Accounting in Europe

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/raie20

Unveiling the consequences of ESG rating

disagreement: an empirical analysis of the impact
on the cost of equity capital

Chiara Mio, Marco Fasan, Francesco Scarpa, Antonio Costantini & Aoife
Claire Fitzpatrick

To cite this article: Chiara Mio, Marco Fasan, Francesco Scarpa, Antonio Costantini & Aoife
Claire Fitzpatrick (21 May 2024): Unveiling the consequences of ESG rating disagreement:
an empirical analysis of the impact on the cost of equity capital, Accounting in Europe, DOI:
10.1080/17449480.2024.2350955

To link to this article: https://doi.org/10.1080/17449480.2024.2350955

Published online: 21 May 2024.

Submit your article to this journal

Article views: 50

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=raie20
Accounting in Europe, 2024
https://doi.org/10.1080/17449480.2024.2350955

Unveiling the consequences of ESG rating

disagreement: an empirical analysis of the
impact on the cost of equity capital

CHIARA MIO *, MARCO FASAN *, FRANCESCO SCARPA *,

ANTONIO COSTANTINI * and AOIFE CLAIRE FITZPATRICK**
*Department of Management-Venice School of Management, Sustainability Lab, Ca’ Foscari University of Venice,
Venice, Italy and **Leibniz-Institute for Financial Research SAFE and Goethe University, Frankfurt am Main, Germany

(Received: November 2022; accepted: April 2024)

ABSTRACT Recent academic research exhibits considerable disagreement among ESG ratings from
different agency providers. The consequences of this disagreement on the market are still under-
explored; thus, we investigate whether this disagreement impacts the cost of equity capital. Using a
sample of 23,201 firm-month observations from January 2019 to March 2021, we find that ESG
disagreement positively moderates the negative relationship between the average ESG score and cost of
equity. By disentangling the aggregate ESG score, we find that the moderating effect of this
disagreement does not hold for any pillar. Furthermore, the association between ESG rating
disagreement and cost of equity is more pronounced in the presence of high analyst information
uncertainty. Overall, our findings highlight that ESG rating disagreement jeopardizes investors’
confidence in ESG ratings and weakens the role of these ratings in reducing the cost of equity, pointing
to the need to improve convergence across agency providers.

Keywords: ESG; rating; disagreement; information uncertainty; cost of equity capital

1. Introduction
The rise of sustainable investments has led to an increased demand for and use of ESG ratings or
scores (Abhayawansa & Tyagi, 2021; Widyawati, 2020). These ratings are formulated by agency
providers by drawing on public and/or private data and information, and are widely regarded as a
proxy for assessing a firm’s ESG performance. Depending on the agency providers, ESG ratings
are intended to assess the social and environmental impacts of companies or their exposure to
sustainability-risk and opportunities. ESG ratings enable users, primarily investors, to make
well-informed decisions regarding companies’ capabilities in managing ESG-related risks,
impacts, and opportunities (Christensen et al., 2022; Serafeim & Yoon, 2023).

Correspondence Address: Francesco Scarpa, Department of Management-Venice School of Management, Sustainability

Lab, Ca’ Foscari University of Venice, Cannaregio 873, Venice 30121, Italy. E-mail: [email protected]

Paper accepted by Francesco Mazzi

© 2024 European Accounting Association

2 C. Mio et al.

In line with the growth of the ESG rating market, a large number of agency providers have
emerged and established themselves as key players in the field (Avetisyan & Hockerts, 2017;
Billio et al., 2022). While the abundance of ESG ratings is not per se a negative issue, academic
research reveals the substantial divergence in the ESG scores provided by different agencies to
the same firms (Billio et al., 2021; Capizzi et al., 2021; Chatterji et al., 2016; Dorfleitner et al.,
2015; Gibson Brandon et al., 2021). This disagreement may have adverse consequences for users
and rated companies. It complexifies integrating sustainability information into investment
decisions, creates confusion, undermines confidence in ESG ratings, and decreases firms’ incen­
tives to become more sustainable (Berg et al., 2022). A similar issue can be observed in the
context of corporate sustainability reporting. Stolowy and Paugam (2023) note that the landscape
of sustainability reporting is characterized by confusion and lack of compatibility, where the
possibility of convergence is limited by the diversity of the definitions and concepts of sustain­
ability, existence of competing standard-setting organizations, and organizations promoting sus­
tainability practices.
While there are several studies documenting the lack of convergence in ESG ratings and
examining the causes (e.g. Berg et al., 2022; Billio et al., 2021; Chatterji et al., 2016; Christensen
et al., 2022; Kimbrough et al., 2022), the literature on the broader economic effects of ESG rating
disagreement is still limited (Christensen et al., 2022; Gibson Brandon et al., 2021; Kimbrough
et al., 2022; Serafeim & Yoon, 2023). Overall, existing evidence suggests that this disagreement
is relevant to market participants. While Gibson Brandon et al. (2021) reveal that environmental
rating disagreement is positively related to stock returns, Christensen et al. (2022) provide evi­
dence that ESG disagreement is associated with higher stock return volatility and larger absolute
price movements, and Kimbrough et al. (2022) find that ESG disagreement is positively associ­
ated with disagreement and uncertainty in the capital market. Serafeim and Yoon (2023) find that
the ability of ESG ratings to predict future ESG news is much weaker in the presence of signifi­
cant ESG disagreement, which is consistent with disagreement hindering the predictive value
and usefulness of ESG ratings. Despite this evidence, the consequences of ESG disagreement
on the cost of equity are not fully understood. The cost of equity is the required rate of return
given the market’s perceptions of a firm’s riskiness. If ESG ratings and the lack of convergence
between ESG agency providers affect the perceived riskiness of the firms, then equity financing
costs should be affected to the extent to which ESG ratings disagree.
This study contributes to the debate on the economic consequences of the lack of convergence
among ESG ratings by examining the impact of ESG rating disagreement on the cost of equity. Pre­
vious studies have consistently documented that firms with better ESG performance face significantly
lower constraints to obtaining financing in the capital markets (Cheng et al., 2014) and exhibit lower
costs of equity capital (Dhaliwal et al., 2011; El Ghoul et al., 2011). Therefore, we hypothesize that the
same relationship holds for firms scored by multiple ESG agency providers. We expect a negative
relationship between a firm’s average ESG rating and its cost of equity capital. Next, we conjecture
that this relationship is contingent upon the level of ESG rating disagreement. Drawing on the litera­
ture on the influence of information uncertainty on investor risk perceptions (Erickson et al., 2012; He
et al., 2013; Zhang, 2006), we hypothesize that the ESG rating disagreement positively moderates the
negative relationship between the average ESG rating and cost of equity capital. More specifically, we
conjecture that the beneficial effects of the ESG ratings on the cost of equity capital are less
pronounced in firms with higher ESG rating disagreement.
To test our hypotheses, we used a pooled panel regression with control variables, fixed effects,
and standard errors clustered at the firm level. We conducted our analyses on a sample of 1,278
European firms that have ESG ratings from at least two of a pool of major rating agencies (i.e.
Sustainalytics, RobecoSAM, Refinitiv, and Bloomberg) between January 2019 and March 2021,
yielding a total of 23,201 firm-month observations.
 Accounting in Europe 3

Our focus on European firms has a twofold justification. First, we intend to fill a gap in the
literature that is not addressed by the extant studies that examined similar questions related to
ESG rating disagreement and cost of equity using US samples (Avramov et al., 2022; Gibson
Brandon et al., 2021). Second, the EU is a unique research setting, due to the proliferation of
regulation on ESG-related issues – such as the Corporate Sustainability Reporting Directive
(CSRD), EU Taxonomy for sustainable activities, and proposal of ESG rating regulation –
which has contributed to increasing investors’ attention on the ESG performance of European
companies.
Our results corroborate existing evidence that firms with higher ESG scores have a lower cost
of equity capital, even in the presence of multiple ESG ratings. We also find that ESG rating dis­
agreement positively moderates the ESG rating-cost of equity relationship. As hypothesized, our
results suggest that the negative relationship between the average ESG rating and cost of equity
capital is weaker when firms have high ESG rating disagreement. In our additional tests, we find
that the cost of equity has a negative relationship with the individual environmental, social, and
governance scores and that the level of disagreement is not a significant moderator for any indi­
vidual pillar. Furthermore, we find that the association between ESG rating disagreement and
cost of equity is more pronounced in the presence of high analyst information uncertainty and
that ESG disagreement is a significant moderator only for firms that operate in environmentally
sensitive industries.
Overall, our study makes timely contributions to different streams of research. First, this study
corroborates prior evidence on the value-relevance of ESG scores for market participants. We
document that in the case of firms rated by multiple providers, the cost of equity is negatively
related to the average ESG score. This implies that investors carefully consider the existence of
multiple ratings for the same firm, and that they monitor and use the average score to assess a
firm’s level of risk. Second, we contribute to the academic debate on the economic consequences
of ESG rating disagreement (Avramov et al., 2022; Christensen et al., 2022; Gibson Brandon et al.,
2021; Kimbrough et al., 2022; Serafeim & Yoon, 2023). Our study reveals the moderating effect of
the ESG rating disagreement on the relationship between the average ESG score and cost of equity;
thus, our findings support the view that ESG disagreement is perceived by investors as valuable
information on the uncertainty surrounding the measurement and value relevance of a firm’s sus­
tainability actions that leads to undermining their confidence in ESG ratings. Finally, our study has
some implications for the recent debate on the lack of convergence in sustainability reporting
(Stolowy & Paugam, 2023), providing evidence on the adverse consequences that may follow
from the persistence of the state of confusion and incompatibility in the measurement of sustain­
ability issues and supporting recent regulatory initiatives (e.g. the EU Corporate Sustainability
Reporting Directive and the International Sustainability Standards Board) aimed at strengthening
standardization and comparability in sustainability reporting.
Our findings should be of interest to managers and policymakers as they highlight that solving
the divergence among ESG rating providers is urgent to enable investors to make informed
decisions in regards ESG-related risks, impacts, and opportunities. This disagreement may jeo­
pardize a firm’s efforts to achieve good ESG performance and compromise the usefulness of
ESG scores; therefore, firms are suggested to care about and monitor the universe of ESG
ratings, rather than engaging in cherry-picking strategies, while policymakers should more effec­
tively regulate to improve the reliability, comparability, and convergence of ESG ratings.
The remainder of the paper is organized as follows. We review the literature and develop our
hypotheses in Section 2 and present the research design in Section 3, and sample selection and
descriptive statistics in Section 4. The main results are reported in Section 5, while Section 6 pre­
sents the additional analyses, and Section 7 presents the robustness checks. We discuss the con­
tributions and implications for research and practice in Section 8.
4 C. Mio et al.

