Quantitative ESG Disclosure and Divergence of ESG

TYPE Original Research
PUBLISHED 05 August 2022

DOI 10.3389/fpsyg.2022.936798
Quantitative ESG disclosure and

OPEN ACCESS divergence of ESG ratings
EDITED BY
Mark Anthony Camilleri,
University of Malta, Malta Min Liu*
REVIEWED BY
School of Business, Renmin University of China, Beijing, China
Monica Billio,
Ca’ Foscari University of Venice, Italy
Michele Costola,
Ca’ Foscari University of Venice, Italy Over the past decade, sustainable finance has been a topic of burgeoning
*CORRESPONDENCE significance for investors, and ESG ratings have become commonly used
Min Liu to implement ESG investment strategies in practice. Strikingly, it is widely
[email protected]
documented in both academic literature and investment practices that ESG
SPECIALTY SECTION
This article was submitted to
ratings of a given firm can be extremely different across rating providers.
Organizational Psychology, However, despite the disagreement in ESG ratings being subject to a lot of
a section of the journal criticism, only few studies have examined the sources and determinants of
Frontiers in Psychology
rating divergence. This study examines whether quantitative ESG disclosure
RECEIVED 05 May 2022
ACCEPTED 08 July 2022
is conducive to rating convergence among agencies. Based on ESG rating
PUBLISHED 05 August 2022 data of Chinese A-share listed companies, the author finds that greater
CITATION quantitative ESG disclosure, especially disclosure on environmental and
Liu M (2022) Quantitative ESG social pillars, results in greater divergence of ESG ratings. When employing
disclosure and divergence of ESG
ratings. Front. Psychol. 13:936798. a difference-in-differences design with a quasi-experiment of disclosure
doi: 10.3389/fpsyg.2022.936798 guidance introduced by Hong Kong Exchange, the results show that if
COPYRIGHT ESG disclosure is standardized and comparable, more numerical information
© 2022 Liu. This is an open-access reduces agencies’ rating disagreement instead. Further analyses show that
article distributed under the terms of
the Creative Commons Attribution the lack of agreement is related to a low rating in the future. The author
License (CC BY). The use, distribution also finds that the effect of quantified ESG disclosure on rating divergence
or reproduction in other forums is
is more pronounced when firms are single businesses rather than diversified
permitted, provided the original
author(s) and the copyright owner(s) businesses with poor ESG performance rather than good ESG performance.
are credited and that the original The results are robust to alternative measures of ESG rating divergence,
publication in this journal is cited, in
accordance with accepted academic
alternative sample, two-way clustering, and additional control variables. Taken
practice. No use, distribution or together, the results indicate that quantitative ESG disclosure degenerates
reproduction is permitted which does rating disagreement.
not comply with these terms.
KEYWORDS
ESG disclosure, ESG ratings, rating divergence, rating agency, quantitative disclosure
Introduction
In recent decades, companies have been under enormous pressure to be sustainable
not only because of the obligation to create a favorable social impact for stakeholders
other than shareholders but also because of the notion that environmental, social,
and governance (ESG) issues are important for firms’ competitiveness and legitimacy
(Lins et al., 2017; Camilleri, 2018; Cao et al., 2019; Grewal and Serafeim, 2020).
The 2020 Global Sustainable Investment Review (GSIR) reports the booming
popularity of sustainable investments, with $ 35.3 trillion assets under management
and 15% growth in 2 years. This implies that a remarkable demand for ESG
ratings as a third-party assessment of ESG issues. Investors, firms, researchers, and
even regulators rely on ESG rating agencies to evaluate firms’ ESG performance.
Frontiers in Psychology 01 frontiersin.org

Liu 10.3389/fpsyg.2022.936798
Ratings are regarded as a tool to address climate risks for and decompose the divergence into three dimensions: using
institutional investors, equally as important as firm valuation different categories (scope divergence), measuring the same
models, shareholder proposals, and hedging (Krueger et al., categories with different indicators (measurement divergence),
2020), and thus increasingly shape investment decisions and taking different weights on the relative importance of
(Hartzmark and Sussman, 2019). Moreover, an extensive stream the categories (weight divergence). They provide evidence that
of academic studies also draws conclusions from ESG ratings scope divergence and measurement divergence are the main
such as studies on economic consequences of ESG performance drivers of rating divergence. More interestingly, Christensen
(Servaes and Tamayo, 2013; Cheng et al., 2014; Khan et al., 2016; et al. (2022) focus on how the extent of firms’ ESG disclosure
Hubbard et al., 2017; Lins et al., 2017) and influence on asset drives the agreement or disagreement of ESG ratings and
pricing (Engle et al., 2020; Pedersen et al., 2021). find that greater disclosure leads to greater rating divergence,
However, a lot of attention from investors, academic which is totally opposite to earnings forecasts in the equity
researchers, and the financial press has been drawn to the ESG markets and credit ratings in the debt market. The most
rating disagreement among different rating agencies for the important difference between these studies and the author’s
same firm (Chatterji et al., 2016; Gibson Brandon et al., 2021; study is that they focus on explaining how the whole
Berg et al., 2022; Christensen et al., 2022). Taking the hot stock disclosure influences rating disagreement, while the author is
in Chinese A-share listed companies for example, Kweichow interested in exploring whether quantitative ESG disclosure
Moutai got a high “AA” from Sino-Securities Index but a low brings about some changes. On the one hand, consistent with
“C+” from SynTao Green Finance for its ESG performance by ESG disclosure (Christensen et al., 2022), analysts may also have
June 2020, which thoroughly confused investors. Regulators, more opportunities for different interpretations of quantitative
including the United States SEC and the European Commission, disclosure, giving rise to greater rating dispersion. On the
have expressed a heightened concern on the quality and other hand, quantitative disclosure not only markedly enhances
precision of ESG ratings. In general, the consistency of ESG the comprehensibility and comparability of ESG disclosure
ratings across different data providers reached an amazingly low but also decreases analysts’ subjectivity and discretion, thus
correlation from 0.30 to 0.66 (Chatterji et al., 2016; Billio et al., setting stage for rating consensus. Hence, it is ex ante unknown
2021; Gibson Brandon et al., 2021; Jørgensen and Ellingsen, whether numerical information would improve or deteriorate
2021). The lack of consistency calls into question the validity ESG rating convergence.
of ESG ratings, which could cause a whole set of adverse To shed light on this question, the author compiles ESG
consequences as follows: first, the disagreement may potentially rating data from six ESG rating providers in China (SynTao
taint sustainable investment decisions, raising challenges for Green Finance, Sino-Securities Index, China Alliance of Social
investors to integrate ESG dimensions into investment strategies Value Investment, WIND ESG, FTSE Russell, and Rankins),
and leading to inefficiencies in the capital market. Uncertainty which represent the major players in Chinese ESG rating
emerging from ESG rating divergence commands an uncertainty space. Given raters’ coverage of public listed companies, the
premium to compensate for the additional exposure, as well author restricts the sample to A-share firms listed on the
as discourages ESG-sensitive investors’ participation in the Shanghai and Shenzhen stock exchanges in China and the
market (Gibson Brandon et al., 2021; Avramov et al., 2022). sample period goes from 2014 to 2020, with 4,966 firm-year
Second, rating disagreement could reduce incentives for firms observations. Consistent with the average cross-correlations of
to improve ESG performance, because there is no sense or ESG ratings from different agencies in prior studies (Gibson
commercial logic in spending substantial resources on activities Brandon et al., 2021; Jørgensen and Ellingsen, 2021), most cross-
that would result in obscure ratings. Third, the inconsistency correlations of ratings are between 0.4 and 0.6 in this study. After
of ESG ratings may shake the foundation of data analysis in controlling for firm-level characteristics and some fixed effects,
academic research, resulting in inconsistency of conclusions. a preliminary regression analysis shows that greater quantitative
In short, it is crucial to establish a deeper understanding of ESG disclosure results in greater divergence of ESG ratings.
the fact and reasons of ESG rating disagreement and explore Furthermore, the author disentangles the ESG disclosure and
how to improve the validity and convergence of ratings across divided it into three pillars: E (the environmental pillar), S (the
rating agencies. social pillar), and G (the governance pillar). The regression
Recently, several studies have examined the sources and results are consistent with the conjecture of the author, who finds
determinants of rating divergence. Chatterji et al. (2016) that disclosures on environmental and social issues contribute
document a surprisingly low agreement of social ratings across more to disagreement than governance disclosures.
six well-known information intermediaries and find that raters To address the potential endogeneity problem and examine
not only define CSR in various ways (theorization is low) whether standardized quantitative disclosure helps to reduce the
but also use different methods and metrics to measure the rating disagreement, the author conducts a quasi-experiment on
same construct (commensurability is low). Furthermore, Berg the implementation of Environmental, Social, and Governance
et al. (2022) identify sources of ESG rating disagreement Reporting Guide introduced by Hong Kong Exchanges (HKEX)

