and reinforces the substi-
tution effect.5 This latter case would apply if leisure was an inferior good. The
opposite effects could apply with a wage rate fall.6
4
Indifference curves are not observable, because they are maps of individual preferences.
This means that empirically measuring income and substitution effects, as described in
theory, is impossible. However, to overcome this problem, economists often make use of the
“Slutsky approximation.” This implies viewing the income effect as the amount of income
required not to place the individual at a point of tangency on the original indifference curve
which cannot be observed but, rather, at the level of income that is required to make the
previous choices affordable at the new relative prices. Comparing the choices made by the
consumer would reveal the substitution effect. This would modify the diagram in Figure 3.4
by moving the hypothetical income–time constraint so that it passed through the original
equilibrium choice. Logically, this would mean that the hypothetical point of tangency,
which identified the new choices following the substitution effect, would also be on a new,
slightly higher indifference curve. It is in this regard that the Slutsky analysis is an
approximation. By comparing the value of choices at the original equilibrium, with regard to
the original and subsequent prices, indicates the extent to which real income has changed.
5
In the absence of the consumer having access to unearned income, it can be argued that the
relationship between the decision to engage in leisure and the wage rate is likely to be
unambiguous. Higher wages will unambiguously make leisure more costly. The ambiguity
arises in the relationship between the hours of leisure and the wage rate, as it is here that a
certain level of income can help to anchor consumption decisions.
6
The astute reader will perhaps be aware that this suggests that preferences and tastes are
clearly linked to the concept of goods being “normal” or not. As discussed later in the chapter,
this raises issues about the role of tastes as independent influences on sports demand.
3.2.3 The derived demand for sports

As discussed above, it can be argued that the structure of sport demands is
hierarchically linked. In this regard, while the demand for leisure does logically
imply some reduction in available income for the consumption of other goods
and services, nonetheless an element of the goods and services that are con-
sumed will reflect the prior participation decision. For example, consumers'
tastes and constraints determine a feasible desire to participate in a leisure
activity, say, keep fit. The act of participation calls forth a set of derived
demands for equipment, such as exercise mats, appropriate clothing, for ex-
ample flexible Lycra wear, as well as facility demand such as room space, sound
system and so on. Likewise, the demand to participate in hill walking as a
leisure activity will involve a demand for travel to access the appropriate venue,
such as a national park, as well as demands for clothing, footwear and possibly
footpaths. Clearly, therefore this hierarchical structure of demands is a useful
organizing schema.7
To understand the economic logic of the behaviour of such derived demand,
and indeed the demand decision in economics generally, involves a straight-
forward modification of the principles already presented in the income–leisure
trade-off model. Complications can be added, as discussed in Appendix 3.2, to
reflect the durability and multiple uses of goods. However, for now, the main
assumptions of that model are simply restated as follows:
3.2.3.1 Motives
1. Individuals seek to maximize their utility “U” by consuming goods and
services, x1 and x2. Formally:
U ¼ Uðx1 ; x2 Þ ð3:9Þ
7
There are, of course, complications associated with this simple model. The main one is
that it implicitly assumes a one-way causality from participation to demands for equipment,
etc. This naturally follows from Marshall's original analysis. In the case of sport and leisure,
however, it could be argued that matters are more complicated. For example, it is well known
that sports clothing has fashion value. This means that the demand for sports clothing can
be entirely unrelated to the level of participation in sports or leisure activities. Likewise, if
one thinks of the leisure demands associated with tourism it may well be that travel is an
integral part of the main “product,” rather than a derived demand reflecting the need to
travel to consume the tourism product. This, of course, does not necessarily undermine
Marshall's analysis. The problem arises in the above examples because of the varying degrees
by which demands are considered to be “separable” or “joint” respectively. These are issues of
which Marshall was well aware and, in practice, show that the definition of specific markets
and products will be of paramount importance.
2. The individual always prefers more of goods and services to less of both.
This implies that they are “normal goods” and that the individual is
willing, if necessary, to substitute more of one for less of the other.
Diminishing marginal utility also applies, so that relatively abundant
goods or services are of less value to the individual than relatively scarce
goods or services.
3.2.3.2 Constraints
1. Individuals are constrained in their consumption activities by the income
that they have earned, I, and the price of the goods and services that they
wish to consume, p1, p2. This implies:
I ¼ p1 x 1 þ p 2 x 2 ð3:10Þ
2. Thus, if the consumer has earned I = e200 in the week and the cost of a
pair of training shoes, x1, is e40, i.e., p1, this means that, in extreme cases,
the individual could:
(a) spend all of their income on five pairs of training shoes;
(b) retain all of their income for consumption on other goods and services
and spend nothing on training shoes.
3. Consider that sports vests, x2, also cost e10, i.e., p2. In this regard the
consumer also has the choice to:
(a) spend all of their income on 20 vests.
(b) retain all of their income for consumption on other goods and services
and spend nothing on vests.
FIGURE 3.5 Equilibrium demands for vests and training shoes.

4. Consequently, if we restrict our attention to the consumption of vests and

training shoes, to facilitate a graphical analysis, the consumer faces a
budget constraint implying options to:8
(a) spend all of their income on five pairs of training shoes and no vests.
(b) spend all of their income on 20 vests and spend nothing on training
shoes.
As implied in Equation 3.10. As with the income–leisure trade-off model,

however, these are corner solutions to the consumer choice problem and, given
the assumptions about utility, unlikely. Figure 3.5 illustrates a more appro-
priate equilibrium choice of vests and training shoes. An indifference curve is
drawn to reflect preferences of the consumer over alternative consumption
combinations of training shoes and vests. The slope and shape of the indif-
ference curve reflects the assumptions made about preferences, as discussed
above and in connection with the income–leisure trade-off model.
The budget line is drawn as a straight line connecting the extreme
feasible choices facing the consumer, as identified in point 4 above. Each
of these extreme possibilities of consumption are given by dividing the in-
come available to the consumer, that is e200, by the price of vests or
training shoes, that is e10 or e40. Consequently 20 vests or five pairs of
training shoes could be purchased. It is important to recognize that the slope
of the budget line is given by the ratio of these prices. The slope is negative,
of course, as increases in income allocated to training shoes must be met be
by a reduction in income allocated to vests. Equation 3.11 formally states
this proposition.
8
It is possible to consider the consumption of all other goods and services as the alternative
to the consumption of either vests or training shoes. This would require treating the other
goods as a composite commodity. Technically, this implies assuming that the relative prices
of the commodities within the group remain fixed. If this is the case, then the quantities of
each of the goods and services can be valued in the aggregate, using a price index in which the
prices of the individual goods and services are proportionately related. This is an important
concept as it suggests that groups of goods and services can be separated, that is preferences
over them are independent of preferences for other groups of goods and services. In this
regard, the goods would not be substitutable and the market would be segmented. Of course,
the relative value of the composite commodity and vests and training shoes will change if
prices change in the latter. Consequently, the decision process is more complex. Here
consumers must allocate incomes to particular batches of goods and services, and then
reallocate these subcomponents of income to goods and services within the segment. An
optimal allocation of income to each subgroup of demands implies that the marginal utility
of income is equal in each case. Appendix 3.1 outlines elements of the consumer choice
problem more technically. The marginal utility of income is identified in this exposition as
the Lagrange multiplier l.
Income allocated to vests 200=10 40

¼ ¼ ð3:11Þ
Income allocated to training shoes 200=40 10
This equation indicates that the rate at which income is transferred be-
tween vests and training shoes, as one is substituted for another, is governed by
the prices of training shoes and vests. This is, of course, eminently logical, as
training shoes cost four times as much as vests. So, each pair of training shoes
is equivalent to four vests. As seen in the income–leisure trade-off model, the
consumer will maximize utility where the indifference curve is tangent to the
income constraint. Formally this implies:
Max U ¼ Uðx1 x2 Þ subject to I ¼ p1 x1 þ p2 x2 ð3:12Þ
As the slope of the indifference curve, the marginal rate of substitution, is

equal to the ratio of the marginal utilities of vests and training shoes, this
implies that in equilibrium the consumer chooses to purchase vests and
training shoes up to the point at which the ratio of marginal utilities is equal
to the ratio of prices. This is an important result as it suggests that, in general,
the subjective utilities attached to goods and services are measured by their
market prices.
3.2.3.3 Predictions from the model

As with the income–leisure trade-off model, predictions of demand can be
made. If, for example, the consumer decides to reduce the amount of leisure
that they wish to consume and consequently increase their income, then this
would shift the budget constraint to the right, as implied in Figure 3.6. The
intuition of this change is given by considering that that numerator of the
extreme possibilities both increase, meaning that more vests or training shoes
could be purchased. A higher indifference curve is now attainable for the
consumer, so, assuming that vests and training shoes are normal goods, more
of both goods are likely to be purchased, as indicated by the arrows that
indicate an increase in the demand of both goods following a move to a new
equilibrium position.
This suggests that demand for sports equipment, facilities and so on, is
expected to vary directly with consumer income. If the price of one of the
goods changes, then for analogous reasons to the price of leisure, i.e., the
wage rate changing in the income–leisure trade-off model, an income and
substitution effect will take place. Figure 3.7 illustrates this possibility for
a decrease in the price of a pair of training shoes. Note here that, follow-
ing the price reduction, the budget line pivots out to the right around
the extreme value of possible vest consumption. This is because the fall in
price of training shoes makes it possible to consume

more pairs, but if no training shoes were purchased
then only the same maximum number of vests could
be purchased. The income and substitution effects are
once again identified by constructing a hypothetical
budget line, with the same slope as the new price
ratio, but tangent to the original indifference curve.
The movement between equilibrium points on the
original indifference curve identifies the substitution
effect, that is to say, the change in demand purely
FIGURE 3.6 The effects of higher income on following from a change in prices. The slope of the
consumer demand. indifference curve implies that demand increases with
a price fall and vice versa. The income effect is then
identified by examining the move from the hypothetical equilibrium on
the original indifference curve implied by the result of the substitution
effect, to the new equilibrium on the new indifference curve. Analogous to
Figure 3.6, the shift between equilibrium points is identified by the par-
allel shift in the budget constraint, indicating that incomes have risen but
prices remain constant. Conceptually, this identifies the income that is
gifted to the consumer simply because the price of one of the goods has
fallen. Because the same number of pairs of training shoes could be
bought now for less income, this leaves additional income for the con-
sumer to allocate between other purchases, as well as training shoes.
The figure is drawn consistent with the assumption that vests and train-
ing shoes are normal goods, to indicate that both the substitution effect
and the income effect reinforce one another to generate extra demand for
training shoes following their fall in price.
FIGURE 3.7 The effects of a price change on consumer demand.

FIGURE 3.8 A theoretical demand curve for training shoes.
The above analysis suggests that the demand for sports equipment, clothes
and facilities (x) can be summarized in Equation 3.13:
x ¼ xðp; p*; I; PÞ ð3:13Þ
Demand is likely to increase or decrease with decreases or increases in the

price of the good or service, i.e., p; increase or decrease with an increase or
decrease in the price of other goods or services, p* because of substitution, and
increase or decrease with increases or decreases in consumer incomes, I.9 Such
predictions are dependent on consumer tastes and preferences, P, remaining
constant.
The basic predictions from this analysis are illustrated in Figure 3.8, where
a demand curve for training shoes is drawn. Changes in the price of training
shoes, other things being equal, will cause demand to increase or decrease by
movements along the curve, whereas changes in income or preferences, or the
price of other goods will, for each change other things being equal, cause the
demand curve to shift to the left or right as fewer or more training shoes are
demanded at any given price respectively.
9
This assumes that goods are essentially gross substitutes. Goods could be complementary,
in which case changes in the price of one of these goods will have the opposite effect on the
level of demand for other goods. One might argue that tennis racquets and tennis shoes are
complementary, as both are required to play tennis and changes in the price of one item will
affect the demand for tennis and thus the demand for the other product. The substitution or
complementarity of equipment demands is thus likely to be mediated by a change in
activities, or adjustment to the level of the activity respectively.
3.3 AN INTEGRATING FRAMEWORK FOR CONSUMER

CHOICE
The above analyses provide the traditional economic analysis of choice. How-
ever, there are a number of limitations with these models, as made transparent
by the economist Gary Becker in what has become known as “New Household
Economics.” As Downward and Riordan (2007) note, in Becker (1965) the
allocation of time is explicitly integrated into the consumption decisions of
individuals. The individual, in turn, is explicitly theorized to be part of a
household. This results in the distinction between consumption and produc-
tion being removed. In contrast, the resources of “time” and “market goods”
combine in “household production” to generate the basic commodities that
yield utility from consumption. As all economic activities involve time and
other goods, purchased via markets, economic agents essentially make choices
involving the relative intensity of these inputs in both producing and consum-
ing commodities. The basic commodities that are consumed, Z, are thus
functions of goods bought on markets, x, and time, t. Formally household
production is implied in the following production function:
Z ¼ Zðx; tÞ ð3:14Þ
Utility, moreover, is then a function of the consumption of these produced

goods and services:
U ¼ UðZ 1 ; Z2 . . .Þ ð3:15Þ
Consequently, Equation 3.14 is an additional constraint that the individ-

ual faces when consuming goods and services along with Equations 3.2 and
3.10. Appendix 3.1 illustrates the implication of these changes more specifi-
cally, but the logic of the changes is to make the choice problem one of
maximizing utility, which depends on the “basic” goods and services, subject
to a resource constraint also based on these.
Max U ¼ UðZ1 ; Z2 . . .Þ subject to GðZ 1 ; Z 2 Þ ð3:16Þ
Figure 3.9 illustrates the main implications of Becker's analysis. If Z1 is a

relatively time intensive activity, say golf, and Z2 a relatively goods intensive
activity, say weight training, then an increase in the wage rate “w” would make
the cost of golf relatively more expensive and generate a substitution of more
weight training for golf, because the resource constraint would change slope.
The resource constraint depends, in part, on the relative price of the commod-
ities and also on the relative cost of time used to produce the commodities.
A significant aspect of Becker's work is that in the main it confines analysis
An Integrating Framework for Consumer Choice 83
FIGURE 3.9 A change in the relative price of good- and time-intensive commodities.
to “substitution” effects resulting from the change in the relative price of

resources, including time. If the consumer experienced an increase in un-
earned income, then a parallel shift in the resource constraint, analogous to
Figure 3.6, would take place, but this would be due to exogenous changes to the
resources at the consumer's disposal.
Does the New Household Economics approach imply an income–leisure trade-off?
Hint: Think about what the cost of leisure must be.
Becker, specifically, explores the relationship between this analysis and the
traditional approach. He writes:
“What, then, is the relation between our analysis, which treats all
commodities symmetrically and stresses only their differences in relative
time and earning intensities, and the usual analysis, which distinguishes a
commodity having special properties called “leisure” from other more
commonplace commodities? It is easily shown that the usual labour–leisure
analysis can be looked upon as a special case of ours in which the cost of the
commodity called leisure consists entirely of foregone earnings and the cost
of other commodities entirely of goods.” (Becker, 1965, p. 98)
More generally Becker (1976, 1992) emphasizes that his approach is not
defined by the specific nature of the material under investigation, but rather its
method – the optimizing behaviour of individuals. Consequently, economic
agents maximize welfare, as perceived by them, subject to income, time,
information and other limiting resource constraints. It is assumed that eco-
nomic agents exhibit stable preferences, which therefore should not be used to
explain demand changes, as implied in Equation 3.13, and that in addition to
markets or arenas of formal contractual exchange, as typically identified in

economics, the social structures in which agents operate, such as households,
act as if they are markets to allocate resources according to their shadow prices,
i.e., opportunity costs.10
An important dimension of this approach is that, as elaborated in Becker
(1974), individuals can allocate time and market goods to invest in personal
capital, skills and capabilities, or social capital and reputation which provide
the greatest return for the household. This implies that individuals invest in
their human capital (see Box 3.1). In this regard, preferences become partially
endogenous to the analysis. As Downward and Riordan (2007) note, this has
obvious implications for the demand for sport. On the one hand, participation
in sports activities requires the acquisition of consumption skills. This sug-
gests that previous consumption in the same activity can increase current
consumption of the same activity. Likewise, consumption of one sports activ-
ity may increase the consumption of another sports activity because of, for
example, implicit skill transfer, such as agility, timing, hand-eye coordination,
etc. There is an obvious rationale, therefore, for expecting the demand for any
one sport to be positively related to the demand for other sports in the context
of personal consumption capital investment (Appendix 3.1 illustrates these
propositions).
In summary, therefore, this new economic framework suggests that utility
maximization can embrace a variety of specific objectives, not simply the
consumption of physical resources, and can easily apply to sports participation
and expenditure on sports goods and equipment. While equipment demand
may be derived from participation, it is also essentially maintained that both of
these are derived from investment in the stock of personal, social or health
capital of individuals. The price of resources is likely to be the main influence
on demand, as socio-economic factors such as income levels, health and
education are also, at least partially, the outcomes of decisions by consumers.
10
In this sense, naive views that economics is concerned only with material objects or
formal exchange and contract are rejected. Moreover, as far as interpreting empirical work is
concerned, there are a number of interrelated points to note regarding Becker's analysis.
Results from econometric work on any given sample need to be interpreted carefully. On the
one hand, the measured income and price effects may mask the true resource allocation
issues at stake. This is because they only crudely capture the true shadow prices of economic
allocation. On the other hand, the usual proxy variables for tastes, such as socio-economic
characteristics, should be viewed at least partially as the results of agents' decisions. It
follows that variations in their impact should not necessarily be seen as evidence of unstable
preferences, but potentially resource adjustment over time in line with decision making
based on stable preferences.
An Integrating Framework for Consumer Choice 85
BOX 3.1 FORMS OF CAPITAL

A number of versions of “capital” and investment in capital & social capital, which reflects relationships and networks
are important for understanding sports participation. Gary of influence of either a formal institutionalized or infor-
Becker has been instrumental in developing the concept of mal mutual acquaintance nature.
human capital in economics, distinguishing it from tradi-
Social capital has also been explored by Putnam (2000) as
tional economic concepts. He writes:
the value associated with being a member of a social net-
“To most people capital means a bank account, a hundred work. A distinction is made between bonding and bridging
shares of IBM stock, assembly lines or steel plants in the social capital:
Chicago area. These are all forms of capital in the sense that
& bonding social capital refers to social networks of rela-
they are assets that yield income and other useful outputs
over long periods of time. tively homogenous individuals;
& bridging social capital refers to social networks of rela-
But these tangible forms of capital are not the only ones. tively heterogenous individuals.
Schooling, a computer training course, expenditures of
medical care, and lectures on the virtues of punctuality and Significantly, also, consumers may invest in characteristics
honesty also are capital. That is because they raise that make them more similar or less similar to other persons.
earnings, improve health, or add to a person's good habits This is suggestive of the establishment of “lifestyles” and
over much of his lifetime. Therefore, economists regard distinct consumer groupings – reflecting social capital for-
expenditures on education, training, medical care and so mation. This suggests that socio-economic characteristics
on as investments in human capital. They are called human may reflect the outcome of choices, which is important in
capital because people cannot be separated from their interpreting the empirical evidence on participation, dis-
knowledge, skills, health or values in the way they can be cussed in the next chapter. For current purposes, however,
separated from their financial and physical assets.” what is important to note is that once utility depends on
previous consumption or the consumption of others,
Gary Becker The Concise Encyclopaedia of Economics, through interdependent preferences, then it is possible to
http://www.econlib.org/library/Enc/HumanCapital.html explain why participation takes place in particular groups of
Accessed 7 January 2007. activities and by particular groups of individuals. Signifi-
cantly, it is possible to extend this analysis to cover invest-
From a broader intellectual perspective other forms of cap-
ment in health. Grossman (1972a, b), provides a seminal
ital exist. Pierre Bourdieu has been particularly influential.
analysis that augments the analysis of Becker to show that
Bourdieu (1986) distinguishes between:
the demand for healthcare is derived from an investment in
& economic capital, which is the traditional resources ex- the stock of health. Viewing the demand for sports or sports
plored by economists, such as land and other assets; participation as a demand for healthcare naturally sees it as
& cultural capital, which reflects the knowledge and skills deriving from a broader concern to accumulate health. It is
possessed, typically, endowed by parents and education. likely that, empirically-speaking, health will affect sports
It might comprise, linguistic acumen, understanding or participation, although care needs to be exercised in inter-
possession of works of art or academic attainments; preting cause and effect.
3.4 POLICY IMPLICATIONS

Chapter 1 indicated that the main economic policy proposition is that the
market allocation of resources in sport is considered to be, in a normative sense.
Consequently, mass participation should be derived from the exercise of freely
choosing individuals, and provided for by the private sector. From this theoret-
ical perspective, the only possible grounds for active policy seeking to promote
participation or to provide facilities for participation might lie in arguments that
the market will fail because of the presence of monopoly, an overt ethical
concern with the equity of access to participation opportunities or the presence
of externalities or public goods that need to be corrected, or accounted for.
The analysis of the decision to participate above is predicated on these not
being central problems and this raises problems for policy makers who might,
as discussed in Chapter 2, want a theoretical rationale for actively seeking to
intervene in the economy to raise participation levels. To begin with, and
discussed in detail in Chapter 5, it is difficult to argue that monopoly supply
of mass sports participation opportunities exists. While the history of mass
sports provision has been, for example in the UK, one of public sector provi-
sion and subsidy, this has recently shifted towards a more commercial basis,
with a lesser monopolistic character. Inevitably this carries with it an argu-
ment that the costs of participation have increased rather than fallen, reflect-
ing a degree of public subsidy for sports provision being removed. From a policy
perspective this is a de facto indication that current policy thinking is not as
concerned with equity, or that externalities or public goods are viewed as
general characteristics of sport.
Moreover, the theories discussed above broadly assume away concerns
with equity, externalities and public goods. On the one hand, the utility
maximizing choices made reflect the individual's preferences and desires,
which are assumed to be fixed (by the agent). This suggests that policy
aimed at directly manipulating preferences is unethical and inconsistent
with an individual's right to choose. On the other hand, while policy could
target “income constraints” on participation, there are subtle but compel-
ling theoretical arguments against such an initiative. The analysis above
has presented the equilibrium choice of the consumer in isolation. This
might be referred to as a partial equilibrium approach. Yet, as Section 3.2.1
indicates, such decisions are really part of an interconnected set of deci-
sions in a general equilibrium. Here, the logic of free choice carries over to
the individual simultaneously determining their income levels through the
choice to work more or less hours. Indeed, the logic of the New Household
Economic approach suggests that other socio-economic characteristics are
also, at least partially, chosen by the agent, and in this sense do not, of
Policy Implications 87
necessity, provide a basis on which active policy promoting participation

should be pursued.
In addition to these normative concerns, the practical impact of policy is
challenged by the New Household Economics approach. This is because
Becker (1974) presents a version of the Coase theorem, which was discussed
in Chapter 1. The Coase theorem states that when an externality affects few
parties, and if property rights are well-defined, then bargaining can promote
economic efficiency because externalities become internalized, thus forcing a
market solution to the resource allocation problem. This theorem can apply in
the New Household Economic approach under the heading of the “Rotten Kid”
theorem as part of the framework in which investment in characteristics, i.e.,
personal or social capital, is analyzed. Here consumption decisions are inter-
dependent because of interdependent preferences.
The Rotten Kid theorem suggests that any interdependency of utilities
between consumers will not affect the consumption patterns or welfare of
either in the face of a redistribution of resources between them. It essentially
argues that policy will have an invariant outcome. In the context of sport, this
situation might arise when a policy maker's utility depends on the utility of
policy recipients. Consequently, a policy maker might receive benefit, i.e.,
political satisfaction, for raising participation levels of policy recipients. How-
ever, any transfers of income between the policy maker and the recipient, for
example to facilitate sports participation, will not affect the consumption or
welfare of either. This is even if the policy maker intends to enhance the
welfare of the recipient and the latter does not share this desire.
The reasoning behind such a conclusion is as follows. If the transfers of
income to recipients require tax revenue to fund them, the future tax burden
on recipients would be rationally taken account of by utility maximizing
individuals. Consequently, their current behaviour would account for this
possibility already. Hence, the current situation cannot be improved on. Like-
wise, even if transfers were made conditional on the consumption of specific
activities, for example, to promote consumption of sport because it was felt to
be a merit good, incentives exist to undermine this policy.
A merit good exhibits positive externalities, but will be under-consumed because the
individual's decision does not take account the benefits that further consumption
would have for society. The positive externalities thus need to be internalized to
increase efficiency.
If the recipient did not want to spend the transfers on the targeted activity,
the transfers would clearly be worth less to them than to the policy maker. This
might result in further reductions in consumption of what is perceived to be an
inferior good. This might in turn undermine the incentive to give transfers.
The upshot of these discussions is that there can be no strong rationale for
active sports policy other than ensuring equal access to information about
sports for the population to allow agents to make free choices. If these choices
are deemed undesirable to policy makers then, so be it! A theoretical motiva-
tion for policy activism, therefore, requires an alternative theoretical analysis.
This could be found in heterodox economics, which challenges elements of
the methodological and theoretical basis of the approaches discussed above.
Downward (2007) and Downward and Riordan (2007) provide an overview of
these approaches and argue that three main contributions are relevant: psy-
chological theories of consumption; post-Keynesian theory; and sociological
theory. An important feature of these approaches is that they focus more on a
descriptive elaboration of the processes by which participation decisions are
made, rejecting rational choice under perfect information. Theoretically, a
direct interdependence between preferences that is the basis of agent choice,
and consumption activity of both the agent and other agents is also suggested
as opposed to relying on stable preferences that are in essence independent of
the choices made. These are important distinctions to note because they
provide a rationale for policy activism.
If one views consumption activity as emerging from processes under
which preferences can actually change as a result of a new opportunity to
participate in previously unfamiliar activity, and that constraints face
voluntary action, for example because agents do not possess optimizing
capability or particular characteristics and social circumstances, such as
income differentials, act to exclude consumption opportunities, then active
policy becomes an option. It could target both the constraints and agent
choice in seeking to promote greater participation. Obvious examples
would be to ensure that facilities are available to all, preventing exclusion
on personal or social criteria through legislation, coupled with the flow of
resources to support areas in which choices are desired but not attainable,
for example because of economic underdevelopment or cultural restraint. It
is important to be aware of these alternative perspectives, as discussed in
the next chapter. The empirical evidence does not present a clear arbiter on
these issues, although common general factors associated with participa-
tion do emerge. Policy prediliction can thus be rationalized, to some ex-
tent, by a prior theoretical predilection.
3.5 CONCLUSION
This chapter has begun an examination of mass participation in sports by
discussing various economic theories of choice. Sports participation has
Conclusion 89
been rationalized as part of a broader process of sports consumption in

which time is allocated to participation in activities and then expenditures.
Following this discussion various critiques of, and complications with, the
basic model have been reviewed. These include: a review of theories that
challenge the motives and choice constraints implied in the model; the
implied direction of influences on choices; and, consequentially, the impli-
cations of the choices that are made. Based on this discussion it is argued
that support for policies that promote mass participation can be found.
Consequently, it is argued that awareness of the alternative approaches to
mass participation is important for suppliers of sport in both commercial
and non-commercial settings. The next chapter examines these issues
further by exploring the empirical evidence on sports participation.
Appendix 3.1 Utility-Maximizing Behaviour

In this appendix the basic elements of solving constrained optimizing problems, as
implied in the choice problem of neoclassical economics, are presented. The Lagrange
multiplier method is most often used to solve these problems. It is named after Joseph
Lagrange, the French mathematician who developed the method. The method implies
taking a constrained optimization problem and converting it into an unconstrained
optimization problem. This is achieved by producing a linear combination of both the
objective function and the constraint, such that their values are equal, for a given value
of the constraint l, which is known as the Lagrange multiplier (LM). For any given value
of the LM, differentiating the optimization problem and solving for the value of the
variables when the differential is equal to zero will identify the optimal solution, i.e., when
infinitesimally small changes in the variables will not affect the value of the function
being optimized. Technically speaking, this produces “first order conditions” (FOC) for
the optimum. Therefore, a maximum or minimum solution might be identified. In utility
maximization, however, assuming that the utility function exhibits diminishing marginal
utility, i.e., the utility function is quasi concave, is sufficient to produce a maximum
solution. Diminishing marginal productivity would imply maximum profits in the case of
the constrained optimization problem in the case of firms' profit maximizing. Therefore,
in the case of the income–leisure model the choice problem implied in:
Max UðI; LÞ subject to T ¼ W þ L and I ¼ wW ðA3:1:1Þ
or:
I ¼ wðT LÞ
Would become:
Z ¼ UðI; LÞ lðI wðT LÞÞ ðA3:1:2Þ

FOCs would be:
LZ=LL ¼ LU=LL lw ¼ 0 or LU=LL ¼ w

LZ=LI ¼ LU=LI l ¼ 0 or LU=LI ¼ l ðA3:1:3Þ
LZ=Ll ¼ I wðT LÞ ¼ 0 or I ¼ wðT LÞ
The first equation shows that the marginal utility of leisure is equal to the wage rate, as
discussed in the text. This means that the value of leisure is implied by the wage rate.
The second equation shows that the marginal utility of income is implied in the LM. This
is the “shadow price” of income, reflecting how utility would change following a change
in the budget constraint implied in the final equation. The optimal conditions are thus
conditional on a value of the LM. In the more general case of the consumer choice
problem:
Max U ¼ Uðx1 ; x2 Þsubject to I ¼ p 1 x1 þ p2 x2 ðA3:1:4Þ
Would become:
Z ¼ Uðx1 ; x2 Þ lðI p1 x1 p2 x2 Þ ðA3:1:5Þ
FOCs would be:
LZ=Lx1 ¼ LU=Lx1 þ lp1 ¼ 0

LZ=Lx2 ¼ LU=Lx2 þ lp2 ¼ 0 ðA3:1:6Þ
LZ=Ll ¼ ðI p1 x1 þ p2 x2 Þ ¼ 0
An important feature of this solution is that the first two FOCs imply:
LU=Lx1 =LU=Lx2 ¼ p1 =p2 ðA3:1:7Þ
This shows that the ratio of marginal utilities of consumption of goods and services is equal to
the ratio of their prices, as implied in Figure 3.5.
Solving for the specific demand functions associated with a utility maximizing problem
requires a specific form of utility function. A common function that is often used is the
Cobb–Douglas function, where utilities are increased through consumption of goods and
services according to parameters which appear as exponents. Other forms of utility
functions are often used. They impose particular restrictions on the patterns of demand
that can be derived, which can then be tested. In general, however, they all retain the
essential property that utility increases at a diminishing rate following the consumption
of goods and services. Equation A3.1.5 can be written using a Cobb–Douglas utility
function as:
Z ¼ xa1 x b2 lðI p1 x 1 p2 x2 Þ ðA3:1:8Þ

Conclusion 91
and the following derived:
x1 ¼ ða=a þ bÞI=p1 ðA3:1:9Þ
This equation, known as a Marshallian demand curve after Alfred Marshall, shows that the
demand for x1 will rise with consumer income “I” and fall with the price of x1, i.e., p1. An
analogous solution procedure would produce a demand for x2 as:
x2 ¼ ðb=a þ bÞI=p2 ðA3:1:10Þ
The New Household model of consumer demand can be analyzed using the tools of
constrained optimization. A more specific form of Equation 3.16 may be written as a
utility function of the basic commodities that are dependent on time, “t” and goods x1
and x2, as implied by Equation 3.14. In turn, the constraint facing households then
becomes one of the cost of goods from the market, which will depend on their relative
price, and the cost of time, as implied by the wage rate. Specifically one has to recognize
that the consumer, as a producer, has to commit time or goods to produce either Z1 or
Z2 so that total money income will be equivalent to either expenditure on goods or
forgone income from allocating time to the production of goods as opposed to work.
Equation A3.1.11 presents the LM formulation of the optimization problem. The first
term on the right-hand side is the utility function. The term in brackets is the modified
constraint. Here b1 and b2 represent the proportions of Z1 or Z2 that are derived from x1
or x2 respectively; t1 and t2 likewise represent the proportions of time committed per unit
of Z1 or Z2, which has a price of w. In this regard, the production functions are assumed
to be of a simple proportionate form. S is “full income” which can either be spent on
market goods or represents forgone money income.
L ¼ UðZ1 ; Z2 Þ lðp1 b1 Z1 þ p2 b2 Z2 þ wt1 Z1 þ wt2 Z2 SÞ ðA3:1:11Þ
FOCs for this problem imply:
LU=LZ 1 =LU=LZ2 ¼ ðp1 b1 þ wt1 Þ=ðp2 b2 þ wt2 Þ ðA3:1:12Þ
which implies that the price of the commodities Z1 and Z2 depend on the prices of x1
and x2, as well as the wage rate as the opportunity cost of time. The prices also depend
on the coefficients b1 and b2 and t1 and t2. A goods-intensive activity such as weight
training, Z1, would thus have a lower value of “t” and a higher value of “b” compared to
a time intensive activity such as golf, Z2. The slope of the constraint in Figure 3.7 is given
by A3.1.12. It follows that an increase in “w” will increase the cost of the time-intensive
activity more, i.e., the denominator of the equation, showing that the marginal utility of Z2
will now be greater than for Z1 in equilibrium. This must imply substitution away from
Z2 towards Z1, because of diminishing marginal utility substitution towards more goods
intensive activities.
Investment in human or social capital implied in Gary Becker's analysis can be illustrated. For
a Cobb–Douglas utility function Downward and Riordan (2007) derive:

b M b
C2 ¼ þ C0 ðA3:1:13Þ
a þ b p2 aþb 1
which is the Marshallian demand for a commodity, “C,” augmented by the last term on the
right. The interpretation of this term suggests that consumption is higher than it would have
been, because of the presence of previous consumption, i.e., investment in consumption
skills, or the acquisition of social characteristics. In this respect, previous consumption of the
same or similar activities is predicted to increase current consumption of them, and predicts
that sets of consumer characteristics are likely to be associated with distinct groupings over
time as social capital accumulation.
Appendix 3.2 Durable Goods, the Time Value of Money and

InvestmentDemand
Durable goods
Durable goods yield a flow of services or utility over time, rather than being used up with one
consumption act. It is clear that both training shoes and running vests are durable goods that
are “stocks” or assets, from which flows of utility can be derived through their repeated use.
This makes the purchase of these goods “investments.” The analysis of the chapter should
be seen as a long-run equilibrium set of choices in which the stock of assets is adjusted.
More specifically, however, the investment “Iv” in each asset, “A” for each time period “t”
can be represented as:
Ivt ¼ At ð1 uÞAt1 ðA3:2:1Þ
Or:
Ivt ¼ dAt þ uAt1 ðA3:2:2Þ
This means that investment is equal to the change in the capital stock or net investment, plus
replacement investment. The term “u” represents the amount of depreciation of the asset for
the last period's use. To retain the current stock of the asset would require compensating for
this depreciation by adding a proportion of the last period's asset stock. In non-industrial
contexts such depreciation may not be explicitly accounted for, but it is present nonetheless.
This analysis shows that time matters for investment. However, because time is passing, the
effects of interest rates on the value of flows of income and expenditure should really be
accounted for. To do this one needs to understand the time value of money.
Conclusion 93
The time value of money

As discussed above in the chapter, the rate of interest measures the time value of money. It
should not be confused with the rate of inflation, which measures the change in the
purchasing power of money because of a change in the general level of prices. Even if
inflation were zero, the rate of interest would be positive. The nominal rate of interest is the
rate of interest excluding the rate of inflation. The real rate of interest is the nominal rate of
interest minus the rate of inflation. Because the value of money varies over time, it would be
wrong to simply compare the flows of incomes and expenditures across different time
periods, as one would not be comparing like with like. In contrast, it is usual to convert flows of
expenditures and incomes into present values, i.e., their values expressed in the current time
period.
To give an example, suppose e10 is saved in a bank, and the current rate of interest is 10%
per annum. At the end of this year this will produce e10(1 + 0.1) = e11. What about the end
of year two? If the savings are left in the in the bank, interest would compound with earnings
(e10(1 + 0.1))(1 + 0.1) = e10(1 + 0.1)2 = e12.10. In principle, this process can continue
and in general the future value of any amount “A” saved at interest rate “r” over “t” periods will
be:
Future value ¼ Að1 þ rÞt ðA3:2:3Þ
Now, it follows that the present value of any amount “A” saved at interest rate “r” over “t”
periods will be:
Present value ¼ A=ð1 þ rÞt ðA3:2:4Þ
Consequently, e12.10 received in two years time at a market rate of interest of 10% is
equivalent to e12.10/(1 + 0.1)2 = e10 today.
This means that, if a consumer was consuming goods and services over time, say two
periods, with x1 referring to the current period's consumption and x2 to the next period's
consumption, the utility maximizing problem would become:
Max U ¼ Uðx1 ; x2 Þsubject to W ¼ p1 x1 þ ðp2 =1 þ rÞx2 ðA:3:2:5Þ
Note that the price of “next” period's consumption, x2, is discounted into present values.
Technically, too, as the sum of two different periods' flows of consumption, the constraint is
now labelled “W,” to represent wealth. As should be clear now from the discussion of these
appendices, wealth is a stock and the sum of a series of income flows. Once this complication
is allowed for, the analysis of consumer choice can proceed as before. Figure A3.1.1
reproduces Figure 3.3, but translates it into consumption over time, i.e., use, of training
shoes.
The slope of the budget line is now a wealth constraint with a slope determined by the present
value of future prices and the current price. Based on these relative prices the consumer
chooses to “buy,” i.e., use, combinations of training shoes over time. In this case, the training
shoes or assets constitute wealth that finance a certain use of them. In this case, the prices of
p1 and p2 are really implied from the use of the training shoes rather than simply by their
actual market prices. Nonetheless, it follows that the values of It and p1 and p2 in Equation
A3.3.7 should really be discounted to period t 1 to make the analysis more robust. Chapter
12 discusses investment decisions in sport in more detail. There are further complications,
however, when considering the consumption of investment goods.
Investment demand
The above analysis suggests that consumption over time can undergo smooth transitions.
This might be reasonable with non-durable goods, for example substituting a visit to the gym
next period for one this period if p1 falls relative to p2. In Figure A3.3.1 this would make the
budget line steeper, pivoting around the horizontal intercept, to recognize that more current
consumption could be undertaken for any amount of future consumption. As detailed in the
chapter, income and substitution effects (over time) would take place with the prediction that
with normal goods more current consumption than before would take place.
However, with durable goods, the substitution is unlikely to be a smooth transition. This is
because the asset can essentially “fund” a number of particular incidences of consumption.
For example, if training shoes last approximately six months and are used to run three times a
week, then the single purchase of training shoes can fund 72 different runs. As Gratton and
Taylor (2000) recognize, sports goods consumption is likely to be unstable and volatile. A
simple model of accelerator effects is used to illustrate this purpose11.
If “Ivx” represents investment in a sports good, such as training shoes, and the change in
demand “dD” represents the change in use of the shoes, which will be a certain number of
FIGURE A3.1.1 Consumption of training shoes over time.
11
A variety of economists proposed an accelerator mechanism to link changes in the trade
cycle to investment decisions. See, for example, Clark (1913), Samuelson (1939) and Hicks
(1950).
Conclusion 95
times per given period, then if the accelerator coefficient “a” is the number of times a pair of
shoes can be used to support an activity (which is the capital/output ratio) over the same
period, the accelerator model is:
Ivx ¼ adD ðA3:2:6Þ
Thus if a = e2, because say the shoes cost e100 and could be used 50 times before
depreciating beyond use, and is currently stable, there will be no investment. If demand
increases, say, from currently 20 runs to 36 runs (dD = 16) over the period, then Ivx will rise
“suddenly” to e32. If the number of runs increases further, to 48 from 36 (dD = 12), then Ivx
will actually fall to e24. Note that, despite the additional increase in runs, investment value
has actually fallen, but is still positive. To retain the current level of demand for training shoes
would have required the number of runs to increase from 36 to 52 (dD = 16, again), i.e., a
constant growth of demand would be required. It follows that, if the growth of demand is
successively larger, investment will increase, if growth reduces, even though it does not
become negative, investment will fall. If demand stabilizes, then once dD = 0, investment will
also return to zero.
It should be noted that the model assumes that product differentiation is constant. It follows
that technological improvement or marketing activity may increase Ivx through adjustments
in “depreciation” of a substantive or perceived kind. Indeed, given the relatively stable
participation rates in sport, which are noted in the next chapters, marketing behaviour may
explain the growing expenditures in these industries as sports equipment, particularly
clothing and footwear, are presented as general fashion items.
CHAPTER 4

Participation: Evidence
OBJECTIVES
& To understand how economists model the demand for sports participation
and consumption
& To understand how sports participation is measured by official sources
& To appreciate some of the main empirical patterns of international sports
participation and consumption
4.1 INTRODUCTION
In the previous chapter various economic theories of choice were presented.
The aim was to indicate how participation decisions in sport are fundamen-
tally linked to expenditure decisions, and to explore the likely efficacy of
policies aimed at promoting sports participation. In this chapter attention
turns towards the evidence associated with sports participation and expendi-
ture and, consequently, to explore the potential targets for policy intervention.
Section 4.2 discusses the main sources of official and widely-used unofficial
data on sports participation and expenditures, and presents some descriptive
statistics from these sources. Section 4.3 then identifies how these data have
been modelled using econometric methods, with reference to the theories of
sports demand discussed in Chapter 3. Section 4.5 reviews the implications of
the findings from the literature for sports policy, before the chapter concludes.
4.2 EMPIRICAL EVIDENCE ON PARTICIPATION AND

EXPENDITURE IN SPORTS
4.2.1 Mass sports participation: Sources of official and unofficial
data
As discussed in Chapter 2, there are a variety of official data sources interna-
tionally investigating mass sports participation. This reflects the nature of the 97
98 CHAPTER 4 The Economics of Sports Participation: Evidence
agency collecting the statistics, the policy purpose to which the statistics
might be put, and convention. Different activities might be investigated be-
cause of their heritage or the particular collective definition of sport employed.
There may also be some terminological ambiguities connected with the activ-
ities associated with country-specific nomenclature, for example over defini-
tions of “football.” How participation is measured may also vary. This not only
concerns the type of activities that are investigated, but the frequency of
participation, the standard period scrutinized, and measures of the intensity
or duration of the activity. Finally, the periodicity of data collection also varies.
In part this reflects the sources of data. Traditionally, data may have been
collected as part of a particular module appended to a regular social survey
or alternatively it may reflect a specifically commissioned piece of work.
Table 4.1 illustrates some international official data sources. They are
considered official data because they are either produced directly by govern-
ment agencies or through policy bodies funded by them. The first column
indicates the country in which the data applies; the second column includes
both the name of the survey and the sponsoring agent. The subsequent col-
umns indicate how the data were collected, some dates over which the survey
was carried out, the age used to define an adult, the approximate sample size
and, finally, how participation was measured. It is clear that some variety is
evident across these characteristics.
There have been some attempts to produce a greater degree of commonality
of data collected, particularly within the European Union. van Bottenburg et al.
(2005) note three main initiatives for the promotion of harmonized statistics
through standardized questionnaires: COMPASS (Coordinated Monitoring of
Participation in Sports); HETUS (Harmonized European Time Use Studies);
and IPAQ (International Physical Activity Questionnaires). However, it should
be noted in the context of Chapter 2, that as sports have figured much more
prominently in recent policy discourse, and because policy initiatives require
targets to be met for participation, more dedicated context-specific data collec-
tion is also apparent. For example, while in many respects the General House-
hold Survey monitors participation for an accepted list of activities described as
sport or leisure in Britain, the survey “Active People,” commissioned by Sport
England in 2005, represents a significant investment of £6 million aimed at
producing data that are representative at local authority level, hence its very
large sample size. The motivation for such data is clear, which is to provide the
means to monitor how the specific policy agency meets its targets and policy
priorities by delivering sport through community sport networks.
In addition to the official data, various other “unofficial” data sources on
participation and expenditure exist. In the UK, Mintel market intelligence
reports are available on a wide range of specific sports and the sports sector as a
TABLE 4.1 Some official sports participation data
Country Survey (1) and sponsor (2) Type Typical Adult age Approximate Activity measure
dates sample size
USA National Health Interview Household 1957– 18 years 36 000–44 000 Type, frequency,
Survey (NHIS) interview current intensity over two
National Centre for Health weeks
Statistics Centers for Disease
Control and Prevention
1. Behavioural risk factor Telephone 1984– 18 years 35 000–107 000 Type, frequency,
surveillance system (BRFSS) interview current intensity over past
2. National Center for Chronic month
Disease Prevention and
Health Promotion
Canada 1. National household survey Telephone 2004 16 years 2408 Type, frequency,
Empirical Evidence on Participation and Expenditure in Sports

on participation in sport1 interview duration over the
2. The conference board of past 12 months
Canada
1. General Social Survey 1985–2005 Type
2. Statistics Canada
Australia 1. Exercise, Recreation and Telephone 2001–2005 15 years 13 726 Type, frequency,
Sport Survey (ERASS) interview duration over the
2. Australian Sports past 12 months
Commision
1. Population Survey Monitor Household 1993–2000 18 years 3000 Any sport over the
interview past 12 months
2. Australian Bureau of
Statistics
Great 1. General Household Survey Household 1987, 1990, 16 years 19 529–14 819 Type over the past
Britain (GHS) interview 1993, 1996, 12 months
2002
2. National Statistics Type, frequency
over past four
weeks
1. Active People Telephone 2005 16 years 363 724 Type, intensity
interview over the past four
2. Sport England weeks
1. Taking Part Household 2005 16 years 28 117 Type over the past
interview 12 months
2. Department of Culture
Media and Sport
99
100
CHAPTER 4 The Economics of Sports Participation: Evidence

Country Survey (1) and sponsor (2) Type Typical Adult age Approximate Activity measure
dates sample size
European 1. Various European 1983 15 years varies Type, frequency

Union Barometers (EB) (EB19), and duration over
1987 (EBs 12 months
28, 28.1),
1990 (EBs
33, 34.2),
1997
(EB47.2),
1998
(EB50.1),
1999
(EB52.1),
2001
(EB55.1),
2003 (EBs
58.2,
2003.1, 60)
2004
(EB62.0)
2. European Commission
1
This survey also measures the economic expenditure of households on sport.
Empirical Evidence on Participation and Expenditure in Sports 101
TABLE 4.2 Top ten and team sport participation rates1
USA activity Canada activity Australia activity Great Britain activity
NHIS % NHSPS % ERASS % GHS 2002 %

1991 2004 2005
Walking 44.1 Ice hockey 6.6 Walking 37.3 Walking 35.0

(other)
Gardening 29.4 Golf 6.5 Aerobics 18.5 Swimming 12.0
(indoor)
Stretching 25.5 Baseball 4.7 Swimming 14.4 Keep-fit 12.0
Cycling 15.4 Skiing 4.0 Cycling 10.3 Snooker 9.0
Weight 14.1 Soccer 3.8 Tennis 7.8 Cycling 9.0
lifting
Stair 10.8 Volleyball 3.1 Running 7.7 Weight 6.0
climbing training
Running 9.1 Basketball 2.8 Golf 7.1 Running 5.0
Swimming 6.5 Tennis 2.6 Walking 5.7 Golf 5.0
(bush)
Basketball 5.8 Curling 2.6 Football 3.8 Soccer 4.0
(outdoor) (outdoor)
Golf 4.9 Bowling 2.4 Basketball 3.5 Bowls 3.0
Baseball 3.5 Included Rugby 1.2 Basketball 1.0
above league
Football 1.5 Rugby 1.0 Cricket 1.0
union
Soccer 0.9 Cricket 1.0 Rugby 0
1
Data are taken from the following sources: U.S. Department of Health and Human Services, Centers for Disease Control
and Prevention, National Center for Chronic Disease Prevention and Health Promotion, (1996); The Conference Board of
Canada (2005); Australian Sports Commission (2005), Fox and Rickards (2004). Zero figures correspond to participation
rates less than 0.5% of the population.
whole. The Sports Industry Research Centre (SIRC), based at Sheffield Hallam
University, also produces market reports and forecasts for the sports market.
In the US and Australia SMGA International and Sweeney Research, respec-
tively, are commercial agencies which provide such data.
4.2.2 Mass sports participation: Descriptive evidence

To give an overview of participation, Table 4.2 presents some statistics from a
selection of the data sets described above for the top ten activities in each
country, plus the rates for the team sports that dominate the professional
industry.
Notwithstanding the different sampling periods and questions, the data
reveal a number of interesting points. The first is that participation rates are
greatest for leisure and/or recreational activities and then they rapidly decline
TABLE 4.3 Overall participation rates1
Country %
Finland 85
Sweden 82
Denmark 69
Ireland 60
Netherlands 59
UK 55
France 54
Belgium 51
Luxembourg 48
Germany 47
Austria 45
Spain 43
Italy 33
Greece 32
Portugal 27
Australia 69
US 76
Canada 31
1
Data are taken from Van Bottenberg et al (2005) for the European countries. The same
data sources are used for Australia, US and Canada.
through various forms of keep fit activity to more passive activities, such as
golf. The second point is that traditional team sports appear as relatively
small-scale activities. As discussed further in the next chapter, this has poten-
tial implications for sports policy. Table 4.3 presents participation at least once
per month for European countries in 2004, Australia and Canada in 2005, and
the US for at least some activity over the year.
It is clear that, in spite of comparative measurement issues, participation
rates vary quite widely by geographical dispersion. This hints at broader struc-
tural differences in participation. Such structural differences are also evident
within countries. Table 4.4 highlights how overall participation varies accord-
ing to key socio-economic criteria, depending on its availability in published
form. Despite the variation in measurement, some general patterns are clear.
Males tend to participate more than females, and participation falls with
increasing age, but rises with higher incomes and better education. There is
also some evidence of ethnic differences in participation, and that household
status and the presence of children can affect participation.
Policy makers have suggested reasons for such differences. It is argued by
van Bottenburg et al. (2005) that differences in gross domestic product (GDP)
can account for the variation in participation rates across Europe, in other
words that income is a key “driver.” Various other drivers are identified. Sport
TABLE 4.4 Socio-economic determinants of participation
Socio-economic US % Canada % Australia % Great Britain % Europe2%

indicator NHIS 1991 NHSPS 2004 ERASS 2005 GHS 20021
Male 78.6 39.0 83.5 65 41

Female 73.1 23.4 83.1 53 35
Age (18–29) 78.3 (<20) 67.2 (15–24) 93.4 (16–19) 77 (15–24) 60
(30–44) 76.6 (20–29) 53.7 (25–34) 89.9 (20–24) 69 (25–39) 41
(45–64) 74.2 (30–39) 42.5 (35–44) 88.9 (25–29) 70 (40–54) 34
(65–74) 74.3 (40–49) 33.3 (45–54) 88.8 (30–44) 67 (55+) 28
(75+) 66.3 (50–59) 29.6 (55–64) 90.7 (45–59) 59
(60+) 26.1 (65+) 90.1 (60–69) 50
(70+) 30
Ethnicity White 78.5 White 67.9
Black 71.6 Non-white 65.5
Empirical Evidence on Participation and Expenditure in Sports

Hispanic 66.4
Annual household (<$10k) 69.7 (<$20k) 21.7 (<£200) 9.3
income (weekly ($10k–$20k) 69.8 ($20k–$40k) 26.1 (£200–£400) 11.9
income for Great ($20k–$35k) 75.7 ($40k–$60k) 35.5 (£400–£600) 14.5
Britain) ($35k–$50k) 80.5 ($60k–$80k) 41.9 (£600–£800) 11.4
($50k) 85.6 ($80k–$100k) 46.3 (£800–£1000) 7.1
($100k) 55.1 (£1000) 13.7
Marital status (married) 38.4
(other) 29.5
Children in the Children 37.6 Children 22.0
household No child 37.0 No child 45.8
Highest education College (16+) 85.8 <High school 16.7 Degree 22.9 (16–19) 32
High school 34.7 A Levels 9.3 (20+) 50
University 46.7 O Levels 15.9
1
Ethnicity and income figures are based on the authors’ own calculations from the raw data and refer to rates against each group total and the total sample size respectively.
2
Data are taken from Van Bottenberg et al (2005) for the European countries.
103
England (2004a) suggests that “ageing” affects participation by reducing the

physical ability to participate and, moreover, raises the concentration of any
sport “illiteracy” among the population at large. Time pressures associated
with increased work hours for employees in the UK arguably reduce participa-
tion. These time constraints in turn affect the capability of volunteers and
professionals to support and to develop sport. Well-being and obesity are iden-
tified as alternative scenarios associated with increasingly sedentary lifestyles
which would affect participation. The levels of investment connected with the
provision of facilities in the locality can affect participation, as does education,
which has the potential to shape the tastes and provide opportunities for sport.
These opportunities may be through the provision of facilities directly or
indirectly through education's potential to raise employment and income.
Variations in access to sport are thus important, and this will also depend on
any additional constraints implied by any sex and ethnic discrimination.
Similar drivers are identified elsewhere. The Conference Board of Canada
(2005) identifies age, gender, income and education as key drivers of change,
but in addition to this find household composition significant. Similarly, the
Australian Sports Commission (2005) found that age, sex, labour force status
and education level affect participation. They also identify the importance of
regional status in affecting participation. For example, participation rates are
higher in state capitals than in other areas. The identification of these drivers
appears to have occurred only indirectly with reference to theory. Sport Eng-
land (2004b) documents how consultation with academic research and data
scrutiny led to the development of their drivers of participation. In contrast,
van Bottenburg et al. (2005) argue that European patterns of participation
could be accounted for by social-psychological and sociological theory, ac-
counting for the interaction between personal, interpersonal and environmen-
tal factors, such as:
“. . .age, gender, education, socio-economic status, perceived advantages and
barriers, perceived health/fitness, intention, self-efficacy, self motivation,
social support and the subjective experience of the living environment and
everyday surrounds.” (van Bottenburg et al., 2005, p. 187)
However, in both cases, it is notable that no reference to economic theory is
made, despite the discussions in Chapter 3 in which it is clear that many of the
variables mentioned can be interpreted through the lens of economic theory.
4.2.3 Expenditure on mass sports participation

With respect to analyses of expenditure on sport, the main official sources of
data tend to be expenditure surveys that are used for monitoring national
economies and, for example, to underpin such data as presented in Table
Empirical Evidence on Participation and Expenditure in Sports 105
TABLE 4.5 Consumer spending on sport
Great Britain £ million (2001) Australia $ million (2003)
Item 1999 2004 Item 1998 2003
Sport clothing 2817 4054 Bicycles 40.4 52.4

and footwear
Sport equipment 900 1386 Boats 406.9 395.3
Health and Fitness 1339 1531 Camping equipment 111.0 133.1
Boats 701 850 Fishing equipment 141.2 185.5
Participant sports 2036 1923 Golf equipment 80.7 80.7
Spectator sports 801 750 Sports footwear 380.7 459.8
Sport gambling 2025 3404 Swimming pools 611.5 1375.4
Sport TV and video 1345 1852 Other sports equipment 363.1 818.8
Sport related 633 551 Hire of sports equipment 60.4 24.2
publications
Sport related travel 1013 965 Health and fitness charges 276.9 580.8
Other expenditure 428 418 Sports club subscriptions 483.1 419.5
Total 14 036 17 684 Spectator admission fees 349.3 294.4
Sports facility hire 1041.8 927.7
Sports lessons 437.8 423.5
Total 4784.9 6172.2
2.4 concerning the overall scale of sport in the economy. Disaggregation is,
therefore, constrained to reflect the categories investigated by particular sur-
veys. There are, of course, some exceptions. Taks and Kesenne (2000) com-
bined data from a survey of 512 households in Flanders with government
expenditure data, private expenditure and investment, to identify a Gross
Regional Sport Product of US$4.3 billion, noting a rise of 4.7 times the level
of 15 years previously.
More generally, rising sports expenditure is identified by the available
research from official surveys. In the UK SIRC (2005) estimates, based on
the family expenditure survey, that the value of sports expenditure increased
from £14 036 million in 1999 to £17 684 million in 2004, allowing for
inflation. Similarly, in Australia, the Australian Bureau of Statistics argue that
the household expenditure survey of 2003 reveals an increase in expenditure
from 1998 of AUS$4,784.9 million to AUS$6,172.2 million allowing for
inflation. Some indication of the dispersion of these expenditures is given in
Table 4.5.
In Great Britain it is clear that the main components of increased expendi-
ture are clothing and footwear, sports gambling, which has emerged as a
dynamic market, and sport television and video. Coupled with the static ex-
penditure from participants, these changes in expenditure patterns are
indicative of relatively autonomous changes in demand. Such claims receive

support from unofficial sources of data. Mintel (2005), for example, indicates
that leisurewear for general, as opposed to sporting, use is a growing market,
and that satellite television spectatorship has affected participation levels in
sport. It is also argued that the growth in expenditure on sports equipment is
indicative of purchases of items, such as bicycles, golf and keep fit equipment.
Such developments, coupled with the growth of expenditure on health and
fitness, are suggestive of a more individualistic and informal participation in
sports than through organized clubs. Similar patterns appear to be the case in
Australia, with expenditure on hire charges falling but equipment purchases
rising, along with swimming payments and increases in expenditure on foot-
wear and bicycles. These issues are discussed further in the next chapter as they
indicate how supply may well have changed to meet changing demands, par-
ticularly as there has been growth in the private sector provision of sports
facilities.
In terms of the general drivers of expenditure, rather than its distribution,
The Conference Board of Canada (2005) reports that in 2004 households
spent approximately $15.8 billion on sports compared to $8.9 billion in
1994. This represents a rise in the proportion of GDP spent on sport from
0.9% of GDP to 1.2% of GDP. It is argued that rising incomes, coupled with
participation in sports contribute towards the higher spending. In contrast,
family size can increase spending, but this reaches a plateau with more than
two children in the household. Likewise, SIRC (2005) argues that increases in
GDP have helped to increase the spending on sports in Britain. Results such as
these suggest that sport is either a normal or a luxury good.
In economics, and as implied in Chapter 3, a normal good is one whose consumption
rises with increases in income. A superior good will experience a greater
proportionate increase in consumption than the increase in income. This means
that the share of household budgets spent on the superior good will increase.
While an economic interpretation can be placed on such data, none of the

official or commercial reports specifically test economic theories in connec-
tion with their claims. The next section concludes the chapter by discussing
the evidence of sports participation and expenditure that has this objective.
4.3 MASS SPORTS PARTICIPATION: THEORETICAL AND

EMPIRICAL ANALYSIS
In this section, analyses of data directly concerned with testing hypotheses
about sports consumer behaviour in the mass participation context are dis-
cussed. Drawing on the results shown in Figure 3.1 concerned with the
Mass Sports Participation: Theoretical and Empirical Analysis 107
allocation of leisure time, participation in sport and subsequent expenditures

are discussed.
4.3.1 Leisure time

There is a voluminous amount of economic literature that examines, indi-
rectly, the demand for leisure time, as the counterpoint to the supply of labour
in the labour economics literature. As discussed in Chapter 3, in the tradi-
tional income–leisure trade-off model leisure is seen as the residual of work
and the wage rate as the opportunity cost of leisure. In this section for brevity,
therefore, attention turns to studies that were directly motivated by a concern
for leisure as the prime unit of analysis. The direct concern for the demand for
leisure time can be noted as corresponding with the growing acceptance of the
New Household Economics of Gary Becker, discussed in Chapter 3. For
example, Gronau (1973) analyzes the intrafamily allocation of time and
argues that:
“. . . the classical dichotomy of ‘work in the market’ versus ‘leisure’ may
serve as a good approximation to the role the husband plays in the
production activity of the household, but does gross injustice to the wife.
To call the whole of the time spent by the wife outside the market sector
‘leisure’ is to overlook the production activities she engages in at home . . .
and the wife's allocation of time should therefore be analyzed in terms of a
three way division of work in the market, work at home and leisure.”
(Gronau, 1973, p. 634)
Notwithstanding the presumption about gender roles implied in this research,
it is also noted that:
“. . . (a)n empirical estimation of the demand for leisure and supply of work at
home calls for detailed data concerning the time budgets of the various
family members. The existing data are too crude to provide conclusive
results.” (Gronau, 1973, p. 641)
Consequently, the value of housewives’ time is estimated on US census data
for 1960 by predicting the potential wage rate of females who do not work for
given income, and the age and education characteristics of those who do
work, controlling for the probability of the individual working by imputa-
tion from the normal distribution. This potential wage is then regressed on
income and the probability of working. It is identified that income increases
the value of time for white females, and more so generally in the presence of
children.
More directly, Abbott and Ashenfelter (1976) and Phlips (1978) make use of
utility functions including leisure time directly and a “full income” constraint
to explicitly derive the demand for leisure, defined as hours not spent working,
BOX 4.1 LINEAR EXPENDITURE MODELS

It was shown in the last chapter that the typical demand for Two common specific forms of analysis of expenditures that
sports equipment, clothes and facilities (x) can be pre- are often employed in economics, or implicitly underpin the
sented as dependent on the price, p, the price of other econometric analysis of expenditures are:
goods or services, p*, consumer incomes, I and prefer-
1. the Engel curve: px = b1 + b2I
ences P. P
2. the linear expenditure system: pixi = piy + bi (I pjxj)
x ¼ xðp; p*; I; PÞ The Engel curve models the impact of income, and possibly
other factors, on expenditure on a good. The linear expen-
It follows logically that demand can also be represented in diture system is similar, but accounts for the fact that ex-
terms of expenditure, i.e.: penditure on good “i” can only be funded out of income
minus the expenditure on all other goods “j.” In this equa-
px ¼ pxðp*; I; PÞ
tion, “y” represents some minimum amount of consump-
tion of good “i” that is required.
as part of a system of equations including the demand for other goods and, in
Phlips (1978), the demand for money also.1 Based on the same US data
ranging from 1938 to 1967, linear expenditure systems, as discussed in
Box 4.1, are estimated to reveal that the demand for leisure time rises with
income.
Two consequences of the theoretical and modelling strategy are that first,
specific restrictions are imposed on the demand functions. Leisure and all
other goods or money are treated as gross substitutes, because they account
for elements of a full income constraint. The second point is that goods and
leisure demands are not derived in a hierarchical way, as they reflect the joint
allocation of resources by the consumer. This is a direct consequence of the
New Household Economic approach.
Since these earlier studies, research has probed the specific factors that
affect the use of time. For example, Gronau (1977) used the 1972 panel of
the Michigan Study of Income Dynamics to analyze the determinants of hours
at work in the market, work at home and hours’ leisure. Using OLS analysis, it
was identified that females’ leisure time increased with age, education, and her
husband's income and education, but declined with the presence of children
aged 0–17 in the household, the presence of children at school, the number of
1
The estimated model is referred to as “seemingly unrelated regression” (SUR) because a
series of regressions for each element of expenditure is estimated with no expectation that
dependent variables for one equation appear in other equations. The correlation between the
equations occurs through the random error terms and, in this case, arises because the budget
constraint must equal the sum of expenditures on the different elements.
rooms in the house (proxying the need for housework) and her previous
employment experience.
Similar analysis was undertaken by Kooreman and Kapteyn (1987), who
also made use of US University of Michigan Survey Data, but undertook an
improvement in the analysis by formally controlling for sample selection by
using a Heckman model to first impute the (shadow) wage of nonworking and
working women, recognizing that wage data only applies to those who work.
In this respect, the probability of working being conditional on a set of the
individual's characteristics is used to impute the wage rate, in contrast to
Gronau (1973).
It is identified that the presence of children affects female participation in
leisure, unlike males. Male leisure demand falls and then rises after 50 years of
age. Education can lead to some substitution of passive activities like reading,
watching television, etc., for males, but the reverse is the case for females. As
wage rates rise, it can be shown that demand for leisure increases for males,
but falls for males in sports-related activities. The opposite is the case for
females, whose demand for leisure falls. Interestingly, such findings are syn-
onymous with the descriptive results presented in Table 4.4 and indeed more
recent research. Table 4.6 summarizes the results of more recent literature on
the demand for leisure time, where the sign indicates the influence on leisure
time by variables in the analysis.
In summary, it appears that males tend to have more leisure time than
females, and that education and income can raise the time spent on leisure,
although work time and the constraints of family can reduce this. The former
TABLE 4.6 The determinants of leisure time
Author Influences on leisure time
Zuzanek (1978) Male (+); female employment (); education (+);

age (+ then )
Solberg and Wong (1991) Male (+ with wage, wife's wage, with travel, wife's
travel); female (+ with husbands wage, with travel,
wife's travel)
Altergot and McCreedy (1993) Children (); married male (+) for passive leisure
Dardis, Soberon-Ferrer and Age ( active leisure; + passive leisure); education
Patro (1994) (+); male (+)
Robinson and Godbey (1997) Age (+ then ); education (+); children ()
Bittman and Wajcman (2000) Male (+)
Thrane (2000) Male (+); full-time employment (); age (+ then );
children ()
Lee and Bhargava (2004) Married (); children (); full-time or part-time
work (); male (+); black ()
is the case particularly for males and the latter is the case for females. Inter-
estingly, it should be noted that by no means are all of the studies above
directly motivated by the New Household Economic approach. Most of the
studies in Table 4.6, in fact, adopt a more sociological approach consistent
with an heterodox economic perspective (for example, Thrane, 2000; Bittman
and Wajcman, 2000; Robinson and Godbey, 1997; Altergot and McCreedy,
1993; Zuzanek, 1978), consequently the commonality of results is actually
coupled with a variety of theoretical explanations, which is important in
considering policy options. Similar results also apply to analyses of the deci-
sion to participate or not in specific activities.
4.3.2 Participation in sport

Table 4.7 presents the results of a survey of studies that have all used
official data to examine the choice to participate in various activities for
a variety of countries. A number of general findings appear from the re-
search, although naturally there is some variation reported in the detail of
the studies. Once again, it is evident that males tend to participate more in
sport and more frequently than females, but in particular activities, such as
aesthetics, and sports such as skiing and skating, this is not the case. It is
also evident that lower age, income and socio-economic status, for example
being at a professional or managerial level or being a skilled worker, raise
the participation rate in sports. The same is true of better health, and
higher levels of education, as well as having access to vehicular transport.
This is also the case where there is participation in other activities or other
family members are active.
Notably, drinking alcohol is associated with higher participation in sport as
opposed to smoking. More generally there is evidence that increased work
hours can reduce participation rates, as can being of non-white ethnicity. A
variety of household characteristics also reduce participation in sport. These
include being married or a couple and, particularly, the presence of children in
the household.
A number of the studies also investigate the frequency of participation.
It is not uncommon for the signs on the variables noted above to be
reversed in this case. Thus age, being married, unemployed, part-time
employed or retired, raise the frequency of participation, while education
and being in full-time employment reduce it. There are also some common
features, for example, the number of activities participated in raises the
frequency of participation for given activities. In short, it appears to be the
case that the decision to participate and its frequency are relatively discrete
choices.
TABLE 4.7 Results of a survey of studies using official data into the choice to participate in various activities
Country Author Sample characteristics Theory and estimator Findings
US Cicchetti et al. 1960 National Recreation Neoclassical demand

(1969) Survey; n = 16 000; combined
with 1965 Bureau of Outdoor OLS on participation or not Age (); non-white (); male (+); income (+); education (+); facility
Recreation; n = 7200; 24 supply (+)
activities OLS on days participated in 1959 Age (+); non-white ()
US Adams et al University of Michigan Survey Neoclassical demand
(1966) Research; n = 1352; 3
activities OLS on participation or not Age (); income (+); male (+); education (+); white (+)
OLS on days participated in 12 months Age (); male (+); white ()
US Stemple (2005) 1998 NHIS; n = 32 240–22 Heterodox (Bourdieu)
500; 25–80 years old; 21 sport Logit on participation or not in last two Income (+); education (+) (variance according to combinations of
and physical activities weeks economic and cultural capital; dominant classes have broad range)
US Humphreys 2000 BRFSS n = 175 246; >18 Becker Heckman model Age (+); married (+); income (); employed (+); education ();
and Ruseski years; 56 activities female (); white (); urban ()
(2006)
Participate or not in some activity in the Age (); married (); children (); income (+); employed (); retired
past month (+); education (+); female (); white (+); urban (+); health (+)
Mass Sports Participation: Theoretical and Empirical Analysis

Frequency
Great Gratton and Tice 1985 Health and Lifestyle Heterodox
Britain (1991) Survey; n = 9003; GHS 1997,
1980, 1983, 1986, 1987 Logit on participation or not in the last Male (+); age (); socio-economic status (+); income (+); illness
two weeks in any sport (); number of activities (+)
OLS on frequency energy expended Male (+); number of activities (+); unemployed (+); retired (+)
Logit on participation or not in last four Age (+); male (+); education (+); socio-economic group (+); full-
weeks time employment (); number of activities (+)
OLS on frequency in last four weeks Age (+); male (+); number of activities (+); part-time employed (+);
illness (+); student (+); keep house (+); separated (); married ();
children (); unskilled ()
Great Farrell and 1997 Health Survey of (Implicit Becker household preferences) Male (+); age (); married () children for males (+); infant ();
Britain Shields (2002) England; n = 3811 ethnic minority (); education (+); drinling (+); smoking ();
households = 6467 Random effects probit on participation health (+); income (+); unemployment (+); household
individuals; 16–65 years; 7 in the last four weeks for >15 minutes membership (+)
activities in all sports plus the top seven
Great Sturgis and 2000 Time Use Survey; 8 Logit on membership of activity groups
Britain Jackson (2003) * years; 42 sports activities
Active aerobic Age (); number of adults in the household (); income (+); male
(+); London/SE (+); own house (+); education (+)
Non-active competitive Age (); income (+); male (+); car ownership (+); north ();
education (); married (); co-habitation ()
Outdoor competitive Number of adults in the household (); income (+); male (+); car
ownership (+); separated (+)
Outdoor non-competitive Age (+ then ) number of adults in household (); children ();
income (+); male () managerial (+); intermediate mangaerial (+);
car ownership (+); east (); carer (+); own house (+); education (+)
Great Downward 2002 GHS n = 14 819; >16 Neoclassical and heterodox Working (+); skills/professional (+); education (+); married (+);
Britain (2007) years; top 10 sports activities regions not SE (+); male (+); white (+); health (+); smoking ();
Logit on participation in any sport , drinking (+); access to vehicle (+); age (); children (); number of
walking and individual actitivites in the adults in the household (); income (+); work hours (); unpaid
last four weeks work hours () volunteering (+); number of leisure activities (+)
111
112
CHAPTER 4 The Economics of Sports Participation: Evidence

Country Author Sample characteristics Theory and estimator Findings
Great Downward and 2002 GHS; n = 14 819; >16 Becker and heterodox
Britain Riordan (2007) ** years
Heckman model
Participate or not Age (); skills/professional (+); drinking (+); regions not SE ();
access to a vehicle (+); sports, other club membership (+); volunteer
(); number of sports (+); sport lifestyle ()
Frequency of participation in any sport, Education (); number of males in the household (+); health (+);
recreational sports and specialized employment (); north (+); access to a vehicle (); income ();
sports unpaid work (+); number of sports (+); sport lifestyle (+); recreation
lifestyle (+); leisure lifestyle (); volunteer in sport ()
Flanders Scheerder et al 1979 Leuven growth study; Heterodox (Bourdieu)
(2005a)* Flemish girls; 1989/199 study
on movement activities in Logit on
Flanders; n = 38 376; 19–77 Participation in the year membership of Age (); female (); class (+); family size (+); urban (+)
years activity groups; participation in the year
Solo sports Female (+)
Competitive and outdoor sports Male (+); class (+)
Duo/team sports Female (); age (); class (+)
Club sport Age (); female (); class (+)
Non-organized sport Age (+); female (+); family size ()
Flanders Scheerder et al 1969 Leuven growth study of Heterodox (Bourdieu)
(2005b)* Belgian boys; 1979 Leuven
growth study of Flemish girls
1989/99 university survey; Logit on membership of activity groups;
n = 22 424 participation in the year
Traditional Humanities school (); parents participating in sport (+)
Aesthetic Female (+); age (); humanities school (+); socio-geographical (+);
parents participating in sport (+)
Family Female (+); socio-economic (+); parents participating in sport (+)
Glide Female (+); age (); humanities school (+); parents participating in
sport (+)
Exclusive Female (+); age (+); family size (); socio-economic (+)
Flanders Taks and 1999-2001 Study on Heterodox (Bourdieu) market
Scheerder (2006) Movement Activities in segmentation
*
Flanders; n = 5172; 6–18
years
Logit on membership of activity groups;
participation in the year
Traditional Female (); age (+); socio-economic (+); parents participating in
sport (+)
Family Female (+); socio-economic (+); parents participating in sport (+)
Aesthetic Female (+); age (); humanities school (=); parents participating in
sport (+)
Exclusive/glide Females (+); age (+); family size (+); socio-economic(+); parents
participating in sport (+)
Popular action Age (); socio-economic (+); parents participating in sport (+)
And numbers of sports
One – univores Female (); age ()
Two – bivores Age (+); parents participating in sport (+)
More than two – omnivores Female (+); socio-economic (+); parents participating in sport (+)
Australia Stratton et al. 2002 General Social Survey; No explicit theory
(2005) n = 15 500
Logit on participation in the last 12 Age (); male (+); state (+); suburb (+); professional (+); income
months (+); socio-economic (+); couple no children (+); single (+);
(Could include non-activity education (+); english speaker (+); health (+); easy transport ();
involvement in sport) not safe environment (); weekly contact family friends ()
Norway Skille (2005) Primary data from author Heterodox (Bourdieu)
n = 566
Logit on participation or not in
Sport city program activities Female (); academic school (+); active family members (+);
volunteer family members (+); peer and media information (+)
Conventional activities Female (); academic school (+); active family members (+); active
friends (); volunteer family members (+); peer and media
information (+)
Sport city program versus conventional Academic school (); peer and media information (+)
activities
Germany Breuer (2006) 1984–2003 Socio-economic Becker
panel; n = 98 772
Logit on participation once a week Income (+); working time (); education (+); age (); immigrant
();
Germany Lechner (2008) 1984–2006 Socio-economic No explicit theory Males: German (+); education (+); year (); technical occupation
Mass Sports Participation: Theoretical and Empirical Analysis

panel n = 6751 (); autonomy at work (+); never smoked (+); high life satisfaction
(+); unemployment (+)
Probit on participation at least monthly Females: year (); German (+); children <3 years (); children >10
years (+); family income (+); office work (); low autonomy at work
(); autonomy at work (+); illness (+); unemployment (+);
inhabitants per km2 (+); city centre ()
*
Studies used either factor analysis of cluster analysis to group activities.
**
Study used cluster analysis to identify lifestyles. See Box 4.2 for a discussion of these techniques.
113
BOX 4.2 FACTOR AND CLUSTER ANALYSIS

Section 4.2 outlined the main elements of econometric account for the variance in the individual variables allowing
research. Implicit in this discussion is that theory identifies for “errors.” As a result, the factors can be thought of as
specific independent variables that affect the dependent latent variables and included in the analysis instead of the
variable, and that the sample of data on which analysis is original variables. The number of factors will be less than the
based is fully transparent. However, in practice, indepen- number of original variables.
dent variables may be related, and data sets might comprise Cluster analysis: this is a statistical technique that classifies
sub sets or strata. Techniques of analysis derived from cases in a data set into subsets, known as clusters. The
psychology, the behavioural sciences and biology have clusters will be measured as similar or dissimilar to one
been developed to address such issues further, and are another according to a measurement criterion being ap-
now becoming commonly used in economics. plied to the values of the variables used to characterize
Factor analysis: this is a statistical technique that identifies the cases. Cluster membership is thus an indicator of a
linear combinations of variables, known as factors, which subsample in the data.
4.3.3 Expenditure
The empirical analysis of expenditure in sport is much less developed in
the literature. As a result, the findings are also less consistent. For exam-
ple, Lera-Lopez and Rapun-Garate (2005) review a sparse literature and
note that studies have shown that males tended to spend more money than
females on sports. Spending also appears to be linked to rising income.
However, the effects of age and education are less clear-cut. While Lamb et
al. (1992) report that younger people spent more on sport in the UK, in
Japan and Belgium Oga (1998) and Taks et al. (1999) suggest the opposite
applies, respectively. Taks et al. (1999) also suggest that spending might fall
with education.
Lera-Lopez and Rapun-Garate (2005) undertook their own survey in
Navarra, Spain, to examine the links between participation and expendi-
ture, although no explicit theoretical orientation is noted. Based on a
sample of 700 respondents, they estimated an ordered-probit model to
identify the frequency of sports participation in the previous year, based
on categories of participation. They then estimated a Tobit model to ex-
amine consumer expenditure in euros per year. The Tobit model was used
because expenditure is bounded below by 0. It was identified that partici-
pation frequency echoed the results of the research surveyed above. Thus,
the frequency fell for females and the employed, but increased with age.
Spending also fell for females, but increased with education and income.
This suggests that there is a direct relationship between participation and
sport but, naturally, facilitated by access to income.
4.3.4 Summary
The above review suggests that consumption in facets of sport, the alloca-
tion of time, the decision to participate and the frequency of participation
and subsequent expenditure on sport are indeed linked, as implied in
outline in Figure 3.1. Significantly, there is some broad qualitative consen-
sus as to the determinants of these features of activity. Significantly, how-
ever, a variety of theoretical positions has motivated the research and could
explain the broad findings including, therefore, the nature of these inter-
relationships.
Consequently, the research is consistent with the income–leisure trade-off
model in viewing the results on the effects of income and work hours on leisure
time or participation or with New Household Economics, in which such
findings coupled with the significance of socio-economic characteristics to
participation are viewed as indicative of investment in personal consumption
capital and human/social capital, and specialization in the division of labour
in the household, with child-care being substituted for participation by
females, and particular activities appealing to the different sexes according
to the development of lifestyles.
In contrast, one could view these latter characteristics as products of social
constraints at work which help to shape preferences, as emphasized in het-
erodox accounts. As discussed in the previous chapter, these provide alterna-
tive perspectives for deliberations over policies being aimed at influencing
participation in sport. On the one hand, if the observed patterns are indicative
of the exercise of given stable preferences then active policy becomes a distor-
tion to price signals and indicative of a welfare loss, as discussed in Chapter 1.
This is because economic inefficiency is implied because the marginal benefits
and costs of transaction are not being taken into account.
On the other hand, if it is argued that choices are not, at least in part,
freely made, either through informational or social opportunity deficien-
cies, then policy that seeks to educate about the benefits of sports partic-
ipation and/or to alleviate the resource and social constraints that affect it
become relevant. Significantly, as discussed above in the consideration of
policy drivers of participation, and in the next chapter, there is a current
rhetorical sweep towards advocating policies to promote participation in
sport and physical activities. This suggests that in practice, if not explicitly,
policy makers embrace the latter theoretical perspective more. This pro-
vides a counterpoint to the main discussion in the next two chapters that
examine the changing supply conditions in mass participation sport. Par-
adoxically, there is a movement towards a greater market-oriented supply
of leisure, and it is not entirely clear that policy intervention will have the
desired effects on the voluntary sector, which is hugely important in the

supply of mass participation sport.
4.4 CONCLUSION
In the previous chapter, various economic theories of choice were presented to
suggest how the use of time can influence the decision to participate in sport
and, in turn, undertake expenditure. In this chapter attention turned towards
the evidence associated with sports participation and expenditure and, con-
sequently, exploration of the potential targets for policy intervention. After
presenting the main empirical framework used by economists to examine
economic relationships, the main sources of official and widely used unofficial
data on sports participation and expenditures were presented, along with some
descriptive statistics from these sources, and subsequent policy discussion.
The chapter then examined a variety of empirical evidence on the demand for
leisure time, the decision to participate in sport and its frequency, as well as
expenditures that have been motivated by a direct desire to test various the-
ories of choice. It has been shown that common qualitative understanding of
the influences on sports consumption are identified, but alternative theoret-
ical explanations can support these results. Consequently, there are possible
tensions implied in promoting policies to affect sports participation.
CHAPTER 5
The Supply of Participant

Sport: The Public and
Private Sector
OBJECTIVES
& To understand the economic rationale for the public and private sector
provision of sport
& To appreciate the empirical evidence on the provision of sport
& To assess the rationale for public policy on the provision of sport
5.1 INTRODUCTION
In this chapter, and the next, attention turns towards providing an apprecia-
tion of the supply of mass participation sport. This is a very difficult task to
undertake, because the supply structure is complex and there is a shortage of
detailed theoretical and empirical work, with most literature comprising pol-
icy-related documentation and investigation. There is also some commercial
research available. There has also been rapid growth in the private sector
provision of sport, particularly in the UK, facilitated by political initiative
and following market opportunity.
This chapter aims to outline these developments. Section 5.2 briefly charts
the structure of supply in both the UK and a number of other countries.
Section 5.3 explores some general policy initiatives that have been promoted
to raise sports participation, and the motives for these policy initiatives.
Section 5.4 then outlines the economic rationale for public policy intervention
and, by implication, to explain why private sector provision and informal
participation can be understood as alternative supply structures. The econom-
ic logic underpinning public policy recommendations to promote sports par-
ticipation is discussed. 117
118 CHAPTER 5 The Supply of Participant Sport: The Public and Private Sector
Section 5.5 then provides an indication of the growth and economic nature
of private sector provision of sports in the UK. Section 5.6 discusses informal
sports participation and Section 5.7 the supply of sports equipment. The main
conclusion is that careful reflection is required to make a case that the market
is not providing adequate supply either in the private sector or by informal
activity.
5.2 THE STRUCTURE OF SUPPLY

As discussed in Chapter 2 the economic regulation of sport falls within the
remit of the Federal Sherman Acts in the US and currently, throughout Eur-
ope, the Treaty of Rome. However, notwithstanding such federal or supra-
national policy making, the organization of sport has some common traits,
although it is not granted any special or distinct status compared to other
activities. Various central/federal ministries contribute to the organization and
development of sports policy, some with specific sports remits and others as
part of broader portfolios. Likewise, direct government involvement in sport is
also mediated at local government levels and, in some countries such as
Germany and Finland, local government involvement dominates. The deliv-
ery of sports policy, moreover, is undertaken by various policy bodies which,
either jointly or distinctly with concerns to promote elite sport, directly seek to
promote and fund mass participation sport. Traditionally this has comprised
working with educational institutions, governing bodies and their clubs, but
as discussed further below, increasingly recognizes that much participation
now takes place by individual consumers who pay to participate in classes at
leisure clubs or who participate in a highly informal and self-organized man-
ner. It is clear, therefore, that there is a mixed economy in the supply of mass
participation sport.
Table 5.1 provides evidence of the relative mix of the sectors supplying
sport by indicating the location of participation, for a number of countries
and varieties of dates. It should be noted that blank cells indicate a lack of
data rather than the sector being of no importance and that in the final
column participation that takes place in sports clubs is, of course, another
form of public-sector provision, supported by large numbers of volunteer
activity.
While there is considerable variability in the data, not least of which
concerns the period of reporting, the data does show that informal sports
participation is the largest sector of activity, followed closely by sports
clubs. Each of these elements of mass participation supply is now investi-
gated in turn.
The Public Sector 119
TABLE 5.1 Location of sports participation
Country Public (including Private Informal Sports club

education and
local authority)
*
England (2002)1 15% 16% 49% 20%
*
Germany (1994) 3% 6% 64% 27%
*
Austria (2000) 5% 74% 21%
Finland (2001) 6% 10% 75% 16%
Denmark (1998) 9% 38% 84% 49%
Ireland (1987–2002) 33%
Netherlands (2001) 52%
Belgium (1999) 68% 49%
Portugal (1998) 36% 24% 15% 43%
Spain (2000) 66% 25%
Lithuania (2001) 49% 29% 23%
1
Source: Carter (2005); others van Bottenburg et al. (2005).
*
Percentage of total.
5.3 THE PUBLIC SECTOR

Table 5.2 indicates some comparative data on the current scale of the public
sector's involvement and financing of sport. While the numbers in Table 5.2
look large, they should be seen in perspective. For example, the Sky television
deal for broadcasting premiership matches between 2002 and 2007 comprised
£1.024 billion. An important feature of Table 5.2 is that both investment in
mass participation sport and elite sport development is included in the public
sector figures. The allocation of funds to these sectors reflects a potential
policy trade-off as is now discussed.
TABLE 5.2 International comparisons of sport funding
Country Central/federal Regional/local Lottery/gaming Total Per

£ billion £ billion £ billion £ billion capita £
*England 0.044 1.914 0.242 2.2 42

Germany 0.08 2.37 0.01 2.46 30
France 1.60 4.85 0.15 6.60 110
Finland 0.12 0.26 0.06 0.44 84
Canada 0.20 1.92 0 2.12 66
Japan 2.23 3.26 0.06 5.55 44
Australia 0.08 0.74 0 0.82 43
Source: Carter (2005) except * DCMS/Strategy Unit (2002).

