J. Account.

Public Policy 46 (2024) 107225

Contents lists available at ScienceDirect

Journal of Accounting and Public Policy

journal homepage: www.elsevier.com/locate/jaccpubpol

Full length article

IFRS 7 adoption and bank risk taking

Gerald J. Lobo a, Chong Wang b, *, Feng (Harry) Wu c
a
C. T. Bauer College of Business, University of Houston, 4250 Martin Luther King Blvd, Houston, TX 77004, USA
b
School of Accounting and Finance, Li Ka Shing Tower, The Hong Kong Polytechnic University, 11 Yuk Choi Rd, Hung Hom, Kowloon, 00000, Hong
Kong
c
Department of Accountancy, Simon and Eleanor Kwok Building, Lingnan University, 8 Castle Peak Road, Tuen Mun, New Territories, 00000, Hong
Kong

A R T I C L E I N F O A B S T R A C T

JEL Codes: We examine whether and how banks’ risk disclosure relates to their risk taking behavior. Effective
G21 from 2007, International Financial Reporting Standards No. 7 (IFRS 7) requires the disclosure of
G28 risk arising from financial instruments, which enhances bank risk disclosure. Using a difference-
G32
in-differences analysis, we document a significant decrease in bank risk taking following IFRS 7
M41
M48
adoption. This effect is more prominent when accounting rules are more strictly enforced and
alternative bank risk constraining means are weak. We also find that after IFRS 7 adoption, banks
Keywords:
are more prudent in loan offering that helps reduce risk. Overall, our results are consistent with
IFRS 7
Risk disclosure the market discipline view of bank risk disclosure and underscore IFRS 7′s role in improving
Bank risk taking financial stability.

1. Introduction

Accounting standard setters posit that a core goal of financial reporting is to aid investors in estimating the riskiness of firms’ cash
flows.1 The importance of fulfilling this goal manifested prominently after the 2007–2008 global financial crisis (GFC) when regulatory
authorities believed that the opacity of banks, especially the riskiness of their operations, contributed to the crisis. Guided by this
belief, in the wake of the GFC, policy makers around the world promoted new rules to enhance the transparency of bank risk taking.2
The rationale for improved transparency is that the unveiled risk exposure tends to invoke investor and depositor responsiveness,
which in turn disciplines banks’ imprudent risk taking (Martinez Peria and Schmukler, 2001; Nier and Baumann, 2006; Stephanou,
2010; Bushman and Williams, 2012). Nevertheless, there is longstanding doubt about the desirability of bank transparency. A sig­
nificant amount of theoretical work suggests that more public disclosure of banks’ negative states is actually undesirable because it
could weaken the informational value of price signals and thereby destabilize asset values (Plantin and Tirole, 2018), impair market

* Corresponding author at: School of Accounting and Finance, Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, China.
E-mail addresses: [email protected] (G.J. Lobo), [email protected] (C. Wang), [email protected] (F.(H. Wu).
1
Refer to, e.g., Statement of Financial Accounting Concepts No. 8 (Financial Accounting Standards Board 2010), available at https://www.fasb.
org/document/blob?fileName=Concepts_Statement_No_8.pdf, and Chapter 1 of The Conceptual Framework for Financial Reporting (International
Accounting Standards Board 2010), available at https://www.ifrs.org/content/dam/ifrs/publications/pdf-standards/english/2021/issued/part-a/
conceptual-framework-for-financial-reporting.pdf.
2
For example, the Dodd-Frank Act of 2010 requires the Federal Reserve Board to publish stress test results of large financial institutions in the U.
S.; the Financial Stability Board created the Enhanced Disclosure Task Force in 2012 to recommend best practices of financial reporting for financial
institutions; the European Central Bank, from 2013, requires loan-level reporting regarding instruments as collateral for repo borrowing.

https://doi.org/10.1016/j.jaccpubpol.2024.107225

Available online 7 June 2024

0278-4254/© 2024 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

discipline (Dang,Gorton,Holmstrom,and Ordonez, 2017), foster excessive risk exposure (Sapra, 2002), and precipitate bank runs
(Morris and Shin, 2002). Historically, bank information disclosure was limited and the tasks of controlling bank risk taking were left
largely to regulatory means (Boro, 1986).3 Such debates call into question the effectiveness of the recent regulatory efforts to promote
bank risk disclosure in constraining risk taking and fostering the stability of financial institutions.
To shed light on this issue, we exploit the mandatory adoption of International Financial Reporting Standards No. 7 – Financial
Instruments: Disclosures (henceforth IFRS 7) among banks around the world, which requires the reporting of information about risk
arising from financial instruments and the management of that risk. IFRS 7 is especially relevant for the banking industry because
financial instruments account for the majority of banks’ total assets and liabilities. Existing studies (e.g., Bischof, 2009) have shown
that IFRS 7 adoption substantially improves risk disclosure and has a sizable impact on the usefulness of the disclosure in banks’
financial statements. Against this backdrop, we examine how the enhanced bank risk disclosure required by IFRS 7 influences bank risk
taking behaviors.
Several features of the implementation of IFRS 7 help us to better understand how risk disclosure mandates in the financial system
affect financial stability. First, IFRS 7 is one of the most significant changes in how firms account for financial instruments in decades
(Conti and Mauri, 2008). The World Bank (2012) notes that IFRS 7 is the only regulation that requires a financial firm to disclose the
level and nature of its risks in both quantitative and qualitative terms. Tweedie (2005, n.p.), the International Accounting Standards
Board (IASB) Chairman at the time, claimed that “IFRS 7 will lead to greater transparency about the risks that entities run … (and) will
provide better information for investors and other users of financial statements to make informed decisions about risk and return.” The
operational module of IFRS 7 has attracted great attention from regulatory authorities, investors, and creditors around the world
(Nadia and Rosa, 2014).4 IFRS 7 is also among the most intensively debated accounting standards and it took years to revise the risk
reporting requirements due to ongoing resistance from companies that attempted to prevent the increased disclosure (Erkens, 2016).
Understanding the efficacy of IFRS 7 in influencing bank risk taking is therefore relevant and necessary.
Second, the IFRS 7 mandate, in particular, risk disclosure, provides an ideal setting to more directly answer the longstanding
question about “how public disclosure of banks’ risk exposure affects banks’ risk taking incentives (italic emphasis added)” (Cordella
and Yeyati, 1998, p.110).5 This distinguishes IFRS 7 from prior studies on regulatory efforts to improve generic bank transparency (e.g.,
Granja, 2018). Recent theoretical development (Heinle and Smith, 2017; Heinle,Smith,and Verrecchia, 2018) shows that the mech­
anism for traditional disclosure models, in which information receivers use noisy signals to indirectly infer and update the uncertainty
in cash flow, does not carry over to risk disclosure regarding the higher moments that sends a signal directly concerning the cash flow
distribution and the possible premium attached to it. In short, risk reporting under IFRS 7 provides more precise and novel information
beyond bank disclosure in general and deserves specific academic attention.
Third, IFRS 7 is an international accounting standard that was mandatorily implemented on some, but not all, banks throughout the
world. Unlike some significant national rule changes (such as the Sarbanes-Oxley Act and the Dodd-Frank Act) that usually affect the
entire economy, the implementation of IFRS 7 provides a unique opportunity to separate the effects of the accounting rule change from
those of concurrent macroeconomic events. In particular, the IFRS 7 setting enables us to conduct a difference-in-differences (DiD)
analysis through which we examine the changes in bank risk taking from before to after IFRS 7′s implementation for IFRS 7-adopting
banks relative to non-adopting banks.
We start our empirical analysis by identifying treatment banks that were required to disclosure risk information under IFRS 7 at the
beginning of 2007 and control banks that were not. We then examine the difference in risk taking measures between the pre- and post-
adoption periods in a DiD framework. We consider three primary measures of bank risk taking, following prior literature (Cheng,Hong,
and Scheinkman, 2015; Kanagaretnam,Lobo,Wang,and Whalen, 2019): the abnormal volatility of return on assets (ROA) (Abnormal
σ(ROA)), the abnormal volatility of net interest margin (NIM) (Abnormal σ (NIM)), and the abnormal Z-score (i.e., distance from
insolvency when losses exceed equity) (Abnormal Z-score). These abnormal measures are the portions of σ (ROA), σ (NIM), and Z-score
that cannot be explained by commonly used bank risk determinants and thus represent banks’ excessive or imprudent risk taking.
Using a sample of 26,084 bank-year observations from 53 countries over the period 2005 to 2010, we document robust evidence
that IFRS 7-adopting banks experience a significant decrease in all of the three risk taking measures from the pre- to the post-adoption
period, as compared to the corresponding changes over the same period for banks without IFRS 7 adoption. This finding indicates that
mandatory disclosure of financial instrument risk under IFRS 7 dampens bank risk taking and suggests that the desirable effect of
market discipline outweighs the potential destabilizing effect of bank risk disclosure. The results maintain after controlling for bank-
level risk taking determinants identified in extant studies and for country-level macroeconomic conditions. They are also robust to the
use of alternative bank risk taking measures.
The effectiveness of IFRS 7 as an information disclosing arrangement hinges on the extent to which its adoption reflects improved
risk disclosure, i.e., how IFRS 7 is enforced. Stricter enforcement renders IFRS 7 adoption a larger disclosure improvement. Consistent
with this notion, we find that the DiD effect of IFRS 7 on bank risk taking is more pronounced when accounting standards are more
strictly enforced. We also find that IFRS 7 implementation reduces bank risk by a larger degree in an environment lacking sufficient

3
For a comprehensive review, see Landier and Thesmar (2011) and Acharya and Ryan (2016).
4
A survey in 2016 by the CFA Institute shows that IFRS 7 disclosures are widely used as part of the valuation and risk analysis process. See
https://www.cfainstitute.org/-/media/documents/article/position-paper/perspective-on-financial-instrument-risk-disclosures-under-ifrs-vol-1.
ashx.
5
Similarly, Ryan (1997, p.83) suggests that “Regulators can encourage the accounting profession to develop a dynamic accounting system with a
focus on risk exposures …”.