2. Literature Review and Hypothesis Development

2.1. ESG Rating and Cost of Equity Capital
There is a rich literature that examines the market reactions to ESG rating. Although exceptions
exist (e.g. Gjergji et al., 2020; Nazir et al., 2022), most studies report a negative relationship
between firms’ ESG performance, proxied by ESG scores, and the cost of equity capital.
Based on a number of theoretical models, there has long been speculation that ESG factors
can affect corporate risks, and ultimately, the cost of capital (Gillan et al., 2021).
For example, exploring the influence of ESG ratings on systematic risk, the cost of capital, and
equity valuations, Giese et al. (2019) demonstrate that high ESG-rated firms experience lower
levels of beta coefficient, and therefore, in the context of the Capital Asset Pricing Model
(CAPM), lower cost of equity. Gonçalves et al. (2022) argue that investors reward firms with
higher corporate and social performance by requiring lower equity premiums. Similarly,
based on a sample of Latin American firms, Ramirez et al. (2022) show that the greater the
ESG performance the lower the economic price of the firm for attracting capital. The inverse
relationship between sustainability performance and cost of capital is corroborated by consistent
evidence (Bhuiyan & Nguyen, 2020; El Ghoul et al., 2011; Gholami et al., 2023; Wong, 2018;
Yilmaz, 2022).
Scholars (e.g. Breuer et al., 2018; El Ghoul et al., 2011; Gillan et al., 2021) suggest two main
channels through which ESG performance influences the cost of equity: the perceived risk of the
firm and investor base. The risk channel reflects the idea that firms with better ESG performance
reduce investor risk perception, and consequently influence investment decision-making.
According to the investor base channel, firms with poor ESG engagement have to offer their
shareholders higher expected returns to compensate them for a narrow investor base.
We hypothesize that the same mechanisms hold valid even in the context in which a firm is
rated by multiple ESG agency providers, with investors monitoring all the scores and requiring
a lower (higher) rate of return to firms with high (low) average ESG ratings. Therefore, we estab­
lish the following baseline hypothesis:

H1: A firm’s average ESG rating is negatively related to its cost of equity capital.

2.2. ESG Rating Disagreement and Cost of Equity Capital

As a consequence of the growth of the ESG rating market, an emerging stream of research has
devoted increasing attention to comparing different ESG models, revealing substantial disagree­
ment across major agency providers (Billio et al., 2021; Capizzi et al., 2021; Chatterji et al.,
2016; Dorfleitner et al., 2015; Gibson Brandon et al., 2021).
For instance, Gibson Brandon et al. (2021) find that the average pairwise correlation between
the ESG ratings from seven providers (i.e. Asset4 (now Refinitiv ESG), Sustainalytics (now
Morningstar), Inrate, Bloomberg, FTSE, KLD (now MSCI), and MSCI IVA) was only approxi­
mately 0.45 for firms in the S&P 500 Index between 2010 and 2017. Billio et al. (2021) find that
only 24% of 1,049 companies listed in the MSCI World Index have the same score from Sustai­
nalytics, RobeccoSAM, Refinitiv, and MSCI. Additionally, scholars document that the degree of
ESG rating disagreement has been increasing over time, suggesting that rating agencies fail to
reach convergence about what constitutes, and how to measure, good ESG performance (Chris­
tensen et al., 2022).
Various causes can explain the lack of agreement between ESG agency providers. The main
reason is related to the different methodologies developed by agency providers; thus, substantial
heterogeneity exists in terms of theorization (i.e. disagreement in defining sustainability and its
 Accounting in Europe 5

dimensions) and commensurability (i.e. disagreement in measuring the same construct) (Chat­
terji et al., 2016). Additionally, the ESG performance metrics, information sources, and weight­
ings generally vary across agencies (Billio et al., 2021; Capizzi et al., 2021; Dimson et al., 2020).
Berg et al. (2022) identify and quantify three distinct sources of disagreement among ratings:
measurement divergence, scope divergence, and weight divergence. Measurement divergence
(contributing 56% of ESG disagreement) refers to a situation where rating agencies measure
the same attribute using different performance indicators. Scope divergence (contributing 38%
of ESG disagreement) refers to the situation where ratings are based on different sets of attri­
butes. Finally, weight divergence (contributing 6% of ESG disagreement) emerges when
rating agencies take different views on the relative importance of ESG attributes. Billio et al.
(2024) find that divergences in ESG ratings are primarily driven by differing accounting
methods.
Analyzing the role of firm-level characteristics, Gibson Brandon et al. (2021) find that ESG
rating disagreement tends to be more prevalent for the largest firms and firms without credit
ratings. Additionally, corporate ESG disclosure has an influence on ESG disagreement, but its
role is still controversial: while Christensen et al. (2022) reveal that disagreement among ESG
data providers is more pronounced for firms with high levels of ESG disclosure, Kimbrough
et al. (2022) show that voluntary ESG disclosure, especially when of higher quality, is associated
with reduced disagreement.
The lack of convergence between ESG agency providers challenges the usefulness and
reliability of ESG scores, and may obfuscate investors’ and other stakeholders’ understanding
of firm ESG performance (Berg et al., 2022). Since ESG rating disagreement reflects
information uncertainty about ESG performance (Kimbrough et al., 2022), we expect the
lack of convergence has the potential to jeopardize the value-relevance of ESG ratings in
reducing the cost of equity capital. Thus, we propose ESG rating disagreement as a
potential moderator of the relationship between a firm’s average ESG rating and cost of
equity capital.
Finance and accounting scholars have widely explored investor behavior under conditions of
information uncertainty, generally conceived as ‘ambiguity with respect to the implications of
new information for a firm’s value’ (Zhang, 2006, p. 105). In general, investors tend to be
averse to uncertainty, and this aversion grows with the level of uncertainty (Williams, 2015).
For instance, previous studies suggest that investors demand a higher cost of capital in the pres­
ence of earnings forecast dispersion (Erickson et al., 2012; He et al., 2013), accounting restate­
ments (Hribar & Jenkins, 2004), or accounting quality uncertainty (Larson & Resutek, 2017).
Lambert et al. (2012) show that the cost of capital is driven solely by the average amount of
uncertainty that investors assess and that this uncertainty is measured by the precision with
which they can assess firm value. Zhang (2006) demonstrates that stocks with higher information
uncertainty experience higher price drift. Furthermore, investors faced with uncertainty base
their decisions on the worst-case scenario (i.e. following maximum expected utility) (Bird &
Yeung, 2012). Therefore, an increase in information uncertainty about firm value causes inves­
tors to demand a higher premium (i.e. higher cost of capital) to compensate for the rise in uncer­
tainty/risk in regards the firm’s fundamentals.
Previous studies document that ESG rating disagreement is relevant for market participants,
influencing stock prices (Gibson Brandon et al., 2021) and generating uncertainty in the
capital market (Christensen et al., 2022; Kimbrough et al., 2022).
Collectively, all these arguments suggest that ESG rating disagreement is likely to jeopardize
the value-relevance of ESG scores. In the presence of high ESG rating disagreement investors are
likely to perceive high information uncertainty on ESG performance, and consequently, to
demand a higher rate of return. In summary, the ESG rating divergence can be interpreted as
6 C. Mio et al.

a source of information uncertainty about firm sustainability performance which prevents a

precise understanding of firm value, especially in the long term, all of which affect the relation­
ship between a firm’s average ESG score and its cost of equity.
Therefore, we hypothesize and test the following:

H2: The level of ESG rating disagreement positively moderates the average ESG rating-
cost of equity relationship. Specifically, the higher the ESG rating disagreement, the
weaker the negative relationship between a firm’s average ESG rating and its cost of
equity.

3. Research Design
3.1. Test of H1
We estimate a pooled panel Ordinary Least Square (OLS) regression model with control vari­
ables, fixed effects, and standard errors clustered at the firm level.
To test H1, which hypothesizes that greater average ESG rating leads to a lower cost of equity
capital, we estimate the following main model:

AVE COEi,t = b1 ESG AVERAGEi,t + b2 gi,t + b3 Xi,t + 1i,t (1)

See Appendix A for variable definitions. AVE COEi,t represents cost of equity capital for firm i
at time t. AVE COEi,t is a proxy for the cost of equity which is the average between the cost of
equity obtained from the CAPM model and the implied cost of equity obtained from the Easton’s
(2004) model of the Earnings-price (EP) ratio. We use this proxy as a universally accepted
measure of cost of equity capital does not exist. The CAPM (Lintner, 1965; Sharpe, 1964)
equates the cost of equity of a firm to the risk-free interest rate plus the firm’s beta times the
market risk premium, while the EP ratio is given by the ratio between forecasted earnings per
share and stock price. We capture the CAPM directly from Refinitiv. The average ESG rating
(ESG AVERAGEi,t ) is equal to the average of all the ESG scores for firm i at time t from the
agency providers included in our analysis.
Control variables are denoted by gi,t ; Xi,t represents the fixed effects, while 1i,t is the error
term. Specifically, we control for the level of ESG rating disagreement (ESG_DISAGREE­
MENT), number of ESG ratings (ESG_NUM), beta (BETA), book-to-market ratio (BTM), firm
leverage (LEVERAGE), average long-term growth forecast (LONGTERMGROWTH), firm size
(SIZE), analyst coverage (ANALYSTCOVERAGE), forecast dispersion (FORECASTDISPER­
SION) and mean earning per share forecast (EPS). We also employ fixed effects for time and
firm level. Time fixed effects are taken at the monthly level as it is the frequency of the data
as well as firm and industry fixed effects.
The number of ESG ratings is controlled because our dataset contains firms with varying
numbers of ESG ratings (2, 3, or 4). The number of ESG ratings could act as an additional
source of information uncertainty; therefore, disentangling this effect from ESG disagreement
is important. Beta is used to control for a firm’s sensitivity to market volatility; Sharpe (1964)
and Lintner (1965) find that a firm’s beta is positively associated with cost of equity. BTM is
controlled for as firms with higher BTM are predicted to earn higher returns (Duong et al.,
2021; El Ghoul et al., 2011; Fama & French, 1992), and if stocks with high BTM are mispriced
below their true value, then they will earn an abnormally high implied risk premium (Gebhardt
et al., 2001). We subsequently control for firm leverage as leveraged firms earn higher stock
returns, and thus, investors require a higher risk premium for firms with greater leverage
(Duong et al., 2021; Fama & French, 1992; Gebhardt et al., 2001). Generally, larger firms can
 Accounting in Europe 7

provide investors with more information than smaller firms, thus reducing information asymme­
try and cost of equity; therefore, we control for firm size (El Ghoul et al., 2011; Gebhardt et al.,
2001; Sharfman & Fernando, 2008). Analyst coverage is also controlled for as it reduces infor­
mation asymmetry and thus cost of equity (Bowen et al., 2008; Gebhardt et al., 2001). We
include forecast dispersion in our set of controls since firms with lower forecast dispersion gen­
erally have lower information asymmetry, and thus, lower cost of equity (Bowen et al., 2008;
Dhaliwal et al., 2006). Finally, we control for long-term growth forecasts since firms with
high expected forecasts can expect stock prices that are valued too high, and thus, low
implied risk premium (Duong et al., 2021; Gebhardt et al., 2001).
We focus on b1 to test H1. If the average ESG rating reduces a firm’s cost of equity capital,
then b1 will be significantly negative.