5.3.1 UK sports policy and provision

Current sports policy and provision in the UK underwent a significant over-
haul following the publication of Game plan, a strategy for delivering Govern-
ment's sport and physical activity objectives in 2002 (DCMS/Strategy Unit,
2002). This document identified two main objectives for government policy
(p. 12) which have recently been reaffirmed in Playing to win: A new era for
sport (DCMS, 2008):
1. a major increase in participation in sport and physical activity, primarily

because of the significant health benefits and to reduce the growing costs of
inactivity;
2. a sustainable improvement in success in international competition, par-
ticularly in the sports which matter most to the public, primarily because of
the “feel-good factor” associated with winning.
5.3.1.1 Participation
A number of potential interacting benefits are noted that arise from the public
provision of mass sport and physical activity. These include enhancing:
& personal satisfaction and social life, by promoting subjective well-being,
citizenship and productive economic behaviour;
& health, by directly tackling issues such as obesity and diabetes, as well as
psychological well-being;
& education, by using sport to improve cognitive, emotional and motivation-
al impacts on academic performance;
& crime reduction, by displacing time, reducing boredom, enhancing self-
esteem, improving cognitive skills, providing alternative peer groups and
creating positive relationships with “significant others;”
& social inclusion, by developing communities, reducing deprivation and
building social capital;
& the environment, through urban and community regeneration.
5.3.1.2 Elite sport

A number of interrelated benefits are also proposed to follow from the public
provision of elite sport. As far as international sporting success is concerned, a
number of impacts are identified:
& feel-good factor, raising well-being, productivity and social capital;
& economic performance, reflected in greater consumer confidence, produc-
tivity and share prices, at least temporarily;
& improving the image of the UK to promote tourism and inward
investment.
The Public Sector 121
These impacts could also be, of course, connected with the hosting of elite
sports events as opposed to merely competing in these events, which is dis-
cussed at length in Chapter 12. Suffice it is to say at this point that, as a result
of these potential impacts (referred to as legacies in the case of the Olympic
Games), it is argued that there is positive feedback between such investment
and participation in sport, thus:
“. . . it is not possible to say that increasing mass participation will automatically
improve international success, or that international success will necessarily
drive mass participation. Both issues need to be tackled separately, leading to
a twin track approach.” (DCMS/Strategy Unit, 2002 p. 84)
As Box 5.1 illustrates, however, there is the presumption of positive feed-
back.
5.3.2 Other sports policy and provision

The approach adopted in the UK is contrasted with three other current
approaches in DCMS/Strategy Unit (2002). These include:
& a laissez-faire approach in the US in which sport is not considered to
be a federal government concern, consequently, no state support is
offered;
& Finland, in which sport is seen as central to social policy and active recre-
ation encouraged; and
& Australia in which elite sporting success is presented as central to national
identity.
BOX 5.1 AIMS AND OBJECTIVES OF THE TWIN TRACK APPROACH
Source: DCMS/Strategy Unit, 2002, p. 84.

FIGURE 5.1 Sports policy trade-off.
This is not to suggest that the benefits of mass participation, as recognized

above, are not perceived to be important in the US and Australia (see also the
Conference Board of Canada, 2005), but that the balance of emphasis is clearly
different, which ultimately depends on the different assumptions made about
the responsibility of the individual or state to produce these benefits, and
indeed, this balance has changed over time as with the UK.
Figure 5.1 illustrates the economic tension involved, making use of a mod-
ified version of the consumer choice framework developed in Chapter 3. For
illustration here, the vertical axis represents the number of medals earned at the
Olympic Games as a direct consequence of funding, while the horizontal axis
represents the number of leisure and fitness facilities provided. The curved line
A–B represents the opportunity cost frontier for the use of public funds, when
the maximum number of either medals or leisure fitness facilities, or a combi-
nation of both, is identified for a given total value of investment, as indicated for
example, by the penultimate column in Table 5.2. Points on the curve represent
points of productive efficiency in the use of resources, as discussed in Chapter 1.
This means that points to the left of the curve indicate an inefficient use of
resources, and points to the right of the curve are unattainable. The curve is thus
described as the production possibilities frontier. It acts as a budget constraint
facing the public policy agency. Unlike the linear budget line facing the con-
sumer presented in Chapter 3, however, the frontier is curved.
Why will the production possibilities frontier be curved concave to the origin?
Hint: Each of the alternatives, medals or leisure and fitness centres, are outputs from a
production function. Each output will be produced, in theory, according to diminishing
marginal productivity.
The Economic Rationale for Public Policy 123
The curved slope reflects the diminishing marginal productivity of

resources being increasingly invested in either the accumulation of medals
or the building of new leisure and fitness centres. As discussed in Chapter 1,
each of the two alternatives of medals or leisure and fitness centres can be
viewed as the outputs of a production function in which the resources of land,
labour and capital are committed by the policy maker. As more of each re-
source is applied to produce either, it is assumed that their output increases at
a decreasing rate. Consequently, if one begins from a point such as “X,”
reduces the building of leisure and fitness centres and reallocates the resources
to producing medals, the resources will have a larger effect in the latter case.
The opposite would occur if beginning at a point such as “Y,” where more
resources were initially invested in winning medals, and then some resources
were reallocated to building leisure and fitness centres. The slope of the curve
thus represents the marginal rate of transformation of resources.
The marginal rate of transformation is the rate at which one good can be transformed
into the other. The slope thus represents the opportunity cost of one good, compared
to another.
The indifference curves, P1 and P2, drawn on the figure, represent policy
makers' preferences for combinations of the use of resources. They can be
viewed as representing a social welfare function, which is discussed further
in the next section. The slope of the indifference curves or marginal rate of
substitution will reflect the relative subjective utilities received by the policy
maker as they switch from one policy objective to another. Any point of
tangency between the indifference curves and the production possibilities
frontier thus represents the optimal choice for the government. As drawn,
they indicate either a change in preferences over time towards medal accumu-
lation rather than leisure and fitness centres from points “X” to “Y” respec-
tively, or the difference between Finland and Australia's policy priorities.
5.4 THE ECONOMIC RATIONALE FOR PUBLIC POLICY

While the above discussion has charted the extent of public sector intervention
in the sports market, and indeed presented some of the specific reasons used to
justify why this is the case, the general economic logic and issues underlying
these presumptions or interpretations of evidence has not been made clear, but
it resides in some of the issues discussed in Chapter 1 under the heading of
normative economics. This is, in fact, an issue explicitly identified in DCMS/
Strategy Unit (2002) in which economic efficiency and equity are identified as
possible reasons for intervention. Each of these concepts is now discussed in
more detail.
5.4.1 Economic efficiency

It was argued in Chapter 1 that economic theory presents maximum econom-
ic welfare for society, or efficiency, as consistent with a competitive market
allocation of resources where both productive and allocative efficiency are
implied. Outputs are produced at lowest possible cost and the prices at which
resources are traded reflect the true benefits and costs to society. Key to facil-
itating these results is that property rights to resources can be allocated clearly
to those engaging in market transactions and those that do not. Property rights
can now be more formally defined as:
“. . . the rules (whether formal and legal or informal custom) which specify
which individuals are allowed to do what with resources and the outputs of
those resources. Property rights define which of the technologically feasible
economic decisions individuals are permitted to make.” (Gravelle and Rees,
2004 p. 9, italics in original).
It should be added here that the permissibility of property rights to facilitate
transactions logically depends on their being defined or measurable. If they
are, then they define private goods which can be bought and sold on markets.
Chapter 1 suggested two contexts in which this was not possible. This is when
consumption decisions are non-rival and when they are non-exclusive. As
presented in Figure 5.2, a variety of economic goods exist.
This suggests that, in practice, according to degrees of non-rivalry or ex-
cludability, a variety of types of goods exist with some spillovers to other
parties than those engaged in a transaction. Externalities, as discussed in
Chapter 1, are the key spillovers. On this basis Table 5.3 indicates the nature
of the externalities involved according to the suggested benefits of increasing
mass participation noted earlier, and how public sector provision of sport may
correct the impacts.
What mechanisms could be used to achieve these adjustments, to harmo-
nize marginal private and marginal social costs and benefits? It is clear now
that three main policy tools could be used. As discussed in Chapter 1, taxes or
subsidies could be employed to adjust the market price either up or down to
correct for the overprovision or underprovision, respectively. Underprovision
could be directly corrected for by the public sector supply of the activity directly
or through incentives and subsidy from the public sector. Finally, information
FIGURE 5.2 Types of economic good.

TABLE 5.3 Externalities and participation in sport
Benefit of State of externalities Corrective impact of sport

participation
Personal satisfaction MSB > MPB Enhance personal and social capital
and social life
Health MSC > MPC Increase human capital/productivity
Decrease healthcare costs
Education MSC > MPC Increase human capital/productivity
Increase social capital
Crime reduction MSC > MPC Increase social capital
Decrease policing costs
Social inclusion MSC > MPC Increase social capital
deficiencies can also be targeted to allow better informed decisions to be made

and to help in establishing property rights.
5.4.2 Equity
Equity in sport has traditionally been promoted in policies of “sport for all,”
such as the explicit commitment to this aim in the 1975 Council of Europe
Charter and implicitly in the European Sports Charter (DaCosta and Mir-
agaya, 2002). In the context of the discussions of Chapter 1, to maintain
that equity provides a rationale for public policy in sport, therefore, requires
a presumption that the current combination of resources available to con-
sumers is first, not adequate for them to consume the amount of sport that
they would otherwise wish to or secondly, that they should consume a
different amount than they currently do. In the former case, concerns about
equity overlap directly with concerns for social exclusion, as emphasized by
Collins (2003, 2004), inasmuch as a lack of income can constrain partici-
pation in sport. It is clear, however, that while income is important, the New
Household Economic approach of Chapter 3 directly recognizes that access
to market goods and time also matter, because they form the inputs to the
household production function. It follows that equity is connected more
generally to the presence of constraints on individual choice and, in this
regard, the policies required to correct for inequity can be the same. Conse-
quently, while taxation or subsidy can be used to adjust the prices paid for
access to the good or service, a direct focus on changing income differentials
might target both horizontal and vertical equity. Willingness to pay is thus
balanced against ability to pay.
Equity is linked to the concept of merit goods, which are goods judged by an
individual or society to be more appropriate than those that would have been
chosen according to consumer preferences, which are measured by willingess
to pay. It is on this basis that it has often been argued that the direct provision
of such goods is more acceptable than a reallocation of income (Musgrave and
Musgrave, 1973; Musgrave, 1987). As Cicchetti et al. (1969) note, merit goods
are essentially a redistribution of income in specific commodity forms.
Notwithstanding these distinctions, however, the above discussion sug-
gests that equity and economic efficiency are to an extent, in conflict, if
willingness to pay is not also an expression of ability to pay. However, the
forms in which this applies are quite complex and depend, to an extent, on
what is implied by equity. The point can be illustrated by examining alterna-
tive concepts of the maximization of social welfare other than the outcomes of
a market allocation of resources, which have been described as efficient.
Box 5.2 indicates some alternative social welfare functions that are discussed
in the economic literature.
A social welfare function ranks alternative social states, as a utility function for
individuals ranks alternative uses of resources.
It is significant to note that they have different implications for both

vertical and horizontal equity. They also refer either to resources or to the
utility experienced from the consumption of resources. Ultimately, policy
makers implicitly or explicitly adopt such a function in making resourcing
decisions. A market-based approach, emphasizing efficiency, could therefore
be identified with a utilitarian approach.
Technically, however, the utilitarian approach had its theoretical origin
in the addition of measurable cardinal utility. Consequently, Bergson (1938)
and Samuelson (1947) proposed a general social welfare function compris-
ing the ordinal utility of all members of society. The Pareto criterion then
argues that if one individual's utility increases, other things equal then so
must social welfare. But this is simply one ethical judgement, and as
Samuelson (1947) argues, the function is consistent with other ethical
beliefs. It should be noted that the social welfare function is always specified
with respect to a given income distribution. Significantly, therefore, com-
plications apply to any attempt to evaluate redistribution policies, discussed
BOX 5.2 CONCEPTS OF SOCIAL WELFARE

Egalitarian: all members of society receive equal goods and Utilitarian: maximizes the sum of individual utilities;
services; Rawlsian: maximizes the utility of the least well-off.
for example by Kaldor (1939), Hicks (1939) and Scitovsky (1941).1 Perhaps
more seriously, Arrow (1950, 1963) has shown that there is no social welfare
function derived from individual preferences that satisfies four relatively
uncontroversial conditions including: that the social welfare function can
apply to any pattern of preferences; that the social welfare function is con-
sistent with the Pareto principle; that the social welfare function ranks
states consistently regardless of how states are presented as choices to
individuals; and non-dictatorship. Dictatorship implies that a particular
individual's choices become decisive.
It follows that attempts to derive a policy preference based on the aggrega-
tion of individual preferences cannot be achieved without compromising at
least one of these conditions. Consequently, Figure 5.1 can perhaps best be
seen as indicative of the policy makers' preferences and beliefs as opposed to
individual's in society in as much that they reflect choices made by elected
representatives.
5.4.3 UK policy revisited: Mechanisms to promote participation

Having explored the economic logic for public sector intervention in the sports
market, an examination of the mechanisms proposed in the UK is now un-
dertaken to reveal the logic of typical policy mechanisms that have been used
or could be used to boost participation. Figure 5.3 reveals that the DCMS/
Strategy Unit (2002) seeks to target both the demand- and supply-side factors
that affect sports participation, also in the context of both human and physical
capital.2
In the context of demand it is clear that one option is to enhance the
information set by which individuals make choices. While this enhancement
1
Kaldor (1939) and Hicks (1939) suggested “compensation” criteria to try to rectify the
compromise to Pareto efficiency implied in a redistribution of resources, which will
inevitably imply that different individuals in society gain and lose as incomes are reallocated.
The criteria imply examining the potential for gainers to compensate losers, or vice versa to
facilitate or avoid changes in resource allocation respectively. The idea has analogies with the
income effect discussed in consumer choice in Chapter 3. However, the criteria developed
were not invariant to the direction of the change in resource allocation, prompting Scitovsky
(1941) to develop a criterion in which potential gainers can compensate the potential losers
and still remain better off, and potential losers could not offer incentives to gainers to forego
the change.
2
The emphasis here is on adults aged 16 years or over. Separate strategies for school children
aged 11 years or under and between 11 and 16 years of age are proposed as part of the
curriculum. The real strategic issue is considered to arise once compulsory schooling
finishes, when the constraints on participation begin to emerge which seek to build on
previous experience of sports at school.
FIGURE 5.3 UK sports participation policy options.
of information can be undertaken through social marketing, that is direct

campaigning, it is also implied that this could take place directly through
the supply side. In this case, those who do not participate in sport might also
be those who have direct contacts with community groups and health profes-
sionals who might be dealing with the consequences of social exclusion,
criminality and poor health. This would provide opportunities to promote
sports participation.
The enhanced training of such staff, and others at various sports facilities
might also send the signal (as discussed in Chapter 1) that enhanced quality of
provision is available, and thereby increase participation. As detailed in Sec-
tion 5.3.1, and revealed in Figure 5.3, moreover, there is a general focus on the
need for increased efficiency in the supply of sports in policy discussion. This
is manifest in a number of ways. In part it is proposed that efficiency gains
could be achieved by making greater use of existing educational public facili-
ties in the community that have previously been solely for use in education.
However, the growth in concern for increasing efficiency has also developed
commensurate with a decline in the emphasis on supplying sports facilities
directly by the public sector. This reflects a policy view that the private sector
may provide services more efficiently than the public sector or at least that the
basis of its decision making, with clear commitment to accountable targets,
has relevance to the supply of sport. This perspective gained momentum in the
1980s in the UK as part of a broad policy sweep towards privatization and the
deregulation of markets that were dominated by monopoly government agen-
cies. In the UK this emphasis has been manifest in policies such as compulsory
competitive tendering (CCT) for local authority services so that the provision
of sports and leisure services could only be retained by the public authorities if
they had competed for the right to provide such services with other private
sector companies according to a sealed bid auction. The only exceptions were
facilities provided solely for educational and community group establish-
ments. As Henry (2001, p. 140) notes, the current policy emphasis is on “best
value”(BV) to remove the excessive emphasis on costs that had occurred with
CCTand:
“. . . to secure continuous improvement in the way in which its functions are

exercised, having regard to a combination of economy, efficiency and
effectiveness.” (Sport England,
1999b, p. 3)
How are efficiency, economy and effectiveness related to economic efficiency?
Hint: Think about resource, use, costs and objectives, say, in perfect competition.
It should be noted that all three of these terms are subsumed in the
economic concept of efficiency. If one believes that Pareto efficiency cur-
rently exists because of the presence of a competitive market, then it
follows that the production function describing the use of resources to
produce outputs must describe maximum technical efficiency, i.e., the
minimum level of inputs are being used to produce the maximum level
of outputs. If the markets for factors of production are perfectly competitive
then the nominal value of the resources expressed as wages, interest rates
and rents will be at their lowest possible level. These two conditions in
turn mean that economic costs, which are the nominal values of factors of
production divided by their marginal productivities (see Chapter 1, Equa-
tion 1.1), will be at their lowest level. Finally, the objective of suppliers, i.e.,
profit maximization, is achieved under the conditions where marginal

revenue is equal to marginal cost.
It should also be noted, however, that this does not mean that BVautomat-
ically proposes a competitive allocation of resources or the aspiration to profit
maximization. The welfare reorientation of BV does not necessarily target the
Paretian ideal. A different objective might be set that does not target profits and
that may involve trading-off the interests of particular stakeholders in society,
for example, because of equity criteria. If this objective is achieved, then
effectiveness is achieved. This is consistent with striving for the “efficiency”
and “economy” of supply provision as far as achieving that particular target
would allow. In this regard BV can be seen as a compromise between seeking
economic efficiency, as defined in technical economic terms, and achieving
policy goals, perhaps associated with the access issues that are not likely to be
of interest to purely commercial facilities.
Figure 5.4 illustrates the main principles involved in which two types of
customers are presented as applicable to a sports centre, those able to pay a
higher price and those able to pay a lower price. This might apply in situations
where there are socially advantaged and disadvantaged potential participants
or customers. The price that they each pay for the use of the facility could give
them access to a particular activity or any bundle of activities and is considered
essentially the same for individuals belonging to either of these groups. It is
assumed that there is a constant AC for the supply of these facilities. In this
case the equivalence of the AC and MC implies that only average variable costs
(AVC) are considered. This would correspond to the required comparison in
which a local authority owned facility was being run either by a private com-
pany or by local authority employees. The implication is that the fixed costs of
the facility are not being treated as part of the comparative profit or loss
calculations concerned.
Recalling the discussions of Chapter
1, it was noted that profit maximizing
behaviour, as might be expected of a pri-
vate company, would imply targeting the
provision of output where MR = MC.
This must occur at the point indicated
“private” in Figure 5.4. The implication
is that further use of the facilities would
require making them accessible to those
not able to pay the relevant price, which
would not be in the interests of the com-
pany. In contrast, a public sector or wel-
FIGURE 5.4 BV and policy outcomes. fare ethos might have as its goal
increasing access to sport, for the reasons discussed earlier. In this respect the
facility might be used up to full potential capacity, to the point indicated
“public.” It is clear, however, that in order to achieve this target prices would
have to be lower to attract such participants. However, this situation would
lead to MR < MC, with the consequence of not maximizing profits. The
facility would make an overall profit or loss depending on the relevant size
of the areas marked “profit” or “loss” if two levels of price were established
(see the discussion of price discrimination below). A larger loss would be made
if the same lower price were charged to all consumers.
What would be the extent of loss if one price were charged?
Hint: Consider the difference between AR = MR and AC = MC.
In the latter case, if the lower price were charged all users would pay less
than AC = MC and as a consequence the loss would correspond to this dif-
ference times the number of activities participants. Significantly, BV suggests
that loss in itself is not a problem. On the contrary, what matters is being
effective, i.e., achieving objectives. In this regard, if the objective is to maxi-
mize participation in sport, such a loss is of no concern. Moreover, it is clear
that regardless of which objective is sought, economy and efficiency will help
to achieve this, because if AC = MC is at its lowest possible value consistent
with the policy target, then either profits are maximized or any losses that
occur are kept at a minimum. The latter case would reduce the scale of any
subsidy required to maintain the use of the facilities and reduce the opportu-
nity cost of investing resources elsewhere.3
A significant implication of this approach is that the ownership of the asset
is not considered to be important per se, and this explains the shift in policy
emphasis away from the provision of facilities. The key issue for policy is
access to facilities as a policy objective. This has helped to reinforce the growth
3
In the current case, because MC = AC there is an assumption that factors of production
exhibit constant marginal productivity, so the efficiency and economy of profit
maximization and participant maximization would be equivalent. However, this would not
be the case if diminishing marginal productivity applied. Here, MC rises with increases in
output, so it follows that output levels above the MR = MC level will be supplied at higher
MC and AC. Losses will be higher than the diagrammatic case for the facility and implied by
perfect competition. So, in economic efficiency terms, any equivalence between productive
efficiency, cost minimization (economy) and allocative efficiency (prices reflecting
opportunity costs) is compromised.
of private sector provision in the UK, as discussed in the next section. The use
of the price mechanism to promote participation is also emphasized in the
other options to promote participation in DCMS/Strategy Unit (2002, p. 108).
These include:
1. part funding participation in activities through employer subsidies or di-
rectly subsidizing employees;
2. encouraging firms to be flexible on work times to help promote participation;
3. subsidizing individuals directly through the use of GP referrals, swipe cards
and vouchers for use at sports facilities.
Each of these mechanisms either reduces the cost of participation directly or
indirectly, for example, by allowing individuals to reallocate work and house-
hold work time. They target specific individuals to whom a price reduction
should apply. If the target for policy is to promote more participation, then it
can be shown that this will be the most effective tool. Consider Figure 5.5; it
presents the consumer choice framework developed in Chapter 3. Here, the
two economic goods being considered are visits to sports facilities and other,
non-sport, consumption goods. To respect individual preferences, a general
government payment might be made to a disadvantaged individual who does
not currently participate much in sport. Alternatively, the price of access to
sports facilities only might be subsidized as implied above. It can be shown
that the latter will promote more use of sports facilities.
The first of the two policies is consistent with the income effect discussed in
Chapter 3 and moves the budget constraint out from the origin in a parallel
fashion. It is possible that some extra sports
participation might follow, but also the con-
sumption of other goods will increase as the
income is distributed across consumption
decisions. In principle, less sport might even
be chosen, which is the case drawn for those
having weak preferences for sport, as is likely
for disadvantaged groups (see DCMS/Strate-
gy Unit, 2002, p. 105). In this case, partici-
pation is considered to be an inferior good by
the individual. In contrast, if only visits to
sports facilities are subsidized, then this is
FIGURE 5.5 Payments versus subsidies. equivalent to a substitution effect. The in-
come constraint pivots through the initial
4
Resources fall if fewer visits are undertaken. This implies additional opportunity costs
from not increasing visits to sports facilities.
The Private Sector 133
equilibrium to indicate that resources have increased if further visits to sports

facilities are undertaken.4 Here more sports participation must follow.
There are a number of caveats that should be added in closing this discus-
sion. The first, and most obvious, is that if such subsidies were applied to all
visits to sports facilities, this might compromise equity objectives and rein-
force benefits to those already participating in sport, despite targeting lower
participating groups. It is clear, therefore, that only those viewed as disadvan-
taged should receive the subsidy. Vouchers would require clear property rights
and not be transferable.5 GP referral schemes would automatically imply this
with named individuals being “prescribed” opportunities to engage in sports
activity. The second qualification is that, just as with equity objectives, such
subsidies place policy priority above consumer preferences generally. This is
because they distort the market prices or values on the use of resources from
those that would have occurred. Despite the policy objectives being seen as
necessary by policy makers or their supporters, there is a direct sense in which
individual preferences are compromised. It is clear that the DCMS/Strategy
Unit (2000) recognizes these issues and argues that:
“We do not suggest that equality of outcomes should be a goal, because people
have different tastes, but we do argue that equality of opportunity should be
a goal.” (DCMS/Strategy Unit, 2002 p. 78)
The general point, however, is that a legitimate case for public policy to
promote sports participation is possible, but its logical underpinnings and
justification should involve making explicit arguments about the nature of
market functioning. In particular, informational problems need to be pre-
sented which provide a basis for the presence of externalities, which might
then be consistent with, resulting from, or resulting in, other constraints on
efficient economic behaviour. An explicit concept of equity might also need to
be made clear. In general, policy solutions that target equity or efficiency
concerns through subsidies to price signals require recognizing that policy
makers preferences are promoted, as opposed to those of all consumers as
would be presented in a market allocation of resources.
5.5 THE PRIVATE SECTOR

It has been noted that there has been growth in private sector provision of mass
participation. This finding, coupled with the concern that there is no evidence
of any major longitudinal growth in mass sports participation in the UK,
5
If vouchers were transferable then the Coase theorem would predict that recipients would
allocate them to others with a greater wish to engage in sport in return for some other benefit,
perhaps cash or goods in kind.
TABLE 5.4 Sports participation 1987–2002 (%)
Activity 1987 1990 1993 1996 2002
Swimming 13 15 15 15 14
Keep fit/yoga 9 12 12 12 12
Weight training* 5 5 5 6 5
Soccer 5 5 4 5 5
Golf 4 5 5 5 5
Tennis 2 2 2 2 2
Badminton 3 3 3 2 2
Lawn/carpet bowls 2 2 2 2 1
Squash 3 3 2 1 1
Table tennis 2 2 2 2 1
At least one activity (excluding walking) 45 48 47 46 43
Source: Fox and Rickards (2004).

*
Combined with weight lifting until 1996.
suggests a changing composition of demand over time as much as any sus-

tained secular change. Table 5.4 presents some comparable time-series data for
mass participation, for adults aged 16 years or above in the four weeks before
the interview for the GHS, for a number of activities that might be associated
with public or private sector facilities provision. The data indicate a general
stagnant level of participation, with some apparent structural shifts of inter-
est, for example, away from squash.
This would suggest that any growth in the private sector supply of this
market has arisen largely from changes in a previously public sector industry
to the mixed economy noted earlier, commensurate with the emphasis on
compulsory competitive tendering and best value, producing market segmen-
tation across suppliers reallocating the derived demand for facilities.
It is difficult to identify general supporting evidence for this proposition,
but it appears to be the case for health and fitness clubs, leisure centres and
swimming pools in the UK (Mintel, 2005a, 2006). The commercial evidence
that is publicly available suggests that private health and fitness clubs operate
at the “top end of the market,” while local authority centres, evolving primarily
from swimming pool provision and the earlier policy desires of the 1970s to
provide public sector facilities, are supplying the other elements of the market.
Significantly, there is evidence that it is in this sector that participation de-
mand is weak. Such inferences are borne out by an examination of the supply
provision indicated in Table 5.5, in which participation in the last 12 months
by adults aged 15 years or over at various locations is reported.
While participation is undertaken at multiple locations, consistent with
more than one sport being undertaken by individuals, it is noticeable that
TABLE 5.5 Location of sport provision
Location 1998 % 2000 % 2003 % 2005 %
Local authority sports centre 50 48 46 38

Private club or gym 32 33 34 32
Local park 23 22 22 24
Home 24 23 21 22
School/college/university 17 16 14 13
Local authority sports ground 12 9 11 8
Place of employment 8 7 6 4
Elsewhere 31 25 25 26
Sample size 1126 1187 1062 971
Source: Mintel (2005a).
participation at local authority centres has actually fallen in relative terms

since the time when BV was introduced, and this is also the case in other
locations. The exceptions are participation in parks, home and private sector
provision per se, which appears to account for a constant share of demand
usage. The stable market sectors appear to be informal participation, dis-
cussed further below, and the private sector. It is important to note, in this
regard, that the declines are not necessarily commensurate with reduction in
supply opportunities in the local authority sector. Mintel (2005b) estimates
that there were 2596 local authority centres in 2004, with an expected addi-
tional 128 in 2005. Moreover, 165, 124, 91, 42 and 79 were each added in the
previous years, suggesting investment and rising supply capacity. Significant-
ly, it is reported that this additional capacity is being operated by private sector
management. While 616 of the centres were run under contract in 2003, 129
were run by in house “Direct Service Organizations,” 332 by private leisure
management contractors and 155 by trusts or volunteers.
These patterns would suggest that this sector is attempting to compete
with the private sector with the suspicion from the data that such
hybrid organizations are struggling to cope with the aims of “sport for all”
as a social and welfare policy, while seeking to operate on commercial
principles. In contrast, a buoyant but consolidating private sector
segment of supply has emerged, as detailed in Table 5.6. There is a
market leader, Fitness First, and a core of multiclub suppliers each offering
clubs of similar size and operating on a membership basis. The economic
principles that can describe both of these market characteristics are now
discussed.
TABLE 5.6 Private sector health and fitness clubs 2007
Company/owner Brand(s) Clubs Members Average

members/club
Fitness First Fitness First/Fitness 183 452 000 2470

Holdings Ltd First for Women/Kaizen
Whitbread plc David Lloyd 59 325 000 5508
Virgin Active Ltd Virgin Active 72 288 301 4004
LA Fitness plc LA Fitness/Promise 87 261 000 3000
Esporta Group Ltd Esporta 53 220 000 4151
Bannatyne Fitness Ltd Bannatyne Health 61 180 000 2951
Club/Just Fitness
Cannons Health & Cannons 52 177 000 3404
Fitness Ltd
JJB Sports plc JJB Health Clubs 39 175 000 4487
Source: Mintel (2007).
5.5.1 Market structure

The characteristics of the market presented in Table 5.6 are common across
many industries and can be described as an oligopoly.
An oligopoly is a market dominated by a small number of sellers, each referred to as
oligopolists. This contrasts with monopoly in which the market is dominated by one
seller. The Greek prefix (oligo or mono) thus distinguishes the market.
Box 5.3 briefly outlines some alternative theoretical models of oligop-

oly that have been used in economics to analyze their behaviour. Unlike
the largely hypothetical models of perfect competition and pure monopoly
presented in Chapter 1, oligopoly models try to capture elements of real-
ism through their descriptions of markets, while retaining the assump-
tions that rational economic agents both have perfect information and
optimize their decisions. A further key difference is that, because oligopoly
models focus on competition between a relatively small numbers of sup-
pliers, the reactions of suppliers to the actions of other “known” suppliers
have to be directly accounted for in the analysis. In the case of perfect
BOX 5.3 TYPICAL OLIGOPOLY MODELS

Cournot model: named after Augustin Cournot, this model Stackelberg model: named after H. von Stackelberg, this
originally addressed competition between two firms, and model is concerned with competition over output in the
concerns competition over output levels in the industry. industry when one of the firms is a market leader.
Bertrand model: named after Joseph Bertrand, this model
is concerned with competition over prices.
BOX 5.4 ELEMENTS OF GAME THEORY

Some key elements of game theory include the following: 3. Nash equilibrium
Named after the famous mathematician, John Nash, a
1. Non-cooperative versus cooperative games
Nash equilibrium is a dominant strategy in which agents
A game is cooperative if binding contracts can be written
are doing the best that they can given the actions of the
as a solution between conflicting objectives. If the cost of
other agents. A Nash equilibrium does not have to exist
a private fitness instructor's services are, say, 10 per
for games and there can be multiple equilibria.
hour, but the instructor was seeking a fee of 20 per hour,
4. Repeated games
in practice any agreed fee between 10 and 20 with
Many economic transactions are undertaken repeated-
customers is possible as both parties could be better
ly. To account for this, rather than viewing the situation
off. Essentially a bargain over consumer surplus and
as a “one-off” game, a repeated game can be modelled.
profit takes place.
If this is the case then reputations for behaviour emerge.
In contrast, in non-cooperative games this is not
Different behaviour can be predicted if a finite or infinite
possible.
horizon for the game is proposed.
2. Dominant strategies
5. Sequential games
To analyze non-cooperative games, the game's out-
If agents make decisions and act simultaneously, then
come will depend on the strategy adopted by economic
the game is different than if agents make decisions in
agents. A dominant strategy is an optimal strategy re-
sequence.
gardless of what other agents choose to do.
competition, suppliers have no discretion to act different from the market,

but take the market price as given. In monopoly the firm is the market,
and can set the price according to the level of possible demand. In
oligopoly situations it is theorized that large firms have a degree of
monopoly power only. It follows that different models will yield different
specific predictions.
To understand the interaction between suppliers, or indeed any agents,
economists make use of game theory. Box 5.4 indicates some general char-
acteristics of game theory, with some more detail presented in Appendix 5.1.
For the purposes of this book, the broad point that matters is that oligopoly
markets can generate competing incentives. On the one hand, profitable
opportunities exist for the suppliers to collude, tacitly if not formally, to act
as a monopoly to raise the profits of the set of firms. The profit can then be
distributed. It remains, however, that there will always be counter incentives
for competitive innovation. An individual firm could, say, cut prices to capture
more of the market and thus attempt to make higher individual profits than
previously, at the expense of the other firms. This can make such collusive
behaviour unstable and coordination in a cartel difficult to manage.6 It
6
The stability of cartel behaviour in professional sports is discussed more generally in
Chapter 7.
should also be noted that explicit coordination is contrary to economic com-

petition regulations. There is no evidence of such behaviour in the UK fitness
market.
In the light of such competing tensions it is often argued that oligopoly
markets will be characterized by product differentiation and non-price com-
petition to avoid “Bertrand” price competition, and to compete over market
shares and to defend market segments as implied in “Cournot” competition.
First mover advantages do exist, however, such that market leaders can emerge
as suggested by Stackelberg.
5.5.2 Pricing
The tariff structure operated by most sports facilities is evidence of a lack of
simple price competition in private sector sports facilities. Typically, a two-
part tariff structure is employed, which can be viewed as a form of price
discrimination. This form of tariff structure implies making a membership
payment and then some subsequent payment for the use of facilities, rather
than just paying a fee for use, which is more prominent in the previously or
current public sector facilities (Mintel, 2005b). Such a system of payment can
be viewed as a mechanism to promote loyalty in the use of facilities through
being a member of the club. As discussed in Chapter 3, the consumer can be
viewed as investing through the purchase of a durable good.
Durable goods yield a flow of services or utility over time rather than being used up
with one consumption act.
Why is a sports club membership like a durable good?
Hint: How does the membership yield utility?
The investment nature of the membership fee is implied via a payment

being required to make use of facilities. This payment is written off, i.e., the
value of the membership depreciates over time, regardless of whether or not
the facilities are actually used. Under such conditions the consumer faces a
modified budget constraint in which they first of all make a decision to join a
club or not and then, subsequent to this, make a decision on how many times
to use the facilities based on the usage fee. It is clear, however, that if the
facilities are used the fixed membership charge can be spread over the subse-
quent usage.
The economic problem facing sports facilities suppliers, therefore, is to

attempt to forecast such behaviour and, in seeking profits, to set a high
membership fee and lower usage cost, or vice versa. There is no simple
formula that indicates how profits would be maximized under such circum-
stances, but there is clearly a trade-off involved. It seems logical to argue that
lower membership fees and lower usage fees might generate a large degree of
market penetration, but this might compromise a supplier's identity with the
“top end of the market.” The discussion of social interactions in Chapter 3, for
example, indicated how consumer prestige might be affected by such behav-
iour, as consumers identify with particular characteristics of membership.
Table 5.7 indicates how in the UK target markets are defined in terms of
membership segments.
To summarize the above discussions, in the UK context the private
sector supply of sports participation can be understood as an oligopoly,
competing for members. It is clear that this segment of the industry has, to
an extent, captured the willingness to pay for sports participation, as
evidenced by the membership structure of such facilities. There is,
TABLE 5.7 Evolving marketing strategies
Company Change in strategy
Fitness First Fitness First Platinum: upmarket with a pool

Fitness First Express: pay-as-you-go with no benefits
Fitness First for Women: women-only clubs
Kaizen: older and non-familiar clients
David Lloyd Leisure More emphasis on the health and fitness plus
racquets facilities: upmarket and the emphasis
of current expansion
Virgin Active Rebranding and refurbishing facilities bought from
Esporta and Holmes Place
LA Fitness Competing in value-for-money segment; low
membership fees, plans for brand promotion
Esporta Plans to consolidate around racquet clubs, which
combine with health and fitness facilities
Cannons Health and Fitness Consolidate existing sites
JJB Health Aggressive promotion of new and some indoor
football facilities
Total Fitness Promotion as exercise, hydrotherapy and
rehabilitation facilities, patient referral, as well
as general health and fitness
De Vere Promotion and investment in Village Leisure
Clubs brand, others remaining stable
Source: Mintel (2007).

therefore, clear evidence of the market internalization of consumer sur-

pluses that existed with purely subsidized public provision. In contrast,
previous public sector facilities, or those still in the public sector, now face
a more challenging market segment, comprising those that participate less
or lack the commitment for membership schemes. While the degree of
private sector supply may vary between countries, it seems likely that
similar potential challenges for promoting and facilitating greater overall
sports participation will apply, unless policy can both attract participation
and also match requisite facilities for participants. The key to achieving
this policy balance will be the management of any segments that exist in
the market.
5.6 INFORMAL PARTICIPATION

All of the descriptive data above suggests that informal sports activity is one
of the most stable and largest segments of the market. This form of partici-
pation is the most difficult to classify and to understand because it lies outside
the official governance of organized sport. Box 5.5 provides a definition of
informal sport.
Consequently, the informal sector is one of changing emphases in which
forms of current, officially recognized sports can take place informally or even
new “sports” can develop. This is an important point to note because, as
discussed in Chapter 2, current formalized sport has its origins in informal
activity and such developments continue to occur. For example, in recent years
in the UK “five-a-side football” as well as being played in an entirely self-
organized manner, has also emerged into a branch of commercially organized
activity, facilitated by companies such as PowerLeague Ltd. This has prompted
the existing governing body, the FA, to seek to work with such providers, which
it has recognized as being the source of much of the investment and growth in
BOX 5.5 INFORMAL SPORT

Informal sport: this term describes an activity that displays stand basketball to be, but some of the defining character-
many of the characteristics of “formal sport,” but does not istics of that sport have been compromised.
involve structured competition governed by rules. So two Informal sports facilities: these can include skateparks,
people playing basketball on their local Outdoor Basketball BMX tracks, basketball courts, kickabout areas and multi-
Initiative site may be said to be engaging in informal sport, use games areas. Associated facilities may include youth
because the activity is very close to being what we under- shelters or equipped play areas for younger children.
Source: Sport England, South West (2006)
Informal Participation 141
participation. The aim is to provide governance and standardization of the

“small-sided” game. Similar patterns of development have occurred with other
activities.
Bicycle motocross (BMX) began as a non-motorized form of motocross in
California in the 1960s, and is still regarded as a street activity associated with
youth culture. However, it became an Olympic sport for Beijing 2008 (http://
www.olympic.org). Since January 1993, BMX has been fully integrated into
the International Cycling Union (UCI). A similar pattern of development has
been experienced in snowboarding, growing again out of “sub-cultural” activ-
ity to become an official sport in 1985, with the staging of the first world cup in
Austria. To harmonize contests, the International Snowboard Association was
founded in 1994. Snowboarding is also now an Olympic sport and stages its
own international events.
Examples such as these suggest that the supply of informal sports is po-
tentially quite fluid, but that some movement towards formal organization
appears with the desire to hold competitions. This issue is discussed further in
the next chapter, when the economic logic underlying the origins of sports
clubs and their governance is presented. As far as activity of a purely informal
and non-competitive form is concerned, supply will either be facilitated by the
public or private sectors, as described above in which many activities, such as
swimming, cycling and keep fit, take place either through the newly-developed
organizations just described or simply at home or other convenient locations
such as parkland and the local neighbourhood.
Under such circumstances, as Chapter 3 discussed in detail, inasmuch
as the consumer of sport can also be presented as the producer of sport,
through combining time and market goods, then the distinction between a
discrete supplier and the consumer disappears. To encourage such partic-
ipation then, in part at least, is to also encourage its simultaneous supply.
Significantly, this might not necessarily involve working through the orga-
nized structures of sport. By definition, their formality does not apply. As
implied in Section 5.4, therefore, any policy aimed at relaxing the socio-
economic constraints on individuals will at least, indirectly, also potential-
ly alleviate the constraints on the supply of informal sport. However, it is
noted in DCMS/Strategy Unit (2002) that targeting preferences would be
required for individuals that have already shown a reluctance to participate
in sports.
It follows also that other policy interventions might be necessary, such as
providing appropriate open spaces in which to engage in the relevant activ-
ities, such as skate parks, exercise trails, etc., (Sport England, SouthWest,
2006) or providing cycle tracks or lanes. It is clear in this context that
“sport” becomes subsumed within the domain of broader social policy in
which transport and urban development play a role. This raises some inter-
esting economic issues, connected with the “externalities” that are being
used to justify the promotion of sports participation, discussed in Section
5.3. Unless the relevant sport and transport agencies work towards common
objectives then a valuable opportunity for progress will be missed. This
provides a rationale for the recent changes in sports policy governance in
the UK, noted in Chapter 2.
Opportunity costs do arise, however. Table 5.8 below reports the sports
club participation rates for a number of activities, including the most
widely participated-in activities and traditional team sport activities. The
second column provides rates of participation across the population. The
last column reports the participation rates of these participants in sports
clubs.
It is clear that sports club activity increases most with more specialized
competitive sports, yet these involve low proportions of the population.
The two largest participation activities, other than walking, are swimming
and cycling yet, relatively speaking, there is very little formal sports club
activity involving these pursuits. While swimming is catered for in leisure
centre provision generally, where activities such as keep fit and even jog-
ging may also be provided, as discussed above these may be more available
through private sector provision, which limits opportunities to those
able and willing to pay for access to the facilities. In contrast, for the
majority of those cycling and others engaging in keep fit and jogging,
formal supply opportunities may be limited. While Figure 5.1 described a
policy trade-off between elite and mass participation sports, it may well be
TABLE 5.8 Sports club participation
Sport Overall participation rate Sports club participation rate
Swim indoors 13.2 3.0

Swim outdoors 3.0 1.7
Keep-fit 12.8 3.5
Cycling 9.5 1.3
Weight training 5.9 3.8
Football indoors 2.0 12.0
Football outdoors 3.8 21.7
Jogging 5.2 3.6
Cricket 0.6 42.4
Rugby 0.5 61.8
Netball 0.3 24.2
Source: General Household Survey 2002 Own Analysis.

Sports Equipment 143
that a similar trade-off applies between formal, club-based activity and

informal activity.
5.7 SPORTS EQUIPMENT

Implicitly the discussion above has centred on facilities provision. As dis-
cussed in Chapter 2, however, the derived demand for equipment will also
reflect current participation rates and, in turn, be connected to the ability of
individuals to supply sports activity. As discussed in Chapter 3 and above, as
market goods, along with time, equipment can facilitate sports participa-
tion/production. It goes without saying that generally speaking, access to
equipment to play golf, sail or mountain bike will be limited according to the
ability to pay. In this respect, the private sector market for equipment is a
highly diverse and global industry in which multinational brands, such as
Nike, Adidas, Trek, etc., compete alongside other more rudimentary brands,
such as Dunlop, Hi Tech and unbranded items. Table 4.8, in Chapter 4,
provided some comparative data on sports expenditure. Suffice it is to say
here that in terms of the UK, Mintel (2005c) suggests that a complex supply
chain provides for such expenditure with channels of distribution as indi-
cated in Table 5.9.
It is clear that the majority of sales are made through sports specialists.
Mintel (2005c) estimates that 35 multi-outlet groups dominate the market,
which also comprises a large number of independent traders. It should be
recognized, however, as discussed in Chapter 4, that despite the location of
the purchases, many of the purchases are unconnected with participation,
particularly in terms of clothing and footwear. Training shoes are often seen
as casual wear and the replica kit market is unconnected with participation per
se. It is more likely that club shops for golf and other team sports and at fitness
TABLE 5.9 Sports good sales
Channel of distribution % of sales
Sports specialists 74
Club shops 6
Department stores 6
Home shopping 5
Clothing specialists 4
Footwear specialists 3
Others 2
Source: Mintel (2005b).

centres are important suppliers of equipment for participants. This is also

more likely with non-club-based activity such as cycling and home gym equip-
ment. As discussed in Chapter 4, however, in general the sports equipment
market seems to be one, in the UK at least, which is probably connected with a
more individualistic and informal participation in sports than through orga-
nized clubs.
5.8 CONCLUSION
In this chapter the supply structure of mass participation sport has been
presented as a counterpoint to the previous two chapters, which explored
participation in mass participation sport as a consumer demand decision.
It is argued that the supply structure is complex, comprising interrelated
and also distinct sectors that have changed over time. Because of the lack
of international research, primary policy and market research documen-
tation has been used to discuss the UK context as a basis for examining
the economics of the supply of mass participation sport. The logic
of public sector involvement in sports provision has been discussed
as being grounded in externalities. An exploration of the oligopolistic
market structure of private sector provision has also been provided. The
informal sports sector has been discussed, along with the market for
sports equipment.
Despite the obvious differences in emphasis that will apply in other coun-
tries, the chapter has highlighted that, in economic terms, policy intervention
requires making the case that property rights are incompletely allocated. If this
is not the case then the private sector, either through the provision of facilities
or individuals engaging informally in sport and purchasing relevant equip-
ment as consumer/producers, will allocate appropriate resources to sport. In
the former case pricing and competitive strategies will emerge as with other
markets. In such a case, regulation of the market would simply require the
usual economic regulation required to ensure that the oligopoly behaviour of
the private sector does not become anti-competitive.
If one argues that property rights are incompletely allocated, then a case for
policy intervention can be made and presented as based on either efficiency or
equity grounds. This is the current policy stance in the UK. In such circum-
stances, the chapter has examined how policy options might correct for the
misallocation of resources by targeting both the demand side and supply side
of the market, with a particular focus on facilities. It has been implied in the
chapter that the current balance between public and private sector facilities
provision leaves the former sector in the potentially difficult role of trying
Conclusion 145
to match competition from the private sector and to meet social welfare
objectives.
The chapter has also hinted at another policy tension which is connected to
the scale of informal activity unconnected to sports clubs, and which may or
may not be met by existing facilities provision. Inasmuch as sports clubs and
facilities are the traditional focus of sports policy makers and policy funding, it
may be that this simply subsidizes multi-sport participants who are able to
access market provision for activities at the cost of other activities that equally
could meet the rationale put forward for intervention in sport, such as health,
but that because of their less specialized nature may also be more accessible to
the currently less active.
Appendix 5.1 Some Elements of GameTheory

Some basic elements of game theory can be illustrated using a hypothetical example of
sports clubs. Consider a market in which there are two main suppliers “Health” and “Fitness”
who seek to maximize their profits. They have to decide whether or not to increase their profits
by attempting to increase their competitiveness, “compete” or remaining as they are “don't
compete.” This might be through increased marketing activities or price reductions. They
recognize, however, that if they undertake such action, then similar competitive retaliation is
likely. The “payoff” matrix, which summarizes the impact on profits for each course of action
is presented below. The columns summarize “Health's” options and the rows “Fitness”
options. Each cell describes the impacts on profits for Fitness and Health, respectively.
Health
Strategies Compete Don't compete
Fitness Compete 10/ 10 +100/ 20
Don't Compete 20/+100 +20/+20
Now if Health doesn't increase competition, Fitness stands to receive either 20 if it doesn't
increase competition or 100 if it increases competition. The same applies for Health. There
are strong incentives, then, for both suppliers to increase competitive activity. However, this
means that both suppliers end up losing 10. This is a Nash equilibrium. In this case the
outcome is suboptimal. It is a well-established result called the “Prisoners Dilemma” and
arises because each supplier acts non-cooperatively with the other, but individual incentives
dominate collective interests.
Should the game be played iteratively, rather than simultaneously, then clearly in a one-shot
context, the firm able to set their strategy first would benefit. This is known as first mover
advantage. The above example also indicates another feature of supplier behaviour, which is
the existence of incentives to collude rather than to be competitive. If suppliers did agree for
example, not to compete, then both would end up better off than the Nash equilibrium. While
this seems unlikely in a “one-shot” context, because of the Nash and first mover scenarios, if
the decisions were made repeatedly over an infinite horizon, then cooperation, not
competition, becomes the rational response. This is because agents realize that any initial
competitive advantage would automatically be removed by sustained retaliatory behaviour in
the next periods. This is known as a “tit-for-tat” strategy. It follows, then, that collusion is
always possible in oligopoly markets, but this may break down with any perceived first mover
advantages becoming apparent. As is discussed in the case of professional sports leagues,
these scenarios have some credibility.
CHAPTER 6
The Supply of Participant

Sport: Volunteers and
Sports Clubs
OBJECTIVES
& To appreciate the structure and diversity of the sports club system
& To appreciate the motivation, profile and value of volunteers in sport
& To understand the economic rationale for volunteering and the origins of
sports clubs
& To appreciate policy debate on the organization and financing of voluntary
sports clubs
6.1 INTRODUCTION
In the previous chapter various mass participation sport supply opportunities
were discussed. One key component of organized sport is the sports club sector
which, as was implied in Chapter 2 (see, for example, Figure 2.1), is where
much of the formal demand and supply of mass participation sport is located
and which acts as a conduit for elite support development. In Section 6.2,
although the literature is relatively sparse and dated, some comparative
insights are provided on the system in various European countries. It is sug-
gested that there is typically a hierarchical structure to sports club systems in
Europe. Section 6.3 then explores the economic logic of the systems, their
hierarchical nature and their general economic rationale. Some comment is
made on their contribution to social welfare. As volunteers are extremely
important to sports club systems, Section 6.4 presents some comparative
insight into the motivation for and scale of volunteering. Following this
descriptive overview of the sector, Section 6.5 explores the economic logic 147
148 CHAPTER 6 The Supply of Participant Sport: Volunteers and Sports Clubs
underlying volunteer activity and its value. The chapter concludes with a
discussion of the policy issues concerning sports club systems.
6.2 THE SPORTS CLUB SECTOR

As implied in Chapter 2, Szymanski (2006a) has argued that sports clubs in
the UK grew out of “associativity,” that is “the tendency of individuals to create
social networks and organizations outside of the family” (p. 1) that became
possible during the enlightenment of the eighteenth century.
The emphasis of the clubs was on competition. In contrast to the UK,
Szymanski (2006a) argues that European sports clubs tended to emphasize
individual mental and bodily health through German Turnen and Swedish
Gymnastics, often overseen by the state with military training as an
objective. Heinemann (1999) argues that these three influences gained
momentum over traditional folk games and pastimes in various European
countries, but ultimately a common form of hierarchical organization
emerged based on the English model, often reflecting the promotion of
“English” sports. In the US, similar associativity developed through sports
clubs, and either evolved into professional activity or remained a community
activity with little formal link between them. In the US, therefore, college
sports became the link between amateur and professional status for athletes,
with local government the agent for overseeing community sport. What fol-
lows, then, is primarily a discussion of a European model of development
carried over to many other countries. However, elements of the model will
apply ubiquitously, such as the links between clubs and their organizing bodies
and links to local government.1
6.2.1 European hierarchical sports organization

Table 6.1 summarizes some (historical) comparisons between country sports
club systems. The main elements of sports governance are listed in the third
column and the number of clubs in the sports club system in the fourth
column. The resources of the clubs are then indicated in subsequent columns.
These are shown to include the providers of facilities, the volunteer base in
aggregate and typically per club, together with the sources of monetary funds.
1
An interesting point to note, that is not pursued further in this book, is that current
emergent professional sports in countries like China, Korea and Japan adopted elements of
the US system of closed leagues and franchises to pool talent, raise competitive levels and
have more commercial viability. This model is increasingly the case with developments in
European sports, as discussed further in Chapter 7.
TABLE 6.1 European sports structure
Country Source Sport governance Clubs Facilities Volunteers Members Typical funding
provision (approximate)
Italy Parro et al. (1999) CONI 39 competitive CONI – 70 535; local authorities 701 795 CONI; 90% from
federations; 16 Others – 4788 betting; sports
associated activities; federations; CONI and
13 multi-sport 30% from sponsorship
networks
Spain Puig et al. (1999) 54 sports federations; 44 509 male predominance federations; 65% from
600 regional (up to 75%); average public sources; clubs;
federations 200 members 32%–59% fees; 68%–
32.8% sponsorship,
lotteries
France Le Roux and Camy CNOSF; 120 160000; 170 000 municipal >1 million average; 60 members federations; 36% public
(1999); federations; 84 finance;; clubs; 60%
Carter (2005) federations public finance
Switzerland Stamm and SOV; 81 sports 28 000 male predominance; SOV and federations;
Lamprecht (1999) 1.5-2 million; average; lottery and sponsorship
60 members; (up to
800 on average)
Germany Horch (1974); DSB; L€ander sport 80 000; 66 500; 60% municipal; 2.6 million; younger male public finance
Heinemann associations (pre-unification); 68%; municipal; 2.8 million predominance;
and Schubert (1999); 88 531; >90 000 average; 300 members;
Carter (2005); Breuer average; 300 members
(2007)
Belgium Taks et al. (1999) Belgian Olympic and 19 000 (Flanders 56%; local authority; male predominance public finance; clubs;
Interfederal 1999); 14 002a 14%; private (60%); average; 67 public finance,
Committee (BOIC); membersa membership,
106 sports federations sponsorship, etc.
(1999); 80 sports
federations (1990)
The Sports Club Sector

149
150
CHAPTER 6 The Supply of Participant Sport: Volunteers and Sports Clubs
Country Source Sport governance Clubs Facilities Volunteers Members Typical funding
provision (approximate)
Denmark Ibsen (1999) Denmark Sports 14 000; and 7000 95% lottery and pools
Federation; Danish company clubs
Gymnastic and Sports
Association; Danish
Companies' Sports
Federation; Team
Denmark
Norway Skirstad (1999) norwegian olympic 13 113 government (37%) male predominance confederations, etc.
committee/norwegian clubs (45%) (61%); typically gambling revenues
confederation of sports 101–500 members clubs facilities 37%
56 sports federations/ income generation
committees approx 500 18% fees 14.1% grants
regional sports
federations sports clubs
Finland Koski (1999); Carter Finnish Olympic 7800 male predominance lottery funding
(2005) Committee/Finish (58%); average 330 federations 13%–74%
Sport Federation members public funds clubs
National district facilities municipal
federations funds
sports clubs
UK Carter (2005) British Olympic 106 423 local authority or average 77 members lottery/public funds
Committee/UK Sport private lottery/public funds/
regional sports councils commercial grants and
national/regional commercial income
governing bodies sports fees
clubs
USA Carter (2005); Slack United States Olympic sponsorship and
(1999) Committee state donation
organization
educational-based
a
Excludes soccer and cycling.
The Economic Foundation of Sports Club Systems 151
FIGURE 6.1 The traditional (English) model of European mass sports organization.
The table reveals that resources in sport are essentially threefold. The first is
voluntary labour provided by the members of the clubs in accordance the
origins of the club systems. Naturally, fees may be paid by volunteers and/or
members. There will be commercial activity by clubs and organizations,
embracing hospitality, sponsorship and the commercial use of any assets
owned. Finally, there will be public sector payments, such as grants or subsi-
dies, typically funding sport via indirect taxation in the form of lotteries and
gambling.
Figure 6.1 provides a stylized model of the system of sport that has emerged
in Europe, which shows that even with a relatively small set of stakeholders, a
complex set of interactions exists. The vertical structure in the figure indicates
the typical governance of sport, while the arrows indicate flows of resources.
From an economic perspective this discussion promotes questions con-
cerning why voluntary sports club systems exist and why they have the broad
hierarchical features that are described above.
6.3 THE ECONOMIC FOUNDATION OF SPORTS CLUB

SYSTEMS
6.3.1 Club goods
The economic rationale for sports clubs was presented in the previous chapter,
where it was argued that the allocation of property rights was important for
defining the character of goods in economics, and Figure 5.2 indicated that
club goods would exist in situations where consumption was excludable, but
non-rival. More formally:
“A club is a voluntary group deriving mutual benefit from sharing one or more
of the following: production costs, the member's characteristics, or a good
characterized by excludable benefits . . . thus, the utility jointly derived from
membership and the consumption of other goods must exceed the utility
associated with non-membership status.” (Cornes and Sandler, 1986
p. 159, italics in original)
Cornes and Sandler's (1986) definition indicates essentially two key elements
of club goods. These are:
1. Voluntary action.
2. Mutual benefit from sharing:
a. production costs
b. members characteristics
c. excludable benefits.
These two characteristics distinguish club goods from public goods. In the case
of public goods, it would be impossible for individuals to exclude themselves
from consumption.2 However, with club goods, individuals choose to become
part of a group to consume the good. Significantly, this suggests that mem-
bership enhances the individual's utility. Yet, as the good is consumable col-
lectively and, as choice is involved, this must imply a degree of excludability or,
synonymously, that shared consumption can impose costs on the individual.
How can sports clubs share production costs, member's characteristics or a good
characterized by excludable benefits?
Hint: Think about how competitive sports need to be supplied, why some memberships are
not simply connected with paying a membership fee and if sports are public goods.
Consider 2c first: it is clear that excludable benefits could apply to any

sports club. Access to facilities and equipment can be controlled by the club
through the prior need to become a member. Members can then share their
use. Membership might also convey less tangible benefits. In the context of
point 2b, it was discussed in Chapter 3 that consumer–producers of sport
might invest in personal consumption capital, and social capital and reputa-
tion, as part of their choices to participate in sport. Clearly, sports clubs
provide an obvious example of this opportunity, in which membership might
2
For example, even if you were a pacifist, you could not avoid “consumption” of a national
defence strategy.
reflect a desire to be associated with other members' characteristics, perhaps at

the expense of non-members, or between particular sports or clubs within
sports. Finally, in the case of 2a it is clear that in sport, competition against
other competitors of necessity implies shared production costs. Sports are an
example of joint production. Of course, individuals might compete against
their own benchmarks, but comparison against other athletes requires, liter-
ally, the production of a competition in which other athletes are directly
involved. Team sports simply extend this basic principle so that athletes need
to work together to produce their own team efforts, to be set against those of
another team. It is clear, also, that these elements of club goods might be
closely interconnected. Consequently, both 2a and 2b imply that member-
ships can affect production costs and also the character of memberships.
In club theory, “congestion” can arise, which acts as an exclusion criterion
by raising the costs of being a member either through having to ration mem-
berships by paying fees, and/or perhaps reducing the “exclusivity” of the club,
and thus potential membership, as discussed in the previous chapter. In
summary, therefore, club goods require voluntary choice being undertaken
in the context of achieving partially rival benefits with memberships limited
because of congestion. It follows that, in terms of social welfare, clubs are
efficient if the marginal benefits of membership are equal to the marginal costs
of “exclusion.”
The basic principles of club theory, and its social welfare implications, can
be illustrated in Figure 6.2. In this simple example the assumptions are:
1. membership comprises homogenous indivi-
duals maximizing their utility;
2. the population is partitioned into a set of clubs
so that all belong to one club supplying the
good, but not to other clubs supplying that
club;
3. all of the supply of the club good is utilized by
members;
4. the club good is partially substitutable by a
private good.3
Under these circumstances one can present the
consumption of club goods for a “representative”
individual and club. In Figure 6.2 private goods,
for example activities provided by a private sector FIGURE 6.2 The demand and supply of club goods.
3
The substitutability that underpins these costs may vary. For example in participating in a
team sport, such as rugby, one might pay to use a gym in the private sector or use gym
facilities provided by the voluntary club in which the team is based, to increase strength.
club, are indicated on the vertical axis, and the club good, for example, activ-
ities of a voluntary sports club, on the horizontal axis. The curved line A–B is
the production possibilities frontier defined by reallocating resources between
the private or club good. The slope of the curve will represent the MRTof the
club good with respect to the private good, as discussed in Chapter 5. It
represents the opportunity costs of providing a private good or a club good
by reallocating resources between the alternative supply opportunities. In this
respect, taking the perspective of the sports participant as a consumer–suppli-
er, the resources committed to consume–supply a sports activity might be
income and time. The balance between these will differ, depending on whether
or not one primarily purchases the activity and simply supplies consumption
time to the private sector activity or, in contrast, primarily supplies time to the
voluntary club, with minor financial expense.
The indifference curve indicates the preferences of the individual between
the goods. Its slope will represent the marginal rate of substitution (MRS)
between the alternatives. The point of tangency between the indifference curve
and the production possibilities frontier will be the consumer–producers op-
timum choice. The equilibrium condition will be:
MRSðClubPrivateÞ ¼ MRTðClubPrivateÞ : ð6:1Þ
Nested within this condition are two outcomes, reflecting the fact the
individual is both a consumer and producer of sports in clubs. The first is that
point “M” represents the membership of the club, i.e., the demand. However,
it also represents the provision of the club, i.e., its supply. In this respect the
demand and supply of the club good is simultaneously determined. If the
entire population can be represented by such behaviour, then Equation 6.1
must imply a Pareto optimum.4
It is clear that the above representation is extremely abstract, notwith-
standing the methodological emphasis of economics described in Chapter 1.
It is likely that sports clubs will operate with some spare capacity, that the
population will not be completely decomposable into a set of clubs providing
that particular activity, and that clubs are also likely to be discriminatory in
terms of membership, with some form of hierarchical structure for their
governance both within specific clubs and across the sport as a whole as a club
good. Under these circumstances, as argued in detail by Cornes and Sandler
4
This proposition is often described as the “integer problem” in club theory (see Cornes and
Sandler, 1986). Pareto optimality requires that the entire homogenous population can be a
multiple of the representative club, with no one left out. No reallocation of members can
then increase efficiency.
(1986) and Sandler and Tschirhart (1997) one cannot

make a case a priori that clubs are Pareto optimal.
6.3.2 Provision of club goods

While the above discussion explains why clubs might
exist, a natural question to ask is why doesn't the pub-
lic sector provide clubs per se?5 Weisbrod (1978, 1988,
1998) argues that this is because the voluntary, i.e.,
club sector acts to correct government failure as much
as market failure, because of an inability of the govern-
ment or market to supply sports for heterogeneous
FIGURE 6.3 Heterogeneous demands and
activities on a relatively small scale. The public sector government failure.
will tend to supply goods according to a political rule.
Under these circumstances it is likely that some con-
sumers will be left dissatisfied, as simple rules are unlikely to cope with the
required variation of preferences. Figure 6.3, derived from Weisbrod (1978),
illustrates this situation.
In this figure the vertical axis indicates the price of sports and the horizon-
tal axis their quantity. The series of sloped lines, from left to right, represent
the demand curves for sports for a series of consumers. With public sector
provision each consumer would pay a (tax-based) price of P. Under this rule it is
clear that the majority of consumers would like supply to be at least equal to Q.
However, this is not the case for the demand by consumers, “NS” who, by
having a small demand both receive and pay too much for sport. Likewise, of
those whose demands are partially met, a number indicated by MS, would
prefer more sport to be supplied at price P. Only the remaining consumer
receives their precisely desired amount. The implication is that the MPB
and MPC faced by consumers are not, in general, equivalent and consequently
are non-Pareto optimal choices.
Consequently, unsatisfied consumers will have an incentive to voluntarily
adjust towards an optimum. In a mixed economy NS might have no option but
to contribute towards sport provision. However, MS will have an incentive
either to enter the market and use the private sector to meet their demands, to
form voluntary clubs to do likewise, or to engage in informal activity.
But which of these might it be? This is not a straightforward question to
answer. For example, under the conditions of the representation of clubs in
Figure 6.2 there is no necessary reason for members of clubs to own the club.
5
Support here is defined generally to include full financing regardless of who actually
provides the good, such as contracted agents.
“Under ideal circumstances, any of a number of institutional forms can

operate the club efficiently” (Cornes and Sandler, 1986, p.188).
One can argue that the preferences of individuals and the technical condi-
tions of supply seem to be important, coupled with the informational require-
ments of decisions. It follows that the more homogeneous the demand the
greater the possibility that both government and the private sector could
supply the sport, because the general ability to meet consumer preferences
increases. However, to the extent that individuals prefer control over their
consumption, for example, as their incomes rise and they prefer to specify
where and when consumption takes place, according to their preferences, then
one might expect the private sector to meet such demands increasingly and
this might explain the current growth of private sector provision of sports as
implied by Weisbrod. This tendency may then be reinforced with excludability
in consumption facilitated by supply conditions. Health and fitness clubs, golf
clubs and tennis clubs might fit this characteristic, with more opportunities to
engage in the activity individually. Synonymously also, the growth in informal
sport, which by its nature is extremely heterogeneous in character, i.e., not
formally circumscribed, may be understood as a shift in demand away from
public sector provision and even that which was previously met by clubs,
towards more minority, i.e., highly specific, activities, which also cannot be
met by the public sector.6 However, where joint production is involved, i.e.,
some form of organized opposition is required, such as in team sports, then
voluntary clubs may remain important.
This discussion highlights a complex policy dilemma. It might be argued as
is the logic of neoclassical economics that preferences are given to consumers.
Under such a perspective it might also be argued that there is inevitability
towards any change in supply observed given the technical conditions of the
production of the sport, and information problems in public sector provision.
On the other hand, if preferences can be viewed as partially endogenous to the
economic system, then policy may help to form preferences, and this might
subsequently affect choices and supply-side evolution (in the absence of public
sector supply of facilities). Yet, as discussed in the previous chapter, in a policy
environment in which individual choice and accountability have been cham-
pioned, then to the extent that this focuses sports participation around indi-
vidual choice this may, inadvertently, squeeze the resources currently invested
in sport through consumer–producers acting as volunteers in the sports sys-
tem. Significantly, also, to the extent that such a mixed economy prevails, the
6
While characteristics of the activity might be common, e.g. using a skateboard or kicking a
football, the context of the activity might be unique and flexible.
welfare implications of changes are not entirely self-evident a priori, as with

club theory.
An interesting feature of Weisbrod's (1978, 1988) work is the suggestion
that one could measure the extent to which an organization can be character-
ized as supplying club goods. This is through the calculation of the “degree of
collectiveness” of an organization's income streams (Weisbrod, 1978, p. 173).
The assumption here is that the forms of income for an organization are
related to the character of their outputs. In this respect, an index of zero would
suggest that the revenue streams are all private sales, and consequently a pure
private good is supplied. If the index was 100, then all of the income would be
donations, and a club good supplied.
6.3.3 Hierarchical form

The above discussion, while addressing the origin and provision of club goods,
does not, of itself, address the question of why a hierarchical governance
structure is ubiquitous in sports club systems. It seems reasonable to argue
that the overall scale of a sports club system will be decided by the limits of
demand from consumer–producers of mass sport, or constraints on supply
imposed from current governance. Thus, the regulation of a sport requires
definition of the characteristics of the sport and its activities to potential
suppliers, as well as control over access to the sports club system. However,
while the “English” model of sport requires the rank ordering of competitors,
which suggests that the organization of competitions will need to be hierar-
chical; this does not of itself imply a need for a hierarchical governance struc-
ture. An explanation for this must depend on other factors.
Economic theory predicts that hierarchies can be an efficient form of orga-
nization relative to voluntary associations or peer groups and market-based
economic activity involving contractual relations between all parties. This
argument has been developed in a number of seminal contributions by
Williamson (1969, 1975, 1981, 1985), and builds on the idea that relevant
organizations emerge to minimize the transaction costs required to undertake
economic activity. The implication is that resource allocation is essentially
facilitated by different forms of “transactions.”
What are transaction costs?
Hint: There are no transaction costs suggested by the economic model of competition and
market exchange discussed in Chapter 1. This is why the “firms” are presented as very small.
How might costs arise if the assumptions of this model are relaxed?
BOX 6.1 THE ORGANIZATIONAL FAILURES FRAMEWORK

1. Human attributes: c. the number of transactors;
d. asset specificity, which concerns the nature of the
a. bounded rationality: implies that economic agents
investment required to undertake transactions and
act rationally to pursue their goals, but lack informa-
can refer to:
tion and ability to fully formulate and solve problems;
b. opportunism: “extends the assumption of self inter- i. human asset specificity, that key individuals with
est seeking to make allowances to self interest seek- tacit knowledge or on-the-job training are essen-
ing with guile” (Williamson, 1981, p. 1545). tial for the transactions;
ii. physical asset specificity, that key equipment or
2. Environmental attributes:
sites are required for a transaction;
a. uncertainty and complexity of transactions;
e. atmosphere, which concerns the perception of the
b. frequency of transactions;
environment in which transactions take place.
Transaction costs are said to arise in economic activity because of

the interaction of certain human and environmental characteristics in the
organizational failures framework proposed by Williamson to examine the
efficiency of economic organizations. Box 6.1 outlines the human and
environmental activities considered as central to Williamson's analysis.
The approach argues that under certain conditions inefficient transactions
can take place, because of inappropriate organizational arrangements. In
general, this follows from economic agents acting “opportunistically” to
pursue their own self-interest at the expense of others, hence compromising
social welfare. This provides incentives, therefore, for new organizational
arrangements to be developed to correct this inefficiency. It is suggested, for
example, that if the number of parties to a transaction is low, and there is a low
frequency of transactions between the parties and that the asset associated
with the transaction is highly specific, then market exchange may be ineffi-
cient. This is because opportunism creates the possibility of either adverse
selection or moral hazard. Examples of these were discussed briefly in Chapter
1.
The organizational failures framework can suggest why hierarchical
organization has emerged for sport generally. Figure 6.4 illustrates the
case.
Given four parties to a transaction, A, B, C and D, the left-hand side
diagram indicates the complexity of transactions that would be required in a
system of market exchange, when each party has to trade with each other, or a
pure peer-group system of voluntary activity, in which communication
FIGURE 6.4 Organizational efficiency.
between all parties is required to organize activities.7 There are a total number
of 12 lines of communication involved in this option. These might be con-
sidered to be the range of possibilities in which opportunistic behaviour,
i.e., transaction costs and organizational failure can occur.
A limited form of organization, i.e., a network with a coordinator A, is
presented in the middle diagram. Here it is clear that the number of lines of
communication are halved compared with bilateral transactions. The coordi-
nator role may, of course, rotate in a democratic way, but may require some
learning by the relevant party to undertake their coordination activity. In
contrast, on the right-hand side diagram is an example of a hierarchy, with
an authorized party, A, coordinating activities between B, C and D. Here, the
lines of communication are halved once again, from a network down to three.
This example indicates the situation in which transaction costs, and the
possibility of organizational failure, are minimized. It is this general principle
that may help to explain why:
1. sports clubs have hierarchical membership structures and management
structures to organize and deliver sport;
2. competition is organized between clubs through governing bodies;
3. different sports, or competitions, are organized by different governing
bodies.8
The latter case can be said to extend the simple hierarchy of Figure 6.4 and
suggests that relative to, say, government funding agencies, different sports
7
This resonates with the earlier discussion in which the market versus voluntary
organization can be seen, in logic, as potentially equal alternatives in the absence of further,
more detailed discussion of the nature of the economic activity undertaken.
8
Consequently, the International Olympic Committee may work with different national
Olympic committees to oversee governance of the “sport” within different countries. In
turn, the national committees work with discrete sport governing bodies.
governing bodies represent the different “divisions” of a multidivisional en-

terprise. The organizational failures framework argues that such multidivi-
sional structures help to economize on the bounded rationality and
information costs faced by strategic managers when seeking to govern com-
plex conglomerate businesses (Williamson, 1971). Economic welfare is thus
enhanced by this evolution of organizations.
Significantly, however, this may not be the case with sport a priori. It may be
the case that, in the absence of the “pure” commercial pressure and financial
objectives in sport, the “economic case” being made by policy makers, as
opposed to genuine commercial pressure for organizational change, can be
limited or muted. This might explain the current hybrid form adopted in the
UK in which clearer higher level strategic accountability has been put in
place with reorganization, but lower level delivery retains a network structure
through county sports partnerships. It may also explain the relative differen-
tial emphasis of the hierarchical structures in the European countries dis-
cussed in Section 6.2.1.
6.3.4 Economic evaluation

The above discussion has revealed that sports clubs have evolved historically
as part of sports systems. In economic terms, their rationale can be under-
stood as meeting the relatively small-scale and heterogeneous demands of
consumer–suppliers of goods which are partially excludable but non-rival.
The growth of organized sport has brought with it pressure to organize in a
hierarchical way; pressures that have increased with a more strategic orienta-
tion for sports policy. However, at the same time divergent patterns of demand
have emerged which has led to a growth in the private sector consumption of
sport and also a growth in informal sports activity. Inasmuch as individual
choice is being exercised by consumer–suppliers of sport in these contexts,
i.e., either to consume sport informally or through voluntary clubs, or through
private sector provision, then it might be argued that the current systems of
sports supply is efficient in a Pareto sense. The organizational failures frame-
work reinforces this view inasmuch as it proposes that efficient organizations
emerge to provide efficient supply opportunities. However, it has been argued
above that the economic efficiency properties of club goods, and particularly
those of a complex hierarchical form, as is the case with sport, cannot be
presupposed. This is particularly the case in the political economy environ-
ment of sports systems, and it seems likely that, in economic terms, sports
systems operate in a second best economic sense.
The second best theorem suggests that the welfare implications of policy
shifts, in a world not characterized by Paretian optima, require case-by-case
Volunteering in Sport 161
consideration and comparison. This is despite a tendency in the economic

literature to assume that piecemeal movements toward a more market-
oriented supply organization necessarily increase social welfare. In this re-
spect, policy movements towards a more competitive market allocation of
resources cannot be presupposed to represent Pareto improvements.
The practical importance of this for policy evaluation is that focus should
be on the question of whether or not any new supply arrangements prevent
participation in sport that previously used to take place but is not now possi-
ble, without a reduction in the welfare of the individuals. It is clear that income
redistribution matters in this regard, and this is why equity of access is an
important issue in sport. More generally, this discussion implies that there is a
real sense in which the policy priorities of public authorities matter and need
to be evaluated.9 Appendix 6.1 outlines the second best theorem.
6.4 VOLUNTEERING IN SPORT

The above discussion has made it clear that voluntary association has been an
essential element of mass sport supply through sports clubs. In this section
some of the main ideas connected with volunteer motivation, its scale and
value is addressed, building on some of the material discussed in Section 6.2.
Subsequent sections then indicate how economic theory can be used to explain
the choice to volunteer, as well as to calibrate its value. To begin with, the
definition of volunteering is discussed.
6.4.1 Defining volunteering

Volunteering is an elusive concept to define as it embraces different kinds of
activity in varying degrees and across a variety of sectors of society. Cultural
and political contexts also clearly affect volunteering traditions (Lukka and
Ellis, 2002; Salamon and Anheier, 1997; UN, 2001). Drawing on Cnaan et al.
(1996), it can be argued that four main elements are associated with various
definitions of volunteerism. These include: the degree of free will involved in
volunteering; the degree of remuneration involved; the structure of
9
Perhaps indirect evidence of this is the tolerance of different societies for different forms of
sports system. Of course, none of this precludes that more general tendencies are evident in
sports systems. The point is that despite these tendencies, say, towards market-based supply,
economic efficiency does not necessarily increase. The exclusion of previous participants in
sport because of, say, price changes that follow a regime change in supply, may thus imply a
welfare loss rather that be indicative of an optimal reallocation of resources by the
consumer–supplier.
volunteering; and the intended beneficiaries of the volunteer act (see Cuskelly
et al., 2006). From the point of view of the limited formal analysis and em-
pirical research that exists, a variety of specific definitions are used. In Aus-
tralia, the Australian Bureau of Statistics (2005) defines a volunteer as:
“. . . someone who willingly gave unpaid help, in the form of time, services or
skills, through an organization or group within Australia. People who did
voluntary work overseas only are excluded. The reimbursement of expenses
in full or part (e.g., token payments) or small gifts (e.g., sports club t-shirts or
caps) are not regarded as payment of salary, and people who received these
are still included as volunteers. However, people who received payment in
kind for the work that did . . . are excluded.” (Australian Bureau of Statistics,
2005, p. 4)
In Canada a volunteer is:
“. . . any individual who volunteered, i.e., who willingly performed a service
without pay, through a group or organization during the 12-month reference
period preceding the survey.” (Research Centre or Sport in Canadian
Society/Centre for Sport Policy Studies, 2005, p. 10)
The latter study argues that the UK has been at the forefront of volunteer
research in sport. In this regard, as indicated by Nicholls (2004), two defini-
tions are potentially relevant. The first reflects the 1997 National Survey of
Volunteering. Here, volunteering is defined as:
“. . . any activity which involves spending time, unpaid, doing something that
aims to benefit (individuals or groups) other than or in addition to, close
relatives, or the benefit of the environment.” (Institute for Volunteer
Research, http://www.ivr.org.uk/researchbulletins/bulletins/1997-national-
survey-of-volunteering-in-the-uk.htm Accessed December, 2007)
The second is specifically concerned with sport as proposed by Gratton et al.
(1997). Here volunteering reflects:
“. . . individual volunteers helping others in sport, in a formal organization
such as clubs or governing bodies, and receiving either no remuneration or
only expenses.” (Gratton et al., 1997 p. i)
It is clear that, with the exception of the Institute for Volunteer Research
definition, the emphasis has been on formal volunteering. In contrast, LIRC
(2003) redefined sports volunteering for the UK to be:
“. . . individual volunteers helping others in sport and receiving either no
remuneration or only expenses.” (LIRC, 2003 p. 6)
The shift in emphasis of the UK definition was to recognize the importance of
informal contributions to sport. It should be noted, however, that given the
relatively small amount of research on sports volunteering, in general the
common emphasis is to emphasize free choice, with volunteers only receiving

reimbursement of expenses at most, and activity taking place within a formal
organization.
6.4.2 Motivation and constraints on volunteering

In descriptive empirical research the specific reasons individuals give for why
they volunteer their time and skills are quite varied. Table 6.2 indicates some
of these reasons according to reviews undertaken by the sources indicated in
the first row of the table.10 Both motivations to volunteer and the benefits from
volunteering are noted. It is possible to review statements such as those in
Table 6.2 and to argue that, while some distinctions in the motivations and
benefits for volunteering exist, they also include common reasons such as:
contributing to the sport and the community, particularly in the context of
family connections; generating friendship, as well as experiencing new things;
and developing new skills. The enhancement of personal development and
employment prospects are also commonly cited. Interestingly, such a perspec-
tive implies a degree of altruism, but also self-interest, in the sense of enhanc-
ing the individuals's own identity and social standing, and generating skills for
employment prospects (see also Cuskelly et al., 2006). This has implications
for the economic analysis of volunteering discussed later.
Much of the research on volunteering also suggests that the voluntary
sector is facing increasing constraints. This was mentioned earlier in Section
6.2, when discussing sports clubs. To begin with, Weed et al. (2005) note that
the changes to the sports club sector have led to tensions, because of the need
to operate in an increasingly commercial environment with the consequent
individualization of sport activity, and because of a focus on professionalism in
supply coupled with the incorporation of sports into public policy, with the
consequent focus on accountability and service provision in the light of re-
ceiving public funds. These tensions particularly centre on a conflict between
the traditional, but informal, associativity facilitated by volunteering, which
might enhance social capital and citizenship as well as support sport, and the
pressures for more formal management and organization, which run counter
to this ethos. The latter may, thus, lead to a reduction in volunteer commit-
ment and recruitment. Table 6.3 presents the types of constraints on volun-
teering stated in the descriptive empirical literature. The interpretation of
such constraints in economic terms is considered below, after the patterns
and scale of volunteering are discussed.
10
It should be noted that these motivations may be associated with countries other than the
source of the claim.
TABLE 6.2 Volunteer motivations and benefits
LIRC (2003) Australian Bureau of Research Centre or Sport in

