All Clear for Takeoff: Evidence

from Airports on the Effects of

Infrastructure Privatization
Finance Working Paper N° 850/2022 Sabrina T Howell
October 2022 New York University and NBER

Yeejin Jang
University of New South Wales

Hyeik Kim
University of Alberta

Michael S. Weisbach
Ohio State University, NBER and ECGI

© Sabrina T Howell, Yeejin Jang, Hyeik Kim and

Michael S. Weisbach 2022. All rights reserved. Short
sections of text, not to exceed two paragraphs, may
be quoted without explicit permission provided that
full credit, including © notice, is given to the source.

This paper can be downloaded without charge from:

http://ssrn.com/abstract_id=4237006

www.ecgi.global/content/working-papers
 ECGI Working Paper Series in Finance

All Clear for Takeoff: Evidence from Airports on

the Effects of Infrastructure Privatization

Working Paper N° 850/2022

October 2022

Sabrina T Howell
Yeejin Jang
Hyeik Kim
Michael S. Weisbach

We are grateful to participants at the LBS Private Capital Symposium, the Private Equity Research Consortium
Spring Symposium, the NYU Stern internal seminar, and the NYU Stern Corporate Governance luncheon, as
well as for helpful comments from Yakov Amihud, Aleksandar Andonov, Geoff Chatas, and Arpit Gupta. We
thank David Gillen for data.

© Sabrina T Howell, Yeejin Jang, Hyeik Kim and Michael S. Weisbach 2022. All rights reserved.
Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission
provided that full credit, including © notice, is given to the source.
 Abstract
Infrastructure assets have undergone substantial privatization in recent decades. How
do different types of owners target and manage these assets? And does the contract
form—control rights (concession) vs. outright ownership (sale)—matter? We explore
these questions in the context of global airports, which like other infrastructure assets
have been privatized by private firms and private equity (PE) funds. Our central find-
ing is that PE acquisitions bring marked improvements in airport performance along a
rich array of dimensions such as passengers per flight, total passengers, number of
routes, number of airlines, cancellations, and awards. Net income increases after PE
acquisitions, which does not reflect lower costs or layoffs. In contrast, in the few cases
where non-PE acquisitions bring some improvement, it appears to reflect targeting
rather than operational changes. Overall, we find little evidence that privatization
alone increases airport performance; instead, infrastructure funds improve perfor-
mance both in privatization and subsequent acquisitions from non-PE private firms.
These effects are largest when there is a competing airport nearby. Finally, we show
that outright ownership rather than control rights alone is associated with the most
improvement after privatization.

Keywords: Infrastructure, Privatization, PE, Airports

JEL Classifications: G32, G38, L32, R42, H54

Sabrina T Howell*
Associate Professor of Finance
New York University, Stern School of Business
44 West Fourth Street
New York, NY 10012, United States
phone: +1 212 998 0913
e-mail: [email protected]

Yeejin Jang
Senior Lecturer
University of New South Wales
High St Kensington
Sydney NSW 2052, Australia
phone: +61 2 9385 4280
e-mail: [email protected]

Hyeik Kim
Assistant Professor
University of Alberta
2-32C Business Building
Edmonton, Alberta T6G 2R6, Canada
e-mail: [email protected]

Michael S. Weisbach
Professor and Ralph W. Kurtz Chair in Finance
Ohio State University
700 Fisher Hall, 2100 Neil Avenue
Columbus, OH 43210-1144, United States
phone: +1 (614) 292-3264
e-mail: [email protected]

*Corresponding Author
 Fisher College of Business
Working Paper Series

Charles A. Dice Center for

Research in Financial Economics

All Clear for Takeoff: Evidence from

Airports on the Effects of
Infrastructure Privatization

Sabrina T. Howell,
New York University and NBER

Yeejin Jang,
University of New South Wales

Hyeik Kim,
University of Alberta

Michael S. Weisbach,
The Ohio State University, NBER, and ECGI

Dice Center WP 2022-10

Fisher College of Business WP 2022-03-010

October 3, 2022

This paper can be downloaded without charge from:

http://www.ssrn.com/abstract=4237006

An index to the working paper in the Fisher College of

Business Working Paper Series is located at:
http://www.ssrn.com/link/Fisher-College-of-Business.html

fisher.osu.edu
 All Clear for Takeoff: Evidence from Airports on the Effects of
Infrastructure Privatization

Sabrina T. Howell, Yeejin Jang, Hyeik Kim, Michael S. Weisbach*

October 3, 2022

Abstract

Infrastructure assets have undergone substantial privatization in recent decades. How do different types
of owners target and manage these assets? And does the contract form—control rights (concession) vs.
outright ownership (sale)—matter? We explore these questions in the context of global airports, which
like other infrastructure assets have been privatized by private firms and private equity (PE) funds. Our
central finding is that PE acquisitions bring marked improvements in airport performance along a rich
array of dimensions such as passengers per flight, total passengers, number of routes, number of
airlines, cancellations, and awards. Net income increases after PE acquisitions, which does not reflect
lower costs or layoffs. In contrast, in the few cases where non-PE acquisitions bring some
improvement, it appears to reflect targeting rather than operational changes. Overall, we find little
evidence that privatization alone increases airport performance; instead, infrastructure funds improve
performance both in privatization and subsequent acquisitions from non-PE private firms. These effects
are largest when there is a competing airport nearby. Finally, we show that outright ownership rather
than control rights alone is associated with the most improvement after privatization.

Keywords: Infrastructure, Privatization, PE, Airports

JEL Classification: G32, G38, L32, R42, H54

* Howell: NYU Stern & NBER, Jang: University of New South Wales, Kim: University of Alberta, Weisbach: Ohio State

University, NBER, & ECGI. We are grateful to participants at the LBS Private Capital Symposium, the Private Equity Research
Consortium Spring Symposium, the NYU Stern internal seminar, and the NYU Stern Corporate Governance luncheon, as well as
for helpful comments from Yakov Amihud, Aleksandar Andonov, Geoff Chatas, and Arpit Gupta. We thank David Gillen for data.

1
1 Introduction

Trade and transportation infrastructure has undergone massive privatization worldwide over the past 50

years. Assets such as seaports, airports, roads, bridges, railroads, water systems, and internet cables have

transitioned from government ownership and operation to the private sector, via either long-term concession

leases or outright ownership transfers. Over time, private equity (PE) investors—-usually through dedicated

infrastructure funds—-have come to play an important role in this process. The privatization of traditionally

publicly-owned infrastructure raises questions about what types of goods private markets can efficiently

provide, a topic that economists have long debated, perhaps most famously in the context of lighthouses.1

Studying infrastructure privatization in the modern era sheds light on the economics of these types of goods,

the public policy questions around who should own these assets, and the role of PE in the economy.

We focus on airports, examining how changes in ownership relate to changes in service quality and

financial performance. Airports are crucial strategic and economic assets, serving as gateways for people

and goods from around the globe to enter a city and its country. Airports and infrastructure more broadly

have distinctive features relative to other assets: They are large, long-term, provide an essential service,

and face little competition and high barriers to entry. Unlike other types of infrastructure, however, airport

revenue can be volatile and linked to the business cycle, as it depends on passenger and freight transport.

In this paper, we document airport privatization patterns over nearly four decades. We examine the

consequences of ownership type as well as variation across regulatory regimes. Airports provide a useful

setting for an international analysis because they share a common business model: Sell to passengers in

terminals and charge airlines for using the terminals and gates. Around the world, airports adhere to

common standards that enable aggregate performance analysis. We consider three ownership types that

have dominated the industry over the past fifty years: Public, non-PE private, and PE. Under public

ownership, the government owns and manages the airport. When the airport privatizes, a firm acquires the

right to operate, invest in, and earn residual cash flows from the airport in either a sale or a concession.

Following industry standards, we define a sale as either an outright sale or a long-term (>30 years) lease,

while concessions are shorter-term leases. We segment the private firms into those owned by infrastructure

funds (PE) and those that are not (non-PE private).

Relative to other work on infrastructure or privatization, one contribution of this paper is to consider
1
See Mill and Robson (1965), Pigou (1938), Samuelson (1964), and Coase (1974).

1
PE separately. PE represents a different economic model from independent private ownership, including

higher-powered incentives to maximize profits and shorter time frames for creating value. Infrastructure

funds have been growing rapidly and are now a major asset class within private capital markets. Between

2015 and 2019, these funds invested $388 billion and have more than $300 billion of capital raised and ready

to invest ("dry powder") as of 2022, up from $69 billion in 2011.2 The real effects of PE have been studied

in other sectors such as retail and manufacturing, but there is little evidence on whether PE creates value in

infrastructure, which is characterized by longer holding periods and intensive government monitoring. With

long term, stable cash flows, privatized infrastructure has proven an attractive class to institutional investors.

However, Andonov et al. (2021) document that in practice infrastructure funds have failed to outperform

the market on average, and value creation has been driven by short-term capital gains from quick exits, not

long-term holdings. This contrasts with evidence of strong returns in PE overall (Harris et al., 2014). It is

an open question whether infrastructure funds have similar real effects as PE in other sectors.

We combine a wide array of sources to paint a reasonably holistic picture of airport ownership and

operations. We begin with an expansive dataset of 2,444 unique airports in 217 countries. In the most

comprehensive, hand-collected privatization data effort to date (to our knowledge), we document that 437

have been privatized. Of these, 102 have at least once been owned by an infrastructure fund. As government

plays a crucial role in determining airport outcomes, even after privatization, we collect information on

airport price regulation as well as local government quality and business climate.

In our main analysis, we use airport-year panel data, with information on traffic and passengers at

airports with more than 10,000 passengers and 100 flights a year. We first examine determinants of private

ownership. Notably, PE but not non-PE buyers avoid airports with cost-based price regulation, where

prices must reflect costs and thus profits are limited. Infrastructure funds also target airports with higher

trade volumes. Motivated by Lerner and Schoar (2005), who show how cross-country governance affects

PE investment approaches, we consider local governance indices. Infrastructure funds target airports in

countries with better judicial effectiveness, and more financial freedom. A takeaway from this exercise is

that these two types of private ownership have different predictors, suggesting a more systematic approach

on the part of the infrastructure funds.

We next explore whether the ownership type is associated with operational changes. Our empirical
2
See https://www.ey.com/en_us/private-equity/how-pe-infrastructure-funds-are-getting-new-options,
https://www.ft.com/content/b00cb6b4-a015-493f-8cf6-ddc89e93ff34, and the Pitchbook 2021 Q1 Real Assets Report
(does not include Oil & Gas).

2
model, following the approach commonly used in the literature, is a differences-in-differences design. As we

are interested in both how airport performance changes post-privatization, and also whether PE ownership

has distinct effects from non-PE private ownership, we employ a single specification that makes use of the

four types of ownership changes we observe in the data. The first two are the privatization events, where

the airport transitions from government to either private non-PE or PE ownership. The second two are

post-privatization transactions where an airport transitions from non-PE private to PE or vice-versa.

Clearly, airports are not randomly targeted for ownership changes. PE and non-PE private firms acquire

airports where they believe they will earn the highest returns. Our case for causal effects is based on dynamic,

fully saturated differences-in-differences event studies which assess whether airports were already on track

to the outcomes we observe. To the degree the event studies suggest no pre-trends, we argue that our results

have a causal interpretation within the treated population. That is, they would not necessarily generalize to

any random airport by itself. Across broad swathes of the world, airports have in practice transitioned to

private and PE ownership, making the treated population economically relevant.

The results, estimated during our sample period of 1996 to 2019, paint a consistent picture in which

infrastructure funds improve airport performance. First, we consider traffic. Passengers per flight is a key

efficiency metric, enabling the airport to serve more customers with the same runway and gate infrastructure.

Under PE ownership, both in privatization events and in subsequent acquisitions from non-PE private firms,

the number of passengers per flight increases, for example by 20% in privatization events. In contrast, non-

PE privatization has no effect. This appears to reflect encouraging airlines to bring larger planes, as the share

of jets increases at the expense of regional and small aircraft. Overall passenger traffic increases under both

ownership types, but by more than four times as much–84%–under PE ownership. The number of flights

exhibit a similar pattern.

Airports are the sole gatekeepers to a key downstream sector: commercial airlines and the routes they

offer. Although we find average increases in the number of airlines and routes for both non-PE and PE

ownership, the event studies have strong pretrends for non-PE, suggesting that targeting rather than

operational changes explains the average effects. For PE, the increase in airlines reflect more low-cost

carriers. Greater competition among airlines may lead to better service and lower prices. Both in

privatization and post-privatization transactions, PE also increases the number of routes by much more than

non-PE. This is is driven by international routes, which increase by 46% after PE privatization. This

benefits passenger welfare and the local economy, as access to more routes creates new economic

3
opportunities (Bernstein et al., 2016).

We find dramatic declines in the flight cancellation rate under PE ownership, again both in privatization

and subsequent transactions. This presumably reflects improved operations. The event studies indicate that

under PE ownership, there is a decline in the share of departures that leave on time in the deal year, but by

the second year after the deal there is a large increase in the on-time departure rate relative to before the deal.

However, the average effects suggest declines in the on-time departure rate for both PE and non-PE. This

might reflect airport congestion, driven by the higher volume documented above, or more delayed incoming

flights. The airport can improve the former, but the latter is out of the airport’s control.

There are a number of aspects of the passenger’s experience in an airport that cannot be directly

observed, such as the quality of the stores, waiting areas, and overall cleanliness. To evaluate the impact of

privatizations on these factors, we employ ACI Word’s annual ASQ awards, which recognize airport

excellence in customer experience, based on surveys of passengers. We find that post-privatization

transitions to PE ownership increase the chances of winning an award.

We consider two financial dimensions: prices and airport financials. First, consistent with the airline

industry’s dim view of privatization (see Section 2.1), fees charged to airlines increase after privatization

by non-PE firms and in non-PE to PE transactions. Relatedly, there is a strong relationship between PE

acquisitions from non-PE private firms and the removal of price regulation, which could reflect PE owners

lobbying for deregulation. In a rare look at the income statements of private, PE-owned firms, we see that

net operating income increases by 108% after PE privatization. This appears to reflect higher revenues rather

than significant cost-cutting, as we see increases in operating expenditure and no change in employees per

passenger. We observe revenue increases, driven by both aeronautical (i.e., charged to airlines) and non-

aeronautical (i.e., terminal retail) sources. After transitions from PE to non-PE private ownership, we see

significant declines in net income and revenue, as well as in operating expenditure.

An important but understudied question in corporate finance is the extent to which ownership leads

to more efficient investments than control rights alone (Hart, 1995b). Airport privatization presents an

interesting setting for distinguishing between the roles of ownership and control by comparing concessions

(which confer control) with sales (which confer ownership). We find that across the main outcome variables,

sales lead to larger efficiency improvements than concessions, especially when they the new owner is a PE

fund. This suggests that ownership rights may lead to better aligned incentives, at least in our context.

Finally, we examine the role of the competitive landscape, which is especially relevant given antitrust

4
authorities’ increased scrutiny of PE. Our estimates indicate that improvements are much larger in the

presence of a competing airport, consistent with competition playing a beneficial role in the airport industry

for inducing the most productive outcomes under private ownership. However, this relationship is

magnified for PE relative to non-PE private ownership. PE ownership may be more responsive to

competitive incentives than other types of ownership because it entails higher quality managers and

higher-powered incentives to maximize profits.

In general, event studies indicate no pre-trends for our key results before PE acquisitions, consistent

with PE owners introducing operational changes. In contrast, there are strong pre-trends for outcomes

where we do see average improvements after non-PE private acquisitions, suggesting that these

improvements reflect targeting. In other words, non-PE private firms tend to purchase airports already on

track to better performance. We perform a number of robustness tests. For example, we use three methods

to address potential bias from multiple treatment periods (i.e., staggered roll-out), even though the vast

majority of observations in our sample are never-treated: the Callaway and Sant’Anna (2021) estimator, a

stacked regression (Cengiz et al., 2019; Baker et al., 2022) and a matching estimator (Huntington-Klein,

2021). Our key findings showing productivity gains following PE buyouts are robust to all of these

approaches.

Overall, we find strong evidence that infrastructure funds improve efficiency of the airports they

acquire. Of course, this conclusion comes with some caveats. Airports are (as we document) not randomly

targeted for privatization and PE acquisition, so despite our tests and the absence of pre-trends, it is

possible that the airport would have experienced the changes we observe in the absence of the ownership

change. Furthermore, we do not observe all dimensions of airport operation. Each data source does not

cover all airports, and some samples have relatively small numbers of PE-owned airports, though the main

results are robust to the overlapping samples. Nonetheless, we believe our sample and outcome variables

represent by far the most complete picture of ownership and operations for a class of infrastructure to date.

These data allow us to document for the first time how privatization and PE ownership affect the

operational and financial performance of infrastructure. Future research is needed and we hope that our

data will be useful in those efforts.

One takeaway from our analysis is that privatization consistently improves productivity only with PE

involvement. This likely reflects a combination of new strategies, including more equity-based

compensation for management and investment in better passenger services and technology alongside

5
well-targeted cost-cutting. One reason a hard-nosed airport owner may benefit the airport is the nature of

airport customers. Airports negotiate access, prices, and capital expenditure with airlines, which are

well-informed and well-resourced corporate stakeholders. Conversely, the customer population is more

vulnerable and faces information asymmetry in settings where PE has been found to have detrimental

effects, such as for-profit colleges and nursing homes (Eaton et al. (2020), Gupta et al. (2021)). For airports

bargaining with airlines, power varies depending on airline market share, ability to add new carriers, and

whether the airline can hold up the airport by using common new resources but refusing to pay higher fees.

Airports also often face competition from nearby airports or high-speed rail, limiting their monopoly

power, and their relatively wealthy, sophisticated passengers are likely to purchase less in terminals or go

elsewhere if quality is poor. Finally, airports face intensive local and central government scrutiny, creating

an ever-present threat of regulation. For these reasons, airports may be a setting in which PE’s capacity to

finance and orchestrate complex operational improvements benefits at least one key non-investor

stakeholder: passengers. Airlines may suffer from higher fees, but seem likely to benefit from higher

passenger volume. Importantly, the absence of strong effects for non-PE firms may not generalize to other

privatization settings, as it may be that the private firms—which were often founded to operate a newly

privatized airport—are not especially productive.

This paper sits at the juncture of three literatures: the political economy of privatization, infrastructure

economics, and the real effects of PE. Our results offer contributions to all three. First, while the privatization

literature has typically found large positive effects on firm performance of privatization (e.g. Dinc and

Gupta (2011)), we find weak effects in the context of airports for non-PE private firms.3 One possibility,

in addition to less-productive private firms mentioned above, is that the essential and highly salient nature

of airports may lead to decent performance under government ownership compared to other infrastructure

assets. Existing work on airport privatization has used small samples or case studies, with more descriptive

methods, and finds mixed effects of privatization on efficiency.4 There is also work on congestion in airports

(Brueckner (2002), Mayer and Sinai (2003)).

Second, we contribute to the literature on who should own and invest in infrastructure. This literature
3
Dinc and Gupta (2011) study drivers of firm privatization in India. While their study is mostly focused on targeting, they also
find evidence of efficiency improvements after privatization. Other work on privatization includes La Porta and Lopez-de Silanes
(1999), Megginson et al. (2004), Biais and Perotti (2002), and Dastidar et al. (2008).
4
In a sample of about 100 mostly North American airports, Oum et al. (2008) find evidence that privatization increases efficiency.
Van Dender (2007) studies determinants of airport revenues in the U.S. Other work includes Oum et al. (2006), Assaf and Gillen
(2012), Adler and Liebert (2014), Gutiérrez and Lozano (2016), Olariaga and Moreno (2019), Aguirre et al. (2019).

6
has focused primarily on the role of government, asking how political agendas affect investment, and how

that investment affects macroeconomic growth (Gramlich (1994), Milesi-Ferretti et al. (2002), Esfahani

and Ramırez (2003), Cadot et al. (2006)). For example, Donaldson (2018) shows the large impact of

India’s colonial rail network. Motivated by these findings and the fact that air travel represents the modern

equivalent of the rail, boat, and road systems studied in previous works, we take a different approach,

focusing on the nature of ownership. Our results shed light on control rights versus ownership rights and

highlight that especially in regulated contexts, high-powered incentives matter for performance.

