1 s2.0 S1057521923003277 Main
1 s2.0 S1057521923003277 Main
A R T I C L E I N F O A B S T R A C T
JEL classification: We examine the effect of mandatory environmental, social and governance (ESG) disclosure on firms' price
G14 discovery efficiency around the world. Using data from 45 countries between 2000 and 2020 and a difference-in-
G15 differences method, we find that mandatory ESG disclosure increases firm-level stock price non-synchronicity
G18
and timeliness of price discovery, suggesting more firm-specific information is incorporated into stock prices
G28
G30
in a more timely manner. Mandatory ESG disclosure improves price discovery efficiency more in countries with
strong demands for ESG information and in firms with poor disclosure incentives. Mandatory ESG disclosure also
Keywords:
Mandatory ESG disclosure
leads to other real market changes, such as lower stock returns, greater changes in institutional ownership and
Price efficiency higher firm valuation.
Governance
* Corresponding author.
E-mail addresses: [email protected] (Q. Zhang), [email protected] (R. Ding), [email protected] (D. Chen), [email protected]
(X. Zhang).
1
For example, socially responsible investors implement a “negative screening” approach that excludes firms operating in “sin” industries such as alcohol and
tobacco, which creates an additional demand for firms to disclose ESG information.
https://doi.org/10.1016/j.irfa.2023.102811
Received 27 March 2023; Received in revised form 27 June 2023; Accepted 12 July 2023
Available online 17 July 2023
1057-5219/© 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
diversity of topics, policies and activities (Amel-Zadeh & Serafeim, incorporated into the stock price, and improved timeliness of price
2018). Given the fundamental differences between ESG and financial discovery, with forward-looking value-relevant information being
disclosure, and the potential divergent effects on stock markets, there is incorporated into the stock price in a more timely manner. We further
high risk in relying on prior research on mandatory financial reporting identify the conditions under which a mandatory ESG requirement could
to predict the consequences of mandatory ESG reporting (Bénabou & impact share price discovery efficiency. A mandatory ESG disclosure
Tirole, 2010; Christensen et al., 2017; Grewal et al., 2019). policy is expected to generate more capital market benefits when the
Motivated by the fundamental difference between ESG disclosure demand for ESG information is stronger. We find that the net benefits of
and financial disclosure, and the limited empirical evidence on the im mandatory ESG disclosure on price discovery efficiency are more pro
pacts of ESG disclosure on stock markets, we focus on mandatory ESG nounced in countries with a strong demand for ESG information
disclosure and its economic consequences for share price discovery ef (countries making environmental protection a high priority, or with
ficiency around the world. Share price discovery efficiency, which poor-quality investor protection) and in firms with poorer disclosure
measures the extent to which the stock price incorporates all available incentives (firms with poorer ESG information disclosure records,
public and private information in a timely manner, is crucial to realising poorer ESG performance or poorer corporate governance quality). Our
the benefits of an efficient capital market (Easley & O'Hara, 2004). Price difference-in-differences (DID) test strengthens the causal inference of
discovery efficiency is collectively influenced by all market participants our results. Our findings are robust to our use of an alternative mandate-
(Easley & O'Hara, 2004; Piotroski & Roulstone, 2004; Verrecchia, 2001) effective event window, exclusion of countries with other institutional
and varies in terms of different amounts of information being incorpo reforms in the event window of the ESG disclosure mandate, exclusion of
rated into share prices within different time horizons (Back, Cao, & observations from Japan and the U.S. (representing >55% of the entire
Willard, 2000; Chen, Kelly, & Wu, 2020; Holden & Subrahmanyam, sample), a policy timing analysis, exclusion of ESG-sensitive industries,
1992; Kyle, 1985; Qian, Sorensen, & Hua, 2007; Yan & Zhang, 2009). placebo tests using pseudo effective years of mandates, and estimation of
ESG information has been found to be value-relevant (Bolton & Kac pooled OLS regressions. Finally, we further reveal that mandatory ESG
perczyk, 2021; Pedersen, Fitzgibbons, & Pomorski, 2021) and manda disclosure leads to real stock market changes in terms of reduced firm-
tory ESG disclosure, by increasing the amount of information available level future stock returns, improved institutional investor participation
to all market participants, may improve information environments for and improved firm valuation outcomes.
sophisticated information users such as analysts and institutional in Our study makes important contributions to the ESG disclosure and
vestors (Bolton & Kacperczyk, 2021; Krueger, Sautner, Tang, & Zhong, corporate governance literature, in a few ways. First, we respond to the
2021; Pedersen et al., 2021; Schiemann & Tietmeyer, 2022). However, it call by Christensen et al. (2021) for more research on whether manda
may not necessarily improve information environments for other in tory non-financial reporting generates market-wide benefits and costs.
formation users, leading to unclear information asymmetry outcomes Our study is the first to identify the capital market impacts of mandatory
between informed and uninformed investors (Brown, Hillegeist, & Lo, ESG disclosure on stock price discovery efficiency, measured by price
2004; Easley & O'Hara, 2004), and unclear trading tactics being adopted non-synchronicity and timeliness. More specifically, our research is
by sophisticated information users to influence price discovery processes related to, but differs significantly from, Krueger et al. (2021) and
(Back et al., 2000; Chen et al., 2020; Holden & Subrahmanyam, 1992; Schiemann and Tietmeyer (2022), who focus on the impacts of
Kyle, 1985; Piotroski & Roulstone, 2004; Qian et al., 2007; Verrecchia, mandatory ESG disclosure on analysts' forecast quality. This is because
2001; Yan & Zhang, 2009). Therefore, whether mandatory ESG disclo mandatory ESG disclosure may improve analysts' forecast quality but
sure regulations improve firm-level share price discovery efficiency re not necessarily the capabilities of all market participants to incorporate
mains unresolved, suggesting that identifying the relative magnitudes of ESG information into share prices (Chen et al., 2020; Easley & O'Hara,
the various benefits and costs arising from such regulations is largely an 2004; Piotroski & Roulstone, 2004). In addition, our focus on a dynamic
empirical issue that deserves investigation. price discovery process, in which all market participants participate, not
To capture the dynamic share price discovery process and its effi only reveals its capabilities in terms of incorporating different amounts
ciency, we use two measures. The first measure is stock price non- of ESG information from all market participants into share prices, but
synchronicity (PSI), capturing the proportion of variation in the firm- also its capabilities in terms of shortening the time horizon needed for a
level stock return that cannot be explained by market and industry- more efficient share price (Beekes et al., 2016; Beekes & Brown, 2006;
wide information but is driven by firm-specific information (Bennett, Fama, 1965; Holden & Subrahmanyam, 1992; Qian et al., 2007; Zhang
Stulz, & Wang, 2020; Chen, Goldstein, & Jiang, 2007; Chen, Huang, Li, et al., 2022). Thus, our analysis can help us understand this dynamic
& Yuan, 2022; Crawford, Roulstone, & So, 2012; Durnev, Morck, Yeung, price discovery process, as a crucial transmission mechanism linking
& Zarowin, 2003; Fernandes & Ferreira, 2008, 2009; Gul, Kim, & Qiu, mandatory ESG disclosure to its tangible capital market benefits (Brown
2010; Gul, Srinidhi, & Ng, 2011; Kacperczyk, Sundaresan, & Wang, et al., 2004; Christensen et al., 2021; Easley & O'Hara, 2004).
2021; Morck, Yeung, & Yu, 2000; Roll, 1988). Greater values of PSI Second, we reveal potential spill-over effects (Admati & Pfleiderer,
suggest more efficient stock prices incorporating more firm-specific 2000; Baginski & Hinson, 2016; Shroff, Verdi, & Yost, 2017) of
public and private information, which reveals a stronger capability of mandatory ESG disclosure in countries where firms have flexibility in
this more efficient price discovery process to incorporate larger amounts complying with these mandates. This extends previous literature
of information into share prices. Our second measure is the intra-year focusing on the economic consequences of voluntary ESG or CSR
timeliness of price discovery (TIMELINESS), capturing the timeliness disclosure (see Christensen et al., 2021 for a comprehensive review).
with which forward-looking information is incorporated into the stock Our analysis suggests market participants may infer information about
price throughout a fiscal year (Alford, Jones, Leftwich, & Zmijewski, non-complying or non-disclosing firms from their complying or
1993; Ball & Brown, 1968; Beekes & Brown, 2006; Beekes, Brown, Zhan, disclosing peers, leading to such non-disclosed but inferred firm infor
& Zhang, 2016; Beekes, Brown, & Zhang, 2015; Haß, Vergauwe, & mation ultimately being incorporated into share prices, but in a less
Zhang, 2014; Zhang, Zhang, Chen, & Gu, 2019; Zhang, Zhang, Chen, & timely manner.
Strange, 2022). Greater values of TIMELINESS suggest that the market is Third, the empirical literature exhibits a heavy focus on disclosure
slower in incorporating forward-looking value-relevant public and pri regulation in the United States, while a global setting could provide
vate information into the current share price. opportunity for tighter research designs to strengthen the causal in
Using data collected from 45 countries between 2000 and 2020, we ferences of mandatory ESG disclosure policy that are absent in a U.S.
find that, in countries where the mandatory ESG disclosure becomes setting (Leuz & Wysocki, 2016). We respond to the call from Leuz and
effective, stock price non-synchronicity increases, suggesting a higher Wysocki (2016) and focus on the country-level mandatory ESG disclo
price discovery efficiency outcome, with more firm-specific information sure policy in a global setting. Our analysis, using data from 45
2
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
countries, allows us to construct a robust control group from countries impact on stock returns. Pedersen et al. (2021) propose the ESG-efficient
without mandatory ESG disclosure policies, which is less likely to be frontier and show how required returns are related to the ESG perfor
affected directly or indirectly by mandatory ESG disclosure policies mance of firms. Second, mandatory ESG disclosure can address the se
adopted by other countries, and thus helps mitigate identification issues lective disclosure issues related to voluntary ESG disclosure (Dhaliwal,
in our DID method used to reveal the real impacts of mandatory ESG Li, Tsang, & Yang, 2011; Dhaliwal, Radhakrishnan, Tsang, & Yang,
disclosure. What is more, our analysis not only reveals the various im 2012; Lins, Servaes, & Tamayo, 2017) and force firms to release previ
pacts related to different firm-level factors that moderate the impacts of ously unavailable information to the market. As a result, stock prices are
ESG disclosure on price discovery efficiency, but also those related to likely to become more informative after ESG disclosure mandates come
different country-level factors. Our analysis should have useful impli into force. Third, mandatory ESG disclosure can improve information
cations for regulators, helping them to evaluate the benefits and costs of environments for sophisticated information users such as analysts and
mandatory ESG disclosure regulation. institutional investors, by increasing the volume and quality of ESG
Finally, a caveat regarding many of the prior studies is that they tend information (Ioannou & Serafeim, 2019; Krueger et al., 2021) and
to concentrate on specific disclosure items (e.g., mine safety records or reducing the costs of private searching, collection and verification of
greenhouse gas emission) in a single country.2 Our study looks at ESG information (Christensen et al., 2021).3 When more sophisticated
mandates regarding ESG disclosure, which comprehensively cover all information users become informed through this disclosed ESG infor
ESG activities taking place within a firm around the world, and thus mation, competition among informed investors leads to aggressive
should add to the literature more powerful evidence of the trade-off trading, which incorporates their superior insights on this value-relevant
between benefits and costs related to ESG disclosure regulation. ESG information into prices (Bushman & Smith, 2001; Chen et al., 2020;
The remainder of the paper is organized as follows. Section 2 reviews Fishman & Hagerty, 1992; Kyle, 1985; Massa, Zhang, & Zhang, 2015;
the literature and develops the main hypotheses. The sample and Qian et al., 2007; Yan & Zhang, 2009; Zhang et al., 2022). Thus,
research design are described in Section 3. Section 4 presents the results mandatory ESG disclosure may shorten informed investors' private in
of baseline models, robustness tests, heterogeneous treatment effects formation horizon (Back et al., 2000; Holden & Subrahmanyam, 1992;
and additional tests. The final section concludes. Kyle, 1985) and improve share price discovery efficiency by leading to
firm-specific information being incorporated into the share price in a
2. Related literature and hypothesis development more timely manner. Based on the above discussion, we propose our
hypotheses H1a and H1b as follows:
Recent literature has moved away from a static market efficiency
H1a. Mandatory ESG disclosure increases share price informativeness.