Liu 10.3389/fpsyg.2022.936798
and conducts a difference-in-differences (hereafter DID) of related research and develops testable hypotheses. Section
estimation. The results corroborate the main finding and suggest Research design and data describes the research design and
that consistent disclosure does not lead to disagreement, and data. Section Results presents the main empirical results
that inconsistent disclosure does. Furthermore, the author and interpretations. Sections Cross-sectional analyses and
explores the consequences of rating divergence and finds that Robustness test report ion the cross-sectional analyses and
the lack of agreement is associated with low rating in the future, robustness tests, respectively. Section Robustness test discusses
implying that disagreement represents a source of uncertainty the conclusions of this study in detail. Section Discussion
about potential risks. Cross-sectional analyses show that the concludes the article.
effect of quantified ESG disclosure on rating disagreement is
more significant when firms are single businesses rather than
diversified businesses with poor ESG performance rather than Related research and hypothesis
good ESG performance. Lastly, the main results are robust development
to alternative measures of ESG rating divergence, alternative
sample, two-way clustering, and additional control variables. Related research
This study contributes to several streams of literature.
First, it extends the literature on non-negligible divergence ESG Ratings
of ESG ratings among different rating agencies (Chatterji Over the past decade, sustainable finance has been a topic
et al., 2016; Billio et al., 2021; Gibson Brandon et al., 2021; of burgeoning significance for investors, and ESG ratings
Jørgensen and Ellingsen, 2021). The findings of the author have become commonly used to implement ESG investment
suggest that even quantitative ESG disclosure brings about strategies in practice. In the meantime, a growing literature
multifarious interpretations of ESG performance and hence on management, economics, and finance also derives their
inconsistent ratings, which provide further empirical evidence conclusions from ESG ratings (Cheng et al., 2014; Khan et al.,
for the presence and determinants of rating disagreement. This 2016; Avramov et al., 2022; Christensen et al., 2022). Most users
study is closely related to Christensen et al. (2022) and has are incapable of assessing the ESG performance of companies on
a key conclusion that rating divergence is larger when firms their own and thus widely rely on ESG ratings from specialist
disclose more. Specifically, this study differs from theirs in that rating agencies. The agencies, as information intermediaries,
the author places importance on how quantitative disclosure use a set of methods to identify risks and opportunities
(i.e., numerical metrics) influences the evaluation process pertaining to ESG issues, transforming massive and complex
and results of information intermediaries about ESG issues. information to aggregate scores (Escrig-Olmedo et al., 2019).
Second, this study complements research studies related to ESG ratings support institutional investors with trillions of
sociology of valuation and evaluation. Although quantification dollars in assets under management to integrate ESG dimensions
enhances comprehensibility and comparability by condensing into investment strategies and screen portfolios for risks and
information, restricting discretion, and simplifying decision- opportunities, which are previously unforeseen from financial
making (Porter, 1995; Espeland and Stevens, 1998, 2008), the performance (GSIR, 2020). Evidently, ESG ratings, like credit
author finds that rating disagreement is more pronounced ratings in debt markets, can and do play a paramount role in
when firms disclose more numerical information, indicating capital allocation.
that low commensurability still poses a serious challenge to the In China, there are several authoritative third-party data
convergence of ratings. In contrast, a shared cognitive system, providers with expertise in evaluating ESG performance and
including common definition and similar measurement of ESG localizing international scoring methodologies such as SynTao
performance (Sauder and Espeland, 2009; Hsu et al., 2012; Green Finance, Sino-Securities Index, and China Alliance
Chatterji et al., 2016; Kotsantonis and Serafeim, 2019), plays of Social Value Investment (CASVI). Rating agencies collect
a paramount role in driving convergence on social evaluation ESG information from various sources, assess it in their
and judgment. Third, this study has implications for literature unique evaluation system, and produce ESG ratings or scores.
on economic consequences of ESG rating disagreement. Gibson Taking SynTao Green Finance for example, it collects self-
Brandon et al. (2021) and Avramov et al. (2022) predict and disclosed information, such as annual reports and sustainability
find that firms with high ESG rating disagreement require an reports, and negative ESG information, such as formal media
equity premium because disagreement is perceived as a source reports and penalties announced by regulatory authorities. The
of uncertainty. This study complements some evidence and rating framework encompasses general indicators applicable
demonstrates that divergence is related to a lower average rating to all companies and industry-specific indicators applicable
in the future, implying that quantitative disclosure degenerates to companies within industry classification (Broadstock et al.,
rating disagreement but exposes potential ESG risks. 2021). However, despite a great deal of time and energy spent on
The reminder of the article proceeds as follows: Section assessment, the validity and convergence of ESG ratings released
Related research and hypothesis development provides a review by agencies are criticized in practice and research.

Liu 10.3389/fpsyg.2022.936798
ESG rating disagreement among ESG rating analysts because of little agreement on
It has been debated critically in both academic literature how to interpret and judge the meaning of ESG performance
and investment practices that the ESG ratings of a given disclosures. According to this viewpoint, Christensen et al.
firm can be extremely different across rating providers. Prior (2022) find empirical evidence that greater ESG disclosure gives
studies (Chatterji et al., 2016; Billio et al., 2021; Gibson rise to greater ESG rating disagreement. Nevertheless, what
Brandon et al., 2021; Jørgensen and Ellingsen, 2021) document about quantitative ESG disclosure?
a surprising lack of rating agreement between worldwide On the one hand, greater quantitative ESG disclosure may
well-established information intermediaries, with an average lead to greater ESG rating divergence. In the ideal situation, if
correlation from 0.3 to 0.66. What is worse, Gibson Brandon rating agencies adopt the same approach to evaluate firms’ ESG
et al. (2021) and Avramov et al. (2022) find that ESG rating performance, the consistency of ratings should be fairly high no
disagreement seems to mislead even professional investors in matter how much firms disclose. Nonetheless, the fact is that
their investment decisions and then discourages them from there are thousands of ways that firms report their ESG data
sustainable investment and active engagement in corporate ESG and raters make their assessments, with different terminologies
issues. The same applies to academics that draw plenty of and different units of measure. To illustrate, Kotsantonis and
influential conclusions. As such, it is crucial to explore why raters Serafeim (2019) selected a random sample of 50 listed firms
diverge widely; as a response, a growing number of studies have in Fortune 500 and found that the 50 firms use more than 20
been dedicated to this topic. distinct metrics to report on the issue of employee health and
Researchers have found some theoretical underpinnings in safety, e.g., lost time frequency rate, total case incident rate,
the literature related to sociology of valuation and evaluation. number of severe accidents that occurred, occupational disease
There are two prerequisites to converge on assessments for rate, and number of lost workdays. These indicators are similar
raters: theorization and commensurability (Durand et al., 2007; but not identical and numerical but incomparable. Evidently,
Espeland and Sauder, 2007; Sauder and Espeland, 2009; Hsu there is no consensus on which of these metrics best capture
et al., 2012). A common theorization implies convergence good ESG performance and how they are aggregated into an
on a common definition of ESG across ESG rating agencies. assessment system (Chatterji et al., 2016; Christensen et al.,
Despite a broad rhetorical agreement on the components of 2022). For firms with low level of quantitative ESG disclosure,
ESG performance, there are actually dramatic differences in agencies commonly have little dispute about their performance
the way raters theorize ESG, and high-level scope divergence ranking near the bottom, whereas for firms disclosing many
(e.g., different sets of attributes between the scope of ratings) numerical information, things begin to get cluttered. Raters need
is one of the main drivers of rating divergence (Chatterji et al., to interpret and judge whether figures present good or bad ESG
2016; Berg et al., 2022). However, after adjusting for differences performance and compare the figures with those of peer firms
in theorization, rating divergence may remain high as a result despite incommensurability of different metrics. Accordingly,
of low commensurability. High commensurability means that the author formulates the following hypothesis:
raters employ similar measurements and interpretations of the H1a: quantitative ESG disclosure is positively related to the
same construct. In practice, different raters generally measure divergence of ESG ratings.
the same attribute with various indicators, which make it On the other hand, greater quantitative ESG disclosure may
impracticable to compare across ESG ratings and hence lead bring about lower ESG rating divergence. In many areas of social
to great disagreement. Berg et al. (2022) provide convincing sciences, the logic of quantification is implicit in a broad range
empirical evidence that measurement divergence is the most of valuation/evaluation systems1 . Quantification transforms
important force of rating divergence based on the data set of ESG quality into quantity and difference into numbers, integrating
ratings from six prominent agencies. information into a shared cognitive system (Espeland and
Stevens, 2008). The dominant reason why quantitative ESG
disclosure may mitigate rating disagreement is that by
Hypothesis development condensing information and simplifying decision-making,
quantification markedly enhances the comprehensibility and
Drawing on studies on investor disagreement in financial comparability of ESG disclosure (Espeland and Stevens, 1998).
markets, two main sources of disagreement are differences Compared with “establish a timely and quick consumer response
in information sets and differences in interpretations of system oriented to consumer needs”, it is apparently easier
information (Hong and Stein, 2007; Cookson and Niessner,
2020). Financial disclosures usually reduce dispersion among 1 Commensuration is an important concept in sociology of valuation
equity analysts, as well as credit rating analysts owing to and evaluation. Generally, most quantification can be understood as
widespread agreement on the meaning of financial information commensuration (Espeland and Stevens, 2008). In the context of this
(Morgan, 2002; Hope, 2003; Bonsall and Miller, 2017; Akins, article, the author does not make a distinction between “quantification”
2018), while ESG disclosures are likely to exacerbate dispersion and “commensuration” and uses the term “quantification”.