Finally, we contribute to work on the real effects of PE.5 Studies have found positive effects on firm

performance and productivity in sectors with relatively little government interference, transparent product

quality, and high levels of competition, such as grocery stores, fast food restaurants, and manufacturing

(Davis et al. (2014), Fracassi et al. (2020), Bernstein and Sheen (2016)). In contrast, in sectors with opaque

product quality, intensive subsidy, and which traditionally rely on implicit contracts with consumers and

the government, there is evidence of negative effects (Eaton et al. (2020), Gupta et al. (2021), Phalippou

(2020), Liu (2021), Ewens et al. (2022)).6 We offer a setting that helps to reconcile some of these findings.

We find largely positive effects of PE ownership in a sector with intensive government regulation and little

competition. However, airport quality is salient to the local population and airports’ national strategic

importance and safety concerns yield motivated and politically empowered regulators, with other

sophisticated stakeholders—such as airlines—also monitoring. Thus it appears that PE can work well in a

highly regulated setting when incentives are well-aligned.

2 Institutional Context

In this section, we describe the history of airport ownership, privatization, and regulation, introducing the

key concepts peculiar to the strategic infrastructure assets that we will analyze later in the paper. We first

briefly discuss the history of privatization, then explain why infrastructure funds represent an economically

interesting and important class of ownership that is distinct from other private ownership.
5
A strand of the PE literature studies returns to investors, including Franzoni et al. (2011), Sensoy et al. (2014), Cavagnaro et al.
(2019), Harris et al. (2014), Robinson and Sensoy (2016), Andonov et al. (2021), and Gupta and Van Nieuwerburgh (2021). Other
work considers how PE structures transactions to create value (Ivashina and Kovner (2011)).
6
Other work on the real effects of PE includes Boucly et al. (2011), Olsson and Tåg (2017), Bellon (2020), Cohn et al. (2021),
Fang et al. (2021), Gao et al. (2021), Gornall et al. (2021), and Liu (2021). See Jenkinson et al. (2021) and Gompers and Kaplan
(2022) for surveys.

7
2.1 Privatization Background

Airports have historically been government-owned and operated, reflecting their role as vital strategic

assets for national economic growth, prestige, and security, with further implications for the local

environment and economy. Regardless of ownership, these features lead airports to be closely monitored

and regulated. Although today most airports remain owned by national or regional public sector agencies,

there has been a wave of privatization over the past fifty years, along with the privatization of many other

types of infrastructure, such as seaports, railroads, and utilities. This wave was inaugurated in 1987 when

the UK floated the British Airports Authority (BAA)—which consisted of Heathrow and six other

airports—on the London Stock Exchange. The rationales for this move and subsequent early privatizations

were reducing public sector inefficiencies and improving service quality (Graham (2020)). In the BAA

privatization, parliament explained that:

“The Government is committed to converting as many as possible of Britain’s airports into

private sector companies as part of its policy of reducing the role of the State. . . It will also

encourage more innovative management, and lead to efficiency gains and greater responsiveness

to customers.”7

After the 2008 financial crisis, the rationale for privatization shifted towards emphasizing financial gains

for the public sector (i.e., reducing public debt) and accessing private capital for needed new investment

(Van Nieuwerburgh et al. (2015), Cruz and Sarmento (2017)). Overall, airports have transitioned from

being considered public utilities to being considered as firms delivering services to airlines and retail stores

(Gillen (2011)).

Alongside these ideological trends, privatization manifested in three main ways. In the early phase,

IPOs were a common means, as in the case of BAA or Malaysia in 1999.8 In the late 1990s, governments

began to rely on two other forms: concessions and sales. In a concession, also called a public-private

partnership (PPP), governments grant rights to operate an airport and control the airport’s activities to a

private sector company for a specific period. During this concession period, the concessionaire operates the

existing facility and undertakes necessary capital investment to expand the airport while paying series of

concession fees to government, which can be fixed or variable (e.g., as a percentage of revenue). At the end
7
Secretary of State for Transport (1985), Airports Policy, Cmnd 9542, London: HMSO
8
Other examples are Vienna (1992), Copenhagen (1994), Auckland (1998), Zurich (2000), Fraport (2001), Thailand (2004),
Paris (2005), and Navegacion Aerea (AENA) in Spain (2015).

8
of the contract period, the airport typically reverts to the government. The government remains the airport

owner and has ultimate control rights (IATA (2018)). Examples of concessions include London Luton in

1998, as well as the Delhi and Mumbai airports in India in 2006.9 The third form, a sale, usually involves a

very long lease of the whole airport (say, 100 years) to a private company. For example, in 1997 and 2002,

Australia privatized the Brisbane, Melbourne, Perth, and Sydney airports this way.10 In a sale, the private

company obtains ultimate control rights over the airport. However, in both concessions or sales the private

firm typically has de facto control over operations and rights to residual cash flows conditional on the local

regulatory regime. Government typically closely regulates aspects of operations, sometimes limiting the

prices the airport may charge (discussed further below).

Early private airport owners were entities created to run one privatized airport, which then expanded,

purchasing others as they privatized in turn. One example of this is Fraport, which was initially established

to operate Frankfurt Airport and now runs most airports in Greece, among other countries. These private

conglomerates sometimes retain a tie to government; for example, local German governments—the State

of Hesse and the city of Frankfurt—together own just over 50% of publicly-traded Fraport. In most cases,

however, the private companies have minimal, if any, government ownership. Following the financial

crisis, the composition of airport buyers shifted away from private operators and international infrastructure

companies towards international funds sponsored by financial institutions, notably infrastructure-specific

PE.11 This new type of owner brought larger amounts of capital for improvements and, in theory, more

professional operational knowledge. As an early example, in 2001, asset manager H.R.L Morrison & Co

purchased Glasgow Prestwick Airport in the UK. More recently, in 2015, Corsair Capital purchased the

Lynden Pindling International Airport in the Bahamas.

Beyond the three broad categories of public, non-PE private, and PE ownership, there are many other

ways in which airports have private sector involvement, for example, by engaging private firms to develop

or run specific terminals. For the purposes of our study, since we do not observe terminal-level performance,

we are interested in overall airport ownership. This is also the most interesting unit of observation from the

perspective of control rights and residual claim to profits.

Privatization has not always been well-received by the airline industry, which complains of higher
9
Further examples of concessions, which sometimes involve new terminal construction contracts, include Lima (2000), Ankara
(2003), Montego Bay (2003), Brasilia/Sao Paulo (2012), Zagreb (2012), ANA in Portugal (2013), and Kansai (2015)
10
Other examples are Dusseldorf (1998), South Africa (1998), Turin (2000), Rome (2000), Milan (2011), and Toulouse (2015),
among many others, including regional airports in the UK.
11
See Condie (2015), Graham (2020).

9
charges without commensurate service improvements. During our sample period, the airline trade

association, IATA, reports that about 11% of total airline revenue is paid to airports.12 IATA also pointed

out in 2017 that average airport per-passenger charges in Europe increased from 16 to 33 Euros between

2006 and 2016. They noted that “The share of fully privately owned airports in Europe increased from 9%

to 16% between 2010 and 2016. . . While publicly owned airports may be considered as benign

monopolists, often pursuing economic and social goals in support of their local region, this is not the case

with privately-owned airports which are driven by investor returns” (IATA 2017).13 Airline welfare is

beyond the scope of this paper, but it is worth highlighting their opposition to privatization.

The U.S. has not historically privatized its airports because of strong federal government incentives

to remain publicly owned and operated, which do not exist in most other countries. First, public sector

airports in the U.S. can raise tax-exempt revenue bonds, while investors in bonds issued by private companies

must pay tax on the interest they earn. Second, airports can receive large federal Airport Improvement

Program grants if they commit to not making a profit from airport operations.14 Over the past 10 years, the

federal government has offered a limited number of exceptions to the normal grant restriction, but the only

successful instance of privatization has been the airport in San Juan, Puerto Rico. In other cases, such as

that in Westchester County Airport and Chicago Midway, local opposition to expansion or to privatization

on principle has derailed the efforts.15

A new model in the U.S., pioneered by the New York City airports, employs PPPs for the financing,

development, and operation of new airport terminals. PE has played an active role, with for example

Carlyle leading a $9.5 billion development at JFK.16 Unfortunately, operations and financials are not

typically available at the terminal level, and these terminal-specific deals are also too recent for evaluation.

For this reason, we do not evaluate these deals and instead treat the few airports in this category as

government-owned and operated.17 That said, it appears likely that PE’s footprint in the U.S. airport sector

will expand. For example, in 2017 Carlyle’s then-president Glenn Youngkin said: “There’s an
12
International Air Transport Association-IATA. (2007). IATA economics briefing 6: Economic regulation. Geneva: IATA.
13
https://www.iata.org/en/iata-repository/publications/economic-reports/airport-competition-myth-
or-reality/
14
https://reason.org/wp-content/uploads/annual-privatization-report-2021-aviation.pdf
15
For example, see https://www.theexaminernews.com/opposition-to-privatization-strong-at-airport-
hearing-in-armonk/a
16
https://centreforaviation.com/analysis/reports/jfks-new-terminal-one-accentuates-the-appeal-of-
the-airport-p3-590729
17
Another reason why U.S. airports are more challenging to study is that airlines sign complex, bespoke contracts in which
the airlines often manage and finance airport assets (such as terminals), and these contracts determine the payments they make
(Van Dender (2007)).

10
extraordinary amount of investment needed in airports. . . That’s probably going to be the top prospect for

investing in infrastructure over the near term.”18

2.2 Private Equity

This paper studies the evolution of airport ownership across three major categories: government, non-PE

private, and PE. PE represents a fundamentally different model for creating value. PE funds are financial

intermediaries, with capital raised from limited partners such as pension funds and endowments, who are

not involved in day-to-day investment and operational decisions. The general partners (GPs), who own the

PE firm and manage its funds, are responsible for the lifecycle of a deal: choosing the company to acquire,

negotiating the transaction, adjusting operations at the target firm, and finally harvesting value, usually via a

liquidation event in which they sell the portfolio company. The traditional transaction in PE is the leveraged

buyout (LBO), where the target firm is acquired with funds comprised mostly of debt—which is placed on

the target firm’s balance sheet—and a small portion of equity from the limited partners.19

PE is associated with particularly high-powered incentives to maximize profits in part because the GPs

who manage PE funds are compensated through a call option-like share of the profits (Kaplan and Stromberg

(2009)). Specifically, their compensation stems primarily from the right to 20% of profits from increasing

portfolio company value between the time of the buyout and an exit, when the company is sold to another

firm or taken public. The funds that purchase airports in our data are mainly closed-end, with roughly 10-

year time frames for liquidating assets and delivering returns to investors. Therefore, the PE managers are

considering avenues for exit at the time of purchase, since they must accomplish the sale within, typically,

3-8 years. In contrast, private firms purchasing airports often have very long-term time frames for generating

returns on the basis of sustainable cash flows. GPs also can receive transaction and monitoring fees, which

are not tied to performance. However, deals are typically not successful if the business continues as-is,

motivating aggressive and short-term value-creation strategies. In contrast, a traditional business owner

running the firm as a long-term going concern with less leverage may prefer lower but more stable profits.

There is evidence that PE buyouts increase productivity, operational efficiency, and generate high returns.

Kaplan and Stromberg (2009) argue that PE owners increase firm value through three channels, which
18
https://www.infrastructureinvestor.com/carlyle-lead-12bn-modernisation-new-yorks-jfk/
19
Kaplan and Stromberg (2009), Jenkinson et al. (2021) and Gompers and Kaplan (2022) provide detailed discussions of the
PE business model and review the academic evidence on their effects. In the interest of brevity, we limit our discussion. See also
Kaplan (1989), Kaplan and Schoar (2005), Gadiesh and MacArthur (2008), Guo et al. (2011), Acharya et al. (2013), Harris et al.
(2014), Robinson and Sensoy (2016), Korteweg and Sorensen (2017).

11
they call financial, governance, and operations engineering. The first channel includes alleviating credit

constraints, which may enable more investment (Boucly et al. (2011)) and exploiting the favorable tax

treatment of debt (Chambers et al. (2021)). Governance engineering includes changes to the compensation,

benefits, and composition of the management team at the target firm to align their incentives with those of

the PE owners: for example, instituting equity-based compensation (Gompers et al. (2016)). Bloom et al.

(2015) show that PE-owned firms are better managed than similar firms that are not PE-owned. In operations

engineering, GPs apply their business expertise to add value to their investments. For example, they might

invest in new technology, expand to new markets, and cut costs.

One motivation for our study is that infrastructure funds have distinctive characteristics that may lead

to different real effects, relative to the rest of the PE market whose real effects have been studied previously.

Infrastructure funds tend to be large, with more than 70% of capital raised since 2012 going to funds that

raise at least $1 billion. The average infrastructure fund in recent years is $2.7 billion, while the average

fund size in all other classes is $700 million. Infrastructure funds purport to offer the high returns of PE but

with more stable cash flows, less business cycle correlation, and lower volatility. Preqin reports that

institutional investors allocate assets to infrastructure overwhelmingly because they believe it offers

diversification benefits, with low correlation with other asset classes.20 Andonov et al. (2021) document

that limited partners investing in infrastructure funds do so because they believe the funds will have these

benefits. However, using cash flow data, they find that in contrast to the common narrative, infrastructure

funds deliver below-market returns and have similar volatility and business cycle exposure as other PE

vehicles. They believe that one reason is a focus on quick exits and the standard closed-ended structure

with periods of 10-12 years in which to invest, create value, and liquidate assets. It remains to be seen

whether this misalignment in time frames leads to negative real effects.

Infrastructure funds merit study in part because their footprint as an asset class has grown dramatically

in recent decades; in 2000, they invested just $2.2 billion, while in 2018, they invested $119 billion.21 It

seems likely that investment will increase in the medium term given the more than $300 billion in dry

powder.22 This massive growth reflects both large increases in private capital supply from pension funds,

insurance companies, sovereign wealth funds, and other institutional investors, but it also reflects
20
https://www.preqin.com/academy/lesson-4-asset-class-101s/infrastructure
21
Investment declined to $39 billion in 2021, which may reflect the pandemic. Based on data from Pitchbook. We use this for
data beyond airports as we do not have access to Preqin outside of airport deals.
22
https://www.ft.com/content/b00cb6b4-a015-493f-8cf6-ddc89e93ff34

12
governments increasingly seeking out private capital to make needed infrastructure investment under

constrained public resources. Infrastructure funds appear to place a premium on GP experience and

competence in the sector. Specialty knowledge about regulatory and operational issues is especially

crucial, not least because winning a project typically requires much more than simply the highest bid,

including a long negotiation process with the local government and a strategic plan for the airport.23

Because of their importance and visibility, changes to airport ownership are typically politically sensitive,

are closely monitored by the local government and generally are objects of close media scrutiny. This

requires special expertise, including the ability to partner with government over the long term, on the part

of the PE managers.

2.3 Airport Revenue

One distinctive feature of infrastructure as an asset class is that governments often regulate how the private

owner or operator of an infrastructure asset may earn revenue, sometimes with the justification that the asset

represents a natural monopoly. Therefore, the price regulation regime and its capacity for change are crucial

inputs to profitability. From the government’s perspective, there is a need to design schemes that will (a)

induce effort on the part of the private firm; (b) keep prices within politically acceptable limits; and (c)

ensure the private firm does not risk being held up. Regulators and politicians often face pressure from local

users for “fair” prices. The constituent base may prove more impactful than the interests of a few, often

non-local, investors. However, holding up investors may make it more difficult to attract new infrastructure

capital. Van Nieuwerburgh et al. (2015) suggests that investors price regulatory risk into their required rate

of return, which has led some jurisdictions, especially middle-income countries, to build a reputation for

contract reliability. Low-income areas may be less politically stable, while high-income countries may have

less urgent need for the external investment. Both situations could lead consumer demands to win the day.

Acharya et al. (2020) formally model these hold-up problems in the infrastructure investment setting.

Airport revenue and pricing follows from the two basic functions of an airport: To enable the airplane

to safely take off and land, and to move the passenger through the terminal. There are two primary sources

of revenue. The first and larger source is aeronautical revenue, which comes from airlines. This includes

per-passenger charges as well as per-landing and per-takeoff runway charges. The per-passenger charges

represent fees for using the terminal building, while the runway charges represent fees for the airplane to
23
CohnReznick. “Infrastructure Investment Report.” 2021.

13
use the runways and gates. Passenger- and aircraft-related charges reflect features such as the size of the craft

and the timing (weekday vs. weekend, morning vs. evening) of departure and arrival. Passenger charges and

runway charges represent 41% and 21% of aeronautical revenues, respectively.24 Parking charges are the

third largest source of fees (12%), and the rest varies by airports, including noise and environmental charges

and government fees.

Second, non-aeronautical (commercial) revenue comes from retail leases as well as ancillary passenger

services such as parking garages and transportation (taxis, buses). Airports earn rents from retail space

in the terminals for stores such as duty-free shops and restaurants. These leasing contracts are generally

structured as minimum fixed rents plus additional profits from sales above specific thresholds. Retail leases

represent 28% of non-aeronautical revenue and car parking and property (e.g., rental cars) revenue are

secondary sources (20% and 18%, respectively). Non-aeronautical revenues were on average 38.8% of the

total revenues in 2013.

An obvious way to improve revenue is to capitalize on passenger growth. An increase in the number

of flights and passengers directly leads to improvement in per-passenger revenues, boosting retail income

growth at the same time. The ACI Airport Economics Survey (2014) highlights the importance of economies

of scale in profitability by documenting the positive relationship between net profit margin and airport size.

While airports with less than one million passengers have average profit margins of 11.9%, airports with

more than 15 million passengers earn 19-20% of net profit margins.

When it comes to reducing cost and expenses to improve profits, it is worth noting that airports are

subject to high fixed costs, and the world’s major airports are run at almost full capacity (Gelhausen et al.

(2013), Dray (2020)). Large-scale expansion such as adding runways or terminals is often risky because

it requires lumpy and irreversible capital investment with much uncertainty about future demand growth.

New additions also typically must clear high local land use, noise, and environmental hurdles. Thus, for

boosting economic profits, it is typically most desirable to increase operating efficiency by accommodating

more passengers per flight and handling a higher volume of aircraft movement per runway.25 Although it is

not always possible to increase the number of international passengers, these are often the most profitable,

especially in developing countries. Airport owners can achieve higher volume without adding space in

various ways, including by investing in technology, management, and better use of existing space.
24
According to the ACI Airport Economics Survey (2014) of 653 international airports.
25
Based on the authors’ conversations with industry executives, including from RDC Aviation.

14
 What determines the prices that airports charge to airlines (aeronautical revenue) and to the concessions

in the terminals (non-aeronautical revenue)? At privatized airports, airport revenues are sometimes

regulated, with price regulation applying either to both aeronautical and non-aeronautical charges

(“single-till”) or to only aeronautical charges (“dual-till”). Regulation takes the form of either rate of return

limits or price caps. Although it is thought that there is not a major difference in practice between these

forms, price cap regulation may encourage cost reductions, potentially at the expense of service quality

(Starkie (2004), Gillen and Niemeier (2008)). Finally, some governments—notably in Australia—use an

explicit threat of regulation to mitigate monopoly pricing at wholly unregulated airports (Forsyth (2008)).

Airport and airline market power both play a role in price setting. Airports on islands, without nearby

competing airports, without strong regional transport alternatives (e.g., high-speed rail), that are more

congested, and those with more international traffic tend to have more market power (Basso (2008), Bel

and Fageda (2010)). One reason that airports can be profitable is that airlines are believed to have low

demand elasticity (Bel and Fageda (2010)). However, airlines also have negotiating power, which increases

with their share of airport traffic (Borenstein and Rose (2014)).

3 Data Sources and Summary Statistics

This section describes the data sources and samples used in analysis.