view and suggests that market efficiency is dynamic and that an efficient
price discovery process, incorporating all available public and/or pri H1b. Mandatory ESG disclosure increases share price timeliness.
vate information from all market participants in a timely manner, is
Another strand of literature suggests that ESG reporting is hard for
crucial to strengthening market discipline (Christensen et al., 2021;
investors to process (Bingler, Senni, & Monnin, 2022; Park & Ravenel,
Edmans, 2009; Grossman & Stiglitz, 1980; Holmström & Tirole, 1993;
2013), thereby increasing long private information horizons among
Zhang et al., 2022; Zhang, Yang, Strange, & Zhang, 2017) and thus
differently informed investors (Back et al., 2000; Choi, Larsen, & Seppi,
reducing the cost of equity capital (Brown et al., 2004; Easley & O'Hara,
2019; Ferreira, Ferreira, & Raposo, 2011; Grossman & Stiglitz, 1980;
2004; Verrecchia, 2001).
Holden & Subrahmanyam, 1992; Zhang et al., 2017), which may ulti
One perspective suggests that making ESG disclosure mandatory
mately dampen share price discovery efficiency. Without clear guidance
mitigates the long-lived private information available in the market
on the metrics of ESG information disclosure that firms have to provide,
(Back et al., 2000; Holden & Subrahmanyam, 1992; Kyle, 1985), in
it is difficult to standardize and regulate the disclosure of ESG infor
centivizes competitive informed trading (Chen et al., 2020; Qian et al.,
mation, which provides managers with the flexibility to manipulate it
2007; Yan & Zhang, 2009), and thus should improve share price dis
(see Hermalin & Weisbach, 2012 for a review). Consequently, manda
covery efficiency. First of all, ESG information is value-relevant. Bolton
tory ESG disclosure may improve the information environment for a
and Kacperczyk (2021) show that carbon emissions have a significant
limited number of sophisticated and capable information users, but is
unlikely to contribute to the information processing and verification of
2 unsophisticated investors (Bingler et al., 2022; Park & Ravenel, 2013).
For example, Christensen et al. (2017) examine the real effect of the mine
Thus, mandatory ESG disclosure may attract the attention of many un
safety disclosure required by the Dodd-Frank Wall Street Reform and Consumer
Protection Act of 2010 in the U.S., and find that the safety of coal mines im
informed investors but enlarge the information asymmetries between
proves but productivity declines. Chen, Hung, and Wang (2018) exploit the CSR them and informed investors (Brown et al., 2004; Easley & O'Hara,
disclosure mandate issued by the Shanghai and Shenzhen Stock Exchanges in 2004; Merton, 1987), stimulating informed investors to engage in more
China, which took effect for fiscal years ending on or after 31st December 2008. discreet trading tactics so as to extend their long private information
They find a decrease in overall industrial wastewater and CO2 emissions in horizon and extract rents through liquidity support from uninformed
cities with more regulated firms. They further document that firms subject to investors (Back et al., 2000; Choi et al., 2019; Ferreira et al., 2011;
the mandate experience a deterioration in profitability. Grewal et al. (2019) Fishman & Hagerty, 1992; Grossman & Stiglitz, 1980; Holden & Sub
focus on short-window returns to events leading to the passing of an EU rahmanyam, 1992; Zhang et al., 2017). The discreet trading tactics used
directive making the disclosure of non-financial CSR information mandatory
(EU Corporate Social Responsibility Directive: NFRD 2014/95/EU). They show,
on average, a negative market reaction but positive returns for firms with more
3
CSR disclosure and better CSR performance before the mandate came into For example, Ioannou and Serafeim (2019), who compare firms from four
force. Downar, Ernstberger, Reichelstein, Schwenen, and Zaklan (2021) inves countries with CSR disclosure mandates before 2011 (China, Denmark,
tigate whether a mandate regarding greenhouse gas emissions introduced by Malaysia and South Africa), find that firms in countries with the mandates in
the UK government generates pressure for firms to decrease their emissions. crease their volume and quality of CSR disclosure in the post-mandate period.
They show that firms affected by the mandate reduce their emissions by about Krueger et al. (2021) explore the effect of mandatory ESG disclosure on firms'
8%, which is accompanied by a significant increase in production costs. Finally, information environments. They find that mandatory ESG disclosure increases
Fiechter, Hitz, and Lehmann (2022) also examine the EU Corporate Social the accuracy of analysts' earnings forecasts, and lowers analyst forecast
Responsibility Directive (NFRD 2014/95/EU) that requires qualifying firms to dispersion. In addition, mandatory disclosure has regular disclosure frequency,
disclose non-financial information from the fiscal year 2017 onwards. They transparency and comparability arising from standardization, which reduces
report that firms increase their CSR activities and did so even before the the costs of obtaining, processing and comparing ESG information (Christensen
mandate took effect. et al., 2021).
3
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
by informed investors, based on their superior information collection observations for the price non-synchronicity sample, and 79,689 firm-
and processing ability, may further compromise market discipline year observations for the price timeliness sample. Table 1 reports the
regarding managerial ESG disclosure (Edmans, 2009; Grossman & Sti sample distribution by country. As shown in the table, China has the
glitz, 1980; Holmström & Tirole, 1993; Zhang et al., 2017, 2022), largest number of observations in the treatment group, with mandatory
leading to poor price discovery efficiency (Ferreira et al., 2011; Gross ESG disclosure reforms (3005 and 3029), and the United States has the
man & Stiglitz, 1980; Zhang et al., 2017).4 Based on the above discus largest number of observations in the control group, without mandatory
sion, we propose H2a and H2b as follows: ESG disclosure reforms (27,299 and 27,700). Japan and the U.S.
contribute >55% of observations in the full samples.
H2a. Mandatory ESG disclosure decreases share price informativeness.
We obtain share price and financial data for public firms from the
H2b. Mandatory ESG disclosure decreases share price timeliness. Refinitiv Datastream and Refinitiv Worldscope. Data on analyst
following is collected from the Institutional Brokers Estimate System (I/
In our empirical testing, if the main impact of mandatory ESG
B/E/S). Annual CO2 emissions (metric tons per capita) are collected
disclosure on the price discovery process is to mitigate the long-lived
from the Emissions Database for Global Atmospheric Research
private information available in the market, and incentivize competi
(EDGAR).8 The effective years of mandatory ESG disclosure reform are
tive informed trading so that more information is incorporated in the
collected from the study of Krueger et al. (2021).9 Major corporate
share price in a more timely manner, H1a and H1b will be supported. If
governance reform years are collected from the study of Fauver, Hung,
the main impact of mandatory ESG disclosure is to increase uncertainty
Li, and Taboada (2017). Firm-level corporate governance data come
and opportunistic trading by informed investors, H2a and H2b will be
from Refinitiv Eikon. Institutional ownership data are obtained from the
supported.
Refinitiv Ownership Database. The status of firm-level ESG reporting is
collected from the Refinitiv ESG Database. Data on legal institution
3. Research methods
quality and other country-level financial and macroeconomic variables
are collected from the World Bank. The anti-self-dealing index is from
3.1. Sample and data
Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2008). Data with
respect to country-level attitudes on environmental protection are
Our sample includes public firms from 45 countries for the period
collected from the World Values Survey (WVS).
from 2000 to 2020. The sample countries are the constituents of the
MSCI World Index and the MSCI Emerging Markets Index. The MSCI
indices include major developed and emerging countries around the 3.2. Measuring price discovery efficiency
world. The selected countries must have data/information available for
all country-level variables.5 We include companies delisted during the An efficient price discovery process, incorporating all available
sample period. We exclude firm-year observations with missing data on public and/or private information in a timely manner, is crucial for
the dependent or explanatory variables. We exclude firm-year obser dynamic market efficiency (Brown et al., 2004; Easley & O'Hara, 2004;
vations with negative book value of equity or negative sales. We also Grossman & Stiglitz, 1980; Verrecchia, 2001). We measure price dis
exclude firms in the financial and utilities sectors, with standard in covery efficiency using two proxies, including stock price non-
dustrial classification (SIC) codes 6011–6799 and 4900–4949, because synchronicity, to capture the extent to which the price discovery pro
these two sectors have unique business models and regulatory re cess incorporates firm-level information into the price, and the timeli
quirements.6 To ensure that our analysis does not suffer from selection ness of price discovery, to capture to what extent forward-looking
bias, our sample includes 27 countries that have enacted ESG disclosure information is incorporated into the share price in a timely manner.10
mandates (i.e., the treatment group) and 18 countries without any ESG Stock price non-synchronicity has been widely used as a measure of
disclosure mandates (i.e., the control group). Because a mandatory ESG stock price informativeness in the literature (Bennett et al., 2020; Chen
disclosure policy is likely to occur around the same time as other insti et al., 2007; Crawford et al., 2012; Durnev et al., 2003; Fernandes &
tutional reforms, we restrict the window of the treatment group (i.e., Ferreira, 2008, 2009; Gul et al., 2010, 2011; Kacperczyk et al., 2021;
countries with the implementation of mandatory ESG disclosure) to the Morck et al., 2000; Roll, 1988).11 Price non-synchronicity measures the
two years before the policy effective year, the policy effective year, and amount of firm-specific information incorporated into stock prices,
the three years after the policy effective year.7 The full sample period (i. based on the R2 obtained from asset pricing regressions. Higher levels of
e., 2000–2020) is used as the window for countries in the control group. price non-synchronicity signal more information-laden stock prices and
The above procedure results in a final sample of 78,815 firm-year more efficient stock markets (Durnev et al., 2003).
We construct the stock price non-synchronicity measure based on the
4
In addition, making ESG disclosure mandatory may not change firms'
8
disclosure behaviour. Instead, they may continue at the same level of ESG The EDGAR database can be accessed through the website https://edgar.jrc.
disclosure if their prior voluntary disclosure satisfies the mandatory regulatory ec.europa.eu/report_2022.