2
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

disciplining of excessive risk taking, i.e., that caused by deposit insurance guarantee.
To explore potential channels through which IFRS 7 adoption relates to bank risk taking, we investigate bank lending behaviors
that could facilitate risk reduction. Specifically, we find that after observing the risk disclosure mandate, banks decrease risk exposure
by requiring more collateral and covenants in loan agreements. We further find that the ex post incidence of covenant violation is
reduced. These results suggest that banks better control and supervise credit risk after adopting IFRS 7. To the extent that lending
comprises a significant part of core banking business, the above loan-level evidence is consistent with risk reducing efforts by banks.
Our study assesses the effectiveness of mandatory IFRS 7 adoption by providing relevant inference for bank risk disclosure
regulation. This consideration is increasingly important, especially after the GFC. In 2012, the Financial Accounting Standards Board
(FASB) emphasized that “users of financial statements overwhelmingly indicated that … understanding a reporting entity’s exposures
to risks that are inherent in financial instruments and the ways in which reporting entities manage these risks is integral to making
informed decisions about capital allocation” (FASB, 2012, p. 4). The spirit of this statement aligns with the disciplining of bank
behavior by market players that use the disclosed risk information; our study supports this argument by showing more prudent risk
taking by banks after adopting IFRS 7. In a broader sense, our study also contributes to the fundamental issue regarding accounting
transparency and bank stability, an issue that has been long and hotly debated among academicians.6
By examining the influence of risk disclosure on disclosers’ behavior, our study adds to the literature that examines the capital
market effects of risk disclosure under different regulatory regimes, which focuses on the impact on information receivers (e.g., in­
vestors, financial analysts) rather than on information senders as in our study (Kravet and Muslu, 2013; Campbell,Chen,Dhaliwal,Lu,
and Steele, 2014; Hope,Hu,and Lu, 2016; Chiu,Guan,and Kim, 2018; Campbell, Cecchini, Ciani, Ehinger, and Werner, 2019). Our study
also supplements prior research on supervisory disclosure of the results of regulatory examination of banks, e.g., bank stress tests
(Bischof and Haske, 2013; Goldstein and Sapra, 2014; Goldstein and Yang, 2019). Unlike this strand of research, the setting of our
study does not face the challenge of separating the disclosure effect from the effect of undergoing the stress testing itself (Beatty, Iselin,
and Liao, 2023). Our study thus helps shed new light on the economics of disclosure, particularly the economic consequences of
financial reporting regulation.7

2. Conceptual framework

Our investigation addresses the real effects of IFRS 7, i.e., the change in risk taking of disclosing banks as a result of the bank risk
disclosure mandate. There is a significant literature on the real effects of accounting disclosure (Kanodia, 2006; Lambert,Leuz,and
Verrecchia, 2007; Kanodia and Sapra, 2016; Leuz and Wysocki, 2016), which identifies “situations in which the disclosing person or
reporting entity changes its behavior in the real economy … as a result of the disclosure mandate” (Leuz and Wysocki 2016, p. 530).
Kanodia and Sapra (2016) and Leuz and Wysocki (2016) note that understanding the real economic consequences of disclosure
regulation is of first-order importance to the accounting discipline. In general, accounting theories suggest that the real effects of
accounting disclosure come from the feedback of information receivers’ responses to information disclosers’ behaviors, which imposes
pressure on the disclosing firms and inhibits their undesirable behaviors. This notion is consistent with the theory of targeted trans­
parency as advocated by Fung, Graham, and Weil (2007) in that the information sent to the disclosure audiences changes their ex­
pectations regarding the disclosing firms and actions against the firms, and the disclosers’ perceptions of these expectations and actions
incentivize them to adopt behaviors that mitigate the negative impact of the issues that prompt the disclosure in the first place.
For banks’ risk disclosure in particular, its real effects relate to the desirability of bank transparency regulations regarding their
impacts on the stability of the financial system. The theoretical literature offers conflicting predictions (Beatty and Liao, 2014;
Bushman, 2014; Goldstein and Sapra, 2014; Acharya and Ryan, 2016; Ryan, 2017; Bischof,Laux,and Leuz, 2021). Improved reporting
of embedded bank risk can prevent excessive risk build-up and eventual market crash, but information (especially negative infor­
mation) disclosure could also precipitate bank runs that destabilize the banking system. These literatures set up the basic theoretical
background for the potential effects of banks’ risk disclosure via IFRS 7 on their risk taking behaviors.
One key reason IASB issued IFRS 7 is that IASB believes the disclosed information via IFRS 7 can help information receivers better
assess the financial position and performance regarding the amount, timing, and uncertainty of the disclosing firm’s cash flows, and
more information transparency allows investors and other financial statement users to make more informed judgments and decisions
after balancing risk and return. This policy objective is consistent with the creation of a real effect framework that results in an
efficiently operating market discipline on banks’ imprudent risk taking. In general, market discipline refers to a situation in which
private sector stakeholders, including stockholders, depositors, or other creditors, take actions on the basis of the costs arising as banks
undertake risks (Mathewson, 1986; Doty,Mahaffey,and Goldstein, 1991). For example, uninsured depositors are exposed to bank risk
taking and may penalize riskier banks by withdrawing their deposits or requiring higher interest rates, which pressures the relatively
risk inefficient banks to become more efficient or to exit the industry. Consistent with this rationale, some studies (e.g., Billett,
Garfinkel, and O’Neil, 1998; Martinez Peria and Schmukler, 2001; Calomiris and Jaremski, 2019) show that a deposit insurance policy
that reduces the incentives of depositors to monitor banks diminishes the degree of market discipline. The threat of market discipline
on banks also hinges on public disclosure of risk information. Calomiris and Kahn (1991), Cordella and Yeyati (1998), and Christensen,
Nikolaev, and Wittenberg-Moerman (2016) argue that the misuse or disregard of disclosure requirements is critical to the efforts to

6
See Acharya and Ryan (2016) for a review, in which they note, “the strongly opposing views in this debate indicate that additional empirical
research is needed to determine when and how more transparent financial reporting enhances stability” (p. 295).
7
See Leuz and Wysocki (2016) for a review of related literature.

3
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

mask irresponsible risk taking problems in banks.8 The role of information disclosure-induced market discipline in constraining the
disclosers’ undesirable behavior has been widely documented in the literature.9
However, another strand of literature suggests that information transparency regulation may be inefficient or even detrimental
with regard to banks’ excessive risk taking and, more generally, the impairment of financial stability. Historically, there was a mistrust
of the public’s ability to evaluate information pertaining to bank conditions, leading to “confidential supervision” rather than public
disclosure.10 This notion is rooted in the premise that banking is a complex business that requires significant expertise and costs to
comprehend the disclosed bank information.
An important consideration is price informativeness which could be degraded by improved disclosure and thus diminish the
effectiveness of market discipline. This consideration is relevant for illiquid and senior claims that dominate the key balance sheet
items of banks (Plantin,Sapra,and Shin, 2008). For instance, providing insights into the current, short-term risk profile of firms may
alter the process by which information is impounded into prices, making prices reflect liquidity concerns or trading generated by other
firms’ asset sales rather than the firm’s own future earnings power. This process creates endogenous, artificial risk that destabilizes
asset values and induces banks to retain excessive risk exposure (Allen and Carletti, 2008; Plantin et al., 2008; Bleck and Gao, 2012;
Plantin and Tirole, 2018). The damage of price discovery in the market also hampers firm learning from stock prices by preempting
traders’ information advantage established from information acquisition and, as a result, the feedback from market to firm, which is a
critical mechanism for any real effects to manifest, becomes less efficient (Gao and Liang, 2013; Goldstein and Sapra, 2014; Banerjee,
Davis,and Gondhi, 2018). More disclosure about risks may magnify the effect of a pessimistic signal (Morris and Shin, 2002; Banerjee
and Maier, 2016), and Corona, Nan, and Zhang (2019) argue that, with the prospect of a bailout, banks knowing the failure likelihood
of other banks through the disclosure of stress-test results may take risks that make a concurrent failure more likely.
Gorton (2015) and Holmstrom (2015) show that if depositors also care about other market participants’ beliefs when judging a
bank’s capitalization and solvency conditions, disclosures may impair market discipline. In this case, reporting discretion inflates
creditors’ beliefs about the fundamentals and reduces panic bank runs (Gao and Jiang, 2014). Relatedly, Dang et al. (2017) show that
keeping investors symmetrically uninformed maintains liquidity, and bank opacity carries a benefit of preventing ex post stability
compromising actions by banks’ claimants such as bank runs, which essentially are systemically dysfunctional forms of market
discipline (Acharya and Ryan, 2016).11 The reduced market discipline under these conditions implies a weakened real effect of risk
disclosure in constraining bank risk taking. Lastly, given the complex nature of information regarding bank risk, such information, if
publicly disclosed, may be ambiguously interpreted by shareholders and other stakeholders, a situation in which more information
asymmetry could arise because of more disclosure (Indjejikian, 1991; Morgan, 2002; Blankespoor,deHaan,and Marinovic, 2020),
which is detrimental to outsider monitoring of banks and thus to financial stability.12 These arguments suggest that enhanced risk
disclosure via IFRS 7 could foster risk taking.
Overall, prior literature has conflicting propositions about the desirability of bank transparency, which suggests different associ­
ations between bank risk disclosure and risk taking. Given that the predictions are theoretically unsettled, we state our hypothesis in
the null form, as follows:
H1: The adoption of IFRS 7 is not associated with a change in bank risk taking.

3. Institutional background and research design

3.1. Institutional background

IFRS 7 is a consolidated bundle of reporting requirements issued by IASB that are applicable to risks arising from financial in­
struments on– and off-balance sheet, specifically, the nature and the extent of as well as the techniques (objectives, policies, and
processes) for managing credit risk, currency risk, interest rate risk, liquidity risk, market risk, and other risks, along with information
about fair value measurement, reclassification of financial assets, collateral, offsetting and transfer of financial assets and liabilities,
and hedging accounting. The disclosure provides an overview of the significance of financial instruments in a firm and the risk

8
Bushman and Williams (2012) find that loan loss provisioning practices to smooth earnings dampen discipline over bank risk taking, which is
consistent with diminished information transparency inhibiting outside monitoring. Their discipline-related aspect focuses on discretionary ac­
counting practices rather than mandatory information disclosure.
9
For example, in the context of the National Banking Era, Granja (2018) shows that increased disclosure of state-regulated banks is associated
with lower bank failure rates and development of commercial banks. Ertan, Loumioti, and Wittenberg-Moerman (2017) find that the European
Central Bank’s disclosure requirement regarding banks’ securitization activities entails better loan quality in terms of lower default probability,
delinquent amount, and losses upon default. Madsen and McMullin (2020) document a decrease in backer support for high-risk projects in a
crowdfunding platform after the platform adds a “risks and challenges” section to project pages, which arguably increases the salience of risks.
10
“‘Confidential supervision’ means … the control of the flow of information regarding a bank’s affairs by the bank regulators through restrictions
on the quantity of information made available to the public and controls on the timing of the availability of the information that is released”
(Mathewson 1986, p. 141). Also refer to Mathewson (1986, p. 140) for the statement of the Acting Secretary of the Treasury in 1955 who opposed
the publication of the national bank examiner report in a lawsuit against an examined bank, which emphasized that depositors cannot be trusted to
act responsibly upon learning adverse information about banks.
11
The rationale is that the central role of banks is providing money-like debt claims in the financial system which should be information
insensitive. To ensure that this can be achieved requires a certain level of bank opacity rather than complete transparency.
12
We thank an anonymous reviewer for suggesting this important point.