Liu 10.3389/fpsyg.2022.936798
to assure consistency and reach consensus on “customer indicators disclosed by firms who have to meet the mandatory
satisfaction is 95%” for rating agencies. In addition, numerical disclosure requirements. Therefore, the author proposes the
information, characterized as mechanical objectivity, restricts second hypothesis:
discretion especially when the credibility is challenged (Porter, H2: Standardized quantitative ESG disclosure is negatively
1995; Espeland and Stevens, 2008). Tang et al. (2021) find related to the divergence of ESG ratings.
that companies held by the same owners as the rating
agency receive higher ESG ratings, suggesting that the conflicts
of interest degenerate the quality of ESG ratings. Hence, Research design and data
quantitative ESG disclosure is so impersonal and comparable
Research design
that raters would decrease discretion and make consistent
judgments. Collectively, the author proposes the following
To test the hypothesis that greater quantitative metrics of
alternative hypothesis:
ESG disclosure will give rise to greater divergence of ESG ratings,
H1b: quantitative ESG disclosure is negatively related to the
the main identification model is as follows:
divergence of ESG ratings.
Whether quantitative ESG disclosure alleviates or
exacerbates rating divergence is unknown. However, when ESG_Divergencei,t = a0 + a1 ESG_Qmetricsi,t
it comes to standardized quantitative disclosure, the situation X
has become relatively clear. If companies disclose their ESG + ak Controli,t + FixedEffects + εi,t (1)
activities and performances in standardized and comparable
indicators, rating agencies are more likely to absorb the where i indexes firm, t indexes year. ESG_Divergencei,t is
commonly used indicators into their evaluation systems the standard deviation of ESG ratings of firm i in year t,
and make horizontal comparisons between companies. and ESG_Qmetricsi,t is the natural logarithm of the number
Standardized quantitative disclosure, following sustainable of quantified indicators about ESG disclosure plus one. The
reporting instruments and standards developed by regulators construction of the two variables will be explained in detail later.
or non-governmental organizations, is conducive to the rating The author predicts that the coefficient of interest α1 will be
convergence among agencies for the following reasons: (1) significantly positive.
rating agencies tend to incorporate the indicators required Controli,t contains a vector of firm-level control variables,
by reporting instruments or standards into their assessment including the average ESG rating of a firm received from
process, which greatly narrows the differences between agencies’ different rating agencies (ESG_mean), the number of rating
evaluation systems and then converge rating outcomes; (2) agencies following the firm (ESG_N), state-owned enterprise
when horizontal comparisons between companies are easily or not (State), total assets (Size), capital structure (Leverage),
implementable, different rating agencies would choose similar profitability (ROA), book-to-market ratio (BM), and Tobin’s Q
companies’ portfolio as highly rated companies. (TobinQ) (refer to detailed definitions in Appendix A). Fixed
To examine this conjecture, I use the Environmental, effects consist of firm, year, and rating agency. For the robustness
Social and Governance Reporting Guide introduced by Hong of the results, the author uses industry, year, and rating agency
Kong Exchanges (HKEX) as a setting of improvement of fixed effects. Standard errors are clustered at the firm level.
ESG data consistency. Similar to government-initiated policies Next, to examine whether standardized quantitative
on ESG in the European context (Camilleri, 2015), this disclosure mitigates rating disagreement (H2), the author
Guide requires a series of standardized key performance constructs the following DID model, taking Environmental,
indicators, and firms must report on the “comply or explain” Social and Governance Reporting Guide introduced by Hong
provisions of this Guide. Obviously, the new guidance Kong Exchanges as an exogenous shock:
greatly elevates the formalization and comparability of ESG
reports. Although there are no explicit regulations for A-
share firms listed on Shanghai and Shenzhen stock exchanges, ESG_Divergencei,t = γ0 + γ1 ESGQmetricsi,t × Di,t
a small part of these firms is cross-listed on HKEX. From + γ2 ESGQmetricsi,t + γ3 Di,t
July 2020 onward, companies cross-listed on A-shares and X
HKEX have to disclose standardized and comparable ESG + γk Controli,t + FixedEffects + τi,t (2)
performance measures following the guidance of HKEX,
while companies only listed on A-shares are not subject where Di,t is a dummy variable equal to 1 for A-share
to it. The author predicts that firms cross-listed on HKEX companies cross-listed on HKEX after July 2020 and 0 otherwise.
disclose more standardized quantitative indicators and get ESG_Divergencei,t , ESG_Qmetricsi,t , Controli,t , and fixed effects
more consistent ESG ratings from agencies. In other words, are the same as in Equation (1). γ1 is the coefficient of interest,
agencies are more likely to reach a consensus about standard and it is predicted to be significantly negative.

Liu 10.3389/fpsyg.2022.936798
TABLE 1 Rating overview using available information in the database.
Rating agency Rating score Coverage Frequency Source
SynTao Green Finance D to A+ HuShen 300 (2014–2020), CSI 500 (2018–2020) Yearly WIND database
Sino-Securities Index C to AAA A-share companies (2009–2020) Quarterly WIND database
CASVI D to AAA HuShen 300 (2015–2020) Semiyearly WIND database
WIND ESG 0 to 10 CSI 800 (2016–2020) Quarterly WIND database
FTSE Russell 0 to 5 (2017–2020) Yearly WIND database
Rankins CCC to AAA CSI 800 (2019–2020) Yearly Annual ESG rating reports
Data of research design, the author excludes: (1) firm-years with <2
ESG ratings, (2) finance companies, and (3) firm-years with
The common challenge faced by ESG research is that data missing variables. The final sample includes 1,024 companies
points of ESG ratings are few in both the cross-section and time in China and 4,966 firm-year observations. All continuous
series, and access to these data has not been sufficient. To include variables are winsorized at the upper and lower 2% levels,
as many raters as possible, the author uses Chinese ESG rating avoiding the impact of extreme outliers.
data provided by six ESG rating agencies: SynTao Green Finance,
Sino-Securities Index, CASVI, WIND ESG, FTSE Russell, and
Rankins. Most are available through the WIND database, and Descriptive statistics
a few are extracted from annual ESG rating reports. Together,
these agencies represent the major players in Chinese ESG rating Table 2 Panel A reports summary statistics and Pearson
space, which cover a substantial part of the ESG ratings market. correlations between ESG ratings from each rating agencies in
As Table 1 shows, these agencies have different frequencies the sample. After standardization ranging from 0 to 10, Sino-
in releasing updated ratings: Sino-Securities Index and WIND Securities Index tends to issue high ESG scores (average of
ESG are quarterly, CASVI is semi-yearly, and others are yearly. 7.846), while FTSE Russell and Rankins tend to issue much
Because of the annual ratings of some agencies as well as rare lower scores (average of 2.46 and 2.652). In view of the
varieties of ratings within a year, the author bases the scores different scoring tendencies of these ESG rating agencies, it is
on firm-year level. According to the information disclosure necessary to control rating agencies’ fixed effects in subsequent
requirement of China Securities Regulatory Commission estimations. The variance across firms of CASVI scores and
(CSRC), Chinese A-share listed companies generally publish Sino-Securities Index scores is relatively high, 1.438 and 1.426
the prior year’s annual financial report and CSR/ESG report specifically. It is worth noting that the mean score issued
in the first half of year. Therefore, if an agency issues more by SynTao Green Finance is around 5 (5.19) and that the
than one rating for a specific firm-year observation, the author variance across firms (standard deviation of 0.981) is least
keeps the rating that is released nearest to the middle of among the six agencies. The right side of Table 2 Panel A
the year. Moreover, to make different ratings comparable, the shows the cross-correlations of ESG ratings from different
author rescales the rating score, ranging from 0 to 10. Based agencies. All these relationships are significant at 1% level.
on the standardized data, variable ESG_Divergence is defined Consistent with the average correlation of each respective cross-
as the standard deviation of ESG ratings of a firm-year from correlation in Gibson Brandon et al. (2021) and Jørgensen
different agencies. and Ellingsen (2021) (0.447 and 0.664 respectively), most
The number of quantified indicators about firms’ ESG Pearson correlations of agency scores are between 0.4 and
disclosure is provided by China Stock Market & Accounting 0.6. The strongest relationship is the pairwise correlation
Research Database (CSMAR). CSMAR extracts substantial between WIND ESG and Rankins scores (0.668), while the
and quantified metrics related to environmental, social, and lowest correlation is between SynTao Green Finance and Sino-
governance activities and performances by scanning through Securities Index.
firms’ CSR/ESG reports. Depending on these data points, Table 2 Panel B displays the summary statistics of the
variable ESG_Qmetrics counts the number of quantified variables used in the main analyses. The mean ESG_Qmetrics
indicators about firms’ ESG disclosure. The firm characteristic is 0.847, meaning that firms disclose approximately 0.847
data used in regressions is also from the CSMAR database. (natural logarithm) substantial and quantified metrics related
The author restricts the sample to A-share firms listed on to ESG performance. These firm-years are followed by 3.556
the Shanghai and Shenzhen stock exchanges in China, and the ESG rating agencies and get a score of 5.966 on average. In
sample period goes from 2014 to 2020. To meet the requirement this sample, an average firm has a leverage of 47.4%, size

Liu 10.3389/fpsyg.2022.936798
TABLE 2 Descriptive statistics.
Panel A: Descriptive statistics and correlations for ESG ratings from each rating agency
ESG ratings N Mean Median Standard Pearson correlations
from agencies deviation
SynTao Sino- CASVI WIND FTSE
Green Securities ESG Russell
Finance Index
SynTao Green 4,277 5.190 5.000 0.981

Finance
Sino-Securities 4,966 7.846 7.780 1.426 0.262
Index
CASVI 1,972 6.523 6.625 1.438 0.404 0.372
WIND ESG 3,252 6.770 6.660 1.059 0.577 0.399 0.520
FTSE Russell 1,905 2.460 2.200 1.036 0.608 0.284 0.458 0.569
Rankins 1,288 2.652 2.860 1.276 0.595 0.544 0.591 0.668 0.581
Panel B: Descriptive statistics of regression variables

Variables N Mean Standard Deviation Percentile 25 Median Percentile 75
ESG_Divergence 4,966 2.051 0.862 1.404 2.073 2.635

ESG_Qmetrics 4,966 0.847 1.014 0.000 0.000 1.609
ESG_mean 4,966 5.966 1.040 5.250 5.984 6.737
ESG_N 4,966 3.556 1.342 2 3 5
State 4,966 0.482 0.500 0 0 1
Size 4,966 23.690 1.234 22.800 23.550 24.440
Leverage 4,966 0.474 0.193 0.324 0.485 0.622
ROA 4,966 0.055 0.064 0.019 0.044 0.086
BM 4,966 0.681 0.283 0.458 0.693 0.919
TobinQ 4,966 1.936 1.350 1.088 1.443 2.182
Panel C: Correlations of regression variables