9
requirements. Fiechter et al. (2022) examine the EU Corporate Social Re Krueger et al. (2021) analyse ESG disclosure policies around the world and
sponsibility Directive (NFRD 2014/95/EU) that has required qualifying firms to create a list of effective years of ESG disclosure mandates. For example, the UK
disclose non-financial information since fiscal year 2017. They report that firms released amendments to the Companies Act 2006 in 2013. According to this
increase their CSR activities and did so even before the mandate took effect. regulation, all public companies are required to produce reports on matters
5
Some constituent countries are excluded from our sample due to the lack of such as strategy and business model, levels of greenhouse gas emissions, human
data. For example, Kuwait, Qatar, Saudi Arabia and United Arab Emirates are rights and diversity in the company.
10
not covered by the anti-self-dealing index of Djankov et al. (2008). We do not use private information risk measurement or PIN here as it only
6
For example, Fama and French (1992) indicate that the high leverage that is reflects the information asymmetries among differently informed investors,
normal for financial firms probably does not have the same meaning as for non- while our focus is on the efficiency of the price discovery process, namely, to
financial firms, where high leverage more likely indicates distress. what extent and in how timely a manner the share price incorporates all
7
Fauver et al. (2017) restrict their sample period to five years before and available public and/or private firm-level information.
11
after corporate governance reform to mitigate the impact of confounding For example, stock price non-synchronicity is related to firm productivity
events. We selected this sample period (two years before the ESG mandatory (Bennett et al., 2020), sensitivity of corporate investment to the stock price
disclosure policy effective year and three years after the policy effective year) in (Chen et al., 2007), analyst initiations of coverage (Crawford et al., 2012),
order to ensure that mandatory ESG disclosure was the only major institutional ownership concentration (Gul et al., 2010) and property rights (Morck et al.,
reform during this period for a given country. 2000).
4
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Table 1
Number of firm-year observations by country and effective years of mandatory ESG disclosure policies.
Number of observations Mandatory ESG disclosure policy year Comply-or-explain regulation? All-at-once disclosure?
Columns 1 and 2 show the number of firm-year observations by country based on the regression models as shown in Eq. (5). The effective years of mandatory ESG
disclosure policies by country are given in column 3. Columns 4 and 5 indicate the reform approaches. The data on effective years of mandatory ESG disclosure reform
and reform approaches are collected from Krueger et al. (2021).
proportion of return variation that cannot be explained by the returns on return variation, into a systematic part that is synchronous to other
the market and the sector in which the firm resides. For each firm i and firms, and a firm-specific part that is informative about the firm itself.
year y in our sample, we run the following time-series regression: We use the log-transformed R2i,y , adjusted for degrees of freedom, to
ri,y,t = β0,i,y + β1,i,y rm,y,t + β2,i rn,y,t + εi,y,t , (1) capture the informativeness of the stock for the firm:
( )
1 − R2i,y
where ri,y,t denotes the daily return time-series of firm i in year y, and PSIi,y = ln (2)
R2i,y
rm,y,t and rn,y,t are the day t value-weighted return indices of the country
and sector in which firm i operates. In the empirical analysis, the market ( )
index is constructed using the value-weighted average return of all the Ti,y − 1 s2 εi,y
R2i,y = 1 − ( ) (3)
constituent firms within a market. Similarly, we construct sector indices Ti,y − ki,y − 1 s2 ri,y
as the value-weighted average return of all firms in a sector, specified by
the two-digit SIC code. where Ti,y and ki,y are the number of daily return observations and
This bi-index model leads to a natural decomposition of the stock number of explanatory variables in the index model respectively and
( )
s xi,y denotes the sample standard deviation of x for firm i in year y. A
5
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
firm has a large PSI when its stock price moves less synchronously with GSHOCK is a dummy variable for global shocks that equals one if the
the market and sector index, and therefore contains a larger idiosyn year is 2008 or 2020, and zero otherwise.13 CO2 is CO2 emissions in
cratic component and more firm-specific information. metric tons per capita. CGRF is a dummy variable that equals one if a
To measure the timeliness of price discovery, we adopt Beekes and country-year is after the year when a major corporate governance re
Brown (2006)’s methodology to examine the process whereby value- form becomes effective in that country, and zero otherwise. We obtain
relevant, private information becomes impounded into a stock's mar the information on corporate governance reforms from the work of
ket price over a fiscal year. The timeliness of price discovery refers to Fauver et al. (2017). LIQ is the sum of three World Governance In
how quickly that process takes effect. The Beekes and Brown (2006) dicators (WGI; government effectiveness, regulatory quality, and the
metric is adapted in subsequent studies to accommodate international rule of law) and the anti-self-dealing index from Djankov et al. (2008).
comparisons (Beekes et al., 2016, 2015; Haß et al., 2014; Zhang et al., The WGI and anti-self-dealing indices are rescaled to be between 0 and
2019).12 The genesis of the timeliness measure lies in the seminal work 1. MKTCAP is the market capitalisation of listed domestic companies
of Ball and Brown (1968) and Alford et al. (1993). The measure assesses divided by GDP. GDPG is the annual percentage growth rate of GDP. We
how accurately a firm's daily share price (Pt), observed throughout the control the following firm-level variables: SIZE, the natural logarithm of
year, approximates its terminal value (P0). Ball and Brown (1968) sug total assets of a firm in U.S. dollars; IO, the number of shares held by
gest that the annual financial report is not a highly timely venue for the institutional investors divided by the total number of shares
provision of value-relevant information, since most of the information outstanding; PROFIT, earnings before interest, taxes and depreciation
has been reported on traditional and social media before the earnings divided by total assets; LEV, total debt divided by total assets; BM, the
announcement day. This justifies the calculation of the differences be book value of equity divided by the market value of equity; CASH, cash
tween the “final” price (P0) 14 days after the earnings announcement and short-term investments divided by total assets; CAPEX, capital
day and the market-adjusted price from 365 days to 1 day before the expenditure divided by total assets; TURN, the share trading volume
earnings announcement day for year t (Pt). divided by adjusted shares outstanding; NUMEST, the natural logarithm
For each fiscal year, the timeliness measure traces the share price of the number of analysts following a firm in a year; VOLTY, the standard
over 365 calendar days ending 14 days after the firm's annual earnings deviation of daily stock returns over the 365 calendar days prior to the
announcement day, which is an important event and common to all fiscal year end date. Detailed definitions of all variables and data re
firms in all countries. The measure captures the speed with which sources are provided in Table A1 in Appendix A.
forward-looking information contained in the forthcoming annual Table 2 provides descriptive statistics for the variables used in the
earnings report is reflected in the stock price up to the day of the annual main tests. All time-varying variables are winsorized at the top and
earnings announcement. Specifically, the timeliness of price discovery bottom 1% to control for outliers. We only report the statistics of the
(TIMELINESS) is calculated as: variables used in our baseline models and the statistics of the explana
(( ) )/ tory variables are based on the price non-synchronicity (PSI) model.
t∑
=− 1
TIMELINESS = |ln(P0 ) − ln(Pt ) | − 0.5 365 (4) Price non-synchronicity (PSI) ranges from − 1.751 to 5.988, with a mean
t=− 365 and median of 1.466 and 1.263, and a standard deviation of 1.457. Price
timeliness (TIMELINESS) ranges from 0.029 to 0.409, with a mean and
where Pt is the daily market-adjusted stock price of a firm from 365 median of 0.136 and 0.117, and a standard deviation of 0.079. The main
calendar days before the annual earnings announcement day until 1 day explanatory variable of interest, ESGPOST, is an indicator variable that
before the annual earnings announcement day. P0 is the price 14 days is equal to one for all years after the mandatory ESG disclosure policy
after the annual earnings announcement day, which reflects the intrinsic becomes effective in a country, and zero otherwise. ESGPOST has 9568
value of the year. In order to reduce the impact of volatility, we deflate observations with a value of one, about 12.14% of all observations
the measure by one plus the absolute rate of return on the share over the (78,815), in the PSI model. ESGPOST has 9895 observations with a value
time window used to calculate the timeliness metric. We forward-fill of one, about 12.42% of all observations (79,689), in the TIMELINESS
prices on days when there is no trading. We set the end date to be 14 model. Table 3 provides the Pearson correlation coefficients of the main
days after the earnings release date, which allows the market to grad variables. The correlation between PSI and ESGPOST is significant and
ually absorb information within a reasonable price discovery process positive, suggesting that stock prices become more informative after the
(Beaver, 1968). The 0.5 adjustment is included to recognize that the implementation of mandatory ESG disclosure reform. The correlation
flow of information is reflected in returns over the day (Beekes et al., between TIMELINESS and ESGPOST is significant and negative, sug
2016; Haß et al., 2014). The larger the value of TIMELINESS, the longer gesting that price timeliness improves after the implementation of ESG
it takes a firm's share price to capture information and converge to P0, disclosure reform.14 In the following sections, we test our hypotheses
suggesting slower price timeliness. using multivariate regressions, controlling for other variables that could
affect market efficiency.
According to prior research, our regression models control for other We begin our regression analysis by estimating the baseline model
country-level and firm-level variables to explain market efficiency (e.g., that examines the effects of mandatory ESG disclosure on PSI and
Beekes et al., 2016; Bennett et al., 2020; Bolton & Kacperczyk, 2021; Gul TIMELINESS respectively. Robustness of the main results is then checked
et al., 2010; Kacperczyk et al., 2021; Lim, Brooks, & Kim, 2008; Piotroski using alternative mandate effective windows, restricted sample analysis,
& Roulstone, 2004). We control the following country-level variables: exclusion of treatment countries with simultaneous reforms, exclusion
of observations from Japan and the United States (representing >55% of
12
the entire sample), policy timing analysis, exclusion of ESG-sensitive
The timeliness measure was first proposed by Beekes and Brown (2006).
The paper estimates the price timeliness of Australian public firms and asso
ciates it with corporate governance quality. The adapted metric is used in
13
subsequent studies which re-examine the effect of corporate governance on We consider the subprime mortgage crisis of 2008 and the COVID-19
price timeliness (Beekes et al., 2015, 2016; Haß et al., 2014). Zhang et al. pandemic of 2020 as global shocks. We also use the years 2007–2008,
(2019) distinguish between the timeliness of good news and that of bad news. 2008–2009, and 2007–2009 as the subprime mortgage crisis period, and the
The authors find that financial integration affects good and bad information results remain stable and robust.
14
timeliness asymmetrically. Note that greater values of TIMELINESS suggest slower price timeliness.
6
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Table 2
Summary statistics.