Statistics (2005) Canadian Society/Centre for
Sport Policy Studies (2005)
Motivations
Shared enthusiasm for the sport Learning new skills The opportunity to help others
Affiliation to their club Gaining training Contributing to a valuable area
Social rewards from volunteering Helping others Having fun
Giving something back/duty Increasing enthusiasm and energy Promoting and maintaining sports
Involvement with own children Sharing talents, abilities and experience Contributing where a family
participates
Enjoyment Fighting boredom Contributing to the community
Wanting the club to do well: Making new friends Using skills
& No replacement/no one else
& Fear club will collapse
Love of the club/sport: Building self-confidence Learning new skills

& Grew up with the club
& Progression of involvement
Wanting to keep the club going Exploring career opportunities Companionship and friendship
Wanting to see youngsters/team get better Feeling needed, useful and appreciated
Using time while children participate Gaining a new direction in life
Wanting a challenge Giving something back to the community
Maintaining involvement after playing Being a team member
Personal development Getting closer to the sport or activity an
organization represents
To find out about the club Contributing to the sport or activity
To do something in retirement Having fun and enjoying themselves
Benefits
Social benefits Gained fundraising skills
Enjoyment Learned coaching, judging,
refereeing and officiating
techniques
Fulfilment from helping others Acquired organizational and
managerial skills
Satisfaction Increased knowledge
Pride in helping the club and its Helped to succeed in paid job
participants do well
Satisfaction from keeping the club going Helped chances of finding a job
Achievement, rewarding, challenge Helped to find employment
Satisfaction from seeing youngsters Made new friends, contacts
succeed/get better
Sense of belonging to the club and wider Promoted services and business
community
Opportunity to gain new skills and Developed network of influence
personal development
Uses time while children participate
TABLE 6.3 Volunteer constraints
LIRC (2003) Australian Bureau of Statistics Research Centre or Sport in

(2005) Canadian Society/Centre for
Sport Policy Studies (2005)
Lack of time Lack of skills, training and Lack of time

experience
Too old Lack of understanding of Lack of interest
volunteering
Demands of job Low income levels Not aware of volunteering opportunities
Family commitments Limited life experience Not asked
Children grew up Lack of acceptance Did not know how to get involved
Help no longer wanted Lack of child care Financial costs
Not enough other volunteers Lack of time Concern over being sued
The work is left to fewer people Cost and availability of transport Unwilling to make a commitment
Requires specialist skills Geographic isolation Contributions not recognised
The club is asking more, Language and cultural barriers Family obligations
because of governing body
organizational pressure
Feel that efforts are wasted A need for higher levels of formal Dissatisfaction with experience
education and training
No appreciation or thanks No supervision
for your efforts for the club
Stopped playing the sport
6.4.3 The profile and scale of volunteering

Table 6.4 presents some broadly comparative results on sport volunteers. It is
clear that sports' volunteering is dominated by males between the ages of 20
and 50 years, predominantly in full-time work and, with the exception of the
UK, with children in the household. Significantly and perhaps related to the
latter point, the weekly hours of volunteering are much higher in the UK.
There is some evidence that other sports activity is more important to sports
volunteers than other activities and volunteering. In fact, the Research Centre
for Sport in Canadian Society/Centre for Sport Policy Studies (2005) identifies
that male volunteering tends to be associated directly with the sport, for
example, coaching, training and officiating, whereas female activities are more
diverse. There is also indirect evidence of this for the UK, according to LIRC
(2005) in which the preponderance of activities are dominated by coaching,
administration, officiating and fund raising, and males make up the majority
of volunteers.
In the light of earlier expressed concerns about constraints in volunteering,
the changing significance of volunteering activity to the sports economy can be
revealed when considering some comparative data on trends in volunteering.
TABLE 6.4 Volunteer profilesa
LIRC (2003) Australian Bureau of Research Centre or Sport in

Statistics (2005) Canadian Society/Centre for
Sport Policy Studies (2005),
Cuskelly et al. (2006) b
Gender Gender Gender

& Male 67% & Male 62% & Male 86 % (64%)
& Female 33% & Female 38% & Female 14 % (36%)
Age Age Age

& 16–24 28% & 18–24 12% & < 25 4% (<24 19%)
& 25–34 22% & 25–34 18 % & 25–34 6% (13%)
& 35–44 21% & 35–44 29 % & 35–44 52% (41%)
& 45–59 19% & 45–59 28% & 45 + 39% (27%)
& 60–69 10% & 60–69 8%
& 70 + 4%
Children Children Children

& Yes 32% & Yes 64% & Yes 94%
& (22% 5–15 years old) & (53 % 5–15 years' old)
& No 68% & No 36% & No 6%
Employment Employment Employment

& Full time 56% & Full time 59% & Full time 65% (Full or
part-time 84%)
& Part-time 14% & Part-time 22% & Education 16%
& Retired 9% & Unemployed 2% & Retired 3%
& Full time education 10% & Not in labour & Non working 13% (16%)
& Non working 11% & Force 17%
Hours volunteered per week Hours volunteered per year

& Formal and informal 4 hours & < 12 hours 15%
& Formal 5.3 hours & 10–25 hours 54%
Non-sport volunteers as well Non-sport volunteering activity & 26–50 hours 22%
& Formal and informal 15% & Participate in sport 69% & > 51 hours 9%
& Formal 14% & Participate in non- organized
activity 14%
& Attended sports events 77%
a
Percentages are rounded
b
The data for Canada from the primary source focus on team sports and are based on a relatively small-scale pilot study. More general data are provided in the
source in parenthesis.
Cuskelly et al. (2006) argue that in Canada volunteering fell by 13% (from 7.7
million to 6.2 million) between 1997 and 2000, although average volunteer
hours rose by almost 9% (from 149 hours per year to 162 hours per year). In
contrast in Australia, while there was a 24% increase in volunteer numbers
between 1995 and 2000, volunteer hours fell by 20% in sports volunteering.
Finally, Gratton et al. (1997) and LIRC (2003) argue that volunteering hours
increased by approximately 45% (from 183 million to 266 million) between
An Economic Analysis of Volunteering 167
1995 and 2002, while the number of volunteers increased by approximately

35% (from 1.49 million to 2.01 million).11 However, it should be recognized
that this comparison includes volunteering in education and events. Focusing
purely on sports clubs and national governing bodies, volunteer hours in-
creased by approximately 40% (from 170 million to 237 million) and volun-
teers by 33% (from 1.19 million to 1.6 million) (LIRC, 2003, p. 22). The
implication is that volunteering is essential to the supply of sports and
“remains the bedrock of opportunity in many sports” (LIRC, 2003, p. 33).
One of the key findings from the comparative data is it would appear that
for the UK, more hours per volunteer are being required to support sport.
However, in Canada total volunteer hours fell from 1147.3 million to 1053
million, while in Australia they increased from 517.7 million to 704.1 million
hours. The broader implication of these trends is that, in the absence of new
volunteers, existing volunteers have to work harder. More generally, more
volunteer effort, if not volunteers, might be required to support at best, as
was discussed in Chapter 4, static participation in sport. It is clear, therefore,
that a “supply-side constraint” is tightening in sports systems.
6.5 AN ECONOMIC ANALYSIS OF VOLUNTEERING

The economic analysis of volunteering can proceed readily from the theory of
the consumer–producer developed in Chapter 3, although this brings with it
some interesting issues for debate. To begin with the general motivation
implied in consumer theory is utility maximization. This implies self-inter-
ested behaviour in consuming as many goods and services that are possible. Is
this reconcilable with voluntary and altruistic motives? Becker (1976, 1992)
emphasizes that utility is the welfare of the individual as defined by them and
is not tied, neither is economic analysis generally, to the specific nature of the
material under investigation. In this respect, drawing on Becker (1974), one
can identify that charity or altruism, which is implied in voluntary acts of
giving donations of goods or time to other individuals or organizations, can be
examples of utility maximization, as individuals optimize on the use of
resources that raise their well-being. This might help to explain why, as noted
earlier, both self-interest and altruism arise as common motives stated by
volunteers.12
11
This is just for formal volunteers, to control for the definitional change of volunteering
between the two pieces of research.
12
This is not to suggest that such an analysis is without criticism. For example, Becker's
analysis is a form of meta ethics, in which morality or virtuous or envious behaviour is
treated as a special case of a broader, but nonetheless specific, utilitarian approach.
In this framework, the resources that the individ-

ual has at their disposal are goods bought on markets,
x, and time, t. It is clear that volunteering will be a
time- rather than goods-intensive activity. Therefore,
it is not surprising, as described above, for time to be
identified as a fundamental constraint by volunteers
on their level of activity.
The other major constraints identified by volun-
teers are connected with organizational experiences,
such as lack of appreciation or supervision, or poor
relationships with other organizations. While recog-
nized as constraints by volunteers, lack of appreciation
can be connected with a reduction of individual wel-
fare identified by volunteers. This might arise directly
from reductions in the benefits received from an ac-
tivity in which tacit or literal appreciation was
expressed, or indirectly from the failure to receive ben-
efit transfers from others, through their allocation of
resources to the individual concerned. This naturally
also applies to lack of supervision, or poor relation-
ships with other organizations which also may imply a
need, on behalf of the volunteer, to commit more of
their own time or goods to activities to overcome these
problems.
Consequently, volunteering can be understood
from the economic perspective of the consumer–pro-
FIGURE 6.5 The volunteer decision. ducer discussed in Chapter 3. This approach also
implies how the value of volunteering can be inferred.
To illustrate this Figure 6.5 is derived, based on a
number of assumptions described below.13
1. The individual has two basic commodities that can be consumed–pro-
duced: volunteering “V” and other activities “OA.”
2. Both volunteering and other activities are produced from the commitment
of time, “t,” and market goods, “g,” which are given in supply.
3. The production functions are:
V ¼ Vðg; tÞ
13
This is an example of the optimizing behaviour outlined in Chapter 3, Section 3.2.4.
and:
OA ¼ OAðg; tÞ
4. The consumer's utility is a function of V and OA so that:
U ¼ UðV; OAÞ
In this analysis the consumer's marginal utility with respect to V and OA are
assumed to be positive but diminishing. Further, in an analogous manner, the
level of output of V and OA can be described by a set of isoquants which
indicate equal levels of output of each commodity, given the use of resources
of time and market goods to produce them. The shape and slope of the iso-
quant reflects that the production function exhibits positive but diminishing
marginal productivity associated with each of the inputs to the production
function. The slope of the isoquant is referred to as the marginal rate of
technical substitution (MRTS) between the inputs.
The box at the top of Figure 6.5 illustrates the production functions for each
of the basic commodities, volunteering (V) and other activities (OA). The
production of V begins at the origin in the bottom left-hand corner of the
box, and increases in a northeasterly direction as more time and goods are
committed to production. The isoquants measuring equal levels of output for
different combinations of input are drawn convex to this origin and slope
down from left to right. In contrast, OA begins at the origin in the top right-
hand corner of the box, and increases in a southwesterly direction as more
time and goods are committed to production. The isoquants measuring equal
levels of output for different combinations of input are drawn convex to this
origin and also slope down from left to right.
Which points in the box will indicate productive efficiency?
Hint: Think about the use of inputs to produce a given level of output.
Note that the box implies that the fixed endowments of goods “g” and time
“t” can be allocated either to the production of V or of OA. An efficient
allocation of these inputs to the production of either V or OA must, therefore,
ensure that all of the resources are employed and that the outputs of either of
the basic commodities are as high as possible. Points in the box that satisfy this
property are on the curved line joining the origins. These are where the MRTS
of time and goods are equal in the production of both V and OA for the
individual. In each of these cases it would be impossible to reallocate goods
and time to increase production of V or OA without reducing the output of OA

or V, respectively.
This curve can thus be used to derive the production possibilities for Vand
OA, which is drawn in the lower part of Figure 6.5. Conceptually, it maps out
the alternative combinations of V and OA implied for alternative efficient
allocations of “g” and “t” for the production of alternative levels of V and
OA. The slope of the curve is the marginal rate of transformation (MRT) of
Vand OA, and is the rate at which each good can be substituted by reorganizing
the production activity of the individual, while being efficient. The indiffer-
ence curve of the individual is then drawn to indicate that the optimal choice
between V and OA is where the MRT and MRS of V and OA are equal. Anal-
ogous to the discussion of club goods, this is where the marginal benefits of
consuming Vand OA are equal to the opportunity costs of producing Vand OA.
Indeed, it should be clear that the amounts of V and OA consumed and
produced become equivalent where:
MRSV;OA ¼ MRTV;OA ¼ MRTSV;OA ð6:2Þ
It follows that the marginal value of volunteering can be implied as the

opportunity cost of OA foregone. This suggests a substantial difficulty in
valuing volunteer activity in practice. It might be tempting to use the oppor-
tunity cost of time only as an approach. This is to directly apply the income–
leisure approach of Chapter 3, illustrated in Figure 3.3, where it is identified
that the wage rate is equivalent to the opportunity cost of time in leisure.
In this regard, by analogy, one might argue that voluntary time – as a form
of leisure – has the same opportunity cost. Indeed, this is the approach
often adopted in practice in empirical work in which a volunteer hour is
multiplied by an appropriate wage rate, or volunteers are asked to impute
a value of compensation for their time, if not working. These individual
valuations can then be aggregated. Table 6.5 presents some examples from
sport.14
In this respect LIRC (2003) note that:
“The valuation of the hours of volunteering is at the average hourly
earnings for all industries for 2002 of £11.69 . . . It is the value of the
hypothetical cost of replacing . . . the sports volunteers with paid labour.”
(LIRC, 2003 p. 14)
14
Other examples apply outside sport, such as Wolff et al. (1993). Davies (2004) discusses
some other examples from sport.
TABLE 6.5 The value of volunteering
Author Country Value
Taks et al. (1999) Belgium 10 billion (Belgian francs)

Heinemann (1999) (West) Germany 6 billion (Deutschmarks)
LIRC (2003) England 14.1 billion (£)
The assumption here is that while the individual may not necessarily substi-
tute work and volunteering, one must identify what the alternative might be if
the volunteer did not offer their time, i.e., having to pay the opportunity cost
for someone else with equivalent skills. As Freeman (1997) argues, however:
“A priori, one might expect that volunteers would consist largely of people with
low opportunity cost of time – low-wage workers or the jobless . . . But . . . few
. . . demographic contrasts between volunteers and non-volunteers are
consistent with a simple opportunity cost explanation of volunteering.”
(Freeman, 1997 p. 146, italics in original)
This suggests that a more sophisticated “average cost” based on the volunteers'
socio-economic profiles and the industrial sectors in which they work is
appropriate. The reason for this is because of the skill differences and the
fact that not all volunteers work (Brown, 1999). However, this raises the
important conceptual point that once human capital is considered then
the opportunity cost of time per se is not an adequate representation of the
value of volunteering (leisure in this context). To repeat a discussion was raised
in Chapter 3:
“It is easily shown that the usual labour–leisure analysis can be looked on as a
special case . . . in which the cost of the commodity called leisure consists
entirely of foregone earnings and the cost of other commodities entirely of
goods.” (Becker, 1965 p. 98)
Consequently, unless an attempt is made to estimate the opportunity costs of
volunteering in terms of not just the allocation of work time, but also contin-
gent on human and social capital investments, then the estimates are likely to
be biased.
Conceptually, a model should be specified to analyze the set of decisions to
consume other goods, to go to work or not, and to volunteer within the context
of a household. On the basis of this model one could, in principle, identify a
properly weighted monetary equivalent to volunteer hours. Freeman (1997)
provides an analysis of volunteering generally in which utility depends on
goods, leisure time and charity which, in itself, is produced by the inputs of
volunteer time and donations. Based on data from the US from the 1989
Current Population Survey, and the 1990 Gallup Survey of Giving and Volun-
teering in the US, regression analysis suggests that including a “charitable
production function” is more consistent with the evidence than an approach

in which all volunteer hours are treated as equivalent.
No such work has been undertaken in sport. The very limited multivariate
work that exists focuses less on the value of volunteer time and more on the
choice to undertake volunteer activity. It is also based on small sample sizes,
consistent with the difficulties in data collection. Wilson (2004) examines the
choice of 112 volunteers in Athletic clubs in the southeast of England to enter a
coach education programme. Broadly it is shown that time constraints and
household patterns of activity matter, but these are different for the genders.
For example, with participation in sport females have to trade-off coaching
opportunities. In general, however, as noted above, working is positively as-
sociated with volunteering, which challenges a simple opportunity cost of time
analysis.
Burgham and Downward (2005) analyzed data on 126 individuals (63
males and 60 females) including swimming volunteers, as well as non-volun-
teers, at facilities in which swimming clubs were operating. It was identified
that increasing age, having previously been a swimmer and having children
swimming were the main factors that promoted the likelihood of becoming a
volunteer. There was some indication that full-time work reduced this likeli-
hood. Based on the subsample of 58 volunteers, it was shown that being
female, in full-time work, with a higher salary increased hours per week
volunteered. The same was the case for having previously been a swimmer
and having children currently swimming. However, the hours worked per
week and having children not connected with swimming reduced the hours
volunteering. The upshot is that, as implied with the participation literature
reviewed in Chapter 4, personal consumption capital is generally also impor-
tant to promoting volunteering. The obvious recruitment pool for volunteers
is participants, as implied by the consumer–producer model.
There is some evidence, also, that the decision to volunteer is not the same
decision as committing hours to volunteering in sports participation, i.e., the
duration of the activity, as also argued in Downward and Riordan (2007). In
this respect there is some evidence that the income–leisure trade-off is impor-
tant for volunteering. While higher incomes and full-time work can promote
volunteering, the hours worked constrain this opportunity. Not surprisingly,
females volunteer more as they were less likely to be in full-time work. This
suggests that household allocation of time is important in volunteering, as
also implied by the consumer–producer model.
As the above analyses are small-scale, although their basic descriptive
characteristics are in line with larger descriptive studies, the implication is
that volunteering provides a fertile area for future research.
Policy Implications 173
6.6 POLICY IMPLICATIONS

The previous section suggests that economics would support the view that
if the benefits of volunteering were falling and the opportunity costs increas-
ing, it is likely that individuals will reallocate time and market goods to the
consumption and production of other activities. The discussions above indi-
cated that there is some evidence of the substitutability of sports volunteering
with sports participation and other consumption activities, but also that a
potentially increasing supply constraint in volunteering is evident in the
sense that more total hours are being required to support static participation
levels. Inasmuch as the age profiles of volunteers are concentrated around
the middle aged and that there is a growth in informal and commercialized
forms of sport consumption, then this supply constraint may or may not be a
problem.
If one adopts the perspective that the consumer–producer always allocates
resources optimally then, notwithstanding the discussion of second best ear-
lier, one might argue that the various sectors of mass sports supply could be
said to be adjusted to an optimum level. However, the analysis above sug-
gested that the sports club system is based on the supply of older volunteers.
Chapter 4 indicated that the demand for sports club activities is more likely for
younger participants. Traditionally, therefore, an “intertemporal” supply
chain could be said to have existed in which consumer–producers shift more
from participation in, to provision of, sport. This suggests a possibility of
potential disequilibrium that might persist as participants' age. Over time,
it might be the case that volunteering and participation in non-club sports
activities become more substitutable, because of the growth in emphasis of
individual consumption. In this sense, there is likely to be excess demand
from newer generations of sports participants, because the supply of volun-
teering shrinks. If the sports club system is considered to be important to
social welfare, then active support may be required from the public sector to
prevent its reduction. Sports clubs also need to ensure that participants re-
main involved as volunteers. However, as discussed earlier, it is not entirely
clear that the public sector can adequately supply club goods, or that volun-
teers welcome more formal systems of management as they are required to be
accountable for funding. The outcome may be, then, that a welfare judgement
needs to be made that increased inefficiency in supply, either through infor-
mational problems of matching supply to demands or a lack of monitoring
and reduced formal accountability, is a price worth paying for overall higher
participation levels. Such decisions seem to be inevitable in a second best
economic scenario.
6.7 CONCLUSION
In this chapter, elements of the broadly hierarchical sports systems typical in
Europe, and, by implication components of hierarchical sports systems else-
where, have been presented. It is noted that sports clubs and voluntary labour
are a central feature of these systems. It is argued that the traditional English
concept of sport lends itself to hierarchical sports competitions but, of itself,
this does not account for the hierarchical governance structures in sport. It is
argued that voluntary sports clubs emerge as a result of both government and
market failures, but in theory they can only be efficient in a second best welfare
sense because of the restrictive assumptions implied by efficiency with club
goods. It is argued that hierarchical governance structures can be explained in
terms of the organizational requirements of sport and the need to govern sport
strategically. In this respect, policy initiatives that seek to increase the ac-
countability of sports organizations can add to this impetus and efficiency
gains can be made.
However, there is a potential shortfall in volunteer activity for sports club
systems. This may be exacerbated by the increased time costs imposed on
volunteers in more formal systems of accountability. Furthermore, if the
public sector promotion and support of sports club participation becomes
increasingly required this will involve potential inefficiencies. In this regard,
contrary to the claim stressed that the public sector in the UK should provide
efficiency, economy and effectiveness; these may best be viewed as partially
contradictory objectives.
Appendix 6.1 The General Theory ofSecond-Best

“The general theorem for the second best optimum states that . . . in a situation in which
there exist many constraints which prevent the fulfilment of the Paretian optimum
conditions, the removal of any one constraint may affect welfare or efficiency either by
raising it, by lowering it, or by leaving it unchanged.” (Lipsey and Lancaster, 1956–1957,
pp. 11–12)
To illustrate this proposition, reconsider Equation 6.1. If there are two individuals in an
economy with different tastes then two different club memberships would be derived in each
case. Consider now that person A consumes–supplies more of the club good than person B.
Assume that the government imposes a sales tax on the club and private sector supply of
sports to raise revenue. An optimum tax-raising strategy would be to tax the club good at a
greater rate than the private good for person A and vice versa for person B. Suppose that the
government decides to employ a uniform tax rate to ensure that Paretian conditions are met,
that both consumers face the same set of prices, and that the same amount of revenue be
raised. In this case, both consumers would be forced onto lower indifference curves,
Conclusion 175
reducing welfare. Figure A6.1.1 illustrates the proposition in which it is possible to show that
the equilibrium situation for each person while paying a uniform price, indicated by the
diagonal constraint linking the axes, suggests a lower indifference curve than could be
possible for less consumption of the preferred good at a differential, i.e., higher tax price for
each of them, as illustrated by the constraints of different slopes.
FIGURE A6.1.1 The second-best theorem.

CHAPTER 7
The Market for Professional

Sports: General Themes
OBJECTIVES
& To understand the origin of professional sports

& To analyze sports as economic contests
& To explore and understand the peculiar economics of professional sports
& To compare and contrast European and US leagues and their sources of
finance
& To consider sports leagues as monopolies or cartels
& To consider the evolution of sports leagues
7.1 INTRODUCTION
In this chapter attention turns towards the analysis of professional team
sports, with Section 7.2 discussing briefly the transition of sports from an
amateur to professional status. Section 7.3 then outlines a more general set of
principles that underpin the nature of sports competition as an economic
contest and links this discussion to the design of sporting competitions,
particularly league and cup competitions, and other forms of tournaments.
Drawing on this discussion, Section 7.4 outlines key concepts that have been
applied in the analysis of professional team sports, and provides a sketch of the
development of more recent forms of competitions. The discussion indicates
the importance of the underlying economic components of sports leagues and
their potential need to coordinate results, as well as their sources of revenues
and costs. Section 7.5 closes the discussion with a comparison of European
and US sports leagues.
177
178 CHAPTER 7 The Market for Professional Sports: General Themes
7.2 THE TRANSITION TO PROFESSIONALISM

7.2.1 Sports clubs: Key concepts revisited
Chapter 6 argued that sports clubs evolved historically along with sports systems
to meet the relatively small-scale and heterogeneous demands of consumer–
suppliers of goods which are partially excludable but non-rival. Sports clubs are
examples of club goods, with individuals choosing (in practice subject to exclu-
sion criteria) to become part of the collective group to consume–produce both the
sports activity and membership to enhance their utility. In this respect, the club
acts to internalize the externality of non-rival benefits by controlling access to
the sport, facilities, equipment and reputation benefits provided by the club.1
Of specific importance to sports clubs is that the consumption–production
of the sports activity requires the sharing of production costs in supply, not only
with regard to the running or administration of the club but, more fundamen-
tally, sport is a prime example of joint production, when individuals are com-
peting against others. Here, competitors need to cooperate to facilitate sport,
either within the club or across clubs as part of a sports system. In this regard,
team sports simply extend this basic principle to the case in which opponents
operate as groups rather than as individuals. Equations 7.1 and 7.2 summarize
this relationship in the form of a production function which, as detailed in
Chapter 1, describes the production of output from the use of inputs. Equation
7.1 recognizes that a sports contest within a club (subscript “wc”) requires the
club's land and capital and the efforts of two competitor members, A and B.
Contestwc ¼ Contestwc ðLand; Capital; CompetitorA ; CompetitorB Þ ð7:1Þ
Contest12 ¼ Contest12 ðLandc1 ; Capitalc1 ; Competitorc1 ; Competitorc2 Þ ð7:2Þ
Equation 7.2 shows that in a sports contest between clubs 1 and 2 with
subscript “12”, signifying that club 1 is the “home” club and club 2 the “away”
club that competition emerges from the use of the labour efforts of competitors
from each club, c1 and c2 respectively, and the land and capital of one of the
clubs, in this case c1.
As the discussion above reveals some general conceptual features of sports
club competition, it follows that changes in the form of consumption and
production of contests will occur with a move from voluntary to professional
sports. The following sections address these issues in more detail; the discus-
sion begins with an examination of the transition of clubs to professionalism
through a process of formalization.
1
As discussed earlier, externalities arise whenever the consumption or production acts of
individuals directly affect the utility or objectives of other economic agents, rather than being
mediated by the market system.
The Transition to Professionalism 179
7.2.2 Formalization
In amateur sports participants can be viewed as consumer–producers, because
they combine voluntary time plus membership fees and other donations and
expenses to fund the supply of the sports activity in which they typically
participate. Szymanski and Zimbalist (2005) confirm this in connection with
baseball in the US. In their original format the typical form of contest in clubs
occurred between relatively close communities and, in many respects, the
specific event was significant as a relatively infrequent gathering between
communities. In early baseball games large social events coincided with the
games. In order for the sports to grow and to develop, however, formalization
was required in terms of defining the nature of the sport, i.e., the rules of
the contest, and in deciding how contests could be organized. In economic
terms this meant that the product needed to be defined and made sufficiently
homogenous so that (a potential) large-scale demand could develop and supply
become organized to meet it.
As discussed in Chapter 2, the emergence of the Football Association (FA)
in 1863 began the formalization of association football, and likewise the Rugby
Football Union (RFU) was established in 1871. Much earlier, in 1858, the
National Association of Baseball Players (NABBP) had emerged in the US,
initially:
“. . . to preserve the social values that they believed in and had little interest
in fostering the expansion of the game . . .” (Szymanski and Zimbalist, 2005,
p. 19)
However, three subsequent and interrelated developments can be said to have
eroded this perspective, which heralded the onset of professionalism and
brought about the formation of hierarchical and commercial governance.
The first was the development of organized competition. Commensurate with
the establishment of governing bodies of sport, in England, knockout cup
competitions soon developed, as discussed in Chapter 2. A key function was
to formally amalgamate the informal growth in the supply of clubs in England
that had taken place typically through county-based organizations. Local ri-
valries became popular spectacles and, not surprisingly, gave rise to payment
to watch the games. Similar developments arose in the US with the wide-
spread growth of baseball, although in a less structured way. As Szymanski and
Zimbalist (2005) note, by the 1860s most US states had baseball clubs and
“barnstorming” competitions including touring clubs with crowds in excess of
5000 spectators. In general, teams could organize their own fixtures.2
2
Prior to the Civil War, English cricket sides (not representative) had fairly regularly toured
the US, playing local teams on a relatively informal basis.
However, a growing demand for a more homogenous sports' contest raises

both the incentive for, and the possibility of, organizing a team for commercial
motives, to access payment by spectators. This naturally carried with it the
need to build stadiums and venues to extract revenue.3 In turn, accessing
paying spectators for commercial gain raised the second stimulus towards
professionalism, which was that it provided an incentive to pay players, i.e.,
to attract the best talent to a team to enhance the prospects of success.
For example, while the FA was initially founded on amateurism, by 1888
the English Football League (EFL) was established and accepted by the FA to
prevent schism, consequently the payment of players was accepted in soccer
for those residing within a six mile radius of a club in 1885.4 By 1895 it had
become the case that players could only be transferred between clubs for the
payment of a fee agreeable to the existing employer as part of the “retain and
transfer system,” while in the face of rising wage rates a maximum wage was
established in 1900.5 In contrast, differences occurred in rugby football be-
cause of the northern rugby union clubs offering “broken-time” payments to
compensate for loss of earnings. This led to a split in the RFU and the
formation of the Rugby Football League (RFL) in 1895. The game soon became
fully professional, because it was in direct competition for both players and
spectators from association football.6
Similar schism occurred in the organization of baseball in the US. As
Szymanski and Zimbalist (2005) argue, the first fully professional baseball
team, the Cincinnati Red Stockings, was established in 1869. Naturally, this
team had to leave the NABBP, which was founded on amateurism, but it
prompted other teams to leave also. As a consequence, the National Associ-
ation of Amateur Baseball Players (NAABBP) and the National Association of
3
The commercial imperative in organization was not overt. Indeed, in football it has been
argued by Russell (1997) that the FA deliberately raised prices to try to maintain football's
appeal to the skilled working and lower middle classes. Likewise, Syzmanski and Zimbalist
(2005, p. 33) note initial resistance to admission price cutting in the NL in response to the
formation of the IL to avoid the “patronage of the degraded” but “recognition by the
respectable.”
4
An amateur football association emerged in 1907, but this rejoined the Football
Association in 1914.
5
Both the maximum wage and retains and transfer systems are discussed in more detail in
Chapters 9 and 12.
6
Indeed, amateur rugby league only developed some separate administrative status in 1973
under the auspices of BARLA, the British Amateur Rugby League Association, reflecting a
critical fall in amateur activity under the governance of the professional clubs who, as
discussed in Chapter 2, were also experiencing huge declines in attendance.
Professional Baseball Players (NAPBBP) were established to govern the alter-

native dimensions of the game in 1871.
The final development towards professionalism occurred with the forma-
tion of leagues with the need to pay professional players prompting a require-
ment to extract more regular gate money and to utilize facilities more fully. In
the case of the EFL, the FA retained overall control of football in terms of the
development and interpretation of rules and, indeed, the parameters of com-
mercial activity. As well as the movement of players between clubs and their
salaries, as discussed above, ticket prices were agreed on, limitations on div-
idend payments to shareholders were agreed and directors could not be paid for
their efforts. The league soon developed, flourished and expanded, principally
organized on vertical selection by merit and some geographic boundaries for
the less successful teams, to allow for the problem of travel costs.7 Likewise, in
professional rugby, while traditionally structured county championships were
contested initially a league of two divisions was introduced by 1902, with 18
sides each and promotion and relegation again depending on merit. Further
variations in league format then followed.
According to Szymanski and Zimbalist (2005), a major impetus to the
development of leagues in English sport was the formation of the National
League (NL) for baseball in the US in 1875, which English administrators were
familiar with. However, its defining characteristics were different. The league
emerged as a direct challenge to the NAPBBP over disputes concerning the
signing of players and threats to the longevity of the game emerging from
match fixing, and betting corruption, undermining public confidence in the
sport. The NL was primarily concentrated in the northeastern US, the tradi-
tional home of baseball, and had the following characteristics:
1. a league board of team owners with authority to make all decisions and to
enforce rules;
2. players bound to their contracts with no managerial powers;
3. a closed league of teams granted a territorial monopoly with limitation of
franchise numbers and eligibility of the size of a city to host a franchise;
4. teams paid annual dues;
5. teams play a balanced list of home and away fixtures;
6. imposed ticket prices;
7. alcohol, gambling and Sunday fixtures banned;
8. independent umpires.
7
It is notable that the growth of mass railway transport made football clubs tend to locate
near railway stations to facilitate access to competitors, as well as to provide access for
spectators (http://www.spartacus.schoolnet.co.uk/Fhistory.htm Accessed 31 March 2008).
The strictures of such a league, its relatively small scale initially of six and then
up to 12 teams in the 1890s, coupled with its exclusivity, meant that rivals
developed, such as the International League (IL) in 1877, stemming from the
NAPBBP, in which player cooperatives were common. The NL adopted policies
to make it as popular as possible to avoid threats to its existence. Consequent-
ly, the IL collapsed in 1880, to be replaced by the American Association (AA) in
1883, which made agreements about territorial exclusivities and player re-
serve systems.
The AA folded by the 1890s, but player reserve systems had become a
central part of the competitive process between the NL and any future rival
leagues. These systems recognized the property rights of teams over their
players and led to compensation payments to teams when other larger and
more successful teams sought to buy their better talent, in much the same way
as the retain and transfer system in football. The concentration of talent
helped to develop the distinction between the Major League, i.e., the NL,
and other “Minor League” teams.8 By 1901 one of the minor leagues, the
Western League, had redeveloped into a Major League, the American League
(AL), by capturing the new potential markets in the southwestern US and
offering franchises to cities that lost franchises with the NL. Since 1903 the
winners of each of the major leagues have competed in a “World Series.”
7.2.3 Review
There are clearly distinct similarities in the shift towards professionalism in
these oldest team sports. Consumer–producers operating in a world in which
clubs were essentially self-financing from members, in the manner of mutual
societies, developed into hierarchical organizations, governing activities in
which both professional producers and distinct gate-paying spectators existed.
Drawing on Weisbrod (1978), it can be argued that the degree of collectiveness
of income streams to the outputs of the organization shifted from being close
to 100, as a club good, towards being much closer to zero, as a private good.
Despite these similarities, however, it is also clear that the specific
approaches to professionalism had distinct characteristics. In baseball, orga-
nization developed from commercially-minded owners who, in many
respects, looked to protect the economic viability of their activities. In Eng-
land, soccer and rugby league developed under the governance of one structure,
each with aspirations to cover both the amateur and professional levels of the
8
An important facet of the reserve clause was the option to draw on the talent as needed.
This meant that the larger teams could hoard talent off their “active” team roster in the lower
leagues and not have to pay major league salaries.
TABLE 7.1 Comparison of voluntary and emergent professional sports
Members Private investors
Ownership
Governance & Club: members & US: team owners
& Sport: governing body & Europe: team owners and
governing body Europe: team
owners and governing body
Club objectives Non-profit & US: profit
& Europe: success and financial
survival
Consumption Consumer–producer Gate paying spectators
members
Production Consumer–producer & Professional athletes
members
& Professional management
Europe: non-remunerated directors

Form of competition Cup competitions Cup competitions
Traditional fixtures League fixtures
Ad-hoc tournaments
game and, in contrast to a closed league format, retain the vertical structure of
competition that has come to symbolize European sports. This included ac-
cess to professional status and the possibility of success at that level being
governed by merit, with no obvious institutional recognition that making a
profit was central to activities. This is despite the fact that professional clubs
now had an entirely different financial structure to the amateur clubs. Indeed,
throughout the history of soccer and rugby league, the absence of profit seems
to be accepted as a facet of business reality (Jackson, 1899; Szymanski and
Kuypers, 1999; Downward and Jackson, 2003). In contrast, despite the stated
claims of club owners, it is argued that US sports have tended to make profits,
even if these are not declared (Fort, 2000; Szymanski and Zimbalist, 2005).
Table 7.1 summarizes the typical main economic differences between volun-
tary and professional sports, as implied by the historical emergence of the
latter from the former.
Some more detailed differences between the US and European models of
professional team sports are discussed later in this chapter. At this juncture,
however, attention turns towards more formally outlining some key economic
principles associated with the economic nature of contests. They provide a
basis for understanding some general themes in the economic analysis of
professional team sports, and indicate how they relate to other forms of sport,
such as events. Moreover, the incentives that led towards formalization and
professionalism discussed above can be highlighted.
7.3 SPORTS COMPETITIONS AS ECONOMIC CONTESTS

In two insightful papers (and other contributions) Szymanski (2003a, b) has
proposed that sports contests can be understood as economic tournaments (or
synonymously contests). As defined in Chapter 1, tournaments are concerned
with economic activity whose output is assessed in relative rather than abso-
lute terms. It is because tournaments produce a rank ordering of outcomes
that they are directly applicable to sports.
Why do the rank order outcomes of a contest make tournament theory relevant to sports?
Hint: Think about how the winner of a cup competition, athletic race or league champion is
defined.
It is quite clear that in sporting contests of any type the winners, whether of
a knockout cup competition, a league championship, an athletic race, a high
jump or whatever, win that event by finishing above their opponents according
to some ranking rule that links the performance of the competitor to their
assigned rank order. In the case of some sporting competitions the ranking rule
may carry with it an absolute yardstick. World records for the times of various
races, distances of various throwing events, heights and distances of various
jumping events or weights lifted are examples of absolute yardsticks. In many
cases, however, the performance of competitors is measured purely in a rela-
tive sense, i.e., how they performed against one another in a specific context.
The precise way in which this takes place can vary and, as indicated in
Table 7.2, follow either the knockout form of traditional cup competitions,
round robin leagues or some hybrid version. The reason why the former
arrangement and its hybrids are more likely can be understood to be a direct
result of economic incentives. These will include pooling together the better
talent, i.e., restricting the scale of competition, attracting better quality com-
petitors and rewarding genuine effort more than luck (which is more likely to
affect a knockout competition with randomly matched competitors and each
match being crucial). The details of these incentives are now discussed.
7.3.1 Contest/tournament theory

As Szymanski (2003a, b) argues, tournament or contest theory has its roots in
the seminal work of Tullock (1980), which explores “rent-seeking” behaviour
by “contestants” for public funds. The analysis has subsequently been used in
labour market analysis and other forms of economic activity.
TABLE 7.2 Typical forms of tournament in sport
Form of tournament Description Example
Single rank order tournament Individuals/teams meet for a single specific Stroke play golf, marathon running,
contest time-trial cycling
Single elimination knockout A hierarchical series of rounds of competition FA Cup, Wimbledon
tournament between contestants, with the top ranked in
each case only progressing
Double elimination tournament As above, but subsets of the best losers are Athletics (heats) Rowing
given another chance to compete
Round robin tournament All contestants meet an equal number of times Sports leagues (some)
Multi-stage tournaments Contestants meet initially in groups then
either progress to:
An elimination tournament World cup, Super-bowl, play-offs, e.g.,
divisional promotion in English football,
Heineken Cup in rugby, Champions
League in football
Further round robin tournament World series
Match play golf Contestants meet in a series of contests
Sports Competitions as Economic Contests

each decided by holes won, not strokes played
Challenge Champions taken on by the right to challenge, Boxing
based in other tournaments
185
The concept of economic rent has its origins in the work of David Ricardo regarding
the differential return to more productive land relative to unproductive land. More
generally, it represents the difference in the return to a factor of production and the
amount that would be required to prevent that factor of production being reallocated
to an alternative use. This latter value is referred to as the transfer earnings (or the
reservation wage in the case of labour).
Rent-seeking behaviour refers to the activities of economic agents extract-

ing a return as a factor of production that is not directly connected with their
productivity, but is more a reflection of their control over resources as a result
of regulation or other constraints, such as asymmetric information, on mu-
tually beneficial exchange as defined in the theory of competitive markets.9
In this respect, monopoly power can be viewed as rent seeking, especially if
the monopoly exists as the result of regulation. As discussed above and
further in Chapter 11, it is possible to identify the traditional forms of labour
market in sports, in which maximum wages and reserve clauses have existed,
as forms of rent-seeking behaviour, because they use regulations either to
restrict the mobility of players or the level of their wages, in contrast to what
voluntary exchange would produce. The principal–agent problem also pro-
vides opportunities for rent-seeking behaviour. Rent-seeking is inefficient,
therefore, as it is connected with the transfer of incomes between economic
agents at the expense of possible welfare gains. From a purely economic point
of view, as Frick argues, one of the primary objectives of tournaments is that
they act:
“. . . as devices . . . (that) . . . can be very helpful in eliciting the effort levels that
are necessary to give the principal the highest expected net profit . . . and
subject to the constraint that the agents must be given a remuneration
package attractive enough to enter the competition.” (Frick, 2000, p. 512)
From a sporting point of view, therefore, tournament theory can indicate how
hard a contestant is likely to work relative to their opponents, which adds to
the potential for their success and the overall quality of the competition as this
effort is balanced against opponents. It also highlights that this process entails
an economic trade-off between the competitor, the agent, who needs to be
enticed to take part in the contest, and the principal or organizer of the contest,
who is seeking an economic return from organizing the activity, but who needs
to employ the competing agents. Furthermore, tournament theory indicates
that a balance needs to be struck between relative contestant effort and
9
Profit in this sense is seen as a result of mutually beneficial exchange. Marxist theory
would argue that profit is connected with the regulations governing the private ownership of
property, and is also rent-seeking. Consequently, there is some ambiguity about these
distinctions.
Sports Competitions as Economic Contests 187
TABLE 7.3 Assumptions and implications of tournament theory
Assumption Commentary
1. Athlete's output depends on their own 1. This identifies that each competitor
deterministic efforts, plus random can influence the outcome by their
influences beyond their control but own endeavour, but subject to their
may be common to all competitors, competitors' actions and also other
such as weather, the state of factors
equipment or facilities, or design of
tournament
2. The strategies of contestants can be 2. Recall from Chapter 5 that game
analyzed using game theory and, in theory is used to examine interacting
particular, Nash (non-cooperative) economic agents. Nash Equilibrium
Equilibrium reflects economic agents optimizing,
given the actions of their competitors
3. The stability of the Nash Equilibrium 3. If this is not the case, the contest
requires some uncertainty over the would become predictable and
outcomes undermine its own existence. This is
discussed further in the text
4. The rewards from any competition 4. Without a differential return to
should be predetermined and increase performance, competitors would not
with rank to provide incentive to increase the effort put into
compete competitiona
5. With equally talented competitors 5. This situation comprises a case of
and with equal economic pure competitive balance
endowments, athletes will choose the
same level of effort and the probability
of winning is equal
6. If contestants are heterogeneous more 6. In a case of unbalanced competition,
able competitors are more likely to both weaker and stronger competitors
win the contest. If competitors do not would economize on their efforts,
know their relative abilities, no effects recognizing their likely performance
on effort are likely. If they do, then
effort levels will fall
a
There is an important point to emphasize here, which is that the reward structure, which need
not be monetary, aims to increase efforts (reduce rent seeking) rather than generate effort per
se. This means that competition is not impossible without inducement, but that higher level
competition is.
randomness in affecting outcomes because of joint production. Frick (2000)

outlines the main assumptions and implications of tournament theory, which
are briefly summarized in Table 7.3.
Drawing on these assumptions, a simple version of the theory provided in
Szymanski (2003b) can be presented to illustrate the main predictions. As-
sume that there are two competitors, 1 and 2. It can be argued that their
success in a competition “S” can be produced by their effort “P,”10 as detailed

in Assumption 1 of Table 7.3. In this way, Equations 7.3 and 7.4 provide a
production function for success in a competition. The equations are written in
Cobb–Douglas form, as presented in Appendix 3.1, because it is recognized
that the productivity of the effort is captured in the exponential parameter “g.”
The parameter g can be viewed as reflecting the tournament's design and
consequently its ability to convert the competitor's efforts into success. It is the
same in both cases, to reflect a lack of bias in the tournament as indicated in
Assumption 5 of Table 7.3.
S1 ¼ P1 g ð7:3Þ
S2 ¼ P2 g ð7:4Þ
The probability, P, of success then becomes:
P1 g
P1 ¼ ð7:5Þ
P1 þ P2 g
g
P2 g
P2 ¼ ð7:6Þ
P1 þ P2 g
g
Equations 7.5 and 7.6 describe a logit form of contest success function for
the competitor. This specifies that their individual probability of success
depends on the exercise of their effort relative to the total effort in the contest,
i.e., that provided by both contestants, as given in the denominator of the two
equations. An important feature of these two equations is that they capture the
externality essential to sporting competitions through joint production (as
discussed earlier with Equations 7.1 and 7.2).
Two further important results can be demonstrated from these equations,
and they concern the tournament design parameter g. As shown in Appendix
7.1, if this parameter is set equal to zero, the tournament cannot discriminate
between the efforts of the contestant and each competitor ends up with the
same, equal, chance of success. The tournament becomes a pure lottery.
Equations 7.5 and 7.6 would reduce to:
1
P1 ¼ P2 ¼ ð7:7Þ
2
10
The random impact “z” on success could be introduced explicitly by a function such as
S1 = P1g expz, where exp is the base of natural logarithms. If the effect of random elements
on success on average is 0, then exp becomes 1 and average success becomes equivalent to
the expression in the text.
If, on the other hand, the tournament could perfectly discriminate effort
from random chance, then g would tend toward infinity. In this case, however,
should any competitor exhibit more effort, then this competitor would have a
probability of:
P11 ð7:8Þ
That is tending towards “1,” or complete certainty. The tournament
becomes “an auction in which the largest contributor of effort is certain to
win” (Szymanski, 2003b, p. 469). This actually applies for sufficiently large
finite values of g. This highlights, as emphasized earlier, that effective tourna-
ments require balancing the systematic efforts of competitors with random
influences, as also stated in Assumption 3 in Table 7.3. This is because
Equations 7.7 and 7.8 provide scenarios in which competitors have no incen-
tive to participate in the tournament. Appendix 7.1 shows that, in general,
increases in the value of g impair (improve) the success probabilities of lower
(higher) effort teams. An exception is where each competitor exerts the same
effort. The probabilities of success are then invariant to g and equal.
Why would complete certainty or complete uncertainty undermine the existence of
tournaments?
Hint: What is the motive for athletes to compete in sport?
There are two potential reasons why complete certainty or uncertainty

could lead to market failure for a tournament, and this is connected with
the motives of competitors in sport. If athletes viewed their efforts as not
having an impact on the outcome of the tournament, then clearly the incen-
tive to participate with all but minimal effort would apply, which ultimately
would undermine the tournament. Likewise, as indicated in Assumption 6 of
Table 7.3, if it became known that one competitor had a disproportionately
greater prospect of winning the tournament, then overall effort in the tourna-
ment would fall, again undermining it. The tournament would fail because of
declining quality, caused by moral hazard.
Assumption 2 in Table 7.3 suggests that the solution to the contest can be
identified as a Nash equilibrium (see Chapter 5). This requires specifiying the
payoff, R, to each contestant, recognizing the efforts of the other contestants,
and then finding the optimal effort levels of each contestant, given the efforts
of the other contestants. Equations 7.9 and 7.10 describe the payoffs to each
contestant, where V is the value of winning the tournament, which can be
described in monetary or non-monetary terms, and c is the marginal costs of
the effort, which are assumed to be identical according to Assumption 5 in

Table 7.3.11
R1 ¼ P1 V c1 P1 ð7:9Þ
R2 ¼ P2 V c2 P ð7:10Þ
As Appendix 7.2 shows, the Nash equilibrium effort level for the contest for
each contestant, Pe, will be as implied in Assumption 5 in Table 7.3:
gV
Pe ¼ ð7:11Þ
4c
Recognizing this, a tournament with “n” contestants would produce an
equilibrium level effort of:
gVðn 1Þ
Pe ¼ ð7:12Þ
n2 c
This equation summarizes a number of important interrelated features
about tournaments that can help to understand the economics of sports. If
the quality of the tournament, “Q,” is likely to be a function of the equilibrium
effort of competitors:
Q ¼ QðPe Þ ð7:13Þ
and, in particular, quality rises with effort such that:
dQ
> 0 ð7:14Þ
dPe
Equation 7.14 provides a formal basis for suggesting that, ceteris paribus:
1. The fewer the contestants are in number, n, the greater the effort that will
be exerted by them. This will subsequently raise the economic value of the
contest as it will be more attractive to spectators. This highlights that
broader participation and economic viability are likely to be negatively
related. As discussed earlier, this can explain why the US model of a smaller
closed league has been directly associated with profits, in contrast to larger
European leagues, particularly with football in the UK, where profits have
11
As Szymanski (2003b) notes, “If contestants have different marginal costs of effort
functions, then the contest will be asymmetric – some players will achieve a higher
probability of success for any given level of effort . . . the alternative is to assume that
individual effort contributions of a given amount produce different probabilities in the
contest success function . . . From an economic perspective these are essentially equivalent
problems. . .”
been less common but greater numbers participate. It also suggests that the
largest economic value in individual sport tournaments is likely to be
where smaller numbers of better competitors meet.
2. If the value of the tournament “V,” is higher, the greater will be the effort
invested by competitors. Other things being equal, this indicates that
tournaments funded at a higher level are more likely to be of a higher
quality. This means that there is a greater chance of sporting records or
higher levels of achievement occurring. Likewise, there is greater chance of
positive feedback between sporting performance and economic returns.
3. In contrast, the level of effort, and quality of competition is likely to fall as
the general MC of effort, “c” rises, other things being equal. This is a
sensible economic result, which draws on Equations 7.9 and 7.10, and
indicates that if increased effort has no effect on the chance of winning a
competition, an increase in the cost of effort would be balanced by a
reduction in effort to ensure that payoffs remain the same. This suggests
that the quality of competition would fall and, as a result, its economic
attractiveness would decrease. This result is equivalent, therefore, to hav-
ing a tournament with lower quality participants.
4. The greater the value of the tournament parameter, “g,” the greater the
effort of contestants will be and consequently, the higher the quality of the
tournament and the potential economic return. Recalling that this param-
eter measures the ability of the tournament to distinguish between the
efforts of the contestants and luck, it follows that a tournament such as a
round-robin league will be of a greater overall quality than a knockout cup
competition, with corresponding greater economic value. The same would
be the case for multi-stage tournaments.
Point 4 reveals that the parameter g measures the rate at which effort can be
converted into probable success. Other things equal, a stronger effect (greater
value) of g means a stronger incentive for competitors to exert effort, increas-
ing the competition's quality and likely economic value. When the league is
perfectly balanced, only the incentive effect operates, therefore in a perfectly
balanced league g will deliver most effort, quality and value. In an unbalanced
league the incentive effect can be offset by a reduction in the balance of
competition.
This point is interesting for three reasons. Broadly it suggests a basis for
understanding why professional leagues evolved out of knockout competitions
or became multi-stage tournaments with elements of leagues. The intuition is
that in a pure knockout tournament, the random matching of opponents and
the fact that every game can eliminate a team from competing further means
that the element of chance in the teams' success is relatively large. League
fixtures and “repeated trials” are more likely to reveal systematic and genuine
efforts, and underpin commercial viability.
It also suggests why balance in competition is seen to be potentially im-

portant to sports leagues. For example, in their design home and away games
are typically played to reduce home advantage or bias, as is evident from a
variety of sources such as the influence of crowds on officials (see Downward
and Jones, 2007; Sutter and Kocher, 2004; Neville and Holder, 1999). Histor-
ically, as discussed in Chapter 9, it can indicate why revenue sharing and
labour market restrictions have been justified in leagues to reallocate the
ability to buy labour efforts in a more balanced way.
Finally it suggests why the “calibration” of tournaments changes to seek
to elicit greater efforts. In soccer in 1981, in England and Wales, and
subsequently embraced by FIFA in 1995 and most other soccer leagues,
the metric for ranking results shifted from two points for a win to three.
The aim was to raise attacking play, the quality of the contest and thereby
spectator interest. More recently, in rugby union competitions such as the
Premiership, the Heineken Cup or the Super 14 competitions, bonus points
are awarded for scoring four or more tries or for maintaining losses to
within seven points. Once again, the aim is to raise the quality of compe-
tition through rewarding attacking play, but also to encourage efforts to
attack even if the team is losing.12
What drives these broad predictions is the joint production of sports con-
tests which, in turn, can be understood as having two key dimensions, the
quality and uncertainty (together with its counterpart, success) of a contest. As
such they can provide the basis for the economic analysis of sporting contests,
particularly as they will also affect the value of competitions. This is now
illustrated in connection with a discussion of some specific themes associated
with the analysis of professional team sports.13
7.4 FOUNDATIONS OF THE ANALYSIS OF PROFESSIONAL

TEAM SPORTS
This section begins with a discussion on the basis of the joint production
of sports contests in professional team sports. This requires investigating if
the sports “league,” i.e., the clubs, owners and governing bodies that orga-
nize the competition, can best be understood as a monopoly firm or a cartel
of firms in an oligopoly market. The objectives of the firm are then
12
This is because the numerical aggregation of results to define a rank is arbitrary overall,
but can induce sporting incentives as the significance of a specific result can vary the overall
set of possible results.
13
These concepts are also drawn on in a discussion of sports events in Chapter 12.
Foundations of the Analysis of Professional Team Sports 193
discussed. The chapter closes with a review of league structures in the US

and Europe.
7.4.1 The production of professional team sports

Professional team sports can be analyzed from an economic perspective in
which labour (mainly players), capital and land (stadiums and other facilities)
are combined by clubs who supply teams to produce a saleable product – the
game or contest.14 To obtain revenue to pay for inputs it is necessary to exclude
non-paying spectators, hence the universal use of stadiums and a traditional
suspicion of live broadcasting of sports. For teams to cover their costs, each
must be guaranteed a minimum number of home games for which tickets can
be sold, which suggests league or round robin tournaments rather than knock-
out competitions. As an organizational structure, leagues comprise sets of
individual clubs that fall within the broader governance structures outlined
in Chapter 2. There are potential divisions of interests between league and
club, because of the existence of externalities in production noted earlier in
the chapter. Indeed, it was a result of this difference that prompted one of the
seminal contributors to the literature, Neale (1964), to coin the phrase “the
peculiar economics of sport.”
7.4.2 The peculiar economics of sport

Early contributors to the economics of professional team sports, Rottenberg
(1956) and Neale (1964) stressed the importance of a key externality known as
the uncertainty of outcome (UO) hypothesis:
“. . . that a more or less equal distribution of talent is necessary if there is to be
uncertainty of outcome; and that uncertainty of outcome is necessary if the
consumer is going to be willing to pay admission to the game.” (Rottenberg,
1956, p. 246)
Consequently:
“. . . consider the position of the heavyweight champion . . . He wants to earn
more money, to maximize his profits. What does he need in order to do so?
Obviously a contender, and the stronger the contender the larger the profits
from fighting him . . . since doubt about the competition is what arouses
interest . . . Pure monopoly is disaster: Joe Louis would have no one to fight
and therefore no income.” (Neale, 1964, pp. 1–2)
14
While the emphasis in discussion earlier in this book has been on clubs, consistent with
the literature “teams” are now referred to as the key facet of clubs.
These quotations illustrate the joint production element of sports discussed

above, in which contests depend on the exercise of efforts and economic value
is likely to be linked to cases where greater joint effort occurs, with comparable
skills. The upshot is that public interest in sport (and hence attendance and
revenue) increases, other things being equal, when teams are as closely com-
petitive as possible. Domination of the league by Team A is liable to reduce
interest in and attendance at games involving the remaining teams, e.g., B
versus C and C versus D, even if attendance at games involving Team A does
increase. In the longer run, even Team A would suffer if the standard of
competition declines.
Neale (1964) identified another externality associated with sports, which
he termed the “fourth estate” or “league standing” effect. Inasmuch as interest
in the sport is not just confined to spectators observing the actual contest, but
spills over to the general public, a positive externality exists from the sport.
This permits the media to earn revenue by reporting sport results without any
contribution to the cost of producing sport. Consequently, the sale of rights to
broadcast its matches represents a sport league internalizing some of the
media benefits created by its operations.
The term is derived from a two-centuries-old notion that the press was the “fourth
estate,” being preceded by the aristocracy, the church and the common people.
Aristocracy and church were represented in the upper chambers of national
legislatures, the people in the lower chamber. The press was seen as sufficiently
important, even then, to constitute a distinct interest group in society rivalling the
others in importance. The name of the pre-revolutionary French legislature
translates into English as “the Estates General.” Some of the terminology survives –
e.g., “the House of Commons.”
A similar positive externality exists for the betting industry. The sale of pro-
grammes and “fanzines” exemplify clubs' internalization of a portion of the
league standing benefit inasmuch as they capture elements of interest associ-
ated with their specific club and specific opponents. The presence of unofficial
material, however, reveals that property rights are not fully allocated.
7.4.3 The league as a natural monopoly or as a cartel

Because of UO, Neale (1964) argued that the market for professional team
sports differs significantly from most industries. In Chapter 1, the standard
competitive model of markets suggests that the firm chooses its level of output
to maximize profit. In contrast, the sports team's level of output is determined
by the league, which determines both its membership and the number of
games each plays in the season. This led Neale to consider that the league
constituted a natural monopoly firm, with clubs acting as plants. Natural
Foundations of the Analysis of Professional Team Sports 195
monopolies can be said to exist if one firm can produce output at a lower
(social) cost than two or more firms, or if only one firm seems able to survive
competition. Both of these criteria appeared to exist for Neale, with joint
production allowing the first condition to occur, and the fact that most sports
seemed to be organized in national leagues or that play-offs or subsequent
tournaments existed to produce a national champion, consistent with the
latter condition because it is most profitable. Thus, as noted earlier, separate
leagues for baseball exist in the US but they have always, since becoming major
leagues, organized play-offs in the World Series. Likewise, while boxing has
always had several governing bodies and world titles, most kudos attaches to
the fighter who manages to unify the titles. Neale's analysis depends crucially
on the assumption that the individual team cannot determine its output. The
extent to which one is willing to accept this proposition depends on how one
defines output, which he took to be the number of games played. This assumes
that playing qualities are the same. If they are not then this offers opportunites
to market a different product by some of the “plants.”
That the team has discretion over its output relative to the league was an
important factor in the view, put forward by Sloane (1971), that the league is a
cartel, as defined in Chapter 5. In the cartel model the club behaves as a firm
rather than as a plant, although it is still subject to UO and still unable to offer
saleable output in the absence of the league. The cartel model has the advan-
tage over the natural monopoly model, however, in that it highlights the
relationships between clubs and leagues and the implied need to manage
these. While cooperation between clubs is essential (hence the relatively
light-handed treatment of sports leagues by competition policy) there will be
times when some clubs will perceive that private advantage can be had by
breaking ranks with the membership as a whole, as indicated earlier in the
formation of the first leagues. The cartel model seems to have explanatory
power as far as recent developments in leagues are concerned. Box 7.1 illus-
trates some recent changes to leagues in the UK, in which traditional competi-
tions have evolved separate administration for higher level domestic
competition which provides access (potential in the case of cricket) into inter-
national competition.
As well as internal threats to its existence, a cartel's continued existence
also depends on its ability to protect its business from independent entry,
which is most evident in the US. In the closed-league US sports, the geograph-
ic spread of franchises can be managed to create a latent demand for a fran-
chise. One implication of this scarcity is that it can be used to leverage public
investment for sports facilities (see Chapter 12), although rivals can emerge.
For example, Fort and Quirk (1995) note that between 1920 and 1994 seven
leagues had existed in football, not counting the black American leagues that
BOX 7.1 RECENT EVOLUTIONARY CHANGES TO LEAGUES

Soccer Rugby Union
Premier League (PL): the largest clubs broke away from the Premiership: an equivalent development to the premier
English Football League, financed by the largest share of the league in rugby union. Promotion and relegation to the
new media income from promotion. Relegation to the National League is possible.
remaining league, the championship, is possible. Magners League: a league for the “Celtic” nations of Ire-
Champions League (CL): the European Cup in soccer land, Scotland and Wales comprising “franchises” from
evolves into a multi-stage tournament with a pool stage regional sides. No promotion or relegation.
and a knockout element. Heineken Cup: a champions' league for teams from the
Premiership and French First Division which qualify and
Rugby League Magners League teams, together with Italian teams.
Super League (SL): an equivalent development to the
premier league in rugby league. Initially a closed league Cricket
was proposed, but promotion and relegation was retained County Championship: reformed to include promotion and
with exceptions for development clubs. Currently relegation.
“franchises” for three year membership will commence Limited over games: development of 20-Twenty Cricket will be
in 2009. influenced by the development of the Indian Premier League.
disappeared or were absorbed as racial integration progressed, and that four-

teen of the then extant NFL teams had begun in rival leagues. New US football
leagues have continued to appear and disappear. In major league baseball, the
World Series ties the “rivals” together.
The upshot of this discussion is that it illustrates that cartel theory prob-
ably explains sports leagues better than the natural monopoly model. Further
cartels can suggest that league arrangements can exist with potentially differ-
ent objectives. The literature suggests that they can be understood in terms of
the distinctions between US and European leagues. This comparison is now
explored further by examining firm objectives and some characteristics of the
forms of leagues.
7.5 US AND EUROPEAN LEAGUES

7.5.1 Club objectives
Drawing on the historical development of professional sports in Europe and
the US, most US writers assume that profit maximization is the goal of team
owners. In contrast, Sloane (1971) argued that European club owners might be
viewed more accurately as utility maximizers, where a major determinant of
utility is playing success, although profit is important for the club's long run
financial stability. Forgoing profits, and even underwriting losses, therefore,
suggests that club ownership is an act of consumption more than investment.
US and European Leagues 197
The impetus for this assumption was the relegation and promotion of hierar-
chical sports systems in Europe, coupled with the historic desire in European
football for success in international competition. Sloane's (1971) proposed
team utility function is given in Equation 7.15:
U ¼ UfP; A; XðP PMIN Þg ð7:15Þ
Where “U” is the utility level attained, which depends on performance on

the field “P,” home attendance “A,” UO “X,” and on the excess of post tax profit
“P” over a minimum acceptable level of post tax profit “PMIN.” The notion of
an acceptable minimum profit reflects factors such as the club's need to secure
its long-run future. Any club that has outstanding debt has to display the
ability to service that debt. Over and above servicing debt, the club may need
to generate an acceptable level of profit; if it is a publicly quoted limited liability
company the directors need to produce enough dividends to keep the share-
holders happy. If they cannot, the share price will fall and the company may
become the subject of a hostile takeover. In contrast, if a club has access to a
sympathetic and wealthy owner (Chelsea with Roman Abramovich, Juventus
with the Agnelli (FIAT) family, AC Milan with Silvio Berlusconi, Shaktar
Donetsk with Rinat Akhmetv) or to favourable treatment from local or na-
tional government (Real Madrid, Barcelona) the minimum acceptable level
could actually be negative. Sometimes, a large public company has a support-
ive relationship with a sports team, for example Philips Electrical with PSV,
Bayer Chemicals with both Leverkusen and Munich, and Yukos Oil with
Moscow Dynamo.
Theoretically, one can expect that, other things being equal, an increase in
any one of the four factors identified by Sloane will make the supporters, the
owners and the management happier. As Sloane (1971) notes, the more closely
the four explanatory variables are correlated with each other the more probable
it is that the predictions of the profit maximizing and utility maximizing
models will resemble one another. Unless the predictions differ in some read-
ily identifiable manner, attempts to distinguish the objectives of teams will
fail. This is not to say that the differences are trivial. Chapter 9 explores the
consequences of these differences for league management policies in detail.
7.5.2 League characteristics

Table 7.4 indicates some summary characteristics of US and European lea-
gues, drawing on and extending Andreff and Staudohar (2000, 2002). Based on
these characteristics, Andreff and Staudohar (2000, 2002) indicate that the
amateur model of European sports shared some characteristics of early pro-
fessional sports, particularly through the rise of gate revenue receipts, as
198
TABLE 7.4 Evolution of sports leagues
CHAPTER 7 The Market for Professional Sports: General Themes

Basis of European amateur sport Traditional European Contemporary European US (traditionally and
comparison (traditionally (1900–1980s) (1980s–current) currrent)
and current)
Ownership Members Small number of shareholders Public companies, media Private owners (previously
companies some collectives)
Entrepreneurial investment Media companies
Finances Membership fees, Gate revenenues Gate revenues Gate revenues
subsidies and donations, (e.g., 68–96%) (e.g., 50%–20%)
hospitality
Gate revenue, some advertising Industrial patronage and Sponsorship Sponsorship
and sponsorship subsidy (e.g., Fiat, Bayer, (e.g., 20%–25%)
Peugeot 4%–21%)
(more for highest performers) Sponsorship Television rights Merchandising
(e.g., 5%–35%) (e.g., 15%–39%)
Merchandising Merchandising Subsidy – stadia
(e.g., 1.4%–11%) (e.g., 10%–34%)
Television rights National and local television
(e.g., 0%–33% rights
Redistribution None Gate sharing Some television revenues Gate sharing
Television rights sharing Some salary caps National television rights
sharing
Maximum wages/salary caps Modified transfer system drafts
Transfer systems Reserve option
(modified over time)
Salary caps
Forms of competition Traditional fixtures Domestic leagues Domestic leagues Domestic leagues and play-offs
Knock-out cups National and International multi-stage
international tournaments
Leagues at highest level Knock-out cups
All domestic, though informal
tours
League structure Vertical with promotion and Vertical with promotion and Vertical with promotion and Closed
relegation relegation relegation
Players developed in minor
leagues, college sports
Team location Traditional/multi-team cities Traditional/multi-team cities Traditional/multi-team cities Franchise/local monopoly
Some “franchises”
Labour market Voluntary local Local developed into national/ National/international Traditionally national e.g.,
international football
Some international
development
US and European Leagues 199
discussed earlier, and league competition. Indeed, the vertical promotion and
relegation of the leagues ensured that the professionalism of the teams in-
creased with performance within the hierarchical structure.
The traditional European professional system of organization emerged as
discussed earlier and is labelled by Andreff and Staudohar (2000, 2002) as the
Spectators–Subsidies–Sponsor–Local or SSSL model. Ownership of the club
was either by an individual entrepreneur or a set of local investors receiving
limited financial return, sometimes by regulation of the role of directors. The
main source of trading revenue was typically gate revenues, with some minor
sources of sponsorship and merchandising, but little television rights income.
Its labour market was essentially closed to competition by restrictions on the
number of overseas players that could be fielded, although some international
trade occurred. Product markets, i.e., the leagues, were also closed and based
nationally. By the 1980s it is argued that the contemporary model, labelled
Media–Corporations–Merchandising–Markets–Global or MCMMG by
Andreff and Staudohar, began to appear, of which the developments in
Box 7.1 are a part. In contrast to the traditional SSSL model, the MCMMG
model indicates that clubs were increasingly driven by media income, sup-
ported by large-scale merchandising and sponsorship activities. Gate revenues
fell to minority sources of income for the larger clubs, i.e., 33% for the Premier
League, 13% for Serie A and 25% for the Bundesliga in 2005–2006 (Deloitte,
2007). The clubs have evolved from local ownership into publicly quoted
corporations or into being owned by business consortia or single wealthy
entrepreneurs, many of whom are from overseas. According to Deloitte
(2007) Serie A derived 62% of its revenue from broadcasting rights in 2005–
2006, the Bundesliga 27% and the Premier League 42%. The Big Five, accord-
ing to Deloitte (2005a) account for about 68% of all the broadcast revenue in
European football. The giants of the Big Five, e.g., Real Madrid, Barcelona and
AC Milan, each earn as much, if not more, from broadcasting than the entire
Dutch league. In 2005 even Holland's giants, PSVand Ajax each earned about
8 million from television – approximately a tenth of what Chelsea earned from
that source and about one sixth of what the average European top league club
earned (Vrooman, 2007). While leagues (the product market) are still domes-
tic, the labour market is international in scope; increasing emphasis is placed
on a primary product market funded by media income and then international
(European) club competition at the highest levels, also funded by media in-
come. This illustrates some convergence of European leagues towards the
model provided by US leagues which, as noted earlier, tend to be closed leagues
with admission by geographical franchise, which provides local monopolies to
the teams, although new franchises can be offered and they can move between
cities.
There are also other broad economic similarities between the contempo-
rary European and US sports leagues. In the US, geographic survival of teams
has tended to be linked to their market potential for franchises. In Europe over
the long-run, teams from areas of sustained weak economic and population
growth tend to go out of business or drop into lower leagues. Evidence that the
geographical location of professional football clubs in England has responded
to long-run economic forces is offered by Waylen and Snook (1990). As Fort
(2000) also argues, the play-off form of US sports, plus the feeding of talent
into the major leagues from the minor leagues, are economically equivalent to
the emergent European multi-stage tournaments and vertical structure, in
that both are mechanisms which drive talent to the highest level of competi-
tion in the light of spectator demand.
In summary, it is clear that structural differences are evident between the
leagues, but common economic processes appear to operate. It also remains
that the economic objectives identified for teams may be different between the
US and Europe, which for some authors, like Primault and Rouger (1999),
derives specifically from the relegation and promotion structure in Europe.
While Sloane (1971) acknowledged that it may be difficult to distinguish
between the objectives this issue does matter. If, as contest theory predicts,
economic value is likely to reside in smaller numbers of higher quality teams,
then the ability of a larger cohort of vertically integrated teams to effectively
coexist seems to be doubtful and suggestive of the need for changes in the
current design of tournaments. This could explain the origins of the elite
leagues at national level competition and the subsequent emergence of the
Champions League or Heineken cup as a stepping stone in the direction of
further elite level competition, where more money is available than in the
Premier League, Premiership or Magner's League. In this respect some
authors, for example Hoehn and Szymanski (1999) and Vrooman (2007),
propose a European super league for soccer based on a closed franchise, leaving
national leagues acting as minor leagues. Kesenne (2007b) argues that the
perceived need of many national teams to compete with teams that dominate
access to European competition is currently a stumbling block to cross-subsi-
dization within national leagues. A closed European league would go some
way to removing this obstacle, although it may be unpalatable to fans whose
support is based on a tradition of promotion and relegation.
7.6 CONCLUSION
In this chapter the process of the transition of sports from amateur to profes-
sional status has been discussed, noting the shift in both the economic orga-
nization and motivational dimensions of sports clubs. Contest theory has
Conclusion 201
been used to develop a more general set of principles that underpin the nature
of sports competition as an economic contest and provide the basis for exam-
ining where the economic incentives lie in tournament design. Particular
attention has been paid to the exercise of joint effort that is required to produce
a contest, deriving from a key production externality and the subsequent
tension that emerges based on the link between the exercise of effort by a
competitor to produce success and the uncertainty required to produce a
contest. The chapter argues that sports leagues are probably best viewed as
cartels – groups of teams that have to cooperate to generate and distribute
revenues, but that frequently have divergent interests and can have different
institutional features. Chapters 8 to 11 explore elements of sports leagues in
more detail.
Appendix 7.1 Some Properties of Tournament Discrimination

Given tournament probabilities:
P1 g
P1 ¼ ðA7:1:1Þ
P1 þ P2 g
g
Then, if g is 0, it follows that:
P1 0
P1 ¼ ðA7:1:2Þ
P1 þ P2 0
0
As any value raised to the power 0 is 1, then Equation 7.1.2 becomes:
1 1
P1 ¼ ¼ ðA7:1:3Þ
1þ1 2
which suggests a 50% chance of success in a two-contestant competition. If the
P
denominator of Equation A7.1.1 was Pi for i = 1, 2 . . . n contestants, then it naturally
follows that for any contestant the probability of success would be 1/n. Table A7.1.1 reveals a
little more of how variations in g affects the teams' success probabilities in a two-team league.
The effects are, in all probability, more extreme than those that would occur in a league of,
say, 20 teams, but they are of qualitative interest. Because this is a two-team league, when
Team 1 supplies a proportion x of total effort, Team 2 inputs the proportion (1 – x). The
“standard” model of competitive balance, discussed further in Chapter 9, is the special case
in which g = 1, when each team's success probability equals its share of league effort. The
bottom row reveals that if the teams make identical efforts, their success probabilities are
independent of the design parameter g, although it seems unlikely that they would not adjust
to change in g. The penultimate row shows that, where teams' efforts are very closely
matched, changing the value of g does not greatly affect their probabilities of success, at least
up to the value 4. But, as the differences between their efforts increase, so does the effect on
their success probabilities when g changes. Although g might in theory assume any finite
TABLE A7.1.1 Success probabilities as g varies, for given levels of team effort
Relative Probabilities Probabilities Probabilities Probabilities Probabilities

team when when when g = 1.0 when when
efforts 1,2 g = 0 g = 0.5 standard g = 2.0 g = 4.0
model
0.3 0.5 0.4 0.3 0.16 0.03

0.7 0.5 0.6 0.7 0.84 0.97
0.4 0.5 0.45 0.4 0.31 0.16
0.6 0.5 0.55 0.6 0.69 0.84
0.49 0.5 0.495 0.49 0.48 0.46
0.51 0.5 0.505 0.51 0.52 0.54
0.5 0.5 0.5 0.5 0.5 0.5
0.5 0.5 0.5 0.5 0.5 0.5
value, however large, it is clear that even at g = 4, the success probability of the team that
supplies 30% of league effort is down by a factor of ten compared to the standard model. Only
the values in the last row and in the fourth column are precise; the others are rounded-off to
the nearest two decimal places, except for the third row third column entry where we need a
third place of decimals to show what is happening. As already predicted, g < 1 increases the
prospects of clubs that supply less effort, while g > 1 increases the probability of success for
the team supplying most of the effort.
Appendix 7.2 The Derivation of Optimal Tournament Effort

Given the payoff to one specific contestant, 1:
R1 ¼ P1 V c1 P1 ðA7:2:1Þ
and substituting equation A7.1.1 for P1, leaves:
Pg1
R1 ¼ g V c1 P1 ðA7:2:2Þ
Pl1 þ P2
Using the quotient rule, this suggests that the first order conditions for a maximum return
would be, for a contestant choosing effort:
dR1 Pg Pg1
¼ V g g 2 1 g 2 c1 ¼ 0 ðA7:2:3Þ
dP1 ðP þ P Þ 1 2
Because it is assumed that each contestant is homogenous and competition is symmetric,
then it follows that P1 = P2 = P so:
Conclusion 203
Pg Pg1
Vg ¼c ðA7:2:4Þ
ð2Pg Þ2
or:
P2g1
Vg ¼c ðA7:2:5Þ
ð2Pg Þ2
or:
P2g1
Vg ¼c ðA7:2:6Þ
2Pg 2Pg
or:
P2g1
Vg ¼c ðA7:2:7Þ
4P2g
or:
1
Vg ¼c ðA7:2:8Þ
4P1
or:
1
Vg ¼P ðA:7:2:9Þ
4c
As detailed in the text, noting that the number of contestants is 2 shows that “1” in the
numerator is equivalent to “n 1” and “4” in the denominator is “n2,” as detailed in the text
for “n” contestants.
CHAPTER 8
Uncertainty of Outcome,
Competitive Balance and Bias
in Sports Leagues
OBJECTIVES
& To understand what is meant by uncertainty of outcome and competitive

balance
& To appreciate how uncertainty of outcome and competitive balance are
measured, and the difficulties with this measurement
& To appreciate the balance of evidence for and against the uncertainty of
outcome hypothesis
& To understand the role of home advantage and why it is not easy to account
for it connected with uncertainty of outcome
8.1 INTRODUCTION
Chapter 7 identified that uncertainty of outcome (UO) is theorized to be
potentially a key component of the economics of professional team sports.
This chapter explores, in some detail, how UO is understood and measured
by economists, because it plays a prominent role in the analysis of sports
leagues as shown in Chapter 9. Research into the extent to which UO
influences both spectator and media demand for professional sports is dis-
cussed in Chapter 10 and serves as a rationale for restrictive labour contracts
which enable teams to share revenues at the players' expense, discussed in
Chapter 11.
Consequently, Section 8.2 explores the time-dependent nature of UO and
draws distinctions between the short-run and long-run. Subsequent sections
explore each element of UO in more detail. The chapter concludes with a
discussion of home advantage in sports. Chapter 7 indicated that sports 205
206 CHAPTER 8 Uncertainty of Outcome, Competitive Balance and Bias in Sports Leagues
leagues have tried to account for this in offering balanced schedules of home
and away fixtures. Indeed, Cairns, Jennett and Sloane (1986) argued that the
most attractive game for spectators is that in which the visiting team is just
sufficiently ahead of the home side for home advantage to equalize prospects.
Further, Forrest and Simmons (2000) suggest that home advantage may be
important in permitting weak sides to survive in leagues historically domi-
nated by stronger ones. The link between home advantage and UO thus
requires discussion.
8.2 THE TIME-DEPENDENT NATURE OF UNCERTAINTY OF

OUTCOME
Chapter 7 outlined the UO hypothesis advanced by Rottenberg (1956) and
Neale (1964), and in particular, argued that the nature of sporting competition
requires a balance between competition and cooperation between contestants.
While the UO hypothesis has been deeply ingrained in the economics of
professional team sports literature, this chapter will demonstrate that its
quantitative impact is not at all well-established. It will become apparent that
this is partly due to the difficulty of deciding what precisely is meant by
“uncertainty of outcome” and how it can be measured.
Cairns et al. (1986) distinguished four temporal forms of UO. First, they
identified short-run UO where the emphasis is on the outcome of a particular
game; secondly, medium-term UO where the identity of the season's winners
is unknown; thirdly, within-season UO where several teams are “in con-
tention.” Finally, they recognized long-term UO (competitive balance), which
is concerned with persistent domination that may damage the whole league.
Each of these concepts is now discussed in some detail. The literature reviewed
employs the various measures of UO as an independent variable in a regres-
sion equation that estimates the effect of UO and other factors on attendances
and (more limitedly) broadcast audiences for sports in a research agenda
examining the demand for sports from fans. Chapter 10 examines the impact
of other variables on attendance and broadcast demand in more detail. Focus
here is on the concept of UO as this is referred to in the literature in various
guises.1
1
It is always tempting to assume that more recent studies are improvements on earlier
studies. It is true that the econometric tools available to the researcher have improved over
time (Borland and MacDonald, 2003). However, as will become apparent in Chapter 10,
research often tends to focus on specific issues and refinements, and does not necessarily
address others. Evaluation of the work thus requires a careful reading of the specific studies
concerned.
Short-Run (Match) Uncertainty of Outcome 207
8.3 SHORT-RUN (MATCH) UNCERTAINTY OF OUTCOME

Match UO is the idea that spectators prefer close contests and are more likely
to attend the next game if the teams are of much the same level of ability. There
have been three broad approaches adopted to measure this dimension of UO:
relative league standings; implied prior probabilities (p) of a home win through
the use of betting odds; and direct estimates of the probabilities of home wins.
Other things being equal, two teams immediately juxtaposed to one another in
the league are of similar quality, and ignoring both home advantage and the
probability of a draw suggests that both have an equal chance of winning.
Likewise, this might be measured by betting odds of “evens.” This suggests
that the closer p is to 0.5 the more attractive the match should be and the
greater the attendance. In the presence of home advantage and ignoring draws
the most attractive game would be where p is less than 0.5. This is to allow
home advantage to offset the visitors' greater quality, and research can be
designed to estimate the value of p that maximizes attractiveness in the
presence of home advantage. These alternative approaches are now discussed.
8.3.1 Relative league standings

Hart, Hutton and Sharot (1975), in an early paper on attendance demand for
English soccer, made use of the teams' immediate pre-match league positions
as indicators of the likely result and found the home team's standing insignif-
icant as a determinant of attendance. One flaw with the approach adopted was
that they used the absolute differences in league positions of teams. This is
because they wanted to use logarithmic measures of the variables in their
analysis, which has some statistical advantages. However, this cannot account
for the relative pre-match standings of the teams, as the sign of the difference
between home and away teams is lost (see Downward and Dawson, 2005a,b).
Subsequent studies examining the effect of relative league standings on
attendance are given in Table 8.1. It is worth noting that some recent innova-
tions in measurement have occurred. Forrest, Simmons and Buraimo (2005)
studied the audience for live broadcast Premier League games from 1993 to
1994 and 2001 to 2002. They devised a new index of uncertainty (partly in
order to refine any potential bias problems in betting markets, which are
discussed below) that incorporates both a measure of home advantage and
one of match UO. Home advantage is computed within their sample to be, on
average, about 0.57 points and is defined as the difference between points per
game won by all home sides and points per game won by all away sides,
evaluated over the previous season. They also identify a league performance
component as the difference between the home and away sides' mean points
208
CHAPTER 8 Uncertainty of Outcome, Competitive Balance and Bias in Sports Leagues
TABLE 8.1 Some findings on the effects of match uncertainty
Author/date Sport/period Dependent variable Indicator(s) Comment
Hart, Hutton English soccer; 4 teams; 3 Log of match Log home standing Insignificant
and Sharot 1975 seasons starting 1969–1970 attendance
Log away standing Significant
Log of absolute difference Insignificant
Borland and Australian Rules football Log of match Difference of league Insignificant
Lye 1992 1981–1986 attendance standings
Peel and Football League all teams Log of match Home standing Significant
Thomas 1988 1981–1982 attendance
Away standing Significant
Prob of home win Significant (not a
test of home
advantage)
Dobson and Football League; 24 teams; Log of match Log home standing Significant
Goddard 1992 2 seasons starting attendance
1989–1989
Log away standing Significant
Wilson and Malaysian semi-pro soccer Log of match Absolute difference Insignificant
Sim 1995 1989–1991 attendance of league points
Square of above Insignificant
(not a test of
home advantage)
Baimbridge, Premier League football Log of match Absolute difference Insignificant
Cameron and 1993–1994 attendance of standings
Dawson 1996
Square of above Insignificant
(not a test of home
advantage)
Peel and English rugby league Match attendance Absolute value of the Significant
Thomas 1997 1994–1995 handicap betting spread
Dawson, Dobson, English Premier League Log of disciplinary Home team uncertainty Significant
Goddard and football 1996–2003 points y = 1, r = 2
Wilson 2005 issued per match
Away team uncertainty Significant
Kuypers 1996 English Premier League Match attendance Estimated odds on a Insignificant
football 1993–1994 home win
Proportion of Sky As above Insignificant
subscribers
watching live football
Carmichael, English rugby league, Log of match Pre-match odds Significant
Millington season 1994–1995 attendance
and Simmons
1999
Falter and French soccer 1997–1998 Log of match Home standing Significant
Perignon 2000 attendance
Away standing Insignificant
Goal difference Significant
Absolute difference Insignificant
Price and
Sen 2003
NCAA Div 1-A Football Match Attendance Home wins in last 11 games Significant
Away ditto Significant
Squared difference of above Insignificant
Forrest, Simmons and BskyB live Premier League Log of television Composite index involving Significant
Buraimo 2005 football 1993–194–2001–2002 audience (millions) league form and home advantage*
Probit of decision As above* Significant
to broadcast
Pre-Boxing Day difference Significant
in relative wages**See F S &B
Short-Run (Match) Uncertainty of Outcome