ESG_Divergence ESG_Qmetrics ESG_mean ESG_N State Size Leverage ROA BM TobinQ
ESG_Divergence 1
ESG_Qmetrics 0.161 1
ESG_mean −0.221 0.513 1
ESG_N 0.258 0.244 −0.171 1
State 0.136 0.325 0.255 0.066 1
Size 0.182 0.432 0.236 0.378 0.395 1
Leverage 0.053 0.172 0.062 0.043 0.196 0.593 1
ROA 0.022 −0.053 0.099 0.097 −0.224 −0.225 −0.479 1
BM 0.101 0.207 0.011 0.045 0.315 0.626 0.507 −0.505 1
TobinQ −0.025 −0.149 −0.032 0.05 −0.315 −0.471 −0.439 0.505 −0.836 1
This table presents descriptive statistics of the sample. Panel A reports summary statistics and Pearson correlations of ESG ratings from each agency. Panel B shows descriptive statistics
of regression variables. Panel C presents Pearson correlations of regression variables. Bold numbers in Panel A and Panel C denote Pearson correlations significant at the 5% level. See
Appendix A for detailed definitions of the variables.
of 23.69, ROA of 0.055, book-to-market ratio of 0.681, and Table 2 Panel C shows the Pearson correlation coefficients
Tobin’s Q of 1.936. Besides, a total of 48.2% of firms are state- of the variables. ESG_Qmetrics, ESG_mean and ESG_N are
owned enterprises. positively correlated with State and Size, which suggests that

Liu 10.3389/fpsyg.2022.936798
state-owned enterprises and fairly large firms disclose more disagreement. Specifically, environmental and social issues
quantitative information about ESG performance, attract more have been debated for a shorter period than governance, and
agencies, and get higher ESG ratings. ESG_N is positively there is less of a general consensus on environmental and
correlated with ESG_Divergence, while ESG_mean is negatively social pillars (Christensen et al., 2022). Based on the previous
correlated with ESG_Divergence. The correlation between discussion, the lack of common understanding of the definition
ESG_Qmetrics and ESG_Divergence is 0.161, which provides of ESG, including what are good performances and how to
preliminary evidence for the hypothesis of the author. measure them, plays an essential role in explaining rating
divergence. Therefore, the author expects that disclosures for
environmental and social pillars should be more likely to result
Results in rating divergence.
The author disentangles ESG disclosure and divided it into
ESG numerical disclosure and ESG ratings three pillars: E (the environmental pillar), S (the social pillar),
divergence and G (the governance pillar). The author then re-estimates
Equation (1) and replace ESG_Qmetrics with ESG_Qmetrics_E,
Table 3 shows the regression results of Equation (1) using ESG_Qmetrics_S, and ESG_Qmetrics_G separately. Table 4
different specifications. Column (1) reports the result of a shows the regression results. As expected, after controlling
regression with observable firm-level characteristics, column (2) for firm characteristic variables and year-, agency- and
provides the result controlling for the same characteristics as firm-level fixed effects, the coefficients on ESG_Qmetrics_E
well as ESG rating agency, industry, and year fixed effects, while and ESG_Qmetrics_S are significantly positive, whereas the
column (3) includes firm characteristics and fixed effects of coefficient on ESG_Qmetrics_G is insignificant. These findings
ESG rating agency, firm, and year to control for unobservable imply that disclosures on environmental and social issues
factors in agency-, firm- and year-levels. The coefficient of contribute more to ESG rating divergence.
ESG_Qmetrics is the coefficient of interest. The point estimates
in columns (1) to (3) are 0.242, 0.108, and 0.061 respectively,
statistically significant at the 1% level. Given that the sample
standard deviation of ESG_Divergence is 0.862, the economic Standardized quantitative ESG disclosure
magnitude of the effect is substantial. Overall, these findings and ESG ratings divergence
corroborate H1a that if a firm discloses more non-standardized
numerical information about its ESG performance, it will receive Section ESG numerical disclosure and ESG ratings
more divergent ESG scores from rating agencies. Quantified divergence shows that the more non-standardized numerical
indicators in ESG disclosure pose a great challenge for rating information about ESG performance a firm discloses, the more
agencies to assess the performance and compare with peers. divergent ESG ratings it gets from agencies. In this section,
With regard to control variables, the coefficient of the author explores what could happen if companies disclose
ESG_mean is negative and statistically significant across all quantitative indicators following the reporting standard.
the model specifications reported in Table 3, suggesting that The underlying argument of the main conclusion is that
ESG rating agencies tend to split on firms with poor multifarious quantified disclosure provides rating agencies with
ESG performance. The estimation coefficient on ESG_N is more information to interpret and evaluate the performance,
significantly positive in column (3) (coefficient = 0.393, which leads to a different judgment. For example, firm A’s
t = 7.5), which is consistent with the intuition and the number of accidents with fatal consequences is 1 and firm B’s
Pearson correlation test. When more ratings are available, injury rate is 5%. Since the two metrics are not necessarily the
the disagreement tends to be more considerable. In addition, same thing, agency C may choose firm A as a better performer
two firm characteristics play an obvious role in explaining on health and safety, while agency D may choose the other.
ESG ratings dispersion: the point estimates of State and Size Besides, agency C probably chooses the number of accidents
are positive and almost significant. State-owned firms exhibit with fatal consequences as one of the key indicators to evaluate
higher ESG rating disagreement than non-state-owned firms firms’ ESG performance, while agency D does not. In this case,
because they usually take social responsibility associated with as mentioned in H2, standard ESG data will narrow this kind of
national strategic goals such as targeted poverty alleviation differences markedly.
and employment creation, and agencies may allocate different Table 5 reports the regression results of a DID estimation
weights in these aspects. Also, ESG activities in larger firms are using Equation (2). As shown in columns (1) and (2), the
personalized, and it is hard to find comparable companies, which point estimates of ESG_Qmetrics are positive and significant
may lead to radically different ratings from agencies. at the 1% level, consistent with the main results. Moreover,
Furthermore, the author examines whether disclosure the coefficient on the interaction term ESG_Qmetrics∗ D is
for some pillars (E, S, and G) contributes more to rating significantly negative, which indicates that if ESG disclosure

Liu 10.3389/fpsyg.2022.936798
TABLE 3 Effect of ESG quantitative disclosure on ESG rating divergence.
Dependent variable (1) (2) (3)

ESG_Divergence ESG_Divergence ESG_Divergence
ESG_Qmetrics 0.242*** 0.108*** 0.061***

(10.58) (5.05) (3.10)
ESG_mean −0.361*** −0.088*** −0.063*
(-14.05) (−3.27) (−1.82)
ESG_N 0.024 −0.115** 0.393***
(1.38) (−2.01) (7.50)
State 0.215*** 0.216*** 0.116
(5.23) (5.83) (1.17)
Size 0.102*** 0.085*** 0.159***
(4.02) (3.49) (3.36)
Leverage −0.067 −0.227** −0.336**
(−0.53) (−2.05) (−2.06)
BM 0.071 −0.199* −0.289***
(0.62) (−1.93) (−2.71)
TobinQ 0.038* 0.002 −0.004
(1.95) (0.11) (−0.21)
ROA 1.453*** 1.212*** 0.365
(4.81) (4.65) (1.36)
Constant 1.221** 0.888 −1.421
(2.47) (1.64) (−1.40)
Year FE No Yes Yes
Agency FE No Yes Yes
Industry FE No Yes No
Firm FE No No Yes
Observations 4,966 4,966 4,966
R2 0.198 0.467 0.487
This table presents the regression results of Equation (1) using different specifications. The dependent variable ESG_Divergence is the standard deviation of ESG ratings. In column (1),
the author regresses ESG_Divergence on ESG_Qmetrics, and firm-level control variables. In column (2), the author adds agency, industry, and year fixed effects. In column (3), the author
replaces industry fixed effects with firm fixed effects. Definitions of the variables are provided in Appendix A. The standard errors are clustered at the firm level. Values of t-statistics are in
parentheses. Statistical significance at the 1, 5, and 10% levels is indicated by ***, **, and *, respectively. Bold text denotes the key variables and their estimated coefficients.
is standardized and comparable, more numerical information of consensus on a firm’s ESG rating: a positive ESG activity is
reduces agencies’ rating disagreement instead. The results highly rated by an agency and undervalued by another or a
suggest that due to the inconsistency of disclosure, more negative ESG activity is identified as an insignificant thing by an
quantitative metrics of ESG disclosure exacerbate disagreement agency and a warning sign by another. Material and sustainable
across ESG rating agencies. ESG activities do not usually lead to disagreement, while ESG
issues with potential risks are more likely to bring about different
interpretations from rating agencies. If potential risks do exist,
Consequences: ESG rating in the future firms with ESG rating disagreement will perform worse in the
following years on average. To test this conjecture, the author
Gibson Brandon et al. (2021) and Avramov et al. (2022) constructs Equation (3) as follows:
find that ESG rating disagreement is positively associated
with cost of equity capital. They conjecture that higher ESG
rating disagreement is perceived as a source of uncertainty,
which requires an uncertainty premium. In this section, the
author examines whether ESG performance in the future would ESG_Meani,t + 1 = β0 + β1 ESG_Divergencei,t
embody the uncertainty risks implied by rating disagreement. X
+ βk Controli,t + FixedEffects + µi,t (3)
There are two possible circumstances to account for the lack

Liu 10.3389/fpsyg.2022.936798
TABLE 4 Effect of E/S/G pillar-specific quantitative disclosure on ESG ratings divergence.
Dependent variable (1) (2) (3)