3.1 Ownership Type and Deals Data

We begin by constructing a list of airports around the globe with more than 10,000 passengers as of 2016.26

This initial sample includes 2,444 international or regional airports located in 217 countries. We hand-collect

the historical ownership structure of these airports. We consider three ownership types, which capture the

key markers along the continuum from wholly public to high-powered private incentives:

1. Public: Government owns and manages the airport;

2. Non-PE Private: A private firm owns the majority of the airport and its management rights; and

3. PE: A private equity fund owns the majority of the airport and its management rights.27

The ownership breakdowns are depicted in Figure 1. While other types of investors, including sovereign
26
This list is from: https://ourairports.com/data/.
27
We identify a deal as PE when the PE firm leads the transaction and has the single largest stake among the acquiring syndicate.
In practice, the PE stake varies from 10% to 100%. Table 1 contains statistics on the stakes purchased by the PE firm.

15
wealth funds, insurance companies, and pension funds have direct stakes in airports, these tend to be smaller

minority stakes and passive. We do not consider them since we are interested in airport operations.

To identify historical ownership changes, we combine various data sources including Preqin,

privatization case reports by ICAO, annual privatization reports by the Reason Foundation, airport annual

reports from airport websites, and online news reports of airport transactions. Preqin provides

transaction-level data on infrastructure deals that covers asset investors, funds, assets, and deal dates. Of

the 2,444 airports, 437 have been privatized during our sample period. Of these, 102 have been at least

once owned by an infrastructure fund. About 43% of these 437 airports are concessions. Table 1 describes

the ownership changes by parties and transaction type. In our analysis, we exploit the fact that we observe

privatization to non-PE and to PE, as well as sales from non-PE private to PE. If an airport is

government-owned (i.e., public) at the time of our data collection, we assume it has always been so since

governments rarely buy back privatized airports.28

Table 1 characterizes the ownership transitions in our sample of airports. We split the transition deals

into concessions in Panel A and sales in Panel B. Out of 2,444 airports, 437 (18%) were privatized. Most

of these deals (401) were of government-owned airports that were privatized by a non-PE group. There are

36 PE privatizations of government-owned airports and 82 cases of a PE fund acquiring an airport that was

previously privatized.29 The privatizations occurred at an increasing rate over our sample period, with 15

before 1990, 97 in the 1990s, 160 in the 2000s, and 148 in the 2010s. Figure 1 shows how privatization

has proceeded in terms of the number of airports (Panels A and B), the share of global passengers (Panel

C) and the share of global flights (Panel D). We see an increase in the role of private and specifically PE-

owned airports, even though the U.S.—with the second-largest number of airports (after China)—has no

privatized airports. For example, Panel C shows that the share of total passenger volume at PE-owned

airports increased from about 1% in the late 1990s to 11-12% in the 2010s.

Privatizations have taken place across the globe, though they are concentrated in certain regions. Figure

2 and Appendix A.1 present statistics on the number of privatizations in each decade for each country.

Airport privatizations are particularly common in Latin America, with 42 in Mexico, 29 in Argentina and 19

in Brazil. In general, these events are correlated with broader privatization initiatives and adoption of more
28
We were only able to observe 4 cases that went back to government-owned from private out of 436 privatized airports in our
sample.
29
Many of the airports (186) were privatized through concessions, in which the firm received the right to operate the airport’s
facilities but do not own it. These acquisitions are mostly by non-PE.

16
liberal, market-based economic policies.

Appendix Table A.2 lists the top PE acquirers (by number of airports acquired) as well as characteristics

of funds and deals. In our data, infrastructure funds acquiring airports tend to be closed-ended (85% are

closed-ended and the remainder are open-ended). However, the mean holding period in our data is 8.3 years

with a median of 7, conditional on exit. However, only 27% of deals have exited by year 10. This pattern

suggests much longer holding periods than the traditional standard in LBOs. This may reflect the longer

investment lead times of airport infrastructure improvements, as well as the different types of risk, notably

macroeconomic demand changes. PE may also be more adept at implementing new aviation and consumer

retail technology. The top non-PE private firms are described in Appendix Table A.3.

3.2 Regulation and Governance Data

We obtain histories of airport price regulatory regimes from David Gillen at University of British Columbia.

The data cover 79 major airports from 1990 to 2018 in Asia, Europe, and Oceania. Twenty-four of these

airports were at one time owned by PE owned and 30 by non-PE private firms.

We also employ national governance indices from the Heritage Foundation. These data cover 186

countries from 1995 to 2019 and include 12 quantitative and qualitative factors of national governance

measures, grouped into the following categories: rule of law (property rights, government integrity, judicial

effectiveness), government size (government spending, tax burden, and fiscal health), regulatory efficiency

(business freedom, labor freedom, and monetary freedom), and open markets (trade freedom, investment

freedom, and financial freedom). Each of the 12 measures is graded on a scale of 0 to 100. Details on the

definition of each variable are in Appendix B.

3.3 Airport Performance Data

Traffic Passenger and flight traffic data come from the International Civil Aviation Organization (ICAO).

As these data begin only after 2014 for some airports, we supplement them with data from Official Aviation

Guide of the Airways (OAG). We consider airports with more than 10,000 passengers and 100 flights a

year to focus commercial airports, excluding airports that are exclusively for general aviation or military

activities. These filters give us an airport-year level sample from 1996 to 2019 consisting of 2,353 airports,

including 395 that have been privatized by non-PE private firms, and 92 that have been PE-owned.

17
Fees We obtain information on annual airport fees levied on airlines from RDC aviation, an analytics

company. The fees that airports charge to airlines are on a per-aircraft-event basis (i.e. takeoff and landing)

and are related to services provided to passengers and airlines, such as passenger service fees, runway fees,

plane parking fees, infrastructure fees, aircraft security fees, and noise fees. By far the largest are passenger

and runway fees. One of the advantages of the charge data from RDC is that the charges are quantity-free. In

other words, the RDC database provides information on the calculated fees that airports charge to airlines at

the aircraft level for various categories of aircraft types, departing and landing time and seasons. However,

realized revenue for the airport depends on the passenger traffic and the volume of aircraft traffic, which

would be sensitive to the economic cycles.

The fees in the data are based on three different aircraft types based on their route and size: small

domestic jet, short-haul international mid-sized jet, and long-haul international jumbo jet. We consider two

standard aircraft types for analysis: DH4 (small domestic jet) and 77W (long international jumbo jet). DH4

can carry up to 78 passengers per flight, and 77W can carry up to 348 passengers. Per passenger charges are

calculated assuming that aircraft is 80% full. Charges with fixed aircraft type and the number of passengers

make it convenient to examine the change in the fixed price charged to airlines after privatization. The data

set includes each fee type by aircraft type. Applying the same filters as above, we are left with an airport

year level sample from 2010 to 2020 consisting of 1,508 airports, including 79 airports acquired by PE, and

336 airports acquired by non-PE private companies.

Financials We obtain information on airport financials from the Air Transport Research Society (ATRS) at

Embry-Riddle Aeronautical University. These data span 2001 to 2017. The database includes financials of

225 airports worldwide, which are focused on large international airports in the US, Europe, and Asia. The

data mainly cover financial variables including total operational revenue, total operational expenditure, net

operational income, total aeronautical revenue, total non-aeronautical revenue, and number of employees.

Punctuality We obtain information on airport on-time performance from the monthly reports of airport

punctuality produced by the Official Aviation Guide of the Airways (OAG). The report provides punctuality

performance of 1,689 airports worldwide from 2016 to 2020, including 97 airports that were acquired by

PE, and 404 airports that were acquired by non-PE private companies.

18
Airlines and Routes We obtain information on route information and the number of passengers and flight

frequencies of each route, and the operating airlines produced by the Official Aviation Guide of the Airways

(OAG). The data include airport-year-airline-route level traffic information. We drop airports with less than

10,000 passengers and 100 flights a year.

Awards We obtain information on airport awards from Airports Council International (ACI). ACI gives

awards every year the airports with the best service quality through the ASQ Awards program, a airport

passenger satisfaction program with airports in 95 countries. The awards data span 2006 to 2021. According

to their website: “ACI World’s annual ASQ Awards recognize airport excellence in customer experience

worldwide based on data from ASQ’s renowned Departures and Arrivals Surveys.”30

Accidents and Fatalities We obtain information on number of accidents and fatalities by scraping the

Flight Safety Foundation’s Aviation Safety Network (ASN).31 The variables are the number of accidents

and fatalities for flights that took off from the airport. These data span 1996 to 2019.

Table 2 presents summary statistics describing the data introduced above, separately for government,

non-PE private and PE ownership. PE-owned airports tend to be in richer countries than publicly-owned

airports, while non-PE owned airports tend to be in poorer countries; the average GDP per capita for PE-

owned airports is around $34,020, $25,480 for publicly-owned airports, and $17,260 for Non-PE owned

airports. The PE-owned airports also have better scores on measures of governance. In particular, PE deals

tend to be in countries with less corruption and better government effectiveness. At the airport level, there are

also systematic differences. On average, PE-owned airports are larger, have a higher fraction of international

flights, and charge higher fees.

3.4 Targeting Analysis

To more rigorously analyze these differences, we explore which airport characteristics predict PE- and

non-PE deals, relative to government ownership. Specifically, we use a sample of airport-year observations

that omits all years after privatization. The results are reported in Table 3. The first two columns focus on

economic factors. They suggest that both PE and non-PE target airports with more international
30
See: https://aci.aero/programs-and-services/asq/asq-awards-and-recognition/
31
https://aviation-safety.net/database/

19
passengers, and PE funds target countries with more trade. The third and fourth columns examine price

regulation. PE appears to avoid cost-based regulation, which is natural since this would place restrictions

on profits. Governance variables are in columns 5-6. PE targets airports in countries with more judicial

effectiveness, financial and investment freedom, and less government spending. In unreported analysis, we

do not find that other observed variables significantly predict acquisitions, nor do they substantially affect

the relationships described here. Overall, these results suggest that private firms in general target airports

with more profitable passenger characteristics, while PE appears to discriminate when it comes to local

governance and regulation.

4 The Effect of Privatizations on Airports’ Performance

The canonical question in the PE literature, dating to Kaplan (1989), concerns the extent to which

acquisitions by PE funds lead to real changes in the firms they acquire. Do PE funds merely acquire firms

at good prices, add leverage to increase expected returns and then realize the returns on their investments,

or do their returns come at least in part from real improvements in the firms they acquire? This issue is

particularly important in the case of sales of government assets such as airports, since there is concern

about corruption leading to sales at prices below their value and rent extraction (Hoffman, 2011).

The ideal experiment would randomize privatization events across airports and ownership type.

Unfortunately, in practice this is impossible and indeed the targeting analysis revealed that there are in fact

systematic predictors of privatization that likely relate to the airports’ capacity for delivering value in the

future. From a policy perspective, the desirability of airport privatizations clearly depends on the extent to

which the transactions that do occur lead to real improvements in the airports’ operations. This is the

question we seek to shed light on in our analysis.

4.1 Empirical Specification

Our interest in this paper is asking how airport performance changes post-privatization, and also whether

PE ownership has distinct effects from non-PE private ownership. We accomplish both of these goals in a

single specification that makes use of the four types of ownership changes we observe in the data. The first

two are the privatization events, where the airport transitions from government to either private non-PE or

PE ownership. The second two are secondary transactions where an airport transitions from non-PE private

20
to PE or vice-versa. To capture these four transitions, we employ the following differences-in-differences

estimating equation.

Yi,t = β1 1(Privatization by PE)i,t + β2 1(Privatization by Non-PE)i,t + β3 1(Post-Priv Non-PE to PE)i,t

(1)
+β4 1(Post-Priv PE to Non-PE)i,t + Xi,t
0
γ + δi + θt + εi,t

The independent variables of interest are relative to government ownership as the base group.

1(Privatization by PE)i,t is one after an airport transitions from government to PE ownership and zero
otherwise. Similarly, 1(Privatization by Non-PE)i,t is one after an airport transitions from government to

Non-PE private ownership. 1(Post-Priv Non-PE to PE)i,t is one after an airport that is already privatized by

a non-PE firm transitions to PE ownership. 1(Post-Priv PE to Non-PE)i,t is the reverse. We report two
p-values on F-tests for equality of coefficients. The first compares the two privatization coefficients and the

second compares the post-privatization coefficients. As Table 1 shows, 401 privatizations are to non-PE

while 36 are to PE, and 71 post-privatization transactions non-PE to PE while just 18 are PE to non-PE.

This highlights how global airports have shifted over time from government to private and then to a higher

share of infrastructure fund ownership. One downside is that we should expect that estimated effects of

privatization to PE as well as transitions from PE to non-PE ownership will be noisier, since they are

identified off of fewer observations.

To control for macroeconomic growth, airport size, and country governance indices related to the

demand for air transportation, we include in the vector Xi,t log GDP per capita, log trade volume, log total

number of passengers, share of international passengers, government size, open markets, judicial

effectiveness, and government integrity. These may evolve differently at control vs. treated airports. All

models also include airport and year fixed effects. The results are also robust to using deal fixed effects.

Standard errors are clustered by airport.

Following the recent literature on resolving bias in two-way fixed effects models with multiple treatment

periods, we test whether the main results spuriously reflect the staggered nature of the transactions. Since

airports are acquired at different dates, and thus the control group depends on the year and may include not-

yet-treated airports, treatment effect heterogeneity and dynamic treatment effects could lead to bias. The

first test is the Callaway and Sant’Anna (2021) estimator, which estimates treatment effects specific to each

group-time and then averages them together. This estimator permits only one treatment variable, and we

21
have four. Therefore, for this test we consider only an indicator for PE ownership, since our central results

where we find evidence for causality concern PE. The second approach is stacked regression, which (Baker

et al., 2022) explain resolves the concern by creating event-specific datasets and stacking them together. The

estimator then includes dataset-specific unit- and time-fixed effects, enabling entirely “clean" controls. This

approach has also been used in Gormley and Matsa (2011) and Cengiz et al. (2019), among others. One

benefit is that we can replicate our main model and include all four treatment variables. The third approach

is a matching estimator, where we address the concern by using only never-privatized airports as the controls

(Huntington-Klein, 2021).

We also estimate fully saturated dynamic differences-in-differences models that allow us to take an event

study approach to the data. Here, we focus on the two private ownership types, PE and non-PE, and look

for the average effects of each separately, as it is infeasible to estimate all or even two effects by year

in the same equation. The event studies are important for making a causal statements: they will reveal

whether, conditional on our controls, airports that are acquired by either PE or non-PE firms were on track

to experience the effects that we see post-acquisition. We use Equation (2) below for PE.

X
Yi,t = β s PE Deal Yeari,s + Xi,t
0
γ + δi + θt + εi,t (2)
s,0

We employ the same specification for non-PE private acquisitions. Variables are as defined for Equation (1),

and standard errors are again clustered by airport.

4.2 The Volume of Passengers and Traffic

The number of passengers per flight coming into and out of the airport is an important metric of performance

because it means more efficiency on the tarmac and more people in the terminals to shop at concessions,

conditional on tarmac capacity. Note that it can increase either because of larger planes (i.e., with more

passenger capacity), or because of fewer empty seats on a plane. By adjusting fees and agreements with

airlines, airports can induce them to fly more saturated routes with larger planes, which increases both

aeronautical and non-aeronautical revenue to the airport, since aeronautical revenue is in part on a per-

passenger basis.

Figure 3 contains dynamic differences-in-differences estimates, illustrating the change in the number of

passengers per flight for PE-acquired airports around the time of the acquisition. Each coefficient estimates

22
an effect relative to control airports in the same year, which helps to control for factors affecting the overall

economy and the airport industry. The plots indicate that following the acquisition (year 0) there is an

increase in passengers per flight for both domestic and international flights. Importantly, the figure contains

no evidence of pre-trends before the acquisition. While we cannot firmly identify causality, the lack of pre-

trends and discontinuous change after the acquisition suggests that in the absence of the acquisition, these

targeted airports would not have experienced an increase. We observe no similar pattern in an event study

of non-PE acquired airports, which we present in Appendix Figure A.1.

The average differences-in-differences estimates from Equation (1) are reported in Table 4. After PE

privatization, airports increase passengers per flight by about 18 relative to government ownership, which is

20% of the mean of 90 passengers (column 1, top row). This is driven by increases in primarily domestic

flights (columns 2-3). In contrast, there is no overall effect for non-PE privatization (column 1, row 2),

though there is a small positive effect in domestic flights (column 3, row 2). In secondary transactions,

there is a robust increase in passengers per flight for non-PE to PE deals, driven by both international and

domestic traffic (columns 1-3, row 3). The effect of transitions from PE to non-PE ownership is positive

but insignificant (column 1, bottom row). To assess whether these effects on passengers per flight reflect

larger or fuller aircraft, we examine aircraft composition. We find that the share of jets increases after PE

acquisitions while the share regional and small aircraft declines (columns 4-5, rows 1 and 3). The aircraft

composition does not change after Non-PE privatization, but there is an increase in the share of jets in Non-

PE to PE deals (columns 4-5, rows 2 and 4). The results are consistent with PE increasing passengers per

flight at least in part by inducing airlines to bring larger planes to the airport.

Another way that privatization can increase the airports’ value is by increasing the total number of

passengers and the number of flights. The event studies in Figure 3 again show no pre-trends, while after PE

acquisition the number of passengers and number of flights increase discontinuously. This is driven by both

domestic and international passengers (Panels C-D). Appendix Figure A.1 shows little change after non-PE

private acquisition. The average estimates, reported in Table 4, show that following PE acquisitions the

total number of passengers increases by 84% in the initial privatization and 17% in purchasing an already-

privatized airport from a non-PE firm (column 4, note that since the outcome is logged we exponentiate for

interpretation). Non-PE privatization also leads to more passengers, but the lack of an event study result

means we should interpret this with caution. Both types of private ownership increase the number of flights

(columns 7-9), with the largest effect for PE privatization (column 7, first row).

23
 Last, we consider freight traffic in Appendix Table A.4. We see evidence of improved efficiency under

PE ownership measured as freight per flight in columns 1-3. Both for privatization and subsequent

transactions, there is evidence that PE ownership has statistically significantly larger effects. For levels,

measured as log tons of freight, we also see much higher increases for PE ownership, though this is in

some cases driven by a negative effect for non-PE private ownership. Overall, these results on volume and

traffic indicate better performance under PE ownership, with non-PE private ownership being no better than

government ownership.

4.3 Downstream Performance: Routes and Airlines

An airport’s value to the economy depends not only on the total number of passengers, but also on the

choices of routes and competition among airlines, with more routes and airlines being better for passengers,

both because they have more options and likely lower prices. Indeed, the number of locations to which one

can fly on a nonstop flight can materially affect the desirability of a city as a place to live, tour, or to locate

a business. For example, when JP Morgan’s infrastructure fund acquired the airport in Cairns, Australia

(near the Great Barrier Reef), the first thing they did was to add nonstop flights to major Asian cities, which

increased the number of tourists visiting the Great Barrier Reef.32

The event studies in Figure 4 show significant and immediate increases in the number of routes after PE

acquisitions. As in the Cairns example, the increase is primarily in international routes (Panel A), which are

typically more profitable. This is driven by both international routes in privatization and post-privatization

acquisitions by PE firms, shown columns 1-3 of Table 5. Column 2 shows that after privatization by PE, the

number of international routes increases by 46% (note that routes are logged so we exponentiate). Figure

A.2 for non-PE private acquisitions shows strong pre-trends, indicating that non-PE private firms seem to

select airports already on track to adding more routes, and suggesting that any positive coefficients should

not be interpreted as causal effects.

In addition to routes, when more airlines serve an airport, passengers will likely benefit from more

options for subsequent connections and lower prices resulting from increased competition. We evaluate

whether privatization leads to entry of new airlines in Figure 4 Panel C, where there is a striking increase in

the year of the deal, suggesting an immediate focus by the new owners on more carriers, especially by the

low cost carriers. As mentioned in Section 2.3, the airport will have more bargaining power with its primary
32
Source: Private conversations with deal participants.

24
customers—airlines—when there are many airlines at the airport jostling for access rights. Therefore, adding

airlines is an obvious means to increase prices and profits. However, we do not see measurable changes to

airline HHI, which is very noisy (Panel D). When it comes to non-PE private firms, Figure A.2 indicates

strong pre-trends, pointing again to a different economic model from PE and an absence of causal effects.

The average estimates, in Table 5 columns 4-5, paint a less clear picture for PE. PE increases the number

of low cost carriers when they privatize. For all three outcomes we observe large estimates for non-PE

privatization, but these come with a caveat due to the lack of event study evidence.