3.2. Test of H2
To test H2, which hypothesizes that ESG rating disagreement moderates the relationship
between average ESG score and cost of equity capital, we run a moderated regression analysis,
which includes the interaction term (ESG_INTERACTION) between ESG rating disagreement
(ESG_DISAGREEMENT) and average ESG rating (ESG_AVERAGE):

AVE COEi,t = b1 ESG AVERAGEi,t + b2 ESG DISAGREEMENTi,t

+ b3 ESG AVERAGEi,t ∗ESG DISAGREEMENTi,t + b4 gi,t + b5 Xi,t
+ 1i,t (2)

In moderated regression analysis with continuous variables, as in Equation (2), the interpretation
of coefficients significantly differs from that in purely additive multiple linear regression models
(Jaccard et al., 1990). In Equation (2), the regression coefficients for ESG_AVERAGE and ESG_­
DISAGREEMENT reflect conditional relationships: β1 reflects the influence of ESG_AVERAGE
on AVE_COE when ESG_DISAGREEMENT equals zero, and β2 reflects the influence of ESG_­
DISAGREEMENT on AVE_COE when ESG_AVERAGE equals zero.
The coefficient b3 represents an interaction effect such that it estimates the change in the slope
of cost of equity capital on average ESG rating given a one-unit change in ESG rating disagree­
ment. Therefore, the interaction term allows to depict the marginal effect of average ESG rating
on the cost of equity capital of a firm conditional on the level of ESG rating disagreement. In
sum, we focus on b3 to test H2. If the negative relationship between average ESG rating and
cost of equity is less strong in the presence of a high level of ESG rating disagreement, then
b3 will be significantly positive.

4. Sample Selection and Descriptive Statistics

4.1. Data and Sample Selection
To test our hypotheses, we use ESG data from four of the major rating agencies: Sustainalytics,
RobecoSAM, Refinitiv, and Bloomberg. The initial dataset consisted of ESG data collected for
14,626 listed firms in 27 EU countries and the United Kingdom.
We focus on Europe because it is different from other institutional settings such as the USA in that
there is more mandatory ESG disclosures – as requested by the Non-Financial Reporting Directive
and the upcoming Corporate Sustainability Reporting Directive – and the EU has put in place the
building blocks for a sustainable finance framework (European Commission, 2019). In this evolving
8 C. Mio et al.

regulatory landscape, ESG ratings have ‘an increasingly important impact on the operation of capital
markets and on investor confidence in sustainable products’ (European Commission, 2023, p. 1).
To capture ESG rating divergence, we only included firms that have ESG ratings from at least
two of the above-mentioned rating agencies, leaving a sample of 1,278 European firms. The
dataset is an unbalanced panel dataset, taken at a monthly frequency between January 2019
and March 2021. This was done to capture the full scope of ESG rating divergence as Sustaina­
lytics updates its ESG ratings on a monthly frequency and RobecoSAM makes sporadic quarterly
updates.
We obtained financial and cost of equity data from Refinitiv while ESG ratings were collected
from Refinitiv and Bloomberg.
Thus, the main sample with all necessary information includes 1,278 unique firms, with
23,201 firm-month observations from January 2019 to March 2021. Table 1 provides a break­
down of sampled firms by country (Panel A) and industry (Panel B).

Table 1. Sample composition.

Country Number of firms
Panel A: breakdown of companies by country
Austria 36
Belgium 54
Czech Republic 4
Denmark 52
Finland 48
France 159
Germany 204
Greece 24
Hungary 5
Ireland 24
Italy 126
Luxemburg 19
Malta 1
Netherlands 71
Poland 31
Portugal 16
Spain 75
Sweden 128
United Kingdom 201
Global Industry Classification Standard (GICS) Number of firms
Panel A: breakdown of companies by industry
Communication services 82
Consumer Discretionary 157
Consumer Staples 78
Energy 42
Financials 180
Health Care 108
Industrials 270
Information Technology 86
Materials 99
Real Estate 78
Utilities 56
Undisclosed 42
Notes: Table 1 provides a breakdown of the sampled companies by country and industry.
 Accounting in Europe 9

4.2. ESG Rating Disagreement and its Drivers

An oversight of the most important features of the four ESG rating providers is available in
Online Appendix A.
RobecoSAM and Refinitiv (Asset 4) are Swiss-based companies, Sustainalytics is based in the
Netherlands while Bloomberg is a USA-based rating agency. Each rating agency uses a scale
ranging from 0 (worst in class) to 100 (best in class) or in the case of Bloomberg 0–10. To
address Bloomberg’s different scale, we multiply each ESG score by 10, such that the final
adjusted ESG score for Bloomberg is also a rating from 0 to 100. Typically, Sustainalytics
uses an ESG risk score which is an inverted scale where 0 is the best in class and 100 is the
worst, but to make all ratings comparable we use the Sustainalytics rank which ranks firms’
ESG behavior from 0 (worst in class) to 100 (best in class). Refinitiv and RobecoSAM
provide the highest firm coverage with 1234 and 1245 firms, respectively, whilst Sustainalytics
has approximately half the coverage at 640 and Bloomberg has the lowest with just 200 firms.
Regarding the data sources, Sustainalytics, Bloomberg, and Refinitiv use publicly available
information, but Bloomberg also relies on direct contact with the firm. Although the sources
of information can sometimes converge, how it is processed via weighting can also be very
different. RobecoSAM uses a different approach that is entirely based on survey data.
Thus, by using different methodologies, the four rating agencies can disagree in scoring the
same firm, as shown in Online Appendix B.
Panel A in Online Appendix B shows the ratings for the Finnish pharmaceutical company
Orion in May 2020. RobecoSAM rates Orion the poorest along all individual ESG pillars. Sus­
tainalytics rates Orion as best in class in Governance with a score of 92, while RobecoSAM and
Refinitiv rank it below average in Governance with the scores of 41 and 49, respectively.
Panel B in Online Appendix B reports the ESG rating for French Transportation company Bollore
in March 2021. Although Bloomberg provides a score for each ESG pillar, an aggregate rating is not
provided for Bollore.1 This table suggests that different E, S, and G weightings are used by rating
agencies when calculating the combined ESG score. Had Sustainalytics equally weight E, S, and
G the final rating would be approximately 23; however, the firm receives a final rating of 11 demon­
strating a greater weighting of the social and governance pillars. Conversely, RobecoSAM and Refi­
nitiv appear to have equally weighted each pillar in their aggregate score.
Table 2 reports descriptive statistics of the ESG ratings of the four agency providers for the
sampled firms.
Columns (1) and (2) display the mean and standard deviation for each rating, whilst Columns
(3), (4), and (5) show the pairwise cross-correlation between ratings. Statistics are presented for
the aggregate ESG rating as well as the individual environmental, social, and governance pillars.
Table 2 reveals how the level of divergence depends on the rating agencies. For the combined
ESG score, Refinitiv and RobecoSAM have the highest correlation at 0.6476 whilst Sustainaly­
tics and Bloomberg have the least at 0.3504. Subsequently, the combined ESG score has the
highest average correlation, followed by environment, social, and governance with the lowest
average correlation. A potential explanation for this phenomenon could be that social and
governance are based on less objective measures than environment.
Thus, consistent with prior evidence, our data show a high level of ESG rating disagreement
across all ESG pillars for European firms.

4.3. Descriptive Statistics

In this section, we report descriptive statistics and correlations for all the variables used in testing
our hypotheses.
10 C. Mio et al.

Table 2. Descriptive statistics of ESG ratings.

Pearson correlations
(1) Mean (2) StdDev (3) Sustainalytics (4) Bloomberg (5) RobecoSAM
ESG pillar
Sustainalytics 70.465 24.611
Bloomberg 42.800 10.993 0.350
RobecoSAM 48.890 29.221 0.6 0.427
Refinitiv 62.607 17.334 0.598 0.506 0.647
Average correlation 0.522
Environmental pillar
Sustainalytics 67.456 24.699
Bloomberg 31.117 20.447 0.334
RobecoSAM 51.629 27.877 0.571 0.379
Refinitiv 58.995 24.729 0.565 0.434 0.577
Average correlation 0.477
Social pillar
Sustainalytics 70.152 24.902
Bloomberg 29.868 17.615 0.280
RobecoSAM 50.396 27.969 0.51 0.244
Refinitiv 67.743 19.524 0.514 0.236 0.574
Average correlation 0.393
Governance pillar
Sustainalytics 66.368 26.783
Bloomberg 60.816 12.425 0.271
RobecoSAM 45.750 30.377 0.485 0.248
Refinitiv 57.895 21.469 0.356 0.426 0.382
Average correlation 0.361
Notes: Table 2 presents descriptive statistics of the four agency providers’ ESG ratings for our sample.