Variables Mean Median SD Min Max p10 p25 p75 p90
PSI 1.466 1.263 1.457 − 1.751 5.988 − 0.151 0.466 2.257 3.431
TIMELINESS 0.136 0.117 0.079 0.029 0.409 0.052 0.076 0.175 0.245
ESGPOST 0.121 0.000 0.327 0.000 1.000 0.000 0.000 0.000 1.000
GSHOCK 0.134 0.000 0.340 0.000 1.000 0.000 0.000 0.000 1.000
CO2 11.347 9.978 5.301 1.629 21.293 4.199 7.558 15.811 18.985
CGRF 0.872 1.000 0.334 0.000 1.000 0.000 1.000 1.000 1.000
LIQ 2.544 2.712 0.590 0.292 3.841 1.476 2.400 2.921 3.011
MKTCAP 1.041 1.030 0.435 0.231 2.293 0.475 0.681 1.374 1.581
GDPG (%) 2.182 2.161 2.893 − 5.693 10.636 − 1.224 1.086 3.076 5.585
SIZE 13.786 13.733 1.804 9.726 18.253 11.492 12.518 14.980 16.198
IO 0.408 0.297 0.327 0.003 1.122 0.051 0.132 0.696 0.926
PROFIT 0.094 0.108 0.149 − 0.681 0.406 − 0.007 0.064 0.160 0.223
LEV 0.218 0.201 0.177 0.000 0.697 0.000 0.055 0.339 0.466
BM 0.677 0.510 0.578 0.038 3.198 0.157 0.283 0.882 1.396
CASH 0.182 0.122 0.185 0.002 0.903 0.022 0.054 0.240 0.429
CAPEX 0.048 0.034 0.047 0.001 0.255 0.007 0.016 0.063 0.106
TURN 1.525 0.960 1.699 0.010 9.407 0.156 0.405 2.001 3.565
NUMEST 3.972 4.043 1.061 1.386 5.948 2.485 3.178 4.804 5.342
VOLTY 0.022 0.019 0.010 0.008 0.063 0.012 0.015 0.026 0.035
The summary statistics of the variables are based on the PSI sample, except in the case of TIMELINESS. PSI is the price non-synchronicity, which is estimated using Eqs.
(2) and (3). TIMELINESS is the price timeliness, which is calculated using Eq. (4). ESGPOST is an indicator variable that is equal to one for all years after the mandatory
ESG disclosure policy becomes effective in a country, and zero otherwise. CO2 is CO2 emissions in metric tons per capita. CGRF is an indicator variable equal to one for
all years after a major corporate governance reform became effective in the country, and zero otherwise. LIQ is the legal institution quality of a country, which is
measured based on rule of law, regulatory quality, government effectiveness, and protection against self-dealing. MKTCAP is the annual market capitalisation of listed
domestic companies divided by GDP. GDPG is GDP growth (annual %). SIZE is the natural logarithm of the total assets of a firm in U.S. dollars. IO is the number of
shares held by all types of institutions divided by the total number of shares outstanding. PROFIT is earnings before interest, taxes and depreciation divided by total
assets. LEV is total debt divided by total assets. BM is the book value of equity divided by market capitalisation. CASH is cash and short-term investments divided by
total assets. CAPEX is capital expenditure divided by total assets. TURN is the annual share trading volume divided by adjusted shares outstanding. NUMEST is the
natural logarithm of the number of analysts following a firm in a fiscal year. VOLTY is the standard deviation of daily stock returns over the 365 days prior to the fiscal
year end date. All continuous variables are winsorized at the top and bottom 1%. Table A1 provides details on data and variables.
industries, placebo tests using pseudo effective years of mandates, and where i and t stand for firm and year respectively. ESGPOST is an indi
OLS estimation controlling for industry, country and year effects. In cator variable equal to one starting in the first year after the mandatory
order to further strengthen our causal inferences regarding mandatory ESG disclosure policy became effective in the country, and for all sub
ESG disclosure policy, we conduct two additional sets of tests. First, we sequent years, and zero otherwise. CONTROLS includes firm- and
investigate the heterogeneous treatment effects based on important country-level control variables. FE stands for firm and year fixed effects.
country and firm characteristics which affect the demand for ESG Based on hypothesis 1, we expect ESG disclosure to be associated with
disclosure. Second, we confirm the real stock market changes due to ESG more firm-specific information being incorporated into stock prices in a
disclosure, by looking at cross-sectional patterns of future stock returns, more timely manner, and thus we expect β1 to be positive in the PSI
institutional ownership change and firm valuation. baseline model and negative in the TIMELINESS baseline model. In all
regression estimations, we use robust standard errors clustered at the
4.1. The effects of mandatory ESG disclosure on price informativeness country level, because the ESG disclosure policy is a country-level
and timeliness decision.15
Columns 1 and 4 of Table 4 present the results of the baseline models.
We first estimate the baseline regression models shown in Eq. (5). The coefficient on ESGPOST is significantly positive in the PSI model
The dependent variables are price non-synchronicity (PSI) and timeli (β1 =0.118, p < 0.01) and significantly negative in the TIMELINESS
ness of price discovery (TIMELINESS) respectively. The main indepen model (β1 = − 0.009, p < 0.01). The results suggest that the share price
dent variable of interest is the implementation of a policy of mandatory incorporates more firm-specific information in a more timely manner
ESG disclosure (ESGPOST). Among the 45 sample countries, 27 coun following the ESG disclosure reforms. Our hypotheses H1a and H1b are
tries launched ESG disclosure reforms in different years during the therefore supported. The effects are also economically significant. Col
sample period and 18 countries did not make such a change. This allows umn 1 shows that price informativeness increases by 8.05% relative to
us to adopt a DID design with multiple treatment groups and multiple the mean following the reforms.16 Column 4 shows that price timeliness
time periods (Imbens & Wooldridge, 2009). Our DID approach compares increases by 6.62% relative to the mean following the reforms.17 Col
changes in market efficiency after the ESG disclosure reforms with umns 2 and 5 estimate the baseline models after inclusion of an
changes in market efficiency for countries without disclosure reforms
during the sample years. The approach is commonly used in the litera
ture to mitigate endogeneity issues and thereby strengthen the causal 15
We also estimate all regression models with robust standard errors clustered
inferences of the empirical investigation (Bertrand & Mullainathan, at the firm level as a robustness check. The results are unchanged.
2003; Fauver et al., 2017; Gao & Zhang, 2017). The baseline model 16
The magnitude of the impact of ESGPOST on PSI is calculated as 0.118
specification is as follows: (coefficient on ESGPOST in column 1 of Table 4) ÷ 1.466 (the sample mean of
∑ PSI in Table 2) = 8.05%.
PSI i,t or TIMELINESSi,t = α + β1 (ESGPOST) + βm CONTROLS 17
The magnitude of the impact of ESGPOST on TIMELINESS is calculated as
(5)
+ FIRM FE + YEAR FE + εi,t 0.009 (absolute value of coefficient on ESGPOST in column 4 of Table 4) ÷
0.136 (the sample mean of TIMELINESS in Table 2) = 6.62%.
7
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
This table reports the Pearson correlations among the variables. All continuous variables are winsorized at the top and bottom 1%. * indicates that the correlation is significant at least at the 5% level (two-tailed test).
− 0.08*
− 0.10*
− 0.15*
− 0.34*
− 0.45*
− 0.01*
− 0.03*
− 0.18*
(COE). Columns 3 and 6 estimate the baseline models after inclusion of
0.14*
0.54*
0.36*
0.19*
0.04*
0.08*
0.03*
0.08*
0.30*
0.37*
[19] an interaction term between ESGPOST and an all-at-once approach
(ATO). Mandatory ESG disclosure provides flexibility under the COE
approach for firms in disclosing ESG information allowing them to
− 0.37*
− 0.14*
− 0.02*
− 0.05*
− 0.02*
− 0.27*
− 0.08*
0.15*
0.02*
0.11*
0.14*
0.63*
0.40*
0.16*
0.08*
0.07*
0.16*
explain why they did not comply, and this is expected to provide in
[18]
− 0.06*
− 0.11*
− 0.14*
0.22*
0.05*
0.31*
0.13*
0.06*
0.19*
0.06*
0.03*
0.33*
0.04*
0.16*
0.04*
a comparison with the ESG data from those firms which do comply or
[17]
disclose (Admati & Pfleiderer, 2000; Baginski & Hinson, 2016; Shroff
et al., 2017). If so, this potential spill-over effect of mandatory ESG
− 0.06*
− 0.02*
− 0.02*
− 0.04*
− 0.10*
− 0.08*
− 0.07*
− 0.05*
− 0.21*
disclosure should be evidenced by a partially compromised price dis
0.02*
0.08*
0.16*
0.03*
0.21*
0.11*
[16]
− 0.06*
− 0.35*
− 0.37*
− 0.40*
− 0.19*
0.13*
0.18*
0.03*
0.10*
0.06*
0.07*
0.12*
0.02*
[15]
mation is due to the effort and time that needs to be spent by market
participants to infer the withheld information, in deriving it from the
data supplied by complying or disclosing firms. As predicted, column 2
− 0.07*
− 0.11*
− 0.06*
− 0.07*
− 0.22*
− 0.15*
− 0.24*
− 0.14*
0.08*
0.07*
0.09*
0.04*
0.06*
− 0.04*
− 0.03*
− 0.08*
− 0.04*
− 0.01*
0.03*
0.01*
0.31*
0.03*
[13]
0.01
− 0.07*
− 0.12*
− 0.02*
− 0.11*
− 0.11*
0.11*
0.23*
[12]
− 0.15*
0.60*
0.08*
0.46*
0.50*
0.24*
[11]
0.01
flexible ESG disclosure mandates that allow gradual disclosure, but only
after the implementation of strict, all-at-once ESG disclosure man
dates.18 Column 6 shows that the coefficients on both ESGPOST and
− 0.54*
− 0.27*
− 0.06*
− 0.02*
− 0.09*
− 0.01
0.05*
0.04*
0.03*
− 0.36*
− 0.16*
more pronounced for those countries that introduce the mandatory ESG
0.04*
0.38*
0.01
0.01
[9]
disclosure all at once. These results reinforce our hypotheses H1a and
H1b, suggesting that stricter ESG disclosure mandates help the market to
incorporate more value-relevant information in a more timely manner
− 0.18*
0.09*
0.05*
0.05*
0.44*
0.09*
0.52*
− 0.07*
− 0.01
0.13*
0.62*
institution quality (LIQ), GDP growth (GDPG), firm size (SIZE), profit
0.01
[7]
− 0.15*
− 0.16*
Beekes et al., 2016; Bennett et al., 2020; Lim et al., 2008) and indicate
that economic conditions and firm fundamentals can affect price dis
− 0.30*
covery efficiency.
0.06*
0.07*
0.01*
[5]
− 0.12*
0.15*
before the policy effective year, the effective year itself and one year
[3]
after the policy effective year. The narrower event window further re
duces the potential impact of other reforms implemented in the sample
0.19*
+3] sample that requires a treatment firm to appear in at least one year
of the pre-reform period and one year of the post-reform period. The
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
results of using these two alternative samples are similar to those re
Correlation matrix.
MKTCAP
NUMEST
GSHOCK
PROFIT
VOLTY
CAPEX
Table 3
TURN
GDPG
CASH
CGRF
18
SIZE
CO2
BM
PSI
IO
8
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Table 4
The effects of mandatory ESG disclosure on price informativeness and timeliness: Baseline results.