Putting the event studies and the table together, we conclude that there is a robust and likely causal

relationship between PE acquisitions and the number of both routes and airlines, but that the rest of the

results should be interpreted with caution. More airlines at PE-owned airports likely increase consumer

welfare through both improved choices of routes and lower prices.

4.4 Punctuality, Safety, and Awards

We are also interested in the quality of an airport from passengers’ perspectives. While product quality is

typically difficult to measure, we have several useful metrics in the airport setting. First, flight cancellations

are perhaps the largest nuisance to passengers and also disrupt airport operations as they tend to create

unexpected congestion. The event studies, in Figure 5, indicate a clear decline in cancellations after PE

ownership (Panel A). For non-PE private acquisitions, the cancellation rate decreases as well, though there

again appears to be a pre-trend (Figure A.3). The average estimated effects from the multivariate model

presented in Table 6 show a striking decline in the percent of flights canceled after PE acquisitions from

non-PE private firms of 0.98%, which is about 50% of the mean rate of 2% (column 1). This decrease in

cancellation rates is consistent with PE improving runway and gate management operations in a way that

materially benefits passengers, who face fewer canceled flights.

We next examine the fraction of flights that depart on time. There is a decline in the share of departures

that leave on-time in year 0, the year in which the deal occurs, but that reverses so that by year 2, the

fraction of on-time departures higher than it was before the deal (see Figure 5, Panel B). Column 2 of Table

6 shows that the on-time departure rate decreases after acquisitions by both PE and non-PE. On average,

a 1% increase in the number of flights leads to a 2.7% decrease in on-time departures. This decline could

reflect the high correlation of flight delays with airport congestion and delayed incoming flights. The airport

can improve the former, but the latter is out of its control. Therefore, higher traffic volume, routes, and

25
airlines could explain the decrease in on-time departure rates.

There is much more to a passenger’s experience at an airport than the actual flights, including security

wait times, cleanliness of the restrooms and quality of the stores and lounges. All of these vary in quality

across airports and can materially affect passengers’ welfare. We assess these dimensions using the ACI

ASQ awards data. These prizes are offered to airports whose passengers report the most positive and

smooth experiences in surveys. Column 3 of Table 6 indicates that privatization increases the likelihood

that an airport wins an award. The most robust effect, however, is for transitions from non-PE private to PE

ownership, in which the chance of an award increases by six percentage points, which is three times the

mean. The event study in Figure 5 Panel C indicates that there is a distinct jump on the chances of winning

the award after PE acquisitions.

Finally, in columns 4 and 5, we consider the possibility that airport privatization affects safety. These

equations predict the number of accidents and fatalities per 1,000 flights. The coefficients on the privatization

variables are small and not statistically different from zero, suggesting that privatization does not have a

material impact on the safety of flying. We would expect these numbers to increase given higher volume of

passengers and flights, so we can interpret this as a positive result.

4.5 Fees Charged to Airlines and Regulation

During our sample period, fees that airports charge to airlines account for about 75% of airport revenue.

The two main charges are runway fees, which are paid for each takeoff and landing, and passenger fees,

which are paid for processing passengers and security services. Figure 6 presents the PE event studies for

the fees broken down by domestic and international flights. Fees increase immediately and persistently

following the acquisition. The event studies for non-PE private in Figure A.4 suggest that fees at these

airports start systematically higher than at control (public) airports, as indicated by the lower value for the

omitted coefficient at year -1, and increase only slightly after the acquisition.

In Table 7, we estimate the average effects. The results indicate that, consistent with the airline

industry’s complaints (see Section 2.1), total fees increase substantially following privatization by non-PE

deals (columns 1-6, row 2) but slightly decrease following privatization by PE.33 In secondary transactions,

the picture changes; there are fee increases when PE acquires airports from non-PE private owners, while
33
The estimate is the combination of the runway fee and the passenger fee. Although the runway fee increases after PE
privatization, the passenger fee decreases, which is the dominant of the two, so overall, the total fee decreases.

26
there are some decreases and some increases when PE sells to non-PE (columns 1-6, rows 3-4). In sum,

both PE and non-PE ownership are associated with higher fees relative to government ownership.

However, despite the higher fees, traffic increases (Table 4). Presumably, other improvements after

privatization outweigh the higher fees to attract new traffic.

When airports are privatized, the government must decide whether to regulate airport prices, since

airports have some degree of market power. As with other infrastructure and utility assets, regulation can

take two principal forms: revenue caps and cost-based. As explained in Section 2.3, both limit the private

owner’s ability to increase profits and would be anathema to PE, where maximizing cash flows and firm

value in the near-term is a key objective. Consistent with this idea, the estimates we present in Table 3

document that PE investors are more likely to target airports without pre-existing cost-based price

regulation. Column 7 of Table 7 presents estimates of equations in which the dependent variable is an

indicator for no price regulation.34 After PE acquisitions, there is a higher chance of deregulation, with the

relationship after transitions from non-PE private being almost 200% of the mean. This deregulation may

help produce the incentives for the performance improvements that we see in the previous tables. There

may be many drivers for pattern, but one possibility is that PE owners lobby for deregulation.

Consider the example of three Australian airports that were privatized in 1996-7, two with majority

PE ownership. Each had revenue caps for five years subsequently, which were removed in 2002 in part at

the request of investor groups. As Gillen (2011) explains, the airport owners and government settled on a

strategy of price monitoring, which occurs to some degree at all privatized airports without price regulation,

creating an explicit threat of regulation. This essentially amounts to a trigger or “grim” strategy, in which

seemingly excessive profits would lead to long-term regulation. Amid a 2018 reconsideration of airport

regulation, investor owners of Australian airports submitted a brief arguing explicitly against regulation,

noting that the “light-handed regulatory regime encourages commercial outcomes, incentivises innovation

and allows investors to earn appropriate risk-adjusted returns. . . [A]irport owners take the threat of regulation

seriously.”35 While clearly this example does not necessarily apply elsewhere, it shows that investor lobbying

can yield regulatory changes.

34
Since regulatory regimes change only rarely and these changes are undoubtedly related to the causes of the privatization, we
do not attempt a causal analysis here. Also, there are insufficient cases (only one) of PE to non-PE transitions in the dataset with
regulation information so we exclude this transition from the model.
35
Australian Airports Investor Group Submission to the “Productivity Commission Review of the Economic Regulation of
Airports,” September 2018. Available here: https://www.pc.gov.au/__data/assets/pdf_file/0011/231122/sub020-
airports.pdf

27
4.6 Financial Outcomes and Employment

The results on fees and traffic point to higher revenues after PE acquisitions, relative to airports that remain

government owned. We next evaluate the impact of privatization on the financial performance of the subset

of airports for which we have income statements. Increasing cash flows is a key strategy in PE, but it is

rare to observe income statements for privately owned firms, so studies of PE’s operational impacts have

typically been unable to show how cash flows change after buyouts and, in particular, whether revenue

increases, costs decline, or both. Instead, the literature has focused on other observable outcomes. This

paper contributes to the literature on PE by directly studying key elements of the income statement. We are

able to do so because some countries require airports to publicly release this information. However, since

only a subset of airports have financial data available, the sample size is limited and thus the results should

be interpreted with some caution.

The event studies in Figure 7 show that net operating income and total operating revenue increase at the

time of the acquisition, an effect that persists for at least four years (Panels A and B). This appears to reflect

increases in both aeronautical (fees to airlines) and non-aeronautical (retail and parking) revenue (Panels C

and D). We do not see any apparent effect on operating expenditures in Panel E. Finally, we consider the

number of airport employees relative to the number of passengers and see some evidence in a decline in

years zero to three.

The average effects, presented in Table 8, are consistent with Figure 7. Net income increases

dramatically privatization by PE; the coefficient implies a 107.5% increase (column 1, row 1). This appears

to primarily come in part from an increase in operating revenue (column 2), with non-aeronautical revenue

increasing by somewhat more than aeronautical revenue (columns 3-4, row 1). The higher net income does

not reflect lower operating costs, because in fact we see an increase in operating costs and no effect on

employees per 1,000 passengers (column 5-6, row 1). After non-PE privatization, there is a smaller

increase in net income, but no increase in revenue (Table 8 columns 1-2, row 2). Instead, columns 5-6 show

a decline in expenditure and in employees per 1,000 passengers. After transitions from non-PE to PE, we

do not see any measurable effects. However, in transitions from PE to non-PE we see large declines in net

income and revenue, as well as expenditure (bottom row).

Together with our previous results, this analysis suggests that in the context of airports, PE does not

create value primarily by cost-cutting, but rather by increasing growth and efficiency. Perhaps surprisingly,

28
non-PE privatization is associated with lower expenditure as the primary means to higher income. Of course,

as emphasized above, airports and infrastructure generally have many unique characteristics and so the

effects of PE here likely do not necessarily generalize to other sectors.

4.7 Ownership vs. Control

An important distinction in our data is between sales and concessions; in the former, the acquirer has an

incentive to operate airports with a longer-term view. Following industry norms, we identify sales as cases

when the acquirer obtains outright ownership or a concession lasting more than 30 years. Long-term

concessions are grouped together with outright sales because they create similar incentive structures. A

longer lease period may incentivize the concessionaire to undertake capital investment and improve

operating performance. The incentives bestowed by ownership (or ownership-like long-term contracts)

cannot be fully contracted, as established by the Grossman/Hart/Moore literature and summarized in Hart

(1995b). In contrast to a non-owning operator, an owner will have both the ability and the incentives to

make investments that were not contracted on initially but could be valuable. The prediction for our sample

of privatized airports is that if ownership leads to more value-creating investments than control without

ownership, then sales should experience larger improvements than concessions.

We evaluate this hypothesis in Table 9, which employs a new set of four independent variables

representing the transaction type: sales to PE, concessions to PE, sales to non-PE private, and concessions

to non-PE private. The eight outcome variables represent the key performance measures from the previous

analyses. Across the board, sales to PE are associated with the largest and most robust improvements (first

row). In most cases, the coefficient on sale to PE is significantly larger than the coefficient on concession to

PE, as shown at the bottom of the table. For non-PE private transactions, most outcomes also exhibit a

larger effect for sales, but the magnitude of the difference is smaller and usually insignificant.

We assess the robustness of this finding and examine whether it comes from the extensive or intensive

margin in Appendix Tables A.5 and A.6. Table A.5 uses continuous ownership and control stakes, and

continues to find that higher ownership stake is associated with the most positive results. Table A.6 uses

dummies for majority ownership and control stake. These results suggest that majority control is not as

important as the result being a sale vs. a concession. Overall, this analysis supports the idea that an

ownership change to a PE fund improves value more than concessions or sales to a non-PE group.

29
4.8 Competition

The presence of a competing airport may encourage performance improvements in the interest of gaining

market share. This relates to the increased scrutiny that PE is receiving from antitrust authorities, who are

concerned that PE may take more advantage of market power.36 PE ownership may be more responsive to

competitive incentives than other types of ownership because it entails higher quality managers and higher-

powered incentives to maximize profits.

We evaluate this in Table 10 by dividing PE and non-PE ownership according to whether or not there is

a competing airport nearby, which we define as within 200 km.37 This again gives us four independent

variables. Across all the main outcome variables, we see that improvements are notably and usually

statistically significantly larger in the presence of a competing airport for both PE and non-PE private

ownership. For example, under PE ownership with competition, the number of flights and routes increase

by 49% and 42%, respectively, while the number of airlines increases by 50% relative to the mean,

compared to small and insignificant effects of PE ownership without competition (columns 3-5, first two

rows). Of course, this does not mean that competition causes the different operational changes since the

competing airport was typically in place before the acquisition and thus is intertwined with targeting.

However, these results are consistent with competition playing a beneficial role in the airport industry for

inducing the most productive outcomes under private ownership.

4.9 Robustness

We test the robustness of our results using several alternative estimation approaches and modeling

specifications. First, as explained in Section 4.1, we conduce three tests to address possible bias from a

staggered differences-in-differences model. In all cases, we employ for parsimony 10 key outcome

variables that contain the key findings in the above analysis. Our first approach retains the main empirical

model of Equation 1, but employs stacked datasets for each event (i.e. treatment year). Following (Baker et

al., 2022), we include fixed effects for each dataset-by-airport and dataset-by-time group. The results are

reported in Appendix Table A.7. The results are all robust to this approach, generally with more statistical

significance.
36
For example, see here: https://www.jdsupra.com/legalnews/private-equity-subject-to-increased-4754204/.
37
Overall, 37% of total airports have a competing airport nearby. Among PE owned and non-PE privately owned airports, 53%
and 48% of them, respectively, have a competing airport nearby.

30
 The second approach is the Callaway and Sant’Anna (2021) estimator, which has the downside of

permitting only one treatment variable. Therefore, the results cannot be compared exactly to our main

tables. We focus on an effect of PE acquisition on average, since this is where our average effects and

dynamic models suggest there is a meaningful causal effect. The results are reported in Appendix Table

A.8. We see significant effects on the key productivity outcomes, such as passengers per flight (column 1)

and number of routes (column 4). There is no significant effect on fees, consistent with the mixed results in

the main tables. Possibly due to the much smaller sample and stringent fixed effects design of the

estimator, we see no effect on income or expenditure in this model (columns 9-10).

The third strategy is a matching estimator. We match each privatized airport one-to-one with never-

privatized, government-owned and operated non-target airports using Coarsened Exact Matching (Iacus et

al. (2012)). To identify control airports, we employ observations two years before the privatization event.

Each targeted airport is matched to government owned and operated control airport on region, the log GDP

per capita, share of international passengers, government size, open markets, year, and log trade volume

two years before the target date.38 In the estimation, we include match cohort-year fixed effects to compare

target and non-target airports within the matched group. These fixed effects ensure we compare airports

with similar characteristics, since for example the airports targeted by PE have different characteristics from

the airports targeted by non-PE private companies. The matched dataset includes 684 airports. Of these, PE

acquires 90 and non-PE private acquires 324.

In Appendix Table A.9, we present estimates using the matched sample. The results are similar to the

main findings. One downside of matching is that it eliminates many airports from the sample, which makes

it more different to run heterogeneity tests. Our main analysis therefore uses the whole data with controls.

Also, while matching improves homogeneity of the sample, we should keep in mind that privatized airports

are may be chosen precisely because their performance is likely to improve relative to other airports; for

example, they could be in a growing section of the country or have been selected by a prominent airline as

a “hub”. We keep this possibility in mind when interpreting our results.

Together, the consistent results from all three of these diverse methods offer comforting support for the

basic findings. In unreported tests we find similar results using our main models with alternative controls,
38
The matched control airports need to be in the same country and have a similar passenger type, which is why we include the
proportion of international passengers. The number of passengers is the best proxy for size. We follow standard practice and do not
match on outcome variables, and for this reason do not match on regulatory structure because we find evidence it appears to change
after PE buyouts. However, we find similar results when we do match on regulatory regime. We also find similar results when we
match on alternative governance variables.

31
deal fixed effects, and alternative clustering of standard errors.

5 Conclusion

Whether infrastructure should be privately owned—and if so, by whom—is an important policy question

facing governments around the world. In practice, conventionally government-owned infrastructure assets

have increasingly been privatized. One driver of this trend is the growing amount of capital allocated to PE

infrastructure funds. Understanding what these funds do, whether they create or destroy value, and whether

public infrastructure should be privatized at all, are important research questions. To begin to address

these issues, we examine airports, which are an important class of infrastructure asset that have undergone

significant privatization in recent decades. As of 2020, 437 airports have been privatized, which is 18%

of all airports worldwide. Of these privatizations, 102 were at least once owned by PE firms. Because of

their visibility, their increasing rate of privatization, and the availability of data on their operations, airports

provide an ideal place to study privatization of public assets.

Our results suggest that privatizations, especially by PE funds, do well both for their investors and for

the general public. They increase the fees they charge airlines but despite these higher fees, also increase

the number of passengers flying through them. They appear to accomplish this by providing better service,

offering passengers nonstop flights to more places, lower cancellation rates, and better amenities inside the

airports. Our results suggest that PE ownership increases productivity more than non-PE private without

significantly different effects on prices (fees), suggesting overall benefits relative to non-PE ownership.

Privatization of infrastructure and the role of PE in such privatization is clearly an important topic of

research. This paper provides evidence suggesting that PE plays a beneficial role in the privatization of

airports. However, there is much more to be learned. For example, do other types of infrastructure achieve

similar improvements to airports when they are privatized? From an investor’s perspective, how do the

financial improvements observed in our sample of airports translate to risk-adjusted returns? These and

other related questions remain important topics for future research.

32
References
Acharya, Viral V, Cecilia Parlatore, and Suresh M Sundaresan, “A model of infrastructure financing,”
NYU Stern School of Business Forthcoming, 2020.
, Oliver F Gottschalg, Moritz Hahn, and Conor Kehoe, “Corporate governance and value creation:
Evidence from private equity,” The Review of Financial Studies, 2013, 26 (2), 368–402.
Adler, Nicole and Vanessa Liebert, “Joint impact of competition, ownership form and economic regulation
on airport performance and pricing,” Transportation Research Part A: Policy and Practice, 2014, 64, 92–
109.
Aguirre, Julio, Pedro Mateu, and Chrissie Pantoja, “Granting airport concessions for regional
development: Evidence from Peru,” Transport Policy, 2019, 74, 138–152.
Andonov, Aleksandar, Roman Kräussl, and Joshua Rauh, “Institutional investors and infrastructure
investing,” The Review of Financial Studies, 2021, 34 (8), 3880–3934.
Assaf, A Georges and David Gillen, “Measuring the joint impact of governance form and economic
regulation on airport efficiency,” European journal of operational research, 2012, 220 (1), 187–198.
Baker, Andrew C, David F Larcker, and Charles CY Wang, “How much should we trust staggered
difference-in-differences estimates?,” Journal of Financial Economics, 2022, 144 (2), 370–395.
Basso, Leonardo J, “Airport deregulation: Effects on pricing and capacity,” International Journal of
Industrial Organization, 2008, 26 (4), 1015–1031.
Bel, Germà and Xavier Fageda, “Privatization, regulation and airport pricing: an empirical analysis for
Europe,” Journal of Regulatory Economics, 2010, 37 (2), 142–161.
Bellon, Aymeric, “Does private equity ownership make firms cleaner? the role of environmental liability
risks,” 2020.
Bernstein, Shai and Albert Sheen, “The operational consequences of private equity buyouts: Evidence
from the restaurant industry,” The Review of Financial Studies, 2016, 29 (9), 2387–2418.
, Xavier Giroud, and Richard R Townsend, “The impact of venture capital monitoring,” The Journal
of Finance, 2016, 71 (4), 1591–1622.
Biais, Bruno and Enrico Perotti, “Machiavellian privatization,” American Economic Review, 2002, 92 (1),
240–258.
Bloom, Nicholas, Raffaella Sadun, and John Van Reenen, “Do private equity owned firms have better
management practices?,” American Economic Review, 2015, 105 (5), 442–46.
Borenstein, Severin and Nancy L Rose, “How airline markets work. . . or do they? Regulatory reform
in the airline industry,” in “Economic regulation and its reform: What have we learned?,” University of
Chicago Press, 2014, pp. 63–135.
Boucly, Quentin, David Sraer, and David Thesmar, “Growth LBOs,” Journal of Financial Economics,
2011, 102 (2), 432–453.
Brueckner, Jan K, “Airport congestion when carriers have market power,” American Economic Review,
2002, 92 (5), 1357–1375.
Cadot, Olivier, Lars-Hendrik Röller, and Andreas Stephan, “Contribution to productivity or pork barrel?
The two faces of infrastructure investment,” Journal of public Economics, 2006, 90 (6-7), 1133–1153.
Callaway, Brantly and Pedro HC Sant’Anna, “Difference-in-differences with multiple time periods,”
Journal of Econometrics, 2021, 225 (2), 200–230.
Cavagnaro, Daniel R, Berk A Sensoy, Yingdi Wang, and Michael S Weisbach, “Measuring institutional
investors’ skill at making private equity investments,” The Journal of Finance, 2019, 74 (6), 3089–3134.
Cengiz, Doruk, Arindrajit Dube, Attila Lindner, and Ben Zipperer, “The effect of minimum wages on
low-wage jobs,” The Quarterly Journal of Economics, 2019, 134 (3), 1405–1454.
Chambers, David, Christophe Spaenjers, and Eva Steiner, “The rate of return on real estate: Long-run
micro-level evidence,” The Review of Financial Studies, 2021, 34 (8), 3572–3607.