4
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

exposure they create, based on internal information accessed by key managers. In this sense, IFRS 7 enables users of financial
statements to view a firm’s exposure to and management of risks “through the eyes of management” (PwC, 2010).
IASB issued IFRS 7 to replace International Accounting Standards (IAS) 30 – Disclosures in the Financial Statements of Banks and
Similar Financial Institutions, and carried forward the disclosure requirements in IAS 32 – Financial Instruments: Presentation and
Disclosure, both applicable to banks. IFRS 7, however, represents a substantial extension of disclosure scope and thus a significant
improvement in risk reporting. For example, IFRS 7 requires reporting banks to provide quantitative and qualitative disclosures of
market risk and liquidity risk, which were not mandated in prior bank reporting requirements. IFRS 7 also introduces disclosures of fair
value assessment of gain or loss on hedging instruments and requires additional information about the transfer of financial assets. In
general, the required level of disclosure is materially higher under IFRS 7 than under previous standards. Moreover, IFRS 7 disclosure
also helps improve the consistency and comparability of information about risk across reporting entities, facilitating a better judgment
of the risk situation by financial statement users.
IFRS 7 became effective for annual periods beginning in 2007. Bischof (2009) shows that banks that adopted IFRS 7 substantially
improved their risk disclosure quality. Therefore, IFRS 7 adoption serves as a proxy for the change in the level of risk reporting
regarding financial instruments; it is especially true for banks because more than 90 % of these financial institutions’ total assets and
liabilities are financial instruments (Bischof, 2009) and the extent of disclosure required heavily depends on the extent of use of
financial instruments and of their exposure to risks.

3.2. Research design

Our basic research design is to compare banks’ risk taking measures across the pre- and post-IFRS 7 adoption periods for the
adopting banks. To control for the effects of concurrent events, we identify a control sample of banks that are not affected by IFRS 7. We
examine how IFRS 7 adopters’ cross-period change in risk taking differs from the corresponding cross-period change for control banks,
using the following DiD model (suppressing time and bank subscripts):
RiskTaking = β0 + β1 IFRS7 × Post + β2 Ln(BankSize)+
β3 (NIMGrowth) + β4 (TooBig)+
β5 (AssetsGrowth) + β6 LLP + β7 Ln(GDP)+ (1)
β8 Ln(GDPpercapita) + β9 Inflation+
BankFixedEffects + TimeFixedEffects + ε
In Eq. (1), the dependent variable Risk Taking refers to banks’ excessive risk taking, which is measured as the abnormal portions of
commonly used risk taking proxies in prior literature. Specifically, we first compute the standard deviation of ROA, σ(ROA), the
standard deviation of NIM, σ (NIM), and Z-score. We choose σ(ROA) and σ(NIM) because previous studies show that these measures
reflect the degree of risk taking in a bank’s operations (e.g., Laeven and Levine, 2009; Houston,Lin,Lin,and Ma, 2010; Kanagaretnam
et al., 2019), and risk in internal operations, relative to other risks triggered by external economic and political factors, is more
manageable and at the discretion of managerial decisions, and thus is an appropriate indicator for bank managers’ risk taking
behavior. Z-score is a more general proxy for overall bank risk and is estimated as (− 1) times the logarithm of the sum of ROA and
capital-asset ratio divided by the standard deviation of ROA. It measures bank stability because it indicates the distance from insol­
vency (Laeven and Levine, 2009); specifically, it represents the negative of the number of standard deviations below the mean by
which profits would have to fall so as to just deplete bank capital. A larger value of Z-score thus suggests a higher level of risk taking. To
estimate the excessive portions of these risk taking measures, we follow Cheng et al. (2015) and Kanagaretnam et al. (2019) and regress
σ(ROA), σ (NIM), and Z-score on the bank-characteristic variables in Eq. (1) that influence normal risk taking. The residuals of these
regressions are used as proxies for abnormal (or excessive) risk taking, denoted by Abnormal σ(ROA), Abnormal σ(NIM), and Abnormal
Z-score, respectively.
The key independent variable is IFRS7 × Post. IFRS7 is an indicator variable that equals one for banks that adopt IFRS 7 (i.e., the
treatment group), and zero for banks that do not adopt IFRS 7 (i.e., the control group). We hand collect IFRS 7 adoption data for each
jurisdiction; our identification for the treatment sample requires that the jurisdiction endorsed IFRS 7 in 2007 and the banks operating
in the jurisdiction are also subject to the IFRS system.13 As such, we define treatment banks in a granular manner, not simply according
to their country of domicile. In other words, our treatment identification is at the bank-level rather than at the country-level as in most
previous IFRS-related studies. In a similar manner, we identify a control sample of banks from IFRS 7 non-adopting countries and also
require that the banks do not use IFRS for financial reporting. Post is an indicator for the period after the first IFRS 7 effective year, i.e.,
in our sample period, it equals one for 2007–2010 and zero for 2005–2006. Given these definitions, IFRS7 × Post equals one for IFRS 7-
adopting banks in the post-adoption period, and IFRS7 × Post equals zero for IFRS 7-adopting banks in the pre-adoption period and also
for non-adopting banks. In our DiD framework, we include bank fixed effects and time fixed effects. Bank fixed effects differentiate
banks in the treatment and control groups, and also control for differences in static bank-level attributes. Time fixed effects, which are

13
For example, in some EU member States, listed banks, but not private banks, are required to prepare financial statements according to IFRS. See
https://ec.europa.eu/info/business-economy-euro/company-reporting-and-auditing/company-reporting/financial-reporting_en.

5
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

based on the pre- and post-IFRS 7 periods, help effectively compare the cross-period change in risk taking surrounding IFRS 7 adoption,
and also capture time-series trends due to macroeconomic factors and other influences in the overall banking system. Specifying these
two-dimensional fixed effects renders it unnecessary to include the dummy variables IFRS7 and Post separately in the model.14 In this
empirical design, the DiD effect is captured by the coefficient on IFRS7 × Post; a negative value would suggest that IFRS 7-adopting
banks show a lower level of risk taking than IFRS 7 non-adopting banks; that is, the IFRS 7 adoption event that improves risk disclosure
in adopters is associated with less risk taking in the disclosing banks.15
We control for a menu of bank-level and country-level variables identified in prior literature that affect bank risk taking. The bank-
level variables include bank size Ln(Bank Size) measured as the natural logarithm of total assets, the annual growth in net interest
margin NIM Growth, an indicator of too-big-to-fail for a bank with deposit share in the top decile of the banking industry of its domicile
country Too Big, annual growth rate of total assets Assets Growth, and loan loss provision scaled by total loans LLP. The country-level
variables include the natural logarithms of GDP and GDP per capita, denoted by Ln(GDP) and Ln(GDP per capita), respectively, as well
as inflation rate Inflation. The Appendix provides detailed definitions of all variables.

4. Data and main empirical results

4.1. Data and descriptive statistics

Our study uses several datasets. We manually search for information about IFRS 7-adopting countries and regions in 2007 from
various sources, including the IFRS Foundation, PwC, and numerous academic papers and press releases, and carefully cross-check this
information. We obtain data for computing bank risk taking variables and bank-level control variables from BankScope (currently
Orbis Bank Focus), and data about GDP- and inflation-related economic variables from the World Bank. BankScope also provides
information regarding the type of accounting standards (IFRS or local GAAP) a bank uses. We use this bank-level information,
combined with the IRFS 7 adoption status in banks’ domicile countries, to identify treatment and control groups.
Table 1 shows the information about IFRS 7 adoption status, number of observations, and average country-level characteristic
variables for each of the 53 countries and regions in our sample. There are 31 jurisdictions that endorsed IFRS 7 in 2007 (i.e., IFRS7 =
1) and 22 that did not adopt the risk disclosure standards (i.e., IFRS7 = 0). Of the total of 26,084 observations, 11,982 are in the pre-
IFRS 7 period and 14,102 are in the post-IFRS 7 period (not reported in the table).
Table 2 reports basic descriptive statistics of the variables in the baseline model in Eq. (1).16 As shown in Panel A, in the full sample,
the mean of our key risk taking measures of Abnormal σ(ROA), Abnormal σ (NIM), and Abnormal Z-score are − 0.002, − 0.056 and 0.019,
respectively, and their standard deviations are 0.010, 0.050, and 1.156, respectively. Only 7.7 % of the banks are “too big” (Too Big =
0.077), suggesting that a large number of banks in our international sample are not associated with large scales (which is also
confirmed by the moderate mean and median values of Ln(Bank Size)), for which risk information disclosure could be important due to
the inherent opaqueness of small corporates, especially those in developing countries. Panels B and C report the corresponding
descriptive statistics for the treatment and control samples, respectively.

4.2. Baseline results

In Table 3, we report the baseline regression results for the DiD model in Eq. (1). Column 1, where Abnormal σ (ROA) is the
dependent variable, shows that the coefficient on IFRS7 × Post is significantly negative (coefficient = -0.002). The magnitude of the
coefficient is 20 % of the sample standard deviation (0.010, as in Panel A of Table 2) of Abnormal σ(ROA), indicating strong economic
significance. The finding suggests that if a bank adopted IFRS 7 in 2007, its excessive risk taking proxied by Abnormal σ (ROA) is
significantly reduced from the pre-adoption to the post-adoption period, relative to the corresponding change during the same period
for a bank that did not adopt IFRS 7. This evidence highlights the impact on bank risk taking, particularly from the mandatory risk
disclosure requirement of IFRS 7. We obtain similar inferences (i.e., significantly negative coefficients on IFRS7 × Post) in the model
specifications with the risk taking measures Abnormal σ(NIM) and Abnormal σ(NIM) as the dependent variables in columns 2 and 3,
respectively. Collectively, the findings in Table 3 indicate that there is a decrease in risk taking for banks that are mandated to disclose
information about their risk exposure and risk management.

4.3. Robustness of the baseline results

We address several potential concerns regarding our baseline results. Given our specific DiD setting, one concern surrounds the

14
A similar DiD design with fixed effects is adopted in prior research such as Bertrand and Mullainathan (1999, 2003), Bertrand, Duflo, and
Mullainathan (2004), Low (2009), Armstrong, Balakrishnan, and Cohen (2012), Jayaraman and Shivakumar (2013), and Christensen, Floyd, Liu,
and Maffett (2017).
15
We follow Core, Hail, and Verdi (2015) and use a jurisdiction’s implementation of IFRS 7 to proxy for an arguably exogenous enhancement of
bank-level risk disclosure quality, that is, risk disclosure is improved after IFRS 7. Later in this paper, we further explore how the enforcement of
accounting rules influences the association between IFRS 7 implementation and bank risk taking.
16
We follow the method used by Kanagaretnam et al. (2019) and average the time variant bank-level control variables for the pre- and the post-
IFRS 7 periods.

6
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Table 1
Summary of institutional variables by country/region. The full international sample includes 26,084 observations in 53 countries and regions in
the sample period of 2005–2010. This table reports the IFRS 7 adoption status (IFRS7), observation distribution across countries, and the full-sample
mean values of country-level economic variables relating to GDP (Ln(GDP)), per capita GDP (Ln(GDP per capita)), and inflation rate (Inflation). Details
about the variable definitions are provided in the Appendix.
Country/Region IFRS7 Ln(GDP) Ln(GDP per capita) Inflation Obs.