Garcia and Spanish League football Log of match Difference of league positions Significant
Rodriguez 1992–1993 to5–1996 attendance
2002
Square of above Significant,
wrongly signed
Forrest, Beaumont, The Football Log of match PROBRAT (ratio of the Both probability
Goddard and League 1997–1998 attendance probability of a home win ratios were
Simmons 2005 relative to that of an away correctly signed
win) PROBRAT2 and significant
HOMEPPG (home team Significant
prematch points per game) correctly signed
209
210
Author/date Sport/period Dependent variable Indicator(s) Comment
AWAYPPG (away team Insignificant

prematch points per game) FBG&S regard
the PPG's as
indexes of
absolute team
quality rather
than of u/o per se
Meehan MLB 2000–2003 Match attendance WINDIFF (Abs) (absolute Significant
et al 2007 difference in teams' win
percents)
WINDIFFP (+ home team Significant
difference in win percent)
WINDIFFN ( home team Significant
difference in win percent)
Games left Significant**
Interacted with
WINDIFFP and
WINDIFFN
Owen and Super 12s rugby union, Log of match SUMPLACE (sum of the Insignificant
Weatherston New Zealand 1999–2001 attendance places each team lies
2004 behind fourth in the table)
PROBH (probability of Insignificant
home win)
PROBH4 Insignificant
Alavy, Gaskell, Premier League Football, The (minute SQDIFF (squared difference Significant
Leach and 248 games broadcsat on by minute) of probability of home/away win)
Szymanski 2006 Sky 2002–2005 rate of change
of TV ratings
PSDRAW (probability of Significant
score draw)
PNDRAW (probability of Significant
no-score draw)
SUMSQ (sum of squared Significant
deviations of probabilities)
Short-Run (Match) Uncertainty of Outcome 211
per game, evaluated using all available current season data. “Outcome
uncertainty” for Forrest, Simmons and Buraimo (2005) is then defined as
the absolute value of the sum of these components, which is at a maximum
when the sum is close to zero. They find this index statistically significant in
the explanation of television audiences, at least for games played after Boxing
Day, when the broadcaster Sky's choice of matches is more likely to be influ-
enced by current team performance.
8.3.2 Betting odds

Using information on league standings (and form) as an indicator of match
UO is simple and appears intuitive. However, an implication is that the
resulting indicator is based on partial (past) information. Viewing team
performance as a production function, league standings reflect the past
employment of playing talent as labour. However, it is conceivable that
knowledge about injuries, suspensions and loss of form may be available,
and would not be ignored by supporters as they make their decisions to
attend matches. There may even be forward-looking information to hand,
for example, that a particular player will have worked off a three-match
suspension in time for the next game. Bookmakers will have utilized
such information in deciding what odds to offer bets on a given result.
The same will be true of bettors in deciding whether or not to stake a bet
and how much they might stake. In this respect, betting odds can be
associated with UO.
Peel and Thomas (1988) exploited published betting odds to obtain a
forward-looking indicator of the probability of a home win. They estimate
the probability of a home win (HP) directly from betting odds by making two
simplifying assumptions, neither of which can be strictly true in soccer al-
though the first is fine for baseball. First they assume that there is no proba-
bility of a draw. Secondly they assume that bookmakers have zero “spread” –
i.e., margin for covering costs. This is discussed in Box 8.1. Under these
circumstances, quoted odds of 4 to 1 against a home win convert into a
probability of a home win equalling 0.2. While Peel and Thomas exploit this
relationship, they fail to test the UO hypothesis by inquiring if, other things
being equal, the value of the home win probability that maximizes attendance
is significantly different from 0.5, as the theory suggests it should be. Further-
more, a problem with the use of betting odds is that it is not known if the
“probabilities” identified by the betting public differ systematically from those
perceived by potential spectators. In this regard Peel and Thomas' approach
may suffer from sample selection bias. Nonetheless, a number of studies
follow up their lead, as detailed in Table 8.1.
BOX 8.1 THE RELATIONSHIP BETWEEN BETTING ODDS

AND UNCERTAINTY OF OUTCOME
The following examples hold when the bookmaker is as- Team 1 is 0.5 – the match is perfectly balanced. Those who
sumed to provide their services free of charge. Peel and backed Team 2 lose their stakes. Those that backed the
Thomas estimated the probability of a home win on this winners get back their stake money plus an equal amount of
basis, but were under no illusions that real bookmakers winnings. On average, the bookmakers simply recycle mon-
behaved in this way. They accepted the inevitable loss of ey between gamblers.
descriptive accuracy against the possibility of a gain in If the bookmakers are offering 3/1 against Team 1, the
predictive power. implied probability is found by solving:
Suppose, for simplicity, that two teams due to play a game Three to one against = 1/3 = P1/(1 – P1) ) P1 = 0.25
have prior win probabilities P1 and P2 respectively. Ruling In which case, P2 equals 0.75 and the odds are three to one
out the possibility of a draw, it follows that P1 + P2 = 1. The ON of Team 2 winning. If Team 1 wins, those who backed
implied odds on a win by Team 1, assuming that neither the Team 2 lose their stake and the successful bettors get back
bookmakers (nor taxation authorities) take any of the money their stake plus winnings equal to triple the sum staked. Had
staked, are equal to: Team 2 won, the bettors would have got back their stakes
Odds of a win by team 1 P1/(1 – P1) plus winnings equal to a third of the stake. On average the
Thus, if the bookmakers are offering 1/1 (even money) the bookmakers would be simply recycling money between
researcher would deduce that the probability of a win by those who made bets.
8.3.3 Direct estimates

An alternative approach to using betting odds to identify the probabilities of
results is to predict, in advance of each match, the probabilities of a home win,
an away win and a draw. Dawson, Dobson, Goddard and Wilson (2005) take
this approach, addressing the incidence of disciplinary sanctions in Premier
League football. They model the number of disciplinary points earned in a
match (one for a yellow card, two for a red) as a function of team qualities,
match uncertainty, match significance and attendance. UO indices seem to
have significant positive effects on the number of disciplinary points collected
by each team. The inference is that a closer, more uncertain, contest perhaps
tempts players to stretch the rules of play.
Owen and Weatherston (2004) compute an estimate of the probability
of a home win and employ it in an attendance study of Super 12 Rugby
Union, but it is found to be insignificant. Alavy, Gaskell, Leach and Szy-
manski (2006) use data on minute-by-minute changes in television ratings
to inquire whether uncertainty affects the size of the audience for a Premier
League football match while it is being played. The probabilities of score-
less draws, score draws, home and away wins evolve through the game and
their computation forms the basis of the work. They find some support for
the UO hypothesis, but also that viewers prefer to see a game that has a
Medium-Term Uncertainty of Outcome 213
winner and that people are attracted to games that take an unexpected
turn.
In a further innovation, Forrest, Beaumont, Goddard and Simmons
(2005) find that teams' current season pre-match points per game at home
and away, and estimates of the ratio of the probability of a home win
relative to that of an away win to help to explain attendance at football
league matches in the 1997–1998 season. While points per game are
viewed as indicators of team quality, Forrest, Simmons and Buraimo
(2005) regard points per game as a match UO variable; clearly the precise
role played by points per game is somewhat unclear.2
8.3.4 Summary
To summarize this section, Table 8.1 presents some of the studies in summary
form, with the sport, the period analyzed, the dependent variable, indicator of
UO and results summary. As can be seen, a mixed set of results are apparent
although on balance league standings, rather than their differences, and betting
odds appear to be significant determinants of attendance. There is also some
indication that this is more likely to be discovered with more recent studies.
8.4 MEDIUM-TERM UNCERTAINTY OF OUTCOME

8.4.1 Within-season uncertainty of outcome
Within-season UO concerns which team will take the championship title, as
opposed to the UO about the outcome of a particular game. There is also
limited agreement on how to measure this facet of UO. An early classic piece,
however, is Borland (1987) which used four measures of UO in an analysis of
annual attendance at Australian Rules football.3 In each of the four measures
an average of observations was made at four points during the season. The
intention was to obtain measures of UO as the season develops, rather than to
compute a single measure for the whole season. The last two observations
were also given double weight so that a given amount of UO is of more
importance as the season draws to its close. The first of the four measures
2
The authors also simulate how many spectators would be lost if perfect UO is established.
It seems safe to infer from the paper that the pursuit of greater UO, ignoring home
advantage, may reach a point at which further “improvements” in UO (as Cairns et al
hypothesized) begin to reduce attendance and gate revenue. But the point estimate of 2
million fewer tickets sold appears rather more speculative than the authors admit.
3
This paper is also noteworthy, being an early attempt to simultaneously model both long-
and medium-term UO.
considers the spread between the top and bottom teams in the league. The
second measure is the sum of the coefficients of variation of the numbers of
games won by all teams, which incorporates information about the perfor-
mances of all teams.
The coefficient of variation is a standardized measure of dispersion of data. It is
calculated by dividing the standard deviation of the data by its mean value.
The third measure is the average number of games a team is behind the
leader. Other things being equal, the fewer the number of games in hand
the lower the probability that the lead will change. The final measure is the
number of teams that have been in, or at most two games out of, the leading
four (at that time) teams that are eligible for the (play-off) finals, at each of the
four measurement points. There was some evidence that the third measure
affected attendance. As with earlier studies that used similar variables, for
example, Demmert (1973) and Noll (1974), there is inevitably some arbitrar-
iness in how one defines “close to” championship success. Table 8.2 presents
some studies and also measures of UO.
8.4.2 Seasonal (team-specific) uncertainty of outcome

The seasonal (team-specific) UO literature exploits the externality property
of UO inasmuch as while attention is on the seasonal outcome, as opposed
to a specific game as with the measure of UO just discussed, the underlying
concept is that spectators go to matches primarily to see their own teams
win. UO is a by-product of having several teams “in contention” for the
championship, preferably as late into the season as possible to maintain
spectator interest. There are differences of opinion about how to model this
and Appendix 8.1 shows how differences between two of the classic
approaches may be reconciled. It has been observed that researchers have
obtained slightly more robust estimates of the impact of this form of UO
(Borland and MacDonald, 2003).
The pioneering work of Jennett (1984) based on an analysis of Scottish
football attendances merits careful consideration, not least because his meth-
od has been adopted and/or modified by others, among them Borland and Lye
(1992), Dobson and Goddard (1992, 2001) and Kuypers (1996). Jennett pro-
posed a simply-measured variable to capture the development of UO through
the season, match by match. His evolutionary approach stands in stark con-
trast to the computation of a single statistic to represent uncertainty in an
entire season. He aimed to measure the uncertainty that attaches to every
game. In any given match, one or both teams may start with a chance of
championship “glory.” Matches where one team (preferably both) may win
TABLE 8.2 Some findings on the effects of medium-term uncertainty
Author/date Sport/period Dependent variable Indicator/s Comment
Jennett 1984 Scottish Premier League Match attendance Home significance Significant
football 1975–1976
to1980–1981
Away significance Significant
Relegation (H&A) Insignificant
significance
Borland 1987 Australian Rules Log of attendance Average number of Insignificant
football 1950–1986 per round* per games in hand over leader
capita *See Borland
Three other indexes All insignificant
Borland Australian Rules Log of match Sum of the number of Significant
and Lye 1992 football 1981–1986 attendance games required for both
teams to reach the finals
Games where both Significant
are in the top 5
Dobson and Football League; Log of match Log of home significance Significant
Goddard 1992 24 teams; 2 seasons attendance
starting 1989–1999
Log of away significance Insignificant
Baimbridge, Premier League Log of match Both in top 4 Insignificant
Cameron and football 1993–1994 attendance
Medium-Term Uncertainty of Outcome

Dawson 1996
Both in bottom 4 Insignificant
Team has won title Insignificant
Kuypers 1996 English Premier Match attendance Home champ sig, Significant
League football games left times
1993–1994 points behind
Home relegation Significant
sig index
Live Sky football Proportion of Sky Home champ sig Significant
matches 1993–1994 subscribers
watching live football
Home rel sig Significant
Falter and French soccer Log of match Season dummies Significant
Perignon 2003 1997–1998 attendance for each match
215
216
Dawson, Dobson, English Premier Log of disciplinary Home team champ sig* Insignificant
Goddard and League football points y = 1, r = 2
Wilson 2005 1996–2003 issued per match
Away team champ sig* *See Significant
D & G above
Carmichael, English rugby Log of match Pre-season odds on the Significant
Millington and league, season attendance division title
Simmons 1999 1994–1995
Garcia and Spanish League Log of match Product of points behind Significant
Rodriguez 2002 football 1992–1993 attendance and games left
to 1995–1996
Owen and Super 12s RU Log of Match SEASON, numbers Significant
Weatherston 2004 Attendance matches 1–11
Medium-Term Uncertainty of Outcome 217
the championship have more “championship significance.” Ideally research-

ers would be able to observe spectators' subjective expectations of their own
teams' championship prospects prior to every match but, of course, this is
impossible. Consequently, Jennett proposed a measure based on the mathe-
matical possibility of winning the title in any given year, and found statistically
significant results.4 The estimated significance of any forthcoming match in
this model is simply the reciprocal of the number of wins currently needed
(WN) for the team to win the championship. This is indicated in Equation 8.1:
SigJ ¼ 1=WN ð8:1Þ
In subsequent studies, Wilson and Sim (1995) found Jennett's measure

significant in their study of attendance at Malaysian semi-professional soccer
games. Dobson and Goddard (1992) provided a modified measure, converted
into an index that always takes a positive value. They find the home side's
match significance variable has a positively significant effect on English soccer
attendance, but not the away side's match significance variable.
Cairns (1990) noted that Jennett's model is essentially backward-look-
ing inasmuch as it requires one to know how many points the winners
accumulated. In contrast, a forward-looking model would enable the po-
tential spectator (and the investigator) to estimate the championship po-
tential of every game for each side in advance. Some success with a semi-
forward-looking index of match significance was obtained by Dobson and
Goddard (2001, Chapter 3) who examined twenty-seven years' of football
league match data between 1973 and 1999. They estimate a match-level
forecasting model that includes both teams' league success over the previ-
ous 24 months, plus more recent success and indicators of whether the
match has championship and/or relegation significance for the home and
away sides. Their model proved superior to random forecasts in predicting
home and away wins and the coefficients of the match significance vari-
ables were both correctly signed and statistically significant.
Dawson, Dobson, Goddard and Wilson (2005) used the same forward-
looking index of seasonal match significance in their study of the occurrence
of disciplinary sanctions, finding it to have a statistically significant effect on
sanctions imposed on the visitors, but not on the home team. Whether this is
primarily because visitors feel under more pressure playing away from home
and try a little too hard or because (see below) match officials tend to give the
benefit of the doubt to home sides is an interesting question.
4
This model cannot be applied to the current season, since it requires the researcher to
know exactly how many points will be required for championship success.
One of the less satisfactory aspects of Jennett's model, pointed out by Bor-
land and Lye (1992), is that while it gives sensible results for teams that have to
win every remaining game, it produces rather nonsensical results when a team
has more games left than wins required, suggesting the same significance level
for any number of matches left, say if only one more win is required. Borland
and Lye suggest (but do not implement) a revision of the original model that
takes into account the number of games left (GL) as well as WN. It incorpo-
rates both WN and GL, as detailed in Equation 8.2:
SigBL ¼ ðWN=GLÞ ð8:2Þ
Another example is Kuypers (1996), who proposes a number of indices of

forward-looking match significance involving GL and WN. The preferred
index is given in Equation 8.3:
SigK ¼ GL PB ð8:3Þ
PB is points currently behind the leader, which plays a role equivalent to

Jennett's WN. Kuypers (1996) finds SigK statistically significant in a study of
English football from 1993 to 1994. Garcia and Rodriguez (2002) employ SigK
in their study of Spanish league football and find it statistically significant.
One of the problems with these approaches is that they assume that
winning, say, two from two games is the same as, say, ten from ten, and do
not account for the difference in cumulative difficulty. Downward and
Dawson (2005b) offer a suggestion for combining these approaches, as
given in Equation 8.4:
SigDD ¼ ð1=GLÞexpðWN=GLÞ ð8:4Þ
Exp is the base of natural logarithms, while the exponent is clearly seen to
be SigBL so that Borland and Lye's contribution is directly represented in the
exponential model. Jennett's contribution is represented indirectly via GL, in
other words SigDD encompasses the earlier models. Appendix 8.1 indicates
how it resolves the anomalies of each approach.
8.4.3 Summary
As with match UO, within-season UO research produces some mixed results,
although here there is less agreement that the UO hypothesis receives support.
Studies summarized in Table 8.1 and 8.2 suggest that, in the shorter run at
least, a team's success is at least as important as UO for determining match
attendances.
Long-Run Uncertainty of Outcome 219
8.5 LONG-RUN UNCERTAINTY OF OUTCOME

8.5.1 Competitive balance
The desire to promote long run UO or competitive balance in sporting leagues
is due to the fear that domination by a few teams may result in a loss of
spectator interest, revenue and profit, as implied in Chapter 7. As noted above
Borland (1987) produced a seminal piece of work investigating medium-term
UO. However, another contribution was to employ a variable to try to capture
the effects of long-term UO on Australian Rules football match attendance by
the number of teams that have reached the previous three years' play-offs,
divided by the number of places available. A larger value of the variable is
consistent with a lower level of long-run concentration. No significant results
were identified, but it could be that three years does not constitute a sufficient
indication of the long run. In turn Borland and Lye (1992) employed a related
index that consisted of the sum of the number of times both contestants had
made it to the final stages in the previous three years. The significance of this
variable was uncertain.
In a long-run study of English soccer Dobson and Goddard (1992) used the
logs of the teams' mean final league positions and the season prior to that in
which the match took place as indexes of historical success, and found them to
be significant and correctly signed determinants of attendance. Furthermore,
while few attempts have been made to model the long-run significance of
individual games, there is one index that consistently shows the expected
(positive) effect on attendance and is almost consistently statistically signifi-
cant; this measure is to identify derbies.
Most research focusing on competitive balance has examined the evolution
of the dispersion of seasonal rankings of teams over time. Many researchers
focusing on US professional team sports have used the standard deviation of
win percent (a measure of dispersion of success) as an indicator of within-
season UO. For example, Fort and Quirk (1995) inquire whether the sample
standard deviation of win percent is significantly greater than the level that
would characterize a perfectly balanced league. Appendix 8.2 indicates how
this can be measured.
TABLE 8.3 Mean SDWPs in baseball and football by sub-periods, 1901–1990
League 1950–1959 1960–1969 1970–1979 1980–1990 1901–1990 average
AL 97(40) 83(39) 76(39) 65(39) n.a.(40)

NL 80(40) 82(39) 68(39) 65(39) n.a.(40)
NFL 210(144) 222(135) 208(132) 191(125) n.a.(140)
Note: n.a. = not applicable.

FIGURE 8.1 The Lorenz curve and Gini coefficient.
Some standard deviations of win percent (SDWP) covering five US sports

leagues for the period 1901–1990, are presented in Table 8.3, taken from Quirk
and Fort (1992). The numbers not in brackets in columns 2–5 are average
SDWPs computed for the periods indicated. The numbers in brackets are the
corresponding SDWPs for perfectly balanced leagues. Note that every number
has been multiplied by 1000 to avoid decimal points.5 In these leagues win
percent was less evenly distributed than as would occur in perfectly balanced
leagues.
Another way to report this phenomenon is to count the number of
win percents in the tails of the distribution. Fort and Quirk (1995) report
that, for the major league US professional team sports, there are too
many win percents in the tails of the distribution, to be consistent
with the sports being perfectly balanced, suggesting a wide variation in
performance.
Schmidt and Berri (2001) examined how well-balanced major league
baseball has been since 1901, using the Gini coefficient. This is related
to the Lorenz curve, which is illustrated in Figure 8.1. Here, the percentage
of teams and championship successes are plotted against one another. The
45 line indicates a situation of equal distribution of wins between teams,
and would constitute a perfectly balanced league. The Lorenz curve, which
is the curved line, plots the “actual” distribution of championship wins
per club. Because the curve lies above the 45 line this shows that fewer
clubs have won more championships than in the balanced case. The area,
5
These figures are approximations, as is made clear elsewhere in this chapter. Ideally, the
leagues should exclude draws, but the NFL does not satisfy this criterion and every team
should play every other league member the same number of times per season, none of these
leagues possesses this property.
A, divided by the sum of areas A and B can be used to derive the

Gini coefficient, whose value must lie between 0 and 1, with the latter
indicating complete inequality.
Schmidt and Berri's (2001) analysis suggests that concentration is not a
serious problem and has, if anything, declined since 1901. However, their
procedure was criticized by Utt and Fort (2002) on the basis that the Gini
coefficient is useful in measuring horizontal industrial concentration as its
values range from “1,” where the industry is a pure monopoly, to around “0”
where it comprises many very small firms. But, as Utt and Fort point out, no
team can monopolize a season's wins in baseball. They adjust the Gini coef-
ficient by comparing the actual distribution of win percents against the most
unequal case, where the champions win all their games, the runners-up win all
but those lost to the champions, and so on down to the bottom team which
loses every game. Not surprisingly, their adjusted Gini coefficients exceed
those of Schmidt and Berri, suggesting the leagues are in fact rather less
competitively balanced than the latter authors had concluded. Utt and Fort's
criticism applies equally to Michie and Oughton (2004) who estimate Lorenz
curves of team shares of total points earned in English soccer for the seasons
1950–1951, 1993–1994 and 2004–2005.
Why would the Gini coefficient of points share be calculated rather than win percents in
English football?
Hint: Consider the form of game results in soccer compared to baseball.
The calculations are based on point shares to allow for the fact that, in
football, games can be won, drawn or lost, and consequently win percents are
misleading. Michie and Oughton (2004) find that competitive balance in
English soccer has (apparently) declined sharply since 1993–1994 (see also
Szymanski and Kuypers, 1999). Significantly, Utt and Fort suggest that, as
nobody knows how to adjust for these departures from the ideal, researchers
should stick to standard deviations of win percents as indexes of competitive
balance.
The problem with this recommendation, however, is that standard
deviations measured over several seasons do not permit a reliable distinc-
tion between cases where the overall spread of results is more or less
identical year-on-year, but the distribution of success is more highly con-
centrated in one instance than another. This question is addressed
by Eckard (2001) and Humphreys (2002). Using essentially the same
approach, they partition the total league win percent variability into a
component that captures the variation of teams' annual win percents about
their own mean win percent or “time variance” and “cumulative variance”
that captures the variation in teams' cumulative win percents across all
teams.6 Other things being equal, an increase in the first element implies
an increase in competitive balance, whereas an increase in the latter ele-
ment implies a decrease in competitive balance. The ratio of the first
element to the total variation of win percents defines a competitive balance
ratio (CBR). Eckard (2001) finds that competitive balance in the American
League (AL) may have been markedly lower in 1995–1999 than in the other
periods, while in the National League (NL) competitive balance may have
been lower but not markedly so. Applying this approach to European sports
may be more problematic, because promotion and relegation changes
leagues' memberships annually, and draws are commonplace.
Eckard also examines the degree of concentration among the top four and
(separately) the bottom four teams in both the NL and AL in all sub-periods;
with five years in each there are 20 slots available in each sub-period. Herfin-
dahl indexes (HI) of team shares in those leading positions are calculated. The
Herfindahl index is the sum of squares of team shares (win percents in
Eckard's study) in the 20 positions available in each five-year period. For
example, if the same four teams feature every year (the highest concentration
by this definition) each obtains 25% of all the places, so the HI takes the value
HI = 252 + 252 + 252 + 252 = 2500.
What value would the index take if shares of wins rather than percents were used?
Hint: Introduce the appropriate decimal points to the calculation.
If genuine shares of wins were used in the analysis, then the HI would
emerge as 0.25. In this regard, the index reveals the “representative” number of
teams that dominate the championship, in this respect 0.25 implies four
teams. Eckard finds that the HI for the AL is highest in 1995–1999, which
is also the five-year period in which the number of teams securing places in the
top four is at its lowest, i.e., six. This is consistent with the findings based on
his decomposition of win percents. In every sub-period more NL (than AL)
teams made it into the top four, while the HI tended downwards consistent
with his findings that competitive balance in the NL improved over the period.
6
Humphreys uses standard deviations.
Humphreys (2002) computes decade-by-decade CBRs and HIs of first place

finishes, again only for teams that played each year in the decade. Note that the
HIs were computed from shares not percents (see Reflection Question 8.2).
The CBRs trend slightly upward and the HIs downward over the period,
suggesting improved balance in both the NL and the AL. Humphreys exam-
ines whether CBR, HI and the standard deviation of win percent affects
attendances, and finds that only the CBR is statistically significant and pos-
itively signed suggesting that, other things being equal, more competitive
balance drew more spectators to MLB. Elsewhere in the world Lee (2006)
models the attendance at Korean professional baseball for the 21 seasons
1982–2002, making use of the win percents of the top and bottom 20% of
teams and finds that the more evenly win percents are distributed the greater is
attendance, as expected.
More informally, Szymanski and Kuypers report (1999) that between the
start of the English football league in 1889–1890 and 1939, seventeen teams
won the championship; Aston Villa and Sunderland won on no fewer than six
occasions each, while Everton and Arsenal each won five. This constituted
almost half of all the seasons then contested – the World Wars disrupted the
Football League's programme. Since the resumption of professional football in
1946, Liverpool has taken their championship tally to 18, Manchester United
to 15 and Arsenal to 13. In the Netherlands Ajax with 29, PSV with 19 and
Feyenoord with 14 dominate the top division, while the Spanish domestic
league championship has been won 29 times by Real Madrid, 18 times by
Barcelona and 9 times by Atletico Madrid. Anderlecht on 28 occasions and
Brugge on 13 have won the league championship of Belgium. At the top level,
the European Cup/European Champions' League has been won nine times by
Real Madrid, six times by AC Milan, five times by Liverpool and four times
each by Bayern Munich and Ajax. The major domestic knockout cup competi-
tions have been similarly dominated. In cricket the England and Wales County
Championship – first contested in 1890 – has been won by Yorkshire on thirty
occasions with four shared wins, while Surrey have compiled 18 wins and four
shares. Of the 18 teams now participating Durham, Northamptonshire, Glou-
cestershire and Somerset have never won a championship.
Further, the Sheffield Shield interstate championship in Australia and the
comparable series (the Currie Cup) in South Africa have been dominated by
the most populous provinces; New South Wales and Traansvaal/Gauteng
respectively, with the next most populous provinces Victoria and Natal/Kwa
Zulu the closest rivals. Tasmania, the smallest and least populous Australian
state, has never won the Sheffield Shield. In these latter cases, however, one
should bear in mind that these data tell only a part of the story as (unlike win
percents) they ignore the teams that finish the season in lower positions.
Different conclusions might be reached if all teams' annual points' totals were
computed and analyzed. The more general point is that, as with UO in the
short- and medium-term, there is no ready-made index that one could argue
comes closest to capturing the multi-dimensional phenomenon that is com-
petitive balance. Moreover, what one wants to know is not so much whether
imbalance exists in sports leagues, but whether it is of a sufficient degree to
warrant concern. Nonetheless, Table 8.4 summarizes the effects of some
studies examining competitive balance on attendance. There is some evidence
that attendances are affected by competitive balance.
8.5.2 Evolution of competitive balance

Finally, although there is limited research there seems to be no convincing body
of evidence that sports leagues do evolve towards perfect balance. Dobson,
Goddard and Ramlogan (2001) inquire whether English football between
1926 and 1997 obeys Gibrat's law of proportional effect (Gibrat, 1931).
Gibrat's law suggests that the size of a firm and its growth rate are independent. If
Gibrat's law holds, one would expect the size distribution of the firms in an industry
to remain constant.
Clearly, if the bigger members of an industry systematically grow more

rapidly than smaller ones, and if the process continues, the industry will
become more concentrated. If the small firms grow faster and if the process
continues, the industry will tend to become less concentrated. Given enough
time, the sizes of all firms (as measured by, say, revenue or assets) would
converge on an “equilibrium” value, although whether this occurs also
depends on random factors. Dobson, Goddard and Ramlogan find that from
the 1920s until the late 1950s there was a period of steady adjustment towards
equilibrium. From the late 1950s until the late 1970s it is suggested that there
was no tendency to revenue convergence, but this occurred from the late 1970s
until about 1990 and convergence also seems to have broken down in the
1990s; the authors conclude there may have been two periods of relative
equilibrium, followed by two of major change.
The causes are not identified from the model itself, but the earlier breakdown
of stability came during a period of marked social change. Population drift
towards southern England and also from the inner cities, and increased access
to private transport may have been important, as well as the influence of the
media in the creation of national support for some teams. It also begins around
the time that the maximum wage was abolished, forcing clubs to increase
admission charges that had been virtually unaltered since the 1920s. The
1990s breakdown of convergence took place against the simultaneous effects
TABLE 8.4 Some findings on the effects of competitive balance on attendance
Borland 1987 Australian Rules Log of attendance Number of teams in the Insignificant
football 1950–1986 per round* per final stages previous three
capita *See Borland years divided by places available
Borland Australian Rules Log of match Sum of times both teams Borders on significance
and Lye 1992 football 1981–1986 attendance in finals previous three years
Humphries, Baseball, AL and Log of total league Comp balance ratio Significant
2002 NL 1901–1999 annual attendance
H Index of win percent Borders on significance
Std deviation of win percent Insignificant
Schmidt and Baseball AL and NL, Log of team Lagged Gini measures* * Significant
Berri 2001 annual data 1901–1998 annual attendance See S&B
Log of league As above Significant
annual attendance
Lee 2006 KPBL 1982–2002 Log of aggregate LTL, captures the dispersion
season attendance of the win percents of the
Long-Run Uncertainty of Outcome

top and bottom 1/5 of teams
Significant
225
of the Bosman Ruling that enhanced players' bargaining power, the influx of
money from broadcasters, and the (related) setting up of the Premier League.
Add the transition to all-seater stadiums and the necessary rise in ticket prices,
and there are plenty of factors that would be expected to shake up the industry
structure.
8.5.3 Summary
In summary, as Downward and Dawson (2000) argued, the upshot of this dis-
cussion and that of previous sections is that the UO hypothesis seems to be rather
overworked as a phenomenon that affects attendances inasmuch as it is not
ubiquitously important, despite the ingenuity with which it has been measured.
8.6 HOME ADVANTAGE

As indicated in Chapter 7, and noted above, UO is to an extent inextricably
bound up with the potential occurrence of home advantage. The phenomenon
of home advantage is well-established in some sports, e.g., soccer, baseball and
ice hockey, although not in major golf and tennis tournaments. Neville and
Holder (1999) survey the field discussing measurement issues and potential
causal factors. The simplest cases to analyze are balanced leagues, where it
might be expected that, in the absence of home advantage, the home team
should win 50% of games (exclusive of draws). Win percents that are signifi-
cantly in excess of 50% are noted based on two (alternative) measures. HWPa
includes drawn games; a team gains two points for a win, one for a draw and
zero for losing. HWPb is the number of games won at home as a percentage of
decided (won) games. Combining the results of nine studies of soccer (40 493
games) HWPa is equal to 63.9% and HWPb is equal to 68.3%. (Clearly the
second measure must always be at least as great as the first.) In balanced
leagues the home advantage passes around competitors, which tends to elim-
inate differences in team quality, so this factor does not need to be accounted
for before computing HWP.7
7
Golf, athletics and tennis championships and other unbalanced tournaments create new
problems. For example, it is not safe to infer from the fact that US golfers often win the Open
and (almost without exception) the USPGA that there is positive home advantage in the US
and a negative home advantage in Britain. The observation is perfectly consistent with the
alternative hypothesis that, by and large, American (male) golfers are better at stroke play
than their British counterparts. This may be allowed for, using players' pre-match
performance rankings as quality indexes, when the apparent ground advantage tends to
diminish, Nevill and Holder (1999). Chapter 11 discusses sample selection issues.
Home Advantage 227
Downward and Jones (2007) found four main causal factors that have
been identified and their influence investigated in the literature: familiarity
with the location; the effects of travel; location-related rules; and the
crowd. It seems the crowd is likely to be the most important of these in
European sports. Barnett and Hilditch (1993) found a significant home
advantage in English soccer accruing to clubs, such as Luton Town, that
had invested in artificial pitches. These have since been banned in English
and Scottish football. On the other hand, there was no pressure in the US
to standardize playing surfaces in baseball, suggesting that it is not seen
there as a serious problem. Koning (2004) finds evidence of a small but
significant level of home advantage in women's speed skating, probably
related in part to familiarity with the locations. There is some suggestion
that travel, especially across time zones, may contribute to home advan-
tage in the US, but this can hardly be a factor in European national sports
leagues (apart from Russia). Nevertheless, these sometimes show consid-
erably higher degrees of home advantage (see Pollard, 2006a), although it
might operate in UEFA competitions. There is no European equivalent of
the rule in baseball that the home team bats last, which is often assumed
to confer an advantage.
Some studies have suggested that home advantage increases with crowd
size, e.g., Schwartz and Barsky (1977) in baseball, Dawson, Dobson, God-
dard and Wilson (2005) in English soccer, while Downward and Jones
(2007) and Pollard (2006b) find that the effect may be nonlinear. One
possible explanation is that the most attended matches are ones that
attract more away and neutral spectators, so the crowd is less supportive
of the home side. Nobody is quite sure just how the crowd effect operates;
it might be that players are more confident in front of their home crowds or
that match officials' decisions are sensitive to the crowd. Neville, Balmer
and Williams (1999a) report that crowds may well influence officials' deci-
sions, as indicated by experimental tests of observees watching video
replays of 52 soccer tackles, 26 by the home and 26 by the away sides,
with half the participants hearing no sound. The participants who heard
sound tended to judge statistically significantly but slightly to the benefit of
the home side. Mascarenhas, Collins and Mortimer (2005) show consid-
erable inconsistency in the rulings of even well-qualified match officials
examining videos of rugby union tackles chosen from 60 hours of tapes
shown to three groups of expert officials (referees, referee coaches and line
judges) ranked by their experience. Dawson, Dobson, Goddard and Wilson
(2005) also find that there is inconsistency in soccer referee's decisions. If
(say) two or three decisions in a single match go in the wrong direction, the
effect on one of the clubs might be quite serious, e.g., elimination from the
228
TABLE 8.5 Some findings on home advantage
Morley and One-day English Cricket Win or loss (logistic Venue, attendance, home and 57% home wins but
Thomas (2005) 1996–1997 regression) away performance inconclusive on why
Boyko et al (2007) English Premier Goal differential Attendance, referee identity Evidence of bias
league 1992–2005
Johnston (2008) English Premier Goal differential Attendance, referee identity No evidence
League 2006–2007
Morton (2006) Super-14 and Tri-Nations Points difference Venue Home advantage
rugby 2000–2004
Clarke (2005) Australian Rules football % of games won; Venue 80% home advantage
1980–1998 margin of victory
Pollard and NL 1876–2002
Pollard (2005)
AL 1901–2002
NHL 1917–2003
NFL 1933–2002
English Football 1888–2003 % of home wins Time series description Greater in early years
(baseball and
basketball) % of
total points for
others
Jacklin (2005) English football 1946–2002 Ratio of home Time series description Reduction in home
to away wins advantage since 1945
Pollard (2006) Association Football % of total points Geographic area Home advantage
1998–2003 72 countries
Page and Champions League/European Probability of win Significant but decreasing
Page (2006) cup 1955–2006 (knockout when second-leg (53–59% approx)
matches) UEFA Cup is at home (logistic
1971–2006 Inter-Cities Fairs regression)
Cup 1955–1971 Cup
Winners Cup 1960–1999
Conclusion 229
European Champions' League. More worrying (although the sample size is

small), in view of the immense match-fixing scandal that has recently
shaken up Italian soccer, is the potential that inaccuracy and disagreement
between qualified officials may create for corruption. Table 8.5 summarizes
some of the literature.
The above brief discussion suggests two important factors connected with
uncertainty of outcome. The first is that, clearly, home advantage should be
controlled for in seeking to examine the UO hypothesis. The second point is
that, to the extent that bias affects results and thus UO, and inasmuch as
attendance affects the bias, then it suggests a degree of simultaneity exists
between attendance and UO which should be allowed for in studies of the
demand for professional team sports. As Chapter 10 illustrates, this has not
always been accounted for.
8.7 CONCLUSION
The above discussion suggests that there has been considerable innovation in
the design of indices of UO, but disagreement between studies remains as to
the effect of outcome uncertainty on match attendance (the traditional focus of
attention). However, there is stronger evidence that home bias affects sports
encounters. The main welfare concern is over long-term domination, its
possible effect on interest and ultimately on league revenue and profitability,
and it is these issues and their potential sources which now need further
analysis, which is undertaken in subsequent chapters.
Appendix 8.1 Reconciliation of Jennett's and

Borland and Lye's Models of Medium-term Uncertainty of Outcome
Downward and Dawson (2005b) suggest a simple equation determining match significance
to encompass both Jennett's and Borland and Lye's indices of UO. This is presented in
Equation A8.1.1:
SigDD ¼ ð1=GLÞexpðWN=GLÞ ðA8:1:1Þ
Exp is (as earlier) the base of natural logarithms, while the exponent is clearly seen to be SigBL
so that Borland and Lye's contribution is directly represented in the exponential model.
Jennett's model is represented indirectly via GL, in other words SigDD encompasses the
earlier models. To show how Jennett's model is incorporated requires some explanation of,
and justification for, the underlying model.
The rationale behind SigDD is relatively simple. It is assumed that the championship winner is
the team that does most cumulative work during the season; in this context the one that
accumulates most wins. Jennett's index may be seen primarily as an index of duration, WN
wins cannot be acquired in fewer than WN games, the successful team is the one that stays
the course. In contrast, Borland and Lye's index tracks the required work rate, crudely
expressed as (WN/GL). Work rate and cumulative work are different concepts; Manchester
City took four of a possible six Premier League points from Chelsea in 2004–2005, a unique
work rate against Chelsea. However, they did not sustain that effort long enough to mount a
title challenge. Neither the duration nor the work rate by itself measures work completed,
although as total work is essentially a sum of products of work rates and durations some
combination of the two is the minimal requirement for an index that aims to capture
cumulative work. This is what SigDD is designed to accomplish. GL in the denominator acts as
a clock, determining how long any required work rate (WN/GL) must be sustained. A team
remains in contention only as long as (WN/GL) 1. The clock mechanism captures Jennett's
model. High significance near the end of the season identifies a team that is close to
championship success; early in the season it signifies a team that is heading for failure. This
model differs in several respects from either of its progenitors. First the maximum
significance level (one to win, one to play) is exp, not one. In this regard the incorporation of a
“glory”dummy as with Jennett makes sense. That the model successfully avoids the pitfalls
discussed in the chapter is simply demonstrated in Table A8.1.1.
Table A8.1.1 shows that the encompassing model gives sensible outcomes whether Team A
has to win every game or not. The second column represents the case in which every game
must be won and the team enters a winning streak; this is the case in which Borland and Lye's
model fails, because it attaches the same score to each match. Given that the number of
games left declines, Jennett's argument is that each successive game should be accorded
more significance and the encompassing model satisfies this requirement. The third column
shows that, as expected, match significance goes to zero when the team loses its first game.
The fourth column shows that, needing to win one of three, if Team A succeeds in the first
game it plays out the season as champions; again there is no problem here. In the fifth
column, Team A is advancing towards the title very uncertainly, dropping two games. The
significance of each remaining game increases just as Jennet requires, although his own
model did not have this characteristic. The encompassing model (SigDD) clearly outperforms
its antecedents, considered as a theoretical construct; how it might fare in empirical work
remains to be seen.
TABLE A8.1.1 Significance of Team A's last three games under various assump-
tions about how many wins are needed (encompassing model)
Match Team A, WN = 3, Team A, WN = 3, Team A, WN = 1, Team A, WN = 1,
wins first two loses the first wins the first loses the first two
16th 0.33exp 0.33exp 0.33exp0.33 0.33exp0.33

17th 0.50exp 0.0 exp Champions 0.50exp0.5
18th exp Potential 0.0 exp Champions exp Potential
Champions Champions
Conclusion 231
Appendix 8.2 Calculating the Dispersion of Sporting Success in Sports

Leagues Using StandardDeviations
Computing the sample SDWP for a given season
Let the league have N members. At the close of the season every team's win percent Xi (for
the ith team) is known.
& STEP 1 Compute the deviation of each team's win percent from the mean; viz., for the ith
team compute (Xi – 0.5) – the ith team's mean deviation.
& STEP 2 Square each of the N mean deviations.
& STEP 3 Sum the squares.
& STEP 4 Divide the sum of squares by N*.
*For small samples one often divides by (N 1).
& STEP 5 Take the square root of the number you calculated at STEP 4. The resulting figure
is the SDWP of the league for that season.
& To arrive at the SDWP for a decade, compute the ten relevant SDWPs and take the
average.
Computing the sample SDWP for a balanced league

The key to understanding the mean and standard deviations of win percents in a
competitively balanced league lies in noting that sporting leagues in the US can be
understood in terms of the binomial probability distribution. This distribution provides a
means of answering binomial probability questions that are common in many aspects of
business and economics. Typically, binomial problems have the following characteristics:
& there are “n” independent trials;
& there are two possible outcomes to each trial, success or failure;
& the probability “p” of a particular outcome occurring is the same in each trial.
Under these conditions, the binomial random variable is the number of successes that occur
in “n” trials. This distribution describes some US sporting leagues that are balanced. For
example, taking the league just described, there are:
1. “n” fixtures;
2. fixtures have to be won or lost;
3. crucially, the probability of a win for a team is the same in each fixture and must equal 0.5
in a balanced set of fixtures.
This means that one can define win percents as a binomial random variable. The mean of a
binomial distribution is “np” which, in a sporting context, must be equal to 0.5n. The
standard deviation of a binomial distribution is Hnp(1 p), which in a sporting context is
H0.25n. As the win percent for any team is w/n, the mean win percent will be given by n(w/n)
= w = 0.5. The variance of the win percent will be equal to Var (w/n) = Var (w)/n2. As Var (w) =
np(1 p) = 0.25n, Var (w/n) = 0.25n/n2 = 0.25/n. The standard deviation of a balanced
league win percent will be equal to 0.5/Hn.
CHAPTER 9
Cross-Subsidization in
Professional Sports Leagues
OBJECTIVES
& To understand the economic theory of sports leagues

& To understand the rationale for, and forms of, cross-subsidization in
professional sports leagues
& To assess the impacts of cross-subsidization on competitive balance in sports
leagues
& To appreciate the empirical evidence relating to the effectiveness of cross-
subsidization
9.1 INTRODUCTION
Chapter 7 introduced the “peculiar economics” of professional team sports,
suggesting that there is a need to coordinate clubs' activities in leagues. In
this chapter the main forms of cross-subsidization policy employed in
leagues, purportedly to ensure that competitive balance is preserved, is
discussed. The next section clarifies the economic rationale for cross-sub-
sidization. Section 9.3 then outlines the variety of policies that have been
employed in sports leagues. Section 9.4 outlines the “standard model” that
has been employed to theoretically analyze sports leagues. It can be argued
that the model has the character of US sports leagues, inasmuch as it is
assumed that teams have a profit maximization objective, as discussed in
Chapter 7, and that the labour market for players is closed. Section 9.4
considers the implication of changing the objectives from clubs to a version
of utility maximization, in which teams seek to maximize their wins
subject to a profit constraint. It is argued that this might be a more
appropriate assumption for European leagues. Section 9.5 then further
relaxes these assumptions to explore the impact of having spectators caring
about absolute playing quality, that the scale of markets may produce
diseconomies of scale from playing talent, and that the supply of talent 233
234 CHAPTER 9 Cross-Subsidization in Professional Sports Leagues
might be variable because of open, international markets. In short, the

chapter argues that the potential for a corrective impact of cross-subsidi-
zation of teams appears to lie with revenue sharing in a win maximizing
league, but remains an area requiring further theoretical and particularly
empirical analysis. The chapter concludes with a discussion of the rele-
vance of the UO hypothesis.
9.2 THE RATIONALE FOR CROSS-SUBSIDIZATION

An economic rationale for cross-subsidization lies in the market structure
of sporting leagues which, as discussed in Chapter 7, can be viewed as
oligopoly markets operating as producer cartels within which clubs need to
compete in sporting terms, but also to cooperate to ensure that the sports
are managed effectively. To attain these ends, leagues have traditionally set
the terms of sporting competition and have directly influenced the eco-
nomic aspects of competition, influencing admission price structures, ne-
gotiating television deals and sponsorship arrangements and last, but not
least, controlling the terms on which players may move between clubs. It is
important to note in this regard that sports leagues have practised these
policies in spite of them being generally prohibited in other industries.
Public policy agencies have traditionally taken the view, informed by eco-
nomic theory, that an industry's attempt to regulate the terms under which
its members operate will act contrary to the public interest. This is
enshrined in legislation such as the Sherman Act (1890) and the Clayton
Act (1914) in the US, the Monopolies and Restrictive Practices Act (1948)
and the Restrictive Trade Practices Act (1956) in the UK. Nowadays, UK
competition policy, as with other European countries, is subsidiary overall
to EU competition policy, as discussed in Chapter 2.
Despite these legal frameworks there has been little historical govern-
ment intervention in sporting markets. In 1922 the US Supreme Court
ruled that baseball was exempt from the provisions of the Sherman Act,
because it did not represent interstate commerce and although this ruling
has often been criticized it has never been overturned. While other US
team sports have never attained the same degree of immunity from the
provisions of competition policy, they have been relatively leniently trea-
ted. Matters have tended to change, however, since the 1960s with a
general reduction in the degree of regulation in both the product and
labour markets in which cross-subsidization has typically operated. This
has generally arisen concurrently with the reorganization of league finan-
cial arrangements, as well as a general political shift towards the advocacy
of free market allocation of resources.
Methods of Cross-Subsidization 235
9.3 METHODS OF CROSS-SUBSIDIZATION

While differing in detail, leagues have traditionally intervened in sporting
labour markets or regulated the distribution of club revenues. Targeting
sports labour markets may seem to be an unusual way to affect competitive
balance insofar as one is seeking directly to influence the price and avail-
ability of inputs (players) rather than output (sporting results), which is the
ultimate object of the policy. However, players' salaries and wages comprise
the bulk of professional sports teams' costs in both the US and Europe.
Secondly, unlike other forms of production, it is labour in the production
function of sport, the players, who ultimately determine performance on
the field.
9.3.1 Sporting labour markets

Leagues have attempted to influence clubs' talent, finance and results by
applying three major types of labour market policy instrument: drafting sys-
tems; salary caps; and reserve option arrangements. The best known example
of a drafting system is the “Rookie Draft” in American football, introduced by
the National Football League (NFL) 1936. The National Basketball Associa-
tion (NBA) employed a similar system in the 1950s and Major League Baseball
(MLB) introduced a draft in 1965. The National Hockey League (NHL) fol-
lowed this pattern a few years later.
For example, the source of recruitment to the NFL is college football,
which comprises amateur players. The reverse draft ensures that every
year the professional teams obtain bargaining access to the new crop of
graduates in reverse order of how they finish in the professional league.
The intention is to give last year's least successful teams the first option
to pick new recruits. Consequently, the weaker teams have the opportu-
nity to enhance their stock of playing talent at a price that is probably
discounted on the player's genuine market value. Over time the intention
is that such a system reallocates talent between the teams to promote
competitive balance. Of course, in hierarchical leagues, with promotion
and relegation, the replacement of weaker teams with stronger teams
reallocates talent in a similar way (Fort, 2000). There is, however, one
potentially major flaw in the thinking behind drafts or, for that matter,
how promotion and relegation affects talent distribution. For example, in
the draft system, despite the opportunity for weak teams initially to sign
new, higher quality talent, there remains an economic incentive for the
stronger teams to buy talent from the weak ones, and the weak teams to
sell talent to the stronger teams.
Why would there be an incentive to trade rookie players?
Hint: Think about the relative cost of the players.
Leaving aside the personal ambitions of players to wish to play for

stronger teams, the clubs would have incentives to trade because the
player will probably be offered a contract with a wage below their market
value by the team allocated first pick in the draft. Consequently, the gap
between what the player can earn for the weaker team and the true market
value of the player becomes an economic rent over which teams can
bargain. The stronger team can offer the weaker team more than they
have paid for the player, while not paying the player their market value. It
follows that both clubs can make a relative financial gain through the
trade. The player may or may not receive some of this gain; this will
depend on the player's contract.
Salary caps, unlike drafting systems, directly target the financial cost
of talent. Salary caps can either involve fixing the cost of individual
and thereby aggregate player costs by establishing a maximum wage,
such as existed in UK in soccer from 1901 until 1960, they can
represent a maximum absolute amount that clubs can spend on players
in total, or they can be expressed as a percentage of team (or league)
turnover. In the US, the NBA, hoping to strengthen weaker teams'
finances, adopted a salary cap in 1980, after the alternative of gate
revenue sharing was blocked by the combined efforts of the strongest
drawing teams.
Given that small-town clubs are forced by the salary cap to buy more
success than their revenues can profitably support, the league as a cartel
has a perpetual enforcement problem. In European rugby union, the finan-
cial difficulties that some clubs and unions from smaller countries have
experienced following rugby union becoming professional have led to the
introduction of a salary cap. The relative financial fragility of rugby league,
despite its early professionalism as discussed in Chapter 2, has meant that
a salary cap has been in place in the Super League. Of course, the frequently
offered justification for the control of wages and/or wage bills is that, in
principle, it makes the best talent affordable to all teams and/or it prevents
teams from hoarding talent, which will help to produce competitive
balance. As with the draft, however, potential problems apply to salary
caps. On the one hand, illicit “side payments” may be offered to players
to ensure that they play for the clubs that have the economic resources to
Methods of Cross-Subsidization 237
attract them.1 On the other hand, the “farm” system developed in baseball,
discussed in Chapter 11, to provide opportunities to hoard players, while
not having them appear on the payroll. It follows that economic incentives
do exist to avoid the policy intent of the salary cap.
The other major labour market policy instrument employed by sporting
leagues has been the reserve option clause, the most famous example of which
has persisted in baseball since 1880. Chapter 11 discusses the impact of these
clauses in more detail, however, the main point is that baseball's clause es-
sentially tied players to clubs for their working lifetimes; when a club signed a
player the relevant clause gave the club the option to renew the contract when
it expired. In the early days, the player had little choice but to accept the new
contract, as unless the club released him he could not seek employment with
another club, and thus had no option but to retire from the sport. The “retain
and transfer” system in European soccer is another example; for a player to
move between clubs required the transfer of the registration document of the
player with the governing body. However, it remains, as discussed above, with
the other forms of cross-subsidization that incentives exist for teams to trade
players. Indeed, one can view the fees that are paid between teams to transfer
registration as a literal representation of the rents that exist.
9.3.2 Revenue redistribution

The other main form of cross-subsidization adopted by sports leagues is the
redistribution of revenue. Traditionally, of course, this has concerned the gate
revenue – monies paid by spectators at the turnstiles. At the outset, profes-
sional baseball operated a 50:50 split of gate revenue, one half going to the
away team.
How does gate sharing produce a more equitable allocation of resources?
Hint: Think about the relative scale of demand for different clubs.
Because imbalance is created by teams with different revenue base, as

revenues are then spent on playing talent, to redistribute some home team
revenue implies that a greater absolute amount of income is transferred from a
1
Such payments might be “in kind,” such as the provision of housing, use of a car or other
perquisites. Amateurism might be thought of as an extreme form of salary cap. Much
anecdotal evidence exists of the use of “boot” money, the offer of employment in owner-
firms, etc., to attract players to specific clubs.
larger club to a smaller club than the other way around. This means that the
larger club experiences a net reduction in revenue, while the smaller team
obtains a net gain in revenue.
As in many other team sports, away team shares of gate revenue have fallen
steadily over the years, and currently a variety of arrangements exist in the US
ensuring that the home team receives the largest share of revenue. The NFL
operates an unusually generous 60:40 split in favour of the home club, while
the NBA and NHL have no gate sharing. In Europe, similar arrangements have
applied. In association football, for example, in England between the 1920s
and the 1980s an 80:20 split on gate revenue existed in favour of the home
club. An alternative form of gate revenue sharing is a “pool” system, whereby a
predetermined share of league-wide gate monies goes into a central fund to be
distributed from the centre, one example being the EFLs 4% levy on all
receipts, the proceeds of which were distributed equally. Similar arrangements
existed in rugby league in the UK, but currently no gate sharing arrangements
exist. Gate revenue is still shared in cup competitions.2
As far as television revenue is concerned, in the US local television coverage
provided no revenue for visiting teams (Fort and Quirk, 1995), although
national television revenue was shared. Fort and Quirk hypothesized that in
a one-team-one-vote world, egalitarian distribution of national television rev-
enue is more or less certain, as signing the national television contract requires
virtual unanimity among league members. As discussed in Chapter 7 and
further in Chapter 10, this system broke down in the EFL, whose First Divi-
sion, with FA support, departed to form the Premier League, taking most of the
television revenue with it. Sharing local television revenue is virtually a non-
starter given the supermajorities required to change the rules of US sports
leagues. Notwithstanding these details, it should now be clear that the effec-
tiveness of revenue distribution actually depends on the labour market. The
aim is to endow weaker teams with income streams that can help them to buy
better playing talent. For the reasons noted earlier, there is no reason why
players would move to the weaker teams.
Having discussed the forms of cross-subsidization, attention now turns
toward the theoretical approaches by which economists have sought to
2
Pool and direct gate sharing can have very different effects on club's finances, except in the
pedagogic two-team league where each team participates in all the league's games. In an N
team balanced league where each team plays the others twice a season there are N(N – 1)
matches, which amount to 90 in a ten-team league, of which any one team plays in 18.
Under direct (one to one) gate sharing how much a team gains (loses) is determined
independently of the revenue from the other 72 matches. Under a pool scheme, what each
team gets depends on the gate receipts at all 90 games.
An Economic Framework for Understanding Cross-Subsidization 239
evaluate their effects. This requires outlining the construction and controlled
manipulation of a variety of theoretical models of sports leagues. It should be
noted at the outset that this literature is relatively technical and much debate
exists about the specification of the models and their implications. What
follows, therefore, is an inevitably circumscribed account.
9.4 AN ECONOMIC FRAMEWORK FOR UNDERSTANDING

CROSS-SUBSIDIZATION
A starting point for the analysis is that leagues are constructed in which there
is imbalance, deriving from different scales of demand for their output, defined
as wins, by fans, as teams look to hire playing talent according to its costs.
Following the economic methodology outlined in Chapter 1, the various
policies are then introduced to the analysis and predictions made about the
effects on the league; for example, what happens to competitive balance,
profits, player remuneration, etc.
There are two important zero–sum relationships in these models, one of
which can be relaxed in the process of exploring alternative scenarios. The first,
inescapable, zero–sum restriction arises from the fact that win percents in a
sports league sum to 0.5N in an N team league in which no draws are possible.
In a two-team league, therefore, win percents sum to unity. If N = 4 then the
sum of win percents is 2, etc. This is because if Team A wins, its opponent must
lose; if Team As win percent goes up by 1%, the sum of the win percents of all
the other teams must fall by 1% (in the two team case the other team's win
percent falls by 1%). The second zero–sum restriction is that the sum of
changes in teams' stocks of talent is zero, which arises from the assumption
that the amount of talent is fixed and that teams hire it all at the going market
equilibrium price. If one assumes either that the supply of talent is variable or
that teams do not necessarily hire the entire stock, this second zero–sum
restriction no longer applies. This possibility is discussed in Section 9.5.
To begin the analysis, a highly influential two-dimensional model evolved
by Fort and Quirk (1995) is presented that draws on earlier work by El Hodiri
and Quirk (1971), Quirk and El Hodiri (1974) and Rottenberg (1956). The
two-team league is pedagogically convenient as the questions may be posed
and solutions found graphically rather than algebraically. The original model
assumes that team owners pursue maximum profit and that they buy talent in
a perfectly competitive market,
9.4.1 Profit-maximizing competitive equilibrium

Fort and Quirk's (1995) two-dimensional model assumes that a club's total
revenue (R) is driven by two factors; their win percent (w) and their home
market size (m), which is usually assumed to be proportional to local popula-

tion. The relevance of these assumptions is presented in Chapter 10. The “ith”
club's total revenue Ri (i = 1, 2) is given by Equation 9.1:
Ri ¼ Ri ðmi ;wi Þ ð9:1Þ
It is further assumed that qRi/qmi >0, which means that big market (big
city) teams, other things being equal, have higher revenues than small market
teams. This is consistent with the demand for the output of large market
teams being greater than in the case of small market teams, for any given level
of performance. In general, it is also assumed that the MR of winning is
positive, i.e., qRi/qwi >0, but that it also declines q2Ri/qwi2 < 0 as the win rate
increases. This is consistent with a declining growth of interest by spectators
in a winning team (see for example Vrooman, 1995). Fort and Quirk (1995,
2004) and Kesenne (2004, 2006) also argue that eventually declining short-
run UO causes the team's total revenue to fall. By implication this means that
its MR eventually becomes negative. This is consistent with the externality
effects of UO reducing demand.
Fort and Quirk (2004) further impose the restrictions that total revenue
is zero whenever the team has either no talent or has monopolized the
league's talent. In either event production is impossible. Another property
of the model (see also Appendix 9.1) is that a 1% change in the amount of
talent “t” hired by a team produces a 1% increase in win percent, indepen-
dent of the existing level of win percent. Szymanski (2004) argues this is
the result of a combination of assumption: (a) that the entire fixed supply
of talent is hired; and of assumption (b) that talent can be measured on a
scale such that the sum of talent adds to unity. In combination, they
ensure that it is immaterial whether a club is viewed as choosing its talent
or its win percent.
This can be illustrated by considering that a team's win percent “w” can be
related to the amount of talent it hires relative to total talent, in the two-team
case, where ti is the actual number of talents hired by the ith team, i.e.:
w1 ¼ t1 =ðt1 þ t2 Þ ð9:2Þ
This is a contest success function, as discussed in Chapter 7, in which the
likelihood of winning a contest is directly connected with the relative share of
talent held by a team. Differentiating the above expression with respect to t1
leaves:
ðLw1 =Lt1 Þ ¼ ft1 þ t2 t1 ð1 þ dt2 =dt1 Þg=ðt1 þt2 Þ2 ð9:3Þ
or:
ðLw1 =Lt1 Þ ¼ 1 ð9:4Þ
if dt2/dt1 = 1, which is made possible by the assumption that a fixed amount
of talent is available and is wholly hired by teams, and that the “ti” sum to
unity, i.e., talent, is measured on a scale such that it can be summed to unity.
In other words, whatever the amount by which Team 2 alters its stock of
talent, it is exactly equal and opposite in sign to that of Team 1's change.
Consequently, when the total amount of talent can be normalized on unity,
a 1% increase in t1 produces a 1% increase in w1. Clearly, under these condi-
tions, it is immaterial whether the team chooses a level of talent or a level of
win percent.
The model can be illustrated in Figure 9.1, which depicts the MR
curves (in the positive zone) of big city Team 1, MR1, and small city Team
2, MR2. The horizontal axis measures the teams' win percents and con-
sequent share of talent, because of the discussion above, and the vertical
axis measures revenue and costs. As drawn, Team 1's MR exceeds that of
Team 2 at all win percents, reflecting the former's greater drawing
power. The intercepts of the marginal revenue curves with the horizontal
axis show that at zero MC the teams' win percents sum to a value in
excess of unity. This is consistent with the idea that teams potentially
seek to buy more talent than is available, ensuring that talent will have a
positive price, it is a scarce good, and any price below the market equilib-
rium price will not be sustainable and will be adjusted back towards
equilibrium.
The assumption that all talent is hired in equilibrium enables
Figure 9.1 to be redrawn in a more convenient form as a two-team league.
Figure 9.2 measures Team 1's share of talent reading from left to right and
Team 2's share reading from right to left. This now illustrates nicely that
any increase in Team 1's win percent must be met by a reduction in Team
2's win percent. It is important to note that the teams' MR's are equal at E,
where Team 1 has a win percent of 0.6 and Team 2 a win percent of 0.4.
This is the equilibrium level of production in the league, in which the
larger team produces more wins than the smaller team, and consequently
provides a visual example of an unbalanced league.
Why is position E an equilibrium position in the market?
Hint: think about the conditions required for a profit maximizing firm to set its output level.
Point E is the equilibrium level of “output” (for each team) in the league,
because here any point to the left implies that for any given win percent
Team 1 has a higher MR than Team 2, and vice versa with any point to the
FIGURE 9.1 Two-team league with Team 1 as the large market team.
right. Under such circumstances the team with the greater marginal rev-
enue would be able to hire more talent relative to the other team, who
would likewise seek to sell talent. Because talent produces wins, this
means that the win percent moves towards the position associated with
E, which here is Team 1 at 0.6 and Team 2 and 0.4. As MR falls with
increases in win percents, it is only at point E that talent will cease to be
hired and hence become an equilibrium, because here both teams are pre-
pared to pay the same price for talent. In this sense the market determines
the MC of talent, as illustrated.
More specifically, one can argue that the team's total cost “C” is a
function of the talent it hires “t” as indicated in Equation 9.5. Here, C
also includes a fixed cost component “k.” Fixed cost is irrelevant to short-
run profit maximization, which depends solely on marginal (ultimately
variable) cost, thus fixed cost is often left out of the equation or subsumed
under variable cost on the grounds that clubs' fixed costs tend to be pro-
portional to their variable (mainly labour) costs:
Ci ¼ cti þ k ð9:5Þ
In this chapter it is assumed that k = 0, following convention.

The parameter “c” is then the price (to be determined) of a unit of
talent.3 Implied in Equation 9.5 is that all clubs pay the same price per unit
of talent, which is equivalent to assuming that in the labour market clubs
behave like perfect competitors. As the league hires all the available talent,
c is the market clearing price of talent, as described earlier.4 An important
point to note in this regard is that win percents could be driven to equality
(0.5), but then the marginal revenue of Team 1 would exceed that of Team
2 by the amount FG. Clearly the market is not going to produce this
outcome.5 In short, at any other distribution of wins than point E, both
teams could increase their profits by, and thereby have an incentive to
engage in, trading talent.
Figure 9.1 (and by implication Figure 9.2) represent the MR curves as
parallel lines (having the same slope). This implies that each team's supporters
have identical dislike of short-run predictability of outcome. Should Team 2's
supporters have a weaker distaste for predictability than Team 1's, its MR
curve will be flatter, in which case for some range of win percents Team 2's
MR may exceed that of Team 1. If Team 2's MR exceeds Team 1's at win percent
0.5 (see Kesenne 2004), the small city team will dominate the league. Further-
more, it is unlikely but not impossible that a small town club may, by superior
management on and off the field, dominate the big city team at all win
percents; in terms of Figure 9.1 this would mean that Team 1 is the small
town side. Having set up the standard model it is now possible to examine how
the introduction of cross-subsidization affects the distribution of talent, win
percent and competitive balance.
3
A player is viewed as a bundle of talent, the more gifted player simply has more talent than
the average player and as each is paid according to the talent she/he possesses individuals
(and teams) earn differing amounts (see Appendix 9.1).
4 P
The stock of talent t is given at any moment, which is probably not too unrealistic an
assumption for US team sports, which are relatively unique to that country, although Latin
America is increasingly producing talented baseball players, some of them employed by the
major US leagues. Should soccer finally take off in America, the fixed stock of talent
assumption would be less realistic for US team sports. European sports teams, especially in
soccer and rugby union and league, find talent worldwide, i.e., in elastic supply, therefore the
standard model is not so obviously applicable to Europe. Vrooman (1995) allows for the
possibility that the big city team may have to pay more per unit for its talent, which might act
as a restraint on the amount it wants to hire. There may be something in this as the earnings
of superstars in sports and entertainment often seem to be much higher than their talent
would suggest is appropriate. Rosen (1981) provides a plausible explanation of the
phenomenon.
5
Most models build on the implicit assumption that causation flows in one direction only,
from performance (win percent) to revenue. There is some statistical evidence, see for
example Davies et al (1995), that causation could also flow in the opposite direction; teams
that enjoy higher revenues can invest more in talent and therefore obtain more success.
FIGURE 9.2 Equilibrium win percent and no incentive to trade players.
9.4.2 Cross-subsidization in a profit-maximizing league

9.4.2.1 Reserve option clause
Consider what ensues if a reserve clause is introduced into the sport, implying
that players' contracts are owned by their employers. This has the effect of
depressing the price of talent below its equilibrium value, c, therefore the teams
will both want to hire more talent than is available, but the reserve clause
prevents the price being bid up to c. Suppose once more that Team 1 has bought
enough talent to attain a win percent of 0.7. Irrespective of whether its MR of
winning is above or below the constrained price of talent, its MR is below that
of Team 2. Provided teams are permitted to buy and sell players' contracts,
Team 1 can sell talent to Team 2 and both will become more profitable, and this
process can continue until their marginal revenues are equal and league reve-
nue is maximized. It can be concluded, therefore, that the effect of a reserve
clause on a league where teams are profit maximizers is not to redistribute
talent in favour of weak drawing sides, but to restrain players' wages. Weak
drawing teams will benefit financially from the introduction of a reserve clause,
at the expense of players and of strong drawing teams to whom they can sell
talent. Consistent with their prior expectations, Fort and Quirk found no
evidence from standard deviations of win percent that competitive balance in
MLB improved following the introduction of (partial) free agency in 1975.
9.4.2.2 Player draft

Consider the effect of a successfully operating reverse order draft as a result of
which weak drawing teams will have acquired more talent at a lower price than
the market would suggest, and will obtain a higher win rate than they can
profitably employ using talent that could be profitably employed by the strong
drawing teams. As implied earlier, profit maximizing teams have strong incen-
tives to trade talent (since the 1960s the NFL has restricted cash sales). If
trading should continue until both teams MRs are equalized, the distribution
of players will be what it would have been in the absence of the draft. Once
again, the real effect of the draft is to lower salaries in favour of profits; weak
teams are made more profitable than they would otherwise be, through paying
lower wages and selling talented players to stronger teams. Fort and Quirk
again found no evidence (comparing standard deviations of win percent) that
the draft introduced by the NFL had any significant effect on competitive
balance; they found mixed evidence for MLB where there did seem to be
a slight reduction in the AL. They also found contrary evidence, in that the
concentration of championships increased in both leagues.
9.4.2.3 Salary cap

Consider the effects of introducing a salary cap to the sports league and
suppose that it succeeds in pushing win shares closer to equality than they
would otherwise be in a profit maximizing league. This would be the case if the
weaker team bought more talent than otherwise might have been the case
relative to the stronger team. Once again, the problem now arises that teams
collectively and individually could make more profit by redistributing talent
from the weak drawing sides (where it adds less to revenue). There are perpet-
ual incentives for weak drawing teams to spend less than the agreed minimum
on talent and for strong drawing ones to spend more, by the expedient of
trading players. It is clear under such circumstances that if the salary cap
can be enforced – and historically leagues have permitted exceptions – as Fort
and Quirk argue, competitive balance in profit maximizing leagues can be
affected by the adoption of a salary cap. However, their data on the NBA
suggested that the introduction of the salary cap had no significant effect on
standard deviations of win percents in basketball. They hypothesize that part
of the reason might have been relatively poor adherence to the salary cap.
9.4.2.4 Revenue sharing

The final case to consider is revenue sharing. A formerly widespread practice is
for the league to fix a proportion (a) of gate monies that the home side retains,
the balance (1 a) going to the visitors. The logic of what follows applies to
any direct revenue sharing, however, such as television revenue sharing or

total revenue sharing. The team's post-redistribution total revenues (R*), for
Teams 1 and 2 are given by Equations 9.6 and 9.7:
R*1 ¼ aR1 þ ð1 aÞR2 ð9:6Þ
R*2 ¼ aR2 þ ð1 aÞR2 ð9:7Þ
In competitive equilibrium the teams select win percents that maximize

profit, by setting MR equal to MC as set out in Equations 9.8 and 9.9. MR
is the change in total revenue that follows from a change in win percent,
i.e., qR1/qw1 >0. As changes in win percent for a given team affect those of
the other team, in a two-team league, but in the opposite direction, because
of the zero–sum property discussed earlier, then qR2/qw1 <0. This means
that the MR for Team 2 falls as the MR for team 1 rises, so:
MR*1 ¼ aMR2 ð1 aÞMR2 ¼ c ð9:8Þ
MR*2 ¼ aMR2 ð1 aÞMR2 ¼ c ð9:9Þ
Because of the negative sign on the second term in each equation, revenue
sharing reduces both teams' marginal revenues below what they would have
been otherwise, that is in the absence of the reallocation of playing talent.
Equilibrium with gate sharing occurs where the post-sharing MRs are equated
and league revenue is maximized, i.e., MR*1 = MR*2. Expanding the terms and
cancelling common items reveals that this is equivalent to MR1 = MR2. The
solution is shown in Figure 9.3, which shows, that each club's MR curve falls by
the same amount. Equilibrium in the presence of revenue sharing is at E* where
win percent is identical to that at E and the price of talent has fallen to c* below c.
An alternative form of revenue sharing is “pool” sharing, where teams
contribute to a central pool which the league makes a redistribution from.
This system is identical in its effects to gate sharing with N = 2, but not for
N > 2. For example, a 20 team no-draw league provides N(N 1) = 380
fixtures per season. Therefore, what Team 1 gets under a pool arrangement
depends on gate receipts from 380 matches. If the league adopts gate sharing,
the amount that Team 1 gets depends on receipts from the 38 games in which
it is directly involved. Unless all teams MR's fall by the same amount, which is
highly improbable, redistribution will affect the competitive balance. The
main conclusion that Fort and Quirk draw from their analysis is that, of
the policy instruments considered, only a salary cap can generally be
expected to assist the finances of weak drawing teams and simultaneously
FIGURE 9.3 Effect of revenue-sharing.
improve competitive balance. Reserve option clauses and reverse order

drafts affect players' wages and therefore act primarily to redistribute
incomes between clubs and players – they might be expected to improve
weak teams' finances, but not to improve competitive balance, suggesting
that competitive balance is invariant to changes in league policies. This is
known as the “invariance proposition,” which is an example of the Coase
theorem. It is shown that, because property rights can be established,
market forces seek an efficient (profit maximizing) solution to talent allo-
cation regardless of the ownership of the property rights.6 This idea was
noted in the seminal work of Rottenberg (1956).
6
It should be noted that the evidence on the effect of such policies is potentially problematic.
For example, Larsen et al (2006) report a significant impact on competitive balance of the
NFL's simultaneous introduction of free agency and a salary cap in 1993. The salary cap was
brought in to help teams survive the likely inflationary effect of free agency on players' wages.
As Larsen et al (2006) state, the simultaneous introduction of two measures made it
impossible for them to determine whether the effect was due to the salary cap alone
(consistent with prior expectations) or to free agency (inconsistent). Economists frequently
find themselves in similar situations – due essentially to the inability to conduct controlled
experiments (ones in which causal factors are varied one at a time). Similar comments apply
to the “before and after” studies reported by Fort and Quirk (1995), and others. One cannot
be sure that the only “significant” differences between the period before and after the
introduction of (say) free agency in baseball is the introduction of free agency; conceivably
some unobserved factor or factors may have changed, leading to erroneous interpretation.
9.4.3 Equilibrium in a win-maximizing league

As discussed in Chapter 7, Sloane (1971) argued from the perspective of
European football that clubs pursue utility maximization, i.e., mainly success
on the field, subject to a minimum satisfactory level of profit. In a large body
of work Stefan Kesenne (1996, 2004, 2007) has championed a particularly
tractable form of the utility maximizing model – win maximization subject to
a zero profit constraint – to determine how different owner's objectives affect
the results of cross-subsidization measures.
Market equilibrium in the win-maximizing league (subject to zero profit)
occurs where the clubs' ARs equal the marginal cost of talent c, shown as Ew in
Figure 9.4, compared to equilibrium in the profit maximizing league Ep. Since
AR always exceeds MR, as discussed in Chapter 1, it follows that talent is more
highly priced in the win-maximizing league. Assuming that big city Team 1
dominates Team 2, the distribution of talent is less equal under win maximi-
zation. However, if Team 1 is a highly successful small city team and Team 2 is
a comparatively less-well-run big city team, the small team will be more
dominant under win maximization than under profit maximization. In this
sense, win-maximization leagues can have closer competitive balance. This
consideration led Kesenne (2004) to distinguish between “good” and “bad”
over-domination; while one might be happy to redistribute revenue to correct
an imbalance that currently favours the large city team, one might be reluctant
to take an action that effectively punishes a dominant small city team for its
success.
FIGURE 9.4 Equilibrium in profit-maximizing (Ep) and win-maximizing (Ew) leagues.

Relaxation of Assumptions 249
Another implication of win-maximizing behaviour is that total league

revenue is not maximized. Figure 9.4 shows that at win-maximizing equilib-
rium Ew the teams' MRs are unequal, implying that if players transfer from
Team 1 the revenue it loses is less than the revenue gained by Team 2.
9.4.4 Cross-subsidization in a win-maximizing league

It has been established that reverse order drafts and reserve option clauses have
no effect on the balance of competition under profit maximization. It is now
shown that the same applies under win maximization. In Figure 9.4 a win-
maximizing league has win percent share Ew; if, say, a combination of draft and
option clauses force the cost of talent below its (win percent) equilibrium
value, teams can trade talent until Ew is attained, at which point the incentive
for further trade is gone. The effectiveness of these instruments is zero in win-
maximizing leagues, thus they are not considered further. There is little point
in further considering the effect of a salary cap for the opposite reason; it is
generally expected (although Vrooman, 1995, disagrees) to be effective if it can
be enforced. In the interests of brevity, therefore, the remaining discussion
focuses on the effectiveness of revenue redistribution.
It can be shown that, provided the redistribution benefits Team 2, its AR
function will rise and that of Team 1 will fall, but relatively less than Team 2's
rises, since Team 1 began with more revenue. Therefore, the distribution of
talent and success will shift in favour of Team 2 and the cost of talent will
increase. Other things being equal, revenue redistribution is expected to in-
fluence the balance of competition in win-maximizing leagues. This is an
important result, as it indicates that the invariance proposition need not apply
if owner objectives change. This means that cross-subsidization can, in prin-
ciple, affect competitive balance. This issue is now further addressed by relax-
ing some of the other assumptions of the model.
9.5 RELAXATION OF ASSUMPTIONS

9.5.1 Introducing the effect of spectator preference for
absolute quality
To begin the analysis take a quadratic revenue equation model from Kesenne
(2000a, 2004, 2006a) to produce the simplest equivalent model to the Fort and
Quirk case above. This is presented in Equation 9.10:
Ri ¼ mi wi bi w2i ð9:10Þ
The function implies that, as win percents increase, revenues increase
linearly according to the first term on the right-hand side, and decrease
non-linearly according to the second term on the right-hand side, assuming

b >0, at an increasing rate. In fact the combination of these terms produces an
inverted curve, in which revenue achieves a maximum and then declines. This
is an indication of spectators' aversion to the increasing certainty of outcome.
In this model the corresponding cost function is Equation 9.5.
The value of Team 1's (and by implication Team 2's) win percent can be
derived, assuming that both owners pursue win maximization. The derivation
under the alternative assumption of profit maximizing follows in Appendix
9.2. Both solutions are presented in Kesenne (2006). The teams' AR functions
are given as Equations 9.11 and 9.12. Total revenue is divided by w to arrive at
AR and at equilibrium, AR equals MC, c (which is consequently higher than
under profit maximization).
Why would marginal cost be greater under win maximization compared to profit
maximization? Hint: Consider how the win-maximizing club decides on its win percent
A win-maximizing team will hire as much talent as it can to generate

wins, provided it does not produce an economic loss. This will be indicated
where the ARs and not MRs are equal to MC, which are equivalent to AC if
MCs are constant. As ARs are always above MRs, this implies that a talent
cost greater than under profit maximization will occur, as well as a greater
degree of competitive imbalance. AR is then given as:
AR1 ¼ m1 b1 w1 ¼ c ð9:11Þ
AR2 ¼ m2 b2 w2 ¼ c ð9:12Þ
Equating and rearranging leaves the solution for Team 1's equilibrium
share of talent (equivalently its win share) as Equation 9.13 where, to simplify
notation, Sb is used to denote the sum of the b parameters:
X X
w1 ¼ fðm1 m2 Þ= bg þ b2 = b ð9:13Þ
If Team 1 has the larger market and if both teams' spectators have the same
degree of taste for UO, then Team 1 has the larger share of the talent, because
the first term on the right is positive and the second equals 0.5 whenever the b
parameters are equal. If Team 1's spectators are exceedingly more averse to
sporting certainty than Team 2, the small city team may dominate the league,
consistent with the message of Fort and Quirk (2004). Kesenne gives the
condition for this “good” imbalance where, even though m1>m2, it is also true
that w2>w1, i.e., it is required for the smaller team (Team 2) to have greater
win percents as:
b1 b2 > 2ðm1 m2 Þ ð9:14Þ

which means that for “good” imbalance to appear in this league the large
drawing team supporters' extra aversion to certainty of outcome must double
its team's advantage in market size. The comparable condition for a “good”
imbalance in a profit-maximizing league is as derived in Appendix 9.2:
ðb1 b2 Þ > ðm1 m2 Þ ð9:15Þ
This suggests that the small market team is more likely (but still relatively
unlikely) to dominate the league if team owners are profit maximizers. Play-
ing quality can be added to the analysis, following Kesenne (2000b, 2006a),
Vrooman (1995) and others in adding the assumption that teams' revenues
depend also on the visitors' playing strength. Revenue now becomes:
Ri ¼ mi wi bi w2i þ ei wj ð9:16Þ
Here “e” is a parameter that converts the visiting teams' success, wj (indi-
rectly via its travelling support), into gate revenue for the home club. It would
be expected that “e” is positive, although not identical across teams, i.e., the
greater the drawing power of the home side the greater is ei (in this context e1
>e2).
Kesenne (2001) found that adding a taste for absolute quality to the basic
model, revenue redistribution can affect competitive balance even under profit
maximization. If neither club's supporters value absolute quality both values
of “e” are zero; if both sets have the same taste for quality then the values of “e”
are equal and non-zero. In either case, Team 1's win percent depends on
market size and the taste for UO, as in the basic model. When the values of
“e” are equal, Team 1 has the same win percent as in the basic model under
profit maximizing, but a smaller share than under win maximization.
9.5.2 Market size diseconomies

Vrooman (1995) suggests that the advantages in sports leagues conferred by
market size might be offset by either the declining marginal productivity of
talent and/or by size diseconomies. These assumptions imply that, for exam-
ple, a given player may be more valuable to the weaker team since its win rate is
lower than the stronger team. Recall that diminishing marginal productivity
suggests that more employment of a variable factor of production leads to an
increase in output, but at a declining rate. Consequently, teams with more

talent can employ that talent at the margin in a less productive manner than a
team with less overall talent. This might apply across all resources, suggesting
that larger teams are relatively less productive in producing wins than smaller
teams. This may result in less competitive imbalance than the basic model
suggests and it also suggests that revenue redistribution can affect competitive
balance. Figure 9.5 illustrates the equilibrium.
Team 1 maximizes profit at “A” paying a premium price “cp” for every unit
of talent hired. Team 2 maximizes its profit at “B” paying a discount price “cd.”
Industry revenue would be maximized at “E,” where both teams pay the
competitive equilibrium price “c” for their talent, but E is not attainable given
the imperfect market for talent that results in a dual price. Profit-maximizing
equilibrium under market diseconomies is more balanced than at E. Talent
hired by Team 1 is paid more than in equilibrium and that of Team 2 is paid
less. It should be emphasized that this is a very naı̈ve model of a market in
which different prices are paid for units of identically productive talent.
Revenue sharing causes both teams' MR curves to fall, but unlike the
standard model the actual win percents do change in Team 1's favour –
why? An equal fall in both teams' MR curves lowers both the competitive
equilibrium price of talent c and MCs, the latter equiproportionally. But, as
Team 1 has a higher level of MC at w1, the absolute fall in its MC exceeds that
of Team 2's. Thus, the same w1, w2 cannot be a post-sharing win percent
combination for profit-maximizing teams. At the new profit-maximizing
FIGURE 9.5 Market size diseconomies (Team 1 maximizes profit at A, wages Cp, Team 2
maximizes profit at B, wages Cd).
win combination both teams' MCs must have fallen by the same absolute
amount, equal to the reduction in their MRs. To reach this new win percent
combination, Team 1 must start buying extra talent to raise its MC slightly
while Team 2 sells talent to lower its MC. At the end of this process both teams'
MCs will have fallen (relative to pre-sharing) by the same amount, equal to the
common reduction in their MRs. Each team will have restored equality be-
tween its MC and its MR and will be once again maximizing its profit at the
new win percent combination. In the course of moving to the new equilibrium
win percent the differential between “cp” and “cd” is squeezed. Declining
marginal productivity is a factor that might amplify the results of market size
disadvantage, but to pursue that line of inquiry demands a “standard” model
not based on constant returns.
9.5.3 A variable supply of talent

So far the analysis has maintained that teams compete to hire a fixed amount
of talent at a constant MC, c, in a closed market. Clearly, if the supply of talent
is variable and/or it is no longer assumed that the entire stock is traded, the
restriction that dt2/dt1 = 1 no longer applies and the choice of win percent is
now distinguishable from the choice of talent; in other words qwi/qti „ 1. A
team chooses to buy a number of talents, but it is unable to select any partic-
ular percentage of the total; its actual share of talent is known only after the
whole industry has made its purchases. In this case the concept of MC relevant
to the profit-maximizing team is not qC/qw, the MC of a win, but it is qC/qt,
the MC of a unit of talent. European sports teams operate in open markets for
talent, they are free to buy from clubs inside and outside of their own national
leagues, and from other continents. To a lesser extent, there is an international
market for players in US professional team sports.
To relax these restrictions Szymanski (2004) and Kesenne (2006b, 2007a)
model leagues using game theory to take decision makers' interdependence
into account, and argues that this is more appropriate than a Walrasian (com-
petitive equilibrium) theory as a descriptor of behaviour when there are few
participants, in other words an oligopoly market as discussed in Chapters 5
and 7. Szymanksi and Kesenne (2004) proved that in a Cournot–Nash profit-
maximizing league revenue sharing worsens competitive balance – in effect
the sharing of revenue goes some way to internalizing the negative external-
ities of teams' decisions, so the league moves closer to joint profit maximiza-
tion. As the number of teams increase the outcomes of Cournot–Nash models
get closer to competitive models; the misallocation of talent in a 20-team
Cournot–Nash league is likely to be much less marked than in the two-team
league.
9.6 WHEN DOES THE “INVARIANCE PROPOSITION” HOLD?