33
Coase, Ronald H, “The lighthouse in economics,” The journal of law and economics, 1974, 17 (2), 357–
376.
Cohn, Jonathan, Nicole Nestoriak, and Malcolm Wardlaw, “Private Equity Buyouts and Workplace
Safety,” Review of Financial Studies, 2021.
Condie, Stuart, “Airport ownership trends in Europe,” Journal of Airport Management, 2015, 10 (1), 14–
23.
Cruz, Carlos Oliveira and Joaquim Miranda Sarmento, “Airport privatization with public finances under
stress: An analysis of government and investor’s motivations,” Journal of Air Transport Management,
2017, 62, 197–203.
Dastidar, Siddhartha G, Raymond Fisman, and Tarun Khanna, “Testing limits to policy reversal:
Evidence from Indian privatizations,” Journal of Financial Economics, 2008, 89 (3), 513–526.
Davis, Steven J, John Haltiwanger, Kyle Handley, Ron Jarmin, Josh Lerner, and Javier Miranda,
“Private equity, jobs, and productivity,” American Economic Review, 2014, 104 (12), 3956–90.
Dender, Kurt Van, “Determinants of fares and operating revenues at US airports,” Journal of Urban
Economics, 2007, 62 (2), 317–336.
Dinc, I Serdar and Nandini Gupta, “The decision to privatize: Finance and politics,” The Journal of
Finance, 2011, 66 (1), 241–269.
Donaldson, Dave, “Railroads of the Raj: Estimating the impact of transportation infrastructure,” American
Economic Review, 2018, 108 (4-5), 899–934.
Dray, Lynnette, “An empirical analysis of airport capacity expansion,” Journal of Air Transport
Management, 2020, 87, 101850.
Eaton, Charlie, Sabrina T Howell, and Constantine Yannelis, “When investor incentives and consumer
interests diverge: Private equity in higher education,” The Review of Financial Studies, 2020, 33 (9),
4024–4060.
Esfahani, Hadi Salehi and Marıa Teresa Ramırez, “Institutions, infrastructure, and economic growth,”
Journal of development Economics, 2003, 70 (2), 443–477.
Ewens, Michael, Arpit Gupta, and Sabrina T Howell, “Local journalism under private equity ownership,”
Technical Report, National Bureau of Economic Research 2022.
Fang, Lily, Jim Goldman, and Alexandra Roulet, “Private Equity and Pay Gaps inside the Firm,” 2021.
Forsyth, Peter, “Airport policy in Australia and New Zealand: Privatization, light-handed regulation and
performance,” Aviation infrastructure performance: A study in comparative political economy, 2008,
pp. 65–99.
Fracassi, Cesare, Alessandro Previtero, and Albert W Sheen, “Barbarians at the store? Private equity,
products, and consumers,” Technical Report, National Bureau of Economic Research 2020.
Franzoni, Francesco A, Eric Nowak, and Ludovic Phalippou, “Private equity performance and liquidity
risk,” Journal of Finance, Forthcoming, Swiss Finance Institute Research Paper, 2011, (09-43).
Gadiesh, Orit and Hugh MacArthur, Lessons from private equity any company can use, Harvard Business
Press, 2008.
Gao, Janet, Merih Sevilir, and Yong Seok Kim, “Private equity in the hospital industry,” Available at
SSRN 3924517, 2021.
Gelhausen, Marc C, Peter Berster, and Dieter Wilken, “Do airport capacity constraints have a serious
impact on the future development of air traffic?,” Journal of Air Transport Management, 2013, 28, 3–13.
Gillen, David, “The evolution of airport ownership and governance,” Journal of Air Transport Management,
2011, 17 (1), 3–13.
and Hans-Martin Niemeier, “The European Union: evolution of privatization, regulation, and slot
reform,” 2008.
Gompers, Paul A and Steven N Kaplan, Advanced Introduction to Private Equity, Edward Elgar
Publishing, 2022.

34
Gompers, Paul, Steven N Kaplan, and Vladimir Mukharlyamov, “What do private equity firms say they
do?,” Journal of Financial Economics, 2016, 121 (3), 449–476.
Gormley, Todd A and David A Matsa, “Growing out of trouble? Corporate responses to liability risk,”
The Review of Financial Studies, 2011, 24 (8), 2781–2821.
Gornall, Will, Oleg Gredil, Sabrina T Howell, and Xing Liu, “Do Employees Cheer for Private Equity?
The Heterogeneous Effects of Buyouts on Job Quality,” The Heterogeneous Effects of Buyouts on Job
Quality (August 23, 2021), 2021.
Graham, Anne, “Airport privatisation: A successful journey?,” Journal of air transport management, 2020,
89, 101930.
Gramlich, Edward M, “Infrastructure investment: A review essay,” Journal of economic literature, 1994,
32 (3), 1176–1196.
Guo, Shourun, Edith S Hotchkiss, and Weihong Song, “Do buyouts (still) create value?,” The Journal of
Finance, 2011, 66 (2), 479–517.
Gupta, Arpit and Stijn Van Nieuwerburgh, “Valuing private equity investments strip by strip,” The
Journal of Finance, 2021, 76 (6), 3255–3307.
Gupta, Atul, Sabrina T Howell, Constantine Yannelis, and Abhinav Gupta, “Does private equity
investment in healthcare benefit patients? Evidence from nursing homes,” Technical Report, National
Bureau of Economic Research 2021.
Gutiérrez, Ester and Sebastián Lozano, “Efficiency assessment and output maximization possibilities of
European small and medium sized airports,” Research in Transportation Economics, 2016, 56, 3–14.
Harris, Robert S, Tim Jenkinson, and Steven N Kaplan, “Private equity performance: What do we
know?,” The Journal of Finance, 2014, 69 (5), 1851–1882.
Hart, Oliver, “Corporate governance: some theory and implications,” The economic journal, 1995, 105
(430), 678–689.
, Firms, contracts, and financial structure, Clarendon press, 1995.
Hoffman, David E, The oligarchs: Wealth and power in the new Russia, Hachette UK, 2011.
Huntington-Klein, Nick, The effect: An introduction to research design and causality, Chapman and
Hall/CRC, 2021.
Iacus, Stefano M, Gary King, and Giuseppe Porro, “Causal inference without balance checking:
Coarsened exact matching,” Political analysis, 2012, 20 (1), 1–24.
Ivashina, Victoria and Anna Kovner, “The private equity advantage: Leveraged buyout firms and
relationship banking,” The Review of Financial Studies, 2011, 24 (7), 2462–2498.
Jenkinson, Tim, Hyeik Kim, and Michael S Weisbach, Buyouts: A Primer, Vol. 1 of Handbook of the
Economics of Corporate Finance: Private Equity and Entrepreneurial Finance, Elsevier, 2021.
Kaplan, Steven, “The effects of management buyouts on operating performance and value,” Journal of
financial economics, 1989, 24 (2), 217–254.
Kaplan, Steven N and Antoinette Schoar, “Private equity performance: Returns, persistence, and capital
flows,” The journal of finance, 2005, 60 (4), 1791–1823.
and Per Stromberg, “Leveraged buyouts and private equity,” Journal of economic perspectives, 2009,
23 (1), 121–46.
Korteweg, Arthur and Morten Sorensen, “Skill and luck in private equity performance,” Journal of
Financial Economics, 2017, 124 (3), 535–562.
Lerner, Josh and Antoinette Schoar, “Does legal enforcement affect financial transactions? The
contractual channel in private equity,” The Quarterly Journal of Economics, 2005, 120 (1), 223–246.
Liu, Tong, “Bargaining with private equity: implications for hospital prices and patient welfare,” Available
at SSRN 3896410, 2021.
Mayer, Christopher and Todd Sinai, “Network effects, congestion externalities, and air traffic delays: Or
why not all delays are evil,” American Economic Review, 2003, 93 (4), 1194–1215.

35
Megginson, William L, Robert C Nash, Jeffry M Netter, and Annette B Poulsen, “The choice of private
versus public capital markets: Evidence from privatizations,” The Journal of Finance, 2004, 59 (6), 2835–
2870.
Milesi-Ferretti, Gian Maria, Roberto Perotti, and Massimo Rostagno, “Electoral systems and public
spending,” The quarterly journal of economics, 2002, 117 (2), 609–657.
Mill, John Stuart and JM Robson, “The collected works of John Stuart Mill, volume III,” Principles of
political economy, part II, 1965.
Nieuwerburgh, Stijn Van, Richard Stanton, and Leo de Bever, “A review of real estate and infrastructure
investments by the Norwegian Government Pension Fund Global (GPFG),” Report to the Norwegian
Ministry of Finance, 2015.
Olariaga, Oscar Díaz and Luis Pulido Moreno, “Measurement of airport efficiency. The case of
Colombia,” Transport and Telecommunication, 2019, 20 (1), 40–51.
Olsson, Martin and Joacim Tåg, “Private Equity, Layoffs, and Job Polarization,” Journal of Labor
Economics, 2017, 35 (3), 697–754.
Oum, Tae H, Jia Yan, and Chunyan Yu, “Ownership forms matter for airport efficiency: A stochastic
frontier investigation of worldwide airports,” Journal of urban economics, 2008, 64 (2), 422–435.
, Nicole Adler, and Chunyan Yu, “Privatization, corporatization, ownership forms and their effects on
the performance of the world’s major airports,” Journal of air Transport management, 2006, 12 (3), 109–
121.
Phalippou, Ludovic, “An inconvenient fact: Private equity returns and the billionaire factory,” The Journal
of Investing, 2020, 30 (1), 11–39.
Pigou, Arthur C, “The Economics of Welfare. 1920. Rev. ed,” 1938.
Porta, Rafael La and Florencio Lopez de Silanes, “The benefits of privatization: Evidence from Mexico,”
The quarterly journal of economics, 1999, 114 (4), 1193–1242.
Robinson, David and Berk Sensoy, “Cyclicality, performance measurement, and cash flow liquidity in
private equity,” Journal of Financial Economics, 2016, 122 (3), 521–543.
Samuelson, Paul A, “Economics—an introductory analysis McGraw-Hill Book Co,” New York, 1964.
Sensoy, Berk A, Yingdi Wang, and Michael S Weisbach, “Limited partner performance and the maturing
of the private equity industry,” Journal of Financial Economics, 2014, 112 (3), 320–343.
Starkie, David, “Testing the regulatory model: The expansion of Stansted airport,” Fiscal Studies, 2004, 25
(4), 389–413.

36
Figure 1: Airport Privatization over Time
Panel A shows the share of all airports and the number of airports owned by PE firms over time in our sample of airports. Panel
B shows ownership dynamics of airports. Panel C shows the share of total number of passengers in PE or NonPE private airports.
Pancel D shows the share of total number of flights in PE or NonPE private airports. We drop airports that were newly built during
2006-2019.

(A) Airport Ownership (1987–2019) (B) PE Ownership of Airports (1997–2019)

(C) PE and non-PE Private Passenger Shares (1997–2019) (D) PE and non-PE Private Flight Shares (1997–2019)

37
Figure 2: Airports That Experienced Privatization
This figure shows countries that experienced airport privatization by their types.

(A) Countries with Privatized Airports (As of 1984)

(B) Countries with Privatized Airports (As of 2019)

38
Figure 3: Event Studies – Privatization Effect on Airport Volume of Passengers and Traffic by PE
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Int’l per Flight Passenger Traffic (B) Dom per Flight Passenger Traffic

(C) Log(Number of Int’l Passengers) (D) Log(Number of Dom Passengers)

(E) Log(Number of Int’l Flights) (F) Log(Number of Dom Flights)

39
Figure 4: Event Studies – Effect of PE Ownership on Downstream Performance: Routes
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Number of International Routes (B) Number of Domestic Routes

(C) Number of Low Cost Carriers (D) Airline HHI

40
Figure 5: Event Studies – Effect of PE Ownership on Punctuality, Safety, and Awards
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Flight Cancellation Rate (B) On-Time Departure Rate

(C) Any Award (D) Number of Accidents

(E) Number of Fatalities

41
Figure 6: Event Studies – Effect of PE Ownership on Fees Charged to Airlines
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Log(Int’l Fee) (B) Log(Dom Fee)

(C) Log(Int’l Passenger Fee) (D) Log(Dom Passenger Fee)

(E) Log(Int’l Runway Fee) (F) Log(Dom Runway Fee)

42
Figure 7: Event Studies – Effect of PE Ownership on Airport Financials
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Log(Net Operating Income) (B) Log(Total Operating Revenue)

(C) Log(Total Aeronautical Revenue) (D) Log(Total Non Aeronautical Revenue)

(E) Log(Total Operating Expenditure) (F) Number of Employees per 1000 Passengers

43
Table 1: Ownership and Control Statistics about Airport Acquisitions
This table shows the distribution of the number of transactions by privatization form and the summary statistics on the ownership
and control stakes of the privatization type. Panel A shows the number of transactions and the summary statistics on control stake
and concession years of the concession deals by privatization form. Panel B shows the number of transactions and the summary
statistics on ownership stake and concession years of sales by privatization form. We consider a very long-term lease (concession
deals that are awarded to private parties to operate for more than 30 years) as a sale.
Panel A: Concessions

Number of Transactions Percent Control Stake Duration (Years)

N Mean Median SD Min Max Mean Median
Privatization
Total 186 94.2 100 15.37 34 100 23.04 25
Govt to non-PE Private 178 96.8 100 11.52 34 100 23.1 25
Govt to PE 8 55.73 45.26 14.05 45 80 21 15

Post-Privatization
Total 50 88.28 100 24.5 14 100 16.66 15
Non-PE Private to PE 36 85.21 100 25.72 14 100 17.13 17
PE to non-PE Private 11 99.24 100 1.24 97.26 100 12.75 15
PE to PE 3 65.36 63 15.31 49 80 20.82 23

Panel B: Sales

Number of Transactions Percent Ownership Stake Duration if

Lease(Years)
Total Outright Long- Mean Median SD Min Max Mean Median
Sales Term
Leases
Privatization
Total 251 129 122 86.37 100 18.98 30 100 51.1 50
Govt to non-PE Private 223 110 113 88.64 100 17.43 30 100 50.82 50
Govt to PE 28 19 9 71.88 66 22.09 36 100 55.11 45

Post-Privatization
Total 50 43 7 70.67 66 25.93 10 100 59.14 46
Non-PE Private to PE 35 28 7 72.06 68 25.92 10 100 59.14 46
PE to non-PE Private 7 7 0 67.85 51 32.04 28 100 NA NA
PE to PE 8 8 0 69.2 66 22.47 45 100 NA NA

44
Table 2: Summary Statistics
This table reports the summary statistics for the main variables used in our study at the airport-year level. We gather variables from
multiple data sources. Country-level economic characteristics are from World Bank, country-level governance characteristics are
from the Heritage Foundation, airport-level traffic is from The Official Aviation Guide (OAG), airport-level project stage is from
Preqin, airport-level regulation is provided by David Gillen from University of British Columbia, airport-level price charged on
airlines is from an aviation data vendor called, RDC, airport-level financials is from Embry-Riddle University, and airport-on-time
performance is from OAG. There are four data types we use in the analyses. Traffic data covers years 1996 to 2019, Charge data
covers years 2010 to 2020, Financials data covers 2001 to 2017, and On-time-Performance data covers from 2016-2020.

Gov’t Non-PE Private PE

N Mean Median SD N Mean Median SD N Mean Median SD
Economic Characteristics (Country-Level)
GDP per Capita (Th) 36,190 $ 25.48 $ 24.17 $ 21.74 3,272 $ 17.26 $ 10.20 $ 16.25 524 $ 34.02 $ 40.36 $ 17.43
Trade Volume (B) 36,190 $ 1,038.85 $ 427.83 $ 1,295.34 3,272 $ 426.67 $ 208.27 $ 495.42 524 $ 613.90 $ 619.23 $ 439.52
Governance Characteristics (Country-Level)
Gov’t Integrity 36,190 57.36 61.00 24.33 3,272 47.23 37.00 22.14 524 67.41 78.00 21.67
Judicial Effectiveness 36,190 62.90 65.40 19.03 3,272 53.00 46.53 20.17 524 71.61 85.90 21.52
Gov’t Spending 36,190 60.20 60.03 20.85 3,272 62.39 67.50 23.08 524 50.85 49.60 22.67
Investment Freedom 36,190 58.72 70.00 19.66 3,272 61.65 70.00 17.68 524 75.82 80.00 13.92
Financial Freedom 36,190 59.09 60.00 20.72 3,272 59.48 60.00 17.00 524 74.09 80.00 16.25
Open Markets 36,190 64.16 66.00 15.92 3,272 65.54 67.47 12.94 524 77.51 81.03 10.70
Traffic (Airport-Level)
Share Passengers Int’l 36,190 0.19 0.01 0.30 3,272 0.38 0.29 0.35 524 0.43 0.38 0.35
Total Passengers per Flight 36,190 85.71 76.91 53.81 3,272 117.54 120.09 41.44 524 126.92 133.22 35.77
Int’l Passengers per Flight 36,190 68.47 26.12 78.56 3,272 119.18 132.13 66.45 524 130.96 147.32 68.64
Dom Passengers per Flight 36,190 77.97 67.34 51.80 3,272 106.56 107.53 50.13 524 115.95 116.99 47.88
Total Passengers (M) 36,190 1.71 0.15 5.21 3,272 2.67 0.73 5.09 524 3.46 1.72 4.85
Int’l Passengers (M) 36,190 0.58 0.00 2.59 3,272 1.48 0.20 3.68 524 2.16 0.56 4.12
Dom Passengers (M) 36,190 1.14 0.10 3.81 3,272 1.20 0.38 2.29 524 1.30 0.62 1.89
Total Flights (Th) 36,190 13.57 2.56 37.19 3,272 19.28 7.02 32.00 524 25.65 14.75 32.31
Total Freight Tons per Flight 36,190 2.44 0.79 4.14 3,272 3.98 2.72 4.56 524 5.87 3.22 8.06
Int’l Freight Tons per Flight 36,190 2.36 0.00 5.16 3,272 4.59 2.83 5.60 524 6.43 3.88 8.24
Dom Freight Tons per Flight 36,190 2.00 0.50 3.57 3,272 3.10 1.45 4.31 524 5.19 2.21 8.94
Total Freight Tons (Th) 36,190 57.28 1.82 243.07 3,272 97.11 16.51 219.46 524 185.13 29.43 441.56
Int’l Freight Tons (Th) 36,190 24.83 0.00 144.96 3,272 57.10 4.28 144.72 524 128.78 11.57 372.23
Dom Freight Tons (Th) 36,190 32.45 0.93 160.97 3,272 40.01 4.48 110.61 524 56.35 12.80 143.97
Quality Measures (Airport-Level)
Number of Routes 36,190 18.17 6.00 32.91 3,272 35.67 17.00 47.26 524 49.71 30.00 52.74
Number of Int’l Routes 36,190 9.01 1.00 23.70 3,272 25.71 9.00 42.80 524 38.32 17.00 51.61
Number of Dom Routes 36,190 9.16 4.00 17.21 3,272 9.96 6.00 11.50 524 11.39 9.00 9.31
I(Award) 36,190 0.02 0.00 0.12 3,272 0.04 0.00 0.21 524 0.07 0.00 0.26
Number of Accidents 36,190 0.05 0.00 0.24 3,272 0.08 0.00 0.30 524 0.11 0.00 0.35
Number of Fatalities 36,190 0.44 0.00 15.55 3,272 0.49 0.00 6.45 524 0.08 0.00 0.64
Airlines (Airport-Level)
Number of Operated Airlines 36,190 7.70 3.00 11.92 3,272 15.72 8.00 18.80 524 17.25 11.00 18.63
Number of Low Cost Carriers 36,190 0.86 0.00 2.03 3,272 2.27 1.00 3.11 524 3.17 2.00 3.47
Airline HHI 36,190 5,727.97 5,063.13 3,045.26 3,272 3,880.68 3,212.78 2,528.96 524 3,759.94 3,131.45 2,350.43
Share of Largest Airline 36,190 66.48 63.14 25.92 3,272 51.32 47.94 23.56 524 51.00 47.32 22.12
Price Regulation (Airport-Level)
No Regulation 777 0.08 0.00 0.27 371 0.22 0.00 0.42 197 0.32 0.00 0.47
Cost Based 777 0.52 1.00 0.50 371 0.15 0.00 0.36 197 0.02 0.00 0.12
Revenue Cap 777 0.06 0.00 0.23 371 0.11 0.00 0.32 197 0.03 0.00 0.16
Hybrid 777 0.11 0.00 0.31 371 0.10 0.00 0.30 197 0.10 0.00 0.30
Fees Charged to Airlines (Airport-Level)
Int’l Fee (Th) 7,566 $ 7.77 $ 6.46 $ 5.31 1,465 $ 11.01 $ 10.41 $ 5.37 339 $ 10.75 $ 9.54 $ 5.93
Dom Fee (Th) 7,566 $ 0.77 $ 0.51 $ 0.76 1,465 $ 0.89 $ 0.72 $ 0.78 339 $ 1.36 $ 1.18 $ 1.05
Int’l Passenger Fee (Th) 7,566 $ 5.22 $ 4.21 $ 4.68 1,465 $ 8.47 $ 8.01 $ 4.97 339 $ 7.81 $ 6.36 $ 4.86
Dom Passenger Fee (Th) 7,566 $ 0.61 $ 0.34 $ 0.69 1,465 $ 0.74 $ 0.58 $ 0.68 339 $ 1.08 $ 0.92 $ 0.91
Int’l Runway Fee (Th) 7,566 $ 2.56 $ 2.01 $ 2.02 1,465 $ 2.54 $ 2.12 $ 1.82 339 $ 2.94 $ 2.01 $ 2.77
Dom Runway Fee (Th) 7,566 $ 0.17 $ 0.12 $ 0.19 1,465 $ 0.15 $ 0.08 $ 0.19 339 $ 0.28 $ 0.18 $ 0.30
Financials (Airport-Level)
Total Op. Rev (M) 2,005 $ 317.74 $ 171.61 $ 392.54 363 $ 382.32 $ 180.10 $ 455.40 204 $ 219.72 $ 100.95 $ 261.31
Total Op. Exp. (M) 1,982 $ 189.35 $ 101.84 $ 240.14 361 $ 227.33 $ 99.96 $ 337.22 204 $ 118.74 $ 55.25 $ 155.92
Net Op. Income (M) 1,981 $ 130.42 $ 65.71 $ 191.57 359 $ 158.95 $ 81.20 $ 166.53 204 $ 100.98 $ 53.24 $ 117.49
Total Aero Rev (M) 2,033 $ 162.99 $ 88.20 $ 197.73 364 $ 182.60 $ 90.67 $ 184.59 204 $ 107.04 $ 55.23 $ 126.79
Total Non-Aero Rev (M) 2,006 $ 155.75 $ 78.47 $ 230.02 364 $ 199.17 $ 93.87 $ 298.40 204 $ 112.26 $ 50.85 $ 138.22
Num of Employees per 1000 psg 2,015 50.72 0.06 1,791.09 369 2,346.08 0.09 20,540.33 205 1,346.17 0.05 19,273.17
On-Time Performance (Airport-Level)
Cancellation (%) 3,011 2.16 1.44 3.92 673 1.56 1.01 2.27 142 1.26 0.93 1.14
Departure (%) 3,011 77.89 80.00 9.73 673 77.21 78.80 8.40 142 72.80 73.51 10.25

