The Influence of IFRS Adoption on Banks’ Cost of Equity: Evidence from European Banks

This study examines how mandatory adoption of International Financial Reporting Standards (IFRS) in European countries affects banks’ cost of equity. Supporters of IFRS argue that its adoption improves the quality of accounting information, which in turn decreases the cost of equity. However, banking regulators could intervene in the implementation of new accounting standards to protect the stability of the banking system, which would deteriorate banks’ information environment and thereby increase the cost of equity. Using a regression analysis of European listed bank data, I find that banks’ cost of equity increases after the adoption of IFRS in countries with strong bank supervisory offices. I also find that strong legal enforcement and additional disclosure requirements jointly reduce banks’ cost of equity, but pre-IFRS inconsistencies between local accounting standards and regulatory standards jointly increase banks’ cost of equity. This study contributes to the literature on market discipline in banking and has policy implications: The findings suggest that, when implementing new accounting standards, potential conflicts between financial reporting and banking regulations should be considered.


Introduction
The 2007 US subprime mortgage crisis shows the importance of the banking system for sustainable economic growth. The adverse effects of this crisis not only impacted the banks and debtors as parties to the mortgage loan contracts, but it also spread across the entire financial system and the real economy. As a result, despite the US government's efforts to stabilize the financial system, real domestic production per capita in the United States decreased by more than 5% from the fourth quarter of 2007 to the second quarter of 2009. This shows how the soundness of the banking system is critical to sustainable economic growth.
While several government regulations have been adopted to maintain the stability of the banking industry, innovations in financial instruments have been developing quickly, and regulatory bodies are playing catch up with the financial market. Consequently, the role of market discipline is crucial because market mechanisms can adapt more flexibly and promptly to change.
Unlike government regulators, market participants are not authorized to access banks' private information. Therefore, the public information environment is critical for the market discipline of banks. Financial statements are a reliable and comprehensive source of public information. Hence, this study investigates the influence of changes in accounting standards on European banks' financial statements post-International Financial Reporting Standards (IFRS) adoption.
Researchers argue that IFRS adoption improves accounting quality because it requires more disclosure than most local European accounting standards pre-IFRS. They suggest that IFRS adoption improves both earnings quality and the information environment. Consequently, post-IFRS, security trading by foreign investors increased, and equity values increased. These studies mainly focus on the Third, this study shows the interaction between the institutional environment of banks and changes in accounting standards. Researchers have pointed out that the institutional environment influences financial reporting [9][10][11]. Supporting this argument, studies on mandatory IFRS adoption suggest that investor protection facilitates IFRS adoption [6,12]. However, few studies have examined the role of bank regulation in adopting new accounting standards.
The remainder of this paper proceeds as follows. Section 2 summarizes prior studies regarding the effect of IFRS adoption on the cost of equity, institutional environments of the banking industry, and the economic consequences of IFRS adoption. Section 3 documents hypothesis development. Section 4 presents the research design, sample selection, and descriptive statistics. Section 5 documents the results of regression analyses. Section 6 concludes the paper.

The Effect of IFRS Adoption
Prior studies argue that IFRS adoption improves several aspects of financial reporting, the information environment, and capital markets. Empirical studies find that earnings quality [13,14] and the information environment [12,15,16] are improved following IFRS adoption. Consequently, security trading [5,17,18] and equity valuation [5,6] improve post-IFRS.
Theory expects that the quality of disclosure is negatively related to the cost of equity [19][20][21], which is backed by empirical evidence [22,23]. Since IFRS adoption improves the transparency of accounting information and the information environment, researchers expect that IFRS adoption decreases the cost of equity. Li [6] finds evidence supporting this using European non-financial firm data.
However, IFRS adoption also has a negative consequence because it increases a manager's choice of accounting policy, which reduces the contractibility of the accounting information. Supporting this argument, Ball et al. [1] report that IFRS adoption reduces accounting-based debt covenants. This study implies that IFRS adoption is not welcomed by bank regulators because it reduces the contractibility of debt contracts, which results in instability in the financial markets. In addition, the findings of Ball et al. [1] also suggest that banks' risks increase post-IFRS because banks' lending contracts that utilize accounting information in debt covenants become inefficient.