The above discussion has attempted to offer an overview of some of the main
issues from a complex literature in which, with considerable ingenuity,
economists have attempted to explore sports leagues. However, an important
question remains, which is under what conditions may we assume that com-
petitive balance is invariant to the application of (or changes in) cross-subsi-
dization policies?
A summary might be as follows. First it appears that reserve clauses (trans-
fer systems) and rookie drafts affect the distribution of value added between
players and owners, but do not affect competitive balance, whatever the goals
of owners, i.e., profit or win maximization or whatever the preferences of fans.
Secondly, it appears that, with some exceptions (for example, see Vrooman
1995), salary caps are the one cross-subsidization policy that most writers
believe can affect competitive balance, if it can be enforced. This leaves revenue
sharing as the one method of cross-subsidization worthy of further discussion,
as its ability to affect competitive balance does appear to depend on factors
such as owner objectives and spectator preferences, the number of teams in the
league, whether teams behave like oligopolists, and the particular form of
cross-subsidization. In most of the models reviewed it appears that revenue
sharing in win-maximizing leagues benefits the smaller teams relatively more
than it harms the dominant ones and thus promotes more competitive bal-
ance while simultaneously raising the price of talent. Table 9.1, therefore,
summarizes some findings on the effectiveness of revenue sharing as a means
of influencing competitive balance in profit-maximizing leagues about which
debate remains.
Table 9.1 is organized to illustrate the outcomes under four model specifi-
cations. A is Kesenne's quadratic version of the (Fort and Quirk) standard
model. B differs from A by introducing fans' preferences for absolute quality.
C departs from A by bringing in market size diseconomies. D differs from A by
abandoning the notion of a fixed supply of talent, all of which is hired, and by
treating clubs as oligopolists who compete via quality variations. The fourth
column shows exceptions to the general rule, and gives references.
9.7 HOW FAR DOES COMPETITIVE BALANCE MATTER?

THE EUROPEAN CHAMPIONS' AND OTHER LEAGUES
An implication of the above research is that only salary caps and revenue
sharing may possibly affect competitive balance. Kesenne (2007b) is pessimis-
tic about the prospects for a return to revenue sharing, which has been allowed
to wither in, for example, European soccer leagues. Given the international
How Far Does Competitive Balance Matter? The European Champions' and Other Leagues 255
TABLE 9.1 Revenue sharing in leagues with profit-maximizing teams
Model specification Effect on Effect on win Qualifying remarks

talent price percent with respect to effects
on win percent
A. Standard model Lowers None Requires all revenues

equations and qwi/ shared and by same
qti = 1 method (Fort and Quirk,
1995)
Gate and pool sharing
may give different
outcomes. If MRs do not
shift identically may
have an effect (Kesenne,
2000a)
Requires N>2 for an
effect on win percentage
B. Standard model Lowers Improves If all e's zero or all
plus taste for identical – very unlikely
absolute quality indeed – no effect.
and qwi/qti = 1 (Kesenne, 2000)
No effect under pool
sharing (Kesenne, 2001)
Requires N>2 for there
to be any effect,
irrespective of the e's
(Kesenne, 2000)
C. Standard model Lowers the No effect on the Operates irrespective of
plus market size revenue revenue maximizing whether MR's fall by the
diseconomies and maximizing win percent same amount so works
qwi/qti = 1 equilibrium even when N = 2
price c (Vrooman, 1995)
Reduces the Worsens balance,
spread shifting it towards
between the the revenue
unit prices maximizing value
of star and where all MR's are
ordinary equalized
talent
D. Cournot–Nash Lowers Worsens balance, Differs from A above in
and qwi/qti „ 1 shifting it towards that qwi/qti „ 1,
the revenue permitting the stock of
maximizing point talent to vary or some to
where all MR's are be unemployed
equalized (Szymanski, 2004)
nature of the single European and broader market in soccer talent, and the
extent and economic importance of supra-national competition, it is hard to
see how any league acting on its own could resume more revenue sharing.
Revenue sharing works for clubs in part through reducing the prices teams pay
for talent; any single European league seeking financial salvation through
revenue sharing would quickly find its leading members unable to attract
enough talent to compete at supra-national level. Kesenne (2007b) notes that
football clubs in European leagues are unlikely to give serious consideration to
national revenue sharing schemes in view of a perceived need to compete with
clubs from Spain, Italy, Germany, England and France, the leagues that have
historically produced most of the supra-national title wins. The same impli-
cation would follow from the imposition of a salary cap in a single league.
Likewise, it seems difficult to conceive of the effectiveness of salary caps in a
sport with a truly international labour market. In other professional sports,
such as rugby, this is possible as evidenced by their implementation.
The only realistic chance for revenue sharing or salary caps to be introduced
into European soccer is for all leagues to act in concert – most probably under
pressure from UEFA. In this context the important difference between Euro-
pean and North American professional sports leagues may not be so much that
the latter have no equivalent to the formers' supra-national competition, but
rather that the latter have fewer leagues, making agreement on cross-subsidi-
zation schemes easier to reach, even in the presence of what is a de facto supra-
league as noted by Fort (2000) through the existence of the World Series and
the Super Bowl, which may exert similar effects on competitive balance in
American leagues and conferences.
It can also be hypothesized that outcomes in the Champions' League
and the UEFA Cup have recently become more predictable, thanks to the
partial replacement of the two-game knockout format of tournament by
multi-stage semi-league formats and by the admission of larger numbers of
competitors, which implies that a team must play more games than for-
merly to reach any particular stage or round. As discussed in Chapter 7,
increasing the minimum number of games required to attain any given
stage tends to enhance the prospects of the most successful teams. Other
things being equal, the joint probability of multiple wins is lower than the
probability of a single outcome against any one of the opponents and this
will make it much less likely that teams with lower probabilities of wins
are successful. In general, the more matches there are per round and the
more rounds there are to any competition the greater the probability that
favourites will reach the final stages. This makes intuitive sense, in that
teams with a larger playing staff, through growing attrition, enhance the
probability of their success.
Conclusion 257
Simultaneously, of course, such competitions produce a greater number of

hours of live sport that can be sold. This perspective may help to explain why
league titles in team sports tend to be more concentrated and more valued than
knockout cup wins and why in an era when television exercises a strong
influence on sports organizations fewer competitions continue to use the pure
knockout format. It remains, however, to note that while sports economists
are agreed that competitive balance is theoretically important to the survival
and development of leagues, they are by no means agreed on how important it
is. As Sanderson (2002) argues, while leagues impose measures on their
members that recognize mutual interdependence they simultaneously pursue
others that tend to exacerbate competitive imbalance. The evolution of the
European Champions' League from the two-game knockout format involving
only national league champions into a fully blown league format (and more
entrants) is a case in point. As discussed in Chapter 8 one form of UO, deriving
from Jennett (1984), argues that sports leagues are populated by supporters
whose main object is to see their own teams win. Sports leagues have survived
and expanded while displaying plenty of imbalances. This suggests that com-
petitive balance is only one of many factors that determine their long-run
health and may, in fact, be an outgrowth of their evolution rather than neces-
sarily what causes them to change.
9.8 CONCLUSION
Chapters 7 and 8 have argued that there is a need to coordinate clubs' activities
to ensure that UO or, as indicated in Chapter 8, competitive balance is pre-
served. Because it is argued that spectators prefer closer contests, the public's
welfare may, therefore, be enhanced and the potential failure of leagues averted
if stronger teams cross-subsidize weaker teams. In this chapter the main forms
of cross-subsidization have been outlined, together with an exploration of
their impact in leagues characterized by clubs maximizing profits and maxi-
mizing win percents. These respective objectives, it is argued, apply to US and
European leagues respectively. It is shown that under conditions of profit
maximization when the supply of talent is fixed, in general a version of the
Coase theorem applies, making competitive balance invariant to any policy. It
is argued that salary caps may affect competitive balance, but may fail because
of enforcement problems. The chapter then argues that changes in the
assumptions connected with modelling sport leagues produce challenges to
the invariance proposition. These assumptions include shifting owner objec-
tives to win maximization, having spectators caring about absolute playing
quality, that the scale of markets may produce diseconomies of scale from
playing talent, and that the supply of talent might be variable because of open,
international markets. The chapter concludes that, under such circum-

stances, any potential for a corrective impact of cross-subsidization of teams
appears to lie with revenue sharing in a win maximizing league, but the ability
of domestic leagues to implement this remains dubious given the need to
coordinate across countries. Attention now turns towards an empirical anal-
ysis of the demand for professional team sports.
Appendix 9.1 Why Win Percent and Talent Distribution are not
FormallyEquivalent
The logic of cross-subsidization policies is to try to produce competitive balance by ensuring
a more equitable distribution of talent. This presupposes that production functions for wins
“W” are as described in Equation A9.1.1, a function of labour “L” for team “i”:
Wi ¼ Wi ðLi Þ ðA9:1:1Þ
If playing talent “t” is a function of labour, i.e.:
ti ¼ ti ðLi Þ ðA9:1:2Þ
this implies that:
Wi ¼ Wi ðti Þ ðA9:1:3Þ
or, if wins are proportional to talent:
Wi ¼ ti ðA9:1:4Þ
Win percents “w” will then become:
wi ¼ Wi =Wi þ Wj ¼ ti =ti þ tj ðA9:1:5Þ
which is a “contest success function,” as described in Chapter 7. Competitive balance will

then be given by:
wi =wj ¼ ti =tt ðA9:1:6Þ
In which the relative supply of talent determines competitive balance. The use of talent rather
than labour enables the consideration of homogenous talent units, as opposed to players that
might differ in skills. However, the results in Equations A9.1.5 and A9.1.6 are potentially
false. Consider a team that hires X% of the talent; obviously the number X is drawn from the
large set of rational numbers.
Rational numbers may be expressed as ratios of integers e.g., 4/5 and 193/26079, integers
being “whole” numbers. Numbers, such as p and the square root of 2, belong to the set of
irrational numbers; i.e., numbers that cannot be expressed as ratios of integers (whole
numbers).
Games, points and wins are measured in integers, based on the number of teams in the
league N – usually less than 30. In an N team balanced league where teams play each other
twice a season the number of games each club plays is equal to 2(N 1). Therefore, win
Conclusion 259
percents are restricted to a very small subset of the rational numbers between zero and one,
implying that it is impossible to hire X% of the talent and enjoy exactly X% of the wins for every
possible value of X.
Appendix 9.2 The Quadratic Model Under Profit Maximization

The basic quadratic model under profit maximization is:
Ri ¼ mi wi bi w2i ðA9:2:1Þ
Ci ¼ cti þ k ðA9:2:2Þ
or:
Ci ¼ cti ðA9:2:3Þ
if k = 0.
Profit maximization requires the teams to equate MR and MC leaving:
m1 2b1 w1 ¼ c ðA9:2:4Þ
m2 2b2 w2 ¼ c ðA9:2:5Þ
Using the fact that win per cents sum to unity, one can substitute (1 w1) for w2, set the
equations equal and obtain:
m2 2b2 w2 ¼ c ðA9:2:5Þ
m1 2b1 w1 ¼ m2 2b2 ð1 w1 Þ ðA9:2:6Þ
and, on rearranging, arrive at:

X X
w1 ¼ fðm1 m2 Þ=ð2 bÞg þ b2 = b ðA9:2:7Þ
As one would anticipate, the only difference between Equation A9.2.7 and the equivalent
with spectator preferences for absolute quality, is that the former has no terms “e” to
represent consumer preference for better quality of play. Comparing A9.2.7 with the
expression for the value of w1 in a win maximizing league, as presented in the text, it is clear
that the denominator is larger under profit maximization, suggesting that other things being
equal, competitive imbalance is worse under win maximization.
CHAPTER 10
The Demand for Professional

Team Sports: Attendance and
Broadcasting
OBJECTIVES
& To understand how the economic theory of demand is applied to attendance

at professional sports matches
& To appreciate the evidence on the determinants of attendance demand
& To understand how broadcasting affects attendance demand and the
determinants of broadcast demand
& To appreciate the role that broadcasting has had on the supply structure of
professional sport
10.1 INTRODUCTION
In this chapter the demand for professional team sports is investigated in some
detail, beginning in the next section with a brief resume of the theoretical
framework presented in Chapter 3. Some theoretical and empirical extensions
are then discussed in Section 10.3, to indicate how the theory requires refine-
ment to apply to professional sports. Section 10.4 reviews evidence of the
demand for professional sports, providing a selective summary of recent stud-
ies. The impact of television broadcasting of sport on attendance demand is
discussed as part of this review and Section 10.5 then explores the sources and
growth of broadcast sport in more detail. Initial research exploring the broad-
cast demand for sport is reviewed, as well as the implications of funding from
the growth of this sector. It is argued that this growth has been primarily driven
by supply-side changes in the media market and that the funding arrange-
ments involved have helped to reshape the supply structure of professional
sports leagues, enabling the league evolution discussed in Chapter 7. 261
262 CHAPTER 10 The Demand for Professional Team Sports: Attendance and Broadcasting
10.2 THE DEMAND FOR PROFESSIONAL TEAM SPORTS

Chapter 3 presented the “derived” demand for sports goods and equipment as a
result of a primary or parent demand for sports participation. The same theo-
retical framework used to examine the demand for goods and equipment can be
applied to the decision of utility-maximizing sports fans, as consumers, to
demand professional team sports. The basic conceptual relationship and qual-
itative set of predictions are illustrated in Figure 10.1, where a demand curve for
match tickets is drawn. Changes in ticket price, other things being equal, will
cause demand to increase or decrease by movements along the curve, whereas
changes in income or preferences or the price of other goods will, for each
change other things being equal, cause the demand curve to shift to the left
or right, respectively, as less or more tickets are demanded at any given price.
10.3 THEORETICAL AND EMPIRICAL ISSUES

While the above analysis suggests that application of the standard economic
theory of demand to professional team sports appears to be relatively straight-
forward, it should be noted that extensions to the theory and empirical com-
plications of measurement and analysis need to be considered.
10.3.1 Measuring quantity

The above analysis, and indeed that of Chapter 1, suggests that the demand for
sports is the match or event whose price is the ticket price. Yet, as Chapter 7
indicates, the core product of sport is really a
contest, a meeting of competitors in a sports
arena allied to many multifaceted characteristics
concerning the timing and significance of the
contest, such as the arena in which it takes place
and its location, as well as the opponents and
the historical and organizational elements of the
contest (see also Borland and MacDonald,
2003). In this respect the researcher has to make
some assumptions about the nature of demand.
One of the first and established assumptions
is that gate paying attendance is the proxy for the
demand for the contest as a live event. While this
seems to be reasonable, it carries with it certain
FIGURE 10.1 A theoretical demand curve for match presumptions. For example, counting the aggre-
tickets. gate number of seats at a match as a measure of
Theoretical and Empirical Issues 263
demand presupposes that the experiences of the

season ticket holder and the corporate client enter-
tained in the firm's own “box,” and the casual spec-
tator buying a “one-off” ticket are all the same.
A further complication is that now, in many
sports, events such as the Champions' League Final,
the Superbowl and the World Series or games be-
tween local or international rivals (for example
Celtic and Rangers, Manchester United and Man-
chester City or AC Milan and Inter Milan), can
generate so much demand that stadiums are sold
out. Indeed regular fixtures in the English PL are
now mostly played in front of full capacities. Under FIGURE 10.2 Actual and latent spectator demand.
such circumstances attendance for such matches
measures supply (as constrained demand) and not
demand as the free choices of fans. This is illustrated in Figure 10.2 where, for
the lower ticket price, “demand” is indicated by supply capacity.
A related question is how to treat the demand for sport on television. It is
possible to add live gate paying spectators to television viewers to get a total
match audience. One justification for this approach might be that the demands
are “complementary” inasmuch as given the opportunity, the spectator would
rather attend the live match. This suggests the prediction that screening live
sport will at worst leave attendances unaffected, but at best may increase them
by galvanizing interest in the sport. An alternative perspective would be to view
television broadcasts as a substitute good that offers spectators the opportunity
to watch live sport in settings that are more amenable. From this point of view
one would expect live broadcasting to reduce attendances.
10.3.2 Measuring price and income

It might be reasonably assumed that the price of a stadium ticket is readily
apparent. However, the price of attendance, i.e., the full opportunity cost of the
match, is not. There will also be other costs, such as travel costs. Furthermore,
the effects on demand of changes in substitute goods should be accounted for
in price effects. Intuitively, however, it seems likely that “cross price” effects on
demand are likely to be non-existent for fans of a particular team, or even
sport. If their own team's ticket prices rise fans are much more likely to
substitute other activities for watching their particular sport rather than
switch to another team in the same sport.
Finally, it is clear that the real personal disposable income of fans should be
the measure of income used in analysis. In the absence of primary research
determining this information, it follows that any assumed proxy, such as

industrial earnings in the region will tend to normalize earnings, rather than
reveal their genuine variance and hence their potential impact on attendance.
10.3.3 Consumer preferences

The basic theory of consumer demand suggests that preferences could be
treated as constant or fixed, in order to derive standard predictions. However,
in the context of professional sports demand it seems clear that spectators'
preferences may be affected by a number of factors. These include the quality of
the experience associated with the sports contest, as noted above, and the fact
that fans preferences, almost by definition, may be shaped by others in the
community or by past experiences of watching their team. These last two
factors are directly analogous to and further examples of, the social and con-
sumption capital that has been argued to play a role in participation demand.
In this respect, fans may exhibit interdependent preferences and habit persis-
tence. Finally, there may be a spatial context to preferences. This suggests that
attendance demand will be influenced by the geographical area for two rea-
sons. On the one hand, it is likely that loyalty and habit persistence are likely
to emerge from a local environment and historical connections with a club, as
discussed in Chapter 7. On the other hand, the local geographic area approx-
imates the scale of potential demand.
10.3.4 Estimation issues

As discussed in Chapter 1, the typical strategy employed in economic research
is to estimate a regression equation to identify the impact of the various
determinants of behaviour on the choices of individuals or firms. Stylized
examples of the forms of regression models used in professional team sport
literature to explain attendance demand are given in Equations 10.1 to 10.3.
These represent a cross-sectional regression equation, a time series regression
equation and a pooled/panel data regression equation, respectively.
Ai ¼ b1 þ b2 Pi þ b3 Yi þ b4 UOi þ b5 Di þ b6 MSi þ i ð10:1Þ
At ¼ b1 þ b2 Pt þ b3 Yt þ b4 UOt þ b5 Dt þ b6 MSt þ b7 At1 þ t ð10:2Þ
Ait ¼ b1 þ b2 Pit þ b3 Yit þ b4 UOit þ b5 Dit þ b6 MSit þ b7 Ait1 þ vit ð10:3Þ
In these equations, “A” represents the dependent variable, which could be

measured as the numbers of spectators, for example in thousands, at games for
Theoretical and Empirical Issues 265
a club. In accordance with economic theory this variable is assumed to depend

on the following causal or independent variables. “P” represents admission
price, measured in a particular currency, for example pounds sterling, US
dollars or Euros. “Y” represents the income of supporters, measured in the
same currency, UO represents uncertainty of outcome that could be measured
in a variety of ways, as discussed in Chapter 8, and “D” refers to a dummy
variable that could measure any quality characteristic that could affect the
game, such as whether the match was a derby match, whether it was raining or
not, whether the match involved the league champions or not and so on. In
practice, a large number of such variables are employed. MS refers to the
market size or local population in some spatial dimension around the
location of the match; e, y and v represent other factors that randomly
cause attendance to vary; and i and t represent indices of observation, with
the former referring to cases, such as clubs at a point in time, and the latter a
particular case, for example, a club or set of clubs, over time. Where both
appear, this indicates that observations are on a set of cases, such as clubs,
over time.
In Equations 10.2 and 10.3 the lagged value of attendance is included as an
independent variable. This is often used by researchers to measure “habit
persistence” or the loyalty of support, which is a proportion of last period's
attendance directly affecting this period's attendance. The degree of habit
persistence would be given by the estimate of b7 from the data. More generally,
the estimate of b2 would indicate the movement along the demand curve, that
is change in attendance, following a unit change in the ticket price, with
estimates of the other coefficients b3 . . . b5 indicating how much the demand
curve shifts, i.e., attendance changes, following a unit change in income, UO
or if the qualitative characteristic is applied or not.
10.3.4.1 Simultaneity
There are some potential problems associated with the econometric strategy
noted above. To begin with, Equations 10.1 to 10.3 are examples of reduced
form equations, measuring attendance as the joint outcome of a set of struc-
tural equations describing different aspects of behaviour. Box 10.1 provides an
example of this in a simple examination of attendances, and Appendix 10.1
provides an example illustrating why it is problematic to assume that a lagged
attendance variable, as in Equations 10.2 and 10.3, measures habit persis-
tence. In a strict sense it can be shown that the presence of the lagged atten-
dance variable can simply be viewed as providing a distinction between the
short run or impact effect of a change in one of the (other genuine) independent
variables on attendance, as given by b3 . . . b6, from the long run effect which is
given by b3 . . . b6 divided by 1 b7.
BOX 10.1 THE IDENTIFICATION PROBLEM

Consider a sports league in which the planned demand for To identify the separate demand and supply (attendance)
seats “Ad” depends inversely on ticket prices, “P” as functions would require specifying additional variables that
AD = a bP, and the planned supply of seats, “As” only appear in the opposing equation. Additional data on
depends directly on ticket prices as AS = c + dP, i.e., ignor- supply would therefore help to identify demand and vice
ing capacity constraints. When demand equals supply in versa. The intuition is that shifts in one of the functions help
equilibrium, i.e., AD = AS, this means that a bP = c + dP, to trace out the other, as illustrated below in identifying the
or that the equilibrium ticket price of a seat is P = (a c)/(d demand curve. A more detailed example of this issue is
+ b). Substituting this price into either AD or AD to identify presented in Coates and Humphreys (2005).
equilibrium attendance would yield A = (da+bc)/(d+b).
This shows that any set of observations on ticket prices and
attendances, which must represent market equilibrium
(where demand and supply cross) or the actual purchases
and sales of seats, is a set of values representing the com-
bination of the demand and supply coefficients and thus not
what a fan alone might be prepared to pay for a ticket, as
theorized in the demand curve. In this simple model, there-
fore, price and attendance are simultaneously determined
and for any observed change in price and attendance one
could not identify whether this was due to demand or supply
separately. A regression of these equilibrium attendances
on these equilibrium prices cannot be presumed to repre-
sent an estimate of attendance demand!
Estimating a single equation that is actually part of a set of equations also

raises statistical problems. An estimated coefficient that is derived from an
equation in which the independent variable is actually determined from an-
other equation will suffer from simultaneous equation bias. As a consequence,
a number of general strategies are open to the researcher to avoid simultaneous
equation bias. They are beyond the scope of this book, but involve replacing
the offending variable (often using its predicted value that will not contain a
random element, unlike the original variable) with an appropriate alternative.
This is referred to as instrumental variable estimation or two-stage least
squares (2SLS) estimation. Finally, the researcher can use (multivariate) esti-
mation techniques that allow for more than one dependent variable in a set of
equations. Examples would be cointegration analysis, used in time series
analyses to explore the possibility of a long-run equilibrium relationship be-
tween variables. An important feature of this analysis is that it directly
addresses the problem of spurious regression occurring, such that apparent
relationships between variables arise because of common random trends in
the data. Another form of system modelling is to use seemingly unrelated
Evidence 267
regression (SUR), that attempts to allow for common random influences

across a set of equations.
10.3.4.2 Multicollinearity
As well as simultaneous equation bias, the possibility of interaction between
the regressors may raise problems of multicollinearity. This would, potential-
ly, make it difficult to identify the impact of specific variables on attendance, as
multicollinearity suggests that there is not really a problem of simultaneity of
the variables but rather that the regressors are measuring different dimensions
of the same influence on demand. In this context, it might well be that market
size is closely correlated with the purchase of quality players and also with
playing success and UO, as it provides a key resource base for clubs. It might be
the case that price and travel costs are closely related or income and unem-
ployment. A number of strategies are open to researchers to deal with these
issues, such as combining the variables using factor analysis to identify latent
variables representing combinations of the set of original variables. Tests of the
joint significance of sets of variables could also be employed.1
10.3.4.3 Heteroscedasticity and serial/autocorrelation

Finally, heteroscedasticity could arise because of the potential for variability in
the behaviour of fans from different population segments. A general form of
adjusting the standard errors of estimates to produce more robust, i.e., con-
sistent, t-ratios on which to base statistical inferences, such as White's adjust-
ed standard errors, have now become a standard feature in econometric work.
Similarly, generalized least squares procedures and other approaches have
become common for estimating equations rather than adjusting the standard
errors. The same applies to treating serial correlation which occurs when the
random errors affecting attendance are linked over time.
10.4 EVIDENCE
Historically, the scope and scale of research has been relatively limited, reflect-
ing the novelty of the economic analysis of sports. For example, Borland and
MacDonald (2003) reviewed 57 studies since 1974 covering the following
sports (and number): soccer (20); cricket (2); rugby league (7); Major League
Baseball (13); Australian Rules football (4); National Football League (6);
hockey (2); National Basketball Association (2); and all four US major league
sports (excluding Major League Soccer) (1). It is noted that 23 of the studies are
1
This would use the F-distribution in OLS or equivalent with other estimators, as discussed
in Chapter 1.
single season analyses, 15 cover between 2 and 6 seasons, 5 cover between 7

and 10 seasons, 3 cover between 11 and 25 seasons, and 11 studies cover
between 26 and 98 seasons. To update this work, an additional 22 studies have
been reviewed for this chapter. These cover Spanish soccer (2), German soccer
(2), English soccer (4), Belgian soccer (2), French soccer (1), Korean baseball (1),
Super 12 rugby union (1), Major League Baseball (5), Minor and Junior League
Hockey (1), Major League Baseball, National Basketball Association and Na-
tional Football League combined (2) and National Basketball Association (1).
Of these studies, 6 cover 1 season only, 6 cover between 2 and 6 seasons, 2
between 7 and 10 seasons, 1 between 11 and 25 seasons and 7 between 26 and
53 seasons. Table 10.1 provides some details.
There is some evidence, therefore, of a broadening scope of study, but
soccer and baseball still dominate the literature. Furthermore, the typical
study is a pooled cross-section–time series analysis of less than 10 seasons,
with a substantial number being only one season. However, there is evidence
now of a growing body of longitudinal analysis, with 18 studies covering
periods in excess of 26 seasons.
10.4.1 Measuring quantity

As discussed earlier, the main variable used to measure the quantity of demand
for professional team sports is gate attendance. However, the quality and detail
of data varies. Most studies examine sets of individual clubs' match atten-
dances within and across a number of seasons. Recent examples are Owen
and Weatherston (2004), Forrest, Simmons and Szymanski (2004), Buraimo,
Forrest and Simmons (2006) and Meehan et al. (2007). As the study period
increases, aggregate league attendance tends to be examined, for example, in
the more recent studies by Winfree et al. (2004), Coates and Harrison (2005),
Lee (2006), or seasonal club attendances are used across longer time periods,
such as in Feddersen et al. (2006) and Feddersen and Maennig (2007). Earlier
studies that pioneered this approach are Simmons (1996) and Dobson and
Goddard (2001).
Simmons (1996) is also significant in that it was one of the first studies to
disaggregate attendance demand by considering the difference in behaviour of
season ticket and casual spectators, thereby recognizing that attendance is not
necessarily homogenous. Garcia and Rodriguez (2002) concentrate on non-
season ticket holders and Falter et al (2008) distinguish between season and
non-season ticket holders. A number of other studies attempt to control for
different types of ticket holders implicitly by, as noted further below, making
use of weighted ticket prices being paid by different sets of spectators; such an
example is the study by Lee (2006). The issue of capacity constraints is now
TABLE 10.1 A selection of recent econometric findings on team sport demand
Author Context, sample size Model, method and estimator Significant variables and comments
Butler (2002) MLB 1999, 2428 games Aims to determine effect of Weekday dummies (+) (Monday is
interleague play on attendance default)
MLB seasons occur within a OLS Away teams starting pitcher's wins
single calendar year minus losses (+)
(inferred) Hot (); bad weather ()
Dependent variable Closing stadium (+)
Log attendance Home team games behind leader ()
Log away team payroll (+)
Inter-league game (+)
Inter-league games few in number
but very attractive
Weekday dummies (+) (Monday is
default)
Forrest, Simmons Premier League soccer 1995– Aims to measure sensitivity of General travel cost (including time)
and Feehan (2002) 1996. attendance (by club) to travel ()
cost
Fan survey by Norman Chester Weighted OLS Proportion zone population within
Centre socio-economic classes 1–2 (+)
20 479 questionnaires Dependent variable Proportion of zone population who
returned out of 58 000 attend 3 or more away games (+)
Log of per capita local ticket Fixed effects - 4 clubs (+)
sales
Evidence
269
270
CHAPTER 10 The Demand for Professional Team Sports: Attendance and Broadcasting
Garcia and Spanish First Division soccer Aims to explain match OLS
Rodriguez (2002) 1992–1993 to 1995–1996 attendance
1580 games OLS (Column 1) Log real ticket price ()
Dependent variable Log real income (+) (inelastic)
log match attendance Weekday dummy ()
IV log real ticket price is No rain (+)
instrumented (Column 6)
Dependent variable Distance ()
log match attendance UO (various) (+)
IV preferred as it does not Team budget (+) home, away
ignore the endogeneity of the
ticket price
On public television ()
IV
As above but income no longer
significant
Allan (2004) Aston Villa, home league Aims to explain home Distance () and squared (+)
games 1995–1996 to 2000– attendance
2001
114 games OLS Cumulative away trophies (+)
Dependent variable Away Euro run (+)
log home attendance England qualifier win (+)
England qualifier draw ()
Bank Holiday (+)
Weekday evening ()
Sky dummy ()
Forrest, Simmons English soccer, PL and First Aims to estimate effects of TOBIT
and Szymanski (2004) Division 1992–1993 to 1997– broadcasting on match
1998 PL 3214 games, FD 3312 attendance and revenue
The large number of PL sell- TOBIT (PL) Dist () and squared (+)
outs suggests TOBITas an
estimator
For both divisions, the facility Dependent variable Weekday dummy ()
fee likely makes up for the
home team's lost ticket
revenue
Log attendance April–May dummy (+)
688 right censored Lpos () and squared (+) home, away
observations, 2526 other
OLS fixed effects (FD) Log last season mean attendancer (+)
home, away
Dependent variable Promoted dummy (+)
Log attendance Log last season mean attendance if
just promoted H ()
Season effects (+)
Weekday dummy ()
OLS FE
Similar to the above
Owen and Weatherston Super 12 RU, NZ fixtures Aims to investigate UO effects Significant variables from the final
(2004) involving the five NZ teams, on attendance log-linear estimate N = 70; liberal
each played at one of 17 strategy
(rotating) grounds. 83 games of
which 13 were capacity
constrained. Seasons 1999,
2000 and 2001.
Evidence
271
272
General-to-specific NB Lagged attendance relates PCGETS Log of lagged attendance (+)

modeling (GETS) in to the GROUND. Not to the
contrast to informal home team. Thus it is NOTan
selection and estimation index of team loyalty
of equations
A liberal strategy permits At each stage the model is Match-day nearest recorded rainfall
more explanatory simplified IFF it passes a ()
variables to “survive”. battery of tests for functional
See the corresponding form, normality of the errors,
final “conservative” homoscedasticity,
strategy model simultaneity, etc.
Dependent variable Cake Tin stadium (+)
Log attendance Home tries in last game (+)
Traditional NZ rival (+)
Away from overseas (+)
Away's proportion of possible points
earned so far (+)
Winfree, McCluskey, MLB 1963–1998, 884 team- Aims to measure the effect on a Distance to nearest rival ()
Mittelhammer and Fort seasons in all, regular play club's annual attendance of the
(2004) only. distance from its nearest rival,
and the impact effect of a new
team.
New Team dummy is unity if a NL-GLS allows for the New Team dummy ()
franchise opens within 500 nonlinear form and for
miles of the team in any year. heteroscedastic errors
Dependent variable Real ticket price () (inelastic)
Annual attendnce (100 000) Real per cap inc (+) (inelastic)
Local pop (+)
Win% (+), change win% (+)
New stadium dummy (+)
Lagged attendance (+)
Current season runs (+)
Div champs (+)
Strike ()
Attendance growth (+)
Team fixed effects (+)
Clapp and Hakes (2005) MLB 1950–2002, 1232 Aims to measure the effect of Results for whole-period regression.
potential franchise years, new stadiums on mean annual
(normal play) attendance.
N = 1179
53 lost due to lack of data on GLS Team win% (+)
stadium construction costs
New team dummy captures Allows for serial correlation Lagged win% (+)
the arrival in a given year of a and for heteroscedastic errors
new franchise within the city
Dependent variable Ln real ticket price () (inelastic)
Log of mean annual game Ln real stadium cost (+)
attendance.
Allows for the effect of league Other teams in city ()
expansion on the number of
games in normal play
Ln stadium capacity (+)
No of all stars (+)
Classic stadium dummies ()
Last year of stadium ()
Stadium aged<6 (+)
New team impact (+)
Ln local pop (+)
Evidence
273
274
Coates and Harrison (2005) MLB 1969–1996 all US based Aims to measure effect of Results for whole period regression.
franchises, 725 franchise years industrial disputes on
franchises' mean season
attendance (normal play)
Strikes and lockouts are IV Ln lagged attendance (+)
themselves endogenous. A
model that takes this into
account might produce
different conclusions.
Compare Putsis and Sen, who
endogenized NFL TV black
outs.
Recognizes the endogeneity of Ln population (+)
ticket price and that it is
observed with error
Dependent variable Ln stadium age ()
Ln of mean season attendance Win% (+)
Lagged win% ()
Post-season last year (+)
Strike-lockout impact dummies ()
Disputes have no effects beyond the
year in which they occur
Coates and Humphreys MLB, NFL and NBA, 1969– Aims to measure effect of new Per capita income (+) not NBA
(2005) 2001 780, 891 and 729 facility provision on mean
franchise years respectively season attendance, from
reduced form equations
NHL excluded due to the The authors use franchise OLS Win% (+) ALL
number of Canadian duration in the current city to
teams pick up “loyalty”
Looks like NBA may be an Rather confusingly they call it One reduced from per sport Stadium age () ALL
inferior good “team trend”
“Other franchises” means the Dependent variable Stadium age squared (+) MLB
number of teams in ANYof the
three sports within the city
Season mean attendance Team duration (+) MLB and NFL,
() NBA
Team duration squared () MLB
Other franchises () MLB and NBA
Last year post-season (+) ALL
Stadium age < 10 (+) ALL
Forrest, Simmons and English soccer, 1993–1994, Aims to explain how The paper also presents an OLS study
Buraimo (2005) 2001–2002 Sky audiences broadcasters select matches to (qv) of the audiences for the 522
televise televised games
*The author's Table 3 has 552 games were broadcast live
confusing nomenclature out of 3346
Do these dummies pick up Derbies and weekend Probit All season variables
games based on current scheduling work all season
PL standings, as Table 3 Matches played before Boxing On the 3346 games Derby (+)
suggests or (as the names Day matches had been selected
suggest) do they pick up pre-season
games that involve PL After BD selection is sensitive Dependent variable Weekend (+)
teams that currently to current form
were, or had been, PL and Dichotomous dummy = 1 if Pre-BD variables
Euro Champions? match is broadcast, otherwise
=0
UO as defined is expected to Combined wages (+)
have a negative effect on
audience
Broadcasters' Audience Wage differential (()
Research Board (BARB) data
are used in the complementary
OLS study of live audiences
Post-BD variables
Combined wages (+)
UO proxy ()
Dummy 1 (+)*
Evidence
Dummy 2 (+)*
The first seems more Dummy 3 (+)*
probable
275
276
Dummy 4 (+)*
Dummy 5 (+)*
*See first column.
Alavy, Gaskell, Leach and English PL soccer, Jan 2002 to Aims to detect UO effects on Results at Step Two
Szymanski (2006) May 2005 changes (per minute) in the
television audience during the
game
248 games for which BARB AB GMM estimator for the Dr at lags of 1 to 3 minutes (+)
possessed audience figures dynamic model of the TV
taken at one minute intervals audience (Step 2)
An extremely interesting NB Vars other than within Dependent variable Dtotal goals (+) home, away
extension to the concept game time dummies appear in
of SRUO rate of change (D) form in the
regression
The per minute change in the Dgoal in last minute (+)
BARB rating Dr
Dprob no score draw ()
Dprob score draw ()
D(prob home win – prob away win)2
()
Dsum of squares of changes in
probabilities since kick-off (+)
Multinomial logit model is Some minute dummies are
used to estimate the significant
probabilities of home wins
away wins score draws and
no-score draws. From these
UO measures are derived
for modelling the television
audience at Step Two
Buraimo, Forrest and FL Championship soccer Aims to model match Years in Champ (+) home, away
Simmonds (2006) 1997–1998 to 2003–2004, attendance to assist clubs in
2884 games decision-making.
Innovative approach (qv) to The authors' Table 2 probably Hausman-Taylor RE Model LnPop < 5 mls (+) home, away
measuring market size stacks only the more
significant variables from a
regression that may have
involved many more, e.g.,
lagged values
That they refer to “results from Allows for unobserved variable Close rivals ()home, away
our estimation” is suggestive effects
Dependent variable Relative wage (+) home, away
Ln of match attendance Derby (+)
Dist (), dist squared (+)
Midweek not television ()
Bank holiday (+)
Month dummies
Terrestrial television English team in
Euro comp ()
Points per game thus far (+) Home,
away
Game on ITV ()
Game on ITV Digital ()
Game on Sky ()
Forrest and Simmons Football league normal play Aims to measure the extent to Lagged home attendance (+) all
(2006) 1999–2002, 4320 games which FL attendances suffer divisions
from match congestion and
live broadcasting of higher
league matches
Suspect Table 2 contains the Prais-Winsten with panel Lagged Away attendance (+)
most significant regressors corrected errors – same Divisions 2, 3
only equation includes all divisions
Dependent variable Promoted dummy interacted with
lagged attendance (+) all divisions
Log of match attendance Derby dummy (+) all divisions
Distance (+), distance squared (-) All
divisions
Home points (+), away points (+) all
divisions
Recent home form (+) Divisions 2, 3
Evidence
Game on ITV () Division 1
ITV CL dummy () Divisions 2, 3
England on television () Division 3
277
278
Midweek dummy ()

Divisions 1, 2
Interactive fixture congestion
dummies
Lee (2006) Korean professional baseball, Aims to identify the main Ln real per capita income (+)
1982–2002, total annual causes of the decline in total
attendance, 21 observations attendance
Korean born C H Park pitched OLS (inelastic)
in MLB
Until the end of 2002 only Dependent variable Park I ()
these games were televised in
Korea. ParkI is a proxy for MLB
competition with KPBL
ln of annual mean attendance 2002 W Cup dummy ()
per game. Takes into account
changes in the number of
normal play games
Coates and Humphreys MLB, NBA and NFL Aims to investigate the From Table 2
(2007) attendance 1991–2001; 264, elasticity of attendance
298 and 308 franchise years demand w r to the prices of
respectively complementary items as
captured by the Fan Cost Index
Table 2 reports only some of GMM Ln FCI () NBA
the authors' results
To allow for simultaneity and Ln real ticket price (() MLB, NBA
unobserved effects
One equation per game Win% (+) ALL
Dependent variable Stadium age years (() NBA
Ln of mean annual home gate Ln real per capita income (+)
per franchise (inelastic) NBA
Ln lagged attendance (+) ALL
Population (+) MLB, NBA
Feddersen and Maennig Bundesliga soccer, 1963–1964 Aims to identify the effect on Real monthly earnings (+)
(2007) to 2005–2006, 13 100 games, attendance of a dedicated
772 club seasons arena relative to an all-purpose
stadium
Panel unit root tests applied The fixed (club) effects are very Tobit fixed effects and OLS End season league position ()
before estimation significant (qv) fixed effects
The absence of any price Borussia Dortmund, Bayern Dependent variable %games midweek ()
variable might suggest that Munich and Schalke 04 have
the authors regard their much larger than average base
equation as a reduced form support
Or that they are imposing Bribery scandal 1970–1971, The club's mean season Dcapacity (+)
the restriction that dummy set 1971–1972 to pick attendance
spectator demand is up possible effect in following
completely price-inelastic season
The paper shows no marked No of rivals within 100km ()
differences in the two sets of
estimates
As only 7 observations were Financial scandal dummy ()
right censored, this is not
surprising.
Novelty effect (+)
Dedicated arena (+)
Meehan, Nelson and MLB 2000–2002, 7189 Aims to test for significance of The results below are taken from
Richardson (2007) normal season games, after 45 UO effects on attendance and their Model 3
had been deleted due to lack of whether symmetrical
attendance data
This paper (and others Every team's first 10 games Cf issue of home advantage Weekday dummies () Sat is default
herein) could have made deleted each season. UO not (Chapter 8) Also to test for daytime (+)
advantageous use of very significant early on, see time-varying significance of
Jennett's and Kuypers' work Chapter 8 UO
to arrive at a more
systematic treatment of
medium term UO
See Chapter 8 and ****Must be a statistical Censored regression – 344 sell- Temp(+) rain ()
appendices artifact; an incorrect sign outs, out of 6828 games under
produced by a particular study
combination of errors
It is not credible that MLB is a Dependent variable match Covered stadium ()
Giffen good attendance
Support for asymmetry of UO New stadium (+)
effects
Evidence
Income (+)
Fan cost index (+)****
Inter-league (+)
279
280
Playoffs last year (+) home, away

Allstars (+) home, away
Games behind leader () home, away
Home win% (+)
Abswin%diff ()
Poswin%diff ()
Negwin%diff (+)
Rascher and Solmes (2007) NBA 2001–2002 all 1189 Aims to model UO effects on Probability home win (+)
matches attendance.
****Another statistical artifact Censored regression Square probability home win ()
Three models (qv) set up Estimated from January There were 446 sell-outs M Jordan (+)
and estimated. For brevity onward FCI has no significant
only the second is effect
considered.
Dependent variable All stars (+)
Game attendance Weekend (+)
Fan cost index (+)****
Median local household income ()
%non-white local pop ()
Local population (+)
Local unemployment rate (+)
Coates and Humphreys MLB, NBA and NFL
(2007) attendance 1991–2001; 264,
298 and 308 franchise years
respectively
Aims to investigate the From Table 2
elasticity of attendance
demand w r to the prices of
complementary items as
captured by the Fan Cost Index
Table 2 reports only some of GMM Ln FCI () NBA
the authors' results
To allow for simultaneity and Ln real ticket price () MLB, NBA
unobserved effects
One equation per game Win% (+) ALL
Dependent variable Stad age yrs () NBA
Ln of mean annual home gate Ln real per cap inc (+) inel NBA
per franchise
Ln lagged att (+) ALL
Pop (+) MLB, NBA
Brown and Link (2008) MLB 1995–2001, 204 Aim to detect possible Comments relate to their Model 5 of
franchise seasons only. bandwagon effects on local Table 3
Arizona and Tampa Bay joined revenues (gate, local television,
MLB in 1998 etc. and concessions)
In economics a Brown and Link take 84 wins OLS Time trend (+)
“bandwagon effect” is as the threshold at which
observed when the number bandwagon effects may begin
of people who buy a product
depends positively on the
number who currently buy
it
Brown and Link use it in a It will normally get the team Dependent variable Popwins (+)
different sense into post-season play
Successful teams attract Several models and estimators Local annual REVENUE* in Poppostseason games (+), lagged
people who are not used 1999 $ (+)
traditional fans
Retro state of art stadium (+)
Post-season games are intended to
capture the bandwagon effect Not so
easy to distinguish the two
bandwagon effects
Falter, Perignon and French first division soccer Aims to measure the effect on Model 5 Table 3
Vercruysse (2008) 1996–1997 to 1999–2000, attendance of France winning
1298 games the World Cup in 1998
The authors leave ticket No evidence of a WC effect on OLS Total attendance
price out on the grounds of the total number of transient
simultaneity. Better to spectators although if the
incorporate and use IV to model (qv) is estimated with a
deal with the problem percentage capacity audience
figure this conclusion is
reversed
But if demand is totally Dependent variable WC dummy (+)
price inelastic their
Evidence
specification is correct
Ln attendance Host city (+)
OLS Lpos () home, away
281
282
Dependent variable Team budget (+) home

ln non season ticket holders Last score (+) home
Sunshine (+)
Transport cost ()
Derby (+)
Fixed effects home
N-S attendance
Similar except WC, host city and
team budgets insignificant
Winfree and Fort (2008) US Minor and Junior Hockey Aims to measure the extent to Time trend (+) ML, JL
leagues 1998–1999 to 2005– which NHL fans substituted
2006. 707 ML games and 531 ML and JL hockey during the
JL games in the sample lockout of 2004–2005
A lockout during 2004– 25% ML teams with local NHL GLS Ln points per game (+) ML, JL
2005 caused the loss of the team are in Canada
entire NHL season
Time trend to capture 80% JL teams with local NHL Some evidence of serially NonNHL teams (() ML
effects of unobserved team are in Canada correlated errors
change in prices,
population, incomes, etc.
Dependent variable Strike dummy (+) ML
Ln per game attendance NHL team (+) JL
Seems as if whatever substitute there
is goes into ML
ML teams (not JL) are affiliates of
NHL
Evidence 283
increasingly addressed in demand studies. One of the first studies to account

for this problem was Kuypers (1996), who recognized that 10% of English
football matches were sell-out fixtures. Explicit treatment of capacity con-
straints is now, however, much more common. This is the case in Price and
Sen's (2003) Tobit analysis of match attendance at US college football, Forrest
et al's (2004) analysis of Premier League attendances, Meehan et al's (2007)
“Censored Normal Regression,” i.e., Tobit, analysis of Major League Baseball,
Feddersen and Maennig's (2007) analysis of the Bundesliga, and Falter et al's
(2008) analysis of French soccer.
Finally, the earlier research that exists offers mixed results on the effects
of television broadcasting on attendance demand. The ambiguity persists
with more recent research; in Falter and Perignon (2000) and Falter et al
(2008) the broadcasting of French football matches is not significantly
associated with attendances, which is echoed by Verhaegen (2006) in Bel-
gian football, while Price and Sen (2003) show that the scheduled broadcast
of college football on television is positively correlated with attendances. In
contrast, Garcia and Rodriguez (2002) find that televised Spanish soccer
matches have lower attendances. The same is true for English soccer for
Aston Villa according to Allan (2004) and for matches outside of the
Premier League according to Forrest et al (2004), Buraimo et al (2006)
and Forrest and Simmons (2006). A significant contribution of these latter
studies is that they do control for the scheduling of matches and the
possibility of substitution effects, because it is Premier League clubs com-
peting in the Champions' League. Overall, the results show that televised
matches are more likely to generate substitute effects when higher quality
and lower quality matches are compared and that it may be difficult to
disentangle the effects from scheduling changes.
10.4.2 Measuring price and income

As noted above, measuring the price of sports matches typically depends on
relatively imprecise data. Indeed, as a result of this, many studies end up
having to ignore prices because of lack of data. Of the more recent studies this
applies to Jones et al's (2000) study of British rugby league, Owen and Weath-
erston's (2004) study of Super 12 rugby union, Forrest et al's (2004) study of
Premier League and First Division soccer, Barajas and Crolley's (2005) study of
Spanish soccer, Coates and Humphrey's (2005) study of Major League Base-
ball, Buraimo et al. (2006) and Forrest and Simmons' (2006) studies of
the English Football League, Feddersen and Maennig's (2007) study of the
Bundesliga, and Brown and Link's (2008) and Winfree and Fort's (2008)
studies of Major League Baseball.
BOX 10.2 PRICE ELASTICITY OF DEMAND

A price elasticity of demand of exactly 1.0 means Price elasticities less than 1, for example 1.5, suggest that
that a 1% price rise or fall in ticket prices (other things a rise or fall in ticket prices of 1% would decrease or increase
being equal) reduces or increases attendance by 1%. attendance by 1.5%, hence sales revenue falls or rises. This
As revenue is equivalent to price attendance, i.e., means that it pays to cut prices and demand is referred to as
quantity, the increase or decrease in revenue that would price elastic. A price elasticity of between 0 and 1, for
follow from a price rise or fall is exactly offset by the example 0.5, suggests that a rise or fall in ticket prices of
reduction or increase in revenue from the reduction or 1% would decrease or increase attendance by 0.5%, hence
increase in attendance. Total sales revenue remains sales revenue rises or falls. This means that it pays to raise
constant. prices and demand is referred to as price inelastic.
When price has been measured, in earlier studies minimum adult admis-
sion prices have been used or an average price based on revenues and atten-
dance. It follows that, if sports engage in price discrimination, such an average
price will be a biased estimator of the actual prices paid by any given supporter.
More recent research studies, such as Falter and Perignon (2000), identify that
ticket price changes do not affect attendance at French soccer and Coates and
Harrison (2005) found the same in Major League Baseball. However, the
conventional wisdom is that most studies (see also Fort 2004) identify that
attendance and ticket prices are negatively related and, moreover, that demand
is price inelastic. Elasticity is a unit-free measurement indicating the percent-
age change in the dependent variable that will follow from any given percent-
age change in the independent variable. Price inelasticity suggests that sports
fans are not sensitive to changes in ticket prices, as illustrated in Box 10.2.
The finding that sports fans have price inelastic demand raises two impor-
tant issues. The first is that it provides a rationale behind teams raising prices
to increase their revenues from price increases. In this regard, Dobson and
Goddard (2001) chart the extremely close relationship between ticket price
increases and revenue increases in English soccer.2 The second implication,
however, calls into doubt the assumption of profit maximization for teams.
The reason for this is illustrated in the figure in Box 10.3, where it is shown
that profit-maximizing teams should price in the elastic portion of their
demand curve. The finding that teams face price inelastic demands has, thus,
generated debate. For example, Fort (2004) suggests that in the US profit
maximization still follows if the analysis allows for the impact of larger-
than-average local television revenues being earned by a team. This suggests
that MR from gate attendance can be negative if offset by a positive MR from
2
See also Dejonghe (2007).
Evidence 285
BOX 10.3 INELASTIC PRICING

Profit maximization occurs where marginal cost (MC) and fall must imply that demand is inelastic. As can be seen
marginal revenue (MR) are equal. With ticket prices (p) on from the diagram below, for teams to set their ticket prices in
the vertical axis and attendance (A) on the horizontal axis, the inelastic portion of their demands must mean that MR
attendance demand, i.e., average revenue (AR) and MR and MC are not equal, as is implied in profit maximization.
can be illustrated as shown in the figure. As shown in Some profit must be forsaken.
Appendix 10.1, MR will have twice the slope of a linear
AR curve. If MC is positive this means that MR has to be
positive. Given that the formula for the ticket price elasticity
of attendance demand is (qA/qp p/A), with a straight
line demand curve of constant slope, price elasticity of
demand can be inferred from the ratio p/A. Intuitively, elas-
tic demands will lie at higher prices/lower attendances and
inelastic demands at lower prices/higher attendances.
More specifically, because MR measures the change in
revenue following a change in price, as discussed in
Box 10.2, a falling price can only reduce revenue with an
inelastic demand. Reducing revenue means that MR must
be negative, so it follows that a negative MR following a price
television. However, this argument has been formally challenged, suggesting

that “the paradox of inelastic sport ticket pricing persists.” (Porter, 2007,
p. 158.)3
Qualifications can be made to the general claim; Simmons (1996) shows
that it is season ticket holders that have inelastic demands relative to casual
spectators. Researchers have also attempted to identify the full economic price
of attendance that might suggest that inelasticity masks the true effects of
price on attendance. One strand of research developed in the US has been to
make use of the “Fan Cost Index” instead of just the ticket price. This covers
the cost of two adults and two child tickets, four small soft drinks and hot dogs,
two small beers, programmes, two adult sized baseball caps and parking. This
index clearly captures a broader set of costs. A priori, however, it is not apparent
which sign this variable would take according to the theory of demand. On the
one hand, as a broader set of opportunity costs are associated with game
attendance one might expect to see a negative sign. On the other hand, it is
an example of product bundling, providing a discount on individual tickets. If
3
Formally it is argued that Fort's (2004) analysis confuses the MR attributed from price
changes with the MR associated with the acquisition of playing talent. See also Chapter 9.
this is the case, then one might expect to see a positive stimulus to demand.
The literature reflects this ambiguity, with Meehan et al. (2007) identifying a
positive impact on attendances in Major League Baseball, while Coates and
Humphreys (2007) and Rascher and Solmes (2007) found a negative and
insignificant impact in the National Basketball Association, respectively.
The second and more traditional approach has been to integrate some
indication of the travel costs associated with match attendance. In the earlier
studies, such as Baimbridge et al (1995, 1996), the distance between venues
acts as an indication of travel cost in studies of English soccer and British rugby
league, with a negative sign being consistent with a “price” effect. More recent
studies, such as Garcia and Rodriguez (2002) for Spanish soccer, Allan (2004)
for Aston Villa soccer club, Winfree et al (2004) for Major League Baseball, and
Forrest et al (2004) Buraimo et al (2006) and Forrest and Simmons (2006) for
English soccer, identify similar effects. Other studies, such as Falter and Peri-
gnon (2000) and Falter et al (2008), use return second-class rail fare as a direct
proxy for travel cost in studies of French football. In both cases a significant
negative sign is detected.
A particularly innovative study by Forrest et al (2002) combines both
distance and direct cost of travel measures to produce a “generalized cost of
attendance” of Premier League soccer, based on a large-scale fan survey in 1998
and 1999. This comprised a weighted average ticket price for season and non-
season tickets, plus a measure of the direct- and time-costs of attending. Based
on this generalized cost it was revealed that 8 clubs had price elastic demands
and 12 had inelastic demands.
There has been less detailed research into the availability of substitute
products on attendance demand. However, where it exists, the evidence
suggests a broadly negative influence on attendance which could also help
to account for price inelastic demands. Consequently, Dobson and God-
dard (2001) note that the presence of professional rugby league in the north
of England could account for the lower attendances, other things being
equal, of northern English football clubs compared to their southern coun-
terparts. Price and Sen (2003) also note that the number of college teams
per state population can reduce attendances at college football games,
Clapp and Hakes (2005) that the number of teams in the city can reduce
Major League Baseball attendances, a view supported by Coates and Hum-
phreys (2005). Likewise, Buraimo et al (2006) show that other English
soccer teams in the same city can reduce attendances for clubs, and Fed-
dersen and Maennig (2007) demonstrate that the number of rival teams
within 100 kilometers of Bundesliga teams reduce attendance. Further
pertinent results are provided by Winfree et al (2004) and Winfree and Fort
(2008), who note respectively that attendances at Major League Baseball
Evidence 287
teams can fall following the relocation of a franchise to within a 500 mile
radius of a team, or that an increasing number of non-national hockey
teams in the same standard metropolitan statistical area (SMSA) can re-
duce attendances in Minor League Hockey. In contrast, Owen and Weath-
erston (2004) do not identify a significant effect of other Super 12 rugby
union franchises being on the same land mass on their attendances, and
Winfree and Fort (2008) that the presence of a National Hockey League
team in the same SMSA can increase attendance in Junior League Hockey.
The latter results suggest complementarity in consumption.
Similar problems of data availability apply to attempts to measure the
income of fans and the impact that this has on attendance. As the theory of
demand refers to the individual's disposable income, it follows that lack of data
or the fact that it varies inconsequentially over the study period has meant that
many studies exclude its influence. Of the more recent literature this applies to
Forrest et al (2004), Buraimo et al (2006) and Forrest and Simmons (2006) for
English soccer; Clapp and Hakes (2005) and Brown and Link (2008) for Major
League Baseball; Winfree and Fort (2008) for Minor and Junior League Hockey
and Falter et al (2008) for French soccer. Where attempts have been made to
include income, researchers have tended to rely on averages of earnings,
employment status or some classification of fans according to socio-economic
status.
Most of the more recent studies identify a significant positive but income
inelastic affect on attendances, but make use of a measure of earnings. These
include Garcia and Rodriguez (2002) who use regional per capita income in
Spanish soccer; Winfree et al (2004) who use real per capita income in the
state for Major League Baseball; Lee (2006) and Meehan et al (2007) who use
per capita income for Korean baseball and Major League Baseball respectively;
Coates and Harrison (2005) who use real income for Major League Baseball;
and Feddersen and Maennig (2007) who use real gross earnings in trade and
industry for German soccer. An exception is Falter et al (2000) who find a
negative effect using average earnings in the department, suggesting that
French soccer is an inferior good, a finding that is confirmed when using the
unemployment rate as an alternative proxy for income. A positive sign on
unemployment is identified.
Why does a positive relationship between unemployment rates and attendance suggest an
inferior good?
Hint: Think about the relationship between income and employment.
Because one would expect to have an inverse relationship between unem-

ployment and income, the fact that the latter is associated with a rise in
attendance suggests that attendance rises with falls in income and vice versa.
This is the definition of an inferior good.
10.4.3 Consumer preferences

As well as the typical economic determinants of demand suggested by theory,
sports researchers have adapted the theory to the sports context, which might
be understood as reflecting elements of fans preferences which include quality
characteristics, as well as interdependent preferences.
10.4.3.1 Quality characteristics

As far as the qualitative attributes connected with the viewing of a contest are
concerned, to begin with, aspects of the stadium have been investigated. For
example, Owen and Weatherston (2004) note increased attendance in Super
12 rugby union for games played in the “Cake Tin,” a distinctive stadium in
New Zealand. Winfree et al (2004) and Meehan et al (2007) suggest that new
stadiums generate a positive stimulus for Major League Baseball attendances,
but the latter also find that this is not the case for domed or roofed stadiums.
This finding is supported by Brown and Link (2008), who identify increases in
Major League Baseball attendances for “retro state-of-the-art” stadia. More-
over, Coates and Harrison (2005) identify that older stadiums reduced atten-
dances in Major League Baseball. This is a result confirmed by Coates and
Humphreys (2005, 2007) in Major League Baseball, the National Basketball
Association and the National Football League. These results are contrasted by
Clapp and Hakes (2005), who identify that older stadiums increased atten-
dances in Major League Baseball, but classic stadiums reduced them, as did the
final year of old stadiums. Finally, Feddersen and Maennig (2007) identify that
dedicated soccer stadiums receive greater attendances than other German
stadiums in which soccer is played, and that construction and novelty can
decrease and increase attendances respectively. In general, the results suggest
that fans appreciate dedicated stadiums, and that they are loyal to their sta-
diums but also receptive to new construction.
In general, the scheduling of fixtures appears to matter, with most studies
identifying that moving games from traditional weekend slots reduces atten-
dances. For example, Garcia and Rodriguez (2002) identify reductions in
attendance at Spanish soccer for games during the week, Forrest et al (2004)
find likewise for the Premier League and English Football League First Divi-
sion, Allan (2004) for midweek games for Aston Villa football club, Forrest and
Simmons (2006) for midweek games for Divisions 1 and 2 for English soccer,
Evidence 289
Meehan et al (2007) for Major League Baseball for weekday games compared to
Saturday, and Feddersen and Maennig (2007) for midweek German soccer
games. Natural exceptions to this are public holidays where there is a ubiqui-
tous finding that attendances increase, as illustrated by Allan (2004), Buraimo
et al (2006) and Forrest and Simmons (1996) for English soccer.
Predictable results are identified for the impact of the weather on atten-
dances, with them increasing with higher temperatures and reducing with
rain. Garcia and Rodriguez (2002), for example, show that attendances at
Spanish soccer increase in the absence of rain, and Owen and Weatherston
(2004) found that attendances at Super 12 rugby union matches decrease
when it rains. The same result is identified for Major League Baseball by
Meehan et al (2007). Likewise, Falter et al (2008) show that French soccer
attendance rises with the percentage of sunshine. Many studies also include
significant seasonal dummy variables which, in part, may also identify the
effects of weather on attendance, as well as factors such as championship
contention.
As far as the quality of the contest is concerned, naturally UO is of great
importance to researchers. The effect of UO on attendance was reviewed in
Chapter 8, where it was argued that the evidence was mixed, although it was
noted that support for the hypothesis is increasingly evident in more recent
studies. The only point worth adding to this discussion is that many studies
identify that “derby” matches involving local rivalries can increase atten-
dances. For example, this is the case for Garcia and Rodriguez (2002) in
Spanish soccer, Owen and Weatherston (2004) in Super 12 rugby union,
Forrest et al (2004), Buraimo et al (2006) and Forrest and Simmons (2006)
in English soccer and Falter et al (2008) in French soccer. However, the liter-
ature balances this hypothesis with the finding that attendances for clubs
increase with the success of the specific club.
This suggests that the quality of the team is important for attendance. As
well as the success of the team, a number of other methods have been used to
identify team and specific player qualities, such as budgets for players (Garcia
and Rodriguez, 2002), wage bills (Buraimo et al. 2006) and the visit of some
historically dominant teams (Buraimo and Simmons, 2007; Verhaegen,
2006). The importance of star players being present in a team has also been
investigated (Rascher and Solmes, 2007; Meehan et al. 2007) and shown to
increase attendance.
10.4.3.2 Interdependent preferences and habit persistence

There is a large body of evidence which asserts interdependent preferences in
professional team sports. On the one hand, it is argued that bandwagon effects
are identified by Brown and Link (2008), because of the impact of post-season
play increasing local revenues in Major League Baseball. Strictly speaking,

however, this is not a bandwagon effect, through consumption being deter-
mined by other's consumption, but is more properly connected with some fans
attending because of the team's success, as discussed above.
Likewise, there are numerous studies indicating that loyalty and habit
persistence are demonstrated by the significance of lagged attendance on
current attendance. Borland (1987) and Simmons (1996) are earlier examples
in Australian Rules football and English soccer respectively. Owen and Weath-
erston (2004) in Super 12 rugby union, Winfree et al (2004), Coates and
Harrison (2005) and Coates and Humphreys (2007) in Major League Baseball,
Forrest et al (2004) and Forrest and Simmons (2006) in English soccer are
more recent examples. As noted earlier, it is by no means clear that such a
measure correctly identifies habit persistence, and hence interdependent pre-
ferences, i.e., fans consuming more games because of past experience or be-
cause others do. However, Dobson and Goddard (2001) do note that the age of
the club is connected with increased attendances in English soccer, and Bur-
aimo et al (2006) that the duration of a fan's club membership increases
attendances. Intuitively, therefore, the results are suggestive of habit persis-
tence and tradition, although Appendix 10.1 cautions against this interpreta-
tion.
10.4.3.3 Spatial context

The spatial nature of the demand for professional sports is, of course, clearly
recognized in theories of sports leagues, as discussed in Chapter 9 in which it is
argued that teams with larger potential market sizes draw higher gates. The
evidence supports this argument with most of the literature identifying the
market size of the home team having a positive impact on attendances. More
recently, Dobson and Goddard (2001) make use of population census data for
the city in which the club is located to examine English soccer, Garcia and
Rodriguez (2002) use population in the province of the club in Spanish soccer,
Owen and Weatherston (2004) make use of the regional population in Super
12 rugby union and Buraimo and Simmons (2006) use population within five
miles of the home and away club in English soccer. The population of the away
team is often included in studies, but is typically deflated by the distance
between the clubs. This suggests that market potential dissipates with an
increase in travel costs.
In the US it is common to use the population of the Standard Metropolitan
Statistical Area to proxy market size. This is the case in Meehan et al's (2007)
analysis of Major League Baseball and Coates and Humphreys' (2007) analysis
of Major League Baseball, National Basketball Association and National Foot-
ball League attendances. Rascher and Solmes' (2007) analysis of the National
Evidence 291
Basketball Association makes use of the population in the Consolidated Met-

ropolitan Statistical Area (CSMA), which is an updated definition of popula-
tion. Some studies, such as Winfree et al (2004) make use of the city
population in their analysis of Major League Baseball.
If more than one club is located in the area, researchers use some mecha-
nism to allocate the population. For example Kesenne and Janssens (1987)
divided the population of the arrondissement by the number of teams; Garcia
and Rodriguez (2002) used the proportion of season ticket holders. This mea-
suring criterion may not be the most appropriate, however. Professional team
sport allegiance may vary spatially and Dejonghe et al (2006), for example,
note that for the Netherlands the market areas of most of the clubs may be
more strongly related with “marketing zones,” for example as reflected in
coverage by local newspapers. It remains, however, that “local” population
seems to be ubiquitously associated with attendances and larger population
bases raise attendance.
10.4.4 Estimation issues

While it is noted above that much of the earlier literature makes use of rather
elementary applications of OLS estimation, there is evidence of an increas-
ingly sophisticated econometric approach to both estimation and diagnostic
checking. Developments are not consistent, however. In terms of simultane-
ity, studies such as Davies et al (1995), Dobson and Goddard (1998) and Jones
et al (2000) identify the possibility of “reverse causality” between attendance
and success in studies of English rugby league and soccer, suggesting simul-
taneity. A number of studies also note that prices and attendances may be
simultaneously related. On this basis, recent work by Falter and Perignon
(2000) drops the price variable from their analysis of French soccer, while
Garcia and Rodriguez (2002) account for it in their instrumental variable
estimation. It is explicitly accounted for in the application of multivariate
systems by Owen and Weatherston (2004) for Super 12 rugby union, and in
Schmidt and Berri's (2001) time series analysis of Major League Baseball.
Negligible attention is paid in the literature to multicollinearity, although it
is tacitly accepted by researchers inasmuch as many researchers routinely
categorize sets of variables together, such as “economic,” “sporting success,”
“UO” or “controls.” One of the few studies to address such issues empirically
is Garcia and Rodriguez (2002) who perform F-tests on these blocks of vari-
ables to test for their joint significance, consistent with their conceptual group-
ings in an analysis of Spanish soccer.
Finally, the more recent research now routinely addresses serial correlation
and heteroscedasticity. For example, Winfree et al (2004) used generalized
least squares to estimate attendance in Major League Baseball because of

heteroscedasticity; Lee (2006) undertakes generalized least squares to correct
for serial correlation in estimating attendances in Korean baseball; Forrest and
Simmon's (2006) use a Prais–Winsten estimator to control for autocorrelation
and adjusted standard errors to control for heteroscedasticity in an analysis of
English soccer; Winfree and Fort (2008) tested for heteroscedasticity in their
study of Minor League and Junior League Hockey, finding none; and Falter et al
(2008) corrected their standard errors for heteroscedasticity in an analysis of
French soccer.
10.5 BROADCAST DEMAND

The discussions above indicated that the broadcasting of sports events may
affect gate paying attendance, suggesting that the demands are substitutes.
This issue is worth further investigation because, as discussed in Chapter 7,
the development of sports leagues has become increasingly dependent on
media income, as evidenced by the conceptualization of the contemporary
European sports league as a Media–Corporations–Merchandising–Markets–
Global or MCMMG model, in which for the largest clubs the single biggest
source of income derives from the media. Understanding how and why this
change in the “product” market occurred is an important issue that is
addressed in the remainder of this chapter.
10.5.1 A brief outline of sports and broadcasting history

The evolution of modern sports leagues is directly linked with the growth of
media income for sport and its distribution has fed directly back into the
structure of sport. For example, the Premier League (PL) was directly con-
nected with the offer of £43 million for the rights to broadcast live games of
the larger, more powerful, clubs by the emerging satellite broadcaster, Sky.
Until then the BBC and Independent Television Network (ITV), which began
as a commercial provider in 1955, shared the televising of sport, which was
relatively limited, as a “public” good. Indeed, the BBC had its rights to broad-
cast major sporting events enshrined in the 1981 and 1984 Broadcasting Acts.
This benign cartel was broken with the arrival of Sky, with developments then
following a model that had been experienced in the US with CBS funding the
AFL, a rival league to the NFL, which ultimately led to the merger of the
leagues. The PL consequently was established as an independent league, al-
though promotion to and relegation from the EFL remained. Other sports have
followed this model of development. The EFL attempted to emulate the fund-
ing model of the PL by securing a contract with ITV Digital for £315 million.
Broadcast Demand 293
However, the market for championship football failed to attract subscribers,

leading to the bankcruptcy of the broadcaster and considerable financial dis-
location for the league.
In addition to the Premiership, Broadcasting Rights provide the financial
backing for the Champions League in European football which evolved, as
discussed in Chapter 7, from the European Cup. Outside of soccer in the UK,
both rugby league and rugby union have followed the Premier League broad-
casting model for football, as have the Super League and Premership respec-
tively, while rugby union also has a European club championship, the
Heineken Cup, in which teams from France, Italy and regional teams from
Scotland, Ireland and Wales compete. In these smaller rugby unions tradition-
al teams compete in national leagues, while the franchises compete in the
Celtic (now Magner's) League.
Elsewhere in Europe, as noted by Szymanski (2006b), the evolution of
broadcasting has been similar, with the removal of monopolies by public
broadcasters and a general migration to pay television through satellite broad-
casting, although in August 2006 the monopoly position of Sky in the UK was
broken following a European Commission ruling, with Setanta sports emerg-
ing as a rival. Interestingly, however, as Andreff and Bourg (2006) point out,
different models of rights allocation exist. In the UK and Germany cartels of
clubs negotiate rights' deals. In France, the league owns and negotiates rights,
whereas in Spain and Italy individual clubs undertake this function. Regard-
less of these differences, soccer has been the main driver, accounting for over
50% of sports broadcasting rights. Furthermore, a common implication of
these developments is a redirection of funds towards the bigger clubs in leagues
(this is programmed into Sky's funding formula for the PL) and towards the
bigger leagues. While all clubs in the PL receive an equal fixed fee from 50% of
the television revenues, for example approximately £743 000 in 1993–1994,
rising to £8 885 000 in 2006–2007, the remaining 50% of revenues are
allocated unequally. One half of this (25% of the total) is allocated according
to the teams' positions in the league. This is known as a Merit Award. In
1993–1994 the Premier League champions, Manchester United, received ap-
proximately £856 000. They received £9 357 000 as champions in 2006–
2007. Likewise, Swindon Town, the team relegated in 1993–1994, received
approximately, £38 000 and Watford, relegated in 2006–2007, received ap-
proximately £487 000. The remaining 25% of television revenues are allocated
according to the number of live matches that a team features in. Each team
must appear in at least three live fixtures. This is known as the facility fee. Not
surprisingly, above the minimum number of televised matches, fixtures in-
volving more successful teams are broadcast more often (as discussed further
below). Consequently, for the seasons noted above, Manchester United
received a facility fee of approximately £1 011 000 and £9 357 000, respec-
tively. Likewise, Swindon Town and Watford received approximately £243 150
and £3 271 000, respectively (Baimbridge et al, 1996; Findlay et al, 1999;
Deloitte, 2007).
Across Europe, however, while the growth in funding has been similar, the
distribution has varied. Vrooman (2007) estimates that total earnings of the
Premier League grew at an annual average (compound) rate of 16% during the
nine years from 1995–1996 (when the Bosman Ruling came into play), reach-
ing 1987 million in 2004–2005. Of this total, broadcasting revenue comprised
862 million, about double the broadcasting revenues in Germany, France and
Spain, and substantially above that in Italy. In turn, broadcasting revenue in
the PL increased at an average annual rate of 33%, increasing as a proportion of
total PL revenue from about 12% in 1995–1996 to 43%. Average revenue
growth rates have been slower in the rest of the “Big Five” leagues in Spain,
Germany, France and Italy. The price paid annually for television broadcasting
rights to the Bundesliga increased from DM255 million in 1997–2000 to
DM750 million in 2000–2004. The annual price of television broadcasting
rights to the French Ligue rose from 125 million in 2001–2004 to 600 million
in 2004–2007.4 In this regard, the PL is the biggest sports league in Europe.
Elsewhere in the world similar situations apply with, for example, Fox
Sports broadcasting Australian Rules football, the National Rugby League
and the emergent A-league for soccer, as well as the Super 14 and Tri-nations
rugby union competitions in Australia. The increased funding in rugby league,
which provided the largest Australian market, and which threatened to absorb
the best rugby talent, led to the formation of the latter rugby union competi-
tions. A significant part of these deals was to repackage games such as rugby
union into a franchised competition that aggregated talent into regions and
was not broadcast on free-to-air television.
Similar developments are beginning to occur in cricket. The duration of
cricket games has always led to problems attracting spectators, leading to the
development of one-day cricket in the 1960s, in contrast to multiple-day
County and test matches. “Twenty-20” cricket was developed in 2003 by the
England and Wales Cricket Board with the direct intent of courting new fans
and being attractive to the media as a complement to existing forms of the
game. In June 2008, however, an Indian Premier League (IPL) emerged, devel-
oped by the Board of Control for Cricket in India, and underwritten with a
television rights deal of $1.026 billion over ten years from a consortium of
businesses from India's Sony Entertainment Television network. The IPL
4
According to “The Political Economy of Football” web address (http://www.
footballeconomy.com/stats/stats_tv_03.htm).
franchises have been signing the star players from around the world, which has
caused consternation in countries such as the UK, as the Indian season over-
laps their traditional season. This could limit the availability of players for
their traditional teams and competitions, but nonetheless they stand to earn
far more money from playing in the IPL. Matters are complicated further
because of a rival unofficial league, the Indian cricket league, underwritten
by a rival broadcaster in India. To meet this challenge, the English and Wales
Cricket Board is, at the time of writing, looking for financial backing from Sir
Allen Stanford, a billionaire who has underwritten Twenty-20 cricket in the
West Indies. The English Premier League would look to involve teams from its
counties plus others from India, Australia and South Africa. The England and
Wales board faces the dilemma of seeking to protect and to promote all ele-
ments of the game, but faces the potential leakage of its best players.
The upshot of these discussions is that media income has grown substan-
tially for sport and its distribution feeds back directly into the structure of
sport. While this situation persists, it is difficult to see how sports leagues and
competitive structures can evolve independently of the demands of the media,
as the media provides financial incentives to change. It also suggests that
uncertainty of outcome may not play the role traditionally thought in sports'
league evolution, although it may be an outcome of restructuring with the
intent of drawing together the best talent available in a specific competition.
The economics underlying these issues are now discussed.
10.5.2 The economics of broadcasting and sport

The successful inroads made by broadcasting companies into the sports in-
dustry can be understood in terms of Neale (1964), who identified that pro-
fessional team sports confer an externality on the public by the league standing
effect, i.e., widespread interest in a successful sports team exceeds those that
pay gate fees. A distinct element of this, the “fourth estate” benefit carries over
directly to the media, who can earn income at no direct cost to themselves. It
can be argued that public television broadcasting was a means of accommo-
dating this externality. However, as discussed in detail in earlier chapters, the
Coase theorem states that if property rights can be established for a resource
and if a market can be organized to price the resources correctly, optimal
allocation or something akin to it should follow. As the property rights to
broadcast sporting fixtures can be easily established, because of restricted
access to sporting stadiums, it follows that there was always the option to
supply sports through such media. What ultimately mattered for the more
recent changes to broadcasting was the technical ability to supply broadcast
coverage with restricted access and an economic incentive to do so. The
general growth of broadcasting technology into cable and satellite (digital)

forms has made restricting access to broadcasts possible, which was not the
case with traditional broadcasting.5 Likewise, the economic regulatory climate
of the 1980s and 1990s was conducive to the deregulation of industries. In
turn, the scale of demand for sports has provided the economic conditions to
promote these forms of supply. Historically, then, it is not surprising that
sports leagues have acted collectively to sell the rights to broadcast live
matches and highlights. This “internalizes” the external benefits. Moreover,
the rise of the new media required explicit subscriptions to the appropriate
technology to generate effective demand. As an expression of the derived
demand for sport, broadcasters had to channel these funds towards sporting
leagues to provide the incentives for leagues to supply them with access to
fixtures in opposition to rival broadcasters.
As noted above, one of the ways in which this might have been prevented is
through explicit economic policy. However, regulators have been relatively
muted on the need to rule against collective selling of television rights. The
collective selling of media rights is sometimes defended, on the grounds that it
is likely to produce more revenue in total and to produce a more equal distri-
bution of broadcasting revenue than if clubs sell rights individually. On the
other hand, it creates the potential for monopolistic restriction of output and
over pricing, and the EU has expressed concern. The logical justification for
the league, rather than clubs selling television rights, is that no club can create
a saleable product in the absence of the league which therefore has a property
right in every league match. This is clear in Neale (1964) who went so far as to
regard the league as a monopoly supplier and the clubs as its “plants.” What is
interesting is that the US and Europe have developed different models, in spite
of the sustained increases in value of the markets noted earlier.
In the US, collective bargaining exists with revenue distribution, while in
Europe sports rights have become associated with premium pay television
channels. Szymanski (2006b) argues that this has been due to four main
reasons. The first is that, unlike the US, one major sport dominates Europe,
which means that the willingness to pay of European consumers will be greater
than in the US, where all four main sports share the market. The second is that
the US has economies of scale, in that markets for sports are national. This
means that national carriers cannot pay as much as pay television networks,
5
There are, of course, still potential leakages to property rights. Signs are emerging that the
owners of rights face problems similar to those experienced by the owners of recording rights
in the music industry. Websites exist that offer (for a small charge or none at all) software to
connect your PC via telephone to the People's Republic of China where, by agreement with
the rights holders (Sky), Serie A matches are broadcast free-to-air.
who can operate in separate segments, i.e., national markets, in Europe. These
points imply that sports are not substitutes in Europe, and that national
leagues are also not substitutes. The third point is that US leagues have
been receptive to designing their competitions around media needs. The final
point offered, and the one favoured as an explanation, is that US regulations
have remained tighter on opening up the media market to competition than
those in Europe. As discussed earlier, in Europe there has been an attempt to
deregulate the market, but this has taken place in an environment in which
some leagues have collective selling of rights and others have decentralized
selling.
This naturally raises the question of which model is most appropriate. It is
clear that some form of decentralized selling is preferred by the European
Union, yet the impacts on consumer welfare are not that straightforward
(Jeanrenaud and Kesenne, 2006). On the one hand, as discussed earlier, decen-
tralized selling has led to UK rights values actually increasing. This could be
the results of multiple downstream monopolies being created, although over-
all more games are being televised. On the other hand, non-exclusivity in some
form for broadcasters would undermine the pay television market, because of
the presence of perfect substitutes. Moreover, one could argue that having
individual clubs make their own deals could benefit the larger clubs most,
thereby affecting competitive balance. However, this is already happening in
the UK with pooled rights. What seems to matter then, for competitive bal-
ance, is revenue sharing or not, rather than if rights are pooled or not, as
discussed in Chapter 9. As noted there, moreover, the impacts on competitive
balance will also depend on the objective of clubs, with revenue sharing po-
tentially contributing to competitive balance in a league of win maximizing
clubs.
What does seem to be clear is that there is a form of tripartite trade-off of
rents involved, in which rights deals extract some rents from consumers and
transfer these in part to the sport. Any wholesale reintroduction of public
broadcasting, then, may increase consumers' surplus at the expense of funds
in the industry. At the same time, as Szymanski (2006b) argues fans want to
watch the highest level of competition that sports have to offer and broadcast-
ing money may well have helped to provide this.6 Significantly, this may well
have helped to promote uncertainty of outcome as a market solution.
There is evidence to support these claims. In the studies of attendance
demand noted above, it is shown that, while televised games reduce atten-
dances, overall televised games correspond with increased revenues for the
6
However, as noted in the next chapter, it is also the case that much of this rent has found its
way to the players.
clubs in the Premier League and First Division (see Baimbridge et al, 1996;
Forrest et al, 2004). The current deals thus benefit the clubs as a transfer from
consumers. This naturally provides incentives for the clubs to have the deals,
because it is able to cross-subsidize its gate paying demand from broadcast
demand. Secondly, in two innovative studies, Forrest et al (2004) and Alavy
et al (2006) examine the choice of broadcasters to televise a game and broad-
cast viewing figures on a minute-by-minute basis respectively. In the former
case it can be shown that in the second half of the season when broadcasters
have more discretion over which games to show, uncertainty of outcome
increases the likelihood of a game being shown live. In the latter case it is
shown that viewers prefer eventful contests with a result rather than UO per se
and certainly are not interested in “tame draws.” Under such circumstances, it
can be suggested that broadcasters and their audiences appear to be much
more interested in exciting close games, probably between high-level compe-
titors. Both of these studies point to the strong influence of UO on broadcast
demand for sport, in contrast to the relatively mixed results for attendance
demand. It is perhaps not surprising that in a closed league system such as the
US broadcasters will be happier with a more centralized system, but in Europe
broadcasters have also been happier to encourage development of the games to
promote higher-level and more equal competition. In the absence of closed
leagues this has meant producing supra-national competition. In this regard,
UO has become an outcome, rather than an initial driver, of league develop-
ments, as suggested earlier.
10.6 CONCLUSION
In this chapter the demand for professional team sports has been investigated
and shown to involve a modified version of the theoretical framework pre-
sented in Chapter 3. In general, it is argued that the literature has increasingly
come to terms with some conceptual and econometric issues associated with
modelling demand, such as the need to account for capacity constraints and
the endogeneity of key independent variables, such as price, and that a number
of general findings emerge. These include that demand is generally found to be
price and to a lesser extent income, inelastic consistent with a normal good.
Market size is ubiquitously significant along with team quality, the success of
teams, favourable weather, local rivalries, significant matches, and reschedul-
ing games away from traditional times and days. As detailed in Chapter 8,
there is some growing evidence that uncertainty of outcome matters, but the
results are mixed, and the same can be said for habit persistence, but this time
because of problems of interpretation. Finally, much of the evidence suggests
Conclusion 299
that broadcast and attendance demands are substitutes if the effects of sched-
uling are accounted for. The evidence that is available suggests that teams still
benefit, however, because of the revenue differentials. The chapter concludes
by noting that the broadcast industry has had increasing effects on the devel-
opment of sports leagues, made possible because of technical and regulatory
developments in the media market and the incentives that media deals offer to
clubs. It appears that UO matters more with broadcast demand.
Appendix 10.1 Problems of Measuring Habit Persistence