Argentina 0 26.620 9.126 12.760 182

Australia 1 27.690 10.840 4.976 73
Austria 1 26.670 10.750 2.437 54
Bulgaria 1 24.570 8.725 8.888 44
Canada 0 28.080 10.780 3.239 118
Chile 0 25.950 9.352 7.025 41
China 0 29.080 8.084 6.629 156
Colombia 0 26.250 8.651 5.202 32
Costa Rica 0 24.380 8.909 11.780 132
Croatia 1 24.850 9.540 3.817 64
Czech Republic 1 26.010 9.862 2.011 32
Denmark 1 26.520 11.010 2.607 38
Ecuador 0 24.830 8.378 7.203 49
El Salvador 0 24.380 8.145 4.413 48
Finland 1 26.270 10.790 2.356 18
France 1 28.600 10.630 2.441 127
Germany 1 28.850 10.620 1.365 68
Greece 1 26.500 10.280 2.976 53
Guatemala 0 24.380 7.940 7.004 34
Hong Kong 1 26.040 10.300 1.822 85
Hungary 1 25.630 9.506 4.012 32
India 0 27.840 7.586 5.098 203
Indonesia 0 27.140 7.881 12.600 126
Ireland 1 26.170 10.890 1.663 35
Israel 0 26.030 10.260 1.153 39
Italy 1 28.430 10.550 2.424 748
Jamaica 1 24.380 8.520 11.390 27
Japan 0 29.380 10.720 − 0.883 1,533
Latvia 1 24.380 9.412 15.430 65
Lithuania 1 24.420 9.371 7.785 32
Malaysia 0 26.130 9.046 4.906 187
Mexico 0 27.640 9.115 5.165 113
Morocco 0 25.070 7.807 2.356 33
Netherlands 1 27.440 10.830 2.052 51
Norway 1 26.780 11.230 3.935 71
Pakistan 0 25.800 7.586 7.170 71
Peru 0 25.470 8.326 2.348 54
Philippines 1 25.860 7.586 3.936 68
Poland 1 26.750 9.291 3.332 60
Portugal 1 26.200 10.030 3.020 43
Russia 0 27.980 9.205 15.690 1,369
Singapore 1 25.980 10.660 4.685 34
Slovakia 1 25.100 9.605 1.606 33
Slovenia 1 24.590 10.080 3.296 39
South Africa 1 26.570 8.854 7.799 68
South Korea 0 27.580 9.882 1.810 100
Spain 1 28.000 10.380 3.501 256
Sweden 1 26.910 10.880 2.654 27
Switzerland 1 27.030 11.200 1.768 16
Thailand 0 26.430 8.423 3.688 97
Turkey 1 27.280 9.234 6.582 73
U.K. 1 28.540 10.610 2.596 163
U.S. 0 30.320 10.810 2.940 18,770

unobserved, and therefore omitted, confounding factors that could contaminate our identification of risk disclosure enhancement from
IFRS 7. These factors may rest in various aspects of a sample bank, the banking industry, and the country, and could drive the dif­
ference between risk taking changes across IFRS 7-adopting and non-adopting banks and bias our DiD results. We mitigate this po­
tential problem in two ways. First, we break the appropriate matching of banks with the corresponding treatment/control
identifications by randomly assigning sample banks to treatment and control groups regardless of whether or not they adopted IFRS 7
in 2007. This process randomizes the impact from IFRS 7 but maintains the influences from hidden factors, if any. If these factors drive
our baseline results rather than the IFRS 7 adoption, we would observe that the main results persist; otherwise, this randomization
exercise would wipe out the DiD effects along with the removal of the correct identification of IFRS 7 adoption. Panel A of Table 4
reports the randomization results, with IFRS7 actually representing a pseudo treatment indicator in this case. The coefficient on the

7
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Table 2
Descriptive statistics of main testing variables. This table reports descriptive statistics of testing variables in the baseline model of Eq. (1),
including the dependent variables Abnormal σ(ROA), Abnormal σ (NIM), and Abnormal Z-score, DiD variables, as well as bank- and country-level
control variables. The statistics are computed from observations in the full sample, the treatment sample, and the control sample, as reported in
Panels A, B, and C, respectively. Details about the variable definitions are provided in the Appendix.
Panel A: Full sample

Mean Std. P25 Median P75

IFRS7 0.100 0.299 0.000 0.000 0.000

Post 0.541 0.498 0.000 1.000 1.000
IFRS7 × Post 0.078 0.269 0.000 0.000 0.000
σ(ROA) 0.007 0.010 0.001 0.003 0.007
σ(NIM) 0.010 0.032 0.002 0.004 0.007
Z-score − 3.654 1.328 − 4.430 − 3.625 − 2.833
Abnormal σ(ROA) − 0.002 0.010 − 0.007 − 0.004 − 0.001
Abnormal σ(NIM) − 0.056 0.050 − 0.063 − 0.052 − 0.038
Abnormal Z-score 0.019 1.156 − 0.699 0.017 0.760
Ln(Bank Size) 12.840 2.101 11.359 12.411 13.879
NIM Growth 0.132 0.366 0.017 0.058 0.125
Too Big 0.077 0.267 0.000 0.000 0.000
Assets Growth 0.037 0.079 − 0.002 0.015 0.047
LLP 0.826 1.429 0.141 0.387 0.918
Ln(GDP) 29.592 1.386 29.367 30.299 30.343
Ln(GDP per capita) 10.514 0.705 10.730 10.795 10.819
Inflation 3.705 3.472 2.661 3.218 3.218
Panel B: Treatment sample

Mean Std. P25 Median P75

IFRS7 1.000 0.000 1.000 1.000 1.000

Post 0.786 0.410 1.000 1.000 1.000
IFRS7 × Post 0.786 0.410 1.000 1.000 1.000
σ(ROA) 0.008 0.011 0.002 0.004 0.007
σ(NIM) 0.016 0.046 0.002 0.005 0.010
Z-score − 3.290 1.136 − 3.854 − 3.273 − 2.671
Abnormal σ(ROA) − 0.002 0.011 − 0.008 − 0.005 − 0.001
Abnormal σ(NIM) − 0.074 0.061 − 0.111 − 0.067 − 0.040
Abnormal Z-score 0.033 1.024 − 0.572 − 0.050 0.650
Ln(Bank Size) 15.054 2.199 13.338 15.034 16.659
NIM Growth 0.136 0.367 0.007 0.073 0.184
Too Big 0.055 0.228 0.000 0.000 0.000
Assets Growth 0.035 0.092 − 0.016 0.000 0.050
LLP 1.012 1.615 0.272 0.589 1.104
Ln(GDP) 27.311 1.345 26.286 28.015 28.435
Ln(GDP per capita) 10.230 0.732 10.035 10.552 10.623
Inflation 3.605 2.769 2.435 2.580 3.423
Panel C: Control sample

Mean Std. P25 Median P75

IFRS7 0.000 0.000 0.000 0.000 0.000

Post 0.514 0.500 0.000 1.000 1.000
IFRS7×Post 0.000 0.000 0.000 0.000 0.000
σ(ROA) 0.007 0.010 0.001 0.003 0.007
σ(NIM) 0.010 0.030 0.002 0.004 0.007
Z-score − 3.694 1.341 − 4.482 − 3.677 − 2.860
Abnormal σ(ROA) − 0.002 0.010 − 0.007 − 0.004 − 0.001
Abnormal σ(NIM) − 0.054 0.048 − 0.061 − 0.051 − 0.038
Abnormal Z-score 0.017 1.169 − 0.717 0.027 0.772
Ln(Bank Size) 12.595 1.941 11.272 12.229 13.546
NIM Growth 0.132 0.366 0.018 0.057 0.119
Too Big 0.079 0.271 0.000 0.000 0.000
Assets Growth 0.037 0.078 0.000 0.016 0.047
LLP 0.806 1.406 0.134 0.365 0.890
Ln(GDP) 29.844 1.138 30.299 30.299 30.343
Ln(GDP per capita) 10.546 0.694 10.795 10.795 10.819
Inflation 3.716 3.541 2.661 3.218 3.218

interaction term of this pseudo IFRS7 and Post becomes insignificantly positive (the p-value is larger than 0.190) no matter whether
Abnormal σ (ROA), Abnormal σ (NIM), or Abnormal Z-score is used as the dependent variable. These non-results suggest that our main
findings are unlikely to be driven by confounding variables.
Second, we follow the approach in Frank (2000) and calculate the Impact Threshold for a Confounding Variable (ITCV) for IFRS 7

8
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Table 3
IFRS 7 adoption and banks’ risk taking: Baseline results. This table reports the estimation results for the baseline model of Eq. (1). The dependent
variables are Abnormal σ(ROA) (in column 1), Abnormal σ(NIM) (in column 2), and Abnormal Z-score (in column 3). The key independent variable is
the interaction term between the treatment bank indicator IFRS7 and the indicator Post for the period after the implementation of IFRS 7. Bank and
country characteristic variables are controlled as in Eq. (1). Details about the variable definitions are provided in the Appendix. The regression
coefficients on independent variables are reported, followed by the p-values (in the parentheses) based on standard errors clustered by country. *, **,
and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
(1) (2) (3)
Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7 £ Post ¡0.002** ¡0.048*** ¡0.194*

(0.010) (0.000) (0.095)
Ln(Bank Size) − 0.003* − 0.022*** − 0.352***
(0.057) (0.001) (0.000)
NIM Growth − 0.004*** 0.006*** − 0.233***
(0.000) (0.000) (0.000)
Too Big 0.002 − 0.041** 0.167
(0.475) (0.036) (0.588)
Assets Growth − 0.014*** − 0.050 − 1.610***
(0.000) (0.120) (0.000)
LLP 0.004** − 0.011*** 0.342**
(0.015) (0.000) (0.018)
Ln(GDP) − 0.017 − 0.223* − 4.091***
(0.160) (0.074) (0.000)
Ln(GDP per capita) 0.039*** 0.135 1.637*
(0.001) (0.204) (0.076)
Inflation − 0.000* − 0.004* − 0.042*
(0.059) (0.095) (0.066)
Bank and Time FE Included Included Included
Observations 26,084 26,084 26,084
Adj. R2 0.475 0.230 0.271