ESG_Divergence ESG_Divergence ESG_Divergence
ESG_Qmetrics_E 0.058*
(1.96)
ESG_Qmetrics_S 0.057**
(2.51)
ESG_Qmetrics_G 0.048
(1.06)
ESG_mean −0.033 −0.048* −0.022
(−1.43) (−1.87) (−1.00)
ESG_N −0.093 −0.099* −0.090
(−1.62) (−1.72) (−1.56)
State 0.230*** 0.222*** 0.231***
(6.19) (5.96) (6.22)
Size 0.092*** 0.091*** 0.094***
(3.71) (3.68) (3.80)
Leverage −0.218* −0.224** −0.219**
(−1.95) (−2.02) (−1.97)
BM −0.203* −0.193* −0.201*
(−1.93) (−1.85) (−1.91)
TobinQ 0.001 0.001 0.000
(0.04) (0.07) (0.02)
ROA 1.120*** 1.145*** 1.102***
(4.27) (4.38) (4.22)
Constant 0.197 0.321 0.086
(0.40) (0.65) (0.17)
Year FE Yes Yes Yes
Agency FE Yes Yes Yes
Industry FE Yes Yes Yes
Observations 4,966 4,966 4,966
R2 0.452 0.453 0.451
This table presents the regression results of the effect of environmental pillar, social pillar, and governance pillar disclosure on rating disagreement. The author re-estimates Equation (1)
and uses ESG_Qmetrics_E, ESG_Qmetrics_S, and ESG_Qmetrics_G as the key independent variables in columns (1) to (3), respectively. ESG_Qmetrics_E is the natural logarithm of the
number of quantified indicators for environmental pillar about firm i’s ESG disclosure plus one. ESG_Qmetrics_S is the natural logarithm of the number of quantified indicators for social
pillar about firm i’s ESG disclosure plus one. ESG_Qmetrics_G is the natural logarithm of the number of quantified indicators for governance pillar about firm i’s ESG disclosure plus one.
Firm-level characteristic variables, as well as agency, industry, and year fixed effects are controlled. The definitions of the variables are provided in Appendix A. The standard errors are
clustered at the firm level. The values of t-statistics are in parentheses. Statistical significance at the 1, 5, and 10% levels is indicated by ***, **, and *, respectively. Bold text denotes the key
variables and their estimated coefficients.
The dependent variable is ESG_mean for firm i in year t + 1, and Cross-sectional analyses
the independent variable is ESG_Divergence for firm i in year
t. Controls contain a vector of firm characteristics, including In Section Cross-sectional analyses, cross-sectional analyses
ESG_N, ESG_Qmetrics, State, Size, Leverage, BM, TobinQ, and are conducted to offer additional evidence on the main
ROA (refer to detailed definitions in Appendix A). Table 6 shows hypothesis. H1a suggests that if a firm discloses more non-
the regression results using two specifications: (1) with year, standardized numerical information about its ESG performance,
agency, and industry fixed effects and (2) with year, agency, and agencies are more likely to use their own expertise and put
firm fixed effects. The negative coefficients of ESG_Divergence their personalized interpretations and judgments on these pieces
in columns (1) and (2) indicate that firms with great ESG rating of information, thereby leading to more rating disagreement.
divergence will experience considerably lower ESG ratings in the Therefore, the author predicts that numerical disclosure of
future, providing further evidence of risk premium being related firms that concentrate on a specific industry would generate
to ESG uncertainty. more disagreement. Compared with diversified companies,

Liu 10.3389/fpsyg.2022.936798
TABLE 5 Effect of standardized quantitative ESG disclosure on ratings divergence: using the shock of disclosure guidance in HKEX.
Dependent variable (1) (2)

ESG_Divergence ESG_Divergence
Coefficient t-statistic Coefficient t-statistic
ESG_Qmetrics*D −0.156*** (−2.62) −0.137** (−2.02)

ESG_Qmetrics 0.092*** (5.02) 0.065*** (3.32)
D 0.088 (0.57) 0.172 (0.98)
ESG_mean −0.081*** (−2.94) −0.060* (−1.73)
ESG_N 0.119** (2.51) 0.392*** (7.50)
State 0.190*** (4.98) 0.110 (1.11)
Size 0.129*** (5.24) 0.155*** (3.27)
Leverage −0.319*** (−2.91) −0.344** (−2.10)
BM −0.330*** (−3.51) −0.291*** (−2.74)
TobinQ −0.003 (−0.18) −0.005 (−0.28)
ROA 0.925*** (3.79) 0.370 (1.38)
Constant −0.524 (−0.96) −1.338 (−1.32)
Year FE Yes Yes
Agency FE Yes Yes
Industry FE Yes No
Firm FE No Yes
Observations 4,966 4,966
R2 0.476 0.488
This table presents the regression results of the mechanism. The dependent variable ESG_Divergence is the standard deviation of ESG ratings. Based on Equation (2), the estimations
include ESG_Qmetrics, D, and the interaction term ESG_Qmetrics* D. D is equal to 1 for cross-listed companies on A-shares and HKEX after July 2020 and 0 otherwise. In column (1), the
author controls for firm-level variables, as well as year, agency, and industry fixed effects. In column (2), the author replaces industry fixed effects with firm fixed effects. The definitions
of the variables are provided in Appendix A. The standard errors are clustered at the firm level. Statistical significance at the 1, 5, and 10% levels is indicated by ***, **, and *, respectively.
Bold text denotes the key variables and their estimated coefficients.
those companies may disclose more industry-specific ESG rating divergence in single business companies rather than in
information, which exacerbates disagreement among raters. diversified business companies.
Taking SynTao Green Finance for example, its evaluation Next, the author examines whether the rating dispersion
system consists of general indicators and industry-specific stemming from numerical ESG information is heterogenous
indicators. General indicators are relatively easy to agree among among firms with different ESG performances. Rating agencies
raters, where as industry-specific indicators depend entirely on are easier to reach an agreement on positive ESG events rather
the rating agency’s own expertise and therefore vary widely than negative ESG events, especially if the overall level of ESG
among raters. performance is poor. Different rating agencies are likely to
The author uses the variable Diversification, defined as identify the same negative ESG event of a firm to be different
the Herfindahl Index of operating income in a company (the levels of severity. For instance, CASVI will evaluate a firm
quadratic sum of the ratio between income of business k with a major violation of laws or regulations as “D” level (the
and total income), to divide the full sample into diversified lowest level) directly, while Rankins may lower the grade to
business companies (high-Diversification group) and single some extent. Based on variable ESG_mean, the author partitions
business companies (low-Diversification group). The columns the full sample into two subgroups: firms with better ESG
(1) and (2) in Table 7 report the regression results for performance than average (high-ESG_mean group) and firms
diversified business companies and single business companies, with poorer ESG performance than average (low-ESG_mean
respectively. The coefficients of ESG_Qmetrics are 0.051 in the group). The columns (3) and (4) in Table 7 show the estimation
diversified business group and 0.144 in the single business results. As predicted, the coefficients of ESG_Qmetrics are 0.025
group, and the difference of the two coefficients is significant in the better ESG performance group and 0.087 in the poorer
at the 5% level. Consistent with the conjecture of the author, ESG performance group, respectively. The difference between
non-standardized numerical ESG disclosure causes greater the two coefficients is significant, suggesting that agencies tend

Liu 10.3389/fpsyg.2022.936798
TABLE 6 Additional analyses: the consequence of ESG rating divergence on future ratings.
Dependent variable (1) (2)

F.ESG_mean F.ESG_mean
Coefficient t-statistic Coefficient t-statistic
ESG_Divergence −0.078*** (−3.26) −0.041* (−1.87)

ESG_N −0.291*** (−5.91) −0.210*** (−4.61)
ESG_Qmetrics 0.472*** (24.25) 0.145*** (6.94)
State 0.173*** (4.16) −0.116 (−0.91)
Size 0.067** (2.22) −0.085 (−1.64)
Leverage −0.123 (−0.93) 0.064 (0.37)
BM −0.285** (−2.16) −0.161 (−1.25)
TobinQ −0.038* (−1.74) −0.035 (−1.64)
ROA 1.432*** (4.74) 0.661** (2.36)
Constant 4.726*** (6.52) 8.797*** (7.38)
Year FE Yes Yes
Agency FE Yes Yes
Industry FE Yes No
Firm FE No Yes
Observations 3,912 3,912
R2 0.577 0.480
This table presents the regression results of the consequences. The dependent variable ESG_mean in the next period is the average ESG rating of a firm received from different rating
agencies. The independent variable ESG_Divergence is the standard deviation of ESG ratings. In column (1), the author controls for firm-level variables, as well as year, agency, and industry
fixed effects. In column (2), the author replaces industry fixed effects with firm fixed effects. The definitions of the control variables are provided in Appendix A. The standard errors are
clustered at the firm level. Statistical significance at the 1, 5, and 10% levels is indicated by ***, **, and *, respectively. Bold text denotes the key variables and their estimated coefficients.
to disagree more on the quantified ESG disclosure of firms with other sectors companies. In this test, the author extended
poor ESG performance. the sample for regression estimation to all industries.
In Table 8, column (3) reports that the coefficient of
ESG_Qmetrics is 0.062 and significant at the 1% level,
Robustness test providing robust evidence in favor of the hypothesis of
author. Additionally, the untabulated results of the DID
The author performs a number of robustness tests in this model (in Section Standardized quantitative ESG disclosure
section. First of all, Equation (1) is re-estimated with alternative and ESG ratings divergence) and Equation (3) (in Section
measures of ESG rating divergence. In the previous tests, the Consequences: ESG rating in future) also corroborate the
author has employed ESG_Divergence, defined as the standard earlier findings.
deviation of ESG ratings of firm i in year t, to proxy for ESG There may be correlations between a given individual’s
rating disagreement. Following prior literature (Christensen model errors in different periods in a panel data (Colin Cameron
et al., 2022), the author constructed the alternative variable and Miller, 2015). To mitigate the potential heteroscedasticity
Divergence1 as the average of absolute values of the difference and sequence-related problems, the author replicates the main
between pairs of ratings that firm I receives from rating agencies test by two-way clustering (both at firm and year levels) and
for its ESG performance in year t, and Divergence2 as the the result in Table 8, column (4), is similar. The coefficient
coefficient of variation of firm i’s ESG ratings in year t. In Table 8, of ESG_Qmetrics remains significantly positive and, again,
columns (1) and (2) show the regression results using alternative supports the conclusions of the author.
dependent variables. The coefficients of ESG_Qmetrics are In the fourth robustness test, the author controls for
still significantly positive, indicating that the main finding potential omitted factors that affect ESG rating divergence.
is robust. Although firm fixed effects and year fixed effects are controlled,
Next, the main analysis is repeated using a redefined unobservable factors may affect rating divergence differently.
sample. Earlier in the article, the author used the sample First, the author controls for whether the ESG report or CSR
excluding observations operating in financial service industry report is verified by a third party (Verification). Southworth
on account of intrinsic differences between financial and (2009) studied corporate voluntary action as a mechanism for

Liu 10.3389/fpsyg.2022.936798
TABLE 7 Heterogeneous effects of ESG quantitative disclosure on ESG ratings divergence: role of diversified operation and ESG performances.
Dependent variable: ESG_Divergence Diversification ESG_mean