45
Table 3: Economic and Governance Predictors of Privatization

This table shows predictors of acquisition by either a PE infrastructure fund (“PE”) or a non-PE private firm (“Non-PE Priv"). The
unit of observation is the airport-year, with years post-acquisition dropped. The model is OLS regression. The dependent variables
are indicators multiplied by 100 for readability of the coefficients. Regulation is only observed for major airports in Asia, Europe,
and Oceana. Regulation variables are indicator variables. No regulation is 1 if there are no price regulation in an airport at that
year, Cost Based is 1 if The governance measures are from The Heritage Foundation, which also provides indices for "tax burden",
"monetary freedom", "labor freedom," and "business freedom," but we find no significant effects of these and do not include them.
All models include year and region fixed effects. Regions are Africa, Asia, Europe, North America, Oceana, and South America.
Standard errors are clustered by airport. Significance: *p<.1 **p<.05 ***p<.01

Economic Regulation Governance Combined

PE Non-PE Priv PE Non-PE Priv PE Non-PE Priv PE Non-PE Priv
(1) (2) (3) (4) (5) (6) (7) (8)
Log Total Passengers 0.039 -0.126∗ 0.036 -0.112
(0.024) (0.071) (0.025) (0.070)
Share Passengers Intl 0.461∗∗∗ 1.269∗∗∗ 0.401∗∗∗ 1.310∗∗∗
(0.128) (0.457) (0.128) (0.481)
Log GDP Per Capita 0.004 -0.181∗∗ -0.022 0.251∗∗
(0.016) (0.076) (0.030) (0.104)
Log Trade Volume 0.052∗∗∗ -0.063∗ 0.033∗∗ -0.009
(0.016) (0.035) (0.015) (0.045)
1(Competing Airports) 0.057 0.379∗∗ 0.039 0.292
(0.056) (0.186) (0.056) (0.185)
Share of Largest Airline -0.001 -0.020 -0.001 -0.021∗
(0.006) (0.012) (0.006) (0.012)
Airline HHI 0.000 0.000∗∗ 0.000 0.000∗∗
(0.000) (0.000) (0.000) (0.000)
Log(Number of Airline) -0.045 0.994∗∗∗ -0.036 0.844∗∗∗
(0.102) (0.244) (0.102) (0.238)
Log(Number of Routes) 0.032 -0.196 0.042 -0.122
(0.053) (0.139) (0.053) (0.143)
I(Award) 0.308 -0.466 0.310 -0.472
(0.329) (0.475) (0.328) (0.483)
No Regulation -0.165 1.447
(2.420) (1.720)
Cost Based -1.962∗∗∗ -0.460
(0.653) (0.820)
Revenue Cap -0.833 1.238
(1.059) (2.065)
Hybrid 0.865 0.256
(1.457) (1.357)
Judicial Effectiveness 0.002 -0.018∗∗ 0.002 -0.012
(0.002) (0.009) (0.002) (0.011)
Financial Freedom 0.010∗∗∗ 0.027∗∗∗ 0.009∗∗∗ 0.027∗∗∗
(0.003) (0.009) (0.003) (0.010)
Investment Freedom 0.007∗∗ 0.017∗∗ 0.007∗∗ 0.019∗∗
(0.003) (0.009) (0.003) (0.009)
Government Integrity -0.004∗∗ -0.015∗∗ -0.002 -0.021∗∗∗
(0.002) (0.006) (0.002) (0.007)
Gov’t Spending -0.003∗ 0.018∗∗ -0.003 0.021∗∗∗
(0.002) (0.007) (0.002) (0.008)
Open Markets -0.010 -0.034∗∗ -0.012 -0.045∗∗
(0.007) (0.016) (0.007) (0.020)
Constant -1.090∗∗∗ 3.171∗∗∗ 2.283∗∗∗ 1.559∗∗ 0.125 1.233∗∗ -0.567 -0.786
(0.343) (0.898) (0.828) (0.640) (0.194) (0.479) (0.378) (0.973)
Observations 41008 37974 1474 1240 42379 39334 41008 37974
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Region FE Yes Yes Yes Yes Yes Yes Yes Yes
Outcome Mean 0.172 0.763 0.172 0.763 0.172 0.763 0.172 0.763

46
Table 4: Effect of Ownership Type on Airport Volume of Passengers and Traffic
This table reports estimates of how four ownership transitions affect passengers and traffic. The sample is an airport-year level
panel from 1996 to 2019. Columns 1, 2, and 3 report changes in per flight passengers, per flight international passengers, and per
flight domestic passengers. Columns 4 and 5 report changes in the share of aircraft types. Jets are aircrafts that can take more than
150 passengers on board. Regional includes smaller aircraft with capacity less than 100 passengers. Columns 7, 8, and 9 report
changes in total number of passengers, number of international passengers, number of domestic passengers. Column 10, 11, and 12
report changes in total, international, and domestic number of air movements. The independent variables capture four ownership
type changes, with government ownership as the base group. Privatization by PE is one after an airport transitions from government
to PE ownership and zero otherwise. Similarly, Privatization by Non-PE is one after an airport transitions from government to
Non-PE private ownership. Post-Priv Non-PE to PE is one after an airport that is already privatized by a non-PE firm transitions
to PE ownership. Post-Priv PE to Non-PE is the reverse. We include the log GDP per capita, share of international passengers,
government size, open markets, and log trade volume as control variables, and include airport, and year fixed effects. We report
two p-values on F-tests for equality of coefficients. The first compares the two privatization coefficients and the second compares
the post-privatization coefficients. Standard errors are clustered at an airport level. Significance: *p<.1 **p<.05 ***p<.01

Dependent Variable: Passengers per Flight Share Aircraft Type Log(Number of Passengers) Log(Number of Flights)
Total In’l Domestic Jets Regional Total In’l Domestic Total In’l Domestic
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
1(Privatization by PE) 17.87∗∗∗ 11.01 20.17∗ 0.14∗∗ -0.14∗∗ 0.61∗∗∗ 0.32 0.71∗∗∗ 0.48∗∗∗ 0.26 0.49∗∗∗
(5.56) (7.46) (10.60) (0.06) (0.06) (0.18) (0.25) (0.26) (0.17) (0.25) (0.15)
1(Privatization by Non-PE) 1.60 0.98 3.68∗∗ 0.00 -0.00 0.17∗∗∗ 0.13∗∗∗ 0.12∗∗ 0.18∗∗∗ 0.12∗∗ 0.11∗∗
(1.45) (2.24) (1.77) (0.02) (0.02) (0.04) (0.05) (0.05) (0.04) (0.05) (0.05)
1(Post-Priv Non-PE to PE) 8.72∗∗∗ 7.55∗ 19.67∗∗∗ 0.09∗∗∗ -0.09∗∗∗ 0.16∗∗ 0.18∗∗ 0.23∗∗ 0.12∗ 0.12 0.08
(2.94) (4.19) (6.58) (0.02) (0.02) (0.07) (0.09) (0.10) (0.07) (0.08) (0.12)
1(Post-Priv PE to Non-PE) 2.69 7.64 -9.84 0.06∗ -0.06∗ 0.02 -0.17 -0.14 -0.01 -0.25 -0.02
(7.40) (10.05) (6.20) (0.03) (0.03) (0.17) (0.20) (0.28) (0.16) (0.20) (0.22)
Observations 40357 40357 40357 40357 40357 40357 40357 40357 40357 40357 40357
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.93 0.87 0.87 0.87 0.96 0.99 0.96 0.94 0.98 0.93
Y-Mean 90.00 76.07 81.42 0.41 0.59 12.46 6.70 11.82 8.21 3.99 7.72
Pr > F Priv Non-PE=Priv PE 0 .2 .12 .04 .04 .02 .45 .03 .08 .57 .02
Pr > F Non-PE to PE=PE to Non-PE .45 .99 0 .53 .53 .42 .11 .21 .46 .08 .69

47
Table 5: Effect of Ownership Type on Downstream Performance: Routes and Airlines
This table reports estimates of how four ownership transitions affect airports’ route systems and airlines. The sample is an airport-
year level panel from 1996 to 2019. Columns 1 to 3 report results on the effect of ownership changes on log number of routes that
are being served in the airport. Column 4 and 5 report results on the effect of ownership changes on the number of airlines being
served in the airport. Column 6 and 7 report results on the effect of the airline HHI measure and the share of the largest airline
in the airport. The independent variables capture four ownership type changes, with government ownership as the base group.
Privatization by PE is one after an airport transitions from government to PE ownership and zero otherwise. Similarly, Privatization
by Non-PE is one after an airport transitions from government to Non-PE private ownership. Post-Priv Non-PE to PE is one after an
airport that is already privatized by a non-PE firm transitions to PE ownership. Post-Priv PE to Non-PE is the reverse. We include
the log GDP per capita, share of international passengers, government size, open markets, and log trade volume as control variables,
and include airport, and year fixed effects. We report two p-values on F-tests for equality of coefficients. The first compares the
two privatization coefficients and the second compares the post-privatization coefficients. Standard errors are clustered at an airport
level. Significance: *p<.1 **p<.05 ***p<.01

Dependent Variable: Log(Number of Routes) Number of Airlines

Total International Domestic Total Low Cost Airline Share of
Carriers HHI Largest
Airline
(1) (2) (3) (4) (5) (6) (7)
1(Privatization by PE) 0.36∗∗ 0.38∗∗ 0.08 1.24 2.13∗∗ 291 2.73
(0.16) (0.17) (0.11) (1.13) (1.04) (353.22) (3.87)
1(Privatization by Non-PE) 0.14∗∗∗ 0.17∗∗∗ 0.03 3.05∗∗∗ 1.18∗∗∗ -245∗∗ -2.50∗∗
(0.03) (0.03) (0.03) (0.47) (0.18) (101.14) (1.02)
1(Post-Priv Non-PE to PE) 0.17∗∗∗ 0.27∗∗∗ 0.01 -0.32 0.20 154 0.78
(0.07) (0.08) (0.04) (0.78) (0.33) (225.09) (2.44)
1(Post-Priv PE to Non-PE) 0.23∗∗ 0.05 0.17 1.07 -0.35 -464 -3.89
(0.09) (0.14) (0.11) (1.59) (1.02) (393.33) (6.33)
Observations 40343 40357 40357 40357 40357 40357 40357
Airport FE Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes
R2 0.94 0.96 0.91 0.93 0.68 0.79 0.76
Y-Mean 2.30 1.26 1.78 8.70 1.01 5,474 64.44
Pr > F Priv Non-PE=Priv PE .17 .21 .67 .14 .37 .14 .19
Pr > F Non-PE to PE=PE to Non-PE .62 .17 .17 .43 .61 .17 .49

48
Table 6: Effect of Ownership Type on Punctuality, Safety, and Awards
This table reports estimates of how four ownership transitions affect other measures of airport performance, relative to government-
owned airports. The dependent variables of the first two columns are related to airport’s on-time performance and use a sample
from 2016 to 2019. The Flight cancellation rate=Number of flights cancelled/Total number of flights * 100. Similarly the On time
departure rate=Number of flights that are on-time/Total number of flights * 100. For the rest of the columns the sample is from
1996 to 2019. The dependent variable in column 3 is an indicator for winning an award. The last two columns report changes in the
number of accidents and fatalities per 1000 flights that took off from the airport. The independent variables capture four ownership
type changes, with government ownership as the base group. Privatization by PE is one after an airport transitions from government
to PE ownership and zero otherwise. Similarly, Privatization by Non-PE is one after an airport transitions from government to
Non-PE private ownership. Post-Priv Non-PE to PE is one after an airport that is already privatized by a non-PE firm transitions
to PE ownership. Post-Priv PE to Non-PE is the reverse. We include the log GDP per capita, share of international passengers,
government size, open markets, and log trade volume as control variables, and include airport, and year fixed effects. We report
two p-values on F-tests for equality of coefficients. The first compares the two privatization coefficients and the second compares
the post-privatization coefficients. Standard errors are clustered at an airport level. Significance: *p<.1 **p<.05 ***p<.01

Dependent Variable: Flight On-Time 1(Award) Number per 1000 Flights

Cancellation Departure
Rate Rate
Accidents Fatalities
(1) (2) (3) (4) (5)
1(Privatization by PE) -1.44∗∗∗ -8.62∗∗∗ 0.12 -0.01 -0.22
(0.12) (0.32) (0.08) (0.01) (0.24)
1(Privatization by Non-PE) -0.77∗∗ -2.33∗∗ 0.03∗∗ 0.00 -0.12
(0.37) (1.04) (0.01) (0.00) (0.09)
1(Post-Priv Non-PE to PE) -0.98∗∗∗ -1.68 0.06∗∗ -0.00 -0.16
(0.25) (1.94) (0.02) (0.00) (0.17)
1(Post-Priv PE to Non-PE) 0.92 -7.49∗∗∗ -0.02 0.00 0.13
(0.57) (1.78) (0.06) (0.01) (0.12)
Observations 3621 3621 40357 40357 40357
Airport FE Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes
R2 0.66 0.82 0.31 0.08 0.06
Y-Mean 2.03 77.53 0.02 0.01 0.08
Pr > F Priv Non-PE=Priv PE .09 0 .26 .48 .71
Pr > F Non-PE to PE=PE to Non-PE 0 .03 .21 .43 .15

49
Table 7: Effect of Ownership Type on Fees Charged to Airlines
This table reports estimates of how four ownership transitions affect fees charged to airlines, relative to government-owned airports.
The sample is an airport-year level panel from 2010 to 2020. Columns 1 and 2 report changes in total international and total domestic
charges. Columns 3 and 4 report changes in international passenger charges and domestic passenger charges. Passenger charges are
fees that are levied on airlines for airports’ services to passengers. Columns 5 and 6 report changes in international runway charges
and domestic runway charges where runway charges are defined as fees that are levied on airlines for using the airport’s runway
when flights land and take off. All charges are calculated per individual flight. Column 7 considers the government regulatory
regime. The sample is an airport-year level panel from 1990 to 2018. 1(No Regulation) is a dependent variable which indicates
whether an airport has no price regulations at that year. The independent variables capture four ownership type changes, with
government ownership as the base group. Privatization by PE is one after an airport transitions from government to PE ownership
and zero otherwise. Similarly, Privatization by Non-PE is one after an airport transitions from government to Non-PE private
ownership. Post-Priv Non-PE to PE is one after an airport that is already privatized by a non-PE firm transitions to PE ownership.
Post-Priv PE to Non-PE is the reverse. We include the log GDP per capita, share of international passengers, government size, open
markets, and log trade volume as control variables, and include airport, and year fixed effects. We report two p-values on F-tests
for equality of coefficients. The first compares the two privatization coefficients and the second compares the post-privatization
coefficients. Standard errors are clustered at an airport level. Statistical significance is indicated by: *p<.1 **p<.05 ***p<.01

Dependent Variable: Log(Total Fee) Log(Passenger Fee) Log(Runway Fee)

International Domestic International Domestic International Domestic No
Regulation
(1) (2) (3) (4) (5) (6) (7)
1(Privatization by PE) -0.02∗∗ -0.05∗∗∗ -0.06∗∗∗ -0.06∗∗∗ 0.13∗∗∗ 0.18∗∗∗ 0.14∗
(0.01) (0.01) (0.02) (0.02) (0.01) (0.01) (0.07)
1(Privatization by Non-PE) 0.28∗∗∗ 0.31∗∗∗ 0.55∗∗∗ 0.49∗∗∗ 0.05 0.11∗∗∗ -0.05
(0.08) (0.08) (0.15) (0.11) (0.03) (0.04) (0.06)
1(Post-Priv Non-PE to PE) 0.17∗∗ 0.18∗∗ 0.54∗∗ 0.43∗ 0.04 0.06 0.28∗∗
(0.07) (0.08) (0.27) (0.22) (0.05) (0.05) (0.11)
1(Post-Priv PE to Non-PE) 0.03 0.05 -0.03 0.04 0.07∗∗ 0.09∗
(0.07) (0.05) (0.09) (0.06) (0.03) (0.05)
Observations 9125 9123 9125 9125 9125 9115 1514
Airport FE Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes
R2 0.95 0.96 0.97 0.97 0.98 0.98 0.72
Y-Mean 8.82 6.23 8.01 5.64 7.54 4.55 0.16
Pr > F Priv Non-PE=Priv PE 0 0 0 0 .03 .05 .04
Pr > F Non-PE to PE=PE to Non-PE .19 .18 .05 .1 .59 .69