adoption. ITCV helps assess how large the endogeneity problem due to a confounding factor (if any) would have to be to overturn the
effect of IFRS 7 adoption on bank risk taking. In our setting, ITCV represents the lower bound of the value of the product of two
correlations that renders the coefficient on IFRS7 × Post insignificant: one correlation is between the dependent variable (i.e., the bank
risk taking measure) and the confounding variable and the other correlation is between the independent variable (i.e., IFRS7 × Post)
and the confounding variable. In Panel B of Table 4, we report ITCVs corresponding to the three dependent variables Abnormal σ (ROA),
Abnormal σ(NIM), and Abnormal Z-score, and find that they are all negative, suggesting that the potential confounding variable is
positively correlated with one of the two variables (i.e., the independent variable and the dependent variable), and negatively
correlated with the other. This result indicates that the confounding variable, if any, would strengthen rather than weaken the effect of
IFRS 7 adoption on bank risk taking, making endogeneity problems implausible. Together, the evidence from Panels A and B implies
that our baseline finding about the IFRS 7–bank risk taking relation is unlikely to be driven by omitted confounding variables.
In addition to the approaches above, we also address the potential influence of some prominent economic and institutional factors
on our main results. One important economic factor is the GFC which could significantly affect banks’ risk taking behaviors. However,
the GFC effect is unlikely to drive the IFRS 7 effect on bank risk taking because IFRS 7 was officially implemented before the GFC.
Consequently, the accounting rule change under IFRS 7 is less subject to potential endogenous problems of the post-GFC regulations
that may result from policy makers’ ex post reactions to the crisis (Granja, 2018). Thus, the IFRS 7 setting enables a better identifi­
cation of the impact of regulation change on bank risk taking and distinguishes this impact from the mere association between
regulation and stability in the aftermath of banking crises. We also note that the GFC tends to impact the banking system of a country as
a whole, whereas our identification of treatment and control groups is at the bank level (rather than the country level), which suggests
that our bank-level finding is unlikely to be completely driven by the differential impacts of the GFC across countries. Moreover, there
is no a priori or empirical evidence showing that the GFC effect on bank risk taking is conditional on how the bank discloses its risks,
which means that although the GFC may affect risk taking, it is unlikely to be a factor confounded with IFRS 7. Given these arguments,
we nevertheless take efforts to filter out the potential influence from the GFC by excluding the years 2007 and 2008 from our sample
such that Post only refers to 2009 and 2010 in the post-IFRS 7 period, which are generally less affected by the banking crisis. We report
the re-estimated results of Eq. (1) in Panel C of Table 4 and find that all the coefficients on IFRS7 × Post remain significantly negative.
Therefore, the GFC is unlikely to subsume the association between IFRS 7 adoption and bank risk taking. Furthermore, in an unta­
blulated test, we randomly exclude a year from the post-IFRS 7 period and continue to find similar results for the IFRS7 × Post co­
efficient, suggesting that the IFRS 7 effect on bank risk taking is not sensitive to any specific periods post IFRS 7 used in our analysis,
regardless of their connection with the GFC.
As another specific institutional factor, we consider Pillar 3 under the Basel II Accord (BCBS, 2004). The Pillar 3 disclosures covered
a similar, but not identical, bank risk profile under the banking regulation framework and were in place in different phases starting in

9
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Table 4
IFRS 7 adoption and banks’ risk taking: Robustness tests. The dependent variables are Abnormal σ (ROA) (in column 1), Abnormal σ (NIM) (in
column 2), and Abnormal Z-score (in column 3) in Panels A to D, and Annual Risk-Weighted Assets (in column 1), Loan Charge-Off (in column 2), and
Loan Loss Reserve (in column 3) in Panel E. The key independent variable in all panels is the interaction term between the treatment bank indicator
IFRS7 and the indicator Post for the period after the implementation of IFRS 7, but in Panel A the proper identification of the treatment and control
groups and thus the coding of IFRS7 is disrupted via a randomization process. Panel B reports the ITCV of IFRS7 × Post in relation to the dependent
variables. In Panels C and D, observations from the financial crisis years of 2007 and 2008 and from Basel II Pillar 3-adopting countries are excluded,
respectively. Bank characteristic variables, country characteristic variables, and bank and time fixed effects, collectively denoted by Controls, are
controlled as in Eq. (1) in all panels. Details about the variable definitions are provided in the Appendix. In all panels except for Panel B, the regression
coefficient on the key DiD independent variable is reported, followed by the p-value (in the parentheses) based on standard errors clustered by
country. *, **, and *** indicate statistical significance at the 10 %, 5 %, and 1 % levels, respectively.
Panel A: Randomization test

(1) (2) (3)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7 £ Post − 0.000 − 0.017 − 0.001

(0.673) (0.190) (0.991)
Controls Included Included Included
Observations 26,084 26,084 26,084
Adj. R2 0.474 0.217 0.270
Panel B: Confounding factor analysis

(1) (2) (3)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

ITCV of IFRS7 £ Post − 0.017 − 0.081 − 0.004

Panel C: Drop financial crisis years of 2007 and 2008

(1) (2) (3)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7 £ Post ¡0.001* ¡0.040*** ¡0.289*

(0.051) (0.000) (0.074)
Controls Included Included Included
Observations 25,412 25,412 25,412
Adj. R2 0.426 0.040 0.239
Panel D: Drop Basel II Pillar 3-adopting countries

(1) (2) (3)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7 £ Post ¡0.004** ¡0.040* ¡0.499**

(0.043) (0.060) (0.015)
Controls Included Included Included
Observations 2,223 2,223 2,223
2
Adj. R 0.340 0.444 0.125
Panel E: Use alternative bank risk taking measures

(1) (2) (3)

Annual Risk-Weighted Assets Loan Charge-Off Loan Loss Reserve

IFRS7 £ Post ¡0.003** ¡0.107*** ¡0.003***

(0.015) (0.000) (0.000)
Controls Included Included Included
Observations 19,678 55,792 68,137
Adj. R2 0.498 0.641 0.636

2008.17 As noted by Bischof, Daske, Elfers, and Hail (2022), the implementation of Pillar 3 as a banking regulation was normally after
IFRS 7 came into effect, i.e., “the banking regulator is almost always second in the line of supervision” (p. 501). This sequence of
regulation implementation suggests that the influence of Pillar 3, which was introduced later, is less likely to explain our inference
regarding IFRS 7, which was introduced earlier. Despite this consideration, we still conduct a robustness check by excluding banks in
Pillar 3 adoption countries from our sample.18 We find that, as reported in Panel D of Table 4, the risk reducing DiD effects of IFRS 7
maintain in all regression specifications with different excessive risk taking measures, as evidenced by the significantly negative
coefficients on IFRS7 × Post. Thus, IFRS 7’s impact on bank risk taking is unlikely to be completely subsumed by the influence from the

17
https://www.bis.org/fsi/fsisummaries/pillar3_framework.htm.
18
Specifically, we drop the observations of banks located in the countries that adopted Pillar 3 of Basel II in or before 2010.

10
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

implementation of Pillar 3 of Basel II.19

Lastly, in Panel E of Table 4, we show that our main results are not sensitive to the use of alternative proxies for risk taking. We
adopt the three primary risk taking measures in the baseline model because they better reflect the overall risk level. For banks in
particular, an important risk concern is regarding bank asset quality and resilience. We therefore adopt three asset-related bank risk
taking measures, which are also widely used in the banking literature (e.g., Shrieves and Dahl, 1992; Delis,Hasan,and Tsionas, 2014;
Khan,Scheule,and Wu, 2017), to check the robustness of the effects of risk disclosure via IFRS 7. The first alternative measure is risk-
weighted assets because it is closely related to capital adequacy, which in turn affects banks’ asset quality and riskiness. We consider
the ratio of risk-weighted assets to total assets estimated each year and denote it by Annual Risk-Weighted Assets. The second alternative
risk taking measure, denoted by Loan Charge-Off, is loan charge-off divided by gross loans, also estimated each year. The rationale is
that a higher charge-off indicates that banks take on more risky assets and have to bear the result of more loan impairments. The third
alternative measure, Loan Loss Reserve, is a bank’s accumulated provisions for loan losses divided by total assets, which relate to the
credit risk exposure of the bank and thus its risk taking behaviors (especially those through bank lending). Using Annual Risk-Weighted
Assets, Loan Charge-Off, and Loan Loss Reserve as the dependent variable in Eq. (1), we re-estimate the regressions in columns 1 to 3,
respectively, and find results largely consistent with those documented for the baseline model. Therefore, our conclusion that risk
disclosure under IFRS 7 is associated with a reduction of bank risk taking holds for various measures of bank risk.20

5. Cross-sectional variation in the association of IFRS 7 adoption with bank risk taking

5.1. The role of accounting standard enforcement

Since we have shown a reduction of bank risk taking associated with mandatary risk disclosure via IFRS 7, our empirical results are
consistent with the market discipline hypothesis. For market discipline to work, timely reporting of reliable risk information is a
necessary condition, and high-quality disclosure in turn heavily depends on the enforcement of IFRS 7. Daske, Hail, Leuz, and Verdi
(2008) and Li (2010) have underscored the central importance of enforcement for the quality of financial reporting, in particular,
stronger enforcement strengthens the impacts of accounting regulations such as the implementation of IFRS. This issue is of particular
pertinence given that IFRS 7 is adopted in substantially different legal, regulatory, and economic environments that could significantly
alter the pressure on bank managers to fairly disclose risk information and also the external agents’ responding behaviors. This po­
tential heterogeneity tends to be more prominent for an international accounting rule like IFRS 7 than for a domestic financial
reporting regulation within either a developed or a developing country. It is therefore important to examine whether and how reg­
ulatory enforcement influences the economic consequences of IFRS 7.
To this end, we check the cross-sectional variation in the association between IFRS 7 adoption and bank risk taking by partitioning
the full sample into two subsamples according to the level of enforcement. We use the indicator for the strictness of accounting
standard enforcement, as constructed by Brown, Preiato, and Tarca (2014), to assign sample countries (and banks therein) to high and
low enforcement subsamples. Brown et al.’s (2014) enforcement indicator collectively reflects a country’s efforts to better enforce its
accounting standards, including compliance promoted through external audit and the activities of independent enforcement bodies.
We re-estimate the baseline model in Eq. (1) using banks in each enforcement subsample and report the results in Panel A of
Table 5. We find that, when Abnormal σ (NIM) or Abnormal Z-score is the dependent variable, the risk reducing effect of IFRS 7 is
substantially more pronounced (i.e., the coefficients on IFRS7 × Post are more negative) in the high enforcement subsample of banks
than in banks from low enforcement countries. For example, IFRS7 × Post has a coefficient of − 0.042 for banks with better enforced
accounting standards when Abnormal σ (NIM) is the dependent variable (column 5), whereas the corresponding coefficient for banks
with less strongly enforced accounting standards is − 0.007 (column 2). The difference between these two values is statistically sig­
nificant, as shown in the bottom row of the panel. The comparison between columns 6 (high enforcement) and 3 (low enforcement) for
the dependent variable Abnormal Z-score shares a similar pattern. In particular, the DiD effect is positive (IFRS7 × Post coefficient =
0.025) in column 3 and is negative (IFRS7 × Post coefficient = -0.173) in column 6. For Abnormal σ(ROA), there is no significant
difference in the risk taking effect of IFRS 7 between the two enforcement subsamples even though the coefficient on IFRS7 × Post is
− 0.003 in the high enforcement case (column 3) and − 0.001 in the low enforcement case (column 1). In general, the results in Panel A
show that stricter enforcement of accounting rules delivers greater influence of risk disclosing under IFRS 7 in mitigating bank risk
taking.

5.2. The role of prevailing alternative disciplining mechanism

We further explore a scenario in which bank risk taking is likely to be a more severe problem and therefore the discipline on banks’
risk taking by IFRS 7 adoption tends to be more important. Market discipline has multi facets, arising from various sources from the
behaviors of different agents. If there are, besides IFRS 7, alternative disciplining mechanisms that constrain excessive risk taking, the

19
Despite this argument, we admit that our robustness test may not be able to completely exclude the possibility that the combination of IFRS 7
and Pillar 3 of Basel II drives the increase in bank risk disclosure, which, in turn, imposes greater discipline on bank risk taking.
20
We also conduct further (untabulated) robustness tests, including excluding the U.S. from the sample, using country-weighted regressions, and
using IFRS countries only. In all these tests, we find that our basic conclusion that banks’ risk disclosure via IFRS 7 tends to reduce their risk taking is
unaltered in general.