High Low High Low
(1) (2) (3) (4)
ESG_Qmetrics 0.051* 0.144*** 0.025 0.087***

(1.95) (4.99) (1.38) (3.32)
Diff. (p-value) in ESG_Qmetrics 0.041 0.088
ESG _mean −0.067** −0.103***
(−2.29) (−3.43)
ESG_N 0.335*** 0.178*** 0.438*** 0.032
(5.55) (3.01) (10.13) (0.44)
State −0.011 0.300* −0.126 0.257**
(−0.07) (1.85) (−0.92) (2.47)
Size 0.182*** 0.147** 0.044 0.198***
(2.97) (2.09) (0.92) (3.86)
Leverage −0.299 −0.075 −0.110 −0.486***
(−1.35) (−0.31) (−0.60) (−2.62)
BM −0.283* −0.473*** −0.086 −0.333**
(−1.72) (−2.80) (−0.68) (−2.40)
TobinQ −0.022 −0.000 0.009 −0.034
(−0.64) (−0.01) (0.43) (−1.58)
ROA 0.582 0.075 0.174 0.570**
(1.60) (0.19) (0.51) (2.08)
Constant −1.810 −0.657 0.823 −2.371**
(−1.32) (−0.41) (0.76) (−2.08)
Year FE Yes Yes Yes Yes
Agency FE Yes Yes Yes Yes
Firm FE Yes Yes Yes Yes
Observations 1,933 1,926 2,487 2,479
R2 0.491 0.506 0.512 0.411
This table presents the regression results of cross-sectional analyses based on Equation (1). Columns (1) and (2) report the estimations of high-Diversification group and low-Diversification
group, and columns (3) and (4) report the estimations of high-ESG_mean group and low- ESG_mean group. Variable Diversification is defined as the Herfindahl Index of operating income
in a company (the quadratic sum of the ratio between income of business k and total income). The definitions of other variables are provided in Appendix A. The standard errors are
clustered at the firm level. The values of t-statistics are in parentheses. Statistical significance at the 1, 5, and 10% levels is indicated by ***, **, and *, respectively. Bold text denotes the key
variables and their estimated coefficients.
addressing the related problems of climate change and energy disclosure and voluntary ESG disclosure differently. Willing
security and found that third-party verification of emission is a dummy variable that is equal to 1 if a firm discloses
data contributes to valuable comparisons across industries and ESG/CSR report under the requirements and zero otherwise.
companies. Accordingly, it is reasonable to speculate that third- Finally, the author further controls for the effect of institutional
party verification of a firm’s ESG disclosure would, to some ownership (Inst). Institutional investors play a prominent role in
extent, narrow rating differences and trigger convergence of shaping ESG performance, and firms with greater institutional
agencies’ ratings. The author constructs an indicator variable, ownership tend to have lower ESG rating divergence (Dyck
Verification, defined as 1 if the firm’s ESG report or CSR et al., 2019; Kim et al., 2019; Christensen et al., 2022). Following
report is verified by a third party and 0 otherwise. Second, previous studies, Inst is measured as the percentage of the
the author controls for voluntary or mandatory ESG disclosure firm’s shares owned by institutional investors. In Table 8, column
(Willing). In December 2008, the Shanghai and Shenzhen Stock (5) reports the regression result after controlling Verification,
Exchange announced regulations that obligated a subset of listed Willing, and Inst. The coefficient of ESG_Qmetrics is positive
companies to file CSR reports along with their annual reports. and significant at the 1% level after controlling these variables,
Rating agencies may conduct an assessment of mandatory ESG confirming the main conclusion.

Liu 10.3389/fpsyg.2022.936798
TABLE 8 Robustness checks: using alternative measures, alternative sample, and alternative specifications.
Dependent variable (1) (2) (3) (4) (5)

Divergence1 Divergence2 ESG_Divergence ESG_Divergence ESG_Divergence
ESG_Qmetrics 0.075*** 0.009*** 0.062*** 0.061** 0.050***

(2.88) (2.63) (3.26) (2.51) (2.59)
ESG _mean −0.047 −0.073*** −0.095*** −0.063 −0.074**
(−0.98) (-11.55) (−2.99) (−1.53) (−2.10)
ESG_N 0.380*** 0.053*** 0.410*** 0.393*** 0.390***
(5.10) (5.75) (8.06) (4.23) (7.46)
State 0.176 0.029 0.078 0.116 0.130
(1.36) (1.50) (0.78) (1.29) (1.32)
Size 0.199*** 0.032*** 0.177*** 0.159** 0.124***
(3.07) (3.74) (4.05) (2.66) (2.63)
Leverage −0.367* −0.066** −0.317** −0.336* −0.321**
(−1.66) (−2.14) (−2.00) (−2.04) (−1.98)
BM −0.357** −0.049** −0.273*** −0.289** −0.240**
(−2.49) (−2.56) (−2.58) (−3.29) (−2.23)
TobinQ −0.000 0.000 0.002 −0.004 −0.005
(−0.02) (0.06) (0.14) (−0.25) (−0.26)
ROA 0.550 0.047 0.442 0.365 0.376
(1.53) (0.96) (1.60) (1.16) (1.43)
Verification −0.051
(−0.63)
Willing 0.226***
(5.19)
Inst −0.001
(−0.53)
Constant −1.386 0.030 −1.675* −1.798 −0.682
(−1.00) (0.16) (−1.76) (−1.46) (−0.67)
Year FE Yes Yes Yes Yes Yes
Agency FE Yes Yes Yes Yes Yes
Firm FE Yes Yes Yes Yes Yes
Observations 4,966 4,966 5,389 4,936 4,961
R2 0.428 0.645 0.459 0.759 0.493
This table presents a battery of robustness checks. Columns (1) and (2) report the regression results using alternative dependent variables (Divergence1 and Divergence2), where Divergence1
is measured as the average of absolute values of the difference between pairs of ratings that firm i receives from rating agencies for its ESG performance in year t, and Divergence2 is measured
as the coefficient of variation of firm i’s ESG ratings in year t. Column (3) presents the results when the sample extends to all sectors, and column (4) presents the results from the estimation
of Equation (1) by clustering firm and year. Column (5) reports the estimation of the results after controlling for a few additional control variables. The definitions of these variables are
provided in Appendix A. The standard errors are clustered at the firm level. The values of t-statistics are in parentheses. Statistical significance at the 1, 5, and 10% levels is indicated by
***, **, and *, respectively. Bold text denotes the key variables and their estimated coefficients.
Discussion ESG, FTSE Russell, and Rankins. The empirical results show
that the divergence of ESG ratings is significantly positive
Using relevant data of the Shanghai and Shenzhen A- with non-standardized numerical ESG disclosure, especially
share non-financial listed companies from 2014 to 2020, with disclosure on environmental and social issues. This
this study examines whether quantitative ESG disclosure is relationship remains unchanged after a series of robustness tests.
conducive to rating convergence among agencies. In particular, To mitigate the potential endogenous problem and examine
to include as many raters as possible, the author uses Chinese whether standardized quantitative disclosure helps to reduce
ESG rating data provided by six ESG rating agencies, which rating disagreement, the author employs the quasi-natural
represent the major players in Chinese ESG rating market: experiment brought by the implementation of Environmental,
SynTao Green Finance, Sino-Securities Index, CASVI, WIND Social and Governance Reporting Guide introduced by Hong

Liu 10.3389/fpsyg.2022.936798
Kong Exchanges (HKEX) and conduct a DID estimation. The adopt the same approach to assess firms’ ESG performance,
results suggest that if ESG disclosure is standardized and rating divergence could come only from different information
comparable, more numerical information reduces agencies’ sets such as raters’ private information about firms. In this
rating disagreement instead. Additionally, the lack of agreement context, greater public disclosure, which means less private
is associated with a low rating in the future, indicating that information, would lead to higher rating consistency. However,
disagreement represents a source of uncertainty on potential the empirical evidence in this study documents a strong
risks. Further analyses show that the impact of quantified positive relationship between quantitative ESG disclosure and
ESG disclosure on rating disagreement is more pronounced divergence of ESG ratings, revealing that rating agencies use
when firms develop a single business rather than a diversified their own expertise and put their personalized interpretations
business and experience poor ESG performance rather than and judgments on firms’ disclosure. This finding is supported
good ESG performance. by previous studies (Chatterji et al., 2016; Berg et al., 2022).
This article draws on Christensen et al. (2022)’s research Chatterji et al. (2016) demonstrate convincingly that raters
methodology, but there are several differences in the main apply diverse theories of social responsibility, and select various
identification model that are noteworthy. First, they use ESG topics (lack of common theorization) and metrics (lack of
disclosure scores provided by Bloomberg to capture the level of commensurability) in their evaluation framework. Berg et al.
firms’ ESG disclosure, ranging from 0.1 to 100. The disclosure (2022) find that rating disagreement stems from different
score is higher if a company discloses more data points sets of attributes (scope divergence), different indicators
that Bloomberg collects. In this study, the author focuses on measuring the same attribute (measurement divergence), and
quantitative disclosure rather than overall disclosure. Therefore, different views on the relative importance of the attributes
the author uses the data of quantified indicators about firms’ ESG (weights divergence).
disclosure provided by the CSMAR database. CSMAR extracts The DID tests (Section Standardized quantitative ESG
substantial and quantified metrics related to environmental, disclosure and ESG ratings divergence) and cross-sectional
social, and governance activities and performances by scanning analyses (Section Cross-sectional analyses) in this article
through firms’ CSR/ESG reports. Depending on these data add further evidence that the lack of common norms
points, the variable ESG_Qmetrics counts the number of for interpreting ESG disclosure brings about the lack of
quantified indicators about firms’ ESG disclosure. The difference agreement across the information intermediaries. Employing
between the research questions determines that the author uses the quasi-natural experiment brought by HKEX, this study
the different data and variable to measure the ESG disclosure finds that despite more quantitative disclosures after the new
than they did. Second, Christensen et al. (2022) obtained guidance, the divergence has been alleviated significantly,
ESG ratings from three agencies, namely, Morgan Stanley because standardized numerical disclosures required by the
Capital International’s IntangibleValue Assessment, Thomson guidance are more comprehensible and comparable. Next,
Reuters’ ASSET4, and Sustainalytics, because these agencies the regression results show that the numerical disclosure of
publicly released ESG ratings and provided international data firms with a single business rather than a diversified business
that covered their sample. In this study, the sample consists generates more disagreement. Taking SynTao Green Finance for
of Chinese A-share public firms, so the rating data available instance, its evaluation system consists of general and industry-
changed. The author uses ESG ratings provided by six raters, specific indicators. General indicators are appropriate for all
SynTao Green Finance, Sino-Securities Index, CASVI, WIND companies and generally regarded as relatively established fields
ESG, FTSE Russell, and Rankins, who together represent with little controversy, while industry-specific indicators are
the major players in Chinese ESG rating space. Since rating more complex and entirely dependent on agencies’ industry
divergence is the interest, we both require observations to have expertise and thus vary widely. Firms developing a single
ratings from more than one rater. It means that the number business are more likely to disclose information with industry
of ratings of a firm-year in Christensen et al. (2022) is 2 or 3, specialization and therefore undertake more rating divergence.
while the number ranges from 2 to 6 in this study. Compared In addition, the author finds that agencies tend to disagree
with them, the author adds the number of ratings available to more on the quantified ESG disclosure of firms with poor
control variables in order to control for their effect on rating ESG performance rather than good ESG performance. There
disagreement. Third, because of the higher potential influence could be two circumstances with a poor ESG performance
of the number of ratings in this study than that in Christensen firm: low degree of disclosure and negative ESG events.
et al. (2022), the author conducts some tests that re-estimate the Although no disclosure almost leads to a general consensus,
main model when the number of ratings is the same or close. The negative events are perceived as different levels of severity.
untabulated results show that the effect of quantitative disclosure For instance, CASVI will evaluate a firm with a major
on rating divergence does not change. violation of laws or regulations as “D” level (the lowest
This study provides further empirical evidence of the level) directly, while Rankins may lower the grade to some
presence and determinants of rating disagreement. If raters extent. Overall, these results all validate H1a that quantitative