Dependent variable PSI TIMELINESS
This table presents the regression results for the impact of mandatory ESG disclosure on price informativeness and timeliness. The dependent variables are price non-
synchronicity (PSI) in columns 1–3, and price timeliness (TIMELINESS) in columns 4–6. ESGPOST is an indicator variable that is equal to one for all years after the
mandatory ESG disclosure policy becomes effective in a country, and zero otherwise. COE is an indicator variable that is equal to one for all years after the mandatory
ESG disclosure policy becomes effective in a country if the comply-or-explain ESG disclosure approach is adopted, and zero otherwise. ATO is an indicator variable that
is equal to one for all years after the mandatory ESG disclosure policy becomes effective in a country if mandatory environmental, social and governance disclosure are
introduced all at once, or zero if the country implements mandatory disclosure gradually. GSHOCK is a dummy variable that is equal to one for the years 2001, 2008
and 2020, and zero otherwise. CO2 is the natural logarithm of CO2 emissions in metric tons per capita. CGRF is an indicator variable equal to one for all years after a
major corporate governance reform became effective in the country, and zero otherwise. LIQ is the legal institution quality of a country, which is measured based on
rule of law, regulatory quality, government effectiveness, and protection against self-dealing. MKTCAP is the annual market capitalisation of listed domestic companies
divided by GDP. GDPG is GDP growth (annual %). SIZE is the natural logarithm of the market capitalisation of a firm in U.S. dollars. IO is the number of shares held by
all types of institutions divided by the total number of shares outstanding. PROFIT is earnings before interest, taxes and depreciation divided by total sales. LEV is total
debt divided by total assets. BM is the book value of equity divided by market capitalisation. CASH is cash and short-term investments divided by total assets. CAPEX is
capital expenditure divided by total assets. TURN is the share trading volume divided by adjusted shares outstanding. NUMEST is the natural logarithm of the number
of analysts following a firm in a fiscal year. VOLTY is the standard deviation of daily stock returns over the 365 days prior to the fiscal year end date. All continuous
variables are winsorized at the top and bottom 1%. Standard errors are clustered at the country level. The t-statistics are reported in parentheses. ***, ** and * denote
significance at the 1%, 5% and 10% levels, respectively (two-tailed tests).
Columns 1 and 2 of Panel B of Table 5 present the results of excluding 2003 respectively. The governance reforms took place within the event
treatment countries with simultaneous reforms. Australia, Canada and windows of the ESG disclosure mandates in those countries. To rule out
France launched major corporate governance reforms in 2004, 2004 and the impact of confounding events, we exclude firms from those three
9
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Panel E: Pooled OLS estimation with industry, country and year effects
Dependent variable PSI TIMELINESS
(1) (2)
ESGPOST 0.112*** − 0.008***
(continued on next page)
10
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
19
ESG-sensitive industries include firms from utility, mining, oil & gas, and
2020. We create an indicator variable, Environment first, that takes the
chemicals industries (Bolton & Kacperczyk, 2021), which produce a signifi
cantly larger fraction of carbon emissions, compared to other industries. value one for countries where more of the population agree that envi
Because we have already removed utility firms from our initial sample, we ronmental protection should be given priority over economic growth
further remove ESG-sensitive industries such as mining, oil & gas, and and zero otherwise. Environment first is estimated based on responses to
chemicals. the WVS question: “Which of the following comes closer to your own
20
The industry dummy variables are based on the Fama-French 48 industry point of view? 1. Protecting the environment should be given priority,
groups.
11
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
even if it causes slower economic growth and some loss of jobs. 2. Table 7
Economic growth and creating jobs should be the top priority, even if the Heterogeneous effects based on corporate governance and institutional quality.
environment suffers to some extent.” We recode the response to this Panel A: Heterogeneous effects based on corporate governance quality
question to one if a survey participant chose statement 1 and zero
Dependent variable PSI TIMELINESS
otherwise. We then calculate the mean score of the responses for each
country-wave. Within a wave, the score is calculated once and applies to (1) (2)
all country-years covered by the wave. An average score is finally ESGPOST × High corporate governance quality 0.037 − 0.006*
calculated for each country across the waves of the WVS. Higher scores (0.56) (− 1.92)
ESGPOST × Low corporate governance quality 0.174*** − 0.011***
suggest that more people put environmental protection ahead of eco
(2.89) (− 3.73)
nomic growth. We create an indicator variable, Environment first, that Number of observations 78,815 79,689
takes the value one if the score is above the sample median and zero Adjusted R-squared 0.333 0.170
otherwise. We re-estimate Eq. (5), replacing the ESGPOST indicator with Control variables Yes Yes
ESGPOST × Environment first and ESGPOST × Not environment first. The Firm fixed effects Yes Yes
Year fixed effects Yes Yes
Not environment first indicator is defined as (1 – Environment first). Panel
A of Table 6 presents the results. The coefficients on ESGPOST × Envi
Panel B: Heterogeneous effects based on legal institution quality
ronment first are significant at the 1% or 5% level, whereas the co
Dependent variable PSI TIMELINESS
efficients on ESGPOST × Not environment first are either insignificant or (1) (2)
significant at the 10% level. This result indicates that the effect of ESG ESGPOST × High legal institution quality − 0.013 − 0.005*
disclosure reform is more pronounced in countries where the value of (− 0.14) (− 1.77)
putting the environment first prevails. ESGPOST × Low legal institution quality 0.169*** − 0.010***
(2.72) (− 3.12)
Second, the effects of ESG disclosure reform may depend on firms' Number of observations 78,815 79,689
status of ESG reporting. If it is the first time a firm produces and/or Adjusted R-squared 0.333 0.170
submits ESG-related reports due to mandatory ESG disclosure regula Control variables Yes Yes
tions, the impact of the mandatory ESG disclosure should lead to a Firm fixed effects Yes Yes
Year fixed effects Yes Yes
completely new set of information being available to market partici
pants, and thus its impacts on price discovery should be more pro Panel A reports the results of examining the relative effects of ESG disclosure
nounced for the firm. We create an indicator variable, ESG reporting reform on market efficiency for different firms based on corporate governance
firms, that takes the value one if a firm has ESG reports uploaded in the quality. High corporate governance quality is an indicator variable that takes the
Refinitiv ESG database (formerly known as ASSET4) in a year and zero value one for firms whose average corporate governance score is above the
sample median, and zero otherwise. Low corporate governance quality is (1 – High
otherwise. In the regression estimation, we replace the ESGPOST indi
corporate governance quality). Panel B reports the results of examining the rela
cator with ESGPOST × ESG reporting firms and ESGPOST × Non ESG
tive effects of ESG disclosure reform on market efficiency in different countries
reporting firms. The Non ESG reporting firms indicator is defined as (1 –
based on institutional quality. High legal institution quality is an indicator variable
ESG reporting firms). Panel B of Table 6 presents the results. The co that takes the value one for countries whose legal institution quality index (LIQ)
efficients on ESGPOST × Non ESG reporting firms are significant at the 1% is above the sample median, and zero otherwise. Low legal institution quality is (1
or 5% level. The coefficients on ESGPOST × ESG reporting firms are not – High legal institution quality). All the control variables used in Table 4 are also
significant. This result indicates that the effect of ESG disclosure reform included. All continuous variables are winsorized at the top and bottom 1%. The
is more pronounced for firms without a previous ESG information results for the control variables and fixed effects are not reported for brevity.
disclosure practice. Standard errors are clustered at the country level. The t-statistics are reported in
Third, it is likely that firms with good ESG performance are more parentheses. ***, ** and * denote significance at the 1%, 5% and 10% levels,
forthcoming with ESG-related news, regardless of the launch of ESG respectively (two-tailed tests).
disclosure mandates. Thus, the effects of ESG disclosure mandates
should be more pronounced for firms with poor ESG performance. We
index of corporate governance quality (CGQ) with 22 underlying
use the ESG score from the Refinitiv ESG database to measure ESG
governance characteristics. Appendix B gives details of the governance
performance. We create an indicator variable, High ESG score, that takes
items and criteria. If a firm meets a characteristic successfully in a given
the value one if a firm has an ESG performance score above the sample
year, it will score one point and zero otherwise. We weight all charac
median in a year, and zero otherwise. In the regression estimation, we
teristics equally to obtain the total CGQ index for a year. The average
replace the ESGPOST indicator with ESGPOST × High ESG score and
CGQ index is then calculated to represent the overall corporate gover
ESGPOST × Low ESG score. The Low ESG score indicator is defined as (1 –
nance quality of a firm during the sample period. In the regression
High ESG score). Panel C of Table 6 presents the results. The coefficients
estimation, we replace the ESGPOST indicator with ESGPOST × High
on ESGPOST × Low ESG score are significant at the 1% or 5% level. The
corporate governance quality and ESGPOST × Low corporate governance
coefficients on ESGPOST × High ESG score are not significant. This result
quality. The Low corporate governance quality indicator is defined as (1 –
indicates that the effect of ESG disclosure reform is more pronounced for
High corporate governance quality). Panel A of Table 7 presents the re
firms with low ESG performance scores.
sults. The coefficients on ESGPOST × Low corporate governance quality
Fourth, prior research documents that good corporate governance
are significant at the 1% level for both the PSI and TIMELINESS models.
improves disclosure and price efficiency (Beekes et al., 2016; Beekes &
The coefficient on ESGPOST × High corporate governance quality is sig
Brown, 2006; Kacperczyk et al., 2021). The mandatory ESG disclosure
nificant at the 10% level for the TIMELINESS model. This result shows
reforms are likely to play a governance role if firms lack sound internal
that the effect of ESG disclosure reform is more pronounced for firms
governance. Thus poorly governed firms are expected to be more
with poorer corporate governance, suggesting that external mandatory
affected by the new rules than well-governed firms. To test the conjec
ESG reform substitutes for internal corporate governance in enhancing
ture, we create an indicator variable, High corporate governance quality,
price efficiency.
that takes the value one for firms whose average corporate governance
Lastly, considering that country-level legal institutions influence
score is above the sample median, and zero otherwise. We adopt the
investor protection, corporate governance and firm value (La Porta,
approach of Chung, Elder, and Kim (2010) and construct a firm-level
Lopez-De-Silanes, Shleifer, & Vishny, 1997, 1998, 2000, 2002), the
12
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
treatment effects are likely to be different across countries with different 1. ΔIOi,t+1 is the absolute value of institutional ownership in year t + 1
legal institution quality. On one hand, better legal institutions may help minus institutional ownership in year t. ESGPOST is defined as for Eq.
ESG disclosure reform to take effect via stronger enforcement of rules (5). The vector of controls includes all the firm- and country-specific
and regulations, and therefore the treatment effect is likely to be more variables controlled in the PSI and TIMELINESS models. We include
pronounced in countries with better institutional quality. On the other firm age (AGE) and stock price momentum (MM1) as additional control
hand, mandatory ESG disclosure reform may substitute for legal in
stitutions in affecting price efficiency because the reforms can be Table 8
implemented through other channels than legal institutions. To explore The impact of mandatory ESG disclosure on stock returns.
the empirical question, we create an indicator variable, High legal insti Dependent variable RET ΔIO TQ
tution quality, that takes the value one if the legal institution quality
(1) (2) (3)
index of a country (LIQ) is above the sample median, and zero otherwise.