Table 3 presents the descriptive statistics for the main sample, which consists of 23,201 firm-month
observations. The mean and standard deviation of ESG_AVERAGE are 56.42 and 21.04, respectively,
indicating considerable variation in firm ESG ratings. The disagreement among ESG ratings measured
by ESG_DISAGREEMENT has a mean and median of 16.41 and 15.56, respectively, with a standard
deviation of 9.03, indicating a substantial ESG disagreement across the entire dataset.
Table 4 presents the Pearson and Spearman correlations for the full sample. As expected
ESG_AVERAGE is positively and significantly correlated with ENV_AVERAGE, GOV_AVER­
AGE, and SOC_AVERAGE. The results indicate that ESG_DISAGREEMENT is negatively
and significantly associated with AVE_COE. However, ESG_DISAGREEMENT is negatively
and significantly correlated to ESG_AVERAGE, suggesting that firms with higher ESG rating
dispersion have lower ESG scores.
Table 5 presents the mean values of our model variables per the number of ESG ratings a firm
has. Of all, 762 firms have 2 ratings, 568 firms have 3 ratings, and 123 firms have 4 ratings. The
cost of equity remains constant at approximately 7% across all three categories, while ESG_DI­
SAGREEMENT and ESG_AVERAGE increase with the number of ratings.

5. Results
5.1. Results for Tests of H1
We now turn to examining H1 regarding whether a firm’s average ESG rating play a role in redu­
cing the cost of equity capital. Table 6 presents the results of our tests based on Equation (1)
 Table 3. Descriptive statistics.
N Mean Std. Dev. 1st Quartile Median 3rd Quartile
AVE_COE 23703 6.459591 51.28409 4.552526 6.331319 8.643405
ESG_AVERAGE 24342 56.42169 21.04338 40.31421 56.55104 73.65031
ENV_AVERAGE 24342 51.59491 21.40592 35.46564 52.46667 68.73383
SOC_AVERAGE 24342 54.44293 19.03062 40.59427 54.61197 69.40008
GOV_AVERAGE 24342 54.40338 19.97798 39.6046 55.16152 70.24426
ESG_DISAGREEMENT 24342 16.40814 9.03276 9.610312 15.5561 22.03649
SOC_DISAGREEMENT 24080 21.78089 10.58806 13.81115 21.98392 29.67825
ENV_DISAGREEMENT 24342 18.71449 10.10287 10.98934 17.93645 25.24089
GOV_DISAGREEMENT 24342 18.6211 9.493976 11.75656 17.74624 24.81556
ESG_NUM 24342 2.716909 0.653472 2 3 3
ENV_NUM 24342 3.119834 0.762916 3 3 4
SOC_NUM 24342 3.056363 0.799035 2 3 4
GOV_NUM 24342 3.119834 0.762916 3 3 4
BETA 22684 0.984455 0.4444536 0.6877637 0.9504887 1.232885
SIZE 23127 22.85811 1.829296 21.63656 22.72677 24.01513
LEVERAGE 23056 1.092197 6.247692 0.347327 0.699688 1.291122
BTM 23512 −13.0513 659.5266 0.284698 0.592012 1.075406
ANALYSTCOVERAGE 22329 13.78185 7.56783 7 13 20
FORECASTDISPERSION 22329 0.288269 1.099444 0.03431 0.098627 0.26243
EPS 22329 1.863664 4.259714 0.31208 0.966782 2.3909
LONGTERMGROWTH 13574 7.571259 17.45166 1.4 6.008875 10.8885
Notes: Table 3 presents the descriptive statistics of the variables used in the main analyses. The full sample consists of 23,201 firm-month observations during the period between
January 2019 and March 2021.
Accounting in Europe 11
 12
C. Mio et al.

Table 4. Correlations (Spearman above/Pearson below).

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21)

(1) AVE_COE 0.01 0.02 0.00 0.00 −0.03 −0.02 0.00 −0.01 −0.01 −0.01 −0.01 −0.01 −0.01 0.01 0.00 0.01 −0.01 −0.01 0.03 0.03
(2) ESG_AVERAGE 0.17 0.88 0.89 0.84 −0.29 0.15 0.17 −0.25 0.31 0.37 0.35 0.37 0.1 0.52 0.14 0.14 0.5 0.12 0.11 −0.08
(3) ENV_AVERAGE 0.16 0.9 0.82 0.69 −0.21 0.02 0.01 −0.22 0.39 0.29 0.27 0.29 0.09 0.54 0.16 0.2 0.48 0.11 0.1 −0.11
(4) SOC_AVERAGE 0.19 0.9 0.84 0.71 −0.22 −0.05 0.07 −0.25 0.34 0.22 0.22 0.22 0.08 0.51 0.17 0.22 0.42 0.13 0.07 −0.09
(5) GOV_AVERAGE 0.08 0.86 0.69 0.71 −0.22 0.07 0.12 −0.26 0.34 0.32 0.31 0.32 0.12 0.44 0.11 0.13 0.43 0.02 0.04 −0.13
(6) ESG_DISAGREEMENT −0.03 −0.24 −0.2 −0.19 −0.19 0.34 0.31 0.43 0.14 0.05 0.03 0.05 −0.03 −0.11 0.01 −0.02 −0.1 −0.08 −0.07 0.01
(7) SOC_DISAGREEMENT −0.01 0.09 −0.01 −0.02 0.02 0.47 0.35 0.1 0.07 0.32 0.29 0.32 −0.02 0 −0.04 −0.18 0.1 0.02 0.1 0.07
(8) ENV_DISAGREEMENT 0.02 0.2 0.12 0.1 0.11 0.35 0.37 0.05 0.18 0.37 0.34 0.37 −0.02 0.05 −0.02 −0.04 0.08 0.04 0.02 0.02
(9) GOV_DISAGREEMENT −0.03 −0.18 −0.2 −0.19 −0.07 0.47 0.11 0.04 −0.05 0 −0.02 0 0.02 −0.15 0.02 −0.06 −0.1 −0.05 −0.06 0.01
(10) ESG_NUM 0.05 0.39 0.41 0.35 0.41 0.02 0.08 0.19 −0.02 0.71 0.67 0.71 0.02 0.38 0.06 0.11 0.39 0.04 0.09 −0.1
(11) ENV_NUM 0.03 0.44 0.33 0.27 0.44 −0.03 0.24 0.32 −0.01 0.77 0.95 1 −0.01 0.24 0 −0.1 0.41 0.11 0.15 0
(12) SOC_NUM 0.01 0.4 0.29 0.24 0.4 −0.02 0.24 0.31 0 0.72 0.95 0.95 −0.01 0.24 0.02 −0.07 0.38 0.14 0.17 −0.01
(13) GOV_NUM 0.03 0.44 0.33 0.27 0.44 −0.03 0.24 0.32 −0.01 0.77 1 0.95 −0.01 0.24 0 −0.1 0.41 0.11 0.15 0
(14) BETA 0.54 0.12 0.09 0.09 0.15 −0.04 −0.04 0.01 −0.03 0.02 0.00 −0.01 0.00 0.15 0.04 0.12 0.08 0.09 −0.06 −0.04
(15) SIZE 0.29 0.5 0.5 0.47 0.44 −0.07 0 0.13 −0.08 0.4 0.34 0.33 0.34 0.16 0.29 0.46 0.55 0.18 0.15 −0.12
(16) LEVERAGE 0.12 0 0.02 0.01 −0.03 0.03 −0.02 0.04 −0.03 −0.03 −0.04 −0.03 −0.04 0.02 0.04 0.11 0.11 −0.01 −0.14 −0.08
(17) BTM 0.41 0.02 0.03 0.04 0.02 0 −0.03 −0.04 0 −0.03 −0.06 −0.05 −0.06 0.24 0.05 0.01 0 0.1 −0.12 −0.25
(18) ANALYSTCOVERAGE 0.08 0.57 0.51 0.46 0.55 −0.12 0.11 0.12 −0.11 0.48 0.55 0.54 0.55 0.11 0.56 0 0 0.17 0.15 −0.04
(19) FORECASTDISPERSION 0.27 0.04 0.04 0.03 0 −0.02 0.01 0.02 −0.03 0.02 0.06 0.07 0.06 0.15 0.08 0.01 0.09 0.06 0.57 0
(20) EPS 0.22 0.08 0.08 0.06 0.04 −0.05 0.04 0.06 −0.04 0.07 0.11 0.12 0.11 −0.08 0.16 −0.03 −0.02 0.13 0.41 0.13
(21) LONGTERMGROWTH −0.02 −0.11 −0.12 −0.11 −0.13 0 0 −0.01 0 −0.08 −0.04 −0.04 −0.04 −0.1 −0.1 −0.08 −0.07 −0.06 0.05 0.1

Notes: Table 4 presents the correlation matrix for the variables used in the main analyses. Correlations with significance levels < 0.05 are in bold.
 Accounting in Europe 13

Table 5. Descriptive statistics per number of ESG rating.

No. 2 ESG scores No. 3 ESG scores No. 4 ESG scores

Number of firms 762 568 123
AVE_COE 6.97 6.12 6.27
ESG_AVERAGE 45.42 63.09 65.91
ENV_AVERAGE 40.83 57.31 64.41
SOC_AVERAGE 46.08 58.96 64.10
GOV_AVERAGE 44.06 60.01 66.20
ESG_DISAGREEMENT 17.10 15.09 19.76
SOC_DISAGREEMENT 16.70 19.43 22.67
ENV_DISAGREEMENT 20.88 22.12 23.35
GOV_DISAGREEMENT 18.71 18.77 17.65
BETA 0.96 1.01 0.95
SIZE 21.82 23.48 23.57
LEVERAGE 1.34 0.95 0.96
BTM 1.32 0.92 0.78
ANALYSTCOVERAGE 9.18 16.11 18.96
FORECASTDISPERSION 0.23 0.34 0.57
EPS 1.37 2.26 1.83
LONGTERMGROWTH 9.45 7.36 5.32
Notes: Table 5 presents the descriptive statistics of the key variables used in the main analyses. The sample is split by the
number of ESG ratings covering a firm (i.e. 2, 3 or 4). The mean value of each variable is presented in this table.