11
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Table 5
Cross-sectional variation in the association between IFRS 7 adoption and bank risk taking. In Panels A and B, the baseline model of Eq. (1) is
estimated in subsamples partitioned by accounting standard enforcement and the strength of alternative disciplining mechanism related to deposit
insurance, respectively. The dependent variables are Abnormal σ(ROA) (in column 1), Abnormal σ (NIM) (in column 2), and Abnormal Z-score (in
column 3). The key independent variable is the interaction term between the treatment bank indicator IFRS7 and the indicator Post for the period after
the implementation of IFRS 7. Bank and country characteristic variables are controlled as in Eq. (1). Details about the variable definitions are
provided in the Appendix. The regression coefficients on independent variables are reported, followed by the p-values (in the parentheses) based on
standard errors clustered by country. The bottom row of each panel reports the p-values for the difference test of the coefficients on IFRS7 × Post
between the two subsamples. *, **, and *** indicate statistical significance at the 10 %, 5 %, and 1 % levels, respectively.
Panel A: Accounting standard enforcement

Without strong enforcement (Strong Enforcement = 0) With strong enforcement (Strong Enforcement = 1)

(1) (2) (3) (4) (5) (6)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7 £ Post − 0.001 − 0.007 0.025 ¡0.003** ¡0.042*** − 0.173

(0.529) (0.617) (0.857) (0.048) (0.001) (0.155)
Ln(Bank Size) − 0.009*** − 0.024* − 0.352** − 0.002 − 0.019*** − 0.283***
(0.000) (0.085) (0.020) (0.102) (0.001) (0.000)
NIM Growth − 0.002** − 0.000 − 0.129** − 0.004*** 0.006*** − 0.279***
(0.032) (0.999) (0.021) (0.000) (0.000) (0.000)
Too Big 0.001 − 0.061** 0.097 0.002 − 0.041 − 0.044
(0.785) (0.035) (0.803) (0.684) (0.216) (0.933)
Assets Growth − 0.005 − 0.002 − 0.706 − 0.013*** − 0.045 − 1.530***
(0.611) (0.926) (0.344) (0.000) (0.161) (0.000)
LLP 0.002*** − 0.007** 0.163*** 0.004** − 0.012*** 0.377**
(0.000) (0.035) (0.000) (0.020) (0.000) (0.020)
Ln(GDP) − 0.005 0.211 − 5.445*** − 0.020 − 0.169 0.052
(0.786) (0.199) (0.001) (0.254) (0.437) (0.958)
Ln(GDP per capita) − 0.013* 0.229** − 0.643 0.051*** 0.070 − 0.118
(0.080) (0.021) (0.454) (0.000) (0.725) (0.882)
Inflation 0.000 − 0.001 0.089*** − 0.000* − 0.007*** − 0.029
(0.248) (0.763) (0.001) (0.063) (0.006) (0.188)
Bank and Time FE Included Included Included Included Included Included
Observations 1,597 1,597 1,597 24,487 24,487 24,487
Adj. R2 0.622 0.595 0.317 0.463 0.196 0.278
p-value of difference in the coefficients on IFRS7 × Post 0.17 0.07* 0.04**

Panel B: Alternative disciplining mechanism related to deposit insurance

Strong alternative disciplining mechanism (High Level of Deposit Weak alternative disciplining mechanism (High Level of Deposit
Insurance = 0) Insurance = 1)

(1) (2) (3) (4) (5) (6)

Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score Abnormal σ(ROA) Abnormal σ(NIM) Abnormal Z-score

IFRS7£Post 0.001 ¡0.016** − 0.014 ¡0.005*** ¡0.062*** − 0.057

(0.566) (0.032) (0.927) (0.002) (0.000) (0.679)
Ln(Bank Size) − 0.008*** − 0.017 − 0.497*** − 0.003* − 0.015*** − 0.330***
(0.000) (0.393) (0.004) (0.059) (0.000) (0.000)
NIM Growth − 0.002* 0.001 − 0.052 − 0.004*** 0.006*** − 0.281***
(0.083) (0.856) (0.569) (0.000) (0.001) (0.000)
Too Big 0.003 − 0.045* − 0.063 0.003 − 0.063** 0.300
(0.591) (0.094) (0.891) (0.602) (0.032) (0.571)
Assets Growth − 0.014** − 0.060 − 0.713 − 0.013*** − 0.013** − 1.618***
(0.049) (0.145) (0.230) (0.000) (0.043) (0.000)
LLP 0.002*** − 0.011*** 0.226*** 0.004** − 0.013*** 0.361**
(0.000) (0.000) (0.009) (0.029) (0.000) (0.036)
Ln(GDP) − 0.028 0.194* − 5.520* − 0.009 − 0.383*** − 4.522***
(0.336) (0.077) (0.059) (0.476) (0.000) (0.001)
Ln(GDP per capita) 0.035 0.003 2.536 0.052*** 0.300*** 1.410
(0.326) (0.977) (0.442) (0.000) (0.000) (0.131)
Inflation − 0.000 0.002** − 0.006 − 0.000 − 0.004 − 0.075*
(0.435) (0.032) (0.774) (0.720) (0.167) (0.062)
Bank and Time FE Included Included Included Included Included Included
Observations 3,689 3,689 3,689 21,929 21,929 21,929
2
Adj. R 0.683 0.558 0.251 0.431 0.200 0.276
p-value of difference in the coefficients on IFRS7×Post 0.00*** 0.00*** 0.41

additional risk reducing effect of IFRS 7 could become marginal and less important; if, on the other hand, alternative external
disciplining is insufficient, the implementation of IFRS 7 should induce a more pronounced risk reducing effect.
We examine an alternative disciplining mechanism influenced by a jurisdiction’s deposit insurance policy. The literature (e.g.,
Calomiris and Jaremski, 2019) has shown that a higher level of deposit insurance relates to a weaker market discipline environment,

12
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

which exerts less constraint on imprudent banking behaviors and drives banks to take more risk. In such an environment, the disci­
plining role of IFRS 7 appears more important and tends to have a stronger constraining effect on bank risk taking.
We test this conjecture by conducting a subsample analysis with the full sample partitioned by the median value of a measure for
deposit insurance as specified by Demirguc-Kunt and Detragiache (2002) and Demirguc-Kunt, Kane, and Laeven (2015), with details
provided in the Appendix. Panel B of Table 5 shows that, when Abnormal σ (ROA) and Abnormal σ (NIM) are used as the dependent
variable, the coefficients on IFRS7 × Post are significantly negative in the subsample with a higher level of deposit insurance. The
magnitudes of these coefficients are also significantly larger than the corresponding ones in the subsample with a lower level of deposit
insurance. These results generally support our conjecture that the risk reducing effect of risk disclosure via IFRS 7 is stronger in an
environment lacking prevailing alternative disciplining schemes from uninsured depositors, which highlights the importance of the
disciplining role of IFRS 7 adoption on bank risk taking.

6. The change of bank lending behaviors: A potential channel connecting IFRS 7 implementation and bank risk taking

Our analyses so far have shown that mandatory IFRS 7 adoption is associated with reduced bank risk taking measured in different
traditional ways. In this section, we focus on banks’ prudent actions in their lending business that help reduce their credit risk exposure
after IFRS 7 adoption. These actions play an important role in defining the nature of bank risk taking. To avoid exposure to excessive
credit risk, banks have the responsibility to include self-protective terms in loan contracts to provide enhanced safety. Following this
rationale, we specifically investigate how IFRS 7 adoption affects bank lending terms by estimating Eq. (2) below:
BankLending = σ 0 + σ1 IFRS7 × Post + σ2 Ln
(LoanSize) + σ 3 Ln(LoanMaturity) + σ 4
(PerformancePricing) + σ5 Ln(FirmSize)
(2)
+σ 6 PPE + σ 7 (FirmROA) + σ 8 (OperationRisk)
+σ 9 Ln(GDP) + σ 10 Ln(GDPpercapita) + σ12 Inflation +
FirmFixedEffects + TimeFixedEffects + ε
We consider the dependent variable, Bank Lending, from three perspectives. First, we examine the collateral requirement in bank
loans (denoted by Secured, an indicator variable that equals one if the bank requires collateral for the loan and zero otherwise). Second,
we examine the number of financial covenants in the loans (denoted by Covenants). These two variables reflect the contractual ar­
rangements embedded in the lending process in the sense that more collateral and more covenants provide better protection of bank

Table 6
IFRS 7 adoption and bank lending. This table reports estimation results of Eq. (2) where the dependent variables are loan collateral
requirement (Secured), number of financial covenants in loan agreement (Covenants), and incidence of covenant violation (Covenant
Violation) in columns 1 to 3, respectively. The key independent variable is the interaction term between the treatment bank indicator IFRS7
and the indicator Post for the period after the implementation of IFRS 7. Other variables in Eq. (2) are introduced in the main text. Details
about the variable definitions are provided in the Appendix. The regression coefficients on independent variables are reported, followed by
the p-values (in the parentheses) based on standard errors clustered by country. *, **, and *** indicate statistical significance at the 10%, 5%,
and 1% levels, respectively.
(1) (2) (3)
Secured Covenants Covenant Violation

IFRS7 £ Post 0.033** 0.062** ¡0.067***

(0.044) (0.010) (0.000)
Ln(Loan Size) − 0.025** 0.010 0.004
(0.014) (0.356) (0.111)
Ln(Loan Maturity) 0.077*** 0.007 0.082***
(0.000) (0.625) (0.000)
Performance Pricing 0.077*** 0.831*** 0.060***
(0.001) (0.000) (0.000)
Ln(Firm Size) − 0.011 0.009 − 0.048**
(0.531) (0.863) (0.030)
PPE 0.018 − 0.053 − 0.0580
(0.728) (0.479) (0.383)
Firm ROA − 0.267*** 0.028 0.480***
(0.000) (0.758) (0.000)
Operation Risk 0.622** − 0.352 − 0.964***
(0.019) (0.155) (0.001)
Ln(GDP) 0.082 − 2.005 1.268
(0.915) (0.228) (0.166)
Ln(GDP per capita) − 0.230 2.367 1.345
(0.814) (0.226) (0.144)
Inflation − 1.064*** 0.216 − 3.052
(0.008) (0.831) (0.130)
Firm and Time FE Included Included Included
Observations 53,420 53,420 11,346
Adj. R2 0.679 0.749 0.739