50
Table 8: Effect of Ownership Type on Financial Outcomes
This table reports estimates of how four ownership transitions affect financial outcomes, relative to government-owned airports.
The sample is an airport-year level panel from 2001 to 2017. The dependent variables are financial measures taken from airport
income statements. They are the logs of total net operational income, total operational revenue, total operational expenditure per
1000 passengers, aeronautical revenue, non-aeronautical revenue, and number of employees per 1000 passengers. The independent
variables capture four ownership type changes, with government ownership as the base group. Privatization by PE is one after an
airport transitions from government to PE ownership and zero otherwise. Similarly, Privatization by Non-PE is one after an airport
transitions from government to Non-PE private ownership. Post-Priv Non-PE to PE is one after an airport that is already privatized
by a non-PE firm transitions to PE ownership. Post-Priv PE to Non-PE is the reverse. There are no airports with data for PE to
non-PE transitions for the dependent variables in columns 3, 4, and 6, so these coefficients are empty. We include the log GDP per
capita, share of international passengers, government size, open markets, and log trade volume as control variables, and include
airport, and year fixed effects. We report two p-values on F-tests for equality of coefficients. The first compares the two privatization
coefficients and the second compares the post-privatization coefficients. Standard errors are clustered at an airport level. Statistical
significance is indicated by: *p<.1 **p<.05 ***p<.01. Significance: *p<.1 **p<.05 ***p<.01

Dependent Variable: Log(Op. Log(Op. Log(Aero Log(Non- Log(Op. Log(Employees

Net Revenue) Revenue) Aero Expenditure) per 1000
Income) Revenue) psg)
(1) (2) (3) (4) (5) (6)
1(Privatization by PE) 0.73∗∗ 0.71∗∗ 0.71∗∗ 0.84∗∗ 0.35∗∗ -0.00
(0.35) (0.33) (0.35) (0.34) (0.15) (0.07)
1(Privatization by Non-PE) 0.30∗∗∗ 0.08 0.20∗∗ -0.03 -0.23∗∗ -0.08∗∗
(0.10) (0.07) (0.09) (0.08) (0.11) (0.03)
1(Post-Priv Non-PE to PE) 0.03 0.03 0.06 0.06 -0.01 -0.02
(0.08) (0.07) (0.10) (0.07) (0.04) (0.02)
1(Post-Priv PE to Non-PE) -0.54∗∗∗ -0.49∗∗∗ -0.61∗∗∗ -0.35∗ -0.41∗ -0.07∗
(0.15) (0.08) (0.07) (0.21) (0.23) (0.04)
Observations 2613 2715 2744 2717 2686 2749
Airport FE Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes
R2 0.91 0.94 0.91 0.93 0.92 0.88
Y-Mean 17.98 18.94 18.24 18.17 9.54 0.13
Pr > F Priv Non-PE=Priv PE .23 .07 .16 .01 0 .36
Pr > F Non-PE to PE=PE to Non-PE 0 0 0 .06 .08 .11

51
 Table 9: Role of the Privatization Form for Main Outcomes
This table reports estimates of how four ownership transitions affect eight primary performance outcomes, relative to government-owned airports. The sample is an airport-year level
panel from 1996 to 2019. Column 1 reports change in total passengers per flight. Column 2 reports change in number of passengers. Column 3 reports change in number of flights.
Column 4 reports change in number of direct routes. Column 5 reports change in number of airlines operated in the airport. Columns 6 and 7 report changes in international charges
and domestic charges where charges are defined as fees that are levied on airlines for using the airport’s passenger service and runway. All charges are calculated per individual flight.
Column 8 reports changes in net operating income. All variables except for passengers per flight and number of airlines are log transformed values. The independent variables capture
two ownership type changes with the variation in ownership and control stake intensity and government ownership as the base group. 1(PE-Sale) is an indicator variable of airport sale
by PE. 1(PE-Concession) is an indicator variable of airport concession deal by PE. 1(NonPE-Sale) is an indicator variable of airport sale by NonPE-private. 1(NonPE-Concession) is
an indicator variable of airport concession deal by NonPE-private. We include the log GDP per capita, share of international passengers, government size, open markets, and log trade
volume as control variables, and include airport, and year fixed effects. We report two p-values on F-tests for equality of coefficients. The first compares PE-Sale to PE-Concession
and the second compares NonPE-Sale to NonPE-Concession. Standard errors are clustered at an airport level. Statistical significance is indicated by: *p<.1 **p<.05 ***p<.01

Dependent Variable: Passengers per Log(Number Log(Number Log(Number Number of Log(Int’l Fee) Log(Domestic Log(Op. Net
Flight of Passengers) of Flights) of Routes) Airlines Fee) Income)
(1) (2) (3) (4) (5) (6) (7) (8)
1(PE-Sale) 12.37∗∗∗ 0.46∗∗∗ 0.39∗∗∗ 0.37∗∗∗ 4.35∗∗∗ 0.67∗∗∗ 0.61∗∗∗ 0.83∗∗
(2.46) (0.06) (0.06) (0.06) (0.98) (0.17) (0.15) (0.32)
1(PE-Concession) 4.65 0.23∗∗ 0.21∗∗ -0.04 0.70 0.22∗∗∗ 0.28∗∗ 0.00
(4.53) (0.10) (0.08) (0.09) (0.94) (0.08) (0.11) (.)
1(NonPE-Sale) 4.77∗∗∗ 0.21∗∗∗ 0.18∗∗∗ 0.16∗∗∗ 2.96∗∗∗ 0.25∗∗∗ 0.27∗∗∗ 0.58∗∗
52

(1.73) (0.04) (0.04) (0.04) (0.50) (0.09) (0.10) (0.26)

1(NonPE-Concession) -2.17 0.14∗∗∗ 0.17∗∗∗ 0.13∗∗∗ 2.43∗∗∗ 0.09∗∗ 0.12∗∗ -0.38∗∗∗
(1.85) (0.05) (0.05) (0.04) (0.72) (0.04) (0.05) (0.09)
Observations 40320 40320 40320 40306 40320 9105 9103 2613
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.96 0.94 0.94 0.93 0.95 0.96 0.87
Y-Mean 89.99 12.46 8.21 2.30 8.71 8.81 6.23 17.82
Pr > F PE Sale = PE Concession .12 .04 .08 0 0 .02 .07 .01
Pr > F Non-PE Sale = Non-PE Concession 0 .27 .93 .54 .53 .11 .17 0
 Table 10: Role of the Competitive Landscape for Main Outcomes
This table estimates how the effects of PE and non-PE private ownership are mediated by having competing airports nearby, for the main outcome variables. 1(PE with Airports
Nearby) is 1 when PE has airports with a comparable size within 200 km. 1(PE without Airports Nearby) is 1 when NonPE-Private does not have airports with a comparable size
within 200 km. 1(NonPE-Private with Airports Nearby) is 1 when NonPE-Private has airports with a comparable size within 200 km. 1(NonPE-Private without Airports Nearby)
is 1 when NonPE-Private does not have airports with a comparable size within 200 km.We include the log GDP per capita, share of international passengers, government size, open
markets, and log trade volume as control variables, and include airport and year fixed effects. We report two p-values on F-tests for equality of coefficients. The first compares the
two coefficients for PE: having competing airports nearby (“w Comp") or not (“wo Comp"). The second compares whether NonPE-Private has airports nearby. Standard errors are
clustered at an airport level. Statistical significance is indicated by: *p.1 **p.05 ***p.01.

Dependent Variable: Passengers per Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Fee) Log(Domestic Log(Op. Net
Flight of Passengers) of Flights) of Routes) Airlines Fee) Income)
(1) (2) (3) (4) (5) (6) (7) (8) (9)
1(PE with Airports Nearby) 9.17∗∗∗ 0.45∗∗∗ 0.41∗∗∗ 0.35∗∗∗ 4.24∗∗∗ -285.98∗ 0.60∗∗∗ 0.46∗∗∗ 1.00∗∗
(2.75) (0.06) (0.06) (0.06) (0.95) (168.95) (0.14) (0.12) (0.45)
1(PE without Airports Nearby) 8.47∗ 0.16∗ 0.13 -0.06 0.29 283.31 0.54∗∗∗ 0.95∗∗∗ -2.31
(4.39) (0.09) (0.10) (0.09) (1.17) (404.84) (0.14) (0.31) (1.67)
1(NonPE-Private with Airports Nearby) 1.84 0.20∗∗∗ 0.21∗∗∗ 0.19∗∗∗ 3.36∗∗∗ -281.92∗∗ 0.29∗∗∗ 0.20∗∗∗ 0.73∗
(1.71) (0.04) (0.04) (0.04) (0.58) (117.59) (0.09) (0.07) (0.38)
1(NonPE-Private without Airports Nearby) 1.69 0.13∗∗ 0.12∗∗ 0.06 1.53∗∗ 68.55 0.03 0.27∗∗ 0.14
(2.12) (0.06) (0.05) (0.04) (0.63) (154.25) (0.04) (0.11) (0.27)
Observations 40320 40320 40320 40306 40320 40320 9105 9103 2613
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
53

Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.96 0.94 0.94 0.93 0.79 0.95 0.96 0.87
Y-Mean 89.99 12.46 8.21 2.30 8.71 5,474.43 8.81 6.23 17.82
Pr > F PE w Comp=PE wo Comp .89 .01 .02 0 .01 .19 .74 .15 .1
Pr > F NonPE Priv w Comp=NonPE Priv wo Comp .95 .33 .16 .03 .03 .07 .01 .55 .27
 Appendix
(For Online Publication)
sss
Figure A.1: Event Studies – Privatization Effect on Airport Efficiency by Non-PE
This figure shows event studies of privatization effect on Airport Efficiency by Non-PE Private

(A) Int’l per Flight Passenger Traffic (B) Dom per Flight Passenger Traffic

(C) Log(Number of Int’l Passengers) (D) Log(Number of Dom Passengers)

(E) Log(Number of Int’l Flights) (F) Log(Number of Dom Passengers)

Internet Appendix 1
Figure A.2: Event Studies – Privatization Effect on Number of Routes by Non-PE Private
This figure shows event studies of privatization effect on Number of Routes by Non-PE Private

(A) Number of International Routes (B) Number of Domestic Routes

(C) Number of Low Cost Carriers (D) Airline HHI

Internet Appendix 2
Figure A.3: Event Studies – Effect of Non-PE Private Ownership on Punctuality, Safety, and Awards
This figure shows dynamic differences-in-differences event studies of transitions to PE ownership.

(A) Flight Cancellation Rate (B) On-Time Departure Rate

(C) Any Award (D) Number of Accidents

(E) Number of Fatalities

Internet Appendix 3
Figure A.4: Event Studies – Privatization Effect on Airport Charges on Airlines by Non-PE
This figure shows event studies of privatization effect on Airport Charges on Airlines by Non-PE Private

(A) Int’l Fee (B) Dom Fee

(C) Int’l Passenger Fee (D) Dom Passenger Fee

(E) Int’l Runway Fee (F) Dom Runway Fee

Internet Appendix 4
Figure A.5: Event Studies – Privatization Effect on Airport Financials by Non-PE
This figure shows event studies of privatization effect on airport financials by Non-PE Private

(A) Log(Total Net Operating Income) (B) Log(Total Operating Revenue)

(C) Log(Total Operating Expenditure) (D) Log(Total Aeronautical Revenue)

(E) Log(Total Non-Aeronautical Revenue)

Internet Appendix 5
Table A.1: The Distribution of Privatized Airports and PE Investor Fund Statistics By Country and Time
This table shows the distribution of privatized airports by country and time. We start with identifying privatization events for 2,444
airports with more than 10,000 passengers in 2016 in 75 countries for the period 1929-2020. The sample includes 436 airports that
were privatized during the period 1929-2020

Country Before 1990s 1990s 2000s 2010s 2020s Total

Albania 0 0 1 0 0 1
Argentina 0 28 1 0 0 29
Armenia 0 0 2 0 0 2
Australia 1 14 11 3 0 29
Austria 0 0 1 1 0 2
Bahamas 1 0 1 0 0 2
Belgium 0 0 1 2 0 3
Bermuda 0 0 0 1 0 1
Bolivia 0 3 0 0 0 3
Brazil 0 0 0 18 1 19
Bulgaria 0 0 2 2 1 5
Cambodia 0 3 0 0 0 3
Cameroon 0 3 0 0 0 3
Canada 0 6 1 1 0 8
Chile 0 3 2 0 0 5
Colombia 0 0 3 7 0 10
Congo 0 0 0 3 0 3
Costa Rica 0 0 1 1 0 2
Cote D’Ivoire 0 1 0 0 0 1
Croatia 0 1 0 1 0 2
Cyprus 0 0 2 0 0 2
Czech Republic 0 0 0 1 0 1
Denmark 0 0 4 0 0 4
Dominican Republic 2 1 6 0 0 9
Ecuador 0 0 2 0 0 2
Egypt 0 1 0 0 0 1
Equatorial Guinea 0 0 2 0 0 2
France 0 1 9 23 0 33
Gabon 1 0 0 0 0 1
Georgia 0 0 2 0 0 2
Germany 0 2 3 2 0 7
Greece 0 1 0 14 0 15
Honduras 0 0 3 0 0 3
Hungary 0 0 1 0 0 1
India 0 1 9 1 5 16
Indonesia 0 0 0 1 0 1
Italy 1 4 6 7 0 18
Jamaica 0 0 1 1 0 2
Japan 0 0 0 6 8 14
Jordan 0 0 1 0 0 1
Kazakhstan 0 0 0 0 1 1
Kosovo 0 0 0 1 0 1
Latvia 0 0 0 1 0 1
Macedonia 0 0 2 0 0 2
Madagascar 0 0 0 2 0 2

Internet Appendix 6
Country Before 1990s 1990s 2000s 2010s 2020s Total

Malaysia 0 0 1 0 0 1
Maldives 0 0 0 1 0 1
Malta 0 0 1 0 0 1
Mexico 0 0 41 1 0 42
Moldova 0 0 0 1 0 1
Myanmar 0 0 0 1 0 1
Netherlands 0 0 1 0 0 1
New Zealand 0 2 0 0 0 2
Nigeria 0 1 0 0 0 1
Norway 0 1 0 0 0 1
Pakistan 0 0 1 0 0 1
Peru 0 0 13 1 0 14
Philippines 0 0 0 3 0 3
Portugal 0 0 2 10 0 12
Puerto Rico 0 0 0 1 0 1
Russia 0 3 0 15 0 18
Saudi Arabia 0 0 0 5 0 5
Serbia 0 0 0 1 0 1
Slovenia 0 0 1 1 0 2
South Africa 0 0 1 1 0 2
Sweden 0 1 0 1 0 2
Switzerland 1 0 1 0 0 2
Tanzania 0 1 0 0 0 1
Thailand 1 1 0 0 0 2
Tunisia 0 0 2 0 0 2
Turkey 0 2 6 0 0 8
UK 7 10 10 2 0 29
US 0 1 0 1 0 2
Ukraine 0 0 1 0 0 1
Uruguay 0 1 1 0 0 2
Total 15 97 163 146 16 437

Internet Appendix 7
 Table A.2: PE Investor Fund Statistics

Panel A: Top 5 Firms by Number of Airports

Number of Airports
Ciclad 16
Macquarie 11
Advent International 7
IFM Investors 6
F2i 5

Panel B: Fund and Deal Statistics

Mean Median N
Fund Size 2.71b 1.17b 43
Closed-Ended 85% 20
Deals Exited 37.78% 90
Deals Exited by Year 10 26.67% 90
Years to Exit 8.32 7 34
Fund Region
EU 41.3% 46
NA 23.91% 46
OC 17.39% 46
AS 6.52% 46
SA 6.52% 46
AF 4.35% 46
Same Region as Airport 76.74% 43

Internet Appendix 8
Table A.3: Government Ownership of Top 10 Non-PE Firms
This table shows the list of top 10 non-PE private firms in terms of the number of airport deals and government ownership of each
firm as of 2022.

Non-PE Private Firms Country # of Gov’t Major owners

Deals (%)
Aeropuertos Argentina 33 15 Corporacion America S.A. 75.65%
Argentina 2000 Government 15.00%
Cedicor 9.35%
Vinci Airports France 29 0 Subsidiary of VINCI SA. VINCI SA Ownership:
Vinci Sa 13.94%
Qatar Holding Llc 3.74%
Partners Group (UK) Ltd. 0.02%
Fraport Germany 20 52.02 State of Hesse 31.31%
City of Frankfurt 20.71%
Deutsche Lufthansa AG 8.44%
Grupo Aeroportuario Mexico 18 0 Fernando Gerardo Chico Pardo 21.0%
del Sureste Grupo ADO 13.3%
TAV Airports Holding Turkey 17 0 Aeroports de Paris SA 46.1%
Tepe Insaat Sanayi AS 5.06%
SNC-Lavalin France 15 0 The Caisse de depot et placement du Quebec 19.9%
Jarislowsky, Fraser Ltd. 10.7%
RBC Global Asset Management, Inc. 10.1%
Grupo Aeroportuario Mexico 13 0 Weston Hill Equity Holdings LP 5.62%
del Pacifico Controladora Mexicana de Aeropuertos SA 4.39%
Grupo Mexico, S.A.B. de C.V. 3.58%
Grupo Aeroportuario Mexico 12 0 Fintech Holdings, Inc. 19.9%
Centro Norte Norges Bank Investment Management 4.11%
Fidelity Management & Research Co. LLC 4.11%
Egis France 8 0 The Vanguard Group, Inc. 2.91%
Norges Bank Investment Management 1.72%
Dimensional Fund Advisors LP 1.34%

Internet Appendix 9
 Table A.4: Effect on Freight

Dependent Variable: Freight per Flight Log(Freight Tons)

Total In’l Domestic Total In’l Domestic
(1) (2) (3) (4) (5) (6)
1(Privatization by PE) 2.94∗∗ 1.03 2.50∗ 0.77 0.46 0.40
(1.25) (0.74) (1.32) (0.54) (0.56) (0.47)
1(Privatization by Non-PE) 0.06 0.25 -0.10 -0.62∗∗∗ 0.40∗∗∗ -0.78∗∗∗
(0.19) (0.27) (0.19) (0.16) (0.14) (0.19)
1(Post-Priv Non-PE to PE) 1.31 2.38∗∗ 2.05∗∗ -0.12 0.98∗∗∗ 0.21
(0.99) (1.07) (0.91) (0.32) (0.37) (0.37)
1(Post-Priv PE to Non-PE) -0.75 2.93 -3.98∗ -1.20∗∗ 0.80 -0.33
(4.05) (3.45) (2.04) (0.49) (0.76) (0.89)
Observations 40357 40357 40357 40357 40357 40357
Airport FE Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes
R2 0.54 0.57 0.47 0.76 0.89 0.72
Y-Mean 2.62 2.65 2.12 12.46 3.88 5.84
Pr > F Priv Non-PE=Priv PE .02 .33 .05 .01 .92 .02
Pr > F Non-PE to PE=PE to Non-PE .62 .88 .01 .06 .82 .57

Internet Appendix 10
 Table A.5: Continuous Ownership and Control Stake

Dependent Variable: Passengers per Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Fee) Log(Domestic Log(Op. Net Log(Op.
Flight of Passengers) of Flights) of Routes) Airlines Fee) Income) Expenditure)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
PE*Ownership Stake 7.74 0.27∗ 0.23∗∗ 0.50∗∗∗ 2.30 -211.87 0.00 0.01 2.45∗∗ 1.23
(6.36) (0.14) (0.11) (0.12) (1.51) (441.12) (0.10) (0.13) (1.21) (0.87)
PE*Control Stake 6.28 0.23∗ 0.20∗∗ -0.07 0.50 47.65 0.11∗ 0.17∗∗ -1.72∗ -0.95
(5.65) (0.13) (0.10) (0.10) (1.00) (350.76) (0.06) (0.07) (1.02) (0.73)
NonPE*Ownership Stake 4.40∗∗ 0.21∗∗∗ 0.19∗∗∗ 0.15∗∗∗ 2.84∗∗∗ -350.01∗∗ -0.00 0.09 0.38∗ -0.14
(1.96) (0.04) (0.04) (0.04) (0.57) (141.89) (0.05) (0.10) (0.21) (0.18)
NonPE*Control Stake -2.70 0.12∗∗ 0.16∗∗∗ 0.10∗∗ 2.63∗∗∗ 54.95 0.05 0.08 -0.38∗∗∗ -0.22∗∗
(2.06) (0.06) (0.06) (0.04) (0.79) (149.94) (0.05) (0.06) (0.09) (0.11)
Observations 40320 40320 40320 40306 40320 40320 9105 9103 2613 2686
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.96 0.94 0.94 0.93 0.79 0.94 0.96 0.87 0.84
Y-Mean 89.99 12.46 8.21 2.30 8.71 5,474.43 8.81 6.23 17.82 9.45
Pr > F PE Ownership=PE Control .9 .87 .87 .01 .42 .73 .48 .38 .06 .17
Internet Appendix 11