using two different specifications: with firm fixed effects2 (column 1) and with firm and month
fixed effects (column 2).
Consistent with our hypothesis, we find that the estimated coefficient on ESG_AVERAGE is
indeed negative and statistically significant across all the models (p < 0.001 in both column
(1) and column (2)). These findings suggest that when a firm is rated by multiple ESG agency
providers, the average ESG score reduces the cost of equity capital. In other words, investors
monitor the ESG ratings provided by different agencies and use the average score to determine
the riskiness of a firm. They perceived firms to be less risky, and thus assign a lower rate of
return, when the average ESG score is higher.
In terms of the magnitude of this effect, the estimated coefficients of ESG_AVERAGE suggest
that a one-unit increase in the ESG_AVERAGE is associated with 0.0524–0.0541 percentage
points decrease in AVE_COE, depending on the model specification. Thus, a firm would
lower its cost of equity by 1.75–1.80 percentage points or (based on a mean cost of equity of
6.46%) by 27.04% to 27.92% if a firm increases its average ESG score from the 25th to the
75th percentile. This appears to be an economically significant effect.
In terms of other factors that help explain the cost of equity, the results for the control variables
corroborate prior studies. The estimated coefficient on BETA is positive and statistically signifi­
cant (p < 0.001) across all specifications. In line with prior literature (Sharpe, 1964; Lintner,
1965), these findings suggest that a firm’s sensitivity to market volatility is positively associated
with cost of equity. Conversely, we find that the estimated coefficient on ANALYSTCOVERAGE
is negative and statistically significant (p < 0.001) across all specifications. This is consistent
with findings from prior literature (Bowen et al., 2008; Gebhardt et al., 2001), as analyst cover­
age enhances transparency, reduces information asymmetry and thus cost of equity. Most of the
other characteristics in the model do not yield significant results. Finally, the SIZE is negatively
associated with the cost of equity (p < 0.05 in the model with firm and month fixed effects)
suggesting that larger firms benefit from economies of scale and perceived stability, thus redu­
cing the risk premium investors require.
14 C. Mio et al.

Table 6. Test of H1 and H2.

(1) (2) (3) (4)

AVE__COE AVE_COE AVE_COE AVE_COE
ESG_AVERAGE −0.0524*** −0.0541*** −0.0668*** −0.0696***
(−0.00903) (−0.00912) (−0.0123) (−0.0127)
ESG_DISAGREEMENT −0.00274 −0.00139 −0.0527* −0.0548*
(−0.00581) (−0.00578) (−0.022) (−0.0227)
ESG_INTERACTION 0.000904* 0.000966*
(−0.00039) (−0.0004)
ESG_NUM 0.333 0.285 0.341 0.294
(−0.214) (−0.216) (−0.214) (−0.216)
BETA 2.280*** 2.326*** 2.276*** 2.322***
(−0.24) (−0.237) (−0.24) (−0.237)
BTM 0.309 0.303 0.303 0.297
(−0.229) (−0.232) (−0.227) (−0.23)
LEVERAGE −0.0152 −0.0191 −0.013 −0.0167
(−0.0551) (−0.0541) (−0.0551) (−0.0542)
LONGTERMGROWTH 0.0257* 0.0237* 0.0256* 0.0236*
(−0.0113) (−0.0111) (−0.0112) (−0.0111)
FORECASTDISPERSION −0.173 −0.2 −0.174 −0.201
(−0.114) (−0.121) (−0.114) (−0.121)
EPS 0.0746 0.0712 0.0752 0.0718
(−0.0575) (−0.0544) (−0.0576) (−0.0545)
ANALYSTCOVERAGE −0.102*** −0.0624*** −0.104*** −0.0641***
(−0.0183) (−0.0184) (−0.0184) (−0.0185)
SIZE −0.297 −0.656* −0.278 −0.637*
(−0.232) (−0.259) (−0.232) (−0.26)
_cons 14.87** 22.75*** 15.28** 23.21***
(−5.521) (−6.155) (−5.607) (−6.27)
Firm fixed effects Yes Yes Yes Yes
Month fixed effects No Yes No Yes
Clustered by firm Yes Yes Yes Yes
N 12722 12722 12722 12722
Adj. R-sq 0.795 0.806 0.796 0.806
Notes: Table 6 reports the results of Equations (1) and (2). Columns (1) and (2) present the results of estimating Equation (1),
which tests the association between average ESG score (ESG_AVERAGE) and cost of equity capital (AVE_COE). Columns (3)
and (4) present the results of estimating Equation (2), which tests the effect of the interaction between average ESG score and
ESG rating disagreement (ESG_INTERACTION) on cost of equity capital (AVE_COE). All t-statistics are in parentheses and
are based on standard errors clustered by firm. ∗, ∗∗, ∗∗∗ indicate significance at the 0.05, 0.01, and 0.001 levels, respectively.

5.2. Results for Tests of H2

Next, we examine whether the ESG rating disagreement has a moderation effect on the relationship
between the average ESG score and the cost of equity capital. Specifically, we rerun our main analy­
sis including the interaction term between the ESG rating disagreement and the average ESG score.
The result of our tests based on Equation (2) are reported in Table 6 using two different specifica­
tions: with firm fixed effects (column 3) and with firm and month fixed effects (column 4).
We find that the estimated coefficients on the interaction term (ESG_INTERACTION) are posi­
tive and statistically significant (p < 0.05) across the specifications. These findings suggest that,
as predicted by H2, the negative relationship between the average ESG score and cost of equity is
positively moderated by the level of ESG rating disagreement. As disagreement above ESG
raters increases, the negative impact of the average ESG score on the cost of equity becomes
less pronounced. In other terms, the lack of convergence between ESG raters impairs the
 Accounting in Europe 15

value-relevance of the ESG scores for investors. Thus, the beneficial effect of the average ESG
score on the cost of equity is weaker in the presence of high ESG rating disagreement.
In terms of the magnitude of this effect, the estimated coefficients for ESG_INTERACTION
indicate that a one-unit increase in ESG_DISAGREEMENT leads to a reduction in the negative
impact of ESG_AVERAGE on AVE_COE by 0.000904–0.000966 percentage points, depending
on the model specification. Increasing ESG_DISAGREEMENT from the 25th to the 75th percen­
tile attenuates by 0.011–0.012 percentage points the reduction in the cost of equity associated
with a one-unit increase in the average ESG score. Thus, increasing ESG_DISAGREEMENT
from the 25th to the 75th percentile is associated with a 16–19 percent reduction in the
impact of a one-unit increase in the average ESG score on the cost of equity, depending on
the model specification. This appears to be an economically meaningful moderating effect.
In terms of other characteristics in the model, we find that the cost of equity has a positive and
statistical significance with BETA and a negative and statistical significance with
ANALYSTCOVERAGE.
Taken together, the results from Table 6 suggest that the average ESG score is associated with a
reduction in the cost of equity capital and such association is weakened by ESG rating
disagreement.

6. Additional Analyses
6.1. Disentangling the Individual ESG Pillars
Previous literature shows that not all the ESG dimensions have the same impact on the cost of
equity (El Ghoul et al., 2011). To advance our understanding of whether our main findings hold
in each ESG pillar, we conduct an additional test to investigate whether the average score and
disagreement on environmental, social, and governance can influence the cost of equity.
Then, we re-estimate Equations (1) and (2) using the average environmental (ENV_AVERAGE),
social (SOC_AVERAGE), and governance (GOV_AVERAGE) scores. We calculate the rating dis­
agreement separately for the environmental (ENV_DISAGREEMENT), social (SOC_DISAGREE­
MENT), and governance (GOV_DISAGREEMENT) pillars. The interaction term for each model
(SOC_INTERACTION, ENV_INTERACTION, and GOV_INTERACTION) is obtained by multi­
plying the average score for each pillar with the corresponding standard deviation.
Specifically, we run the following regressions:

AVE CoEi,t = b1 ENV AVERAGEi,t + b2 gi,t + b3 Xi,t + 1i,t (3)

AVE CoEi,t = b1 ENV AVERAGEi,t + b2 ENV DISAGREEMENTi,t

+ b3 ENV AVERAGEi,t ∗ENV DISAGREEMENTi,t + b4 gi,t + b5 Xi,t
+ 1i,t (4)

AVE CoEi,t = b1 SOC AVERAGEi,t + b2 gi,t + b3 Xi,t + 1i,t (5)

AVE CoEi,t = b1 SOC AVERAGEi,t + b2 SOC DISAGREEMENTi,t

+ b3 SOC AVERAGEi,t ∗SOC DISAGREEMENTi,t + b4 gi,t + b5 Xi,t
+ 1i,t (6)

AVE CoEi,t = b1 GOV AVERAGEi,t + b2 gi,t + b3 Xi,t + 1i,t (7)

16 C. Mio et al.

AVE CoEi,t = b1 GOV AVERAGEi,t + b2 GOV DISAGREEMENTi,t

+ b3 GOV AVERAGEi,t ∗GOV DISAGREEMENTi,t + b4 gi,t + b5 Xi,t + 1i,t (8)

The empirical results are available in Online Appendix C (Equations (3) and (4)), Online
Appendix D (Equations (5) and (6)), and Online Appendix E (Equations (7) and (8)). The coeffi­
cients of average ratings (ENV_AVERAGE, SOC_AVERAGE, and GOV_AVERAGE) are all
negative and significant (p < 0.001). Therefore, H1 is confirmed also when the ESG pillars are
separately examined. The results also suggest that the social score has a larger impact on decreas­
ing the cost of equity than the environment and governance scores.
Finally, the interactions terms are not significant for any of the individual pillars. Therefore,
H2 is not supported when the ESG score is disentangled into its components. This suggests that
investors interpret the divergence in the score of individual ESG dimensions differently com­
pared with the divergence at the level of the overall ESG score. While overall ESG rating dis­
agreement is negatively perceived by investors, the lack of convergence may not affect
investors’ perceptions of firms’ riskiness in case of individual pillars.