We re-estimate the baseline models by replacing the ESGPOST indicator ESGPOST − 0.071*** 0.004** 0.154***
with ESGPOST × High legal institution quality and ESGPOST × Low legal (− 5.72) (2.16) (3.54)
Country-level controls
institution quality. The Low legal institution quality indicator is defined as GSHOCK − 0.047 0.009 − 0.046
(1 – High legal institution quality). Panel B of Table 7 presents the results. (− 0.90) (1.12) (− 0.27)
The coefficients on ESGPOST × Low legal institution quality are significant CO2 0.002 0.002** − 0.14
at the 1% level for both the PSI and TIMELINESS models. The coefficient (0.35) (2.47) (− 0.71)
CGRF − 0.013 − 0.010** − 0.064
on ESGPOST × High institutional quality is significant at the 10% level for
(− 0.61) (− 2.29) (− 1.12)
the TIMELINESS model. This result indicates that the effect of ESG LIQ − 0.043 0.002 − 0.519**
disclosure reform is more pronounced in countries with weak legal (− 0.70) (0.27) (− 2.32)
institutions. MKTCAP 0.051 0.008** 0.474***
(1.02) (2.34) (2.94)
GDPG 0.011** 0.001** − 0.009
4.4. Does mandatory ESG disclosure lead to real stock market changes? (2.40) (2.21) (− 1.09)
Firm-level controls
AGE 0.031*** − 0.005 − 0.074*
The results of our main tests shed some light on the contempora (2.93) (− 1.41) (− 1.70)
neous impact of the ESG disclosure policy on market efficiency. We now MM1 − 0.054*** 0.001** 0.355***
test whether investors indeed care about the changes due to the new (− 12.22) (2.50) (10.24)
policy and whether mandatory ESG disclosure leads to real stock market SIZE − 0.146*** − 0.005*** − 0.249***
(− 16.63) (− 3.50) (− 7.34)
change in the longer term. Relevant theories suggest that, if the ESG
IO − 0.047*** 0.006 0.280***
disclosure reforms help to reduce private information asymmetry, a (− 2.83) (0.69) (8.04)
lower cost of equity capital can be expected (Brown et al., 2004; Easley PROFIT − 0.080** − 0.002 0.891***
& O'Hara, 2004). Previous studies have integrated relevant ESG factors (− 2.42) (− 1.04) (3.80)
LEV 0.237*** 0.017*** − 0.635***
into analyses of the pattern of stock returns. For example, Chava (2014)
(9.01) (2.62) (− 7.47)
finds that firms that derive substantial revenues from the sale of coal or BM 0.175*** 0.003 − 0.617***
oil are associated with a higher implied cost of capital. Bolton and (9.57) (1.11) (− 5.99)
Kacperczyk (2021) find that carbon emissions of U.S. firms are signifi CASH − 0.057 − 0.004** 0.800***
cantly positively related to cross-sectional stock returns, suggesting that (− 1.52) (− 2.41) (5.83)
CAPEX − 0.413*** − 0.009 1.553***
investors have demanded compensation for their exposure to carbon
(− 5.98) (− 1.30) (3.85)
emission risk. The authors also find that institutional investors indeed TURN − 0.019*** 0.001*** − 0.002
divest from firms associated with high carbon emissions. Pedersen et al. (− 4.47) (6.80) (− 0.20)
(2021) sort stocks into quintiles based on individual ESG proxies and NUMEST − 0.046*** − 0.002 0.146***
then form portfolios that go long on the best ESG stocks and short on the (− 7.51) (− 1.61) (3.87)
VOLTY 3.009** 0.456*** 2.468
worst ESG stocks. The authors find that the portfolio based on G (i.e., (2.32) (4.99) (0.90)
governance) earns significant abnormal returns. They also find that the Number of observations 71,198 74,718 76,787
ESG proxies are positively associated with institutional holdings in Adjusted R-squared 0.122 0.048 0.253
favour of greener firms. Firm fixed effects Yes Yes Yes
Year fixed effects Yes Yes Yes
In our first test, we evaluate ESG disclosure reform based on its
ability to forecast future stock returns. In addition, we examine the This table presents the regression results for the impact of mandatory ESG
change in institutional ownership after ESG disclosure reform. We disclosure on the stock return, change in institutional ownership and firm
calculate annual market-adjusted stock returns using the same estima valuation. In column 1, the dependent variable is the stock return (RET),
tion window as that defined in the estimation of TIMELINESS. The measured as the market-adjusted annual returns of a share in year t + 1. In
column 2, the dependent variable is the change in institutional ownership (ΔIO),
change in institutional ownership is calculated as the absolute value of
measured as the absolute value of institutional ownership in year t + 1 minus
institutional ownership in year t + 1 minus institutional ownership in
institutional ownership in year t. In column 3, the dependent variable is Tobin's q
year t.21 We estimate the following fixed-effects regression model: (TQ), measured as total assets minus the book value of equity plus the market
∑ value of equity divided by total assets in year t. RET and ΔIO lead the explan
RET i,t+1 orΔIOi,t+1 = α + β1 (ESGPOST) + βm CONTROLSi,c,t
(6) atory variables by one year, and TQ is contemporaneous with the explanatory
+ FIRM FE + YEAR FE + εi,t variables. ESGPOST is an indicator variable that is equal to one for all years after
the mandatory ESG disclosure policy becomes effective in a country, and zero
where RETi,t+1 refers to the annual stock return of company i in year t + otherwise. AGE is the natural logarithm of firm age. MM1 stands for price mo
mentum, measured as the compounded returns over the previous 12 months.
Other control variables are the same as those used in Table 4. Continuous var
21
We use the absolute value of the change in institutional ownership because iables are winsorized at the top and bottom 1% to control for outliers. Standard
the ESG disclosure mandates can either increase or decrease institutional errors are clustered at the country level. The t-statistics are reported in paren
ownership, and therefore it is their institutional ownership rebalancing activ theses. ***, ** and * denote significance at the 1%, 5% and 10% levels,
ities (including both buying and selling) that matter, which increases price respectively (two-tailed tests).
efficiency after the ESG disclosure mandates.
13
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
variables (e.g., Fauver et al., 2017; Kacperczyk et al., 2021). Firm fixed How, though, does the stock market respond to the changes in relation to
effects and year fixed effects are also included. We cluster standard er ESG developments and what are the real impacts of mandatory ESG
rors at the country level. The coefficient of interest is β1 . For the return disclosure on price discovery efficiency? These are fundamental ques
model, the expected sign of β1 is negative because the ESG disclosure is tions for the emerging field of ESG and accounting/finance.
likely to reduce the risk premium of information asymmetry and ulti This paper examines the impact of mandatory ESG disclosure re
mately reduce investors' expected return and cost of equity capital. For forms launched around the world since the early 2000s on price dis
the institutional ownership model, the expected sign of β1 is positive, covery efficiency in 45 countries. We undertake a DID analysis and find
meaning that investors may either increase their investments in firms robust evidence that the ESG disclosure reform has significantly
associated with good ESG performance or divest from firms associated improved price discovery efficiency. Heterogeneous treatment tests
with poor ESG performance after the ESG disclosure reforms become reveal the conditions under which the reforms take effect. The effects of
effective. the ESG reforms are stronger for firms in countries that value and pri
We are also interested in testing whether the ESG reforms will ulti oritize environmental protection, for firms in countries with low insti
mately influence firm value, given the alleged effects on price efficiency tutional quality, for firms that do not release ESG-related documents, for
and returns. Firm value is measured by Tobin's q, which is calculated as firms with poor ESG performance, and for firms that lack complete
total assets minus the book value of equity plus the market value of corporate governance mechanisms. We find that institutional investors
equity divided by total assets. We estimate the following fixed-effects care about the reforms, such that the real impacts of mandatory ESG
regression model: disclosure lead to reduced future stock returns (cost of equity capital),
∑ improved institutional ownership and higher firm valuations in the post-
TQi,t = α +β1 (ESGPOST)+ βm CONTROLSi,c,t +FIRM FE+YEAR FE+ εi,t reform period.
(7) Our research has a few implications for policy makers. First, stock
exchanges now face global competition in attracting high-quality com
where TQi,t refers to the Tobin's q of company i in year t. The explanatory panies to list and in raising equity capital from investors. In order to
variables and fixed effects used are the same as those defined for Eq. (6). strengthen their competitiveness in the global financial market, it is
We cluster standard errors at the country level. The coefficient of in crucial to prioritize mandatory ESG disclosure regulations, which can
terest is β1 . The expected sign of β1 is positive because the ESG disclosure improve price discovery efficiency and reduce the cost of equity capital.
reform is likely to increase firm value if the reform reduces the cost of Second, flexibility in mandatory ESG regulation, such as a comply-or-
equity capital, all other things being equal. explain (COE) or step-by-step style, is not as effective as the strictly
We report the regression results of the additional tests in Table 8. mandatory approach in promoting price discovery efficiency. However,
Columns 1 and 2 show the results for the regression models described in it is better than no mandatory disclosure regulation at all, because it
Eq. (6). Column 3 presents the results for the regression model shown in could still generate spill-over benefits and improve the overall infor
Eq. (7). We find a negative and statistically significant effect of ESGPOST mation environment and price discovery efficiency. Finally, despite
on future stock returns (RET) (β1 = − 0.071, p < 0.01) in column 1. This mandatory disclosure having net benefits in promoting price discovery
result suggests that mandatory ESG disclosure contributes to the infor efficiency, the challenges remain in creating standardized reporting
mation set, so that risk premiums on ESG factors are reduced after the standards for ESG disclosure (Christensen et al., 2021), which could
implementation of new policies. The ESG disclosure reform decreases further reduce the cost of information collection, processing and veri
stock returns by 7.1% annualized. Our finding echoes those of Bolton fication, and maximize its benefits in promoting stock market develop
and Kacperczyk (2021) and Pedersen et al. (2021). For example, Bolton ment. Our analysis also has valuable implications for corporate
and Kacperczyk (2021) document that a one-standard-deviation in managers, board directors and investors. For management and board
crease in firms' carbon emissions can increase stock returns by up to directors, ESG disclosure practice could be perceived as a quality signal
3.6% per annum. In column 2, the change in institutional ownership by investors, and thus a way for companies to differentiate themselves
(ΔIO) shows a significant increase after the ESG reform (β1 =0.004, p < from their competitors. Although institutional investors may be able to
0.05), suggesting that institutional investors integrate ESG factors and gain insights from ESG disclosure, retail investors may need to be
adjust their portfolios accordingly. The result in column 3 shows that mindful in trading shares with a low level of ESG disclosure compliance.
firm value (TQ) increases in the post-reform period (β1 =0.154, p < Further study could extend our analysis in a couple of ways. We focus
0.01). With respect to the magnitude of the effects, the mandatory ESG on the effect of market-level ESG disclosure regulation on overall price
disclosure reforms are associated with a 9.5% increase in the change in discovery efficiency in terms of the incorporation of both financial and
institutional ownership, and an 8.11% increase in firm value, relative to non-financial information. Further studies could extend this to investi
their respective means.22 gate how firm-level financial disclosure quality and ESG disclosure
quality are affected by ESG disclosure regulation and via which channels
5. Conclusion price discovery efficiency is promoted more. Second, we find some ev
idence of spill-over effects of mandatory ESG disclosure with compliance
Both scholars such as Nobel Laureate in Economics Oliver Hart and flexibilities. Further research could extend ours by looking at the
practitioners such as CEO and chairman of BlackRock, the largest asset channels through which such spill-over effects are realized.
management firm in the world, Larry Fink argue that the purpose of
incorporation goes beyond shareholder value maximization to providing Data availability
products and solving social problems such as climate change and social
inequality. ESG disclosure mandates enable a wide audience to under Data will be made available on request.
stand the implications of firms' activities and policies on social welfare.