This appendix, although not necessarily representing precisely how particular authors
arrived at their models of demand, illustrates some problems that arise in the interpretation of
habit persistence. Consistent with the discussion in the chapter, in a simple case we might
specify a linear demand for attendance as Equation A10.1.1:
At ¼ a1 þ a2 Pt þ a3 At1 ðA10:1:1Þ
Here, A is the level of home game attendance – perhaps by club by season – P is a measure
(possibly ticket revenue divided by attendance) of price, and the subscripts are date
observations. The parameters a1, a2 (a2 < 0) and a3 (0 < a3 < 1) are to be estimated by
some regression procedure. It follows that, by analogy, a set of equations associated with
previous time periods can be written as indicated in Equations A10.1.2–10.1.4.
At1 ¼ a1 þ a2 Pt1 þ a3 At2 ðA10:1:2Þ
At2 ¼ a1 þ a2 Pt2 þ a3 At3 ðA10:1:3Þ
AtN ¼ a1 þ a2 PtN þ a3 AtN1 ðA10:1:4Þ
Based on these equations, successive backward substitution can be used to remove past
attendance levels from the right-hand side (hereinafter RHS) of Equation A10.1.1 in order to
discover what actually drives attendance other than the current price. This leaves:
At ¼ a1 þ a2 Pt þ a3 At1 ¼ a1 þ a2 Pt þ a3 ða1 þ a2 Pt1 þ a3 At2 Þ

¼ a1 ð1 þ a3 Þ þ a2 ðPt þ a3 Pt1 Þ þ a23 At2 ðA10:1:5Þ
which, after further substitutions, resolves into:
At þ a1 ð1 þ a3 þ a23 . . . þ aN1
3 Þ þ a2 ðPt þ a3 Pt1 þ a23 Pt2 . . . þ aN1
3 P tNþ1 Þ
þ aN
3 AtN ðA10:1:6Þ
Equation A10.1.6 states that today's attendance is a geometrically weighted distributed lag
function of price. Unfortunately, Equation A10.1.6 is not consistent with habit persistence,
and the very last RHS term shows why. Since, as N goes to infinity, aN3 goes to zero, past
attendance (AtN) plays no role whatsoever in determining current attendance, At. In a
supposed habit persistence model past attendance insignificant! This is actually a partial
adjustment model with a3 determining the speed with which attendance responds to price
change. Downward and Dawson (2000) indicate how habit persistence can be introduced
into the model.
CHAPTER 11
The Labour Market in

Professional Team Sports
OBJECTIVES
& To understand the institutional structure of the players labour market in

professional sports
& To appreciate profit and win maximizing models of wage and employment
determination
& To become aware of the main results of research into the labour market
11.1 INTRODUCTION
The discussions in Chapters 7 and 9 have suggested that labour plays a
paramount role in the supply of sports. Ultimately, the exercise of labour
creates the sports contest. This chapter explores the labour market in more
detail. In the next section, the labour market in a perfectly competitive
market is briefly re-presented. This will help to indicate conceptually the
novelty of the sports labour market, and to indicate what, in economic
terms, an efficient or theoretically justifiable wage rate would be. Section
11.3 reviews the historic institutional forms of labour market in the US and
Europe, arguing that sports leagues' monopoly power has endowed teams
with monopsonistic power in sports labour markets, but that this situation
has now changed. Some descriptive evidence to support predictions associ-
ated with the labour market changes are presented. Section 11.5 then
examines the formal econometric evidence associated with the labour mar-
ket in the US and Europe. Section 11.6 briefly reviews the market for
coaches and managers in professional sports, before conclusions follow in
the final section. 301
302 CHAPTER 11 The Labour Market in Professional Team Sports
11.2 PERFECT COMPETITION AS A BENCHMARK LABOUR

MARKET
The model of perfect competition is a useful benchmark for the economic
analysis of sports labour markets for two reasons. The first is that it provides a
conceptual basis for deriving a “just wage,” i.e., one that players deserve based
on their productivity. The second is because it highlights some peculiarities
about the labour market in sport. The first claim was presented in Chapter 1,
where it was argued that teams would hire labour up to where the wage was
equal to the marginal revenue product of labour (MRPL), which is the revenue
earned for the firm from the last unit produced by the worker being sold at the
market price.1
Now, in general terms, these conditions for maximizing profits apply to
all factors of production. Appendix 11.1 illustrates this for the case in which
two factors of production are concerned, labour, “L” and capital, “K,” the
latter could represent the stadium and related land, etc. Each factor of
production will be employed up to the point at which its marginal revenue
product (MRP) equals its marginal factor cost or price, i.e., the wage rate “w”
for labour and the profit rate “r” for capital. An interesting possibility arises
from this discussion. By direct analogy with the theory of consumer
demand, if the price of labour rises, i.e., “w” increases, it follows that
profit-maximizing firms will reduce their demand for labour and increase
it for capital, that if they have an incentive to substitute capital for labour,
and vice versa if the cost of capital, “r” rises.2 It is here that an interesting
difference between sports and other industries arises. Because labour ulti-
mately produces the sports contest, it is not possible for teams to substitute
capital for labour if the wage rates increase.3 This is simply a “technological”
restriction of the production of sport. In this respect in sports, both the long-
run and short-run decisions of firms are effectively the same. Under such
circumstances there would appear to be strong incentives to constrain wages
in sports leagues, which suggests a less benign reason than the motive of
1
It is worth noting that, in general, the price at which output is sold is not equal to MR.
Thus, in general, the marginal product of labour multiplied by the price of the firm's product,
known as the value of labour's marginal product (VMP), is not equal to the MRP. For
example, the VMP of a monopoly would be greater than the MRP of a monopoly.
2
In general, investment also tends to bring technological advance with it. These predictions
are thus made assuming technology remains constant.
3
This is not to say that capital does not matter. In sports such as motor racing, yachting,
cycling and other capital-intensive sports, investment in capital does matter to a degree,
although labour is still the key resource.
Historic Evolution of the Labour Market in Sport 303
UO to explain polices aimed at the labour market. The historic institutional

nature of the labour market is now discussed.
11.3 HISTORIC EVOLUTION OF THE LABOUR MARKET IN

SPORT
11.3.1 The United States
As Szymanski and Zimbalist (2005) note, after the foundation of the Na-
tional League (NL) in baseball in 1876, it soon became customary to com-
pensate the owners of other teams for their “reserve” rights over their
players, because of the upward pressure that trading players had on salaries.
Indeed, as the NL became more dominant, the reserve clause was estab-
lished in 1879. This clause gave the club power to rollover the player's
contract, by having them re-sign within a one year period. This “anti-
competitive” practice was upheld in 1922 by the US Supreme Court in
the light of challenges to the policy. The court ruled that baseball was
exempt from anti-trust legislation because it did not comprise interstate
commerce. As they developed, other US professional team sports adopted
similar reserve option policies, although they did not obtain legal exemp-
tion. It is probably fair to say that, initially, team owner profits were a key
consideration in this policy, but as Fort and Quirk (1995) note:
“Over time as the reserve clause faced court challenges, owners of sports
teams developed the argument that, whatever the consequences of the
reserve clause on players' salaries, it was needed to preserve competitive
balance. Owners argued that free agency would allow the richest teams
to acquire a disproportionate share of the playing talent in the league.
Competitive balance would be destroyed, driving weaker franchises out
of business.” (Fort and Quirk, 1995 p. 1174)
Since the 1970s there has been a much freer labour market. Impetus for
these changes began with the 1970 Flood versus Kuhn case in the US Supreme
Court, in which Curtis Flood, a star player with the St Louis Cardinals, was
traded to the Philadelphia Phillies in 1969, but he refused to play, wanting
instead to be considered by other teams. The US Supreme Court ruled that the
case was subject to the 1922 ruling.4 Nonetheless, in 1970, a forceful baseball
players' union negotiated a collective bargaining agreement that brought into
existence an impartial arbitrator to help sort out contractual grievances. Some
crucial developments followed. By 1973 final offer arbitration (FOA) for salary
4
Earlier failed challenges to the ruling exist in the 1950s, but the courts were not entirely
unsympathetic and prevented the anti-trust exemption carrying over to football.
TABLE 11.1 US sport salaries ($)
Sport 2000 2007
National Football League

& Highest & 619 050 & 1 102 880
& Lowest & 361 200 & 440 520
% spread & 71.4% & 150.4%
Major League Baseball

& Highest & 2 225 000 & 3 591 667
& Lowest & 268 500 & 380 000
% spread & 729% & 845%
National Hockey League

& Highest & 1 600 000 & 2200 000
& Lowest & 500 000 & 800 000
& % spread & 220% 175%
National Basketball Association 2001 2006
& Highest & 4 475 688 & 5 215 000
& Lowest & 590 850 & 1 347 000
& % spread & 657% & 287%
Source: http://content.usatoday.com/sports/
grievances was instituted.5 Furthermore, in 1975 two star pitchers, Andy

Messersmith of the Los Angeles Dodgers and Dave McNally of the Montreal
Expos, refused to sign contracts, played for a year without contracts for their
clubs and at the conclusion argued they had worked off their clubs' reserve
option periods, and asked to be regarded and treated as free agents. Crucially,
the arbitrator, Peter Seitz, favoured the players and was consequently sacked by
the team owners. However, the courts refused to reverse the decision. The
outcome of the Messersmith–McNally judgement was to give the club an
exclusive six-year option on the player's services, after which time they auto-
matically become a free agent.
As Sanderson and Siegfried (1997) note, the move back towards free agency
in US sports has culminated in a situation in which “rookie” players of one or
two years' experience are subject to a reserve clause. Intermediate players of
between three and six years experience are eligible for FOA and veterans of
seven or more years' service are free agents or eligible for FOA.
With regard to other policies, Staudohar (1998) notes that the NBA was the
first league to reinstigate a salary cap in 1984 and the National Football League
5
In final offer arbitration, the two sides to a dispute, the players and the team, make salary
offers and then the arbitrator picks one of the two offers as the settlement. This is distinct
from conventional arbitration settlements in which the arbitrator is free to impose any
settlement they deem fit.
initiated a cap in 1994. The impetus came from the asserted financial diffi-
culties experienced by clubs following the rise in player's salaries after the
move to free agency. Indeed, the increases in salary have been quite extraor-
dinary. Sanderson and Siegfried (1997) note that average baseball salaries rose
by approximately 730% between 1975 and 1985; between 1967 and 1977
basketball salaries rose by approximately 615% and football salaries grew by
110%; and between 1977 and 1987 ice hockey salaries grew by approximately
180%. More recent data is presented in Table 11.1, for the median annual
salary. It should be noted that the dispersion in salaries has increased in both
football and baseball.
11.3.2 The UK and Europe

As noted in Chapter 10, in much the same way as baseball has tended to
dominate both academic and general sporting discussion in US sports, this has
been the case with soccer in European sport. However, this is not to say that
other professional team sports such as rugby league and rugby union do not
have active labour markets. Soccer provides the most detailed exemplar, how-
ever, and has driven European sports labour markets. There are strong paral-
lels with the development of US sporting labour markets in which labour
market restrictions have applied to both player remuneration and player mo-
bility. As Szymanski and Kuypers (1999) argue:
“At its foundation in 1888 the Football League expressly set out the twin aims
of imposing a maximum wage and preventing the movement of a player
from one club to another without permission of the former.” (Szymanski
and Kuypers, 1999 p. 99)
However, the post-war boom in interest for soccer, with attendance figures
at their all-time high (Dobson and Goddard, 2001), coincided with increased
player and union militancy. In 1960, Jimmy Hill coordinated a campaign that
led to the abolition of the maximum wage. Player mobility was controlled by
the “retain and transfer” system. Only a player registered with the Football
Association (FA) can play professional football. Because the registration is held
by a club, historically, it could control the player's movements much in the
same way as the reserve option clause in US sports. At the end of a season, for
example, a club could retain players if it wished or let them leave. In principle it
could retain the registration of a player, even if it did not renew the contract.
Moreover, clubs could charge a fee – a transfer fee – for allowing the player to
move to another club. Note that this could apply even in the absence of the
maximum wage so, effectively, the terms and conditions of the players' con-
tracts lay with the club under this system. The first effective challenge to the
transfer system came in 1963 when George Eastham took his club, Newcastle
United, to court for refusing to let him leave the club on a transfer. The courts
upheld his claim.
“From then on, the club holding the registration had to offer a new contract at
least as remunerative and of the same duration as the expired contract . . . in
order to retain the player's registration; if such a contract was not
forthcoming, the player became a free agent.” (Dobson and Goddard, 1998
p. 776)
Moreover, in 1977 players were awarded the right to decide on a move at the
end of their contracts. However, if the club wanted to retain the player or
demand a fee, the player could go to arbitration to the Football League Appeals
Committee. As this increasing “freedom to contract” occurred, not surpris-
ingly, salaries and transfer fees escalated rapidly as larger clubs competed to
sign talent. Nonetheless, in the latter case there was a desire to keep transfer
fees in the game because of their purported revenue-generating ability for
smaller clubs (Dobson and Goddard, 1998, p. 777). It is also worth noting
that while the case of the UK has been discussed, the general thrust of insti-
tutional developments has been similar in other countries.
“In other countries similar though not identical arrangements apply. Thus,
under the rules of the Union of European Football Associations (UEFA) a
board of experts makes a binding judgement in the case of a disputed fee . . .
but differences remained. Thus, within France a transfer fee is payable only
in the case of a player's first change of club and within Spain players aged 25
or more can transfer freely without a fee being required.” (Sloane and
Campbell, 1997 p. 2–3)
It is perhaps for this reason that the Bosman ruling has received much
attention although it was, in many respects, a relatively minor legislative
change to the changes in the system that had already occurred. The ruling
arose because Jean Marc Bosman took two clubs (RC Liege, his current club,
and US Dunkerque, the club which sought to buy him) to the European Court
of Justice under Article 177 of the Treaty of Rome, which enshrines the free
mobility of land, labour and capital in the European Union, for damaging his
employment opportunities by fixing a transfer fee (over which they could not
agree). The court declared that, in the absence of pressing reasons of public
interest, the transfer rules did constitute an obstacle to the free movement of
workers. The substantive outcome of the Bosman ruling was that no fee could
be expected by clubs on the transfer of an out-of-contract player.
A new transfer system was agreed in 2001 which specifies two “transfer
windows” for the purchase and sale of players in which players are allowed to
move, but only once per season unless they are on loan. Unilateral breaches of
contract can take place at the end of the season, but require the payment of a fee
so as to protect the interests of the training clubs (Dimitrakopoulos, 2006). In

addition, for players less than 24 years' of age compensation needs to be paid for
transfers, with players under the age of 18 requiring agreed conditions and
distribution of the fees between clubs. Players over 24 years of age become free
agents at the end of their contract. A player can move in the final year of a four-
or five-year contract if they hand in a request to the club no more than 15 days
after the end of the previous season (FIFA, 2005). Under these circumstances
the club would be entitled to compensation. In this way the labour market has
become structured according to experience, as with the US. The Bosman ruling
has been consolidated by the Kolpack ruling in 2003, which has essentially
made the Bosman ruling international for European clubs trading with coun-
tries with which the EU has association agreements, and it has also had impacts
in cricket and both rugby codes. A European Union association agreement
involves a treaty with a non-EU country to promote cooperation, trade and
other links.
Further liberation of the players' labour market is expected with the Web-
ster ruling. The Webster ruling is a test case in soccer law involving Andy
Webster, a former player with Heart of Midlothian football club in Edinburgh,
Scotland. In September 2006 he became the first player to exploit the updated
transfer regulations of FIFA (Article 17) that players who sign a contract when
aged 28 or under are able to terminate those contracts after three years, and
can do so in two years if aged 28 or over. Compensation should be paid, but
clearly the player can now decide where they wish to play even within the
contract period. The ruling was initiated in response to fears that the EU
would abolish the transfer system altogether. The upshot of this increasing
liberalization has meant that player wages and also transfer fees have in-
creased, as in the US. It is argued that by 1978, more freedom of contract:
“... led to a wage explosion in the late 1970s and early 1980s. Between 1977
and 1983, wage expenditure by clubs trebled in the first, third and fourth
divisions, while it more than doubled in the second.” (Szymanski and
Kuypers, 1999 p. 95)
This led to much concern over the financing of soccer in the UK with the
1983 Chester report examining the growing indebtedness of clubs. More
recently many other sports such as rugby union and rugby league in the UK,
and rugby league in Australia have introduced salary caps. Soccer is an excep-
tion, which is probably connected with the difficulties of enforcement that a
genuinely international labour market brings, as discussed in Chapter 9.
Notably, for Premier League clubs, the range of wage bill increases for clubs
between 1996 and 2001 was 54% to 476%, and between 2001 and 2005 ranged
from a cut of 33% and an increase of 319%. Over the same periods, for Division
TABLE 11.2 Premiership transfer spending
Season English Overseas Total Overseas/

£ million £ million £ million English %
1992–1993 51 n.a. 51 n.a.

1993–1994 67 n.a. 67 n.a.
1994–1995 84 29 113 35
1995–1996 94 69 163 73
1996–1997 95 84 179 88
1997–1998 95 118 213 114
1998–1999 136 142 279 104
1999–1990 105 150 255 143
2000–2001 135 229 364 170
2001–2002 118 195 323 152
2002–2003 86 101 187 117
2003–2004 132 254 386 192
2004–2005 108 232 340 215
2005–2006 134 301 435 225
Source: Deloitte (2007).

n.a. = not available.
1 of the EFL, the range was 21% to 1164% and a cut of 71% to an increase of
175% (see Deloitte, various years). The data reveal not only sustained and large
growth, but also increasing dispersion of growth, with teams facing relegation
having to curtail expenditure absolutely, and teams continuing to maintain
their positions or looking to get promoted having to increase expenditures.
A direct consequence of the Bosman and subsequent rulings has been the
growth in the market in overseas players, particularly in the Premier League
which, as noted in Chapter 7, is the richest European league. As a result of such
trade, FIFA have recommended a “six-plus-five” rule for leagues, with home
country teams needing to field at least six domestic players, but such a restric-
tion would be contrary to European law.6 Table 11.2 provides some compar-
ative data in which transfer spending on English and overseas players is
charted.
11.3.3 Player agents

A potentially important element of the players' labour market that has arisen
with market liberalization in both the US and Europe has been the growth in
6
http://www.dailymail.co.uk/te95xtbased/sport/football/article-1023043/FA-support-FIFA-
rules-stop-foreign-players-dominating-Premi95er-League-bid-strengthen-English-game.
html.
the role of player agents. Semens (2008) argues that in the US agents have been
in operation since the 1920s (see also Shropshire and Davis, 2003), but
became more prevalent in the 1960s (Joyce, 1997; Mason and Slack, 2001,
2003; Mason and Duquette, 2005; Shropshire and Davis, 2003; Karcher,
2006). In the US, as part of the process of labour market changes, agents have
become complementary to player unions in becoming the main vehicle for
salary negotiations (Shropshire and Davis, 2003, p. 11) and tend to act solely
for the player (Holt et al., 2006, p. 5).
In the UK and Europe, Semens (2008) argues that agents have also been in
evidence since soccer became professional, but traditionally clubs internalized
the activity through scouting networks (Roderick, 2001). Moreover, in 1936
FIFA banned the use of agents in transfers, over concerns that they promoted
illegal moves (Berlin Congress, August 1936, displayed in FIFA collection,
NFM). As with the US, it is argued that agents became more prevalent in
the 1960s, particularly in Italy in the 1970s and 1980s when clubs wished to
recruit talent from eastern Europe and employed a middleman to “facilitate”
the deals, overcoming political and legal barriers (Holt et al., 2006, p. 4). Asser
(2005) also notes that the growth of revenues from sport, for example the
broadcasting income noted in Chapter 10, and the Bosman ruling promoted
the widespread use of agents and the widening of their roles to exploit
the playing and commercial opportunities arising from the expanded inter-
national market.
The existence of player agents may also help to explain why transfer
fees have persisted. Under a more liberal labour market one might expect
that players benefit at the expense of clubs, through their wage increases,
but a fee being paid means that some revenue is still directed to the selling
club, as noted earlier. As agents receive a fee based on a percentage of the
transfer, and as Semens (2008) notes, it could be argued that buying clubs
are only interested in the total cost of the player, including wages and fees,
this suggests that the payment of fees remain in the interests of selling
clubs and agents, who potentially benefit at the expense of players. For
example, agents might have an incentive to move a player while they are
within contract to obtain a proportion of the fee. In essence there is a three-
way principal–agent relationship.
11.3.4 Contract duration

The above discussion suggests that contract durations have become a sig-
nificant feature of players' labour markets. Under the reserve clause or
traditional soccer contract, clubs faced no potential financial loss from
losing a player to a rival bid for the player's services from another team.
In essence, there was a long-term contract be-

tween the club and the player – although a par-
ticular contract might be nominally specified for
a set number of years. However, in moving to-
wards a more competitive environment, nomi-
nal contract lengths assume a real economic
significance.
FIGURE 11.1 Labour market structures Commensurate with the risks of injury and
loss of earnings faced by players, clubs face the real
economic risk of losing their better players with no
recourse to compensation in a free market. In the aftermath of the Bosman
ruling, in a context of relative aversion to risk, it would be expected that player
contract lengths would generally increase as a form of insurance, and also to
attract fees, should a player wish to leave a club. Once again, there is descrip-
tive evidence to support this claim according to Szymanski and Kuypers (1999)
and Simmons (1997). However, the more recent changes in the labour market
will have limited this incentive.
11.4 THEORIZING THE PLAYERS' LABOUR MARKET

The above discussion has argued that a growth in the freedom of contract
for players in sports leagues has led to greater player salaries, a greater
dispersion of salaries and also a growth in the total cost of players if
transfer fees are included, as well as an influx in overseas players to leagues
with an open labour market, such as Europe. Explaining such develop-
ments can be made with reference to a variety of different theories of the
labour market.
Figure 11.1 presents a taxonomy of labour market structures, which
essentially maintains that firms, in this case teams, are profit maximizing.
A win maximizing approach is discussed later. Each axis measures power
and, as drawn, suggests that low power of clubs and players results in
perfect competition, as discussed in Section 11.2. Moving horizontally
the diagram indicates increasing club power with player power remaining
constant. This could imply that the number of clubs is smaller than in
the competitive case or that particular clubs are perceived to be so good
that players want to play for the club. At a rudimentary level, there will
always be more players than clubs. It is clear that these assumptions are
closer to sporting leagues than the perfectly competitive assumptions. The
limiting theoretical case of this scenario is a monopsony, which has
been used to describe the historical position of sports labour markets in
Theorizing the Players' Labour Market 311
FIGURE 11.2 Wages and employment under monopsony (W2, E2) and competition
(W0, E0).
their historic context as described above. This labour market leads to

the monopsonistic exploitation of players.7 This exploitation arises
because players will receive salaries below their MRP, as is now illustrated.
11.4.1 Monopsony
The monopsony model differs from the perfectly competitive model inasmuch
as the MC of hiring another employee comprises the wage received by that
person plus the extra wages paid to those who were already on the payroll. This
drives a wedge between the MRPL and the wage. The intuition is that the
higher wages needed to attract additional workers are also paid to other
employees. Appendix 11.2 presents the result more formally.
Figure 11.2 illustrates the comparison between monopsony and compe-
tition, and illustrates the associated equilibrium levels of the wage rate and
7
This is a general case of exploitation. Other forms, such as due to race, would then be
additional extensions of this general case.
employment, defined as hours of labour purchased per week. A1A2 is the

perfectly competitive firm's demand for labour, MRPL, defined over hours
of work, while A1A3 is the corresponding MRPL of the monopsonist. B1B2
is the weekly supply of hours to price-taking firms in the labour market.
B1B3 is the monopsonist's marginal cost of labour (MCL). The perfect
competitor maximizes profit where B1B2 intersects A1A2, giving W0, E0
as the perfectly competitive equilibrium wage–hours combination. The
monopsony firm maximizes its profit where B1B3 (MCL) intersects A1A3,
but the monopsony equilibrium wage–hours combination is W2, E2. Note
that under monopsony the MRPL (given by A1A3) exceeds W2, signifying
exploitation of labour. The general inference is that wages and hours, and
(implicitly) numbers in employment, will be lower (other things being
equal) where the firms exercise increasing power to determine prices. It
naturally follows that as club power, monopsony, diminishes, wages will
increase, with corresponding increases in employment. Moreover, it is
to be expected that with payment commensurate with productivity wages
would increase faster for segments of the labour market associated
with higher skills. This, of course, is what the evidence discussed above
suggests.
The adjustment to wages can be viewed as players recapturing economic
rents. Consider Figure 11.2 once more. A perfectly competitive firm operating
at equilibrium pays a wage W0 equal to the competitive MRP and hires
E0 hours of labour at that wage. But the competitive firm's labour supply curve
B1B2 indicates that most of these hours could have been purchased at lower
wage rates. It is important to distinguish between the actual wage paid (in
equilibrium W0) and the “reservation wage” – the wage that is just enough to
persuade somebody to supply that particular hour, read in Figure 11.2 as the
height of the labour supply curve B1B2 at the hour in question. If the equilib-
rium wage rate W0 is slightly reduced the marginal hour will not be supplied,
but intra-marginal hours will continue to be supplied because a small reduc-
tion will leave the actual wage in excess of most intra-marginal workers'
reservation wages. An earnings equivalent corresponds to the reservation
wage; wage rates are usually measured by the hour, the ensuing earnings are
usually measured on a weekly, monthly or annual basis. A person's “transfer
earnings” is the level of earnings below which they quit the payroll. Earnings
over and above transfer earnings (actual wages in excess of reservation wages)
provide their recipients with “economic rent,” which like consumer surplus is
a pure welfare gain.
With the growth of player power one might expect increasing wages to lie
somewhere between W2 and MRPL, depending on relative bargaining power.
This possibility is now discussed.
Theorizing the Players' Labour Market 313
11.4.2 Monopoly
The predictions connected with wage increases etcetera from an initial
monopsony position could also be explained with reference to other labour
market structures than a simple adjustment towards the competitive case.
Moving diagonally up from the monopsonistic case in Figure 11.1 indicates
increasing player power with club power remaining low. Here it is assumed
that particular players have unique talents and are monopolies as there is no
effective substitute for them. In this case competitive pressure applies to clubs
wishing to hire the player. One can think of many examples of sporting
monopolies such as: Mark McGuire; Deion Sanders; Jerry Rice; Michael
Jordan; Shak O'Neal; David Beckham; Christiano Ronaldo; Daniel Carter.
In the case of sporting monopolies, as Neale (1964) points out, the supply
curve of the athlete is vertical and, as a consequence, wages are demand
driven. This is consistent with the superstar model proposed by Rosen
(1981), where the ability for players to attain monopoly power is related to
the notion of scarcity.
11.4.3 Bargaining
The final quadrant of Figure 11.1 refers to the case of bilateral monopoly.
Here, both clubs and players have market power and under such condi-
tions it is clear that decisions would be made paying explicit attention to
the likely responses of other parties. As discussed in Chapter 5, this
suggests a game theoretic model. As “games” can involve conflict or
non-conflict this requires some deliberation. In a sporting context one
can see that signing a new player to strengthen a team may not produce
conflict between clubs and players. However, if the new player is seen
as a replacement player, then the context is different – conflicts of
interest arise. Likewise, games can be characterized as cooperative or
non-cooperative. Cooperative games imply that agents can make binding
commitments on one another. Non-cooperative games do not. It is clear
that in the former case players' labour markets, which are governed by the
law of contract, have this characteristic. Game theoretical models can also
account for differences in the information possessed by the parties to a
deal. Thus, it can be assumed that information is symmetrically distrib-
uted. In contrast, it is possible that games might be characterized by
asymmetric information, in which one party to a deal has more informa-
tion than the other. In the former case it is likely that, other things being
equal, an experienced player will know their true ability relative to the
competition. Moreover, because of the scrutiny of sports players it is also
likely that teams will have an accurate perception of established players'
abilities. Under such circumstances it seems plausible to think of the

player's contractual conditions as being characterized by symmetric infor-
mation. However, in the case of a relatively new talent, the situation is
clearly different. Finally, an appropriate solution concept needs to be
identified.
As indicated in Chapter 5, a common approach adopted in cooperative
games is to invoke the Nash bargaining solution. The assumption under-
pinning this solution is that the bargained outcome will maximize the
product of the incremental utilities of all parties. Basically, this solution
makes the value judgement that parties to a bargain will settle on a solution
that maximizes overall utility rather than that of a particular individual.
Thus, in the case of monopsony we can note that the difference between the
salary paid to a player and the revenue received by the club, in the form of the
MRP of the player, represents an increment of income that could be reallo-
cated to the player. Clearly there is a disagreement point based on the
current level of player's pay from which bargains can be made. As shown
in Appendix 11.3, in general the split of incremental income will depend on
the form of the utility function. Nonetheless, the model predicts that the
new agreed wage will lie somewhere between the player's MRP and their
monopsony wage.
11.4.4 Non-profit-maximizing behaviour

The above approaches maintain the (implicit) assumption that teams max-
imize profit which, as discussed in Chapter 7, is disputed for European
sports. As discussed in Chapter 9, Kesenne in various publications
addressed the behaviour of a club that seeks to maximize its wins subject
to recovering a predetermined level of cost (k) that conceivably might incor-
porate an element of profit. The conclusion he reached is that the club will
pay a wage equal to its average net revenue (ANR), and that since ANR will
exceed MRP in the neighbourhood of win-maximizing equilibrium, the club
will pay wages in excess of MRP. By implication, win-maximizing clubs
systematically overpay their players, with potentially inimical implications
for the game's survival. The origin of this conclusion is presented in Appen-
dix 11.4. However, it suggests that win maximizers are expected to pay
higher wages and to hire more talent than their profit-maximizing equiva-
lents, which Kesenne argues might explain some of the financial problems
in European soccer. From this perspective it can be argued that soccer in
Europe has evolved from a set of profit-maximizing national monopsony
leagues into a set of win-maximizing leagues that operate in a (largely)
unified talent market.
The Players' Labour Market: Some Econometric Findings 315
11.5 THE PLAYERS' LABOUR MARKET: SOME

ECONOMETRIC FINDINGS
This section examines a selection of econometric studies that have aimed to
test for the historic degree of monopsonistic exploitation in more detail,
whether this has reduced following changes to the labour market and whether
a particular labour market structure fits the evidence best. The potential stock
of econometric studies is large so the aim of this section is to cover seminal
work, and that which illustrates some of the more recent main findings and
methods of investigation. Because the availability of data is different in the US
and elsewhere, the discussion is divided into consideration of the US (typically
baseball) and Europe (typically soccer). It should be noted throughout the
discussion that player characteristics and performance are often central to
the research. One can think of the characteristics as indicative of the human
capital of the player (as discussed in Chapter 3) that governs their productivity.
In the European literature, team characteristics are also measured which, it
will be shown, refer to the relative powers of the teams in terms of controlling
player contracts and availability. In the US literature, the tendency is to focus
on measuring the labour market structure more directly.
11.5.1 The United States

11.5.1.1 The reserve option
The seminal contribution to the econometric analysis of the sports labour
market is provided by Scully (1974), who analyzed the 1968 and 1969 baseball
seasons, predating both arbitration and the movement towards free agency.
Because data on individual player performances and salaries is freely available
in the US, Scully's objective was to estimate the MRPs of players, compare
these to the players' annual salaries, and determine from that comparison the
extent of monopsonistic exploitation suggested by the reserve option clause.
Scully specified a causal chain (recursive) model of two structural equations in
the first of which player inputs influence win percent, i.e., through their
productivity (measured by the team “slugging average” and team's “strike-
to-walk ratio” for hitters and pitchers). In the second equation, win percent
influences revenue by attracting fans.
A recursive model suggests that the values of the dependent variables in a set of
equations are not simltaneously determined, but that the values of one variable
subsequently feed into determining the other variable.
Coefficients were estimated by OLS. By multiplying the coefficient mea-

suring the impact of player performance on win percents with that measuring
the impact of win percents on revenues, MRPs were obtained to be compared
to salaries. It was suggested that players tended to be paid about 10% to 15% of
their gross MRPs. Medoff (1976) argued that Scully's specification of a recur-
sive model was problematic, because it did not address the possibility of
simultaneity between win percents and revenues in the two equations. Using
2SLS to address this possibility it was suggested that players received around
30% to 50% of their MRPs.
MacDonald and Reynolds (1994) researched the period during which FOA
and free agency were both operating, using data for 1986–1987. The authors'
findings suggested that MRPs have no significant influence on the salaries of
those ineligible for FOA, but that they significantly affect the salaries of FOA
eligibles and are most important in determining the salaries of players eligible
for free agency. MacDonald and Reynolds also tested for superstar effects in
baseball salaries, i.e., the presence of the monopoly power of players is such
that as a performer's ability increases so does their salary, but at an increasing
rate. For both batters and pitchers superstar effects were identified.
Marburger (1994) modelled baseball salaries via a one-step process that
takes account of the bargaining structure and dispenses with the estimation of
MRPs, and reached very similar conclusions. Marburger estimated expected
1992 salary levels for eight years of a player's career and identified that attain-
ing FOA status added about 50% to a batter's expected salary and 30% to a
pitcher's. Secondly, between attaining FOA and becoming eligible for free
agency, the player's expected salary increased to the free agent level.
11.5.1.2 Contract length and salaries

Kahn (1993) examined the impacts of free agency and FOA eligibility on MLB
salaries, contract duration and total compensation (defined as duration times
salary) in a reduced-form study derived from a structural model. Data on 1144
non-pitchers and 831 pitchers active between 1987 and 1990 were used.
Perhaps the most important finding was that free agency has a very powerful
positive effect on players' contract durations, but less on salary, which agrees
with earlier findings. Krautmann and Oppenheimer (2002) postulate a struc-
tural model connecting salary and contract duration. The sample consisted of
272 MLB free agent batters covering the seasons 1990 to 1994. They find that
while better performers tend to enjoy longer contracts and higher salaries,
there is a performance-related trade-off between wages and contract length.
Consequently, the literature for the US suggests that removal of labour
market restrictions has tended to produce higher wages, indicative of a reduc-
tion in monopsonistic exploitation, and an increase in contract duration,
consistent with teams insuring themselves against the loss of quality players.
However, the evidence on which form of labour market structure dominates is
unclear, and is probably linked to differences in player attributes and the
TABLE 11.3 A selection of econometric evidence on US labour markets in sport
Author Context Model, method and estimator Results
Scully (1974) Major league baseball Period 1968–1969 148 salaries Estimated that players earned on
Testing for player observed average about 10–15% of their gross
exploitation OLS regression MRPs
Two regressions were used to Dependent variables
compute MRPs for & win percentage
comparison with batters' and & team revenue
pitchers' salaries Independent variables

& slugging average
& strike-to-walk
& NL dummy
& in contention
& SMSA pop
& age of stadium
& percentage non-white players
The Players' Labour Market: Some Econometric Findings

MacDonald and Major league baseball Period 1986–1987 Young players are paid below their
Reynolds (1994) Testing for player 1300 salaries: all players rostered estimated MRPs
exploitation Aug 31 in both years Experienced players are paid in
Two regressions were used to 2SLS accord with their estimated MRPs
compute MRPs Dependent variables Positive significant coefficients on
Two subsidiary regressions & win percentage the squares of MRP terms support
(for batters and pitchers) & team revenue the superstar hypothesis
testing for superstar effects in Independent variables Suggests that as the level of talent
salaries & team runs increases the salaries of the most
& earned run average talented players increase faster than
& contention dummies their talent
& local pop
& personal local income
& local MLB rival
& arbitration dummies
& free agent dummy
& years played as a proxy for experience
Period 1986–1987
792 batters
572 pitchers
OLS
Dependent variables
317
318
CHAPTER 11 The Labour Market in Professional Team Sports
& batters' salaries

& pitchers' salaries
Independent variables
& MRP
& MRP squared
& FOA eligible
& Free agent eligible
Marburger Major league baseball Period 1991–1992 Expected salaries of young players
(1994) Tests whether pay adjusts 1360 salaries observed increase significantly once they
from rookie to FA level OLS* attain FOA eligibility
smoothly or in steps Dependent variables Thereafter salaries increase
MRPs not directly estimated & hitters' salaries (logs) gradually, attaining parity with free
Salaries directly regressed on & pitchers' salaries (logs) agency after six years
independent variables Independent variables
Data separated prior to & own runs
estimation into: & own home runs
& ineligibles
& career runs
& FOA eligibles
& career home runs
& FA eligibles
& years played
as an alternative to using & own ERA
dummies as independent & innings pitched last season
variables & career innings
& career ERA
*We infer from the fact that

Marburger does not specify the
estimator that it was OLS (the default
estimator)
Maxcy and National basketball Period 1951–2004 In the NHL and NFL at least one
Mondello (2006) association, national football 54 end season measures of competitive measure of competitive balance has
league, national hockey balance in each sport improved with free agency. But in
association NFL it may be the salary cap that is
Tests if competitive balance responsible
has improved following free Not so clear-cut in NBA, which had a
agency soft salary cap
Equations estimated for each OLS after unit root tests* The empirical results are broadly
sport, and for every definition Dependent variables consistent with the invariance
of competitive balance the & ratio of SDWP to its ideal value principle; labour market restrictions
authors include & SRCC of end-season standings principally affect the distribution of
*Unit root tests suggested Independent variables economic rent
the errors are stationary, thus & lagged SDWP variable H–T test rejects the null hypothesis
OLS is a suitable estimation & free agency dummies of no unobserved fixed effects
method & salary cap dummies Chow test rejects the hypothesis that
Kahn (1993) Major league baseball & strike dummies all coefficients are the same for white
Author aims to test for effects & rival league dummies and other non-pitchers
on both salary and contract & expansion dummies No clear pattern of differences
duration of eligibility for FOA Period 1987–1990 Overall arbitration eligibility and free
and of free agency Panel data on 1144 non-pitchers and 831 agency had similar effects on salaries
Two markets – one for each pitchers Free agency increased contract
type of player Hausman-Taylor fixed effects duration
Kahn estimates reduced form estimator
equations for salary and Salaries model
The Players' Labour Market: Some Econometric Findings

contract duration Dependent variables
The H–Testimation method & logs of pitchers' salaries
allows for omitted variable & logs of non-pitchers' salaries
bias Independent variables

Independent variables differ & 3 FOA dummies
between the equations for & experience
pitchers and for non-pitchers & performance data
Using the same set would & team chars
assume pitchers and non- & year contract signed
pitchers are almost perfect Duration model
substitutes Dependent variables
Separate equations for black & pitcher contract durations
and white non-pitchers as & non-pitcher contract durations
opposed to a single dummy Independent variables
variable As above
Obviously not identical for both types of
player
The pitcher's salary and his duration are
determined by one set of independent
variables
319
320

Another set determines those of non-

pitchers
Some independent variables, e.g., FOA,
feature in both reduced forms
Krautmann and Major league baseball: free Period 1990–1994 2SLS results reported below
Oppenheimer agent batters only 272 free agents Positive significant coefficient on
(2002) Structural equation DUR, negative on the interaction
2SLS (and OLS)
estimation term favours the insurance
LHS dependent variables
Main purpose is to detect hypothesis
& log of real salary (1990 = 100)
whether contract duration is Experience not significant – may be
traded against salary RHS dependent variables the result of restricting the study to
OLS and 2SLS are & DUR free agents, which reduces within-
inconsistent in the presence & DUR*PERF sample variation in experience
of unobserved quality Independent variables MINORITY dummy “wrongly”
variations & minority dummy signed and insignificant in 2SLS
By sampling free agents only & team revenue (1990 = 100) results
Krautmann and & experience
Oppenheimer reduce the & ABYEAR
scale of this problem

Hard to say what the
“expected” sign is for the race
dummy
The Players' Labour Market: Some Econometric Findings 321
structured nature of the labour market. This suggests something of a transi-

tion from monopsony, through bargaining and arbitration towards competi-
tive outcomes with free agency or superstar “overshoots” in wages. Table 11.3
presents a summary of the literature reviewed.
11.5.2 Europe
Research into the labour market in European soccer has not followed the same
lines as that in the US, partly because of relative lack of data on salaries and
also because of the existence in Europe of a well-organized transfer market
where players trade for cash, other players and on free terms, unlike the US
where teams trade players either for other players or for prior rights in future
drafts. Consequently, researchers have investigated the transfer market for
evidence of rent-sharing between buying and selling clubs in terms of the fees
paid and received.8
11.5.2.1 Monopsony rent-sharing

Carmichael and Thomas (1993) seek to explain the determination of
transfer fees in English soccer as an outcome of two-sided Nash bargaining
between selling and buying clubs. The authors examine transfers that
occurred between the end of the 1989–1990 and 1990–1991 seasons,
during which 255 transfers were reported, although only 214 cases were
analyzed because of the need to exclude free transfers, transfers involving
payment by means of player exchange, and transfers involving clubs from
other leagues, including overseas ones. The former two cases preclude
bargaining, while the latter lacked data on overseas clubs.9 The coefficients
on player characteristics were correctly signed and statistically significant
at or above the 5% level. In general, buying club characteristics had signif-
icant coefficients, but not selling clubs, which suggested to Carmichael and
Thomas that the upward trend in fees had been buyer-led, consistent with
the occurrence of rent sharing.
8
Frick (2007) presents a very useful summary of labour market research in football, noting
that as well as the issues discussed in this chapter, contract duration can affect the
performance of players, as is consistent with tournament theory, and institutional changes
such as the Bosman ruling have lengthened player career durations suggesting that it has
enhanced the need for players to raise their productivity (see also Frick et al., 2006).
9
Free transfers actually suggest some form of truncation in the data, which would make OLS
problematic. This issue has also affected other studies of the transfer market, as discussed in
the chapter.
Dobson and Gerrard (1999) examine six seasons' of data from the EFL
from 1990–1991 through to 1995–1996, plus the closed season of 1996–
1997. During these years 2215 permanent transfers took place involving
football league teams only. Of these, in 198 cases the fee was not disclosed,
another 98 were exchanges, and there were 432 free or nominal fee trans-
fers and others that were problematic. Part-exchange transfers were includ-
ed, where a cash valuation was published. After exclusions 1350 transfers –
about 60% of the total – could be used in the study. The Bosman ruling led
to an increase in the proportion of free transfers, and thus the proportion of
current transfers that could be used fell as time passed. Most player char-
acteristics appeared to be individually significant. Selling club characteris-
tics appeared to be statistically significant. The most important finding
was that both sets of characteristics were significant, supporting the rent-
sharing hypothesis.10
11.5.2.2 Fees in a competitive labour market

A paper by Carmichael, Forrest and Simmons (1999) notes that about 10% of
players were transferred every year and suggested the possibility that those
who do transfer are not a random sample of soccer professionals. To correct for
selection bias the authors use Heckman's (1979) two-stage estimator. In the
first stage they model the probability that a given player will transfer, using the
information on all players in the sample. In the second stage they use the data
on those who have transferred to model fees, adding a correction term derived
from the residuals from the probability model.
The sample studied was all active professionals (i.e., those who played
in at least one league game) in the football league between May 1993 and
May 1994, and who were still contracted to a club at the end of the season.
Of 2100 players, 59 were deleted for lack of information on player attri-
butes and 11 because their transfers involved exchange of players, leaving
2029, of which 240 transferred within the season.11 The selection model
meant that information on the 42 free transfers in their sample could be
used rather than excluding them, although this represents a censored
distribution of fees from below which would make OLS estimates of the
10
Did the authors go too far in assuming that the amount of rent sharing is constant? Drop
this assumption and the simple test cannot be applied. In an earlier (1997) discussion paper
Dobson and Gerrard had split the sample into three groups (N = 450) graded by price. For
each group they found buying club characteristics to be significant, but the Chow test
rejected the hypothesis that the parameters were the same for each subsample. This runs
counter to a constant rate of rent sharing, but not to rent-sharing as such.
11
It is perhaps ironic that this sample selection issue is not addressed!
Managerial Efficiency 323
fee equation inconsistent. Therefore, in the second stage the authors used
the Tobit estimator.
It is shown that support exists for the authors' hypothesis that unobserved
player characteristics make their possessors more likely to be transferred and
to command higher fees when transferred. Traditional player characteristics
are also significant. In the Tobit equation that corrects for sample selection
bias, variables that capture age, experience, scoring ability and international
status attain at least the 5% level of significance. Inasmuch as the emphasis is
on these characteristics, a competitive market explanation has some support.
A decisive test with a bargaining approach in which buying and selling clubs
characteristics are included is not possible because one cannot include the
former for players that have not transferred. Nonetheless, the proliferation of
players' agents, as discussed earlier, suggests that some form of market imper-
fection remains.
11.5.2.3 Salary determination in football

With the emerging availability of salary data a number of studies assessing
salary determination have been published. Lucifora and Simmons (2003)
model individual salaries in Italian football, Serie A and Serie B. Salary data
existed for 730 players; the authors were able to obtain sufficiently informative
career histories for 533. Some evidence of superstar effects for forwards is
identified. Frick (2006) reports estimates of player salaries in the first division
of the Bundesliga. The sample contained 1025 players and 2381 player years,
on average about 2.3 years of a player's first division career during the seasons
2001–2002 through 2005–2006. No superstar effects were identified, but
overseas players (controlling for position, age and experience) were generally
paid more than equivalent native Germans, although the only groups paid
significantly more were other Europeans and South Americans. Table 11.4
summarizes the literature reviewed.
11.6 MANAGERIAL EFFICIENCY12

The above discussion has focused primarily on the contribution of playing
talent to the production of team sports, and then analyzed the impact of
regulations on issues such as their remuneration and contract duration. In
12
In the UK and Europe the term “manager” is used, particularly with soccer, to describe
what the US and other sports would typically refer to as the coach. There is clearly increasing
specialization of managerial and coaching roles in professional sport, but the effects of their
organization have not been researched.
324
TABLE 11.4 A selection of econometric evidence on transfers and wages in European sport
Carmichael English soccer Period 1990–1991 Equation accounts for about 80% of
and Thomas Using transfer fees (£) to 214 intra-league transfers, exclusive sample variation in the log of the fee
(1993) investigate monopsony rent of free transfers BCs generally more significant than SCs
sharing OLS Most significant BCs are gate, goal
Dependent variables difference, league position and division
transfer fee (log) Most significant SCs are goal difference,
Independent variables league position and division
& player characteristics (PC) League appearances and age the most
& age squared significant of the PCs
& goals Hints at asymmetrical bargaining power
& tribunal
& position
& buyer characteristics (BC)
& profit
& gate
& goal difference
& league position squared
& division
& seller characteristics (SC)
& as above
Dobson and English soccer Period June 1990–Aug 1996 BCs statistically significant as a group,
Gerrard Modelling real (1990 = 100) 1350 intra-league transfers exclusive favourable to rent sharing
(1999) transfer fees to investigate rent of free transfers Equation accounts for about 79% of the
sharing OLS in-sample variation in the log of the real
Model specifically designed to Dependent variables transfer fee over six years
test for rent sharing & real transfer fee (log) NB: not directly comparable with
Carmichael and Thomas above, which
is for current price fees in a single year
BCs more significant as a group than
SCs (agrees with Carmichael and
Thomas)
Most significant PCs are age, league
appearances, number of previous clubs,
international goals, interaction terms
with goal variables, premium prices for
forwards and defenders
Independent variables Very low value transfers gave trouble.
& PC Better treated as free transfers
& age, agesquared
Overall some commonality with
& previous club
Carmichael and Thomas's findings
& league total, league total squared
& goals (various)
& internat
& pos
& BC
& pos, possq
& lastpos, lastpossq
& goal diff
& gate
& division
& SC as above
& TIME
& month dummies
& football season dummies
Carmichael, English soccer Period May 1993–May 1994 Heckman estimator rejects the absence
Forrest Modelling transfer fees and the 2029 footballers of unobserved quality variations
and Simmons probability of being transferred 240 intra-league transfers including Most significant of the PCs are age,
(1999) simultaneously 42 free transfers league appearances, some goal terms,
A competitive market Heckman; second stage is Tobit goals conceded by defenders,
determines both the midfielders, under-21s
probability of a transfer and the Most significant club variables are
fee loans, previous clubs and change of
Heckman estimator to deal manager
with possible selection bias in l tests significant, favouring the
the fee equation selection bias hypothesis
Managerial Efficiency
325
326

Probit equation estimates Probit 2029 obs Age, appearances, goals, international
transfer probabilities, using Dependent variables status, division are significant
data on all players. The & dummy, 0 or 1 as transfer does/not Of club characteristics, only the division
residuals are used to construct occur is significant
a term (l) to remove sample Independent variables
selection bias in the fee & PC
equation & age, agesquared
The equation to estimate the & league appearances
transfer fees, estimated by & goals
Tobit to permit CFS to include & position
free transfers & international
*Loans and previous clubs may & club characteristics
indicate human capital & division
& goal ratio
& loans*
& previous clubs*
& change of manager*
& promotion risk*
& relegation risk*
& Tobit 240 transfers
Dependent variables
& fee (£)
& all those from the Probit, plus l,
minus the asterisked club

characteristics
Lucifora and Soccer, Serie A and B Period 1995–1996 Positive sign coefficient on square of
Simmons Modelling the logs of player 533 of 730 players forward strike rate (Equation 11.1) and
(2003) salaries using a reduced form OLS with fixed team effects on superstar variables (Equation 11.3)
Testing for superstar earnings Dependent variables broadly consistent with convexity
Is it possible to identify this as & log of salary Club fixed effects could not be rejected
something different from Independent variables
tournament outcomes? & age, agesquared No test for unobserved player effects
*lagged one year & *appearances Age, appearances and forward scoring
**lagged and related to rates all very significant, also
position international status
***career rate lagged one year
****banded career scoring
rates
& **goal rate
& **assist rates
& ***strike rate squared
& *under 21
& *it int
& *other int
& ****superstar1/2
Frick (2006) Bundesliga Period 2001–2002 to 2005–2006 Hausman test rejects suitability of OLS
Modelling logs of salaries 2381 player years (1025 players) as an estimator
Author also considers the wage Fixed (club) effects No evidence of any convexity in the
bill performance nexus Dependent variables response of salaries to age, career games
Club characteristics are picked & log of salary and international caps
up by the club effects alone Independent variables Tenure a significant positive factor
No club specific indexes are & age (and its square) Native born players, other things being
included & career games (and square) equal, are often paid less than overseas
Goals not included separately, & international caps (and square) colleagues
& 4 position dummies Players from Europe and South America
as the positional dummies
& tenure with current team earn significantly more than native-
identify players that are more
& region of origin dummies born players
or less likely to score
& year dummies
Managerial Efficiency
327
sports, however, the role of the coach/manager is also perceived to be impor-

tant in affecting the productivity of talent, as evidenced by the popular media's
attention to their being hired and fired.
Evaluating a coach/manager's contribution to the team's sporting success is
as important and as difficult as evaluating a player's. In economic terms, the
coach/manager can be assumed to affect the form of the production function
described throughout this book, that is the technology by which labour and
other inputs combine to produce output. One problem is to disentangle this
effect from that of the players' efforts. Another problem is to relate the pro-
ductivity of the coach/manager to either their reward or prospects of retaining
their job.
One possibility is to use the regression residuals from a team production
function as an indicator of technical efficiency. This is the approach under-
taken by Carmichael and Thomas (1995) who estimate a production function
for English rugby league during 1990 and 1991. Teams (and by implication
their coaches/managers) that enjoy higher than expected win percents are
regarded as being more efficient than those that produce lower than expected
win percents on the basis of available resources. A problem with the analysis is
that OLS always produces positive and negative residuals, as defined in Chap-
ter 1. Negative residuals suggest that output is below the maximum possible,
but positive residuals suggest that output can exceed the maximum that the
production function is supposed to determine.
One solution to this technical problem is to model a stochastic frontier
production function, as used by Dawson, Dobson and Gerrard (2000) and
Dawson and Dobson (2002). In the former case this is used to analyze a
sample of 147 coach/managerial observations, covering 72 appointments, to
examine their ability to maximize team performance for a given quality of
playing resources. In the second case this analysis is augmented to include the
possible effects of human capital on 660 coach/managerial observations cov-
ering 318 managerial appointments. Both studies examine English soccer
during the 1992–1993 to 1997–1998 seasons. The general conclusion is that
efficiency has been falling due to increasing demands for success. Dawson and
Dobson (2002) identify that variations in efficiency are linked to the human
capital characteristics of coaches/managers and particularly to experience pri-
or to an appointment.
This raises the important issue that the duration of a coaches/manager's
tenure at a club is likely to be linked to their performance. To this end, Scully
(1994) uses a stochastic production function analysis of coaching efficiency in
basketball, baseball and American football and adds their efficiency to a hazard
model exploring the duration of the coach's tenure. Scully identified that the
expected duration was positively related to efficiency. Audas, Dobson and
Conclusion 329
Goddard (1999) used this approach to examine 918 coaching/managerial

spells between 1972–1973 and 1996–1997 in English soccer and identify that
job security is highly contingent on current team performance, a finding
reinforced in an updated piece of work in Dawson and Dobson (2006) exam-
ining 1100 managerial terminations between 1972–1973 and 2002–2003.
The implication is that, as far as the output of sports teams is concerned,
coaching and managerial inputs matter and in turn team performance is
crucial to the duration of tenure of these inputs. To the extent that this
simultaneous process can affect the trajectory of the impact of playing talent,
it is clear that the previous analysis in this book assumes this complexity away.
11.7 CONCLUSION
Labour is the most important input into professional team sports. The interest
of economists has ranged widely, over wages, monopsonistic exploitation,
transfer fees, discrimination and contract durations. The broad thrust of the
literature suggests that historically endowed monopsonistic exploitation has
been reduced or removed as labour markets have evolved in the light of regu-
latory reform. Superstar effects, competitive markets where players are paid
their MRPs and bargaining effects can be identified from the evidence, suggest-
ing that no one theory describes all aspects of the labour market, but that a
range of ideas can help us to understand the process of their transition, with
bargaining being likely to persist in the presence of active European player
agents, but with competitive outcomes with US athletes moving towards
being free agents. It is also likely that less talented players and younger players
are likely to experience degrees of exploitation, while superstars will be the
exceptions in any sport.
A theoretical conundrum does remain. As with the conceptualization of
sports leagues, either profit or win maximization can be assumed to drive
labour markets. The above discussion broadly presupposes profit maximiza-
tion. If win maximization is dominant then one might expect players to have
always received greater payment than their MRP, as teams hoard talent. Eu-
ropean analysis has not tested this proposition. It could apply if linked to
longer-term profitability problems with European sport that has been exacer-
bated with labour market freedoms for some clubs. Approaching the labour
market from this perspective, though, is also suggestive of a need to examine
how the redistribution of impacts has occurred.
What is clear is that, regardless of profit or win maximization (subject to a
profit constraint), rising player costs affect all teams, but will affect those least
able to raise their revenues most. As discussed in Chapter 7, therefore, this
could help to explain the financial difficulties in large-scale leagues such as

soccer in England, for the lower quality teams and why, as discussed earlier,
salary caps may be perceived to be acceptable for lower drawing sports gener-
ally. The labour market, both in terms of harnessing talent to generate saleable
productivity, i.e., contests for fans, as well as in terms of dominating costs,
thus remains of central importance to sports leagues.
Appendix 11.1 General Conditions for the Demand for Factorsof Production
The general profit-maximizing conditions for perfect competition have a direct corollary in the
general theory of consumer demand presented in Appendix 3.1 It is assumed that the firm
maximizes profits, P, which is the difference between total revenue, R, and total costs, C,
subject to the constraint of the level of output that it can produce according to the production
function. Note that output constrains the firm's decisions because both revenues and costs
depend on output and the price of output and inputs respectively. The latter are given to the
firm in competitive markets. This can be written as:
Maximize P ¼ R C subject to Q ¼ QðL; KÞ ðA11:1:1Þ
as:
R ¼ P Q or PQ ðA11:1:2Þ
and:
C ¼ wL þ rK ðA11:1:3Þ
then substitution for C and Q leaves:
P ¼ PQðL; KÞ C subject to Q ¼ QðL; KÞ ðA11:1:4Þ
First order conditions for a maximum then become:

LP=LL ¼ PLQ=LL w ¼ 0 ðA11:1:5Þ
LP=LK ¼ PLQ=LK r ¼ 0
or:
LP=LL ¼ PLQ=LL ¼ w ðA11:1:6Þ
LP=LK ¼ PLQ=LK ¼ r
which are the conditions where marginal revenue productivities are equal to marginal factor
costs. An important feature of these two FOCs is that they imply:
LQ=LL=LQ=dK ¼ w=r ðA11:1:7Þ

Conclusion 331
That the ratio of marginal products – notice that the constant prices cancel – is equal to the
ratio of the factor costs. This is the equilibrium condition implied in Figure A11.1.
FIGURE A 11.1 Equilibrium demands for labour and capital.
Appendix 11.2 Profit-MaximizingMonopsony

The Monopsonist has to maximize profits as:
Maximize P ¼ TR TC ¼ PðQÞFðL; KÞ wðLÞL rK ðA11:2:1Þ
Partially differentiating Equation A11.2.1, using the product rule for the revenue component
with respect to labour and capital, setting the derivatives to zero and rearranging, we obtain
for marginal revenue productivity of labour (MRPL) and marginal revenue productivity of
capital (MRPK):
MRPL ¼ PðLF=LLÞ þ ðLP=LQÞðLF=LLÞQþðLw=LLÞL ¼ w ðA11:2:2Þ
MRPL ¼ PðLF=LKÞ þ ðLP=LQÞðLF=LKÞQþðLw=LKÞL ¼ r ðA11:2:3Þ
The MRPL under monopsony equals MRPL under perfect competition, the first term on the
left-hand side plus two others. One of these, the last term, is the cost of having to pay a higher
wage rate to the existing labour force when the firm hires extra labour. To the monopsonist,
the marginal cost of labour (MCL) equals the wage rate plus the additional cost of intra-
marginal labour. The other component is the revenue lost on intra-marginal sales when the
price is cut to sell the extra product; (qP/qQ)Q is the rate at which the price falls and Q is
intra-marginal sales. Hence, the whole term corresponds to the change in revenue. Note that
the rate of change is multiplied by (qF/qL) in order to get a measure of the rate at which a
change in L (labour input) affects the product price (P). Other things being equal, the
profit-maximizing monopsonist pays a lower money wage than the profit-maximizing
imperfect competitor.
Appendix 11.3 The Nash BargainingModel

The Nash bargaining solution assumes that the product of the incremental gains to utility is
maximized in striking a bargain. If we define M as the player's MRP, W0 as the player's
monopsonistic wage and W the wage that players seek to achieve above W0, the players
utility, G, will be a function of the wages received and the joint utility function for the team and
the player can be written as:
U ¼ ðM WÞðGðWÞÞ GðW0ÞÞ ðA11:3:1Þ
Differentiating this function with respect to W yields:
LU=LW ¼ MðLG=LWÞ ðLG=LWÞÞ þ GðW0Þ ðA11:3:2Þ
To identify the necessary conditions for a maximum this equation can be set equal to 0.
Rearranging this condition implies that the newly-bargained wage will be equal to:
W ¼ M ½GðWÞ GðW0Þ=ðLG=LWÞ ðA11:3:2Þ
The top line of the term in square brackets will be positive, because W will lie above W0 – or
there would be no bargain – and hence the player's total utility will be higher in the case of W
than W0. The bottom line of the term in square brackets will also be positive, because it is the
marginal utility of the player's wage and this is positive by assumption. This suggests that the
overall term in square brackets is positive. Thus, the newly-bargained wage, W, will lie
somewhere below the marginal revenue product of the player and above the monopsonistic
wage W0.
Appendix 11.4 A Win-Maximizing LabourMarket

The basis of Kesenne's (2002) argument can be illustrated by assuming that the ith club's
total revenue (TR) is a function R(x) of the amount of talent (x) that it employs, with the
property that MR is always positive but that it decreases with talent hired, as presented in
Equation A11.4.1. To simplify the expression the subscript that identifies the ith team is
dropped:
TR ¼ RðxÞ ðA11:4:1Þ
Conclusion 333
To impose Kesenne's restrictions on MR requires MR = qR/qx > 0, and qMR/qx = q2R/

qx2 < 0; i.e., MR is positive but decreases as more talent is added to the team, which the
author argues is consistent with the idea that spectators value UO. The team's total cost (C) is
taken to be a linear function of talent hired plus a fixed cost element (k) that might incorporate
a desired minimum profit component, embodied in Equation A11.4.2, where W is the wage
rate:
TC ¼ Wx þ k ðA11:4:2Þ
If the team's object is to maximize performance (x) subject to the constraint that it recovers
fixed cost k, a constrained maximization (Lagrangean) equation L where x is the maximand
and l (positive) is the Lagrange multiplier can be presented as A11.4.3. Choosing x and l to
maximize function L ensures that the financial constraint (recovery of the club's fixed cost) is
met at the equilibrium value of x:
Minimize L ¼ x þ l½RðxÞ WðxÞx k ðA11:4:3Þ
To find the highest level of talent the club can afford, subject to meeting its cost target,
requires partial differentiation of L with respect to x and l, and setting both derivatives equal
to zero leaves:
LL=Lx ¼ 1 þ lðLR=LxÞ xðLW=LxÞ WðxÞ ¼ 0 ðA11:4:4Þ
LL=Ll ¼ ½RðxÞ WðxÞx k ¼ 0 ðA11:4:5Þ
Equation 11.4.4 may be rearranged (compare Kesenne's 10) to give:
LR=Lx ¼ xðLW=LxÞ þ WðxÞx l1 ðA11:4:6Þ
Given that l is positive, Equation 11.4.4 implies that the MR of talent (qR/qx) is less than the
MC of talent [x(qW/qx) – W(x)] suggesting that cost-constrained win maximizers would hire
more talent in equilibrium than profit maximizers, a result that fits Kesenne's prior
expectations well. Equation 11.4.5 reveals that when qL/ql is zero, the cost constraint holds,
which is precisely why the Lagrange multiplier l is used in simple constrained optimization
problems. The equation may be rearranged into the form below (compare Kesenne's 11) the
interpretation of which is that the equilibrium wage paid per unit of talent equals club net
average revenue ANR. Total net revenue is R(x) minus fixed cost k. Dividing by x gives
average net revenue.
WðxÞ ¼ ½RðxÞ k=x ¼ TNR=x ¼ ANR ðA11:4:7Þ
Under cost-constrained win maximization the wage exceeds that paid by clubs in a profit
maximizing industry. The relevant zone of ANR for cost-recovering win maximizers is always
where ANR>MR, implying that talent tends to be overpaid.
CHAPTER 12

Events and Infrastructure
OBJECTIVES
& To understand the economic rationale for public sector investment in sports
events and infrastructure
& To understand the difference between market and non-market impact of
sports investment
& To understand the difference between economic impacts, the full economic
value and cost–benefit evaluation of sports investment
& To appreciate the main findings of the empirical evidence evaluating public
sector sports investment
& To appreciate the empirical evidence on the provision of sports
& To assess the rationale for public policy on the provision of sport
12.1 INTRODUCTION
In this chapter attention turns away from a specific focus on professional team
sports and returns to consider elements of the economics of sports events more
generally. This is for two main reasons: on the one hand, from the point of
view of policy, as discussed in Chapter 2, a sports “event” has come to repre-
sent a sports contest that is taking place in a context different from a league-
based competition of professional teams, which is less regular and can involve
individuals or teams. This brings with it a need to examine some distinct
elements of economics. On the other hand, as discussed in Chapter 5, one of
the major policy issues connected with sports events is for the public sector to
invest in hosting and/or the infrastructure required for hosting events. How-
ever, the distinction between types of sports contests and the arbitrariness
between team and individual sports are not, of themselves, of major concep-
tual significance. It can be shown that both the economic rationale for, and the 335
336 CHAPTER 12 The Economics of Sports Events and Infrastructure
principles of evaluation of, any public sector investment in sports is common

to all sports contests, although naturally the scale and types of investments
differ.1
To explore these issues in more detail, Section 12.2 revisits the concept of a
sports event discussing the variety of types identified in the literature. It is
noted that, in the absence of professional teams, a central economic problem
with event design is to balance the conflicting incentives required to encourage
participation in an event, but also to promote high level performance which
makes the event attractive, as predicted by contest theory. The role of adverse
incentives, such as cheating (as a form of rent seeking), are discussed.
Section 12.3 then discusses the policy rationale for public sector invest-
ment in sport and, in particular, the role that externalities and other
“spillover” effects are perceived to play in hosting events. The role of adverse
incentives is also discussed. Section 12.4 then focuses in much more detail on
the sports investment decision. Section 12.5 argues that the evidence suggests
that dubious recommendations for investment decisions have often been
made in sport. The chapter concludes by suggesting that, while a coherent
theoretical case can be made for public sector investment in sports, the evi-
dence in support of this case is weak, and suggests considerable caution and
planning is required to harness the spillover effects from the investments.
12.2 SPORTS EVENTS

Depending on the literature or policy organization, the specific definition of a
sports event can vary although, as noted earlier, they tend to be united in
recognizing that events do not directly refer to professional sports leagues.
Many criteria have been taken into account in describing an event but some of
the more common are given in Table 12.1 (see Rahmann et al., 1998; Toohey
and Veal, 2000; Gouguet, 2002; Barget and Gouget, 2007; Dejonghe, 2007b).
In addition in the tourism and management literature, the implicit scale,
impact and prestige of certain events, such as the Olympic Games or world
cups, means that they are referred to as “hallmark” or “mega” events (Getz,
1991; Torkildsen, 1994). In the UK, the Sports Industry Research Centre
(SIRC), based on empirical case studies, has also provided a specific taxonomy
of events connected with their sporting significance, economic impacts, reg-
ularity and scale (see Gratton et al., 2000; Gratton and Taylor, 2000).
1
Consequently this chapter does not focus on any specific event, but assimilates the
evidence. For an authoritative account of the economics of the Olympic Games, the reader is
referred to Preuss (2004).
Sports Events 337
TABLE 12.1 Types of sports events
Frequency Irregular
& Olympic Games
& Commonwealth Games
& World Championships
Regular
& Wimbledon
& Six Nations’ rugby
Level of competition International

& World Cup
& Champions League
National
& FA Cup Final
Regional/Local
& Amateur club championships
Single or multi-sport Single

& Henley rowing regatta
& Tour de France
Multi
& Olympic Games
& Commonwealth Games
& Asian Games
& School/university championships
Economic scale and impact Large- or small-scale (with respect to)

& Attendances
& Media coverage
& Sponsorship
& Investment requirements
Ownership of the event International sport federation

& IOC
& FIFA
& ICCB
Private
& Tour de France
& Paris–Dakar Rally
Location and assignment Same

& National stadia
& London, Paris, New York marathons
Rotating requiring bids

& Olympic and other games
& Champions League Final
& Final stage of Tour de France

SIRC identify:
& Type A: irregular, one-off, major international spectator events generating
significant economic activity and media interest (e.g., Olympic Games,
football world cup, European football championship).
& Type B: major spectator events, generating significant economic activity,
media interest and part of an annual domestic cycle of sports events (e.g.,
the FA cup final, six nations’ rugby union internationals, test match crick-
et, open golf, Wimbledon).
& Type C: irregular, one-off, major international spectator/competitor events
generating limited economic activity (e.g., European junior boxing cham-
pionships, European junior swimming championships, world badminton
championships, IAAF grand prix).
& Type D: major competitor events generating limited economic activity and
part of an annual cycle of sports events (e.g., national championships in
most sports).
Despite these differences in emphasis and classification, from an economic
perspective, two particular issues are of interest for this chapter. The first is
participation in the event. This requires consideration of the decisions of
spectators to watch the event and also of competitors to enter in the event,
which the remainder of this section discusses. The second issue concerns the
economic significance of the event, to which much of this chapter is devoted.
Discussion of this issue commences in the next section.
In the former case there is little specific economic research on attendance at
events, with most emphasis being connected with attendance at professional
sports leagues, as discussed in Chapter 10. However, the empirical grounding
of SIRCs taxonomy suggests that the highest level of sporting competition
does not necessarily attract the highest spectatorship, in the sense that the
economic significance of some of the events are low.2 This suggests that,
outside of the major multi-sport events, most economic activity appears to
be associated with the popular team sports and some individual sports, such as
golf and tennis. The reasons for this are, of course, difficult to identify. It might
be suggested, following the discussion of Chapter 10, that elements of personal
consumption and social capital, i.e., habit persistence, apply to the demand in
this case although, clearly, detailed research is required for a proper under-
standing. In general, moreover, SIRC note that in lower-level competitions,
spectatorship is broadly associated with the athletes, their families and
broader social groups.
As far as attracting competitors is concerned, then naturally some form of
selection mechanism must apply. On the one hand governing bodies will
2
The term “significance” is used deliberately, as will become evident later in the chapter.
Sports Events 339
establish performance criteria, which are required to be met for further repre-
sentation at regional, national and international competition. Selection here
is primarily from “within” the sport and founded on linkages between mass
participation and elite performance as discussed, for example, in Chapter 5.
Naturally, this applies to team and individual sports. In many respects there is
no choice to be made about participation. What is more interesting from an
economic perspective, therefore, is when choice is exercised by a competitor to
enter a specific event, given that they have options of alternatives at their
disposal. As discussed in Chapter 7, such choice problems conform to those
analyzed using tournament theory, where it was shown that a broadly positive
feedback between levels of sporting performance and economic performance
existed. This has obvious implications for the current context. Organizers
need to offer a large enough prize and distribution to attract competitors
and to elicit high levels of effort from them and also to attract spectators,
broadcast and sponsorship revenues, while at the same time not incurring
losses. It is important to note here that the prize need not be monetary. For
many athletes or event holders the opportunity to participate in or offer an
Olympic qualification trial may be considered a high enough prize, which
consequently attracts spectator and related interest and expenditure.
The link between economic and sporting performance is, however, quite
subtle in contest theory. In Chapter 7 it was argued that prizes had to be
increasing with the rank of competitors, to continue to elicit effort. This
suggests that the highest prizes should be associated with the highest level of
performance in tournaments, and that greater prize differentials would be
associated with higher performances. Logically, however, this suggests that
having a single prize is the most appropriate for event organizers, inasmuch
as the highest level of sporting performance can be elicited. In fact, this is not
the case in most sports. A range of prizes are offered and the economic reason
for this can be understood as reflecting the view that competitive balance and
the overall quality of the competition is more important to organizers and fans
than a single record-breaking performance or domination of a competition.
This is, of course, a mirror image of the UO argument in professional team
sports. However, there is another reason for multiple prizes that applies to
events more than professional sports teams, and this is that entry to tourna-
ments may be low if participants have clear or predictable chances of losing.
This was discussed in Chapter 7. The implication is that, from an economic
point of view, organizers are prepared to sacrifice some level of effort to promote
greater participation in the event. In the absence of this other participants may
not be found, or they may engage in very low levels of performance as they have
virtually no chance of success, and hence spectator, broadcast and other
incomes fall which could undermine the economic sustainability of the event.
There is relatively little empirical evidence of the application of contest

theory to sports events, and most has focused on the effort–performance–
reward relationship of competitors.3 Frick (2003) summarizes some of the
earlier research. One of the seminal pieces is an examination of the 1984
US and 1987 European Professional Golf Association tours, in which Ehren-
berg and Bognanno (1990a, b) found that player performance was positively
correlated with the value of prizes, controlling for a wide range of variables
including the quality of other players, the weather and course difficulties.
Bognanno (1990) found similar results in US professional bowling, but a
contradictory conclusion also, that greater skewness in the prize money pro-
duces less performance (measured as pins knocked down per round).
More consistent with the predictions of contest theory, Fernie and Metcalf
(1996, 1999) identified that British jockeys who had signed a contract that
guaranteed them a certain fixed income showed deteriorating performances in
comparison to when they had a contract without a certain guaranteed sum.
Likewise, Lynch and Zax (1998) examined US horse riding and found that
more effort is elicited with greater prizes and prize differentials, but the highest
prize events often attract low quality as well as high quality competitors.
Rationally, one would expect that over time poorer quality horses would not
be entered into races where they expected to do badly. This is, perhaps, an
example of long-shot gambling. In cases in which the prize differentials wid-
ened, however, it is shown that the quality differences in competitors reduced,
as would be expected.
In other sports, broadly consistent results are obtained. More recently
Lynch and Zax (2000) analyzed 135 professional running contests held be-
tween 1993 and 1995 and found that higher prize money resulted in faster
finishing times. This result is supported by Maloney and McCormick (2000)
who analyzed 115 professional running races in the US held between 1987 and
3
There is a large and growing literature that examines the macroeconomic relationships
between medal success and economic activity. Econometric analysis of medal performance,
i.e., medal counts or proportions of medals won or, as recently proposed by De Bosscher
(forthcoming 2007), the market share of points earned in competition, with 3 points
corresponding to a gold medal, 2 for a silver medal and 1 for a bronze, is regressed on
independent variables such as Gross Domestic Product (GDP), population, and host country
status in reduced form equations as discussed in Chapter 10. These determinants act as
proxies for implied demand and supply factors contributing to sports success. For example,
population can be viewed as an indicator of the pool of talent available to countries, a supply-
side factor. In contrast, GDP can be thought of as the ability of countries to “buy” success. In
this regard it acts as a proxy for allocation of resources to sport, a supply-side factor, but also
to the purchase of athletic effort, a demand-side factor. A summary of some of the studies is
presented in Appendix 12.4.
Sports Events 341
1991 and identified that the average prize paid and prize spread had a positive
effect on performances. Lallemand et al. (2008) also found that in women’s
official WTA tennis tournaments between 2002 and 2004 a positive and
significant relationship between prize spread and performance was evident.
However, they also found that players’ ability and the difference in ability
between players, given by their individual ranking, has a bigger influence on
performance than incentives. Becker and Huselid (1992) examined prize mon-
ey differentials in National Association for Stock Car Auto Racing (NASCAR)
and International Motor Sports Association (IMSA) racing, and confirmed
that performances increased with prize money. They also noted that the effects
were possibly non-linear, peaking at a certain level. This result was supported
by Von Allmen (2001) who investigated the team incentives and reward struc-
ture of the 1999 professional NASCARWinston Cup. It was concluded that, in
comparison with golf, rather “flat” prize money is more efficient than a win-
ner-takes-all system, because the costs that may be inflicted on other partici-
pants, due to an increasing crash rate, could be too high. Becker and Huselid
(1992) had also identified that riskier driving occurred with higher prize mon-
ey. However, Schwartz et al. (2007) argued that drivers will increase their
efforts as they see their overall ranking drop, so that unskilled drivers only
tend to drive more aggressively in the early stages of the season.
This latter point is interesting, because it raises the issue of selection biases
in the research. That is, competitors are only likely to enter tournaments
which they perceive they have an opportunity to win and, of course vice versa,
as discussed above. This means that better competitors are more likely to enter
the highly paid tournaments. In this way average performances are likely to be
higher regardless of the direct effects of prizes on effort. To try to account for
selection effects Frick (2003) examined three golf tours in the US, the senior
PGA tour, the PGA tour and the Buy.it.com tour, that differ in prize money and
prestige. Selection into the tours is likely to require demonstrable skill. It is
found that in the lowest value tour, the Buy.it.com tour, the survival rate of
entrants is much lower, suggesting that lower quality players opt out of golf if
they fail to make a living. Further, an earnings premium is identified for
players progressing on to the other tours, in which performance is higher. It
is argued, therefore, that both selection and performance are influenced by
prizes. The upshot of this empirical research is that the evidence that is
available broadly supports the positive feedback that it is proposed exists
between sporting and economic performance, and also supports organizers
offering more balanced tournaments than those that might occur with only
one prize.
There are three important corollaries to this discussion. The first relates
to the prize spread. As noted in Chapter 7, Frick (2003) argues that a key
tenet of tournament theory is that the incentives to elicit effort should

increase with rank. The evidence broadly supports this argument with
respect to financial prizes. However, it seems reasonable to argue that this
can also explain why a range of non-pecuniary prizes exist, such as the gold,
silver and bronze medals in the Olympics and other games.4 In the ancient
games only the winner received the prize of the olive branch! This suggests,
on the one hand, that athletes and their supporting administration place a
differential weighting on these medals. In this sense, a gold medal is not just
two places above a bronze medal. It might also not be appropriate to score its
value 3, versus 2 for silver and versus 1 for bronze, as some medal tables
comparisons suggest.5 On the other hand, having a range of prizes can be
viewed as a mechanism to increase participation of nations across the
events, to avoid traditional specialisms, etc.6
The second point concerns the “darker” forms of incentive offered in
tournaments, which is the possibility of cheating, as discussed in Chapter 7.
Unlike the sports policy and philosophy of sport literature, the economics of
cheating in sport is an under-researched area. Maennig (2002), however, draw-
ing on Becker (1968) and the theory of rational behaviour discussed in Chapter
3, argues that rising prize money and related incomes from participation in
sport raises the incentive to cheat, and consequently more cheating is likely if
the returns for legitimate endeavour remain constant and efforts to tackle
illicit behaviour also remain constant.7 In this way cheating, such as doping,
can be understood as the extraction of economic rents in the face of informa-
tion asymmetries between athlete and monitor.
The third point is that hosting major events has now become a large and
growing business and the supply of such events requires cities to “bid” to host
4
This is not to say that funding for athletes from sports organizations, and related
sponsorship deals etc., are not relevant, but that historically amateur activity even operated
with this prize structure.
5
In an interesting paper Blavatskyy (2004) shows that, in the typical symmetric contest
modified to allow for asymmetric equilibria, that is when some contestants drop out of a
tournament as with multiple elimination tournaments such as the Olympics, aggregate
effort can rise with multiple prizes, whereas it would normally fall as the effort levels of
competitors drop. In the former case, this is argued to be because of the likely greater number
of competitors. The general proposition that effort levels fall with prize sharing is identified
in the evidence above, inasmuch as greater prize differentials increase performance. The
value of medals is discussed in Appendix 12.4.
6
There is currently anecdotal evidence that countries target specific events, however, to
boost their overall medal tally.
7
A forthcoming volume edited by Rodriguez and Kesenne, based on a workshop at Gijon in
Spain in May 2008, addresses the economics of doping and other threats to sport.
Investment Decisions 343
them. This process can also be viewed as a tournament. Tournament theory

predicts that success is most likely to be the case where greater effort is
expended. In this case, this suggests that cities (and governments) that are
most prepared for hosting events are more likely to be successful in meeting
the objectives of the tendering sports organization.
The Olympic Games illustrates this process. According to Preuss (2004)
the IOC derives its power from the monopoly supply of the “product” – the
Olympic Games – which it franchises to a specific Organizing Committee of
the Olympic Games (OCOG) to a city or country, to deliver. Why cities bid to
host the Games is a complex issue, which is partially addressed in the next
section, however, whatever the specific reasons, the strong demand to host the
games coupled with the need to cover infrastructural costs:
“. . . means that bid candidates will become limited to those cities which
already have an adequate infrastructure to stage the Games. It is said that
today there are only 20 countries able to stage the Olympic Games.” (Preuss,
2004 p. 284)
Clearly, therefore, the allocation of the Olympic Games to host cities faces
the same trade-offs between participation in a tournament and the perfor-
mance of participants in the tournament, measured here in terms of its
economic scale. Expenditures of “effort,” i.e., resource and capability, to pro-
duce the Games raise the likelihood of success in bidding to become the host.
This militates against the opportunity for developing countries to host the
Games, although arguably these countries would benefit more.8 As with tour-
naments generally, moreover, incentives to cheat exist. Corruption scandals
have been noted, as indicated in Box 12.1.
This discussion reveals that many of the economic principles used to
analyze professional team sports contests carry over to the process of organiz-
ing sports events when they are understood in the context of contest theory. As
implied above, the similarity of aspects of analysis also applies to the decision
to invest in sports, which is a central component of hosting sports event. This
issue is now discussed at length.
12.3 INVESTMENT DECISIONS