13
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

interests and also better monitoring of borrowers. After a loan is issued, these loan agreement terms serve as a tool for banks to control
their risk exposure and manage the risk. Third, we examine the post-loan outcome relating to banks’ risk, as reflected in covenant
violation that is detrimental to lenders in terms of escalating their credit risk exposure. A lower incidence of violating financial
covenants suggests a less risky output of bank lending. We construct an indicator variable Covenant Violation that equals one if there is
covenant violation occurrence and zero otherwise.
We follow the large literature on bank loan contracting and control for the natural logarithms of loan size (Ln(Loan Size)) and loan
maturity (Ln(Loan Maturity)), the existence of performance pricing terms in the loan agreement (Performance Pricing), as well as
borrowing firms’ size (Ln(Firm Size)), intangible assets (PPE), ROA (Firm ROA), and variability of cash flows from operations (Operation
Risk). We retrieve these loan-level data from Deal Scan and firm-level data from Compustat Global. We provide detailed definitions of
these control variables in the Appendix. Country-level controls are the same as in the baseline model. With firm and time fixed effects,
the model in Eq. (2) maintains a DiD design, in which an IFRS 7-induced improvement in loan safety level translates into positive
coefficients on IFRS7 × Post when Secured or Covenants is used as the dependent variable, and a negative IFRS7 × Post coefficient when
Covenant Violation is used as the dependent variable.
The results reported in Table 6 are largely consistent with the pattern in our conjecture. The coefficient on IFRS7 × Post is 0.033 in
column 1 (Bank Lending is proxied by Secured) and 0.062 in column 2 (Bank Lending is proxied by Covenants), both being significant at
the 5 % level. These results suggest that loans with higher collateral requirements and more covenants are issued by banks after they
adopt IFRS 7, relative to banks without IFRS 7 adoption. These loan contracting terms provide protection for bank risk exposure and
our finding is thus in line with the risk mitigating effect of IFRS 7. Column 3 shows that the incidence of covenant violation is also
significantly reduced after banks adopt IFRS 7, because the coefficient on IFRS7 × Post is significantly negative (coefficient = -0.067; p-
value = 0.000) in the Covenant Violation regression. This result further reconfirms the safety improving efforts by banks in their lending
business. Overall, the findings in Table 6 support the notion that banks take actions to reduce risk taking in a way consistent with
enhanced market discipline induced by IFRS 7.

7. Conclusion

In 2007, a significant number of banks adopted IFRS 7, which mandates disclosure via accounting statements of the extent and
severity of a bank’s exposure to risks relating to financial instruments and the approach to managing these risks. By requiring
disclosure of information regarding risky cash flow prospects, policy makers (e.g., IASB, which issued IFRS 7) aim to assist users of
financial statements to better manage their capital allocation because they can understand risk information known to the internal
managers in a more direct way rather than having to indirectly infer and update the uncertainty in cash flows through other noisy
signals. If so, investors, depositors, and other external stakeholders may be deterred from taking positions that expose them to
excessive risks, and thus impose a market discipline that may constrain imprudent risk taking by the reporting banks. In this study, we
examine the impact of risk reporting under IFRS 7 on risk taking behavior using a large international sample of banks.
We conduct our analysis in a DiD setting in which we identify the improvement in risk disclosure by comparing treatment IFRS 7-
adopting banks with control non-adopting banks. We find that treatment banks show significantly less risk taking after the imple­
mentation of IFRS 7, relative to control banks, especially when the accounting rules are more strictly enforced and an alternative risk
constraining mechanism via depositor monitoring is weak. These results generally support the market discipline argument. We further
document that, after reporting risk information under IFRS 7, banks tend to apply stricter collateral and covenant terms in loan
issuance. These behaviors contribute to more prudent risk taking and managing of risk. Overall, our findings point to beneficial effects
of risk disclosure via IFRS 7 in terms of constrained risk taking that helps improve financial stability.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

The authors do not have permission to share data.

Acknowledgement

The authors thank Marco Trombetta (editor) and two anonymous reviewers for valuable comments. Chong Wang and Feng (Harry)
Wu acknowledge research supports from Hong Kong Polytechnic University and Lingnan University, respectively. All errors are our
own.

14
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Appendix. Variable definitions

IFRS 7-Related DiD Variables

IFRS7 Indicator variable that equals one if a bank adopted IFRS 7 in 2007, and zero otherwise.
Post Indicator variable that equals one for the post-IFRS 7 sample period (2007–2010) and zero for the pre-IFRS 7 sample period
(2005–2006).
Bank Risk Taking Variables

σ(ROA) Standard deviation of ROA (earnings before taxes scaled by total assets) estimated for the pre- (post-)IFRS 7 sample period. Source:
BankScope.
σ(NIM) Standard deviation of net interest margin estimated for the pre- (post-)IFRS 7 sample period. Source: BankScope.
Z-score Natural logarithm of ((ROA + CAR)/σ(ROA)) multiplied by (− 1), where CAR is equity scaled by total assets. ROA and CAR are
calculated as the mean for the pre- (post-)IFRS 7 sample period, and σ(ROA) is estimated for the pre- (post-)IFRS 7 sample period as
described above. A higher Z-score constructed this way represents more risk taking. Source: BankScope.
Abnormal σ(ROA) Residual from the regression of σ(ROA) on the bank-characteristic variables in Eq. (1) estimated for the pre- (post-)IFRS 7 sample
period.
Abnormal σ(NIM) Residual from the regression of σ(NIM) on the bank-characteristic variables in Eq. (1) estimated for the pre- (post-)IFRS 7 sample
period.
Abnormal Z-score Residual from the regression of Z-score on the bank-characteristic variables in Eq. (1) estimated for the pre- (post-)IFRS 7 sample
period.
Annual Risk-Weighted Annual risk-weighted assets scaled by total assets. Source: BankScope.
Assets
Loan Charge-Off Loan charge-off scaled by gross loans. Source: BankScope.
Loan Loss Reserve Loan loss reserve scaled by total assets. Source: BankScope.
Control Variables

Ln(Bank Size) Natural logarithm of bank total assets averaged over the pre- (post-)IFRS 7 sample period. Source: BankScope.
NIM Growth Annual growth in net interest margin, averaged over the pre- (post-)IFRS 7 sample period. Source: BankScope.
Too Big Indicator variable that equals one if a bank’s deposit share is in the top decile of all banks in the country in the pre- (post-)IFRS 7
sample period, and zero otherwise. Source: BankScope.
Assets Growth Annual growth rate in total assets averaged over the pre- (post-)IFRS 7 sample period. Source: BankScope.
LLP Loan loss provision divided by total loans averaged over the pre- (post-)IFRS 7 sample period. Source: BankScope.
Ln(GDP) Natural logarithm of GDP (in constant 2010 US$). Source: World Bank.
Ln(GDP per capita) Natural logarithm of GDP per capita (in constant 2010 US$). Source: World Bank.
Inflation Annual inflation rate in percentage calculated using the GDP deflator. Source: World Bank.
Cross-Sectional Sample Partitioning Variables

Strong Enforcement Indicator variable that equals one for countries that strictly enforce accounting standards (the accounting enforcement index of
Brown et al. (2014) is larger than the sample median), and zero otherwise. Source: Brown et al. (2014).
High Level of Deposit Indicator variable that equals one for countries with a moral hazard index larger than the sample median, and zero otherwise. The
Insurance moral hazard index measures the level of deposit insurance, which is the first principal component of deposit insurance system
features, including the dummies for no coinsurance, foreign currency deposits covered, interbank deposits covered, type of funding,
source of funding, management, and the level of explicit coverage. A larger moral hazard index indicates a higher level of deposit
insurance. Source: Demirguc-Kunt and Detragiache (2002) and Demirguc-Kunt et al. (2015).
Bank Lending Variables and Associated Controls

Secured Indicator variable that equals one if the loan includes a collateral requirement, and zero otherwise. Source: Deal Scan.
Covenants Number of financial covenants. Source: Deal Scan.
Covenant Violation Indicator variable that equals one if there are one or more covenant violation incidents, and zero otherwise. Source: Deal Scan.
Ln(Loan Size) Natural logarithm of loan amount of the facility in million US$. Source: Deal Scan.
Ln(Loan Maturity) Natural logarithm of the number of months to maturity of a loan. Source: Deal Scan.
Performance Pricing Indicator variable that equals one if the loan includes performance pricing provisions, and zero otherwise. Source: Deal Scan.
Ln(Firm Size) Natural logarithm of total assets in million US$ of the borrowing firm. Source: Compustat Global.
PPE Property, plant and equipment of the borrowing firm divided by total assets. Source: Compustat Global.
Firm ROA ROA (earnings before taxes divided by total assets) of the borrowing firm. Source: Compustat Global.
Operation Risk The standard deviation of yearly cash flows from operations divided by total assets estimated over the past five fiscal years. Source:
Compustat Global.

References

Acharya, V.V., Ryan, S.G., 2016. Banks’ financial reporting and financial system stability. J. Account. Res. 54, 277–340.
Allen, F., Carletti, E., 2008. Mark-to-market accounting and liquidity pricing. J. Account. Econ. 45, 358–378.
Armstrong, C., Balakrishnan, K., Cohen, D., 2012. Corporate governance and the information environment: Evidence from state antitakeover laws. J. Account. Econ.
53, 185–204.
Banerjee, S., Davis, J., Gondhi, N., 2018. When transparency improves, must prices reflect fundamentals better? Rev. Financ. Stud. 31, 2377–2414.
Banerjee, S., Maier, M., 2016. Public information precision and coordination failure: An experiment. J. Account. Res. 54, 941–985.
BCBS (Basel Committee on Banking Supervision). 2004. International convergence of capital requirement and capital standards: A revised framework. Basel: Bank of
International Settlements.
Beatty, A., M. Iselin, and S. Liao. 2023. Financial accounting and disclosure in banking. Chapter in forthcoming 4th edition of The Oxford Handbook of Banking.
Beatty, A., Liao, S., 2014. Financial accounting in the banking industry: A review of the empirical literature. J. Account. Econ. 58, 339–383.
Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. J. Polit. Econ. 111, 1043–1075.