22
The magnitude of the impact of ESGPOST on ΔIO is calculated as 0.004 (coefficient on ESGPOST in column 2 of Table 8) ÷ 0.042 (the unreported sample mean of
ΔIO) = 9.5%. The magnitude of the impact of ESGPOST on TQ is calculated as 0.154 (coefficient on ESGPOST in column 3 of Table 8) ÷ 1.898 (the unreported sample
mean of TQ) = 8.11%.
14
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Table A1
Variable Acronym Description Data source
Stock price non- PSI PSI is a measure of stock price informativeness based on the R2 from asset pricing Refinitiv Datastream
synchronicity regressions, as shown in Eqs. (1)–(3). Authors' own calculation
Stock price timeliness TIMELINESS The price timeliness of value-relevant news, based on daily market-adjusted share Refinitiv Datastream & Worldscope
prices, as estimated in Eq. (4). The measure is deflated by one plus the absolute rate of Authors' own calculation
return on the share over the period.
Mandatory ESG ESGPOST An indicator variable that is equal to one for all years after the mandatory ESG Krueger et al. (2021)
disclosure disclosure policy becomes effective in a country, and zero otherwise.
Comply-or-explain COE An indicator variable that is equal to one for all years after the mandatory ESG Krueger et al. (2021)
disclosure policy becomes effective in a country if the comply-or-explain ESG
disclosure approach is adopted, and zero otherwise.
All-at-once ATO An indicator variable that is equal to one for all years after the mandatory ESG Krueger et al. (2021)
disclosure policy becomes effective in a country if mandatory environmental, social
and governance disclosure are introduced all at once, or zero if the country
implements mandatory disclosure gradually.
Global shocks GSHOCK An indicator variable for global shocks that equals one if the year is 2008 or 2020, and Authors' own calculation
zero otherwise.
Carbon emissions CO2 CO2 is CO2 emissions in metric tons per capita. Emissions Database for Global Atmospheric
Research (EDGAR)
https://edgar.jrc.ec.europa.eu/report_2022
Corporate CGRF An indicator variable equal to one for all years after a major corporate governance Fauver et al. (2017)
governance reform reform became effective in the country, and zero otherwise.
Legal institution LIQ Sum of three World Governance Indicators (government effectiveness, regulatory World Governance Indicators
quality quality and the rule of law) and the anti-self-dealing index from Djankov et al. (2008). Djankov et al. (2008)
The WGI and anti-self-dealing indices are rescaled to be between 0 and 1.
Stock market MKTCAP Market capitalisation of listed domestic companies divided by GDP. World Development Indicators
capitalisation
GDP growth GDPG Annual percentage growth rate of GDP. World Development Indicators
Environment first – The environment first score of a country is estimated based on responses to the WVS Four waves of the World Values Survey (WVS) in
question: “Which of the following comes closer to your own point of view? 1. 1999–2004, 2005–2009, 2010–2014 and
Protecting the environment should be given priority, even if it causes slower 2017–2020.
economic growth and some loss of jobs. 2. Economic growth and creating jobs should
be the top priority, even if the environment suffers to some extent.” We recode the
response to this question to one if a survey participant chose statement 1, and zero
otherwise. We then calculate the mean of the response for each country-wave. The
score is calculated for each wave of the WVS. Within a wave, the score is calculated
once and applies to all country-years covered by the wave. An average score is
calculated for each country. Higher scores suggest that people put environmental
protection first.
ESG-reporting firms – An indicator variable that takes the value one if a firm has ESG reports uploaded in Refinitiv ESG Database
the Refinitiv ESG database in a year, and zero otherwise. Missing information in a
year is coded as zero.
High ESG score – An indicator variable that takes the value one if a firm has an ESG performance score Refinitiv ESG Database
above the sample median in a year, and zero otherwise. Missing information in a year
is coded as zero.
Corporate CGQ Refer to Appendix B for our corporate governance standards relating to financial and Refinitiv Eikon
governance quality operating transparency, following Chung et al. (2010). CGQ is the annual ratio of
their CG scores according to these 22 CG standards, divided by the full score of 22.
Missing information in a year is coded as zero.
Firm size SIZE The natural logarithm of the total assets of a firm in U.S. dollars (Worldscope item Refinitiv Worldscope
02999).
Institutional IO Number of shares held by all types of institutions divided by total number of shares Refinitiv Ownership Database
ownership outstanding.
Profitability PROFIT Earnings before interest, taxes and depreciation (Worldscope item 18,198) divided by Refinitiv Worldscope
total assets (Worldscope item 02999).
Leverage LEV Total debt (Worldscope item 03255) divided by total assets (Worldscope item Refinitiv Worldscope
02999).
Book-to-market ratio BM Book value of equity (Worldscope item 03501) divided by market value of equity Refinitiv Worldscope
(Worldscope item 08001).
Cash holding CASH Cash and short-term investments (Worldscope item 02001) divided by total assets Refinitiv Worldscope
(Worldscope item 02999).
Capital expenditure CAPEX Capital expenditures (Worldscope item 04601) divided by total assets (Worldscope Refinitiv Worldscope
item 02999).
Stock trading volume TURN Share trading volume (Datastream item VO) divided by adjusted shares outstanding Refinitiv Datastream
(Datastream items NOSH/AF).
Analyst following NUMEST The natural logarithm of the number of analysts following a firm in a fiscal year. I/B/E/S
Stock return volatility VOLTY The standard deviation of daily stock returns over the 365 calendar days prior to the Refinitiv Datastream
fiscal year end date. Authors' own calculation
Firm age AGE The natural logarithm of the number of years since the firm was incorporated. Refinitiv Worldscope
Stock price MM1 Compounded stock returns over the previous 12 months. Refinitiv Datastream
momentum Authors' own calculation
Stock return RET Market-adjusted annual returns of a share in year t + 1 Refinitiv Datastream
Authors' own calculation
(continued on next page)
15
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Table A1 (continued )
Variable Acronym Description Data source
Change in ΔIO The absolute value of institutional ownership in year t + 1 minus institutional Refinitiv Ownership Database
institutional ownership in year t.
ownership
Tobin's q TQ Total assets (Worldscope item 02999) minus book value of equity (Worldscope item Refinitiv Worldscope
03501) plus market value of equity (Worldscope item 08001) divided by total assets.
Table B1
Corporate Governance Categories and Standards
Audit
1 Audit committee independence is true.
Board
2 Strictly independent board members are >50% of the board directors.
3 Nomination committee independence is true.
4 Compensation committee is compromised solely of independent outside directors.
5 Committee meeting attendance average is greater than zero (i.e., meets at least once during the year).
6 Staggered Board Structure is false.
7 Size of board of directors is at least 6 but not >15 members.
8 Shareholders have cumulative voting rights to elect directors.
9 CEO serves on no more than two additional boards of other public companies.
10 Chairman is not ex-CEO.
11 CEO and Chairman is separated.
12 Any of the following policies is available publicly.
• Board Independence
• Board Diversity
• Board Experience
• Executive Compensation Performance
• Executive Compensation ESG Performance
• Executive Retention
Charter
13 There is no poison pill provision.
14 Supermajority Vote Requirement is false.
15 Shareholder Approval Significant Transactions is true.
16 Written Consent Requirements is true.
17 Limited Shareholder Rights to Call Meetings is false.
18 Unlimited Authorized Capital or Blank Check is false.
Compensation
19 Board Member Long Term Compensation Incentives is greater than zero.
Ownership
20 Executive Compensation Long Term Objectives is true.
21 Shareholders Approval Stock Compensation Plan is true.
Anti-Takeover Devices
22 There is no anti-takeover device.
This table shows the construction method for the corporate governance quality index. Following Chung et al. (2010), we adopt
22 CG standards relating to financial and operating transparency. The CG standards are taken from data compiled by Refinitiv
Eikon.
References Beekes, W., Brown, P., Zhan, W., & Zhang, Q. (2016). Corporate governance, companies’
disclosure practices and market transparency: A cross country study. Journal of
Business Finance and Accounting, 43, 263–297.
Admati, A., & Pfleiderer, P. (2000). Forcing firms to talk: Financial disclosure regulation
Beekes, W., Brown, P., & Zhang, Q. (2015). Corporate governance and the
and externalities. Review of Financial Studies, 13, 479–519.
informativeness of disclosures in Australia: A re-examination. Accounting and
Alford, A., Jones, J., Leftwich, R., & Zmijewski, M. (1993). The relative informativeness
Finance, 55, 931–963.
of accounting disclosures in different countries. Journal of Accounting Research, 31,
Bénabou, R., & Tirole, J. (2010). Individual and corporate social responsibility.
183–223.
Economica, 77, 1–19.
Amel-Zadeh, A., & Serafeim, G. (2018). Why and how investors use ESG information:
Bennett, B., Stulz, R., & Wang, Z. (2020). Does the stock market make firms more
Evidence from a global survey. Financial Analysts Journal, 74, 87–103.
productive? Journal of Financial Economics, 136, 281–306.
Back, K., Cao, H., & Willard, G. (2000). Imperfect competition among informed traders.
Bertrand, M., & Mullainathan, S. (2003). Enjoying the quiet life? Corporate governance
Journal of Finance, 55, 2117–2155.
and managerial preferences. Journal of Political Economy, 111, 1043–1075.
Baginski, S., & Hinson, L. (2016). Cost of capital free-riders. The Accounting Review, 91,
Biddle, G., Hilary, G., & Verdi, R. (2009). How does financial reporting quality relate to
1291–1313.
investment efficiency? Journal of Accounting and Economics, 48, 112–131.
Ball, R., & Brown, P. (1968). An empirical evaluation of accounting income numbers.
Bingler, J., Senni, C., & Monnin, P. (2022). Understand what you measure: Where climate
Journal of Accounting Research, 6, 159–178.
transition risk metrics converge and why they diverge. Finance Research Letters, 50,
Beaver, W. (1968). The information content of annual earnings announcements. Journal
Article 103265.
of Accounting Research, 6, 67–92.
Bolton, P., & Kacperczyk, M. (2021). Do investors care about carbon risk? Journal of
Beekes, W., & Brown, P. (2006). Do better-governed Australian firms make more
Financial Economics, 142, 517–549.
informative disclosures? Journal of Business Finance and Accounting, 33, 422–450.
Brown, S., Hillegeist, S., & Lo, K. (2004). Conference calls and information asymmetry.
Journal of Accounting and Economics, 37, 343–366.