6.2. Analyst Coverage Uncertainty

Although investors and asset managers have recently placed a higher emphasis on ESG ratings,
their decision-making process is also influenced by financial estimates, opinions, and projec­
tions provided by analysts. The analyst coverage may affect the level of information uncer­
tainty in the landscape in which investors operate and impact investor risk perceptions.
Information uncertainty is typically proxied by the dispersion in analysts’ earnings forecasts
(Barron & Stuerke, 1998), which is the standard deviation of one-year-ahead analysts’ fore­
casts of earnings per share (EPS). Previous studies document a significant positive relation
between the analyst earnings forecast dispersion and cost of equity capital (Barron &
Stuerke, 1998; He et al., 2013) because the level of dispersion in earnings forecast is perceived
by investors as valuable information about the level of uncertainty concerning firms’ future
economic performance.
To further understand the extent to which the level of ESG rating disagreement is perceived by
investors as an additional source of uncertainty concerning the future performance of firms, we
conduct an additional test to investigate whether ESG rating disagreement exacerbates the increase
in the cost of equity due to analyst earnings forecast dispersion (FORECASTDISPERSION).
To do so, we perform an additional analysis using two interaction terms (Equation 9): AVE_­
DISP_INTERACTION (the interaction between FORECASTDISPERSION and ESG_AVERAGE)
and SD_DISP_INTERACTION (the interaction between FORECASTDISPERSION and ESG_DI­
SAGREEMENT). Our model thus becomes:
AVE CoEi,t = b1 ESG AVERAGEi,t + b2 ESG DISAGREEMENT i,t
+ b3 ESG AVERAGEi,t ∗ESG DISAGREEMENT i,t
+ b4 FORECASTDISPERSIONi,t ∗ESG DISAGREEMENT i,t (9)
+ b5 FORECASTDISPERSIONi,t ∗ESG AVERAGEi,t
+ b6 FORECASTDISPERSIONi,t + b7 gi,t + b8 Xi,t + 1i,t
The empirical results are available in Online Appendix F. We find that the coefficients of
AVE_DISP_INTERACTION are not significant for both Columns (1) and (2) while the coeffi­
cients of SD_DISP_INTERACTION are significantly positive (p < 0.05 for Column (1) and
p < 0.01 for Column (2)).
 Accounting in Europe 17

These results suggest that ESG rating disagreement amplifies the increase in the cost of equity
due to the uncertainty in forecasting EPS. Thus, firms with high levels of analyst earnings fore­
cast dispersion suffer from a larger increase in the cost of equity in the presence of higher levels
of ESG rating divergence.

6.3. The Role of Industry

Prior research shows that the market perceives controversial business sectors to be riskier and
thus assigns a higher risk premium to firms involved in these industries (El Ghoul et al.,
2011). To explore the industry effect in our study, we perform an additional analysis to investi­
gate how the impact of ESG disagreement on cost of equity capital varies across industries with
different ESG exposures. We hypothesize that investors have a higher sensitivity to ESG uncer­
tainty if the firm operates in an industry with higher environmental footprints, labor/supply chain
considerations, and regulatory stringency.
We draw on the EU Taxonomy for sustainable activities, which defines the criteria for econ­
omic activities that are aligned with broad environmental goals. Not all economic activities and
sectors are covered (eligible) by the classification system introduced by the EU Taxonomy. In the
first phase, the European Commission has prioritized the most environmentally sensitive indus­
tries, that is, those that mostly contribute to GHG emissions in the EU.
We then re-estimate Equation (2) by splitting our sample into two groups. One sub-
sample includes firms that operate in EU Taxonomy-eligible industries (Tax_Ind), while
the other includes firms that do not operate in EU Taxonomy-eligible industries
(Non_Tax_Ind).
The empirical results are available in Online Appendix G. We find that for all industries the
coefficient of average ESG score remains negative and significant, indicating that investors value
the ESG performance of firms irrespective of industries’ environmental sensitivity. This aligns
with our main findings in Table 6. However, the coefficients of the interaction term between
ESG rating disagreement and average ESG score is significant only for firms involved in indus­
tries that are eligible to the EU Taxonomy. This suggests that ESG disagreement has a moderat­
ing effect on the relationship between average ESG score and cost of equity capital only for firms
that operate in environmentally sensitive industries.

7. Robustness Checks
In this section, we report the results of some sensitivity analyses to test the robustness of our main
results on the moderating role of the ESG rating disagreement.
A first potential concern about the robustness of our results is that firms with high levels of
ESG rating disagreement may be systematically different from those with a low level of ESG
rating disagreement, which could result in a potential endogeneity problem. To circumvent
this bias, we perform propensity score matching (PSM). We match a group of control firms
(those with low levels of disagreement, i.e. ESG_DISAGREEMENT < 30) to a treated group
(those with high levels of disagreement, i.e. ESG_DISAGREEMENT > 30) with similar charac­
teristics. We use k-nearest neighbor matching with a maximum of fifteen control firms for each
treated firm and set a caliper width of 0.05 for the propensity score to ensure close matches. The
observations are matched in a random sequence to prevent deterministic patterns based on the
initial order. The firms were matched based on industry, book-to-market ratio, leverage, long
term growth, size, beta, and year.
The matching results are available in Online Appendix H. The t-test shows that the difference
in means between the treated and control group is non-significant, thus indicating a good match.
18 C. Mio et al.

We then used the matched sample to estimate the effect of having high/low ESG rating dis­
agreement on the cost of equity capital using a treatment variable equal to 1 if the firm is part
of the treatment group and 0 if it is part of the control group.

AVE CoEi,t = b1 Treatment∗ESG AVERAGEi,t + b2 Treatmenti,t + b3 gi,t + b4 Xi,t

+ 1i,t (10)

The empirical results of Equation (10) are available in Online Appendix I. We find that for
firms with low level of ESG disagreement, the coefficient for average ESG score remains nega­
tive and significant, whilst for firms with high ESG disagreement the effect of the average ESG
score remains significant but diminishes in magnitude. This corroborates our results showing that
the relationship between ESG score and a cost of equity capital is conditional on the level of ESG
rating disagreement.
Second, to further corroborate our results, we split our data by the level of ESG rating dis­
agreement, to examine the effect of the average ESG score on the cost of equity in the presence
of low and high levels of ESG rating disagreement. Since in our sample, the variable ESG_DI­
SAGREEMENT ranges from 0 to 61, we created two groups of firms: firms with ‘Low disagree­
ment’, with ESG_DISAGREEMENT ranging from 0 to 29, and firms with ‘High Disagreement’,
with ESG_DISAGREEMENT ranging from 30 to 61. Our model thus becomes:

CoEi,t = b1 ESG AVERAGEi,t + b2 gi,t + b3 Xi,t + 1i,t (11)

The empirical results of Equation (11) are available in Online Appendix J. We find that for
firms with low levels of ESG disagreement, the coefficient of ESG_AVERAGE remains signifi­
cantly negative, while for firms with high ESG disagreement, the coefficient of ESG_AVERAGE
becomes non-significant. This corroborates our results showing that the relationship between
ESG score and cost of equity capital is conditional on the level of ESG rating disagreement.

8. Conclusion
This paper examines the consequences of ESG rating disagreements on the cost of equity. We
used a sample of 23,201 firm-month observations from January 2019 to March 2021, and con­
trolled for firm-specific determinants as well as firm and month-fixed effects; our empirical
analysis reveals interesting findings that are consistent with our hypotheses.
First, we find that firms with higher average ESG scores benefit from a lower cost of equity
capital. Second, we document that the level of ESG rating disagreement positively moderates
the negative relationship between average ESG score and cost of equity. This finding suggests
that the beneficial effect of ESG ratings for the cost of equity is weaker for firms with a high
level of ESG disagreement.
To refine our main findings, we conducted three additional analyses. First, we find that the
relationship between the average score and cost of equity capital is negative and significant
for each individual ESG dimension, but that the level of ESG rating disagreement is not a sig­
nificant moderator for any pillar. Second, we split our sample into firms that operate in industries
covered by the EU taxonomy on sustainable activities and those that do not operate in industries
covered by the EU taxonomy, to find that while a high average ESG score reduces the cost of
equity in all industries, the ESG rating disagreement is a significant moderator only for firms
in industries covered by the EU taxonomy. Third, to corroborate our evidence that the ESG
rating disagreement is perceived by investors as a source of uncertainty, we find that the level
 Accounting in Europe 19

of ESG rating divergence is a positive moderator of the relationship between analyst earnings
forecast dispersion and cost of equity capital. This means that firms with high levels of
analyst earnings forecast dispersion suffer from a larger increase in the cost of equity in the pres­
ence of higher levels of ESG rating divergence.
We validate our main results by using propensity score-matching (PSM) to control for differ­
ences in firm characteristics between treatment and control groups, and reach similar conclusions
after matching each firm having a high level of ESG disagreement with its nearest neighbor.
Overall, our study makes timely contributions to different streams of literature.
First, we extend the evidence on the value-relevance of ESG performance for market partici­
pants, by documenting that ESG scores affect investor risk perception and reduce the required
rate of return. Additionally, our study reveals that the presence of multiple ESG raters for the
same firm is not per se negative. Evidence on the negative relationship between the average
ESG score and cost of equity suggests that investors carefully consider the existence of multiple
ratings for the same firm, and that they monitor and use the average score to assess a firm’s level
of risk.
However, our study shows that the presence of multiple ESG ratings has adverse consequences
for users when agency providers significantly disagree in rating the same firm. Indeed, as a
second area of contribution, this study broadens our knowledge and understanding of the econ­
omic consequences of ESG rating disagreement. The lack of compatibility and convergence in
ESG ratings makes it more complex to integrate sustainability information into investment
decisions. As hypothesized, we show that this state of confusion in measuring the same firm
is perceived by investors as valuable information on the uncertainty surrounding the authenticity
and value relevance of a firm’s sustainability actions, and therefore, its real capacity to success­
fully manage impacts, risks, and opportunities associated with sustainability issues. We reveal
that this uncertainty weakens the beneficial effects that a good ESG score has on the perceived
riskiness of a firm, and in turn, on the required rate of return for equity investors.
Our additional analyses suggest that the ESG rating disagreement is monitored by investors
not only when they look at the ESG ratings but also when they make use of financial estimates
and projections by analysts. In both cases, the dispersion in ESG ratings is perceived by investors
as information indicating the level of uncertainty about a firm’s future economic and financial
performance.
Thus, our study corroborates prior evidence that ESG rater disagreements are associated with
uncertainty in the capital market (Christensen et al., 2022; Kimbrough et al., 2022; Serafeim &
Yoon, 2023). This study sheds light on the consequences of the disagreement on the cost of
equity; therefore, it is closely related to Gibson Brandon et al. (2021) and Avramov et al.
(2022), which among other things, address a similar question. While Gibson Brandon et al.
(2021) find that ESG rating disagreement is associated with an increase of 92 basis points in
the annual cost of equity capital, Avramov et al. (2022) show that the negative ESG-CAPM
alpha relationship only exists among stocks with low rating uncertainty. However, the theoretical
framing and the research design we use to address this question are quite different from theirs;
hence, our study contributes to advancing and refining the understanding of the impact of ESG
rating disagreement on the cost of equity in different ways. First, while Gibson Brandon et al.
(2021) and Avramov et al. (2022) use a sample of US firms, our paper builds on European
firms. Second, the empirical evidence we provide (January 2019 to March 2021) is more
recent than the empirics in Gibson Brandon et al. (2021), who use data from 2010 to 2017,
and Avramov et al. (2022), who use data from 2002 to 2019. Additionally, our contribution
lies in revealing that ESG disagreement has a negative moderating effect on the positive relation­
ship between average ESG score and cost of equity capital, while Gibson Brandon et al. (2021)
and Avramov et al. (2022) investigate the marginal effects of disagreement on the cost of equity.
20 C. Mio et al.