Liu 10.3389/fpsyg.2022.936798
ESG disclosure is positively related to divergence of ESG numerical information, indicating that low commensurability
ratings because of different interpretations and judgments still poses a serious challenge in newly emergent areas.
of information. Finally, this study has implications for research on economic
consequences of ESG rating disagreement. Gibson Brandon
et al. (2021) and Avramov et al. (2022) find that firms with
Conclusions high ESG rating disagreement require an equity premium
because disagreement is perceived as a source of uncertainty.
The use of ESG ratings in academic research This study complements some evidence and demonstrates that
and investment practice has continued to accelerate divergence is associated with a lower average rating in the
recently. In the meantime, the issue of ESG rating future, indicating that there would be potential risks with
disagreement has also spurred a considerable concern in those firms.
the financial press, investors, and researchers. Obviously, This study not only contributes to the theoretical debate on
it is warranted to examine the validity of ESG ratings why ESG raters disagree but also has important implications
and explore the dynamics driving convergence among in practice. Researchers should carefully assess and explain
information intermediaries. data analyses based on ESG ratings, especially if using
Using rating data from six ESG rating providers in China, one particular agency’s rating. If possible, use multiple
the author examines whether quantitative ESG disclosure is measures of ratings to proxy ESG performance as robustness
conducive to rating convergence among agencies. The results tests to enhance the confidence of these academic studies.
show that greater quantified ESG disclosure brings about What is more, this study also calls for closer cooperation
greater divergence of ESG ratings. Specifically, quantitative among companies, rating agencies, and policymakers to
disclosure on environmental and social issues plays a greater build a consensus on crucial ESG performance and establish
role compared with governance disclosure in explaining consistent and transparent disclosure standards. Meanwhile,
rating disagreement. rating agencies should become more transparent about their
To mitigate the potential endogenous problem and examine rating methodologies and valuation systems. Improvement in
whether standardized quantitative disclosure helps to reduce transparency will help rating users to locate the source of rating
the rating disagreement, the author uses a shock of the disagreement and evaluate whether the disagreement affects
implementation of Environmental, Social and Governance their decisions.
Reporting Guide introduced by Hong Kong Exchanges This study has several limitations that should be noted.
(HKEX) and conducts a DID estimation. The estimation First, this study uses the number of materials and quantified
results support the conjecture. In addition, further analyses indicators in the ESG or CSR report to measure the degree
show that the lack of agreement is associated with a low of a firm’s ESG disclosure. As such, future studies may
rating in the future, indicating that disagreement represents take account of disclosure from other sources, including
a source of uncertainty on potential risks. The author also company websites, news reports, and penalty information
finds that the effect of quantified ESG disclosure on rating from regulatory agencies. Second, restricted to the relatively
dive is more significant when firms are single businesses small dataset, the study focuses on ESG rating divergence
rather than diversified businesses with poor ESG performance in all non-financial industries. Future research could
rather than good ESG performance. Lastly, the main results examine the industry heterogeneity of ESG disagreement,
are robust to alternative measures of ESG rating divergence, i.e., the tobacco, mining, and weapons industries are worthy
alternative sample, two-way clustering, and additional of attention.
control variables.
This study responds to the heated debate on ESG rating
divergence in the following aspects. First of all, it enriches Data availability statement
the academic evidence of ESG rating disagreement. It should
be noted that this study is related to Christensen et al. The original contributions presented in the
(2022) with the conclusion that rating divergence is larger study are included in the article/supplementary
when firms disclose more. Specifically, this study differs material, further inquiries can be directed to the
from theirs in that the author explores how quantitative corresponding author/s.
disclosure influences the evaluation results of rating agencies
about ESG issues. Next, this study complements the literature
related to sociology of evaluation. Although quantification Author contributions
enhances comprehensibility and comparability (Porter, 1995;
Espeland and Stevens, 1998, 2008), rating disagreement is The author confirms being the sole contributor of this study
more pronounced when firms disclose more non-standardized and has approved it for publication.

Liu 10.3389/fpsyg.2022.936798
Acknowledgments that could be construed as a potential conflict