16
Q. Zhang et al. International Review of Financial Analysis 89 (2023) 102811
Bushman, R., & Smith, A. (2001). Financial accounting information and corporate Hermalin, B., & Weisbach, M. (2012). Information disclosure and corporate governance.
governance. Journal of Accounting and Economics, 32, 237–333. Journal of Finance, 67, 195–233.
Chava, S. (2014). Environmental externalities and cost of capital. Management Science, Holden, C., & Subrahmanyam, A. (1992). Long-lived private information and imperfect
60, 2223–2247. competition. Journal of Finance, 47, 247–270.
Chen, Q., Goldstein, I., & Jiang, W. (2007). Price informativeness and investment Holmström, B., & Tirole, J. (1993). Market liquidity and performance monitoring.
sensitivity to stock price. Review of Financial Studies, 20, 619–650. Journal of Political Economy, 101, 678–709.
Chen, Y., Huang, J., Li, X., & Yuan, Q. (2022). Does stock market liberalization improve Hong, H., & Kostovetsky, L. (2012). Red and blue investing: Values and finance. Journal
stock price efficiency? Evidence form China. Journal of Business Finance and of Financial Economics, 103, 1–19.
Accounting, 49, 1175–1210. Imbens, G., & Wooldridge, J. (2009). Recent developments in the econometrics of
Chen, Y., Hung, M., & Wang, Y. (2018). The effect of mandatory CSR disclosure on firm program evaluation. Journal of Economic Literature, 47, 5–86.
profitability and social externalities: Evidence from China. Journal of Accounting and Ioannou, I., & Serafeim, G. (2019). The consequences of mandatory corporate
Economics, 65, 169–190. sustainability reporting. In A. McWilliams, D. Rupp, D. Siegel, G. Stahl, &
Chen, Y., Kelly, B., & Wu, W. (2020). Sophisticated investors and market efficiency: D. Waldman (Eds.), The Oxford handbook of corporate social responsibility:
Evidence from a natural experiment. Journal of Financial Economics, 138, 316–341. Psychological and organizational perspectives (pp. 452–489). Oxford University Press.
Chhaochharia, V., & Grinstein, Y. (2007). Corporate governance and firm value: The Kacperczyk, M., Sundaresan, S., & Wang, T. (2021). Do foreign institutional investors
impact of the 2002 governance rules. Journal of Finance, 62, 1789–1825. improve price efficiency? Review of Financial Studies, 34, 1317–1367.
Choi, J., Larsen, K., & Seppi, D. (2019). Information and trading targets in a dynamic Krueger, P., Sautner, Z., Tang, D., & Zhong, R. (2021). The effects of mandatory ESG
market equilibrium. Journal of Financial Economics, 132, 22–49. disclosure around the world. European Corporate Governance Institute Working Paper
Christensen, H., Floyd, E., Liu, L., & Maffett, M. (2017). The real effects of mandated Series in Finance, Working Paper No. 754/2021, Swiss Finance Institute Research Paper
information on social responsibility in financial reports: Evidence from mine-safety No. 21–44. https://doi.org/10.2139/ssrn.3832745
records. Journal of Accounting and Economics, 64, 284–304. Kyle, A. (1985). Continuous auctions and insider trading. Econometrica, 53, 1315–1336.
Christensen, H., Hail, L., & Leuz, C. (2021). Mandatory CSR and sustainability reporting: La Porta, R., Lopez-De-Silanes, F., Shleifer, A., & Vishny, R. (1997). Legal determinants of
Economic analysis and literature review. Review of Accounting Studies, 26, external finance. Journal of Finance, 52, 1131–1150.
1178–1248. La Porta, R., Lopez-De-Silanes, F., Shleifer, A., & Vishny, R. (1998). Law and finance.
Chung, H., Elder, J., & Kim, C. (2010). Corporate governance and liquidity. Journal of Journal of Political Economy, 106, 1113–1155.
Financial and Quantitative Analysis, 45, 265–291. La Porta, R., Lopez-De-Silanes, F., Shleifer, A., & Vishny, R. (2000). Investor protection
Claessens, S., & Laeven, L. (2003). Financial development, property rights, and growth. and corporate governance. Journal of Financial Economics, 58, 3–27.
Journal of Finance, 58, 2401–2436. La Porta, R., Lopez-De-Silanes, F., Shleifer, A., & Vishny, R. (2002). Investor protection
Crawford, S., Roulstone, D., & So, E. (2012). Analyst initiations of coverage and stock and corporate valuation. Journal of Finance, 57, 1147–1170.
return synchronicity. The Accounting Review, 87, 1527–1553. Leuz, C., & Wysocki, P. (2016). The economics of disclosure and financial reporting
Dhaliwal, D., Li, O., Tsang, A., & Yang, Y. (2011). Voluntary nonfinancial disclosure and regulation: Evidence and suggestions for future research. Journal of Accounting
the cost of equity capital: The initiation of corporate social responsibility reporting. Research, 54, 525–622.
The Accounting Review, 86, 59–100. Lim, K., Brooks, R., & Kim, J. (2008). Financial crisis and stock market efficiency:
Dhaliwal, D., Radhakrishnan, S., Tsang, A., & Yang, Y. (2012). Nonfinancial disclosure Empirical evidence from Asian countries. International Review of Financial Analysis,
and analyst forecast accuracy: International evidence on corporate social 17, 571–591.
responsibility disclosure. The Accounting Review, 87, 723–759. Lins, K., Servaes, H., & Tamayo, A. (2017). Social capital, trust, and firm performance:
Djankov, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2008). The law and The value of corporate social responsibility during the financial crisis. Journal of
economics of self-dealing. Journal of Financial Economics, 88, 430–465. Finance, 72, 1785–1824.
Downar, B., Ernstberger, J., Reichelstein, S., Schwenen, S., & Zaklan, A. (2021). The Massa, M., Zhang, B., & Zhang, H. (2015). The invisible hand of short selling: Does short
impact of carbon disclosure mandates on emissions and financial operating selling discipline earnings management? Review of Financial Studies, 26, 1701–1736.
performance. Review of Accounting Studies, 26, 1137–1175. Merton, R. (1987). A simple model of capital market equilibrium with incomplete
Durnev, A., Morck, R., Yeung, B., & Zarowin, P. (2003). Does greater firm-specific return information. Journal of Finance, 42, 483–510.
variation mean more or less informed stock pricing? Journal of Accounting Research, Morck, R., Yeung, B., & Yu, W. (2000). The information content of stock markets: Why do
41, 797–836. emerging markets have synchronous stock price movements? Journal of Financial
Easley, D., & O’Hara, M. (2004). Information and the cost of capital. Journal of Finance, Economics, 58, 215–260.
59, 1553–1583. Park, A., & Ravenel, C. (2013). Integrating sustainability into capital markets: Bloomberg
Edmans, A. (2009). Blockholder trading, market efficiency, and managerial myopia. LP and ESG’s quantitative legitimacy. Journal of Applied Corporate Finance, 25,
Journal of Finance, 64, 2481–2513. 62–67.
Fama, E. (1965). The behavior of stock-market prices. Journal of Business, 38, 34–105. Pedersen, L., Fitzgibbons, S., & Pomorski, L. (2021). Responsible investing: The ESG-
Fama, E., & French, K. (1992). The cross-section of expected stock returns. Journal of efficient frontier. Journal of Financial Economics, 142, 572–597.
Finance, 47, 427–465. Piotroski, J., & Roulstone, B. (2004). The influence of analysts, institutional investors,
Fauver, L., Hung, M., Li, X., & Taboada, A. (2017). Board reforms and firm value: World and insiders on the incorporation of market, industry, and firm-specific information
evidence. Journal of Financial Economics, 125, 120–142. into stock prices. The Accounting Review, 79, 1119–1151.
Fernandes, N., & Ferreira, M. (2008). Does international cross-listing improve the Qian, E., Sorensen, E., & Hua, R. (2007). Information horizon, portfolio turnover, and
information environment. Journal of Financial Economics, 88, 216–244. optimal alpha models. Journal of Portfolio Management, 34, 27–40.
Fernandes, N., & Ferreira, M. (2009). Insider trading laws and stock price Raddatz, C. (2006). Liquidity needs and vulnerability to financial underdevelopment.
informativeness. Review of Financial Studies, 22, 1845–1887. Journal of Financial Economics, 80, 677–722.
Ferreira, D., Ferreira, M., & Raposo, C. (2011). Board structure and price Roll, R. (1988). R2. Journal of Finance, 43, 541–566.
informativeness. Journal of Financial Economics, 99, 523–545. Schiemann, F., & Tietmeyer, R. (2022). ESG controversies, ESG disclosure and analyst
Fiechter, P., Hitz, J., & Lehmann, N. (2022). Real effects of a widespread CSR reporting forecast accuracy. International Review of Financial Analysis, 84, Article 102373.
mandate: Evidence from the European Union’s CSR Directive. Journal of Accounting Shroff, N., Verdi, R., & Yost, B. (2017). When does the peer information environment
Research, 60, 1499–1549. matter? Journal of Accounting and Economics, 64, 183–214.
Fishman, M., & Hagerty, K. (1992). Insider trading and the efficiency of stock prices. The Verrecchia, R. (2001). Essays on disclosure. Journal of Accounting and Economics, 32,
Rand Journal of Economics, 23, 106–122. 97–180.
Gao, H., & Zhang, W. (2017). Employment nondiscrimination acts and corporate Yan, X., & Zhang, Z. (2009). Institutional investors and equity returns: Are short-term
innovation. Management Science, 63, 2982–2999. institutions better informed? Review of Financial Studies, 22, 893–924.
Grewal, J., Riedl, E., & Serafeim, G. (2019). Market reaction to mandatory nonfinancial Zhang, Q., Zhang, X., Chen, D., & Strange, R. (2022). Market discipline or rent extraction:
disclosure. Management Science, 65, 3061–3084. Impacts of share trading by foreign institutional investors in different corporate
Grossman, S., & Stiglitz, J. (1980). On the impossibility of informationally efficient governance and investor protection environments. International Review of Financial
markets. American Economic Review, 70, 393–408. Analysis, 79, Article 101965.
Gul, F., Kim, J., & Qiu, A. (2010). Ownership concentration, foreign shareholding, audit Zhang, X., Yang, X., Strange, R., & Zhang, Q. (2017). Informed trading by foreign
quality, and stock price synchronicity: Evidence from China. Journal of Financial institutional investors as a constraint on tunneling: Evidence from China. Corporate
Economics, 95, 425–442. Governance: An International Review, 25, 222–235.
Gul, F., Srinidhi, B., & Ng, A. (2011). Does board gender diversity improve the Zhang, X., Zhang, Q., Chen, D., & Gu, J. (2019). Financial integration, investor protection
informativeness of stock prices? Journal of Accounting and Economics, 51, 314–338. and imbalanced optimistically biased information timeliness in emerging markets.
Haß, L., Vergauwe, S., & Zhang, Q. (2014). Corporate governance and the information International Review of Financial Analysis, 64, 38–56.
environment: Evidence from Chinese stock markets. International Review of Financial
Analysis, 36, 106–119.
17