12.3.1 Theoretical issues
Chapter 3, and particularly Appendix 3.2, introduced investment decisions
in terms of increasing the capacity of an individual to undertake an activity
8
Although, as discussed below, it can be argued that the benefits are more alleged than real
in economic terms.
BOX 12.1 BRIBERY AT THE IOC

1. One of the most serious charges of corruption at the See: http://www.infoplease.com/spot/olympicscandal1.
Olympic Games concerns allegations, in December html
1998, suggesting that members of Salt Lake City’s bid 2. Ivan Slavkov, the Bulgarian IOC member, was investi-
committee spent millions of dollars to improperly influ- gated by the IOC when in 2000 it was alleged he offered
ence the votes of 14 International Olympic Committee support to a businessman who went on to try and solicit
(IOC) members. It is alleged that payments included bribes from Cape Town when they were bidding to stage
cash, medical expenses, travel expenses, gifts and en- the 2004 Games. He was also secretly filmed by the
tertainment, and college tuition payments for IOC mem- BBC arguing that he could arrange votes to back
bers’ children. Policy responses followed including not London’s 2012 bid for money. He was expelled from
allowing members to visit bidding cities and reconstitu- the IOC in 2004.
tion of the bidding process and membership of the IOC.
See: http://news.bbc.co.uk/sport1/hi/other_sports/
In addition, six IOC members were expelled and three
olympics_2012/3531456.stm
others resigned. IOC president Juan Antonio Samaranch
See also Brown, G., Olympic Five-Ring circus: Bribery
ruled out resigning, but pledged to lead an effort to purge
Scandal Tarnishes Olympic Gold.
the corruption, promising the formation of an ethics
commission to review how the IOC members operate.
by enhancing their skills (by acquiring human capital or consumption cap-

ital) or involving the purchase of goods and equipment that can be used
more than once (durable goods) to yield a flow of services or utility over time
that is not used up with one consumption act. More traditionally, invest-
ment decisions are connected with suppliers increasing their capacity to
supply goods or services to consumers. This implies that suppliers incur
costs to increase the capital required to facilitate their operations in the
current period, or at least over some phased set of future periods, while
seeking to at least pay back these costs from revenue streams being earned
from future purchases of their goods or services.
Investment decision making requires consideration of two related points:
1. Establishing the decision rule for making an investment; and
2. Assessing the mutually exclusivity or independence of the decision.
The first point implies that to make an investment requires specifying a
clear objective. Following the discussion of Chapter 5, it is assumed in
economic theory that private sector suppliers seek to maximize profits.
However, this objective was primarily discussed in a situation in which
the current capacity of suppliers was fixed. As investment decisions take
time to earn a return, the nature of profit maximization over time needs to
be discussed. The second point is concerned with whether or not the in-
vestment is unique or part of a portfolio of investment options. For example,
the building of an Olympic stadium or a new professional sports team
stadium is an example of a mutually exclusive investment. Only one of the

investment options for alternative designs and locations can be undertaken.
In contrast, the equipping of a leisure centre constitutes independent in-
vestment decisions in that more than one type of equipment can be invested
in simultaneously, at least up to a budgetary or spatial limit. Once either of
these limits is met of course, options which might include alternative bun-
dles of investment goods, become mutually exclusive. Appendix 12.1
demonstrates that accepting any project which yields a positive “net present
value” (NPV) contributes to maximizing profits over time in theory, by
maximizing wealth.
In economic terms “wealth” is the sum or stock of net income flows of an individual.
12.3.2 Public sector rationale

The discussions in Chapter 5 indicated that profit maximization might not be
relevant for the theoretical analysis of the public sector. In general, however,
the specific economic rationale for public sector investment in sport tends not
to be explicitly discussed, although various benefits are identified as its out-
comes. This is the case in the twin-track approach advocated for the UK, in
DCMS/Strategy Unit (2002), in which investments in both elite sport, that is
in the production of international success and the production of facilities that
could host events, and mass participation mutually reinforce one another.
Specifically, according to this approach, investment in international success
will generate benefits (costs) from:
1. The feel-good (bad) factor stemming from sporting success (failure);
2. The enhanced (diminished) image of the UK;
3. The enhanced (reduced) productivity and consumer confidence following
sporting success (failure).
While hosting sporting events and investing in infrastructure may:
1. Generate ongoing benefits for the community in the form of legacies, includ-
ing urban regeneration, enhanced sports participation, increased tourism;
2. Attract new visitors helping to sustain economic activity;
3. Contribute to future sporting success.
Clearly these features are interrelated. Investment in events and infrastruc-
ture may generate sporting success and its related benefits. The “feel-good”
factor may affect economic productivity, enhance the image of a country’s
tourism and bring success in sports, which in turn further enhance the lega-
cies. Arguments such as these are not unique to the UK (Sandy et al., 2004, p.
187; Gratton et al., 2005, p. 986). There is, therefore, a strong and general
policy impetus to invest in international sporting success, events and
TABLE 12.2 Forms of investment
Aim of investment Type of investment Examples
Facilitating international Support national teams in UK Sport, Sport Canada,

sporting success competitions German Sports
Confederation (see
Chapters 2 and 6)
Facilitating professional Build new/refurbish US major league sports
sports teams stadium franchises
Hosting national teams Build/refurbish a national Wembley Stadium,
stadium Millennium Stadium
Hosting local, national and Build new/refurbish stadia, Olympic venues, Sheffield,
international events arenas and related facilities Indianapolis
infrastructure. Table 12.2 summarizes the four main arenas in which this has
happened and gives some examples of the investments.
Chapters 2 and 6 indicated a need for public investment in elite sports
that do not have the requisite income from commercial activity to be self-
sustaining. The administration of such investment is undertaken by
various non-governmental sports policy bodies. As far as investment in
physical infrastructure is concerned, the US has had a long tradition of civic
authorities providing new, or refurbishing existing, facilities in the major
league sports. Siegfried and Zimbalist (2000) report that since the 1950s
public investment in new facilities has never dropped below 55% on average,
has typically been in the region of about 70% and has, at times, been 100%.
In the case of refurbishing existing facilities the proportion of total invest-
ment has never dropped below 78% on average over the same period and has
typically been in excess of 90%.
As noted by Gratton et al. (2005), elsewhere investment has typically
been in national stadiums or stadiums that have been designed for a
variety of uses. In theoretical terms a defence of public sector invest-
ment is that, first, the private sector does not identify profitable opportu-
nities to make such an investment but secondly, this lack of identification
results from an understatement of the benefits to the economy that could
occur with the investment. There are two main reasons why the economic
benefits may not be accounted for, both of which are connected with
market failure. The first is that the benefits exist as externalities. In this
respect, the argument for supporting public investment in sport parallels
that made for supporting sports participation activities, as discussed in
Chapter 5.
Which of the above impacts of international sporting success and hosting mega events are
examples of externalities?
Hint: From Chapter 5 externalities were defined as occurring when the activities of an agent
directly affect another agent and whose effects are incidental to the activity.
The impact of sporting success on general productivity would be an exter-

nality effect. The same would be the case of sporting success attracting tour-
ists, increasing attendance at events or hosting of sports events raising sports
participation. This is because the effects are incidental to the initial sports
activity and not related through market transactions but through direct effects
on the utility, and hence behaviour, of individuals.9
The second is that even though a specific investment might not be profit-
able, i.e., add to a private sector company’s wealth and hence would be over-
looked, wider economic benefits can be generated by the investment through
multiplier effects, as first discussed in Chapter 1. In this case taking a broader
as opposed to an individual organizational perspective, the costs of the invest-
ment overall might be met or exceeded by the subsequent economic activity
that is generated beyond the investing organization. This makes the multiplier
effect distinct from an externality, as the effects are mediated through market
linkages rather than directly by the utility functions of individuals. The de-
tailed nature of the multiplier and its possibility is discussed further below.
Which of the above impacts of international sporting success and hosting mega events could
bring about or rely on multiplier effects?
Hint: Which activities are likely to be generated from related market transactions?
In the examples above urban regeneration or new visitors sustaining eco-

nomic activity would be examples of multiplier effects. This is because, in the
former case, the initial investment in sports infrastructure stimulates addi-
tional economic activity in the markets providing the inputs to the investment
asset. Likewise, these markets further stimulate economic activity in their
respective input markets and so on. In the latter case, the spending of visitors
at an event will support the investment in the event directly and thus
9
As will become clearer below, it is the additional tourism generated from sports investment
that is central to legitimately identifying economic impact.
stimulate related input markets, as just described, but will also stimulate other
related consumption, such as accommodation, hospitality and transport. This
makes it apparent that the effects of the multiplier are complex. Because of its
centrality to discussions concerning public investment in sports infrastruc-
ture, the concept of the multiplier is discussed in more detail in the next
section.10
12.3.3 The multiplier

As discussed in Chapter 3, a basic feature of economic analysis is that it
constitutes a flow of resources between economic agents as described in the
“circular flow of income.” An important feature of the circular flow of income
is that expenditure will correspond to income at the aggregate levels and, by
implication, the value of economic output. Equation 12.1 summarizes this
result:
Expenditure ¼ Income ¼ Value of output ð12:1Þ
It was also noted in Chapter 3 that there are leakages from the circular flow
in the form of savings by households, taxation by public authorities, and
imports of goods and services. Each of these, other things being equal, will
reduce the circular flow of income. Correspondingly, injections of income can
increase the circular flow of income. These might comprise investment by
firms or expenditures from public authorities and exports. If the injections and
leakages of income balance, then the circular flow of income will be constant
and will be in equilibrium.
The multiplier effect is linked to the injection of resources to the circular
flow of income, and consequently the latter’s expansion. Box 12.2 identifies
the components of multiplier effects that describe this expansion. Because in
theory, expenditures, incomes and the value of outputs must be equal, then
these effects could in principle be calibrated in terms of either of them.
Box 12.2 indicates that the multiplier is the scaling factor that links the value
of the total economic activity generated by an investment to the value of the
initial injection of resources. From a planning or evaluation point of view, once
a multiplier is calculated or estimated, it can be used to scale-up the value of
the initial proposed economic activity that will be generated from a subsequent
investment to identify the overall new level of economic activity.
10
As will become apparent discussed further below, the effects of the investment can either
be understood through the demand side, i.e., the related expenditures involved, or the supply
side, i.e., the incomes received.
BOX 12.2 COMPONENTS OF THE MULTIPLIER

Direct effect: the initial increase in expenditure or income construction company or the company running the sports
connected with the injection of resources by the sports event.
event. For example, the expenditure on building or specta- Induced effect: the increase in expenditures on, or incomes
tors visiting facilities, or income received by facility owners received by, suppliers and employees of the organizations
or event organizers. that supply those building the facilities or running a sports
Indirect effect: the increase in expenditure or income gen- event. In principle the multiplier can be calibrated as:
erated as a subsequent result of the sports facility construc-
tion or hosting the event. For example, the expenditure on, Direct effect þ Indirect effect þ Induced effect
or income received by, suppliers to and employees of a
A final issue to note is that while Equation 12.1 calibrates economic value
in terms of expenditures, incomes or the value of output, it is possible to
identify the employment that might be generated by an investment. Recall
from Chapter 1 and Chapter 11 that in economic terms output “Y” can be
understood as deriving from a production function that depends on the inputs
of land, “Ld,” Labour, “L” and capital “K.” This is represented in Equation
12.2:
Y ¼ YðLd; L; KÞ ð12:2Þ
Logically, the value of output will be equivalent to the value of the inputs.
Consequently, if “p” is the price of output (for example a ticket price), “rt” is the
rental value of land (for example price of land per acre), “w” is the wage rate (for
example of athletes and employees of a stadium) and “r” is the cost of capital
(for example, profit rate) then Equation 12.3 holds, in which the value of
output is equivalent to the total costs of the factors of production:11
pY ¼ rtLd þ wL þ rK ð12:3Þ
This means that for any given moment, with prices known and land and
capital fixed, then employment can be given as a proportion of the value of
output as:
L ¼ ðpY rtLd rKÞ=w ð12:4Þ
11
Note here that profit is viewed in economic terms as a “cost.” The profit rate is the cost of
keeping the capital in the productive activity.
Calculations such as these are

common in the literature and consul-
tancy reports evaluating investment
in sport, with L being forecast from
the total value, pY, estimated to be
generated by an investment and non-
labour costs from current experience
being extrapolated from current levels.
There are problems with these calcu-
FIGURE 12.1 The impact of injections on prices and resource
lations. On the one hand if technology
employment.
changes as a result of the investment
in K, then the form of the production function implied in Equation 12.2 will
change, because of the changes to productivity. On the other hand, even with
similar technology, greater employment of inputs will affect productivity as,
for example, implied in the concept of diminishing marginal productivity. In
both cases, therefore, in the long run as investments are undertaken, changing
productivity will cause the prices of factors of production to change relatively
in connection with the demands for them, and consequently alter the relative
employment of the factors. Proportionate calculations of the likely employ-
ment of labour assume these complications away.12
The empirical calculation of the multiplier is discussed in more detail in
Section 12.4. For now, attention focuses on the possibility of its existence.
This is important because without the multiplier, a central argument for a
public policy of investment in sports infrastructure is removed. Both the
possibility and extent of multiplier effects are dependent on certain conditions
existing in the structure of supply in the economy. To begin with, the basis for
relying on multiplier effects to enhance regeneration and employment is con-
nected to the possibility of subsequent increases in real economic activity, i.e.,
Y as defined above, following an injection of resources. In contrast, as dis-
cussed in Chapter 1, if all economic activity behaved according to the predic-
tions of the model of perfect competition then the economy would always fully
adjust to an equilibrium in which resources were fully employed and no
increase in injections could produce a multiplier effect in the aggregate.
These alternatives are illustrated in Figure 12.1 in which two scenarios are
presented. In both cases the vertical axis measures the price level in the
economy and the horizontal axis the real value of output, Y. On the left hand
side is an economy with fully employed resources, as indicated by a vertical
12
The reader should note that this sort of argument resonates with the discussion of supply
structures below and is central to differences in multiplier calculations according to input–
output and computable general equilibrium models discussed later.
(aggregate) supply curve, Sv. If (aggregate) demand is increased, from D1 to D2,

by an injection to the circular flow of income then prices have to rise, i.e.,
inflation occurs, and there is no increase in real economic output and no
multiplier is possible.
In contrast, if there are unemployed resources in the economy, as indicat-
ed by a horizontal supply curve, then an injection to the circular flow of
income can lead to an expansion of real economic activity, following the
effects identified in Box 12.2, as prices do not have to rise to encourage the
employment of the resources. This latter situation, however, reveals that
markets must have failed to fully allocate resources in order for a multiplier
to occur. In this respect, while externalities and multiplier effects are both
indicative of spillovers from economic transactions due to market failures,
the failures are of a different kind. The multiplier effect relies on an assump-
tion about the failure of the market to allocate resources. In this respect the
multiplier has its origins in the analysis of the sources and remedies of
prolonged unemployment experienced in the Great Depression of the
1930s. The unemployment was characterized as structural, because the mar-
ket system did not seem to have adjusted to re-employ the workers or other
resources. It is this logic that is implied in arguments to regenerate an area. As
traditional industries decline it is felt that the market, i.e., cheaper prices of
land, and the possibility of employing workers more cheaply will not encour-
age new industries to invest in the area.13 Consequently policy agencies
should begin the process of reinvestment.
What slope would the aggregate supply curve have to take for both prices and real resources
to increase with an injection of demand?
Hint: Think of a position between the two extremes just discussed.
Of course, it is possible that both prices and the employment of real

resources can increase following an injection to the circular flow of income,
as implied above. This would require a supply curve that slopes upward from
left to right. In such circumstances an injection of resources, i.e., demand,
13
This can be viewed as a problematic claim. Whereas in the 1930s international trade
collapsed, currently in a global economy many industries now make international
investment decisions and typically manufacturing has been relocated to the newly-emerging
markets in the Far East and Asia, to avail themselves of cheaper labour and also investment
incentives. Consequently, the global market can be viewed as allocating resources but the
process is slow and politically unpalatable.
must raise both prices and the level of real output. Clearly the steeper the curve
the more that prices adjust relative to real resources and hence the lower the
potential multiplier and vice versa.
The point that should be emphasized here is that the value of economic
output, for example, as indicated in Equation 12.1, is the product of the price
level and the level of real economic output as indicated in Equation 12.3. Any
change in expenditure in the economy, i.e., an injection of income, can thus
become manifest as a change in either prices or output (or both) because of
differences in the supply conditions of an economy. The degree of price change
that ensues from an injection of investment limits the potential size of mul-
tiplier effects.
The second supply-side issue to consider in reviewing the possibility and
scale of multiplier effects is the spatial context within which investments take
place. Implicit in the discussions above, and connected to the macroeconomic
origins of multiplier analysis, is that an “aggregate” form of economy is con-
sidered. In this respect, the circular flow of income should be viewed as a
conceptual device that refers to an economic system within certain bound-
aries. Historically, nations provided clear policy-related boundaries for this
analysis. If a smaller region of the economy provides the focus for analysis,
however, for example using sports investment to build a specific facility or to
host an event to regenerate a specific urban area, then it is likely that any
multiplier effects will depend on the spatial context of the venue or event. This
implies that local spending is unlikely to contribute to economic impact, as it
may simply be a diversion from other expenditure in the area. Likewise, an
injection of investment in one area may “crowd out” expenditure in another
area or increase the leakages from that area.
Crowding out is a term used in economics, particularly macroeconomics, to indicate
that if an economy is at full employment then public expenditure can only take place
at the expense of private sector investment. These arguments have been used to
counter the theoretical efficacy of public sector investment as a means of controlling
the level of economic activity and also, empirically, for the growth in the proportion
of public sector activity in the economy.
For example, if a new sports stadium helps to promote further investment

into a region, elements of this may come from suppliers relocating their
businesses. If the stadium helps to promote tourism, elements of this may
include reduction in tourism and expenditure elsewhere.
The final supply-side issue concerns the timescale over which multiplier
effects are likely to be experienced. Theoretically, the multiplier effect is cal-
ibrated by comparing two equilibrium positions: the level of real income before
the investment; and that after it. The model of perfect competition, and indeed
most economic theory, suggests that market prices will adjust rapidly to
reallocate resources and consequently there can be no net increase in real
economic activity. The question arises of what happens if this is not the case.
It is conceivable that the supply side of the economy takes time to adjust to a
demand side change following an injection of resources (and, of course, vice
versa). A number of different scenarios now become possible.
One view might be that in Figure 12.1, the right-hand side diagram may
represent the short-term, but the left-hand side diagram the long-term. The
argument to support such a view is that an injection of investment or other
expenditure initially appears to raise real output. However, this is because
economic agents misconstrue the injection as leading to an increase in real
income. In fact, the market system will generate full employment in the long-
run and so only a nominal increase in income occurs. Consequently, as prices
begin to rise, the economy is forced towards its long-term real output level, but
higher prices will stabilize for the economy over time. This is connected with
the adaptive expectations of economic agents. If economic agents have ratio-
nal expectations this would imply a rapid change in prices consistent with
perfect competition, such that the left hand-side of the diagram applies.
Another view is that the multiplier effects discussed in Box 12.1 implicitly
presuppose that the level of supply capacity or potential is fixed. It is inherently
a relatively short-run perspective. In other words, the multiplier works by the
re-employment of previously unemployed resources. The analogy might be
that a vacant stadium is once again used for activities following some refur-
bishment investment. Such an effect does not address any subsequent invest-
ment in the locality that may occur as a result of the initial injection of
resources.14 If investment decisions depend on increases in income (in the
economic area) then the potential capacity to supply will also increase in that
area. Investment depending on income can be represented as an accelerator
model, a version of which was discussed in Appendix 3.2. The impact of this
accelerator mechanism would be to reinforce the multiplier effect through an
increase in the supply potential of the economy.
While the traditional multiplier–accelerator model relies on the idea that
investment can also be induced from changes in income, suggesting a poten-
tially greater long-run impact of the investment than just the multiplier, the
more general point is that if investment raises productivity generally then
there may be a shift in the supply curve to the right, even if the left-hand
diagram of Figure 12.1 applies, as the capacity of the economy is enhanced. It
seems unlikely, however, that such effects, relying on wholesale productivity
14
As will be seen in the chapter below, this is not an issue typically addressed in the
literature.
changes, could be triggered by sports investment directly in equipment and

facilities and would need to rely on the productivity of employees being en-
hanced. Such claims are often made to support the public investment in
science and technology.
Despite the detail of these different scenarios, a broad difference in per-
spective lies at the heart of the discussions. This concerns the ability of the
market system to fully employ resources. If the market system can provide full
employment, then there is no logical basis for arguing that sports investment
can have wider economic benefits and it may have adverse redistribution
effects. It follows that the case for sports investment having such spillover
effects relies on the alternative position that markets cannot be presupposed to
employ all economic resources fully.15
12.4 TYPES OF SPORTS EVENTS AND IMPACTS

12.4.1 Duration and scale of impacts
The above discussion has focused primarily on the rationale for public sector
investment in sports events and infrastructure. In this section a clearer dis-
tinction is made between the duration and scale of impacts in different sports
contexts, as well as noting important differences in the type of impact that can
be monitored or assessed.
Table 12.3 includes the investment alternatives presented in Table 12.2 and,
notwithstanding their likely interaction, identifies the primary externality and
spillover effects that it can be argued are associated with each option in columns
3 and 4, respectively. Column 5 reports typically measured outcomes of the
options, which are referred to later in the chapter. It can be seen from these
columns that the infrastructural investments indicated in the last three rows
appear to be theoretically equivalent, as was implied in the opening paragraphs of
this chapter. However, there are two senses in which these alternatives can differ.
First, in the final row it is indicated that existing infrastructure could be used to
support economic events. In this respect subsidy of their use or the use of funds
to attract the events, would constitute the investment.
Secondly, and of more importance, in column 2 it is noted that in addition
to the initial investment subsequent economic stimuli can be identified from
visiting spectators and regeneration activity. Different investment options will
15
This concern over the efficacy of prices to allocate resources can be hinted at in noting that
in the multiplier–accelerator model investment depended (primarily at least) on past income
and not the rate of interest (or profit) as the cost of capital. In Chapter 3, it was argued that
investment market needed to be linked to the price of investment funds, to govern economic
behaviour consistently with the model of competition and fully employed resources.
TABLE 12.3 Summary of impacts of investment
Investment Economic investment Externalities Multiplier/accelerator Typically measured

outcomes
Facilitating Support national teams in 1. Increased 1. Rise in consumer

international competitions consumer confidence confidence surveys,
sporting success 2. Increased increase in consumption
productivity 2. Rise in productivity
indices
3. 1 & 2 Rise in share prices
Facilitating Build new/refurbish Multiplier effect Rise in employment,

professional sports stadium 1. Subsequent multiplier income or expenditure
teams 1. Visiting spectators effect
2. Regeneration 2. Subsequent multiplier/
accelerator effect
Hosting national Build new/refurbish a Multiplier effect Rise in employment,

teams national stadium 1. Subsequent multiplier effect income or expenditure
1. Visiting spectators 2. Subsequent multiplier/
2. Regeneration accelerator effect
Hosting local, Build new/refurbish stadia Multiplier effect Rise in employment,

national and and arenas, etc. 1. Subsequent multiplier effect income or expenditure
international 1. Visiting spectators 2. Subsequent multiplier/
Types of Sports Events and Impacts

events 2. Regeneration Hosting accelerator effect
events 3. Subsequent multiplier effect
3. Visiting spectators 4. Subsequent multiplier/
4. Regeneration accelerator effect
355
have different temporal and scales of effect.16 This raises a very important
issue, as indicated by Preuss (2004) in a review of the impact of the Olympic
Games, that notwithstanding the discussions of the supply-side of the econ-
omy discussed in the previous section:
“. . . the effect of non-recurring expenditure weakens in the course of time and
vanishes completely. This means that the increase declines with every new
period and, in the long run, the falling demand leads back to the equilibrium
income that existed before the Games. . .” (Preuss, 2004 p. 41)
This quotation reveals that to support the initial multiplier effects, i.e., on
levels of economic activity, a continuous injection of resources is required. If
these cease then the economy will shrink back to its previous level. In contrast,
if the structure of supply changes, for example accelerator effects promote new
investment and this becomes self-sustaining and supports new levels of activ-
ity, then when the injection of resources ceases the economy will shrink back
to a higher equilibrium level of income.
It follows that a professional sports team is the case most likely to bring
sustained additional expenditures than otherwise might have been the case (by
either expanding the capacity of a stadium or providing attendance expendi-
tures that were not previously available because the sport was not previously
provided for). This is because the facility will be scheduled for regular use
according to league and various other competitions. This will be more frequent
than the provision of infrastructure to host the sports contests of national
teams, and much more so than the investment in infrastructure to host an
event like the Olympic Games or a world cup.
This may help to explain why in the former case, national stadia are often
used by other sports and for events other than sports. It may also help to
explain why the legacies of the Olympic Games are now viewed with such
importance as an argument to justify public investment in the Games. Con-
versely, it may help to explain why in Germany the 2006 World Cup was
primarily hosted at existing facilities, as has been the case with the rugby
union World Cups. Although gains may be more temporary, they are likely
to be larger. In the latter case, the world cup is seen as an opportunity for the
IRB to raise most of its revenue for investing into the development of the sport
(http://www.rugbyheaven.com.au/news/news/irb-retains-world-cup-format/
16
The emphasis here is placed on spectators and regeneration as broad categories. In reality
a whole host of additional activity might be stimulated. For example, as discussed in Chapter
10, broadcast income might be attracted. This, together with the facility, might bring with it
additional sponsorship. Merchandising opportunities might also change as better facilities
are built. These details are suppressed in the interests of conceptual clarity. Preuss (2004)
gives a highly detailed account of the sources of impact from the Olympic Games.
Types of Sports Events and Impacts 357
accessed 15/1/08). Without such gains, as Szymanski (2002) argues, public

expenditure on sports events could be thought of as a consumption rather than
investment decision.
In addition to the potential fragility of longer-term sustainability of eco-
nomic activity generated by sports investment, it has been argued that the
scale of economic activity associated with various sports events is likely to vary
significantly. One-off major sports events such as the Olympic Games and
world cups are likely to have the greatest single value, followed by regular large-
scale sports tournaments such as Wimbledon, the Six Nations competition in
rugby and the Open in golf. Lower level tournaments and regular annual sports
leagues are likely to generate the least value per event. In general, it is argued
that the higher the level and the longer the duration of the event, the larger the
economic value that will be connected with the sports contest (Gratton and
Taylor, 2000; Gratton et al., 2000, 2005). This suggests that there may be a
trade-off between the durability and the scale of activity supported by invest-
ment in sports.
12.4.2 Theorizing economic activity as an impact

Up until now, the specific nature of the impact of sports investment has been
left broadly defined to refer to the impact on the level of real economic activity,
i.e., real expenditure, income or output following an injection to the relevant
circular flow of income. It is important to recognize, however, that there is a
distinction between three main concepts that can be used to evaluate alterna-
tive scenarios of economic activity generated by investment. These are:
1. Economic significance;
2. Economic impact;
3. Economic welfare.
Economic significance is the broadest measure of economic activity and
simply corresponds to the total values of all transaction involved. In contrast,
“economic impact” has become identified specifically with identifying the net
benefits to the economy as a whole as a result of the investment (Crompton,
2006). Finally, as discussed in Chapter 1, economic welfare is maximized
where economic efficiency is maximized. As discussed in Chapter 1, in theory,
perfect competition is consistent with maximum economic efficiency because
market transactions will reflect the full economic costs and benefits of
resources and, as such, guarantee productive and allocative efficiency. There
are two implications of considering the welfare effects of sports investment in
analysis. The first is that even if the main impact of the investment is to
redistribute resources criteria can be employed to assess this redistribution
in welfare terms, as discussed in Chapter 5, in which those who gain from a
redistribution can compensate those who lose. The second is that a focus on
economic welfare suggests that both the costs and benefits of alternatives
should be taken into account in evaluating the investment, unlike an econom-
ic impact analysis. Cost benefit analysis (CBA) is viewed as appropriate in this
case, a point made strongly by Kesenne (2005).
12.4.3 The practical measurement of economic activity

as an impact
If one accepts the above arguments that changes in economic welfare following
an investment are the most theoretically relevant indicator of economic
impacts, then this raises the question of how might this take place. As far as
both appraisal and evaluation are concerned, a focus on economic welfare
suggests that the true opportunity costs and benefits of investments should
be accounted for. There are a number of dimensions to this task.
12.4.3.1 Adjusting prices

The first is to ensure that prices are not distorted by taxation or subsidies, as
these reflect policy interventions rather than the benefits and costs of the
investment. Consequently, taxes should be added back to prices and subsidies
removed. Prices should also be discounted at the appropriate discount rate and
inflation taken into account.
12.4.3.2 Multiplier effects

Multiplier effects should be accounted for to ensure that genuine net benefits
are identified. It follows that this must also apply to an economic impact
analysis. Potential increases in economic activity that would have taken place
without the investment should also be discounted from the analysis.17
There are a number of specific methods for calculating multipliers, but they
essentially either focus on the demand side, i.e., expenditure, or supply-side
structure of the economy (Hudson, 2001; Blake, 2005).18
17
This is known as a “deadweight” effect. In the review of the evidence discussed later, the
trend rate of output, employment or income can be viewed as an indicator of this. Also, the
problems in the use of multiplier analysis in the past can be linked to including the
deadweight value as an impact from the investment or failing to allow for crowding out, etc.
18
These are the methods that tend to be employed in large-scale appraisals. In smaller-scale
projects methods that evaluate the multiplier by investigating the supply chain of an
organization or project exist. A good example is the LM3 model proposed by the New
Economic Foundation in conjunction with the Countryside Commission in the UK (Sacks,
2002).
1. Export injection multipliers;

2. Input–output models;
3. Computable general equilibrium models.
Export injection multipliers draw on the Keynesian analysis of the expenditure
multiplier discussed in Chapter 1 and presented in Appendix 1.2, to demonstrate
that an initial injection of expenditure from outside the economic system – which
constitutes visiting fans in the sporting context, hence the reference to exports – is
magnified through interlinked spending in the (local) economy. Equation 12.5
presents the basic multiplier presented in Appendix 1.2, and Equation 12.6
presents the more developed “export” version, where the impact of imports
and taxes reduces the value of the multiplier. The additional terms reduce the
value of the multiplier because of leakages. The term (1 i) indicates the pro-
portion of initial export-generated expenditure that remains an export. Multiply-
ing by this term deflates the mpc to account for imports. The term (1 t)
indicates the proportion of initial export generated expenditure that is not taxed.
1
Multiplier ¼ ¼3 ð12:5Þ
1b
1
Export multiplier ¼ ¼ 1:25 ð12:6Þ
1 bð1 iÞð1 tÞ
where:
b is the marginal propensity to consume (mpc) (<1), e.g., 0.667

i is the marginal propensity to import (<1), e.g., 0.5
t is the marginal tax rate (<1), e.g., 0.4
The input–output (IO) approach to multipliers is based on deriving the
inter-sectoral, i.e., supply-side, transactions in a defined economic region.
Based on an understanding of how various economic sectors trade with one
another, a change in economic activity in one sector can be examined for its
impact in another sector, and consequently multiplier effects established.
There are a number of assumptions of IO analysis which can be summarized
as in Dwyer et al. (2006):
1. There are no resource constraints;
2. The production function implies constant proportions between inputs and
outputs;
3. Price effects are neutral;
4. Government budgets are neutral;
5. Components of demand (including government budgets) are exogenous
(not given by the model).
These assumptions suggest that the supply-side of the economy is as given

on the right-hand side of Figure 12.1, and that if there is an expansion of
output in a given sector or industry, the derived demand for inputs for other
industries or sectors and the employment of labour will increase in proportion,
and prices will not rise to reduce the value of real output overall. The last
assumption suggests that the impact of any expansion on government
finances is ignored. How are the multiplier effects calculated? The basis of
the approach is to present a set of input–output coefficients in a matrix “A”
that summarize the trading flows of the economy for a unit of output.
A matrix is a table of elements or quantities. They can be added, subtracted,
multiplied or, by multiplying by an “inverse” matrix, equivalent to divided. They are
particularly useful for summarizing a set of linear equations. A vector is a matrix in
which either the number of columns or rows are restricted to one. Likewise, it is
equivalent to a linear subcomponent of a matrix.
It is then possible to show that:
X ¼ ½I A1 D ð12:7Þ
X is a vector of the outputs of the industries, and D is the demand for the
outputs. I is an identity matrix, in which each value on the (principal)
diagonal running from top left to bottom right is 1 and all other values
are 0. This matrix corresponds to the value of 1 in (scalar) arithmetic.
The matrix I A is calculated by subtracting the corresponding elements
of the identity and input–output coefficient matrices. The superscript 1
indicates that the inverse of the resultant matrix is multiplied by D. Mul-
tiplying by the inverse of a matrix is used in mathematical operations to
correspond to dividing in normal arithmetic. [I A]1 is known as the
Leontieff inverse matrix, after the originator of the analysis. It is this matrix
which provides the multiplier effects. Consequently, Equation 12.7 indi-
cates that for any change in consumption demands, D, the effect on industry
output, X, is a multiple of D determined by [I A]1 whose parameters are
the input–output coefficients, A, that describe the structure of inter-indus-
try trading (remember that I only contains 1s or 0s).
The assumptions of the IO approach have been challenged and, as a con-
sequence, it is increasingly argued that computable general equilibrium (CGE)
analysis is a more appropriate method for deriving economic impacts (Dwyer
et al., 2006; Blake, 2005). In contrast to IO analysis, CGE:
1. Constrains the availability of labour and capital so that wages and prices
can change only or as well as, output changes, following an injection to the
economy;
2. Accounts for the feedback on the demand for labour, and capital and corre-
sponding consumption of the changes in wages and prices;
3. Components of (real) final demand are endogenous as prices vary;
4. Changes in economic activity affect government expenditures and tax
receipts.
The implication of these assumptions is to recognize that while input–
output models can measure all of the positive impacts of an event, they are
incapable of modelling most of the negative impacts, so they consistently
overestimate the impact of events (Blake, 2005, p. 12). The possibility of
price changes crowding out the effect of injections is taken into account in
CGE, suggesting that the supply-side might be more akin to the left-hand
side diagram in Figure 12.1, or at least the supply curve has a positive
slope.
12.4.3.3 Non-market valued effects

A concern with economic welfare, in assessing the impact of sports invest-
ment, implies that attempts should also be made to estimate the externality
effects of sports investment, i.e., to account for the wider social and environ-
mental costs and benefits for which there is no market price. Conceptually
speaking, this recognizes that the total economic value of a sports investment
has more than one dimension. Gouguet (2002) and Barget and Gouguet
(2007) argue that, drawing on the field of environmental economics, two broad
domains of value can be identified with the investment, as illustrated in
Figure 12.2. Total economic value breaks down into two main categories,
use and non-use values. Non-use values are connected with the valuation
FIGURE 12.2 The components of total economic value.

placed by individuals, for example, on the existence of the investment. This

might be because they identify with the investment, for example the sports
franchise, facility or team, or because they feel that it is beneficial to them and
the community even although they do not intend to make use of the invest-
ment. In contrast, use value is connected with the actual use or intended use of
the investment. Actual use value would comprise, for example, the entrance or
ticket price paid, but would also include any additional value or consumer
surplus. As discussed in Chapter 1, this is the amount greater than the ticket
price that a consumer would be prepared to pay. Potential use value, on the
other hand, is the valuation placed on the intended future use of the invest-
ment. In the specific language of the Olympic Games, it is also possible that
legacy values exist which correspond to the potential future values that would
be generated from the investment for future generations. Of course, this also
applies to any investment and in environmental economics is often termed
bequest value.
A number of different methods exist to attempt to measure these elements
of value and they were typically developed to value natural resources. The
approaches broadly break down into two types. Revealed preferences, which
makes use of the values of economic variables in related markets to identify
use values, and stated preferences. Travel costs are an example of revealed
preferences discussed in Chapter 10. Stated preferences require survey work to
elicit the values attached to an economic activity or resource. For example,
contingent valuation techniques capture statements of the highest
“willingness to pay” to access an economic resource or lowest “willingness
to accept” compensation for losing access to it (Downward, 2004; Tidsell,
2006).
A final point to note is that explicit concern for equity can be built into
the analysis by allowing for differences in income distribution. For exam-
ple, the marginal utility of consumption can be estimated for particular
income bands and then some average level of income. Identifying the
proportion that the former comprises of the latter provides a weight. With
lower income levels it is expected that the marginal utility of consumption
will be higher than the average, and vice versa for those with higher-than-
average income. This means that for those on lower income, weights will
be >1 and for those on higher income, weights will be <1. These weights
could then be used to adjust estimated utilities so that lower-income
utilities will be increased and higher-income utilities decreased. Naturally,
when aggregating utilities to assess a redistribution of resources in policy,
this means that welfare is reallocated in favour of those with lower
incomes. Appendix 12.3 provides an example of how one might calculate
such weights.
12.4.3.4 Increased productivity and consumer confidence

The above discussion has focused directly on the measurement of econom-
ic welfare, impact or significance as a result of (potential) changes in real
economic resources. However, it was also noted earlier that international
sporting success can yield changes in consumer confidence and productiv-
ity through its impact on feel-good factors, coupled with the expectation of
a positive economic impact. The measurement of productivity and con-
sumer confidence as outcomes of such impacts could be undertaken di-
rectly by examining data on each of these items, however, most of the
literature focuses on the impact of sporting success, or its likelihood, and
subsequent economic benefits on share price returns, because it is argued
that the stock market will capture all available information that affects the
profitability of companies. This might be as a result of enhanced consumer
confidence or increased productivity. The typical approach used to examine
the effects of sporting success or the announcement of the hosting of a
major sporting event on stock price returns is an event study.19 This
approach uses a regression equation to examine the pattern of share prices
and then makes use of a dummy variable to indicate a period of time over
which, say, an international success occurs or the announcement of the
decision to host an event takes place compared to before the success or
announcement.
12.4.3.5 Other impacts

To close this section it is worth noting that research from an economic
perspective into the other external benefits or costs, i.e., legacies, of sport-
ing investment or hosting sports events, is relatively undeveloped. This
might include examining the impacts on community cohesion, communi-
ty image and sports development, including enhanced participation which,
as noted earlier, is considered to be a symbiotic goal for sports policy in the
UK. Indeed, as Crompton (2004) argues, an alternative rationale for asses-
sing the impact of sports events than purely focusing on the impact on real
economic output and employment is possible. This lies in focusing on
the intangible benefit of “psychic” income. Primary research could also
investigate the effects of sports investment on community and sporting
behaviour.
19
This terminology does not derive from the analysis of sports events but rather the event of
changes in the regulations of market trading. The event study approach, as indicated below,
has become common in the economic analysis of the impacts of sports events more generally
than stock prices.
12.5 EMPIRICAL EVIDENCE

The empirical evidence on the impact of investment in sports infrastructure
and events is the basis of much debate and, indeed, is clouded by some
controversy. Despite the undoubted complexity of trying to tease out the
complex effects of large-scale investment activity, the point is made by Cromp-
ton (2006) that while:
“There is a sound conceptual rationale for economic impact studies, and they
have a legitimate political role in informing both elected officials and tax
payers of the economic contributions . . . to community residents’
prosperity . . . this legitimacy is predicated on the studies being undertaken
with integrity. Because the motivation under-girding them usually is to
prove the legitimacy of the sponsor’s economic case, the temptation to
engage in mischievous practices is substantial. In some cases the practices
are the result of ignorance and are inadvertent, but too often they are
deliberate and enacted with intent to mislead and distort.” (Crompton,
2006 p. 67)
This suggests that the evidence is also, at times, deliberately misleading.
The earlier discussion of adverse incentives from contests suggests a ratio-
nale for this behaviour. There is a distinction in the literature, therefore,
between the ex ante research that tends to derive from proponents of in-
vestment in sports infrastructure and events, which Crompton suggests at
times is misleading, and ex post research which tends to come from inde-
pendent academic sources (Coates and Humphreys, 2002; Downward,
2003; Kasimati, 2003).20
12.5.1 Ex ante evidence

The claim that the ex ante evidence is often deliberately misleading is a strong
one, but Crompton presents evidence from US professional team sports. For
example, in consideration of the use of public funds to invest in a new stadium
for the Dallas Cowboys NFL team in Arlington Texas, four economic impact
studies were produced each with widely varying results. Three were supportive
of the investment, citing estimated impacts of between $51 million and $416
million in various cities and counties, whereas one study opposing the invest-
ment estimated a loss of $325.3 million.
20
The point here is not that academics may not have been part of the ex ante research
projects, which are typically funded by governments or organizers and carried out by state
agencies of private consultancy firms or consortia, but that there is an implicit advocacy
position as opposed to a more objective assessment of the impacts of the investment.
Empirical Evidence 365
Variation in results seems endemic to the literature concerned with either

forecasting or evaluating the hosting of major events worldwide.21 Table 12.4
presents some studies. Comparison of the value of the impacts is problematic
not only because of some currency differences, but because the year associated
with the prices in which the overall impacts were measured varies. However, it
is clear that there is considerable variance in the results, even for the same
event and method of assessing impact.
12.5.2 Explaining the variance

The source of variance in economic impact studies could be due to techni-
cal difficulties. However, it has also been systematically researched. Cromp-
ton (1995) has argued that 11 potential flaws can be identified with
economic impact studies, and which are potentially exploited by advocates
of investment in sports facilities and events. These are reported in
Table 12.5. Elements of these flaws apply to some of the studies of major
events, such as focusing on gross impacts, not being clear about the type of
employment generated and not always accounting for the costs in calculat-
ing net benefits (if not using a full cost-benefit analysis because economic
impact is identified as a policy target). Significantly, from an economic
perspective, perhaps of most importance is flaw number 10. No study
attempts to evaluate what investment in an alternative facility or scheme
would provide, although some include sensitivity analyses of their fore-
casts, such as Blake (2005) and the New South Wales (Australia) Treasury
(1997).
In an interesting piece of work, Hudson (2001) conducted a meta-analysis
of 19 economic impact studies connected with Major League Baseball, the
National Football League or the National Hockey League between 1989 and
1997 to try to identify which factors accounted for most of the variation of
economic impact valuations. The value of the economic impact was treated as
the dependent variable and a range of independent variables were used to
identify their contribution to the value estimate. These included the type of
sport, the location of the franchise, the age of the stadium and the assumptions
made about the location and nature of expenditures generated by the franchise
investment. The results indicated the lack of statistical significance of a
21
In addition to these studies that identify impacts in terms of the levels of employment and
income generated by events, others, such as Bohlmann and van Heerden (2008), identify the
percentage change in such variables. Using a computable general equilibrium model, they
identify that the 2010 world cup in South Africa is likely to change GDP by 0.08–0.69% and
employment by 0.72–()0.35%. The range of forecasts is based on low tax and high tax
scenarios, respectively, as a result of the event.
366
CHAPTER 12 The Economics of Sports Events and Infrastructure
TABLE 12.4 The economic impact of major events
Author Event Method US $ billion Jobs Impact

(unless indicated. period
Date of
prices if available)
Economic Research Los Angeles Olympics, 1984 Input output 2.3 (1984) 73 375 Not given
Associates (1984)
Kim et al. (1989) Seoul Olympics 1988 Not given 1.6 336 000 Not given
Brunet (1995) Barcelona Olympics 1992 Not given 30 296 640 Not given
KPMG (1993) Sydney Olympics 2000 Input output 5.1 (1992) 156 198 Not given
Humphreys and Atlanta Olympics 1996 Input output 5.1 (1994) 77 026 Not given
Plummer (1995)
NSW Treasury (1997) Sydney Olympics 2000 Computable general 4.5 (1996) 98 700 1995–2006
equilibrium
Andersen (1999) Sydney Olympics 2000 Input output 4.5 (1996) 63 000–90 000 Not given
Papanikos (1999) Athens Olympics 2000 Macroeconomic 15.9 (1999) 445 000 2000–2010
expenditure
Balfousia-Savva et al. (2001) Athens Olympics 2000 Macroeconomic 10.2 (2000) 300 400 2000–2010
expenditure
Cambridge Policy Manchester Commonwealth Cost benefit analysis Not given 6300 to 2007
Consultants (2002) Games 2002
Madden (2002) Sydney Olympics 2000 Computable general (AUS)$6.5 (1996) Not given 1995–2005
equilibrium
Yu (2004) Asian Games 2002 Not given (HK)$72.8 310 000 Not given
Blake (2005), Price Waterhouse London Olympics 2012 Computable general (£)1.94bn (2005) 8164 2005–2016
Coopers (2005), EEDA (2006) equilibrium
Brunet (2005) Barcelona Olympics 1992 Not given Not given 20 000 1986–2004
Insight Economics (2006) Melbourne Commonwealth Computable general (AUS)$1.6 (2002) 13 600 2002–2022
Games 2006 equilibrium
Maening and Du Plessis (2007) World Cup Germany 2006 Cost benefit analysis $0.997 Not given Not given
Maening (2007) World Cup Germany 2006 Not given Not given 25 000–50 000 Not given
Grant Thornton (2003) World Cup South Africa 2010 Cost benefit analysis (R)21 150 000 Not given
TABLE 12.5 Flaws with economic impact studies
Flaw Implication
1. Using sales instead of income 1. This models gross business turnover and
multipliers not the income created in the locality
which may be of more use to local
planning and forecasts of government
tax revenue
2. Misrepresentation of multipliers 2. Employment multipliers not allowing for
changes in working hours, growth in
casual part-time work only
3. Incremental multipliers used instead 3. The denominator used is direct income
of proportional multipliers rather than visitor expenditure. This
inflates the multiplier, as the latter is
clearly larger
4. Failure to define the region of impact 4. Can overstate the impact in a region, part
of a city, etc.
5. Inclusion of local spectators 5. Overstates expenditure
6. Failure to exclude casual spectators 6. Overstates expenditure
7. Fudging multipliers 7. Borrowing a multiplier, say, from an
official regional source is misleading.
National multipliers will be > state
multipliers which will be > city
multipliers because of leakages
8. Claiming total instead of marginal 8. The return on the potentially
benefits incremental contribution of public funds
should be considered otherwise benefits
are overstated.
9. Confusing turnover with the 9. Aggregating spends incorrectly by double
multiplier counting sales, etc., overstates the
benefits as opposed to “value-added”
10. Omit opportunity costs 10. Would a shopping centre not attract more
visitors and expenditure?
11. Measure only benefits 11. Omits costs of congestion, etc.
variable measuring the assumed multiplier, and that including local spending
in the studies boosts the forecast economic impact by $38 million, while not
identifying that the sport was the sole reason for visiting spectators coming to
the area, i.e., attributing visitors to the area as deriving from the attraction of
the franchise investment per se) can boost the impact by $41 million. This
suggests that, independently of the multiplier calculations, it is important to
get the direct impact of additional spending right in any study, as this can
seriously over-inflate the value of the impact. More recently, Crompton (2006)
has reaffirmed that including local residents in economic impact evaluations
is the “most frequent mischievous procedure” (p. 70).
This is not to say that including local residents in the analysis is always
wrong. If a major event keeps a resident at home, rather than going on a trip
that would normally have taken place, for example, to go on holiday, then
clearly this is additional expenditure that otherwise would not have taken
place. It follows that any visitors that are deterred from the area as a result
of the event should also be accounted for, as a reduction in direct impact.
Likewise, it may be the case that if the event or infrastructure was part of
the reason for attracting visitors to an area then logically a proportion of this
visit can be considered attributable to the investment. Preuss (2004) provides a
detailed discussion of these issues.22
Overall, however, this discussion suggests that the evidence being provided
to justify public investment in sports infrastructure and events is unreliable in
the specific sense in which research is classified, as different researchers, meth-
ods and contexts seem able to produce different results for the same or similar
events. It would seem reasonable to account for this in policy related advice.23
In research, reliability is broadly concerned with the consistency of results deriving
from various measurements or data collection tools. This contrasts with validity in
which a measurement or measurement tool actually measures what it is intended to
measure.
12.5.3 Ex post evidence

Standing in contrast is a body of research that evaluates the impact of invest-
ment in sports infrastructure and events ex post. Here, much more consistent
and reliable results appear to be derived. The approach adopted in the analysis
is also not based on any multiplier analysis or structural economic modelling,
as implied in IO and CGE, but is based on the event study regression approach
noted above. Here, measures of income and employment in a local economy
across cities and/or time are regressed on variables measuring changes in the
sports environment, such as the construction of new facilities or the relocation
of franchises, to see if the latter affect the former. In one of the studies the
authors conclude that:
22
Primary research in which a set of attractor variables to an area could be used to weight the
significance of the facility or event relative to others could be used to disaggregate a unit of
expenditure in much the same way that marketers might attribute value to components of a
product or service.
23
It should be noted that UK Sport’s advice on assessing the impact of sports events
recognizes that there are problems with using multiplier analysis and consequently focuses
on evaluating the direct impacts. However, this is not explicitly linked to the potential
problems of multiplier analysis, but more that multipliers will be location-specific and
consequently, not identify equivalent comparison across events (see UK Sport, undated).
“. . . Far from being engines of growth, these results indicate that at best SMSAs
get nothing from their sports fanchises; at worst they pay dearly for
professional athletic franchises.” (Coates and Humphreys, 1999 pp. 362–363)
These are a gloomy set of results for proponents of sports-led economic
development, because Table 12.6 reveals that they are common. The first
column reports the authors of the work, and the second column the subject
of the analysis. The next column describes the model and estimator used and
the final column reports the conclusions.
The literature suggests that, at best, hosting unique events or hosting post-
season events adds nothing statistically to employment or the value of the
economy, and at worst can cost the locality.24 What is interesting about this
research is the greater reliability, i.e., consistency of finding, that is lacking in
the ex ante literature. This naturally begs the question why this occurs.
A clear inference in the ex post research, as indicated by Crompton, is that
there is an incentive for policy makers to overstate the benefits from hosting
major events or investing in sports infrastructure. A number of economic
factors might promote this state of affairs. An obvious answer is corruption
or deliberate misleading behaviour connected with incentives to win bids in
the light of weak monitoring, as suggested by contest theory. Other answers
can be offered. Fort (1997) provides a discussion of public choice theory applied
to US sports and argues that in “direct democratic” systems of voting such as
referenda, organized lobby groups with economic interests can dominate by
trading votes to achieve a result of mutual interest. A number of specific
economic conditions could explain what make this political outcome possible.
To begin with, public investment in sports events infrastructure might be an
example of adverse selection, as discussed in Chapter 1, in which more risky
economic investment options are presented to the public than otherwise
would have been the case, because of the inability to effectively scrutinize
the options through a political process. Consequently, there is a government
failure in efficiency terms that is masked by political rhetoric. As discussed in
Chapter 5, the lack of frequency of the transactions of public investment in
sports franchises or major investments may add to the information costs of
transactions, i.e., the ability to exercise effective monitoring by the public.
This would stand in contrast to’ the relative efficiency of the stock market.
24
The ethos of this is captured in the following passage which is a book review by Tollison
(2001, p. 207), “. . . the discussion of so-called multipliers in this book is illuminating.
Somehow, local spending on sports creates extra local income. I thought Keynes was dead, but I
guess he lives on in these ‘local effects’ calculations of economists who study tourism and the
like. But, surely, the subsidization of sports is a poor example. At best these transfers from
taxpayers to teams are zero-sum transfers. At worst, they are negative-sum transfers. . .”
370
TABLE 12.6 Econometric evidence on economic impacts
Matheson and Baseball world series Period 1972–2000 “Our detailed regression
Baade (2008) analysis reveals that
25 metropolitan areas hosting
over the period 1997–2000
the play offs
cities appearing in the
OLS regression Dependent
major league baseball
variable:
& change in real income
post-season had higher
than expected income
& lagged change in real income
growth by 0.003%. This
figure is not statistically
& time trend
significantly different
& taxes as a percentage of all
than zero, although a
cities (73) best guess of the economic
& nominal wages as a
contribution of a single
percentage of all cities (73) post-season game is
& oil price booms and related $6.8million, or at most
city specific impacts $75 million, roughly
(e.g. Hurricane Andrew) half that of the typical
& for 25 cities, compared ex ante projection” (p. 333)
model estimates
with actual income
growth to assess impact
of play-offs
Baade and 1994 World Cup Period 1970–2000 “. . . an overall negative
Matheson (2004) impact on the average
13 of 73 largest US cities host city and the US
OLS regression Dependent economy overall” (p. 351)
variable:
& percentage change in real
income
& lagged percentage change in
real income
& time trend
& taxes as a percentage of all
cities (73)
& nominal wages as a
percentage of all
cities (73)
& oil price booms and busts
& for 13 cities compared model
estimates with actual income

growth to assess impact of world
cup
Coates and Post-season play (e.g., Period 1969–1998 “The mean overall impact
Humphreys (2002) World series, Superbowl, of the vector of sports
play-offs) Cities with professional sports variables . . . is negative
teams . . . the parameters on
Fixed effects estimator (panel the post-season appearance
estimator) variables are not
Dependent variable: statistically significant
from zero” (p. 296)
& real per capita income
& economic/demographic factors
& sport environment
Empirical Evidence
& appearance in post-season play
371
372
Baade and 1984 Los Angeles and Period 1969–1997 “The estimated coefficients
Matheson (2002) 1996 Atlanta Olympic for the summer Olympic
Games Games variable did not
57 cities among largest 50 emerge as statistically
metropolitan areas significant in either Los
hosting the play-offs Angeles or Atlanta” (p. 139)
Pooled regression
Dependent variable:
& percentage change in
employment in an area
& population (log value)
& real per capita personal
income
& nominal wages as a percentage
of all cities (57)

& taxes as a percentage
of all cities (57)

& oil boom and bust
& regional location
& metropolitan statistical area
& Olympic host

Lertwachara and Expansions, relocations Period 1969–2000 “. . . we find that a
Cochrane (2007) of MLB, NBA, NFL professional sport
and NHL 33 franchises team does not have a
Pooled regression (2 steps) positive economic impact
Estimate residual, unexpected on the local community
per capita income . . . estimated local
and its growthregress residual income in the presence of
on number of franchises a professional sports
Dependent variable: franchise is lower than
& level and growth of MSA
what would be estimated
per capita income in the absence of a
& unexpected values of i)
professional sports
franchise.” (p. 253)
& US GDP
& number of franchises
Empirical Evidence
373
Further, as discussed in Chapters 7 and 9, sports leagues can be understood

as cartels that can exercise monopoly power. Siegfried and Zimbalist (2000)
argue that in the US the monopoly power of leagues is such that they permit
enough franchises to make the entry of new leagues uneconomic and yet, at
the same time, leave enough outside options to give teams the option to
relocate. A similar argument could be made with respect to event organizers.
Because the IOC and FIFA are the only bodies that can award the Olympic
Games and the soccer World Cup, a state of affairs ascribed to other events and
organizers as well, it is clear that monopoly status is present. Indeed, because
mega events are completely unique, Leeds and von Almen (2002) have de-
scribed the IOC as the “ultimate franchise monopolist.” Under such circum-
stances, it can be argued that bidding cities face an all-or-nothing demand
curve in which over-bidding, or the “winners curse,” can occur.
The winners curse can arise in auctions in which competitive bids are made to buy
an item. Logically, the market price of an item must be the average revenue that is
received by a supplier. In an auction this must be equal to the average value of bids.
However, the winning bid must, logically, be greater than the average bid, which
suggests that it overvalues the item.
Figure 12.3 illustrates the proposition. This reproduces the basic monopoly
diagram introduced in Chapter 1 as Figure 1.3. In contrast to just a profit-
maximizing monopolist, the all-or-nothing monopolist can fix the scale of
output and extract additional expenditure beyond the market demand level.
Clearly this can relate to public subsidy beyond what the paying public will
contribute towards going to a major event.
Beside the potential political motives for overstating the impacts of sports
investment and the potentially dubious assumptions that are made in the
analyses, the reasons why little impact appears to
be identified in the ex post studies requires some
further deliberation. One argument might be that
the ex ante research is invalid, as well as unreliable.
In this regard, it might be argued that it is just not
possible to capture the complex interactions that
underpin possible crowding out. However, in this
respect the validity, if not the reliability, of the ex
post research can also be challenged. Baade and
Matheson (2004) argue that in their analysis:
“. . . the standard error of the estimate for the

typical city is above 1% meaning that one would
expect the models to predict actual economic
FIGURE 12.3 The “all-or-nothing” demand curve. growth for the cities . . . within 1% point about
two-thirds of the time. For the cities in question, a 1% error translates into a
$300 million difference for the smallest cities . . . and over a $2 billion
difference for the largest metropolitan areas . . . While it is unlikely that the
models for any individual city will capture the effects of even a large event,
one would expect that across a large number of cities, any event that
produces a large impact would emerge, on average as statistically
significant.” (Baade and Matheson, 2004 p. 350)
It is not entirely clear that this conclusion follows. Table 12.5 reports
impacts that can lie within these statistical bounds and, as discussed earlier,
it might be expected with sports franchises as opposed to major events that the
impacts will be smaller. It follows that there is some doubt as to the validity of
this line of research. The approach may just not produce sensitive enough
measurements to be definitive. There could also be an issue of causality
associated with the negative findings, that cities with lower incomes require
investment in sports. Causality is not particularly investigated in the analyses.
This is not to suggest that there may not be good reasons for accepting the
results. Baade and Matheson (2002) note that structurally, an event like the
Olympic Games imports “alien investment” to an industrial location, and
probably lacks the linkages that effective clusters of industry have in promot-
ing economic growth, as it requires entirely different physical and human
capital. This is of course precisely the reason why legacies are sought. Further,
in the case of US sports, franchise infrastructure has been described as resem-
bling “walled cities” by Seigfried and Zimbalist (2000). This is because the
possibilities of spillover effects for expenditure are essentially minimized by
tie-in clauses to suppliers of refreshment and hospitality, etc. The effects are
compounded if professional athletes do not live locally and they simply do not
attract new visitors to the area which, given the geographic dimensions of the
US, seems plausible. In this case spending is merely substituted from one
leisure activity to another. This could account for the reduction in income in
an area if the other spending would have been linked, say, to activities that
were more closely articulated to the locality. Suffice it to say that it is not
simply enough to say that the ex post research is necessarily adequate in
rejecting calls for public investment in events.
12.5.3.1 Increased productivity and consumer confidence
As discussed earlier in the chapter, the event study approach has also been
used to assess if investment in international sporting success can yield
changes in consumer confidence and productivity through its impact on
feel-good and that this might be reflected in share price returns. There is an
emergent literature addressing this issue. Ashton et al. (2003) examined
daily data drawn from the Financial Times Stock Index (FTSE) 100 index
for the period 6 January 1984 until 3 July 2002 and regressed the stock
returns on a dummy variable scored 1 if the English national soccer team
won a match, 0 if it drew a match and 1 if it lost a match. The results
suggested that wins were statistically associated with an increase in share
returns and losses with decreases. The implication is that feel-good en-
hanced expectations of profitability.
More qualified impacts have been identified in two studies of the an-
nouncement of the Olympic Games being hosted on host nation stock prices.
Berman et al. (2000) examined daily stock price data in Australia from 4
January 1988 to 29 November 1996 for the All Ordinaries Accumulation
Index and the 23 Australian Stock Exchange Industry Accumulation Indices.
The results indicated no effect on overall stock prices, but significant positive
effects in sectors connected with building, developers, engineering and ser-
vices, i.e., industries connected with the development and hosting of the
Games. Significantly, too, it was revealed that these effects were confined to
industries based in New South Wales.
In a study of the Athens and Milan Stock exchanges, Veraros et al. (2004)
examined the impact of the announcement of the Games being awarded to a
winner (Athens) and not the other main rival city (Milan). Weekly data,
because of limits on daily price fluctuations, were examined between 1 Sep-
tember 1995 and 10 October 2003 in both cases. It was identified that the
Athens stock market returns increased by 7.7%, which was primarily driven by
significant expected returns in construction and industrials. In the case of
Milan, no general (negative) impact on returns was identified, although an
effect was identified for the construction and electrical industry, contrary to
expectations, which is identified as being spurious.
12.5.4 Non-market valued effects

The literature measuring the non-market value of sports infrastructure and
investment is now growing, with a number of contingent valuation analyses
being undertaken. Table 12.7 presents analyses where willingness to pay has
been elicited for individuals in a sample and then aggregated for the relevant
population. Significantly, Barros (2006) also argues that the analysis reveals
that the full costs of Euro 2004 of e282 million are not covered by the
aggregate willingness to pay, which suggests that it would be inefficient to
use public funds to support the project.
12.5.4.1 Other impacts
As far as the effects of investment in sport on community cohesion, social
capital, sports development and other externalities is concerned, the literature
is very sparse. One contribution, Downward and Ralston (2006), however,
TABLE 12.7 Contingent valuation analyses
Author Context Model and data WTP Total WTP
Johnson et al. (2000) New university basketball and Tobit Basketball $6.36 $610 293–$311 249
minor league baseball stadia n = 230 Minor league $59 061–$301 951 (over
for Kentucky baseball $6.17 95 958 residents)
Barros (2006) Stadium construction Tobit e0.106 (2000) e11.874 m (2000) (over
for Euro 2004 n = 1600 Portuguese 3.73 m households)
residents
Johnson et al. (2006) To ensure the NFL Jaguars Probit and truncated NBA
remained in Jacksonville and regressiona Range from
to attract an NBA team n = 367 and n = $37.02–$67.23 (per capita)
139 or 167 NFL
Range from
$94.24–147.70 (per capita)
Heyne, Maennig and Ex-ante and ex-poste Not stated Ex ante Ex ante
S€ussmuth (2007) valuations of $5.66 $467m
the 2006 World Cup
Cited in Maennig and
Du Plessis (2007)
Ex post Ex post
$13.38 $1.1 bn (Over 82 m
inhabitants)
EFTEC (2005) London 2012 Olympics Not stated London £22 £0.67 bn London
cited in Manchester and £3.2 bn outside London
Glasgow £12
Blake (2005), Price
Waterhouse Coopers (2005)
Empirical Evidence
377
378
Author Context Model and data WTP Total WTP
EEDA (2006)
Walton et al. (2007) London 2012 Olympics Weibull estimationb £42.20 £5 833 095
n = 167 or 147 (138 225 residents in Bath)
£173 271 934
(4 105 970 residents in
the southwest)
a
The two equation approach, described as a Heckman model in Chapter 4 allows the determinants of preparedness to pay to differ from the determinants of the amount that respondents are
willing to pay, unlike Tobit. A Probit relationship was used to identify the probability that a respondent will be willing to pay, and then a log-linear truncated regression model is estimated on the
sub-sample of those who presented a wtp value, as well as the inverse mills ratio which controls for the sample selection. The log-linear form was used because of the non-normality of the
distribution of the total wtp.
b
Weibull models are part of a class of models used to model “hazard ” in economics. Typically this occurs in labour economics to model the duration of unemployment. The dependent variable is
thus a discrete interval. Willingness to pay is thus identified as such an interval in this analysis.
Conclusion 379
examines a sample of 403 volunteers from the 2002 Manchester Common-

wealth Games to identify whether their experiences of the games raised their
interest in, and their potential for, greater sports participation and volunteer-
ing in sport. The results suggested that while volunteering at a major event can
raise interest, participation and volunteering in sport generally, particularly for
younger volunteers, there appeared to be a much stronger potential opportu-
nity to generate wider social capital, i.e., volunteering in non-sports contexts,
than necessarily producing changes associated with sport. It is clear that
further research of a longitudinal and wider population is required.
12.6 CONCLUSION
This chapter has explored elements of the economics of events, suggesting that
the design and allocation of non-professional team sports events can also be
understood with reference to contest theory. It has also been argued that
professional sports and events share other economic features, particularly
connected with the impact of sport provision that is, at least partially, funded
by public authorities. It has shown that both the economic rationale for, and
principles of evaluation of, any public sector investment in sports is common
to all sports contests, although naturally the scale and types of investments
differ. The definition of investment decisions has been discussed, as well as the
policy rationale for public sector investment in sport. It has been noted that
externalities and other spillover effects are perceived to play a central role in
evaluating sports investment decisions from a public sector perspective. The
chapter identifies that the most appropriate form of impact to consider is that
on economic welfare and this requires assessing both the tangible impacts of
investments, as well as its intangible aspects.
In the former case it has been argued that ex post evaluation of investment
decisions suggests that multiplier effects are weak. This may be due to the
supply-side of the economy producing crowding out effects or be indicative of
measurement errors. In either case, the evidence appears to be more reliable
than many ex ante studies in which large-scale impacts are often forecast. The
discussion also indicates that more reliability is associated with the measured
impacts of sports investment on intangible impacts, for example value to
communities, and expectations of future consumer confidence and produc-
tivity gains, although it is not clear that such gains would outweigh the costs of
the investments. The chapter concludes, therefore, by suggesting that while a
coherent theoretical case can be made for public sector investment in sports,
the evidence in support of this case is relatively weak, and suggests consider-
able caution and planning being required to harness the spillover effects from
the investments.
Appendix 12.1 NPV, Wealth and Profit Maximization

Recall from Appendix 3.3 that the present value of any amount “A” at interest rate “r” over “t”
periods:
Present value ¼ A=ð1 þ rÞt ðA12:1:1Þ
Consider Figure A12.1.1 below, it represents a production possibilities frontier for all of the
investment opportunities open to a supplier. The current period, “t” is represented by the
horizontal axis, and the future, “t + 1” is on the vertical axis. Note that the concave shape
indicates that the projects are ranked in order of their return. In other words, from a position of
current investment funds, S0, on the horizontal axis, investment of some of these funds, say S0 to
S1 can yield future funds of 0 to S2, etc. Wealth would comprise current resources 0 S1 plus
future resources 0 to S2, from the investment. It follows that each point on the curve indicates an
investment portfolio in which some resources are held over and some invested for the future. The
marginal rate of transformation of this curve measures the yield or internal rate of return of the
investments, “d.” Note that this must decline as more and more funds are invested in projects.
At which point on this curve will any investment decision maximize profits or wealth? To answer
this question recall that Equation A12.1.1 indicates the present value of any amount “A.”
Central to this calculation is the market rate of interest “r.” This represents the opportunity cost
of funds to the supplier, inasmuch as it represents what can be earned from placing resources
in savings, rather than investing in expanding their business, or the cost of borrowing money to
invest in the business. It follows that a profit maximizing supplier would only invest in the
business if the internal rate of return or yield from the business activity exceeded the market
rate of interest. At the margin this implies that investment would stop where the internal rate of
return, “d,” is equal to “r.” Drawing a line at slope “r” at the furthest point to the right of the
origin in Figure A12.1.1 would consequently identify the maximum wealth from any investment
indicated at point M, derived from the point of tangency where the slope of “r” would be equal
to “d.” This shows that the intercept of this line with the current time period axis must indicate
the NPV of any investment combination (and by analogy, the intercept of this line with the
vertical axis must indicate the future value of any investment). Note that the vertical distance
FIGURE A12.1.1 NPV and wealth maximization.

Conclusion 381
(M–S*) represents the future returns from the investment S0 to S*. In time “t,” this would be
equivalent to the horizontal distance (S*–SM). In other words, because for investments up to
S*, d > r, this investment must add more to wealth than simply placing resources in financial
markets. As wealth is the sum of flows of resources, maximizing wealth is simply the long-run
equivalent of maximizing the flow of profits.
Appendix 12.2 A Simple Input–Output Analysis of Multiplier Effects Given a

Simple EconomicSystem
TABLE A12.2.1 A simple input–output account e million

Consuming Industry Household Output
M S consumption total
Producing M 100 10 200 310
Industry S 60 10 60 130
Input total 160 20 260 440
The rows of the table indicate the value of transactions corresponding to the output of the
industry “X,” where “M” is manufacturing and “S” is sport, and the columns show how much
of this output is consumed by the sector itself, and by other sectors and consumers.
Consequently, the two columns of the “consuming industry” part of the table represent how
much output is bought and sold between industrial sectors, i.e., firms in the circular flow of
income, whereas the column “consumption” represents how much is purchased by
households, in the circular flow of income “D.” For simplicity, the government sector and
imports and exports are ignored, as are any other payments by industries to those not
subsequently involved in inter-industry trade. This might involve wages to employees. Note
that the row and column totals must sum to the same overall amount, as any input to an
industry must be balanced by an output.25
A set of input output coefficients can be calculated by dividing the purchases of each
industry, as obtained from the values in the columns of the input–output table of transactions,
by the total value of output of the industry making the purchase.
25
Strictly, each column total should be equivalent to the row total. For example manufacturing
produces e310 million output and provides e160 million to other industries. This leaves e150
million for disbursement to other industries outside of the trading system or to wages and profits
that might fund the household consumption, which is treated as exogenously given here. This
means that in calculating the input–output coefficients, technically, one should divide each
column entry by its total to indicate what proportion of the total inputs provided by the industry go
to each producing industry, including itself. Numerically speaking one gets the same answer. In
this example, to keep the numerical calculations simpler the row totals have been used, with the
same logic that one is illustrating with the input–output coefficients the proportion of the
industries output that is used as an input to itself and other industries.
TABLE A12.2.2 Input–output coefficients

M S
M 0.32 0.08
S 0.19 0.08
In equations this is:

0:32 M þ 0:08 S þ DM ¼ M ðA12:4:1Þ
0:19 M þ 0:08 S þ DS ¼ S
In matrix terms the input–output coefficient matrix is:

0:32 0:08
ðA12:4:2Þ
0:19 0:08
Consequently, given:
X ¼ ½I A1 D ðA12:4:3Þ
If the component of sports expenditure rises, say, by e10 million, but manufacturing demand is
constant, i.e., 0, then:

1:5 0:13 0 0 þ 1:3 1:3
¼ ¼ ðA12:4:4Þ
0:31 1:1 10 0 þ 11 11
There is an increase in manufacturing output of e1.3 million and an increase in sports output of
e11 million. The manufacturing demand rises, of course, because of the inter-industry transactions.
Appendix 12.3 Calculating WelfareWeights

Adopting a Cobb–Douglas utility function means that utility “U” can be derived as a function
of the value of consumption “C” of an individual:
U ¼ Cb ðA12:3:1Þ
or in natural logarithms (ln), as discussed in Chapter 4:

U ¼ blnC ðA12:3:2Þ
where b is the elasticity of utility with respect to consumption. The marginal utility of
consumption will be:
LU=LC ¼ b=C ðA12:3:3Þ
or if b = 1:
LU=LC ¼ 1=C ðA12:3:4Þ
or:
LU=LC* ¼ 1=C* ðA12:3:5Þ
if C* refers to some average level of consumption. This means that utility declines in direct
proportion to consumption from income in the case of Equation A12.5.4. A utility weight
relative to average income can then be defined as C*/C by dividing Equation A12.3.4 by
Equation A12.3.5.
Appendix 12.4 The Determinants of Medal Success
TABLE A12.4.1 Econometric evidence on performance (medals) in Olympic Games
De Bosscher, Olympic Summer Period Athens 2004 “population size is responsible

Heyndels, Games and for 19.3% of the international
De Knop, Van Winter Games Stepwise OLS regression Dependent variable success, wealth add another
Bottenburg and & Ln (Med) weighted points country 19.5% and together with the
Shibli (2008) & (gold = 3, silver = 2, bronze = 1) political system for (former)
communist countries, we end
Independent variables up with a model where 52.4%
& Ln Population, Ln GDP cap, Ln population
of the international success is
density, dummies Muslim, protestants, and explained”
(former) communist countries “the results revealed that only
Period Salt Lake City 2002 wealth and communism are
significant . . . with a
OLS regression determination coefficient of
Dependent variable 54.6%”
& Ln (Med) weighted points country
& (gold = 3, silver = 2, bronze = 1)
& Ln Population, Ln GDP cap, Ln population
density, dummies muslim, protestants, (former)

communist countries and mountains
Rathke and Olympic Summer Period 1952–2004 “GDP is a good predictor of

Woitek (2007) Games success . . . The effect of
Translog product function population, however, is
Dependent variable positive only for relative rich
& (medal share) countries . . . Host and Soviet
& Olympic diploma countries outperform . . .”
Independent variables (p.12)
2 2
& Ln of GDP, GDP , population, population ,
Conclusion
GDP.population
Dummies for host country, Soviet, plan (China,
Norrth Korea, Albania and Yugoslavia)
383
384
TABLE A12.4.1 (Continued )
Mitchell and Olympic Summer Period Athens 2004

Stewart (2007) Games
OLS and Poisson
Dependent variable
& total medals won by country
& weighted medals won by country
& people: population, birth rate, fertility, infant
mortality, life expectancy, literacy, migration

& government: dummy (former) communist
2
& economy: GDP, GDP , GDP/capita, GDP in
agriculture, industry and services, openness,

government budget, donor (development aid),
Gini index income lowest and highest 10%,
poverty,
& technology: cell phone
Other variables
& dummy Africa
Roberts (2006) Olympic Summer Period: Athens 2004 “GDP per capita, a relatively
Games cold winter climate and the age
Poisson and negative binomial count data dependency ratio all have
regression statistically significant
Dependent variable: impact. . . population, health
&total medals won by country expenditure per capita, host/
neighbour effect and rugby and
Independent variables: cricket effects all appear to
& GDP/capita, population, health expenditure per hold no explanatory power.”
capita, age dependency ratio(people <15 + >64 (p. 8)
divided by people 15–64), frost prevalence
(proportion land with more than 5 frost-days a
year), dummies for host or neighbour country,
test playing cricket nations and top 10 rugby
union nations
Hoffman, Chew Olympic Summer Period Sydney 2000 “all variables chosen were
Ging and Games OLS significant at a 90% confidence
Ramasamy (2004) interval”(p. 268)
Dependent variable
& total number medals won by country
(ASEAN countries)
& share of world population, share of world GNP,
dummies for (former) communist country, cooler

humid climate, host country, having hosted since
1948, having hosted twice since 1948
Bernard and Busse Olympic Summer Period Rome 1960–Atlanta 1996 “the results for population and
(2004) Games GDP per capita . . . remain
Tobit regression positive and significant . . . The
Dependent variable Sovjet countries . . . and other
& medal share of country planned economies have
Independent variable shares that are higher. . . The
log population, log GDP/capita, dummies host, host effect . . . is also positive
Sovjet, planned (not Sovjet), lagged medal share and significant . . . lagged
(indicator for current success) medal share . . . is strongly
significant” (p. 415)
Johnson and Ayfer Olympic Summer Period Summer 1952–1998 “per capita income has a
(2004) and Winter positive impact . . . more
Games Winter 1952–1996 populous nations win more
OLS with nation fixed effects medals . . . the home nation
has a strong advantage . . .
Dependent variable neighboring nations sharing
& all participants by nation marginally in the bounty in
& female participants by nation Summer Games . . . single
& all medals by nation party and communist systems
& gold medals by nation outperform . . . Cold nations
Independent variables outperform in both Summer
& GDP/Cap, squared GDP/cap, population, and Winter Games” (p. 984–
dummy host, neighbour, politics(monarchy, 987)
Conclusion
single-party, military or other), light and heavy
frost (share of land with less than 5 days and more
385
386
TABLE A12.4.1 (Continued )
than 20 days of winter), time trend, total medals

that can be won
Tcha and Pershin Olympic Summer Period 1988–1996 “countries with higher GDP
(2003) Games per capita . . . collect medals
Tobit regression and patterns of specializing in
from a diversified range of
sports by using the revealed comparative
sports ... Asian countries did
advantage (in line with neoclassical trade
not perform well in swimming
models)
. . . countries with RCA in
Dependent variable
& cumulative number of medals for each country
athletics are relatively well
endowed with land mass, high
and sport (swimming, athletics, weights, ball
altitude and hogh GDP per
games, gymnastics, others)
capita . . . AFD(Africa) also
& land mass, coast length, altitude, temperature
turns to be significant and
positive in athletics” (pp. 229–
GDP per capita, GDP and population
& dummies for African, Asian and former socialist
234)
countries
Hoffman, Chew Olympic Summer Period Sydney 2000 “significant quadratic
Ging and Games relationship between Sydney
Ramasamy (2002) OLS medal success and mean
Dependent variable temperature,. . . optimal
& number of medals won by country average temp of around
Independent variables 15 C. . . countries in warmer
& policy factors GNP/capita, population,
and colder humid climates are
expected to win an extra 14
dummies for current/previous communism,
and 20 medals respectively . . .
Sydney host, countries having hosted once,
All factors(policy) are
countries having hosted twice,
& deterministic factors average temperature in
significant” (pp. 247–548)
capital, squared average temp, dummies for
cooler humid zone, warmer humid zone, dry zone
and tropical humid zone
Stamm and Olympic Summer Period 1964–1980, 1984–2000, 1992,1996, “of particular importance are
Lamprecht Games 2000 population size . . . and level of
(2001) economic development” (p. 5)
OLS
Dependent variable
& total medals won by country
& final participation
& GDP per capita, secondary education
enrolment, extent of political and civil liberties,

population size, sporting tradition, measured as
degree of institutionalisation of elite sports: year,
duration of IOC membership
& political system: dummy for socialist countries;
effect of authoritarian model of sport promotion