of interest.
The author acknowledges the administrative and financial
support from School of Business, Renmin University
of China. Publisher’s note
All claims expressed in this article are solely those of the
authors and do not necessarily represent those of their affiliated
Conflict of interest organizations, or those of the publisher, the editors and the
reviewers. Any product that may be evaluated in this article, or
The author declares that the research was conducted in claim that may be made by its manufacturer, is not guaranteed
the absence of any commercial or financial relationships or endorsed by the publisher.
References
Akins, B. (2018). Financial reporting quality and uncertainty about credit risk Espeland, W. N., and Sauder, M. (2007). Rankings and reactivity: How
among ratings agencies. Account. Rev. 93, 1–22. doi: 10.2308/accr-51944 public measures recreate social worlds. Am. J. Sociol. 113, 1–40. doi: 10.1086/
517897
Avramov, D., Cheng, S., Lioui, A., and Tarelli, A. (2022). Sustainable
investing with ESG rating uncertainty. J. Finan. Econ. 145, 642–664 Espeland, W. N., and Stevens, M. L. (1998). Commensuration as a social process.
doi: 10.1016/j.jfineco.2021.09.009 Annu. Rev. Sociol. 24, 313–343. doi: 10.1146/annurev.soc.24.1.313
Berg, F., Koelbel, J. F., and Rigobon, R. (2022). Aggregate confusion: the Espeland, W. N., and Stevens, M. L. (2008). A sociology of quantification. Arch.
divergence of ESG rating. Rev. Financ. 26, 1–30. doi: 10.1093/rof/rfac033 Eur. Sociol. 49, 401–436. doi: 10.1017/S0003975609000150
Billio, M., Costola, M., Hristova, I., Latino, C., and Pelizzon, L. (2021). Inside Gibson Brandon, R., Krueger, P., and Schmidt, P. S. (2021). ESG
the ESG ratings:(Dis) agreement and performance. Corp. Soc. Resp. Env. Ma. 28, rating disagreement and stock returns. Financ. Anal. J. 77, 104–127.
1426–1445. doi: 10.1002/csr.2177 doi: 10.1080/0015198X.2021.1963186
Bonsall, S. B., and Miller, B. P. (2017). The impact of narrative disclosure Grewal, J., and Serafeim, G. (2020). Research on corporate sustainability:
readability on bond ratings and the cost of debt. Rev. Acc. Stud. 22, 608–643. review and directions for future research. Found. Trends. Account. 14, 73–127.
doi: 10.1007/s11142-017-9388-0 doi: 10.1561/1400000061
Broadstock, D. C., Chan, K., Cheng, L. T. W., and Wang, X. (2021). The role GSIR. (2020). Global Sustainable Investment Review 2020.
of ESG performance during times of financial crisis: evidence from COVID-19 in Global Sustainable Investment Alliance. Available online at:
China. Financ. Res. Lett. 38, 101716. doi: 10.1016/j.frl.2020.101716 www.gsi-alliance.org/wp-content/uploads/2021/08/GSIR-20201.pdf
Camilleri, M. A. (2015). Environmental, social and governance disclosures in Hartzmark, S. M., and Sussman, A. B. (2019). Do investors value sustainability? A
Europe.Sustain. Account. Mana. 6, 224–242. doi: 10.1108/SAMPJ-10-2014-0065 natural experiment examining ranking and fund flows. J. Finance. 74, 2789–2837.
doi: 10.1111/jofi.12841
Camilleri, M. A. (2018). Theoretical insights on integrated reporting: the
inclusion of non-financial capitals in corporate disclosures. Corp. Commun. 23, Hong, H., and Stein, J. C. (2007). Disagreement and the stock market. J. Econ.
567–581. doi: 10.1108/CCIJ-01-2018-0016 Perspect. 21, 109–128. doi: 10.1257/jep.21.2.109
Cao, J., Liang, H., and Zhan, X. (2019). Peer effects of corporate social Hope, O-. K. (2003). Accounting policy disclosures and analysts’ forecasts.
responsibility. Manage. Sci. 65, 5487–5503. doi: 10.1287/mnsc.2018.3100 Contemp. Account. Res. 20, 295–321. doi: 10.1506/LA87-D1NF-BF06-FW1B
Chatterji, A. K., Durand, R., Levine, D. I., and Touboul, S. (2016). Do ratings Hsu, G., Roberts, P. W., and Swaminathan, A. (2012). Evaluative schemas
of firms converge? Implications for managers, investors and strategy researchers. and the mediating role of critics. Organ. Sci. 23, 83–97. doi: 10.1287/orsc.1100.
Strategic. Manage. J. 37, 1597–1614. doi: 10.1002/smj.2407 0630
Cheng, B., Ioannou, I., and Serafeim, G. (2014). Corporate social responsibility Hubbard, T. D., Christensen, D. M., and Graffin, S. D. (2017). Higher highs and
and access to finance. Strategic. Manage. J. 35, 1–23. doi: 10.1002/smj.2131 lower lows: the role of corporate social responsibility in CEO dismissal. Strategic.
Christensen, D. M., Serafeim, G., and Sikochi, A. (2022). Why is corporate virtue Manage. J. 38, 2255–2265. doi: 10.1002/smj.2646
in the eye of the beholder? The case of ESG ratings. Account. Rev. 97, 147–175. Jørgensen, E. N. J., and Ellingsen, T. H. (2021). ESG Disagreement: Determining
doi: 10.2308/TAR-2019-0506 Factors and Impact on Stock performance (Handelshøyskolen BI).
Colin Cameron, A., and Miller, D. L. (2015). A practitioner’s guide to cluster- Khan, M., Serafeim, G., and Yoon, A. (2016). Corporate sustainability: first
robust inference. J. Human Res. 50, 317–372. doi: 10.3368/jhr.50.2.317 evidence on materiality. Account. Rev. 91, 1697–1724. doi: 10.2308/accr-51383
Cookson, J. A., and Niessner, M. (2020). Why don’t we agree? Evidence from a
Kim, I., Wan, H., Wang, B., and Yang, T. (2019). Institutional investors and
social network of investors. J. Finance. 75, 173–228. doi: 10.1111/jofi.12852
corporate environmental, social, and governance policies: evidence from toxics
Durand, R., Rao, H., and Monin, P. (2007). Code and conduct in French cuisine: release data. Manage. Sci. 65, 4901–4926. doi: 10.1287/mnsc.2018.3055
Impact of code changes on external evaluations. Strategic. Manage. J. 28, 455–472.
doi: 10.1002/smj.583 Kotsantonis, S., and Serafeim, G. (2019). Four things no one will tell you about
ESG data. J. Appl. Corp. Financ. 31, 50–58. doi: 10.1111/jacf.12346
Dyck, A., Lins, K. V., Roth, L., and Wagner, H. F. (2019). Do institutional
investors drive corporate social responsibility? International evidence. J. Finan. Krueger, P., Sautner, Z., and Starks, L. T. (2020). The importance of
Econ. 131, 693–714. doi: 10.1016/j.jfineco.2018.08.013 climate risks for institutional investors. Rev. Financ. Stud. 33, 1067–1111.
doi: 10.1093/rfs/hhz137
Engle, R. F., Giglio, S., Kelly, B., Lee, H., and Stroebel, J. (2020). Hedging climate
change news. Rev. Financ. Stud. 33, 1184–1216. doi: 10.1093/rfs/hhz072 Lins, K. V., Servaes, H., and Tamayo, A. N. E. (2017). Social capital, trust, and
firm performance: the value of corporate social responsibility during the financial
Escrig-Olmedo, E., Fernández-Izquierdo, M. Á., Ferrero-Ferrero, I., Rivera- crisis. J. Fin. 72, 1785–1824. doi: 10.1111/jofi.12505
Lirio, J. M., and Muñoz-Torres, M. J. (2019). Rating the raters: evaluating how
ESG rating agencies integrate sustainability principles. Sustain. Basel. 11, 915. Morgan, D. P. (2002). Rating banks: Risk and uncertainty in an opaque industry.
doi: 10.3390/su11030915 Amer. Econ. Rev. 92, 874–888. doi: 10.1257/00028280260344506

Liu 10.3389/fpsyg.2022.936798
Pedersen, L. H., Fitzgibbons, S., and Pomorski, L. (2021). Responsible Servaes, H., and Tamayo, A. (2013). The impact of corporate social responsibility
investing: the ESG-efficient frontier. J. Finan. Econ. 142, 572–597. on firm value: The role of customer awareness. Manage. Sci. 59, 1045–1061.
doi: 10.1016/j.jfineco.2020.11.001 doi: 10.1287/mnsc.1120.1630
Porter, T. M. (1995). Trust in Numbers Princeton, NJ: Princeton University Press. Southworth, K. (2009). Corporate voluntary action: A valuable but incomplete
solution to climate change and energy security challenges. Policy. Soc. 27, 329–350.
Sauder, M., and Espeland, W. N. (2009). The discipline
doi: 10.1016/j.polsoc.2009.01.008
of rankings: tight coupling and organizational change.
Am. Sociol. Rev. 74, 63–82. doi: 10.1177/0003122409074 Tang, D. Y., Yan, J., and Yao, C. Y. (2021). The Determinants of ESG Ratings: Rater
00104 Ownership Matters. Working paper.

Liu 10.3389/fpsyg.2022.936798
Appendix
TABLE A Variable definitions.
Variable Definition
Panel A: ESG variables

ESG_Divergence The standard deviation of ESG ratings of firm i in year t.
ESG_Qmetrics The natural logarithm of the number of quantified indicators about firm i’s ESG disclosureplus one.
ESG_Qmetrics_E The natural logarithm of the number of quantified indicators for environmental pillar about firm i’s ESG disclosure plus one.
ESG_Qmetrics_S The natural logarithm of the number of quantified indicators for social pillar about firm i’s ESG disclosure plus one.
ESG_Qmetrics_G The natural logarithm of the number of quantified indicators for governance pillar about firm i’s ESG disclosure plus one.
ESG_mean The average ESG rating of a firm received from different rating agencies.
ESG_N The number of rating agencies following the firm.
Divergence1 The average of absolute values of the difference between pairs of ratings that firm i receives from rating agencies for its ESG
performance in year t.
Divergence2 The coefficient of variation of firm i’s ESG ratings in year t.
Panel B: Fundamental variables
State A dummy variable equal to one if the company is a state-owned company zero otherwise.
Size The natural logarithm of total assets.
Leverage The ratio between total debt the book value of assets.
ROA The ratio of net profit to total assets.
BM The ratio between book value market value.
TobinQ Tobin’s Q.
Diversification The quadratic sum of the ratio between income of businessk total income.
Verification A dummy variable equal to one if the firm’s ESG report or CSR report is verified by a third party zero otherwise.
Willing A dummy variable equal one if the firm discloses ESG/CSR report under the requirements zero otherwise.
Inst The percentage of the firm’s shares owned by institutional investors.

Quantitative ESG Disclosure and Divergence of ESG

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quantitative ESG Disclosure and Divergence of ESG

Uploaded by

Copyright:

Available Formats

TYPE Original Research

PUBLISHED 05 August 2022

Quantitative ESG disclosure and

Frontiers in Psychology 01 frontiersin.org

Frontiers in Psychology 02 frontiersin.org

Frontiers in Psychology 03 frontiersin.org

Frontiers in Psychology 04 frontiersin.org

Frontiers in Psychology 05 frontiersin.org

TABLE 1 Rating overview using available information in the database.

Rating agency Rating score Coverage Frequency Source

Frontiers in Psychology 06 frontiersin.org

TABLE 2 Descriptive statistics.

SynTao Green 4,277 5.190 5.000 0.981

Panel B: Descriptive statistics of regression variables

ESG_Divergence 4,966 2.051 0.862 1.404 2.073 2.635

Panel C: Correlations of regression variables

Frontiers in Psychology 07 frontiersin.org

Frontiers in Psychology 08 frontiersin.org

TABLE 3 Effect of ESG quantitative disclosure on ESG rating divergence.

Dependent variable (1) (2) (3)

ESG_Qmetrics 0.242*** 0.108*** 0.061***

Frontiers in Psychology 09 frontiersin.org

TABLE 4 Effect of E/S/G pillar-speciﬁc quantitative disclosure on ESG ratings divergence.

Dependent variable (1) (2) (3)

Frontiers in Psychology 10 frontiersin.org

Dependent variable (1) (2)

ESG_Qmetrics*D −0.156*** (−2.62) −0.137** (−2.02)

Frontiers in Psychology 11 frontiersin.org

Dependent variable (1) (2)

ESG_Divergence −0.078*** (−3.26) −0.041* (−1.87)

Frontiers in Psychology 12 frontiersin.org

Dependent variable: ESG_Divergence Diversification ESG_mean

ESG_Qmetrics 0.051* 0.144*** 0.025 0.087***

Frontiers in Psychology 13 frontiersin.org

Dependent variable (1) (2) (3) (4) (5)

ESG_Qmetrics 0.075*** 0.009*** 0.062*** 0.061** 0.050***

Frontiers in Psychology 14 frontiersin.org

Frontiers in Psychology 15 frontiersin.org

Frontiers in Psychology 16 frontiersin.org

Acknowledgments that could be construed as a potential conflict

Frontiers in Psychology 17 frontiersin.org

Frontiers in Psychology 18 frontiersin.org

TABLE A Variable deﬁnitions.

Panel A: ESG variables

Frontiers in Psychology 19 frontiersin.org

You might also like

ESG_Qmetrics 0.242* 0.108* 0.061***

ESG_QmetricsD −0.156 (−2.62) −0.137 (−2.02)

ESG_Qmetrics 0.051* 0.144* 0.025 0.087*

ESG_Qmetrics 0.075* 0.009* 0.062* 0.061 0.050***