15
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we trust difference-in-differences estimates? Q. J. Econ. 119, 249–275.
Bertrand, M., Mullainathan, S., 1999. Is there discretion in wage setting? A test using takeover legislation. RAND J. Econ. 30, 535–554.
Billett, M., Garfinkel, J., O’Neal, E., 1998. The cost of market versus regulatory discipline in banking. J. Financ. Econ. 48, 333–358.
Bischof, J., 2009. The effects of IFRS 7 adoption on bank disclosure in Europe. Account. Eur. 6, 167–194.
Bischof, J., Daske, H., 2013. Mandatory disclosure, voluntary disclosure, and stock market liquidity: Evidence from the EU bank stress tests. J. Account. Res. 51,
997–1029.
Bischof, J., Laux, C., Leuz, C., 2021. Accounting for financial stability: Bank disclosure and loss recognition in the financial crisis. J. Financ. Econ. 141, 1188–1217.
Bischof, J., Daske, H., Elfers, F., Hail, L., 2022. A tale of two supervisors: Compliance with risk disclosure regulation in the banking sector. Contemp. Account. Res. 39,
498–536.
Blankespoor, E., deHaan, E., Marinovic, I., 2020. Disclosure processing costs, investors’ information choice, and equity market outcomes: A review. J. Account. Econ.
70, 101344.
Bleck, A., and P. Gao. 2012. Where does the information in market-to-market come from? Working paper, University of Chicago.
Boro Jr., A.J., 1986. Banking disclosure regimes for regulating speculative behavior. Calif. Law Rev. 74, 431–490.
Brown, P., Preiato, J., Tarca, A., 2014. Measuring country differences in enforcement of accounting standards: An audit and enforcement proxy. J. Bus. Financ. Acc.
41, 1–52.
Bushman, R.M., 2014. Thoughts on financial accounting and the banking industry”. J. Account. Econ. 58, 384–395.
Bushman, R.M., Williams, C.D., 2012. Accounting discretion, loan loss provisioning, and discipline of banks’ risk-taking. J. Account. Econ. 54, 1–18.
Calomiris, C.W., Jaremski, M., 2019. Stealing deposits: Deposit insurance, risk-taking, and the removal of market discipline in early 20th-century banks. J. Financ. 74,
711–754.
Calomiris, C.W., Kahn, C.M., 1991. The role of demandable debt in structuring optimal banking arrangements. Am. Econ. Rev. 81, 497–513.
Campbell, J., Chen, H., Dhaliwal, D., Lu, H., Steele, L., 2014. The information content of mandatory risk factor disclosures in corporate filings. Rev. Acc. Stud. 19,
396–455.
Campbell, J., Cecchini, M., Cianci, A., Ehinger, A., Werner, E., 2019. Tax-related mandatory risk factor disclosures, future profitability, and stock returns. Rev. Acc.
Stud. 24, 264–308.
Cheng, I., Hong, H., Scheinkman, J., 2015. Yesterday’s heroes: Compensation and risk at financial firms. J. Financ. 70, 839–879.
Chiu, T.-T., Guan, Y., Kim, J.-B., 2018. The effect of risk factor disclosures on the pricing of credit default swaps. Contemp. Account. Res. 35, 2191–2224.
Christensen, H., Nikolaev, V., Wittenberg-Moerman, R., 2016. Accounting information in financial contracting: The incomplete contract theory perspective.
J. Account. Res. 54, 397–435.
Christensen, H., Floyd, E., Liu, L., Maffett, M., 2017. The real effects of mandated information on social responsibility in financial reports: Evidence from mine-safety
records. J. Account. Econ. 64, 284–304.
Conti, C., Mauri, A., 2008. Corporate financial risk management: Governance and disclosure post IFRS 7. Icfai Univ. J. Financ. Risk Manage. 5, 20–27.
Cordella, T., Yeyati, E.L., 1998. Public disclosure and bank failures. IMF Staff. Pap. 45, 110–131.
Core, J., Hail, L., Verdi, R., 2015. Mandatory disclosure quality, inside ownership, and cost of capital. Eur. Account. Rev. 24, 1–29.
Corona, C., Nan, L., Zhang, G., 2019. The coordination role of stress tests in bank risk-taking. J. Account. Res. 57, 1161–1200.
Dang, T.V., Gorton, G., Holmstrom, B., Ordonez, G., 2017. Banks as secret keepers. Am. Econ. Rev. 107, 1005–1029.
Daske, H., Hail, L., Leuz, C., Verdi, R., 2008. Mandatory IFRS reporting around the world: Early evidence on the economic consequences. J. Account. Res. 46,
1085–1142.
Delis, M.D., Hasan, L., Tsionas, E.G., 2014. The risk of financial intermediaries. J. Bank. Financ. 44, 1–12.
Demirguc-Kunt, A., Detragiache, E., 2002. Does deposit insurance increase banking system stability? An empirical investigation. J. Monet. Econ. 49, 1373–1406.
Demirguc-Kunt, A., Kane, E., Laeven, L., 2015. Deposit insurance around the world: A comprehensive analysis and database. J. Financ. Stab. 20, 155–183.
Doty, J.R., Mahaffey, D.C., Goldstein, M.J., 1991. Full disclosure, market discipline, and risk taking: Rethinking confidentiality in bank regulation. Washington Univ.
Law Quarterly 69, 1117–1135.
Erkens, M., 2016. Disclosure behavior of European firms around the adoption of IFRS. Springer Gabler, Wiesbaden.
Ertan, A., Loumioti, M., Wittenberg-Moerman, R., 2017. Enhancing loan quality through transparency: Evidence from the European Central Bank loan level reporting
initiative. J. Account. Res. 55, 877–918.
Financial Accounting Standards Board (FASB). 2012. Proposed Accounting Standards Update: Financial Instruments (Topic 825). Available at https://www.fasb.org/jsp/
FASB/Document_C/DocumentPage?cid=1176160611960&acceptedDisclaimer=true.
Frank, K., 2000. Impact of a confounding variable on a regression coefficient. Sociol. Methods Res. 29, 147–194.
Fung, A., Graham, M., Weil, D., 2007. Full Disclosure: The Perils and Promise of Transparency. Cambridge University Press, New York.
Gao, P., Jiang, X., 2014. Reporting discretion, market discipline, and panic runs. Working paper. University of Chicago and Duke University.
Gao, P., Liang, P., 2013. Information feedback, adverse selection, and optimal disclosure policy. J. Account. Res. 51, 1133–1158.
Goldstein, I., and H. Sapra. 2014. Should banks’ stress test results be disclosed? An analysis of the costs and benefits. Foundations and Trends® in Finance 8, 1-54.
Goldstein, I., Yang, L., 2019. Good disclosure, bad disclosure. J. Financ. Econ. 131, 118–138.
Gorton, G., 2015. Stress for success: A review of Timothy Geithner’s financial crisis memoir. J. Econ. Lit. 53, 975–995.
Granja, J., 2018. Disclosure regulation in the commercial banking industry: Lessons from the National Banking era. J. Account. Res. 56, 173–216.
Heinle, M., Smith, K., 2017. A theory of risk disclosure. Rev. Acc. Stud. 22, 1459–1491.
Heinle, M., Smith, K., Verrecchia, R., 2018. Risk-factor disclosure and asset prices. Account. Rev. 93, 191–208.
Holmstrom, B. 2015. Understanding the role of debt in the financial system. Working paper, Bank for International Settlements.
Hope, O., Hu, D., Lu, H., 2016. The benefits of specific risk-factor disclosures. Rev. Acc. Stud. 21, 1005–1045.
Houston, J., Lin, C., Lin, P., Ma, Y., 2010. Creditor rights, information sharing, and bank risk taking. J. Financ. Econ. 96, 485–512.
Indjejikian, R., 1991. The impact of costly information interpretation on firm disclosure decisions. J. Account. Res. 29, 277–301.
Jayaraman, S., Shivakumar, L., 2013. Agency-based demand for conservatism: Evidence from state adoption of antitakeover laws. Rev. Acc. Stud. 18, 95–134.
Kanagaretnam, K., Lobo, G., Wang, C., Whalen, D., 2019. Cross-country evidence on the relationship between societal trust and risk-taking bank banks. J. Financ.
Quant. Anal. 54, 275–301.
Kanodia, C., 2006. Accounting disclosure and real effects. Found. Trends Account. 1, 167–258.
Kanodia, C., Sapra, H., 2016. A real effects perspective to accounting measurement and disclosure: Implications and insights for future research. J. Account. Res. 54,
623–676.
Khan, M.S., Scheule, H., Wu, E., 2017. Funding liquidity and bank risk taking. J. Bank. Financ. 82, 203–216.
Kravet, T., Muslu, V., 2013. Textual risk disclosures and investors’ risk perceptions. Rev. Acc. Stud. 18, 1088–1122.
Laeven, L., Levine, R., 2009. Bank governance, regulation and risk taking. J. Financ. Econ. 93, 259–275.
Lambert, R., Leuz, C., Verrecchia, R., 2007. Accounting information, disclosure, and the cost of capital. J. Account. Res. 45, 385–420.
Landier, A., and D. Thesmar. 2011. Regulating systemic risk through transparency: Trade-offs in making data public. In: M. Brunnermeier, and A. Krishnamurthy
(Eds.). Risk Topography: Systemic Risk and Macro Modeling. Chicago: University of Chicago Press.
Leuz, C., Wysocki, P., 2016. The economics of disclosure and financial reporting regulation: Evidence and suggestions for future research. J. Account. Res. 54,
525–622.
Li, S., 2010. Does mandatory adoption of International Financial Reporting Standards in the European Union reduce the cost of equity capital? Account. Rev. 85,
607–636.
Low, A., 2009. Managerial risk-taking behavior and equity-based compensation. J. Financ. Econ. 92, 470–490.
Madsen, J., McMullin, J., 2020. Economic consequences of risk disclosures: Evidence from crowdfunding. Account. Rev. 95, 331–363.

16
G.J. Lobo et al. Journal of Accounting and Public Policy 46 (2024) 107225

Martinez Peria, M.S., Schmukler, S.L., 2001. Do depositors punish banks for bad behavior? Market discipline, deposit insurance, and banking crises. J. Financ. 56,
1029–1051.
Mathewson, A.D., 1986. From confidential supervision to market discipline: The role of disclosure in the regulation of commercial banks. J. Corp. Law 11, 139–178.
Morgan, D., 2002. Rating banks: Risk and uncertainty in an opaque industry. Am. Econ. Rev. 92, 874–888.
Morris, S., Shin, H., 2002. Social value of public information. Am. Econ. Rev. 92, 1521–1534.
Nadia, C., Rosa, V., 2014. The impact of IFRS 9 and IFRS 7 on liquidity in banks: Theoretical aspects. Procedia – Soc. Behav. Sci. 164, 91–97.
Nier, E., Baumann, U., 2006. Market discipline, disclosure and moral hazard in banking. J. Financ. Intermed. 15, 332–361.
Plantin, G., Sapra, H., Shin, H.S., 2008. Marking-to-market: Panacea or Pandora’s box? J. Account. Res. 46, 435–460.
Plantin, G., Tirole, J., 2018. Marking to market versus taking to market. Am. Econ. Rev. 108, 2246–2276.
PwC. 2010. A Practical Guide to IFRS 7. Available at https://www.pwc.com/gx/en/ifrs-reporting/pdf/a_practical_guide_to_ifrs_7.pdf.
Ryan, S., 1997. A survey of research relating accounting numbers to systematic equity risk, with implications for risk disclosure policy and future research. Account.
Horiz. 11, 82–95.
Ryan, S., 2017. Do the effects of accounting requirements on banks’ regulatory capital adequacy undermine financial stability? Annu. Rev. Financ. Econ. 9, 1–20.
Sapra, H., 2002. Do mandatory hedge disclosures discourage or encourage excessive speculation? J. Account. Res. 40, 933–964.
Shrieves, R., Dahl, D., 1992. The relationship between risk and capital in commercial banks. J. Bank. Financ. 16, 439–457.
Stephanou, C. 2010. Rethinking market discipline in banking: Lessons from the financial crisis. Working paper, World Bank.
Tweedie, D. 2005. Comment when introducing IFRS 7. International Accounting Standards Board Press Release (Aug. 18, 2005). Available at https://www.yumpu.
com/en/document/read/22427959/iasb-press-release-ias-plus.
World Bank. 2012. Note of Centre for Financial Reporting Reform (retrieved on Aug. 28, 2016).

17