Den Butter and Olympic Summer Period Seoul 1988 and Barcelona 1992 “All regressions prove that the
Van der Tak (1995) Games level of welfare is a major
Bayesian method of Poission regression determinant of the relative
Dependent variable performance of nations at the
& number of medals country OG multidimensional welfare
Independent variables indicators do not outperform
& equation 1,2,3: population, dummy (former) NI . . . per capita income
communist country, GDP/Pop (several forms is elasticity can be set equal to
several equations such as national income/pop unity, but the elasticity with
(eq1), NIpurchasing power/pop(eq2), “real” respect to population size
national income/pop (eq3) appears to be smaller than
& equation 4,5: log pop, Human development one” (pp. 34–35)
index (eq4), log quality of life index (eq5), dummy
communist
Conclusion
387
References
Abbott, M., and Ashenfelter, O. (1976) Labour supply, commodity demand and the
allocation of time. Review of Economic Studies 43: 389–411.
Adams, F., Davidson, P. and Seneca, J. (1968) An Analysis of Recreational Use of the
TVA Lakes, reprinted in Davidson, L. (ed.) (1991) The Collected Writings of Paul
Davidson, Vol. 2, Inflation,Open Economies and Resources. London: Macmillan.
Alavy, K., Gaskell, A., Leach, S. and Szymanski, S. (2006) On the Edge of your Seat:
Demand for Football on Television and the Uncertainty of Outcome Hypothesis.
International Association of Sports Economists Working Paper, 06-31.
Allan, P. (2004) Satellite television and football attendance: The not so super effect.
Applied Economics Letters 11: 123–125.
Altergot, K., and McCreedy, C. (1993) Gender and family status across the lifecourse:
Constraints on five types of leisure. Society and Leisure 16, 1: 151–180.
Andreff, W., and Staudohar, P. (2002) European and US sports business models. In:
Barros, C., Ibrahimo, M. and Szymanski, S. (eds). Transatlantic sport: The com-
parative economics of North American and European sports. Edward Elgar, Chel-
tenham.
Arrow, K. (1950) A difficulty in the concept of social welfare. Journal of Political Econ-
omy 58: 328–346.
Arnout, J., (2006) Independent European Sports Review, European Union.
Arrow, K. (1963) Social Choice and Individual Values. John Wiley and Sons, New York.
Arthur, Andersen (1999) Economic Impact Study of the Sydney 2000 Olympic Games.
Centre for Regional Economic Analysis. University of Tasmania, Australia.
Asser (2005) Professional Sport in the Internal market, Working paper, DG Internal
Policies of the Union – Directorate A – Economic and Scientific Policy, European
parliament, Brussels.
Ashton, J. K., Gerrard, B., and Hudson, R. (2003) Economic impact of national sporting
success: Evidence from the London stock exchange. Applied Economics Letters 10:
783–785.
Audas, R., Dobson, S., and Goddard, J. (1999) Organisational performance and man-
agerial turnover. Managerial and Decision Economics 20: 305–318.
Australian Sports Commission (2005) Participation in Exercise and Sport. Annual
Report 2005.
Australian Bureau of Statistics. (2005) Sport Volunteers and Other Volunteers: Some
Data from the 2002 General Social Survey. National Centre for Culture and Re-
creation Statistics.
Australian Sports Commission (2004) Participation in Exercise, Recreation and Sport:
Annual Report. Belconnen, ACT.
Baade, R., and Matheson, V. A. (2002) Bidding for the Olympics: Fools’ Gold? In: Barros,
C. et al. (eds). Transatlantic Sport: The Comparative Economics of North American
and European Sports. Edward Elgar, Cheltenham.
Baade, R., and Matheson, V. A. (2004) The quest for the cup: Assessing the economic
impact of the world cup. Regional Studies 38, 4: 343–354.
Baimbridge, M., Cameron, S., and Dawson, P. (1995) Satellite broadcasting and
match attendance: The case of rugby league. Applied Economics Letters 2, 10:
343–346.
389
390 References
Baimbridge, M., Cameron, S., and Dawson, P. (1996) Satellite broadcasting and the
demand for football: A whole new ball game? Scottish Journal of Political Economy
43, 3: 317–333.
Balfousia-Savva, S., Athanassiou, L., Zaragas, L., and Milonas, A. (2001) The Economic
Effects of the Athens Olympic Games. Centre of Planning and Economic Research,
Athens, Greece.
Barajas, A. and Crolley, L. (2005) A model to explain support in Spanish football,
University of Vigo September, Online at http://mpra.ub.uni-muenchen.de/3235/
MPRA Paper No. 3235.
Barget, E., and Gouguet, J. (2007) The total economic value of sporting events; theory
and practice. Journal of Sports Economics 8, 2: 165–182.
Barnett, V. and Hilditch, S. (1993) The effect of an artificial pitch surface on home team
performance in football (soccer). Journal of the Royal Statistical Society, A, part 156,
39–50.
Barros, C. (2006) Evaluating sports events at European level: The Euro 2004. Interna-
tional Journal of Sport Management and Marketing 1, 4: 400–410.
Becker, B., and Huselid, M. (1992) The incentive effects of tournament compensation
systems. Administrative Science Quarterly 37: 336–350.
Becker, G. (1965) A theory of the allocation of time. Economic Journal 75, 299:
493–517.
Becker, G. (1968) Crime and punishment: An economic approach. Journal of Political
Economy 76: 169–217.
Becker, G. (1974) A theory of social interactions. Journal of Political Economy 82:
1063–1091.
Becker, G. (1976) The Economic Approach to Human Behaviour. The University of
Chicago Press, Chicago, IL.
Becker, G. (1992) The economic way of looking at life. The Journal of Political Economy
101: 385–409.
Bergson, A. (1938) A reformulation of certain aspects of welfare economics. Quarterly
Journal of Economics 52, 2: 310–334.
Berman, G., Brooks, R., and Davidson, S. (2000) The Sydney Olympic Games an-
nouncement and Australian stock market reaction. Applied Economics Letters 7:
781–784.
Bernard, A., and Busse, M. (2004) Who wins the Olympic Games: Economic resources
and medals totals. The Review of Economics and Statistics 86, 1: 413–417.
Bittman, M., and Wajcman, J. (2000) The rush hour: The character of leisure time and
gender equity. Social Forces 79, 1: 165–189.
Blake, A. (2005) The Economic Impact of the London 2012. Olympics Working paper,
No 5 Christel DeHaan Tourism and Travel Research Institute, Nottingham Uni-
versity Business School, Nottingham, UK.
Blavatskyy, P. (2004) Why the Olympics have three prizes and not just one. Institute for
Empirical Research in Economics, December, University of Zurich, Switzerland.
Bognanno, M.L. (1990), ‘An Empirical Test of Tournament Theory’, Ph.D. Thesis, Depart-
ment of Economics, Cornell University.
Bohlmann, H., and van Heerden, J. (2008) Predicting the economic impact of the 2010
FIFA world cup on South Africa. International Journal of Sport Management and
Marketing 3, 4: 383–396.
References 391
Borland, J. (1987) The demand for Australian Rules football. The Economic Record 63,
182: 220–230.
Borland, J., and Lye, J. (1992) Attendance at Australian Rules football: A panel study.
Applied Economics 24, 9: 1053–1058.
Borland, J., and MacDonald, R. (2003) Demand for sport. Oxford Review of Economic
Policy 19, 4: 478–502.
Bourdieu, P. (1986) The forms of capital. In: Richardson, JG. (ed). Handbook of Theory
and Research for the Sociology of Education. Greenwood Press, New York.
Boyko, R. H., Boyko, A. R., and Boyko, M. G. (2007) Referee bias contributes to home
advantage in English premiership football. Journal of Sports Sciences 25:
1185–1194.
Breuer, C. (2006) Sport Participation in Germany – a demo-economical model. German
Sport University Cologne, Discussion paper 01/06.
Breuer, C. (2007) Sport-entwicklungsbericht 2005/2006: Analyse zur Situation der
Sportvereine in Deutschland. Bundesinstitut f€ ur Sportwissenschaft, Bonn.
Brown, D., and Link, C. (2008) Population and bandwagon effects on local team
revenues in MLB. Journal of Sports Economics 10: 1–18.
Brown, E. (1999) Assessing the value of volunteer activity. Non-Profit and Voluntary
Sector Quarterly 28, 1: 3–17.
Brunet, F. (1995) An economic impact analysis of the Barcelona ’92 Olympic Games:
Resources, financing and impact. http://olympicstudies.uab.es/pdf/OD006_eng.pdf.
Brunet, F. (2005) The economic impact of the Barcelona Olympic Games, 1986–2004.
www.olympicstudies.uab.es.
Buraimo, B., Forrest, D. and Simmons, R. (2006) Insights for clubs from modelling
match attendance in football. Working paper, Lancashire University.
Buraimo, B. and Simmons, R. (2007) A tale of two audiences: spectators, television
viewers and outcome uncertainty in Spanish football. Lancaster University Man-
agement School Working Paper 2007/043.
Burgham, M., and Downward, P. (2005) Why volunteer, time to volunteer? A case study
from swimming. Managing Leisure 10: 79–93.
Butler, M. (2002) Interleague play and baseball attendance. Journal of Sports Economics
3: 320–334.
Cairns, J. (1990) The demand for professional team sports. British Review of Economic
Issues 12, 28: 1–20.
Cairns, J., Jennett, N., and Sloane, P. (1986) The economics of professional team sports:
A survey of theory and evidence. Journal of Economic Studies 13, 1: 3–80.
Cambridge Policy Consultants (2002) The Commonwealth Games 2002: A Cost and
Benefit Analysis, Revised Executive Summary. http://www.gameslegacy.com/files/
CG2002%20-%20SUMMARY%20REPORT%20-%20OCTOBER%202002.doc.
Carmichael, F., Forrest, D., and Simmons, R. (1999) The labour market in association
football: who gets transferred and for how much? Bulletin of Economic Research
51, 2: 125–150.
Carmichael, F., Millington, J., and Simmons, R. (1999) Elasticity of demand for rugby
league attendance and the impact of BskyB. Applied Economics Letters 6: 707–800.
Carmichael, F., and Thomas, D. (1993) Bargaining in the transfer market: Theory and
evidence. Applied Economics 25: 1467–1476.
Carter, P. (2005) Review of National Sport Effort and Resources. DCMS, London.
392 References
Cicchetti, C., Davidson, P., and Seneca, J. (1969) The Demand and Supply of Outdoor
Recreation: An Econometric Analysis. Bureau of Outdoor Recreation, Washington,
DC.
Clapp, C., and Hakes, J. (2005) How long a honeymoon? The effect of new stadiums on
attendance in major league baseball. Journal of Sports Economics 6: 237–263.
Clarke, S. (2005) Home advantage in the Australian football league. Journal of Sports
Sciences 23, 4: 375–385.
Cnaan, R., Handy, F., and Wadsworth, M. (1996) Defining who is a volunteer: Con-
ceptual and empirical considerations. Nonprofit and Voluntary Sector Quarterly 25,
3: 364–383.
Coalter, F. (1999) Sport and recreation in the United Kingdom: Flow with the flow or
buck the trends? Managing Leisure 4: 24–39.
Coase, R. (1960) The problem of social cost. Journal of Law and Economics 3: 1–44.
Coates, D., and Harrison, T. (2005) Baseball strikes and the demand for attendance.
Journal of Sports Economics 6: 282–302.
Coates, D., and Humphreys, B. (1999) The growth effects of sports franchises, stadia
and arenas. Journal of Policy Analysis and Management 18, 4: 601–624.
Coates, D., and Humphreys, B. (2002) The economic impact of post-season play in
professional sports. Journal of Sports Economics 3, 3: 291–299.
Coates, D., and Humphreys, B. (2005) Novelty effects of new facilities on attendance at
professional sporting events. Contemporary Economic Policy 23, 3: 436–455.
Coates, D., and Humphreys, B. (2007) Ticket prices, concessions and attendance at
professional sporting events. International Journal of Sport Finance 3, 2: 161–170.
Collins, M. (2003) Sport and Social Exclusion. Routledge, London.
Collins, M. (2004) Driving up participation. Social Inclusion, in Sport England. (2004).
Cornes, R., and Sandler, T. (1986) The Theory of Externalities, Public Goods and Club
Goods. Cambridge University Press, Cambridge, UK.
Council of Europe (1980) European Sport for All Charter. Strasbourg.
Council of Europe, (1992) European Sports Charter, http://www.coe.int/t/dg4/sport/
SportinEurope/charter_en.asp.
Crompton, J. (1995) Economic impact analysis of sports facilities and events: Eleven
sources of misapplication. Journal of Sports Management 9: 14–35.
Crompton, J. (2004) Beyond economic impact: An alternative rationale for the subsidy
of major league sports facilities. Journal of Sports Management 18: 40–58.
Crompton, J. (2006) Economic impact studies: Instruments for political shenanigans?
Journal of Travel Research 45: 67–82.
Cuskelly, G., Hoye, R., and Auld, C. (2006) Working with Volunteers in Sport: Theory
and Practice. Routledge, London.
Dardis, R., Soberon-Ferrer, H., and Patro, D. (1994) Analysis of leisure expenditures in
the United States. Journal of Leisure Research 26, 4: 309–321.
Davies, B., Downward, P., and Jackson, I. (1995) The demand for rugby league: Evi-
dence from causality tests. Applied Economics 27, 10: 1003–1007.
Davies, L. (2004) Valuing the voluntary sector in sport: Rethinking economic analysis.
Leisure Studies 23, 4: 347–364.
Dawson, P., and Dobson, S. (2002) Managerial efficiency and human capital: An
application to English association football. Managerial and Decision Economics
23: 471–486.
References 393
Dawson, P. and Dobson, S., (2006) Managerial Employment and Re-Employment

Patterns: A Case Study of English Professional Football. In: Jeanrenaud, C. and
Kesenne, S. (eds) conomics Applied to Sports: Five Case Studies, CIES Editions.
Dawson, P., Dobson, S., and Gerrard, B. (2000) Estimating coaching efficiency in
professional team sports: Evidence from English association football. Scottish Jour-
nal of Political Economy 47: 399–421.
Dawson, P., Dobson, S., Goddard, J. and Wilson, J. (2005). Modelling the incidence of
disciplinary sanction in English premier league football. Paper delivered at the
University of Groningen, March 2005.
DCMS/Strategy Unit (2002) Game Plan: A strategy for delivering government's sport
and physical activity objectives.
DCMS/Strategy Unit (2008) Playing to win.
De Bosscher, V., Bingham, J., Shibli, S., Van Bottenburg, M. and De Knop, P. (forth-
coming 2007) Sports Policy Factors Leading to International Sporting Success: An
International Comparative Study. Oxford, UK: Meyer and Meyer Sport.
Dejonghe, T., Van Hoof, S. and Kemmeren, T. (2006) Voetballen in de kleine ruimte: een
onderzoek naar de geografische marktgebieden en ruimtelijke uitbreidingsmogelij-
kheden voor de clubs in het Nederlandse Betaalde Voetbal, Nieuwegein, Mulier
Instituut, Arko Sports Media.
Dejonghe, T. (2007a) Sport en economie: een aftrap. Arko Sports Media, Nieuwegein.
Dejonghe, T. (2007b) Sport in de wereld: ontstaan, verspreiding en evolutie.
Academia Press, Gent.
Deloitte, Touche (2005a) Annual review of football and finance. Deloitte and Touche,
Manchester.
Deloitte (2005b) The Football Money League: the Climbers and the Sliders. Manche-
ster.
Deloitte (2007) Annual Review of Football Finance, Manchester.
Demmert, H. (1973) The Economics of Professional Team Sports. Heath, D.C.
Den Butter, F., and Van Der Tak, C. (1995) Olympic medals as an indicator of social
welfare. Social Indicators Research 35: 27–37.
Dimitrakopoulos, D.G., (2006) More than a Market? The Regulation of Sport in the
European Union, Government and Opposition 41: 561–580.
Dobson, S., and Gerrard, B. (1999) The determination of player transfer fees in English
professional soccer. Journal of Sport Management 13: 259–279.
Dobson, S., and Goddard, J. (1992) The demand for standing and seated viewing in
English non-league football. Applied Economics 24, 10: 1155–1164.
Dobson, S., and Goddard, J. (1998) Performance, revenue and cross subsidis-ation in
the football league 1927–1994. Economic History Review L1, 4: 763–765.
Dobson, S., and Goddard, J. (2001) The Economics of Football. CUP, Cambridge.
Dobson, S., Goddard, J., and Ramlogan, C. (2001) Revenue convergence in the English
soccer league. Journal of Sports Economics 2, part 3: 257–274.
Downward, P.M. (2003), Sport and Sport Constructions in the Regional Competition of
Cities: Some Observations from the USA and UK. In Buch, M.P., Maennig, W. and
Schulke, H.J. Nachhaltigkeit von Sportstatten, Bundesinstitut fur Sportwis-
senschaft, Bonn.
Downward, P. (2004) Assessing neoclassical microeconomic theory via leisure demand:
A post keynesian perspective. Journal of Post Keynesian Economics 26, 3: 371–395.
394 References
Downward, P. (2005) Understanding Participation in Sport, Leisure and Tourism. In:

Downward, P. and Lumsdon, L. (eds). Essential Data Skills: Leisure and Tourism. A
Guide to using Official Data Sources. The Stationery Office, London.
Downward, P. (2007) Exploring the economic choice to participate in sport: Results
from the 2002 General Household Survey. The International Review of Applied
Economics 21, 5: 633–653.
Downward, P., and Dawson, A. (2000) The Economics of Professional Team Sports.
Routledge, London.
Downward, P., and Dawson, A. (2005a) Measuring short-run uncertainty of outcome in
sporting leagues: A comment. Journal of Sports Economics 6, 3: 303–313.
Downward, P. and Dawson, A. (2005b) Anomalies in the economics of professional
team sports: examples from short and medium term models of outcome uncer-
tainty. Paper delivered at the University of Groningen, March 2005.
Downward, P., and Jackson, I. (2003) Common origins, Common Future? A Compar-
ative Analysis of Association and Rugby league Football in the UK. In: Fort, R. and
Fizel, J. (eds). International Comparisons in the Economics of Sports. Praeger.
Downward, P., and Jones, M. (2007) Exploring crowd-size effects upon referee deci-
sions: Analysis of the FA cup. Journal of Sports Sciences 25, 14: 1541–1545.
Downward, P., and Ralston, R. (2006) The sports development potential of sports event
volunteering: Insights from the XV11 Manchester Commonwealth Games. Euro-
pean Sport Management Quarterly 6, 4: 333–352.
Downward, P., and Riordan, J. (2007) Social interactions and the demand for sport: An
economic analysis. Contemporary Economic Policy 25, 4: 518–537.
Dwyer, L., Forsyth, P., and Spurr, R. (2006) Economic Evaluation of Special Events. In:
Dwyer, L. and Forsyth, P. (eds). International Handbook on the Economics of
Tourism. Edward Elgar, Cheltenham.
Dwyer, L., and Forsyth, P. (2006) International Handbook on the Economics of Tourism.
Edward Elgar, Cheltenham.
East of England Development Agency (EEDA) (2006) Economic Impact Study of the
London 2012 Olympic Games and Paralympic Games.
Eckard, W. (2001) Baseball's blue ribbon report: Solutions in search of a problem.
Journal of Sports Economics 2, 3: 213–227.
Economic Research Associates. (1984) Community Economic Impact of the 1984
Olympic Games in Los Angeles and Southern California. Los Angeles Olympic
Organizing Committee, Los Angeles.
Economics for the Environment Consultancy (2005) Olympic Games Impact Study –
Stated Preference Analysis: Final Report, December 2005, Department of Culture
Media and Sport. London.
Ehrenburg, R., and Bognanno, L. (1990a) Do tournaments have incentive effects?
Journal of Political Economy 98: 1307–1324.
Ehrenburg, R., and Bognanno, L. (1990b) The incentive effects of tournaments revis-
ited: Evidence from the European PGA tour. Industrial Relations Review 43: 74–88.
El-Hodiri, M., and Quirk, J. (1971) An economic model of a professional sports league.
The Journal of Political Economy 79, 6: 1302–1319.
European Commission (1999) The European Model of Sport. Brussels.
Falter, J.-M., and Perignon, C. (2000) Demand for football and intra-match winning
probability: An essay on the glorious uncertainty of sports. Applied Economics 32:
1757–1765.
References 395
Falter, J.-M., Perignon, C., and Vercruysse, O. (2008) Impact of overwhelming joy on
consumer demand: The case of a world cup victory. Journal of Sports Economics 9:
20–42.
Farrell, L., and Shields, M. (2002) Investigating the economic and demographic deter-
minants of sporting participation in England. Journal of the Royal Statistical Society
(A) 165, 2: 335–348.
Feddersen, A., Maennig, W., and Borcherding, M. (2006) The novelty effect of new
soccer stadia: The case of Germany. International Journal of Sport Finance 1:
174–188.
Feddersen, A. and Maennig, W. (2007) Arenas versus multi-functional stadia – which
do spectators prefer? Working paper, University of Hamburg.
Fernie, S., and Metcalf, D. (1996) It's not what you pay it's the way that you pay it
–jockeys’ pay and performance. Centre Piece 2: 2–6.
Fernie, S., and Metcalf, D. (1999) It's not what you pay it's the way that you pay it.
That's what gets results – jockeys’ pay and performance. Labour 13: 385–411.
Forrest, D., Beaumont, J., Goddard, J., and Simmons, R. (2005) Home advantage and
the debate about competitive balance in professional sports leagues. Journal of
Sports Sciences 23, 4: 439–445.
Forrest, D. and Simmons, R. (2000) Outcome uncertainty and attendance demand in
sport: the case of English soccer. Discussion Paper, Centre for Sports Economics,
University of Salford.
Forrest, D., and Simmons, R. (2006) New issues in attendances demand: The case of
the English football league. Journal of Sports Economics 7: 247–266.
Forrest, D., Simmons, R., and Buraimo, B. (2005) Outcome uncertainty and the couch
potato audience. Scottish Journal of Political Economy 52, 4: 641–661.
Forrest, D., Simmons, R., and Feehan, P. (2002) A spatial cross-sectional analysis of the
elasticity of demand for soccer. Scottish Journal of Political Economy 49, 3:
336–355.
Forrest, D., Simmons, R., and Szymanski, S. (2004) Broadcasting, attendance and the
inefficiency of cartels. Review of Industrial Organization 24: 243–265.
Fort, R. (1997) Direct Democracy and the Stadium Mess. In: Noll and Zimbalist, (eds).
Sports, Jobs, and Taxes: The Economic Impact of Sports Teams and Stadiums.
Brookings Institution, Washington D.C.
Fort, R. (2000) European and North American sports differences(?). Scottish Journal of
Political Economy 47, 4: 431–455.
Fort, R. (2004) Inelastic sports pricing. Managerial and Decision Economics 25: 87–94.
Fort, R., and Quirk, J. (1995) Cross-subsidization, incentives, and outcomes in profes-
sional team sports leagues. Journal of Economic Literature 33, 3: 1265–1299.
Fort, R., and Quirk, J. (2004) Owner objectives and competitive balance. Journal of
Sports Economics 5, 1: 20–32.
Fox, K. and Rickards, L. (2004) Results from the sport and leisure module of the 2002
General Household Survey, TSO. London.
Freeman, R. B. (1997) Working for nothing: The supply of volunteer labour. Journal of
Labor Economics 15, 1: S140–S166.
Frick, B. (2000) Contest theory and sport. Oxford Review of Economic Policy 19, 4:
512–529.
Frick, B. (2003) Contest Theory and Sport, Oxford eview of Economic Policy 19: 512–
529.
396 References
Frick, B. (2006) Salary Determination and the Pay-Performance Relationship in Pro-

fessional Soccer: Evidence from Germany. In: Rodriguez, P., Kesenne, S. and Garcia,
J. (eds). Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg.
University of Oviedo, Oviedo.
Frick, B. (2007) The football players’ labour market: Empirical evidence from the major
European leagues. Scottish Journal of Political Economy 54, 3: 422–446.
Frick, B., Pietzner, G., and Prinz, J. (2006) The Labour market for Soccer Players in
Germany – Career Duration in a Competitive Environment. In: Jeanrenaud, C. and
Kesenne, S. (eds). Economics Applied to Sports: Five Case Studies. Cies, Neuch^atel,
Suisse.
Friedman, M. (1953) The Methodology of Positive Economics in Essays in Positive
Economics. University of Chicago Press, Chicago.
Garcia, J., and Rodriguez, P. (2002) The determinants of football match attendance
revisited. Journal of Sports Economics 3, 1: 18–38.
Getz, D. (1991) Festivals, Special Events and Tourism. Van Nostrand Reinhold.
Gibrat, R. (1931) Las Inegalites Economiques. Libraire du Recueil Sirey, Paris.
Gouguet, J. (2002) Economic impact of sporting events: What has to be measured? In:
Barros, C., Ibrahimon, M. and Szymanski, S. (2002). Transatlantic Sport: The
Comparative Economics of North American and European Sports. Cheltenham:
Edward Elgar.
Grant Thornton (2003), SA 2010 Soccer World Cup Bid Executive Summary. www.
polity.org.za.
Gratton, C., and Taylor, P. (1985) Sport and Recreation: An Economic Analysis. E. and F.
N. Spon, London.
Gratton, C., and Taylor, P. (2000) Economics of Sport and Recreation. E. and F.N. Spon,
London.
Gratton, C. and Tice, A. (1991) Sports Participation in Britain: An Investigation into
the Changing Demand for Sport. Manchester Polytechnic Working Paper.
Gratton, C., Dobson, N., and Shibli, S. (2000) The economic importance of major
sports events: A case study of six events. Managing Leisure 5: 17–28.
Gratton, C., Nichols, G., Shibli, S., and Taylor, P. (1997) Valuing Volunteers in UK
Sport. Sports Council, England.
Gratton, C., Shibli, S., and Coleman, R. (2005) Sport and economic regeneration in
cities. Urban Studies 42, 5–6: 985–999.
Gravelle, H., and Rees, R. (2004) Microeconomics. Prentice Hall, London.
Green, M. (2004) Changing policy priorities for sport in England: The emergence of
elite sport development as a key policy concern. Leisure Studies 23, 4: 365–385.
Green, M., and Houlihan, B. (2005) Elite Sport Development: Policy Learning and
Political Priorities. Routledge, London.
Gronau, R. (1973) The intrafamily allocation of time: The value of the housewives’
time. American Economic Review 63, 4: 634–651.
Grossman, M. (1972a) The Demand for Health: A Theoretical and Empirical Investi-
gation. National Bureau of Economic Research, New York.
Grossman, M. (1972b) On the concept of health capital and the demand for health.
Journal of Political Economy 80: 223–255.
Hart, R., Hutton, J. and Sharot, T. (1975) A statistical analysis of association football
attendance. Journal of the Royal Statistical Society; Series C (Applied Statistics), vol.
24, 1: 17–27.
References 397
Heckman, J. (1979) Sample selection bias as a specification error. Econometrica 47:

153–161.
Heinemann, K. (1999) Sports Clubs in Europe. In: Heinemann, K. (ed). Sports Clubs in
Various European Countries. Hofmann, Schorndorf.
Heinemann, K., and Schubert, M. (1999) Sports Clubs in Germany. In: Heinemann, K.
(ed). Sports Clubs in Various European Countries. Hofmann, Schorndorf.
Henry, I. (2001) The Politics of Leisure Policy. Palgrave, London.
Heyne, M., Maennig, W. and S€ ussmuth, B. (2007) Mega-sporting Events as Experience
Goods? Working Paper, Bremen University, University of Technologie and Ham-
burg University.
Hicks, J. (1939) Value and Capital. Clarendon Press, Oxford.
Hoehn, T., and Szymanski, S. (1999) The Americanisation of European football. Eco-
nomic Policy 14, 18: 205–240.
Hoffmann, R., Chew Ging, L., and Ramasamy, B. (2002) Public policy and Olympic
success. Applied Economic Letters 9: 545–548.
Hoffman, R., Chew Ging, L., and Ramasamy, B. (2004) Olympic success and ASEAN
countries: economic analysis and policy implications. Journal of Sports Economics
5: 262–276.
Holt, M., Mitchie, J., and Oughton, C. (2006) The Role and Regulation of Agents in
Football. The Sport Nexus.
Horch, H. (1994) Resource composition and oligarchization: Evidence from German
sport clubs. European Journal for Sport Management 2, 1: 52–67.
Houlihan, B. (1997) Sport, Policy and Politics: A Comparative Analysis. Routledge,
London.
Houlihan, B. (2005) In: Slack, T. (ed). The Commercialisation of Sport. Frank Cass,
London.
Hudson, I. (2001) The use and misuse of economic impact analysis: The case of
professional sports. Journal of Sport and Social Issues 25, 1: 20–39.
Humphreys, B. (2002) Alternative measures of competitive balance in sports leagues.
Journal of Sports Economics 3, 2: 133–148.
Humphreys, B. and Ruseski, J. (2006) Economic Determinants of Participation in
Physical Activity and Sport. IASE Conference, Bochum.
Humphreys, J. and Plummer, M. (1995) The Economic Impact on the State of Georgia
of hosting the 1996 Olympic Games: 1995 update. Selig Center for Economic
Growth, Terry College of Business, University of Georgia.
Ibsen, B. (1999) Structure and Development of Sports Organisations in Denmark. In:
Heinemann, K. (ed). Sports Clubs in Various European Countries. Hofmann,
Schorndorf.
Insight Economics. (2006) Triple Bottom Line Assessment of the XVIII Commonwealth
Games: Executive Summary. Insight Economics, Melbourne/Perth.
Jacklin, P. (2005) Temporal changes in home advantage in English football since the
Second World War: What explains improved away performance? Journal of Sports
Sciences 23, 7: 669–679.
Jackson, N. (1899) Association Football. George Newnes, London.
Jennett, N. (1984) Attendances, uncertainty of outcome and policy in Scottish league
football. Scottish Journal of Political Economy 31, 2: 175–197.
398 References
Jeanreneaud, C., and Kesenne, S. (2006) Sport and the Media an Overview. In: Jeanre-
naud, C. and Kesenne, S. (eds). The Economics of Sport and the Media. Edward
Elgar, Cheltenham, UK.
Johnson, B., and Whitehead, J. (2000) The value of public goods from sports stadiums:
The CVM approach. Contemporary Economic Policy 18, 1: 48–58.
Johnson, B., Mondello, M., and Whitehead, J. (2006) Contingent valuation of sports:
temporal embedding and ordering effects. Journal of Sports Economics 7, 3:
267–288.
Johnson, D., and Ayfer, A. (2004) A tale of two seasons: participation and medal counts
at the Summer and Winter Olympic Games. Social Science Quarterly 85, 4:
974–993.
Johnston, R. (2008) On referee bias, crowd size, and home advantage in the English
soccer premiership. Journal of Sports Sciences 26, 6: 563–568.
Jones, J., Schofield, J., and Giles, D. (2000) Our fans in the North: the demand for
British rugby league. Applied Economics 32, 14: 1879–1890.
Joyce, K. (1997) The Ethics and Dynamics of Negotiating a Professional Sports Con-
tract. Texas Entetainment and Sports Law Journal 6: 7–11.
Kahn, L. (1993) Free agency, long-term contracts and compensation in MLB: Estimates
from panel data. Review of Economics and Statistics 75: 157–164.
Kaldor, N. (1939) Welfare propositions of economics and interpersonal comparisons of
utility. Economic Journal 49: 549–552.
Karcher, R., (2005) The NCAA's Regulations Related to the Use of Agents in the Sport of
baseball: Are the Rules Detrimental to the Best Interests of the Amateur Athlete?
Vanderbilt Journal of Entertainment and Law. (http://ssrn.com/ abstract=936898).
Kasimati, E. (2003) Economic aspects and the Summer Olympics: A review of related
research. International Journal of Tourism Research 5: 433–444.
Kesenne, S. (1996) League management in professional team sports with win max-
imising clubs. European Journal for Sports Management 2: 14–22.
Kesenne, S. (2000b) The impact of salary caps in professional team sports. Scottish
Journal of Political Economy 47, 4: 422–429.
Kesenne, S. (2000a) Revenue sharing and competitive balance in competitive team
sports. Journal of Sports Economics 1, 1: 56–65.
Kesenne, S. (2001) The different impact of different sharing systems on the competitive
balance in professional team sports. European Sports Management Quarterly 1:
210–218.
Kesenne, S. (2004) Competitive balance and revenue sharing when rich clubs have poor
teams. Journal of Sports Economics 5: 206–212.
Kesenne, S. (2005) Do we need and economic impact study or a cost-benefit analysis of
a sports event. European Sport Management Quarterly 5: 133–142.
Kesenne, S. (2006b) The win maximisation model reconsidered. Journal of Sports
Economics 5: 981–1006.
Kesenne, S. (2006a) Competitive balance in team sports and the impact of revenue
sharing. Journal of Sport Management 20: 39–51.
Kesenne, S. (2007a) The Economic Theory of Professional Team Sports. Edward Elgar,
Cheltenham, UK.
Kesenne, S. (2007b) The peculiar international economics of professional football in
Europe. Scottish Journal of Political Economy 54, 3: 388–399.
References 399
Kesenne, S., and Janssens, P. (1987) Belgian soccer attendances. Tijdschrift voor Econ-
omie en Management 3: 379–400.
Kim, J., Rhee, S., and Ju, J. (1989) Impact of the Seoul Olympic Games on National
Development. Korea Development Institute, Seoul.
Koning, R. (2004) Home Advantage in Speed Skating: Evidence from Individual Data.
Paper given at the University of Groningen, March 2005.
Kooreman, P., and Kapteyn, A. (1987) A disaggregated analysis of the allocation of time
within the household. The Journal of Political Economy 95, 2: 223– 249.
Koski, P. (1999) Characteristics and contemporary trends of sports clubs in the Finnish
context. In: Heinemann, K. (ed). Sport Clubs in Various European Countries. Hof-
mann, Schorndorf.
KPMG. (1993) Sydney Olympics 2000: Economic Impact Study. Sydney Olympics
2000 Bid Ltd., Sydney, NSW.
Krautmann, A., and Oppenheimer, M. (2002) Contract length and the returns to
performance in major league baseball. Journal of Sports Economics 3: 5–17.
Kuypers, T. (1996) The beautiful game? An econometric study of why people watch
English football. University College London Discussion Papers in Economics ,
pp. 96–101.
Lallemand, T., Plasman, R., and Rycx, F. (2008) Women and competition in elimina-
tion tournaments: evidence from professional tennis data. Journal of Sports Eco-
nomics 9, 1: 3–19.
Lamb, L., Asturias, L., Roberts, K., and Brodie, D. (1992) Sports participation – How
much does it cost? Leisure Studies 11, 1: 19–29.
Larsen, A., Fenn, A., and Spenner, E. (2006) The impact of free agency and the salary
cap on competitive balance in the NFL. Journal of Sports Economics 7, 4: 374–390.
Lawson, T. (2003) Reorienting Economics. Routledge, London.
Lechner, M. (2008) Long-run labour market effects of individual sports activities. IZA
DP Number 3359.
Lee, Y., and Bhargava, V. (2004) Leisure time: Do married and single individuals spend
it differently. Family and Consumer Sciences Research Journal 32, 3: 254–274.
Lee, Y. (2006) The decline of attendance in the Korean professional baseball league.
Journal of Sports Economics 7: 187–200.
Leeds, M., and von Allmen, P. (2002) The Economics of Sports. Pearson, London.
Leibenstein, H. (1966) Allocative efficiency versus x-inefficiency. American Economic
Review 56: 397–409.
Lera-Lopez, F., and Rapun-Garate, M. (2005) Sports participation versus consumer
expenditure on sport: Different determinants and strategies in sports management.
European Sports Management Quarterly 5, 2: 167–186.
Le Roux, N., and Camy, J. (1999) An Essay on the French Sports System. In: Heine-
mann, K. (ed). Sports Clubs in Various European Countries. Hofmann, Schorndorf.
Lertwachara, K., and Cochran, J. J. (2007) An event study of the economic impact of
professional sport franchises on local US economies. Journal of Sports Economics 8,
3: 244–254.
Lipsey, R.G. and Lancaster, K. (1956–1957) The general theory of second best. The
Review of Economic Studies, 24, 1: 11–32.
LIRC. (2003) Sports Volunteering in England 2002. Sport England.
Lucifora, C., and Simmons, R. (2003) Superstar effects in sport. Journal of Sports
Economics 4, 1: 35–55.
400 References
Lukka, P., and Ellis, A. (2002) An exclusive construct? Exploring different cultural
concepts of volunteering. Institute for Volunteering Research.
Lynch, J. and Zax, J.S. (1998) Prizes, selection and performance in Arabian Horse
Racing. Discussion Paper, University of Colorado.
Lynch, J. and Zax, J.S. (2000) The reward to running: Prize structures and performance.
MacDonald, D., and Reynolds, M. (1994) Are baseball players paid their marginal
products? Managerial and Decision Economics 15: 443–457.
Madden, J. R. (2002) The economic consequences of the Sydney Olympics: The CREA/
Arthur Anderson Study. Current Issues in Tourism 5, 1: 7–20.
Maennig, W. (2002) On the economics of doping and corruption in international
sports. Journal of Sports Economics 3, 1: 61–89.
Maennig, W. (2007) One Year Later: A re-appraisal of the economics of the 2006 soccer
World Cup. Hamburg Working paper Series in Economic Policy, No 10.
Maennig, W., and Du Plessis, S. (2007) World-cup 2010: South African economic
perspectives and policy challenges informed by the experience of Germany 2006.
Contemporary Economic Policy 25, 4: 578–590.
Maguire, J. (1999) Global Sport: Identities, Societies, Civilisations. Polity Press, Cam-
bridge.
Maloney, M., and McCormick, R. (2000) The response of workers to wages in tourna-
ments: evidence from foot races. Journal of Sport Economics 1, 2: 99–123.
Marburger, D. (1994) Bargaining power and the structure of salaries in major league
baseball. Managerial and Decision Economics 15, 5: 433–441.
Marshall, A. (1952) Principles of Economics, 8th edition. Macmillan, London.
Mascarenhas, D., Collins, D., and Mortimer, P. (2005) The accuracy, agreement and
coherence of decision-making in rugby union officials. Journal of Sport Behaviour
28, 2: 253–271.
Mason, D. S., and Duquette, G. H. (2005) Globalisation and the evolving player-agent
relationship in professional sport. International Journal of Sport Management and
Marketing 1: 93–109.
Mason, D. S., and Slack, T. (2001) Industry Factors and the Changing Dynamics of the
Player-Agent Relationship in Professional Ice Hockey. Sport management Review
4: 165–191.
Mason, D. S., and Slack, T. (2003) Understanding Principal-Agent Relationships:
Evidence from Professional Hockey. Journal of Sport Management 17: 37–61.
Matheson, V., and Baade, R. (2008) Striking out: Estimating the economic impact of
baseball's world series. International Journal of Sport Management and Marketing
3, 4: 319–334.
Maxcy, J., and Mondello, M. (2006) The Impact of Free Agency on Competitive Balance
in North American Team Professional Team Sports Leagues. Journal of Sport Man-
agement 20: 345–365.
Medoff, M. (1976) On monopolistic exploitation in professional baseball. Quarterly
Review of Economics and Business 16, 2: 113–121.
Meehan, J., Nelson, R., and Richardson, T. (2007) Competitive balance and game
attendance in major league baseball. Journal of Sports Economics 8: 563–580.
Michie, J., and Oughton, C. (2004) Competitive Balance in Football: Trends and Effects.
Football Governance Research Centre, Birkbeck College, University of London.
Mill, J. S. (1900) Principles of Political Economy with some of their Applications to
Social Philosophy. G Routledge and Sons, London.
References 401
Mintel. (2005a) Sports Goods Retailing – UK. Mintel International Group Limited.
Mintel (2005b), Fitness Classes, http://academic.mintel.com.
Mintel (2005c) Sports Goods Retailing, http://academic.mintel.com.
Mintel (2006) Leisure Centres and Swimming Pools. http://academic.mintel.com.
Mintel (2007), Health and Fitness Clubs, http://academic.mintel.com.
Mitchel, H., and Stewart, F. (2007) A competitive index for international sport. Applied
Moorhouse, G. (1995) The Official History of Rugby League. Hodder and Stoughton,
London.
Morley, B., and Thomas, D. (2005) An investigation of home advantage and other
factors affecting outcomes in English one-day cricket matches. Journal of Sports
Sciences 23, 3: 261–268.
Morton, R. H. (2006) Home advantage in southern hemisphere rugby union: National
and international. Journal of Sports Sciences 24, 5: 495–499.
Musgrave, R. A., and Musgrave, P. B. (1973) Public Finance in Theory and Practice.
McGraw-Hill Book Company, New York.
Musgrave, R. (1987) Merit goods. In: Eatwell, J., Milgate, M. and Newman, P. (eds). The
New Palgrave: A Dictionary of Economics. Macmillan and Stockton, London and
New York.
Neale, W. (1964) The peculiar economics of professional sport. Quarterly Journal of
Economics 78, 1: 1–14.
Nevill, A., Balmer, N., and Williams, M. (1999) Crowd influence on decisions in
association football. Lancet (letter) 353: 1416.
Nevill, A. M., and Holder, R. L. (1999) Home advantage in sport: An overview of studies
on the advantage of playing at home. Sports Medicine 28: 221–236.
Nichols, G. (2004) Pressures on Volunteers in the UK. In: Stebbins, R. and Graham, M.
(eds). Volunteering as Leisure/Leisure as volunteering: An International Assess-
ment. CABI Publishing, Wallingford.
Noll, R. (1974) Government and the Sports Business. Brookings Institution, Washing-
ton, DC.
NSW Treasury (1997) Economic Impact of the Sydney Olympic Games, http://www.
treasury.nsw.gov.au/pubs/trp97_10/index.htm.
Oga, J. (1998) Business fluctuation and the sport industry in Japan: An analysis of the
sport industry from 1986 to 1993. Journal of Sport Management 12, 1: 63–75.
Owen, P., and Weatherston, C. (2004) Uncertainty of outcome and Super 12 rugby
union attendances. Journal of Sports Economics 5, 4: 347–370.
Page, L., and Page, K. (2007) The second leg home advantage: Evidence from European
football cup competitions. Journal of Sports Sciences 25, 14: 1547–1556.
Papanikos, G. (1999) Tourism Impact of the 2004 Olympic Games. Tourism Research
Institute, Athens.
Pareto, V. (1906) Manual of Political Economy. Translated by A.S. Schwier (1971). New
York Kelley.
Peel, D., and Thomas, D. (1988) Outcome uncertainty and the demand for football: An
analysis of match attendances in the English football league. Scottish Journal of
Political Economy 35, 3: 242–249.
Peel, D., and Thomas, D. (1997) Handicaps, outcome uncertainty and attendance
demand. Applied Economics Letters 4, 9: 567–570.
Phlips, L. (1978) The demand for leisure and money. Econometrica 46, 5: 1025–1043.
402 References
Pollard, R. (2006a) Worldwide regional variations in home advantage in association

football. Journal of Sports Sciences 24, 3: 231–240.
Pollard, R. (2006b) Home advantage in soccer. Journal of Sport Behaviour 29, 2:
169–189.
Pollard, R., and Pollard, G. (2005) Long-term trends in home advantage in professional
team sports in North America and England (1876–2003). Journal of Sports Sciences
23, 4: 337–350.
Parro, N., Bizzaglia, G., and Conti, D. (1999) The Sports System and Sports Organisa-
tions in Italy. In: Heinemann, K. (ed). Sports Clubs in Various European Countries.
Hofmann, Schorndorf.
Porter, P.K. (2007) The Paradox of Inelastic Sports Pricing Managerial and Decision
Preuss, H. (2004) The Economics of Staging the Olympics. A Comparison of the Games
1972–2008. Edward Elgar, Cheltenham.
Price, D., and Sen, K. (2003) The demand for game day attendance in college football:
An analysis of the 1997 division 1-A season. Managerial and Decision Economics
24: 35–46.
Price Waterhouse Coopers. (2005) Olympic Games Impact Study, Final Report. DCMS,
London.
Primault, D., and Rouger, A. (1999) How relevant is North American experience for
professional team sports in Europe? In: Jeanrenaud, C. and Kesenne, S. (eds).
Competition policy in professional sports: Europe after the Bosman case. Standaard
Editions, Antwerp.
Púig, N., Garcia, O., and Lopez, C. (1999) Sports Clubs in Spain. In: Heinemann, K.
(ed). Sports Clubs in Various European Countries. Hofmann, Schorndorf.
Putnam, R. (2000) Bowling Alone. Simon and Schuster, New York.
Quirk, J., and El-Hodiri, M. (1974) The Economic Theory of a Professional Sports
League. In: Noll, R. (ed). Government and the Sports Business. Brookings Institute,
Washington, DC.
Quirk, J., and Fort, R. (1992) Pay dirt: the business of professional team sports. Prince-
ton University Press, Princeton.
Rahmann, B., Weber, W., Groening, Y., Kurscheidt, M., Napp, H.-G., and Pauli, M.
(1998) Sozio-€ okonomische analyse der fussball-welrmeisterschaft 2006 in
Deutschland: Gesellschaftliche wirkungen, kosten-nutzenanalyse und finanzier-
ungsmodelle einer sportgrossveranstaltung. Sport un Buch Strauss, Cologne.
Rascher, D., and Solmes, J. (2007) Do fans want close contests? A test of the uncertainty
of outcome hypothesis in the NBA. International Journal of Sport Finance 2, 3:
130–141.
Rathke, A. and Woitek, U. (2007) Economics and Olympics: An efficiency analysis.
Working Paper, Institute for Empirical Research in Economics, University of Zur-
ich.
Research Centre for Sport in Canadian Society and Centre for Sport Policy Studies
(2005) Volunteerism: Researching the Capacity of Canadian Sport. Sport Canada.
Robbins, L. (1940) An Essay on the Nature and Significance of Economic Science.
Macmillan, London.
Roberts G., (2006) Accounting for achievement in Athens: a count data analysis of
national Olympic performance, Economic Working Paper EWP0602, University of
Victoria, department of Economics.
References 403
Robinson, J., and Godbey, G. (1997) Time for life: the surprising ways Americans use
their time. Pennsylvania State University Press, University Park.
Robertson, R. (1995) Glocalisation: Time-space and homogeneity-heterogeneity. In:
Featherstone, M., Lash, S. and Robertson, R. (eds). Global Modernities. Sage,
London.
Roderick, M. (2001) The Role of Agents in Profesional Football in Singer and Fried-
lander Football Review: 2001–2002 Season.
Rodgers, B. (1977) Rationalising Sports Policies; Sport in its Social Context: Interna-
tional Comparisons. Council of Europe, Strasbourg.
Rodgers, B. (1978) Rationalising Sports Policies; Sport in its Social Context: Technical
Supplement. Council of Europe, Strasbourg.
Rosen, S. (1981) The economics of superstars. American Economic Review 71, 5:
845–858.
Rottenberg, S. (1956) The baseball players labour market. Journal of Political Economy
64, 3: 243–258.
Russell, D. (1997) Football and the English: A social history of Association football in
England, 1863–1995. Carnegie, Preston.
Sacks, J. (2002) The money trail, measuring your impact on the local economy using
LM3. New Economics Foundation/Countryside Commission.
Salamon, L., and Anheier, H. (1997) Defining the Nonprofit Sector: A Cross-national
Analysis. Manchester University Press, Manchester.
Samuelson, P. A. (1947) Foundations of Economic Analysis. Harvard University Press,
Cambridge, MA.
Sanderson, A. R. (2002) The many dimensions of competitive balance. Journal of Sports
Economics 3, 2: 204–228.
Sanderson, A., and Siegfried, J. (1997) The implications of athlete freedom to contract:
Lessons from North America. Institute of Economic Affairs 17, 3: 7–13.
Sandler, T., and Tschirhart, J. (1997) Club theory: Thirty years later. Public Choice 93:
335–355.
Sandy, R., Sloane, P., and Rosentraub, M. (2004) The Economics of Sport: An Interna-
tional Perspective. Palgrave Macmillan, Basingstoke.
Scheerder, J., Vanreusel, B., and Taks, M. (2005a) Stratification Patterns of Active Sport
Involvement Among Adults. International Review for the Sociology of Sport 40:
139–162.
Scheerder, J., Vanreusel, B., Taks, M., and Renson, R. (2005b) Social Changes in Youth
Sports participation Styles 1969–1999: The case of Flanders: Belgium. Sport, Edu-
cation and Society 10: 321–341.
Schmidt, M., and Berri, D. (2001) Competitive balance and attendance: The case of
major league baseball. Journal of Sports Economics 2, 2: 145–167.
Schwartz, B., and Barsky, S. (1977) The home advantage. Social Forces 55: 551–559.
Schwartz, J. T., Isaacs, J. P., and Carilli, A. M. (2007) To Race or to Place?: An Empirical
Investigation of the Efficiency of the NASCAR Points Competition. Journal of
Sports Economics 8: 633–641.
Scitovsky, T. (1941) A note on welfare propositions in economics. Review of Economic
Studies 9: 77–88.
Scully, G. (1974) Pay and performance in major league baseball. The American Eco-
nomic Review 64, 6: 915–930.
404 References
Scully, G. (1994) Managerial efficiency and survivability in professional team sports.

Managerial and Decision Economics 15: 403–411.
Stamm H., and Lamprecht M. (2001) Sydney 2000 – The best Games ever? World sport
and relationships of structural dependency, Paper presented at the 1st World Con-
gress of the Sociology of Sport, July 2001, Seoul, South-Korea.
Semens, A.P. (2008) Player Representatives in the Football Industry, PhD Thesis Uni-
versity of Central Lancashire, UK.
Shropshire, K. L., and Davis, T. (2003) The Business of Sports Agents. University of
Pennsylvania Press, Philadelphia.
Siegfried, J., and Zimbalist, A. (2000) The economics of sports facilities and their
communities. Journal of Economic Perspectives 14, 3: 95–114.
Simmons, R. (1996) The demand for English league football: A club level analysis.
Applied Economics 28, 2: 139–155.
Simmons, R. (1997) Implications of the Bosman Ruling for Football Transfer Market-
s’in Sloane, P., (ed). Institute of Economic Affairs 17:13–18.
SIRC (2005) Sport Market Forecasts 2005–2009, Sheffield.
Skille, E. A. (2005) Individuality or cultural reproduction. International Review for the
Sociology of Sport 40, 3: 307–320.
Skirstad, B. (1999) Norwegian Sport at the Crossroad. In: Heinemann, K. (ed). Sports
Clubs in Various European Countries. Hofmann, Schorndorf.
Slack, T. (1999) An Outsider looking in or an Insider looking out? In: Heinemann, K. (ed).
Sports Clubs in Various European Countries. Hofmann, Schorndorf.
Sloane, P. (1971) The economics of professional football: The football club as a utility
maximiser. Scottish Journal of Political Economy 17, 2: 121–146.
Sloane, P.J. and Campbell, A. (1997) The implication of the Bosman Case for Profes-
sional Football, University of Aberdeen Discussion Paper 02.
Solberg, E., and Wong, D. (1992) Family time use: Leisure, home production, market
work and work related travel. The Journal of Human Resources 27, 3: 485–510.
Sport England (1999a) Best Value Through Sport: The Value of Sport. London.
Sport England (1999b) Best Value Through Sport: Case Studies. London.
Sport England (2004a) The Framework for Sport in England.
Sport England (2004b) Driving up participation: The Challenge for Sport.
Sport England SouthWest (2006) Active Play and Informal Sport, Somerset.
Stamm, H., and Lampecht, M. (1999) Sports Organisations in Switzerland. In: Hei-
nemann, K. (ed). Sports Clubs in Various European Countries. Hofmann, Schorn-
dorf.
Staudohar, P. (1988) The football strike of 1987: The question of free agency. Monthly
Labour Review 111, 8: 26–31.
Stemple, C. (2005) Adult Participation in Sport as Cultural Capital. International
Review for the Sociology of Sport 40: 411–432.
Stratton, M., Conn, L., Liaw, C. and Connolly, L. (2005) Sport and Related Physical
Activity: The Social Correlates of Participation and Non-Participation by Adults.
Sport management Association of Australia and New Zealand, Conference, Can-
berra.
Sturgis, P., and Jackson, J. (2003) Examining participation in sporting and cultural
activities: Analysis of the UK 2000 Time Use Survey. Department for Culture,
Media and Sport, UK.
References 405
Sutter, M., and Kocher, M. (2004) Favoritism of agents – The case of referees’ home bias.
Journal of Economic Psychology 25, 4: 461–469.
Szymanski, S. (2002) The economic impact of the world cup. World Economics 3, 1:
169–177.
Szymanski S. (2003a) The economic design of sporting contests. Journal of Economic
Literature, 41, December: 1137–1187.
Szymanski, S. (2003b) The assessment: The economics of sport. Oxford Review of
Economic Policy 19, 4: 467–477.
Szymanski, S. (2004) Professional team sports are only a game. Journal of Sports
Economics 5, 2: 111–126.
Szymanski, S. (2006a) A Theory of the Evolution of Modern Sport. International
Association of Sports Economists, Working Paper, No 06-30.
Szymanski, S. (2006b) Why have Premium Sports Rights Migrated to Pay-TV in Europe
but not in the US? In: Jeanrenaud, C. and Kesenne, S. (eds). The Economics of Sport
and the Media. Edward Elgar, Cheltenham, UK.
Szymanski, S., and Kesenne, S. (2004) Competitive balance and gate revenue sharing in
team sports. Journal of Industrial Economics 53: 165–177.
Szymanski, S., and Kuypers, T. (1999) Winners and losers, the business strategy of
football. Penguin Books, London.
Szymanski, S., and Zimbalist, A. (2005) How Americans Play Baseball and the Rest of
the World Plays Soccer. The Brookings Institution, Washington, DC.
Taks, M., Renson, R., and Vanreusel, B. (1999) Consumer expenses in sport: A mar-
keting tool for sports. The European Journal for Sport Management 6, 1: 4–18.
Taks, M., and Kesenne, S. (2000) The economic significance of sport in Flanders.
Journal of Sports Management 14, 4: 342–365.
Taks, M., and Scheerder, J. (2006) Youth Sports Participation Styles and market Seg-
mentation Profiles: Evidence and Applications. European Sport Management
Quarterly 6: 85–121.
Tcha, M., and Pershin, V. (2003) Reconsidering performance at the Summer Olympics
and revealed comparative advantage. Journal of Sports Economics 4, 3: 216–239.
The Conference Board of Canada (2005) Strengthening Canada: The Socio-economic
benefits of sport participation in Canada. Ottawa.
Thrane, C. (2000) Men, women and leisure-time: Scandinavian evidence of gender
inequality. Leisure Sciences 22: 109–122.
Tidsell, C. (2006) Valuation of Tourisms’ Natural Resources. In: Dwyer, L. and Forsyth, P.
(eds). International Handbook on the Economics of Tourism. Edward Elgar, Chel-
tenham.
Tollison, R. (2001) Review of Rosentraub, M. (1999), Major League Losers: The Real
Cost of Sports and Who's Paying For It. New York Basic Books. Journal of Sports
Economics 2, 2: 206–207.
Toohey, K., and Veal, A. (2000) The Olympic Games: A social science perspective. CABI
publishing, New York.
Torkildsen, G. (1994) Torkildsen's Guide to Leisure Management. Longman, London.
Tullock, G. (1980) Efficient Rent Seeking. In: Buchannan, J., Tollison, R. and Tullock,
G. (eds). Toward a Theory of Rent Seeking Society. A and M University Press, Texas.
United Nations (2001) International Year of the Volunteer 2001: International Sym-
posium on Volunteering Final Report. International Symposium on Volunteering,
18 to 21 November 2001, Geneva, Switzerland. www.iyv2001.org.
406 References
U.S. Department of Health and Human Services, Centers for Disease Control and
Prevention, National Center for Chronic Disease Prevention and Health Promo-
tion, (1996) A Report of the Surgeon General, Atlanta, GA.
Utt, J., and Fort, R. (2002) Pitfalls to measuring Competitive Balance with Gini Coeffi-
cients. Journal of Sports Economics 3, 4: 367–373.
Vamplew, W. (1988) Pay Up and Play the Game: Professional Sport in Britain 1875–
1914. Cambridge University Press, Cambridge.
van Bottenburg, M., Rijnen, B., and van Sterkenburg, J. (2005) Sports participation in
the European Union: Trends and Differences. W.J.H. Mulier Institute and Arko
Sports Media, The Netherlands.
Verhaegen, D. (2006) De verklarende factoren van de toeschouwersaantallen in de
Belgische hoogste voetbalklasse, eindverhandeling handelswetenschappen. Lessius
Hogeschool, Antwerpen.
Veraros, N., Kasimati, E., and Dawson, P. (2004) The 2004 Olympic Games announce-
ment and its effect of the Athens and Milan stock exchanges. Applied Economic
Letters 11: 749–753.
Von Allmen, P. (2001) Is the Reward System in NASCAR Efficient? Journal of Sports
Vrooman, J. (1995) A general theory of professional sports leagues. Southern Economic
Journal 61: 971–990.
Vrooman, J. (2007) Theory of the beautiful game: The unification of European football.
Scottish Journal of Political Economy 54, 3: 314–354.
Waylen, P., and Snook, A. (1990) Patterns of regional success in the football league,
1921 to 1987. Area 22: 353–367.
Weed, M., Robinson, L., Downward, P., Green, M., Henry, I., Houlihan, B., and Argent,
E. (2005) Academic Review of the Role of Voluntary Sports Clubs. Institute of Sport
and Leisure Policy, Loughborough University.
Weisbrod, B. A. (1978) The Voluntary Nonprofit Sector: An Economic Analysis. Lex-
ington Books, Lexington.
Weisbrod, B. A. (1988) The Non-Profit Economy. Heath, Lexington.
Weisbrod, B. A. (1998) To Profit or not to Profit: The Commercial Transformation of the
non profit sector. Cambridge University Press, Cambridge.
Williams, G. (1994) The Code War. Yore Publications, Middlesex.
Williamson, O. (1969) Corporate Control and Economic Behaviour.
Williamson, O. (1971) Managerial Discretion, Organisation Form and the multi-
Division Hypothesis. In: Marris, R. and Wood, A. (eds) (1971) The Corporate
Economy, Growth, Competition and innovative Potential. London: Macmillan.
Williamson, O. (1975) Markets and Hierarchies: Analysis and Antitrust Implications.
Free Press, New York.
Williamson, O. (1981) The modern corporation: Origins, evolution, attributes. Journal
of Economic Literature 19: 1537–1568.
Williamson, O. (1985) The Economic Institutions of Capitalism: Firms, Markets, Re-
lational Contracting. Collier Macmillan, London.
Wilson, B. (2004) A Logistic Regression Model of the Decision of Volunteers to Enter a
Sports Coach Education Programme. In: Stebbins, RA. and Graham, M. (eds).
Volunteering as Leisure/Leisure as volunteering: An International Assessment.
CABI Publishing, Wallingford.
References 407
Wilson, P., and Sim, B. (1995) The demand for semi-pro league football in Malaysia
1989–1991: A panel data approach. Applied Economics 27: 131–138.
Winfree, J. and Fort, R. (2008) Fan substitution and the 2004-05 NHL lockout. Journal
of Sports Economics, 10: 1-10. NOTE: The on-line version is not too clear about
pagination, c f Brown and Link above!.
Winfree, J., McCluskey, J., Mittelhammer, R., and Fort, R. (2004) Location and atten-
dance in major league baseball. Applied Economics 36: 2117–2124.
Wolff, N., Weisbrod, B. and Bird, E. (1993) The supply of volunteer labor: The case of
hospitals. Nonprofit Management and Leadership 14, 1: 23–45.
Yu, M. (2004) The Economic and Social Impacts of Hosting Selected International
Games. Research and Library Services Division Legislative Council Secretariat,
Hong Kong.
Zuzanek, J. (1978) Social differences in leisure behaviour: Measurement and interpre-
tation. Leisure Sciences 1, 3: 271–293.
List of abbreviations
AC Average costs
AFL American Football League
AL American League
AR Average revenue
C Total cost
CL Champions League
ECB England and Wales Cricket Board
EFL English Football League
FA Football Association
FIFA Federation Internationale de Football Association
GAA Gaelic Athletic Association
GLS Generalized least squares
ICB International Cricket Board
IOC International Olympic Committee
IRB International Rugby Board
MC Marginal costs
ML Maximum likelihood
MLB Major League Baseball
MP Marginal product
MPB Marginal private benefit
MPC Marginal private cost
mpc Marginal propensity to consume
MPL Marginal product of labour
MR Marginal revenue
MRP Marginal revenue product
MRS Marginal rate of substitution
MRTS Marginal rate of technical substitution
MSB Marginal Social Benefit
MSC Marginal Social Cost
NBA National Basketball Association
NCAA National Collegiate Athletic Association
NFL National Football League
NHL National Hockey League
NL National League
OLS Ordinary Least Squares
P Price
PL Premier League
Q Output or Quantity
R Total Revenue
409
410 List of abbreviations
RFL Rugby Football League

RFU Rugby Football Union
SL Super League
UEFA Union of European Football Associations
Index
A collective bargaining agreement, CCT See Complusory competitive

AA See American Association (AA) 303 tendering (CCT)
‘‘Active People,’’ survey, 98 farm system in, 237 Ceteris paribus, 9
Ad valorem tax, 21 See also Tax formalization, 179, 180–181 Champions league (CL), 45, 196,
Agents, player, 308–309 leagues for, 195 200
Ageing, and participation, 104 competitive balance in, 220– Circular flow of income, 66–68, 350,
AL See American League (AL) 223 353, 354
American Association (AA), 187 MLB See Major League Baseball Clayton Act, 234
American League (AL), 222 (MLB) Clubs, 178
Amsterdam Treaty, 41 National League (NL), 181–182 European structure, 148–151
ANR See Average net revenue (ANR) negro leagues, 37 goods, 151–155
Attendance demand, for profes- Olympics 2012 and, 54 provision, 155–157
sional teams participation rates, 101 participation, 142
consumer preferences, 264 salaries, superstar effects, 315– supply, 148–151
estimations See Estimations, at- 316 Cluster analysis, 114
tendance demand Basketball, 51 Coase theorem, 24, 87, 133, 295
evidence See Evidences, atten- NBA See National Basketball Cobb-Douglas function, 90–92,
dance demand Association (NBA) 188, 384
habit persistence and, 290 participation, 55–56, 57, 101 Cointegration analysis, 266
income measurement, 263–264 BBC, 292 Commonwealth Games (CWG),
MLB and, 288 Bertrand price competition, 139 See 54–55
NBA and, 288 also Price COMPASS, 56, 58
NFL and, 288 Betting odds, 211–212 Competitive balance, 219–226
price measurement, 263–264 uncertainty of outcome (UO) in European and other leagues,
quantity measurement, 262–263 and, 212 256–257
UO effect on, 289 Bicycle motocross (BMX), 140, 141 evolution of, 224–226
weather impact on, 289 BMX See Bicycle motocross (BMX) Complusory competitive tendering
Australian Rules football, 213, Bosman Ruling, 226, 306–307 (CCT), 129
219 contract duration, 322 Congestion, in club goods, 153
AVC See Average variable costs player agents, 309 Consolidated Metropolitan Statisti-
(AVC) Boxing, 195 cal Area (CSMA), 291
Average net revenue (ANR), 314 Broadcasting demand, 263, 283, Consumer preferences, 264,
Average revenue, and marginal rev- 292–298 288–291
enue, 31–32 economics of, 295–298 integrating framework for, 82–85
Average variable costs (AVC), 131 history of, 292–295 interdependent preferences and
MCMMG model, 292 habit persistence, 289–290
B Premier League and, 292 quality characteristics, 288–289
Bargaining spatial context, 290–291
collective, 296–297, 303 C Consumer spending, 105
and economic efficiency, 24, 87 Capacity, explicit treatment of, Consumption
Nash bargaining solution, 313– 283 economic model of, 65–81
314, 333–334 Capital, forms of, 85 derived demand for sports,
Baseball, 52 Cartel, 194–196 76–82
411
412 INDEX
dual decision hypothesis, reserve option clause, 237, Equilibrium

66–69 244 profit-maximizing competitive,
income-leisure trade-off, salary caps, 236–237, 245 239–244
69–75 legal frameworks, 234 win-maximizing league and,
Contest theory See Tournament profit-maximizing competitive 247–248
theory quilibrium, 244–247 win percent, 244
Contingent valuation, see willing- revenue redistribution, 237–239, Equipment, 143–144
ness to pay 245–247 Equity, 18–19, 125–127
Contract for players, 310–314 win maximizing and, 249 Estimations, attendance demand,
bargaining, 313–314 zero–sum restrictions, 239 264–267
duration, 313–314 CSMA See Consolidated Metropol- cointegration analysis, 266
monopoly power, 312–313 itan Statistical Area (CSMA) heteroscedasticity and serial/
monopsony model, 311–312 Cumulative variance, 222 autocorrelation, 267
non-profit-maximizing behav- Currie Cup, The, 223 instrumental variable, 266
iour, 314 multicollinearity, 267
Cooperative games, 313 multivariate techniques of, 266
Costs, 351
D simultaneity, 264–267
Data, mass sports participation
average, 6, 9, 15, 16, 130, 171 2SLS, 266
analysis, 106–114
cost benefit analysis (CBA), 360, Europe
sources, 97–101
367, 368 labour market, 305–308
Demand
fan cost index, 278–280, 285 leagues, 196–200
actual, 263
marginal cost (MC), 5, 6, 8, 13, European Cup/European Cham-
attendance See Attendance de-
130, 153, 189–190, 248, 285 pions' League, 223
mand, for professional teams
marginal cost of labour (MCL), European football club revenues, 61
broadcasting, 292–298
311, 333 European mass sports organization,
club goods, 153
marginal social cost (MSC), 20, stylized model, 151
curve, theoretical, 262
125 EU rulings in sport, 41
derived, 76–82
Cournot model, 137 Events, 53–55
inelastic pricing, 285
Cournot–Nash models, 254 contest theory and, 342
latent spectator, 263
‘‘Covering law’’, 4 effort-performance-reward,
price elasticity of, 284
Cricket 342–345
Diminishing marginal utility, 69
IPL, 294–295 investment in See Public sector
Direct Service Organizations, 136
origin and characteristics, 50 investments
Drafting system, 235–236
‘‘Twenty-20’’ cricket, 294–295 participation in, 340–341
Dual decision hypothesis, 66–69
Cross-sectional regression equation, types of, 339–340
Durable goods, 92–93
264 Evidences, attendance demand,
Cross-subsidization, 233–257 267–292
assumptions, relaxation of, E consumer preferences, 288–291
249–254 Econometrics, of labour market, estimation, major issues,
market size diseconomies, 314–324 291–292
251–253 Economic efficiency, 124–125 income measurement, 283–288
spectator preference, for Economic goods, types, 125 price measurement, 283–288
quality, 249–251 Economic analysis, empirical quantity measurement, 268, 283
variable supply, of talent, framework, 25–30 Ex anteevidence, investment,
253–254 Economics 366–367
economic framework for under- core economic model, 4–14 Exchange rate, 68
standing, 239–249 definition, 2–3 Exclusivity reduction, club, 153
economic rationale for, 234 methodology, 3–4 Exercise, Recreation and Sport
invariance proposition, 247, Egalitarian society, 126 Survey (ERASS), 56
254–256 English Football League (EFL), 44, Expenditure
labour market policies, 235–237 47, 180, 181, 223, 238, 292 on mass sports participation,
drafting system, 235–236, English soccer, long-run study of, 104–106
245 219 rationale for policy, 25
INDEX 413
Externalities, 19–21, 178, 193–194 Herfindahl indexes (HI), 222 duration of, 309–310
See also Uncertainty of outcome Heterogeneous demands and econometric findings, 314–324
(UO) government failure, 155 historic evolution, 303–310
Coase theorem, 24, 87 Heteroscedasticity, 267 Europe, 305–310
and investment, 348–349, 356, HI See Herfindahl indexes (HI) player agents, 308–309
357, 363 ‘‘Hicksian’’ demand, 75 United Kingdom, 305–310
negative, 19, 20 Home advantage, 207, 226–229 United States, 303–305
and participation, 125, 142 findings on, 228 managerial efficiency, 324–330
positive, 19, 20 perfect competition, 302–303
subsidies and, 21, 25 I player agents, 308–309
taxes and, 21, 25 Ice hockey, 50, 305 policies, by leagues
participation rates, 101 drafting system, 235–236
F Imperfect information, 22–24
Income See also Income-leisure
reserve option arrangements,
FA See Football Association (FA) 237
Facility fee, 293 trade-off salary caps, 236–237, 245
Factor analysis, defined, 114 and consumer demand, 79–80 Laissez-faire approach, 121
Fan cost index, 278–280, 285 See full income, 91 Leagues
also Costs measurement, 263–264, characteristics, 197–200
Federation Internationale de Foot- 283–288 club objectives, 196–197
ball Association (FIFA), 45, 192, Income-leisure trade-off, 69–75 cross-subsidization in, 233–257
307, 308, 309, 376 Indian Premier League (IPL), 294 cross-subsidization policy by See
FIFA See Federation Internationale Inelastic pricing, 285 Cross-subsidization
de Football Association (FIFA) Informal sports, 55–56, 140 Europe, 196–200
Final offer arbitration (FOA), 303 facilities, 140 recent evolutionary changes to,
Fitness clubs, 136 participation, 140–143 196
Flood vs. Kuhn, 303 Information asymmetry, 22–24 relative league standings, 207–
FOA See Final offer arbitration Infrastructure, investment in See 211
(FOA) also Public sector investments uncertainty of outcome, 205–
Football Instrumental variable estimation, 229
rugby See Rugby 266 United States, 196–200
soccer See Soccer Integrating framework for consumer US, 61
Football Association (FA), 43, 179, choice, 82–85 Leisure time, 107–110
305 Interdependent preferences, determinants, 109
Formalization process, profession- 289–290 Linear expenditure models, sports
alism, 179–182 Interest rate, 67 participation, 108
International Olympic Committee Long-run UO, 219–226
(IOC), 53–54 competitive balance, 219–226
G bribery at, 346 Lorenz curve, 220
Game theory, elements of, 138,
International Rugby Board (IRB), 49
145–146
Intertemporal supply chain, 173
Gate revenue sharing, 238 See also M
Investment
cross-subsidization Maastricht Treaty, 41
demand, 94–95
General Household Survey (GHS), Major League Baseball (MLB), 235,
in events See Public sector
58–59 288
investments
Gibrat's law, 224 Marginal cost of labour (MCL), 311,
IPL See Indian Premier League (IPL)
Gini coefficient, 220 333 See also Costs
Governing bodies, 43–49 K Marginal costs (MC), 5–6, 13, 130,
Government failure, 24–25 Keynesian expenditure multiplier 131–132, 153, 189–190, 248 See
derivation, 32–33 also Costs
H and average costs (AC), 8, 9
Habit persistence, 290 L inelastic pricing, 285
measurement, 299–300 Labour market Marginal rate of substitution (MRS),
Health clubs, 136 contract for players, 310–314 72, 154
414 INDEX
Marginal rate of technical substitu- MLB See Major League Baseball National Household Survey on Par-
tion (MRTS), 169 (MLB) ticipation in Sport (NHSPS), 58
Marginal revenue (MR), 6, 8, 13, Money, time value of, 93–94 National League (NL), 181, 222, 303
131, 132 Monopolies and Restrictive Prac- National Olympic Committees
average revenue in monopoly, tices Act, 234 (NOC), 53
31–32 Monopoly, 14–17 NBA See National Basketball Asso-
Marginal revenue product of labour average and marginal revenue in, ciation (NBA)
(MRPL), 302 31–32 Negative externalities, 19, 20
Marginal social cost (MSC), 20, 125 power, 312–313 Net present value (NPV), 347,
Market players contract, 312–313 382–383
failures, 14–24 vs.perfectly competitive leagues, New household economics, 82–84
Marginal Rate of Transformation 16 NFL See National Football League
equity, 18–19 Monopsony model, 311–312 (NFL)
externalities, 19–21 MRPL See Marginal revenue prod- NHL See National Hockey League
imperfect information, uct of labour (MRPL) (NHL)
22–24 MRS See Marginal rate of substitu- NL See National League (NL)
monopoly, 14–17 tion (MRS) Non-cooperative games, 313
public goods, 21–22 MRTS See Marginal rate of technical Non-profit maximizing behaviour,
size, 251–253 substitution (MRTS) 314
structure, supply, 136–139 MRT See Marginal rate of Transfor- Normative economics, 10–14
Marketing strategies, 137 mation NPV See Net present value (NPV)
‘‘Marshallian’’ demand, 75 Multicollinearity, 267
Mass participation, 55–59 See also Multiplier effect, 350–356
Participation components, 351 O
data analysis, 106–114 input-output analysis of, Oligopoly models, 137
data sources, 97–101 383–384 OLS analysis See Ordinary least
descriptive evidence, 101–104 Multivariate estimation techniques, squares (OLS) analysis
expenditure, 104–106 266 Olympic Games, 53–54
Match significance See Medium- evidence on performance in,
term UO 385–395
Match-level forecasting model, 217 N operating costs, 59–60
Opportunity costs, 84
Match UO, 207–213 NABBP See National Association of
betting odds, 211–212 Baseball Players (NABBP) Ordinary least squares (OLS) analy-
direct estimates, 212–213 NAPBBP See National Association sis, 28–29
relative league standings, of Professional Baseball Players Organizational efficiency, sports
207–211 (NAPBBP) club system, 159
MCL See Marginal cost of labour Nash bargaining solution, 313–314, Organizational failures framework,
(MCL) 333–334 158
MCMMG See Media–Corpora- National Association of Baseball
tions–Merchandising–Markets– Players (NABBP), 179 P
Global (MCMMG) model National Association of Professional Pareto optimum, 154
Medal success, determinants of, Baseball Players (NAPBBP), Participation See also Supply of par-
385–395 180–181 ticipant sport
Media–Corporations–Merchandis- National Basketball Association data, official, 99–100
ing–Markets–Global (MCMMG) (NBA) definition, 56–58
model, 199, 292 formation, 51 economic model of consump-
Medium-term UO, 213–218 labor market policies, 235 tion, 65–81
forward-looking model, 217 salary caps, 236, 245 evidence, 97–116
findings on the effects of, National Football League (NFL), integrating framework for con-
215–216 235, 288 sumer choice, 82–85
seasonal UO, 214, 217–218 National Health Interview Survey, location of, 119
within-season UO, 213–214 58 policy implications, 86–88
Merit Award, 293 National Hockey League (NHL), 235 rates, 101, 102
INDEX 415
Payments vs. subsidies, 132 broadcasting demand for, R

Players 292–298 Random error, in OLS analysis, 28,
agents, 308–309 broadcasting history and, 29
contract See Contract for players 292–295 Rawlsian approach, social welfare,
draft, 245 habit persistence in, 290 126
football, 47–49 interdependent preferences in, Recursive model, 315
labour market, 310 289–290 Regression analysis
bargaining, 313–314 league as natural monopoly, empirical approach, 29–30
Europe, econometric studies, 194–196 professional teams, 264
316, 322–324 peculiar economics of, 193–194 Relative league standings, 207–211
monopoly, 312–313 production of, 193 Rent-seeking behaviour, in tourna-
monopsony model, 311–312 recent econometric findings on, ment theory, 186
non-profit-maximizing be- 269–282 Reservation wage, 312
haviour, 314 regression models used in, 264 Reserve option
United States, econometric spatial context, 290–291 arrangements, 237
studies, 315–316 Profit maximization, 382–383 clause, 244
soccer, 23–24 Profit-maximizing league Resource allocation, economic
Policy cross-subsidization in, 244–247 model on, 4–14
expenditure-based rationale for, equilibrium in, 248 Restrictive Trade Practices Act, 234
25 Public goods, 21–22 Retain and transfer system, 237
governing bodies, 42–49 Public policy, 40–42 Revenue redistribution, 237–239,
historic and current, 59–62 Public sector, 119–124 245–247
implications, mass participation, economic rationale, 124–134 effect of, 247
86–88, 172–173 investments See Public sector gate revenue sharing, 238
intervention, 14–24 investments pool sharing, 238, 246
limitations, 24–25 other sports policy, 121–124 television revenue, 238
Positive externalities, 19, 20 UK sports policy, 120–121 RFU See Rugby Football Union
Prais–Winsten estimator, 292 Public sector investments, 345, (RFU)
Premier League (PL), 45 347–350 Rookie Draft, 235
Price decision, 345–347 Rugby, 44, 48–49, 180, 182–183,
and consumer demand, 79–80 economic rationale for, 347–350 196
economic impacts and, 360 externalities, 348–349, 356, 357, broadcasting, 293, 294
elasticity of demand, 284 363 IRB, 49
and income, 263–264, 283–287 forms of, 348 IRFB, 49
measurement, 263–264, 283– impacts participation rate, 101
288 duration and scale, 356–359 RFU, 43, 46, 179
shadow, 84, 90 economic activity as, 359– Rugby Football Union (RFU), 43, 46,
talent, 243, 244, 246, 253, 254 365 179
effect of revenue sharing on, empirical evidence, 366 Rugby League International Federa-
255 ex anteevidence, 366–367 tion, 48
ticket, 5–21, 181, 262, 265–266, ex post, 370–380
284–286 non-market value, 380–381 S
Pricing, 139–140 See also Price variance, 367, 369–370 Salary caps, 236–237, 256
Private sector, and supply, 134–140 multiplier, 350–356 and leagues, 245
health and fitness clubs, 136 components, 351 small-town clubs, 236
market structure, 136–139 willingness to pay, 376–378 Sales
pricing, 139–140 goods, 143
Probability of home win (HP), 211 tickets, 14–15
See also Match UO Q SDWP See Standard deviation of
Productivity, 11–13 Quadratic revenue equation model, win percent (SDWP)
Professional team sports, 43–49, 249–256 Seasonal UO See Medium-term UO
50–53 Quantity measurement, 262–263, Second best optimum theorem,
attendance demand for, 262–292 268, 283 174–175
416 INDEX
Seemingly unrelated regression of participant sport See Supply of U

(SUR), 266–267 participant sport UEFA See Union of European
Serial correlation, attendance esti- structure, 118–119 Football Associations (UEFA)
mation, 267 of talent, 253–254 Uncertainty of outcome (UO),
Sheffield Shield interstate champi- Supply of participant sport 205–229, 265
onship, 223 economic efficiency in, 124–125 direct estimates, 212–213
Sherman Act, 234 economic evaluation, 160–161 effect on attendance, 289
Simultaneity, attendance estima- economic rationale for sports findings on the effects of,
tion, 264–267 clubs, 151–155 208–210
SIRC See Sports Industry Research equity in, 125–128 and home advantage, 207,
Centre (SIRC) European hierarchical sports or- 226–229
2SLS See Two-stage least squares ganization, 149–151 hypothesis, 193
(2SLS) estimation facilities provision and, 143–144 long-run See Long-run UO
‘‘Slutsky approximation’’, 75 hierarchical form, 157–160 medium-term See Medium-term
Soccer, 44, 47, 56, 182, 183, 207, informal participation, 140–143 UO
243, 283, 288–289 private sector provision for, 134– forward looking model, 217
broadcasting, 293, 294 135 relationship between betting
EFL See English Football League market structure, 136–139 odds and, 212
(EFL) pricing, 139–140 short-run See Match UO
FA, 43, 179, 180, 305 provision of club goods, 155–157 time-dependent nature of, 206
FIFA, 45, 192, 307, 308, 309, UK sports policy and provision, Union of European Football Asso-
376 120–124 ciations (UEFA), 45, 227, 306
home advantage, 227 mechanisms to promote par- United Kingdom Econometric find-
leagues, 196 ticipation, 128–134 ings on the labour market
participation rate, 101 volunteering See Volunteering labour market, 305–308
in UK and Europe, 305–308 SUR See Seemingly unrelated re- sports policy, 120–121
Social welfare, 126 gression (SUR) elite sport, 121
Socio-economic determinants, of participation, 120–121
participation, 103 T promoting participation,
Spanish domestic league champi- Talent, variable supply, 253–254 128–134
onship, 223 Tax, 21, 25, 87 United States
Spatial context, 290–291 Taxation, 66, 67–68, 126, 151 college football, 283
Spectator preference, for quality, Team-specific UO See Seasonal UO econometrics findings on the la-
249–251 Television revenue, 238 bour market, 315–316
Spectators–Subsidies–Sponsor– Time labour market, 303–305
Local (SSSL) model, 199 series regression equation, 264 leagues, 61, 196–200
Sports value of money, 93–94 UO See Uncertainty of outcome
definition, 36–38 variance, 222 (UO)
economic organization, 38–40 Tournament Utilitarian approach, social welfare,
Sports Industry Research Centre discrimination properties, 201 126, 127
(SIRC), 101 optimal effort derivation, Utility-maximizing behaviour,
SSSL See Spectators–Subsidies– 202–203 89–92
Sponsor–Local (SSSL) model theory, Tournament theory
Stackelberg model, 137 Tournament theory, 184–192, 341, V
Standard deviation of win percent 345 Volunteering, 161–172
(SDWP), 220 assumptions and implications benefits, 164
Stochastic error term See Random of, 187 constraints, 163, 165
error, in OLS analysis rent-seeking behaviour, 186 decision, 168
Subsidiarity principle, 41 types, 185 defined, 161–163
Super league (SL), 47 Twin track approach, policy, 121, economic analysis, 167–172
Supply 122 motivation, 163, 164
monopoly, of sports fixtures, Two-stage least squares (2SLS) profile, 165–167
16–17 estimation, 266 value, 171
INDEX 417
W Willingness to pay. See Public sector Within-season UO, medium-term

Wages, 61, 75 investments UO, 213–214
monopsony model, 311–312 Win-maximizing league WN See Wins currently needed
reservation, 312 cross-subsidization in, 249 (WN)
Walgrave ruling, 41 equilibrium in, 247–248
Wealth, 382–383 Win percent, equilibrium, 244 Z
defined, 347 Wins currently needed (WN), 217 Zero–sum restrictions, 239

Sports Economics Theory, Evidence and Policy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sports Economics Theory, Evidence and Policy

Uploaded by

Copyright:

Available Formats

Sports Economics

Books in the Sport Management Series

Sport and the Media

Sport Funding and Finance

Managing People in Sport Organizations

Introduction to Sport Marketing

More information on the series can be found online by visiting

Sport Management Series

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD

Copyright © 2009, Elsevier Ltd. All rights reserved

No part of this publication may be reproduced, stored in a retrieval system

British Library Cataloguing in Publication Data

Library of Congress Cataloging-in-Publication Data

For information on all Butterworth–Heinemann publications

Printed and bound in Great Britain

Series editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Chapter 1 The Economics of Sport

Chapter 2 The Nature, Organization and Economic Significance of Sport

Chapter 3 The Economics of Sports Participation

Chapter 4 The Economics of Sports Participation: Evidence

4.3 Mass Sports Participation: Theoretical and Empirical

Chapter 6 The Supply of Participant Sport: Volunteers and Sports Clubs

6.4.2 Motivation and constraints on volunteering . . . . . . . 163

Chapter 7 The Market for Professional Sports: General Themes

Chapter 8 Uncertainty of Outcome, Competitive Balance and Bias

8.4.2 Seasonal (team-specific) uncertainty of outcome . . . 214

Chapter 9 Cross-Subsidization in Professional Sports Leagues

Chapter 10 The Demand for Professional Team Sports: Attendance and

Chapter 11 The Labour Market in Professional Team Sports

Appendix 11.1 General conditions for the demand for factors

Chapter 12 The Economics of Sports Events and Infrastructure

Dr Paul Downward is Director of the Institute of Sport and Leisure Policy,

Moreover, because the pedagogy of economics can be technical, drawing

Figure 6.4 Organizational efficiency

The Economics of Sport

& To understand some main features of economics

BOX 1.1 DEFINITIONS OF ECONOMICS

1.2 WHAT IS ECONOMICS?

Consequently, gate and television revenues need to be earned to pay players'

1.3 ECONOMIC METHODOLOGY

by the definition of Robbins, and the empirical relevance of theory as an aid to

1.4 A CORE ECONOMIC MODEL: PERFECT COMPETITION

While this might seem to be a reasonably plausible scenario, notwithstand-

FIGURE 1.1 A perfect competitive sports league.

The theory of perfect competition assumes that clubs wish to maximize

In economic theory labour costs are usually assumed to be the variable

Wage rate of players

BOX 1.2 WHAT IS MARGINAL COST?

illustrated in Box 1.2. In the theory of competitive markets this is assumed to

curve, is received by the club. An important distinction to note here is that

FIGURE 1.2 Increased demand for sports fixtures.

1.4.2 Normative economics

but also that every optimum allocation of resources is a perfectly competitive

A function is a shorthand description that indicates that a dependent

An important corollary of the above discussion concerns the markets for

Wage rate of players

Multiplying both sides of this relationship by the marginal product of

Price Marginal product of players Wage rate of players ð1:4Þ

Marginal revenue product of players Wage rate of players

The marginal revenue product (MRP) of players is their contribution, in

identified by economists as the most desirable situation, as it maximizes

1.5 MARKET FAILURE AND THE RATIONALE FOR POLICY

Moreover, this suggests that the number of

consumer surplus resulting from monopoly organization of the league. The