Pr > F NonPE Private Ownership=NonPE Private Control .01 .16 .68 .37 .83 .04 .44 .94 0 .70
 Table A.6: Majority vs. Minority Stake

Dependent Variable: Passengers per Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Fee) Log(Domestic Log(Op. Net Log(Op.
Flight of Passengers) of Flights) of Routes) Airlines Fee) Income) Expenditure)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1(PE-Sale-Majority) 15.83∗∗∗ 0.47∗∗∗ 0.36∗∗∗ 0.42∗∗∗ 2.91∗∗∗ -135.45 0.27∗∗ 0.30∗∗∗ 0.78∗∗ 0.25
(2.78) (0.08) (0.08) (0.07) (1.11) (272.72) (0.11) (0.11) (0.37) (0.25)
1(PE-Sale-Minority) 9.25∗∗ 0.58∗∗∗ 0.55∗∗∗ 0.39∗∗∗ 7.89∗∗∗ -585.91∗∗∗ 1.13∗∗∗ 0.97∗∗∗ 2.77∗∗ 1.95∗
(4.06) (0.10) (0.09) (0.08) (1.59) (147.24) (0.24) (0.21) (1.37) (1.16)
1(PE-Concession-Majority) 4.23 0.21∗∗ 0.19∗∗ -0.07 0.66 88.13 0.13∗∗ 0.18∗∗∗ -0.28∗ -0.15
(4.85) (0.10) (0.09) (0.09) (0.97) (298.35) (0.06) (0.07) (0.16) (0.12)
1(PE-Concession-Minority) 6.21 0.24 0.27 0.11 1.47 164.45 0.38∗∗∗ 0.47∗∗ 0.00 0.00
(5.77) (0.23) (0.23) (0.17) (1.91) (406.11) (0.11) (0.21) (.) (.)
1(NonPE-Sale-Majority) 4.38∗∗ 0.19∗∗∗ 0.16∗∗∗ 0.14∗∗∗ 2.81∗∗∗ -351.57∗∗∗ 0.17∗∗ 0.21∗∗ 0.30∗ -0.17
(1.76) (0.04) (0.04) (0.04) (0.50) (123.00) (0.08) (0.10) (0.17) (0.14)
1(NonPE-Sale-Minority) 14.08∗∗ 0.52∗∗∗ 0.42∗∗∗ 0.40∗∗∗ 3.93 -107.06 0.60∗∗ 0.53∗∗ 2.37∗∗ 1.58∗
(5.72) (0.17) (0.16) (0.13) (2.46) (395.97) (0.27) (0.23) (1.13) (0.88)
1(NonPE-Concession-Majority) -16.49∗∗∗ -0.40∗∗ -0.26 -0.29∗∗ -1.20 161.22 -0.51∗ -0.40∗ -2.75∗∗ -1.80∗∗
Internet Appendix 12

(6.02) (0.18) (0.16) (0.13) (2.57) (418.02) (0.27) (0.23) (1.15) (0.91)
1(NonPE-Concession-Minority) -13.20∗ -0.15 -0.12 0.03 -5.55∗ 589.17 0.00 0.00 0.00 0.00
(6.90) (0.33) (0.34) (0.24) (2.96) (847.10) (.) (.) (.) (.)
Observations 40320 40320 40320 40306 40320 40320 9105 9103 2613 2686
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.96 0.94 0.94 0.93 0.79 0.95 0.96 0.87 0.84
Y-Mean 89.99 12.46 8.21 2.30 8.71 5,474.43 8.81 6.23 17.82 9.45
 Table A.7: Stacked Regression
This table estimates the main effects in a stacked regression estimator. We include the log GDP per capita, share of international passengers, government size, open markets, and log
trade volume as control variables, and include airport, match cohort-year and year fixed effects. Prob > F is the F-test probability that tests the likelihood of Non-PE Private and PE
coefficients being different from each other. Standard errors are clustered at an airport level. Statistical significance is indicated by: *p.1 **p.05 ***p.01.

Dependent Variable: Passengers Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Log(Domestic Log(Op. Log(Op.
per Flight of of Flights) of Routes) Airlines Fee) Fee) Net Expenditure)
Passengers) Income)
Internet Appendix 13

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1(Privatization by PE) 17.99∗∗∗ 0.61∗∗∗ 0.49∗∗∗ 0.37∗∗ 1.28 288.66 -0.02∗∗ -0.05∗∗∗ 2.38∗∗ 1.66∗∗
(5.55) (0.18) (0.17) (0.16) (1.13) (353.12) (0.01) (0.01) (1.02) (0.71)
1(Privatization by Non-PE) 1.37∗∗∗ 0.17∗∗∗ 0.18∗∗∗ 0.13∗∗∗ 3.11∗∗∗ -248.87∗∗∗ 0.29∗∗∗ 0.31∗∗∗ 0.23∗∗∗ -0.28∗∗∗
(0.32) (0.01) (0.01) (0.01) (0.11) (22.94) (0.02) (0.02) (0.03) (0.03)
1(Post-Priv Non-PE to PE) 8.87∗∗∗ 0.17∗∗ 0.13∗ 0.19∗∗∗ -0.32 156.69 0.17∗∗ 0.19∗∗ 0.09 0.04
(2.94) (0.07) (0.07) (0.07) (0.78) (223.41) (0.07) (0.08) (0.13) (0.10)
1(Post-Priv PE to Non-PE) 2.73 0.02 -0.00 0.23∗∗ 1.07 -465.53 0.04 0.06 -1.71∗∗∗ -1.35∗
(7.45) (0.17) (0.16) (0.09) (1.59) (392.71) (0.06) (0.05) (0.54) (0.72)
Observations 827727 827727 827727 827727 827727 827727 184378 184336 47731 49024
Airport-Group FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year-Group FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.91 0.96 0.94 0.94 0.93 0.79 0.95 0.96 0.89 0.92
Y-Mean 89.37 12.42 8.18 2.27 8.42 5,514.71 8.80 6.20 17.98 9.51
Pr > F Priv Non-PE=Priv PE 0 .01 .06 .15 .11 .13 0 0 .04 .01
Pr > F Non-PE to PE=PE to Non-PE .44 .4 .47 .68 .43 .17 .18 .18 0 .06
 Table A.8: Callaway Sant’Anna Estimator
This table estimates the main effects using the Callaway Sant’Anna estimator. We include the log GDP per capita, share of international passengers, government size, open markets,
and log trade volume as control variables, and include airport, match cohort-year and year fixed effects. Prob > F is the F-test probability that tests the likelihood of Non-PE Private
and PE coefficients being different from each other. Standard errors are clustered at an airport level. Statistical significance is indicated by: *p.1 **p.05 ***p.01.

Dependent Variable: Passengers Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Log(Domestic Log(Op. Log(Op.
per Flight of of Flights) of Routes) Airlines Fee) Fee) Net Expenditure)
Passengers) Income)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1(PE) 7.71∗ 0.23∗∗∗ 0.18∗∗ 0.16∗∗ -0.09 315.17∗ 0.11 0.11 -0.12 -0.22
(4.41) (0.09) (0.08) (0.07) (0.61) (177.38) (0.08) (0.08) (0.19) (0.27)
Observations 40622 40622 40622 40622 40622 40622 9239 9237 2641 2714
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Internet Appendix 14
 Table A.9: Main Outcomes Using Matched Sample
This table estimates the main effects in a matched sample. We match each privatized airport with never-privatized, government-owned non-target airports 2 yrs before privatization
using CEM, on passengers per flight, total number of passengers, total number of routes, GDP per Capita, Trade Volume, the share of international passengers, judicial eectiveness,
and financial freedom. The ratio of treated vs. control is 1:1. Privatization by PE is one after an airport transitions from government to PE ownership and zero otherwise. Similarly,
Privatization by Non-PE is one after an airport transitions from government to Non-PE private ownership. Post-Priv Non-PE to PE is one after an airport that is already privatized
by a non-PE firm transitions to PE ownership. Post-Priv PE to Non-PE is the reverse. We include the log GDP per capita, share of international passengers, government size, open
markets, and log trade volume as control variables, and include airport and year fixed effects. We report two p-values on F-tests for equality of coefficients. The first compares the
two privatization coefficients and the second compares the post-privatization coefficients. Standard errors are clustered at an airport level. Statistical significance is indicated by: *p.1
**p.05 ***p.01.

Dependent Variable: Passengers Log(Number Log(Number Log(Number Number of Airline HHI Log(Int’l Log(Domestic Log(Op. Log(Op.
per Flight of of Flights) of Routes) Airlines Fee) Fee) Net Expenditure)
Passengers) Income)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1(Privatization by PE) 16.45∗∗ 0.62∗∗∗ 0.54∗∗ 0.45∗∗∗ 0.97 300.57 -0.01 -0.10∗∗∗ 2.39∗ 1.73∗∗
(7.22) (0.24) (0.22) (0.17) (1.24) (455.34) (0.02) (0.03) (1.28) (0.85)
Internet Appendix 15

1(Privatization by Non-PE) -0.86 0.14∗∗∗ 0.18∗∗∗ 0.14∗∗∗ 2.40∗∗∗ -206.24∗ 0.20∗∗∗ 0.26∗∗∗ 0.13 -0.24
(1.60) (0.04) (0.04) (0.03) (0.50) (112.63) (0.06) (0.08) (0.21) (0.19)
1(Post-Priv Non-PE to PE) 5.19∗ 0.06 0.05 0.10∗ -0.73 202.14 0.18∗∗ 0.17∗∗ 0.36 0.26
(2.88) (0.06) (0.07) (0.05) (0.79) (242.53) (0.08) (0.08) (0.22) (0.17)
1(Post-Priv PE to Non-PE) 4.16 -0.75 -0.76 -0.13∗∗∗ -2.59 -1,027.44 -0.08∗∗∗ -0.16∗∗∗ 0.00 0.00
(13.43) (1.07) (0.95) (0.05) (1.82) (1,872.89) (0.03) (0.05) (.) (.)
Observations 12520 12520 12520 12518 12520 12520 3429 3429 1094 1127
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.90 0.96 0.94 0.93 0.94 0.80 0.90 0.95 0.90 0.87
Y-Mean 101.80 12.99 8.54 2.61 11.62 4,761.46 8.98 6.32 17.41 9.64
Pr > F Priv Non-PE=Priv PE .02 .05 .1 .07 .28 .28 0 0 .09 .03
Pr > F Non-PE to PE=PE to Non-PE .94 .45 .4 0 .35 .51 0 0 .11 .12
 Panel C: Fees Charged to Airlines
Dependent Variable: Log(Int’l Fee) Log(Domestic Fee)

(1) (2) (3) (4) (5) (6) (7) (8)

1(Privatization by PE) 0.23∗∗∗ 0.23∗∗∗
(0.05) (0.04)
1(Privatization by Non-PE) 0.08∗∗ 0.06∗
(0.04) (0.04)
1(Post-Priv Non-PE to PE) 0.16∗∗ 0.27∗∗
(0.08) (0.12)
1(Post-Priv PE to Non-PE) -0.02∗ -0.03
(0.01) (0.02)
Observations 7121 7259 7162 7111 7119 7257 7160 7109
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.96 0.96 0.96 0.96 0.97 0.97 0.97 0.97
Y-Mean 8.82 8.82 8.82 8.82 6.24 6.24 6.24 6.24

Panel D: Financials
Dependent Variable: Log(Op. Net Income) Log(Op. Expenditure)

(1) (2) (3) (4) (5) (6) (7) (8)

1(Privatization by PE) 0.32 -0.07
(0.27) (0.07)
1(Privatization by Non-PE) -0.07 0.09∗∗∗
(0.06) (0.03)
1(Post-Priv Non-PE to PE) 0.84 0.41
(0.58) (0.69)
1(Post-Priv PE to Non-PE) 0.00 0.00
(.) (.)
Observations 1938 1997 1970 1922 1980 2039 2012 1964
Airport FE Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.94 0.94 0.94 0.93 0.89 0.91 0.83 0.87
Y-Mean 17.80 17.80 17.80 17.80 9.43 9.43 9.43 9.43

Internet Appendix 16
B. Variable Definitions
Governance Variables
All governance variables are from the Heritage Foundation. For details please see: https://www.
heritage.org/index/pdf/2022/book/02_2022_IndexOfEconomicFreedom_METHODOLOGY.pdf.

• Gov’t Integrity: The score for this component is derived by averaging scores for the following three
sub-factors, all of which are weighted equally: Perceptions of corruption, Risk of bribery, and
Control of corruption including “capture” of the state by elites and private interests. Each sub-factor
is converted to a scale of 0 to 100 using the following equation:
S ub − Factori = 100 × (S ub − Factor Max − S ub − Factori )/(S ub − Factor Max − S ub − Factor Min )

• Judicial Effectiveness: The score for the judicial effectiveness component is derived by averaging
scores for the following three sub-factors, all of which are weighted equally: Judicial independence,
Quality of the judicial process, and Perceptions of the quality of public services and the independence
of the civil service. Each sub-factor is converted to a scale of 0 to 100 using the following equation:
S ub − Factori = 100 × (S ub − Factor Max − S ub − Factori )/(S ub − Factor Max − S ub − Factor Min )

• Gov’t Spending: The scale for scoring government spending is nonlinear, which means that
government spending that is close to zero is lightly penalized while government spending that
exceeds 30 percent of GDP leads to much worse scores in a quadratic fashion. The equation used to
compute a country’s government spending score is: GEi = 100˘α(Expendituresi )2 where GEi
represents the government expenditure score in country i; Expendituresi represents the average total
government spending at all levels as a percentage of GDP for the most recent three years; and is a
coefficient to control for variation among scores (set at 0.03). The minimum component score is
zero.

• Investment Freedom: The Index evaluates a variety of regulatory restrictions that typically are
imposed on investment. Points, as indicated in (https://www.heritage.org/index/pdf/2022/
book/02_2022_IndexOfEconomicFreedom_METHODOLOGY.pdf), are deducted from the ideal
score of 100 for each of the restrictions found in a country’s investment regime. It is not necessary
for a government to impose all of the listed restrictions at the maximum level to eliminate investment
freedom. The few governments that impose so many restrictions that they total more than 100 points
in deductions have had their scores set at zero.

• Financial Freedom: The Index scores an economy’s financial freedom by looking at five broad areas:
The extent of government regulation of financial services, The degree of state intervention in banks
and other financial firms through direct and indirect ownership, Government influence on the
allocation of credit, The extent of financial and capital market development, and Openness to foreign
competition. These five areas are considered so that the overall level of financial freedom that
ensures easy and effective access to financing opportunities for people and businesses in the economy
may be assessed. An overall score on a scale of 0 to 100 is given to an economy’s financial freedom.

• Open markets: The index scores free market economy and is derived by equal-weighting indices
like trade freedom, financial freedom, and investment freedom. The trade freedom score is based on
two inputs: The trade-weighted average tariff rate and A qualitative evaluation of nontariff barriers
(NTBs).

Internet Appendix 17
Downstream Performance Measures
• Airlines: Number of Airlines that are operated in the airport.

• Low Cost Carriers: Number of low cost carriers that are operated in the airport.

• Airline HHI: Herfindahl-Hirschman Index created using the share of Airlines.

• Share of Largest Airline: The share of the airline with the largest share.

• Competing Airports: An indicator variable that is 1 if an airport has airports nearby with comparable
sizes within 200 km.

Price Regulation Variables

All regulation variables are provided by David Gillen at University of British Columbia and are only
observed for major airports in Asia, Europe, and Oceania. For details please see: Gillen, D., Niemeier, H.
M. (2008). The European Union: evolution of privatization, regulation, and slot reform.

• No Regulation: Indicator variable of 1 if there is no regulation in that airport-year and 0 if there is a

regulation.

• Cost Based: Cost based regulation charges the same price that would ideally prevail in a perfectly
competitive market, equal to the efficient costs of production, plus a market-determined rate of return
on capital. It is an indicator variable of 1 if there is a cost based regulation and 0 if not.

• Revenue Cap: Revenue Cap regulation sets an overall limit in the allowed average price increase.
This differs from cost based regulation which seeks to regulate individual prices. Revenue Cap can be
regarded as a form of incentive regulation, though the strength of the incentives varies. The Revenue
Cap can be under a single or dual till regime; under a single till all revenues are considered when
setting the Revenue Cap while in a dual till only revenues derivative from aviation operations (landing,
passenger and parking charges) are considered.

• Hybrid: Some revenue capped airports are subject to regular cost based resets, and this form of
regulation can be seen as a combination of cost based and incentive regulation (or hybrid regulation).

Safety Measures
• Accidents: The number of accidents is from over 23,000 airliner (aircraft originally certified to carry
12 or more passengers), military transport category aircraft, and corporate jet aircraft accidents dating
back to 1919. The information is primarily derived from official governmental agencies, such as air
accident investigation boards and civil aviation authorities.

• Fatalities: The number of fatalities within an accident. It is a per flight measure aggregated at an
airport-year level.

• Award: An indicator variable that is 1 if an airport received an award for its service that year.

Internet Appendix 18
Airport Charge Variables
• Int’l Fee: International Fee is the sum of International passenger and runway charges. They are
charged per aircraft movement.

• Dom Fee: Domestic Fee is the sum of domestic passenger and runway charges. They are charged per
aircraft movement.

• Int’l Passenger Fee: International Passenger Fee is levied for processing passengers and includes
security costs. They are charged per aircraft movement.

• Int’l Runway Fee: International Runway Fee is levied for using the runway infrastructure of the
airport. They are charged per aircraft movement.

• Dom Runway Fee: Domestic Runway Fee is levied levied for using the runway infrastructure of the
airport. They are charged per aircraft movement.

Airport Financial Variables

• Total Operating Revenue: Total operating revenue is the total amount of money coming into airport
from both aeronautical and non-aeronautical activities.

• Total Operating Expenditure: Total operating expense is total expense an airport incurs through its
normal operations.

• Net Operating Income: Total operating revenue minus total operating expenditure.

• Total Aeronautical Revenue: Total aeronautical revenue is the total amount of money coming into
airport from aeronautical activities.

• Total Non-Aeronautical Revenue: Total Non- aeronautical revenue is the total amount of money
coming into airport from non-aeronautical activities such as typically includes retail food and
beverages, shopping, car parking, and property and real estate.

• Number of Employees per 1000 passengers: Derived by Number of Employees/Number of Passengers

× 1000

Airport On Time Performance Variables

• Cancellation: Percentage of flights that is cancelled.

• Departure: Percentage of flights that is on-time.

19
 about ECGI

The European Corporate Governance Institute has been established to improve corpo-
rate governance through fostering independent scientific research and related activities.

The ECGI will produce and disseminate high quality research while remaining close to
the concerns and interests of corporate, financial and public policy makers. It will draw on
the expertise of scholars from numerous countries and bring together a critical mass of
expertise and interest to bear on this important subject.

The views expressed in this working paper are those of the authors, not those of the ECGI
or its members.

www.ecgi.global
ECGI Working Paper Series in Finance

Editorial Board

Editor Mike Burkart, Professor of Finance, London School

of Economics and Political Science
Consulting Editors Renée Adams, Professor of Finance, University of Oxford
Franklin Allen, Nippon Life Professor of Finance, Professor of
Economics, The Wharton School of the University of
Pennsylvania
Julian Franks, Professor of Finance, London Business School
Mireia Giné, Associate Professor, IESE Business School
Marco Pagano, Professor of Economics, Facoltà di Economia
Università di Napoli Federico II
Editorial Assistant Asif Malik, Working Paper Series Manager

www.ecgi.global/content/working-papers
Electronic Access to the Working Paper Series

The full set of ECGI working papers can be accessed through the Institute’s Web-site
(www.ecgi.global/content/working-papers) or SSRN:

Finance Paper Series http://www.ssrn.com/link/ECGI-Fin.html

Law Paper Series http://www.ssrn.com/link/ECGI-Law.html

www.ecgi.global/content/working-papers