Thus, our study more clearly shows the interplay between the average ESG rating and ESG rating
disagreement. We document that a high level of disagreement hinders the usefulness of a good
ESG rating as a proxy for a low level of firm’s riskiness, and jeopardizes the beneficial effects on
the cost of equity. Finally, while Gibson Brandon et al. (2021) and Avramov et al. (2022) only
use the overall ESG score, we replicate the analyses with the scores for each E-S-G-dimension,
finding that the rating disagreement does not work as a moderator of the relationship between the
cost of equity and average score for any individual pillar.
Finally, our study has some implications for the recent debate on the lack of convergence in
sustainability reporting (Stolowy & Paugam, 2023). While sustainability reporting is self-narra­
tive disclosure from the firm, ESG rating is an external assessment of the same real sustainability
actions. Providing evidence on the disagreement among major sustainability rating agencies and
on the adverse consequences for users and rated companies, we show how the lack of a level-
playing field in the definition and measurement of sustainability issues creates complexities
for firms, investors, and other stakeholders analyzing corporate sustainability-related activities.
Given that the probability of convergence in sustainability reporting appears limited, at least in
the short term (Stolowy & Paugam, 2023), our findings can be a useful starting point to anticipate
the adverse consequences that may follow from the persistence of the state of confusion in the
sustainability reporting landscape. Additionally, they support recent regulatory initiatives
aimed at enhancing alignment and comparability in sustainability reporting, such as the Corpor­
ate Sustainability Reporting Directive within the European Union and IFRS Sustainability Dis­
closure Standards developed by the International Sustainability Standards Board (ISSB).
The evidence we provide should be of interest to practitioners and policymakers.
Our findings reveal that diverse ESG ratings may have adverse consequences for the rated
companies as they may undermine the value relevance of good ESG scores. Therefore, we rec­
ommend that companies be cognizant of the ESG rating landscape in which they operate, recog­
nizing that rating agencies have their own methodologies that may not always converge. To
effectively leverage their sustainability efforts, firms should actively monitor and understand
the methodologies of each ESG provider, thus identifying relevant ESG factors and how they
are assessed. Firms should avoid selectively choosing favorable ESG raters, as investors tend
to scrutinize all scores assigned to the same firm. Furthermore, companies should strive to main­
tain consistency in their scores across all providers since divergence among ratings is viewed
unfavorably by investors and may diminish the positive effects of a strong sustainability
performance.
Our study offers important implications for policymakers by demonstrating the economic
importance of reducing ESG disagreement. It highlights the need for policymakers to contribute
to developing a shared and unified understanding of what constitutes good ESG performance and
promoting common metrics to measure ESG attributes. Our findings show how the lack of
common rules for ESG ratings may have adverse effects on the economic system. A persisting
state of confusion and uncertainty may discourage investors from integrating environmental,
social and governance factors into their investment decisions, and this may decrease the flow
of investments on sustainable economic activities. This scenario may discourage companies
from investing and engaging in sustainability issues as the possibility of an economic return
from a good ESG profile becoming more remote. In this sense, the EU Taxonomy Regulation
for sustainable activities and the recent proposal from the European Commission (2023) for a
regulation to improve the reliability, comparability, and transparency of ESG ratings, represent
desirable efforts to enhance the quality of information about ESG ratings and enable investors to
make better-informed decisions regarding sustainable investments.
Our study is subject to some limitations. First, our analysis is based on the ESG ratings from
four rating agencies. Therefore, it could be further enriched by including data from additional
 Accounting in Europe 21

ESG raters (e.g. MSCI). Second, our sample is restricted to European firms, and as a result, our
findings may have implications suitable only for this context. The EU is a unique research
setting, where investor attention and concern about the ESG performance of companies may
be exacerbated by the proliferation of regulation on ESG-related issues (e.g. mandatory sustain­
ability/ESG reporting). Future research may investigate the extent to which this regulation con­
tributes to reducing the ESG rating divergence and explore alternative institutional settings.
Finally, our study is limited by the proxy that we adopted to measure the cost of equity. To
further advance the debate on the economic consequences of ESG disagreement, future research
may explore the extent to which the market reactions are dependent upon the causes of the dis­
agreement (i.e. measurement, scope, and weight) or the type of investors.
Contributing to understanding and solving the issues associated with the measurement and
assessment of corporate sustainability is crucial to facilitate the transition to a more sustainable
economy.

Notes
1
Similar to the case of Bollore, for other firms in our database Bloomberg provides only the scores for the individual E, S,
and G pillars, without giving the aggregate ESG score. For our analyses, we only used the ESG scores directly provided
by Bloomberg since we did not deem it appropriate to manually calculate the missing scores.
2
We employ firm fixed effects to address some of the endogeneity prevalent when using ESG ratings; however, we
acknowledge endogeneity cannot be entirely ruled out in our analysis.

Acknowledgments
The authors acknowledge the research support from the EIBURS project ESG-credit and Trans­
pArEEnS project ‘Mainstreaming Transparent Assessment of Energy Efficiency in ESG
Ratings’. Aoife Fitzpatrick gratefully acknowledge research support from Prof. Loriana Pelizzon
and the Leibniz Institute for Financial Research SAFE. The usual disclaimer applies.

Disclosure statement
No potential conflict of interest was reported by the author(s).

ORCID
Chiara Mio http://orcid.org/0000-0003-0945-0334
Marco Fasan http://orcid.org/0000-0002-9426-6090
Francesco Scarpa http://orcid.org/0000-0002-6633-6127
Antonio Costantini http://orcid.org/0000-0002-9143-8125

References
Abhayawansa, S., & Tyagi, S. (2021). Sustainable investing: The black box of Environmental, Social, and Governance
(ESG) ratings. The Journal of Wealth Management, 24(1), 49–54. https://doi.org/10.3905/jwm.2021.1.130
Avetisyan, E., & Hockerts, K. (2017). The consolidation of the ESG rating industry as an enactment of institutional retro­
gression: Consolidation of the ESG rating industry. Business Strategy and the Environment, 26(3), 316–330. https://
doi.org/10.1002/bse.1919
Avramov, D., Cheng, S., Lioui, A., & Tarelli, A. (2022). Sustainable investing with ESG rating uncertainty. Journal of
Financial Economics, 145(2), 642–664. https://doi.org/10.1016/j.jfineco.2021.09.009
Barron, O., & Stuerke, P. (1998). Dispersion in analysts’ earnings forecasts as a measure of uncertainty. Journal of
Accounting, Auditing & Finance, 13, 243–268. https://doi.org/10.1177/0148558X9801300305
22 C. Mio et al.

Berg, F., Kölbel, J., & Rigobon, R. (2022). Aggregate confusion: The divergence of ESG ratings. Review of Finance,
1–30.
Bhuiyan, M. B. U., & Nguyen, T. H. N. (2020). Impact of CSR on cost of debt and cost of capital: Australian evidence.
Social Responsibility Journal, 16(3), 419–430. https://doi.org/10.1108/SRJ-08-2018-0208
Billio, M., Costola, M., Hristova, I., Latino, C., & Pelizzon, L. (2021). Inside the ESG ratings: (Dis)agreement and per­
formance. Corporate Social Responsibility and Environmental Management, 28(5), 1426–1445. https://doi.org/10.
1002/csr.2177
Billio, M., Costola, M., Hristova, I., Latino, C., & Pelizzon, L. (2022). Sustainable finance: A journey toward ESG and
climate risk. SAFE Working Paper No. 349. https://ssrn.com/abstract=4093838.
Billio, M., Fitzpatrick, A. C., Latino, C., & Pelizzon, L. (2024). Unpacking the ESG ratings: Does one size fit all? SAFE
Working Paper No. 415. https://ssrn.com/abstract=4742445.
Bird, R., & Yeung, D. (2012). How do investors react under uncertainty? Pacific-Basin Finance Journal, 20(2), 310–327.
https://doi.org/10.1016/j.pacfin.2011.10.001
Bowen, R. M., Chen, X., & Cheng, Q. (2008). Analyst coverage and the cost of raising equity capital: Evidence from
underpricing of seasoned equity offerings. Contemporary Accounting Research, 25(3), 657–700. https://doi.org/
10.1506/car.25.3.1
Breuer, W., Müller, T., Rosenbach, D., & Salzmann, A. (2018). Corporate social responsibility, investor protection, and
cost of equity: A cross-country comparison. Journal of Banking & Finance, 96, 34–55. https://doi.org/10.1016/j.
jbankfin.2018.07.018
Capizzi, V., Gioia, E., Giudici, G., & Tenca, F. (2021). The divergence of ESG ratings: An analysis of Italian listed com­
panies. Journal of Financial Management, Markets and Institutions, 9(2), 2150006. https://doi.org/10.1142/
S2282717X21500067
Chatterji, A. K., Durand, R., Levine, D. I., & Touboul, S. (2016). Do ratings of firms converge? Implications for man­
agers, investors and strategy researchers: Do ratings of firms converge? Strategic Management Journal, 37(8),
1597–1614. https://doi.org/10.1002/smj.2407
Cheng, B., Ioannou, I., & Serafeim, G. (2014). Corporate social responsibility and access to finance: CSR and access to
finance. Strategic Management Journal, 35(1), 1–23. https://doi.org/10.1002/smj.2131
Christensen, D. M., Serafeim, G., & Sikochi, A. (2022). Why is corporate virtue in the eye of the beholder? The case of
ESG ratings. The Accounting Review, 97(1), 147–175. https://doi.org/10.2308/TAR-2019-0506
Dhaliwal, D., Heitzman, S., & Zhen Li, O. (2006). Taxes, leverage, and the cost of equity capital. Journal of Accounting
Research, 44(4), 691–723. https://doi.org/10.1111/j.1475-679X.2006.00214.x
Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. (2011). Voluntary nonfinancial disclosure and the cost of equity
capital: The initiation of corporate social responsibility reporting. The Accounting Review, 86(1), 59–100. https://
doi.org/10.2308/accr.00000005
Dimson, E., Marsh, P., & Staunton, M. (2020). Divergent ESG ratings. The Journal of Portfolio Management, 47(1), 75–
87. https://doi.org/10.3905/jpm.2020.1.175
Dorfleitner, G., Halbritter, G., & Nguyen, M. (2015). Measuring the level and risk of corporate responsibility – An
empirical comparison of different ESG rating approaches. Journal of Asset Management, 16(7), 450–466. https://
doi.org/10.1057/jam.2015.31
Duong, H. K., Fasan, M., & Gotti, G. (2021). Living up to your codes? Corporate codes of ethics and the cost of equity
capital. Management Decision, 60(13), 1–24. https://doi.org/10.1108/MD-11-2020-1486
Easton, P. D. (2004). PE ratios, PEG ratios, and estimating the implied expected rate of return on equity capital. The
Accounting Review, 79(1), 73–95. https://doi.org/10.2308/accr.2004.79.1.73
El Ghoul, S., Guedhami, O., Kwok, C. C. Y., & Mishra, D. R. (2011). Does corporate social responsibility affect the cost
of capital? Journal of Banking & Finance, 35(9), 2388–2406. https://doi.org/10.1016/j.jbankfin.2011.02.007
Erickson, M., Wang, S.-W., & Zhang, X. F. (2012). The change in information uncertainty and acquirer wealth losses.
Review of Accounting Studies, 17(4), 913–943. https://doi.org/10.1007/s11142-012-9184-9
European Commission. (2019). Communication from the commission on the European Green Deal. COM (2019) 640
final. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52019DC0640.
European Commission. (2023). Proposal for a regulation of the European parliament and of the council on the trans­
parency and integrity of Environmental, Social and Governance (ESG) rating activities. https://eur-lex.europa.eu/
legal-content/EN/TXT/?uri=CELEX:52023PC0314.
Fama, E. F., & French, K. R. (1992). The cross-section of expected stock returns. The Journal of Finance, 47(2), 427–
465. https://doi.org/10.1111/j.1540-6261.1992.tb04398.x
Gebhardt, W. R., Lee, C. M. C., & Swaminathan, B. (2001). Toward an implied cost of capital. Journal of Accounting
Research, 39(1), 135–176. https://doi.org/10.1111/1475-679X.00007
Gholami, A., Sands, J., & Shams, S. (2023). Corporates’ sustainability disclosures impact on cost of capital and idiosyn­
cratic risk. Meditari Accountancy Research, 31(4), 861–886. https://doi.org/10.1108/MEDAR-06-2020-0926
 Accounting in Europe 23

Gibson Brandon, R., Krueger, P., & Schmidt, P. S. (2021). Esg rating disagreement and stock returns. Financial Analysts
Journal, 77(4), 104–127. https://doi.org/10.1080/0015198X.2021.1963186
Giese, G., Lee, L.-E., Melas, D., Nagy, Z., & Nishikawa, L. (2019). Foundations of ESG investing: How ESG affects
equity valuation, risk, and performance. The Journal of Portfolio Management, 45(5), 69–83. https://doi.org/10.
3905/jpm.2019.45.5.069
Gillan, S. L., Koch, A., & Starks, L. T. (2021). Firms and social responsibility: A review of ESG and CSR research in
corporate finance. Journal of Corporate Finance, 66, 101889. https://doi.org/10.1016/j.jcorpfin.2021.101889
Gjergji, R., Vena, L., Sciascia, S., & Cortesi, A. (2020). The effects of environmental, social and governance disclosure
on the cost of capital in small and medium enterprises: The role of family business status. Business Strategy and the
Environment, 30(1), 683–693. https://doi.org/10.1002/bse.2647
Gonçalves, T. C., Dias, J., & Barros, V. (2022). Sustainability performance and the cost of capital. International Journal
of Financial Studies, 10(3), 63. https://doi.org/10.3390/ijfs10030063
He, W. P., Lepone, A., & Leung, H. (2013). Information asymmetry and the cost of equity capital. International Review
of Economics & Finance, 27, 611–620. https://doi.org/10.1016/j.iref.2013.03.001
Hribar, P., & Jenkins, N. T. (2004). The effect of accounting restatements on earnings revisions and the estimated cost of
capital. Review of Accounting Studies, 9(2/3), 337–356. https://doi.org/10.1023/B:RAST.0000028194.11371.42
Jaccard, J., Wan, C. K., & Turris, R. (1990). The detection and interpretation of interaction effects between continuous
variables in multiple regression. Multivariate Behavioral Research, 25(4), 467–478. https://doi.org/10.1207/
s15327906mbr2504_4
Kimbrough, M. D., Wang, X., Wei, S., & Zhang, J. (2022). Does voluntary ESG reporting resolve disagreement among
ESG rating agencies? European Accounting Review, 1–33. https://doi.org/10.1080/09638180.2022.2088588
Lambert, R. A., Leuz, C., & Verrecchia, R. E. (2012). Information asymmetry, information precision, and the cost of
capital. Review of Finance, 16(1), 1–29. https://doi.org/10.1093/rof/rfr014
Larson, C. R., & Resutek, R. J. (2017). Types of investor uncertainty and the cost of equity capital. Journal of Business
Finance & Accounting, 44(9-10), 1169–1193. https://doi.org/10.1111/jbfa.12283
Lintner, J. (1965). Security prices, risk, and maximal gains from diversification. The Journal of Finance, 20(4), 587–
615. https://doi.org/10.2307/2977249
Nazir, M., Akbar, M., Akbar, A., Poulovo, P., Hussain, A., & Qureshi, M. A. (2022). The nexus between corporate environ­
ment, social, and governance performance and cost of capital: Evidence from top global tech leaders. Environmental
Science and Pollution Research, 29(15), 22623–22636. https://doi.org/10.1007/s11356-021-17362-0
Ramirez, A. G., Monsalve, J., González-Ruiz, J. D., Almonacid, P., & Peña, A. (2022). Relationship between the cost of
capital and environmental, social, and governance scores: Evidence from Latin America. Sustainability, 14(9),
5012. https://doi.org/10.3390/su14095012
Serafeim, G., & Yoon, A. (2023). Stock price reactions to ESG news: The role of ESG ratings and disagreement. Review
of Accounting Studies, 28, 1500–1530. https://doi.org/10.1007/s11142-022-09675-3
Sharfman, M. P., & Fernando, C. S. (2008). Environmental risk management and the cost of capital. Strategic Manage­
ment Journal, 29(6), 569–592. https://doi.org/10.1002/smj.678
Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. The Journal of
Finance, 19(3), 425–442. https://doi.org/10.1111/j.1540-6261.1964.tb02865.x
Stolowy, H., & Paugam, L. (2023). Sustainability reporting: Is convergence possible? Accounting in Europe, 20(2), 139–
165. https://doi.org/10.1080/17449480.2023.2189016
Widyawati, L. (2020). A systematic literature review of socially responsible investment and environmental social gov­
ernance metrics. Business Strategy and the Environment, 29(2), 619–637. https://doi.org/10.1002/bse.2393
Williams, C. D. (2015). Asymmetric responses to earnings news: A case for ambiguity. The Accounting Review, 90(2),
785–817. https://doi.org/10.2308/accr-50866
Wong, C. (2018). Rate the raters 2018: Ratings revisited. SustainAbility. https://www.sustainability.com/thinking/rate-
raters-2018-white-paper/.
Yilmaz, I. (2022). ESG-based sustainability performance and its impact on cost of capital: International evidence from
the energy sector. International Journal of Applied Economics, Finance and Accounting, 12(2), 21–30. https://doi.
org/10.33094/ijaefa.v12i2.529
Zhang, X. F. (2006). Information uncertainty and stock returns. The Journal of Finance, 61(1), 105–137.
24 C. Mio et al.

Appendices
Appendix A. Variable Definitions

Variable name Description

AVE_COEi,t The cost of equity for firm i at time t.
ESG_AVERAGEi,t The average of the ESG scores of a firm received from the four rating
agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics) for
firm i at time t.
ENV_AVERAGEi,t The average of the Environment scores of a firm received from the four
rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics)
for firm i at time t.
SOC_AVERAGEi,t The average of the Social scores of a firm received from the four rating
agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics) for
firm i at time t.
GOV_AVERAGEi,t The average of the Governance scores of a firm received from the four
rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics)
for firm i at time t.
ESG_DISAGREEMENTi,t The standard deviation of the ESG scores of a firm received from the four
rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics)
for firm i at time t.
ENV_DISAGREEMENTi,t The standard deviation of the Environment scores of a firm received from
the four rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM,
Sustainalytics) for firm i at time t.
SOC_DISAGREEMENTi,t The standard deviation of the Social scores of a firm received from the four
rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM, Sustainalytics)
for firm i at time t.
GOV_DISAGREEMENTi,t The standard deviation of the Governance scores of a firm received from
the four rating agencies (i.e. Bloomberg, Refinitiv, RobeccoSAM,
Sustainalytics) for firm i at time t.
ESG_NUMi,t The number of ESG ratings for firm i at time t.
ENV_NUMi,t The number of Environment ratings for firm i at time t.
SOC_NUMi,t The number of Social ratings for firm i at time t.
GOV_NUMi,t The number of Governance ratings for firm i at time t.
BTMi,t Ratio of book value of equity to market value of equity for firm i at time t.
LEVERAGEi,t Ratio of total debt to book value of equity for firm i at time t.
LONGTERMGROWTH The average long-term growth forecast for firm i at time t.
FORECASTDISPERSIONi, Standard deviation of one-year-ahead analyst forecasts of earnings per
t share for firm i at time t.
EPSi,t Average of one-year-ahead analyst forecasts of earnings per share.
ANALYSTCOVERAGEi,t Number of unique analysts issuing earnings per share forecasts for firm i at
time t.
SIZEi,t Logarithm of total assets for firm i at time t.