The Institutional Environment of Listed Banks
This study focuses on the equity capital in European listed banks, which are exposed to two different types of institutional environments: Bank supervision and disclosure regulation. Listed banks are regulated by banking supervisory offices. Although the detailed structures of bank supervisory systems vary by country [24], the ultimate goal of supervisory offices is the same, namely, to safeguard the stability of the financing system because the stability of financial markets plays a critical role in the economic growth of the country. As a publicly listed firm, listed banks are also bound to the disclosure requirements of investors, whose main concern is not in protecting the stability of the markets, but in protecting investors' private interests.
The difference in policy objectives between bank regulation and corporate disclosure creates conflicts between accounting policy and bank regulation. Skinner [3] investigates the adoption of deferred tax accounting in Japan in 1998, during which Japanese banks' regulatory capital was insufficient. Thus, to maintain banks' solvency, the Japanese government and bank regulators decided to use a deferred tax asset as regulatory capital. Because maintaining the solvency of banks is more important for the country's economy, the quality of accounting information was sacrificed during the adoption of deferred tax asset accounting. Skinner [1] implies that banking regulations limit or distort the adoption of new accounting standards if the standards negatively affect the solvency of banks. IFRS adoption could necessitate regulatory intervention, as in the case investigated by Skinner [3]. Bischof [25] also pointed out that there are incentives to prevent European bank regulators from introducing new accounting standards that affect banks' financial statements, which means that the intervention of bank supervisory offices in the adoption of new accounting standards is not limited to a specific country.
To implement accounting standards, support of institutional environments is necessary [9]. However, as prior studies show [3,25], listed banks face a potential conflict between bank regulation and financial disclosure; therefore, how IFRS adoption influences listed banks is unclear.
Basel II is a set of guidelines that shaped the mandatory adoption of IFRS in European counties. Although replaced by Basel III, Basel II is still useful in understanding the mechanism of the banking regulations. Basel II is based on the following three pillars: (1) Minimum capital requirement that requires safer capital as banks' risky assets increase; (2) review process by a government supervisory office; (3) market discipline that relies on sophisticated investors' monitoring. For the first and second pillars, private information can be required from banks or banks' auditors. Frequently, these two pillars have priority over accounting standards [3,25]. For the third pillar, market discipline penalties include direct penalty by investor activism and indirect penalties through market prices of securities, including stock prices. Due to the high information efficiency of market prices, market discipline can reflect the bank's health information at a rate that bank supervisors cannot follow. This means that market discipline is superior to bank supervisors in reflecting the consequences of financial instruments, which are rapidly becoming increasingly complex as they undergo innovation, on the health of banks. Therefore, the importance of market discipline is in an increasing trend [7].
Banking regulations affect banks' financial reporting. Bank supervisors can require banks to disclose private information found during the review process [26,27]. Furthermore, regulatory capital requirements enhance disclosure by providing timely and extensive information that is not required by accounting standards [28]. Stringent banking regulations could conflict with accounting information in that the regulations safeguard the banking system, whereas the accounting information focuses on capital providers. Moreover, banking regulators could sacrifice the quality of accounting information to stabilize the financial system [3,28] or to avoid rapid changes in the accounting numbers to minimize the negative impact on debt contracts based on accounting information [1,25].
The influence of investor protection is the same for banks and non-financial firms. Strong investor protection provides incentives to managers to provide transparent accounting information [10,11,29]. Consequently, IFRS adoption reduces banks' cost of equity [6,30,31].

IFRS Adoption in the Banking Sector
Compared with the previous local accounting standards, IFRS adoption brought several changes. The two most important changes for this study [2,14] include an increase in fair value measurement and an increase in accounting choices.
As the fair value measure increases, the statement of financial position (balance sheet in US Generally Accepted Accounting Principles terminology) increases in relevance for equity valuation. However, the market volatility included in fair value increases the noise in measuring banks' future cash flows. Even though fair value measures do not directly rely on level 1 inputs, which are market values, level 2 or level 3 inputs of fair value measures do not alleviate the information risk because they discretionary. In sum, extended use of fair value measures increases information risks, which are unfavorable for both investors and bank regulators.
In addition, although the increase in accounting choices enhances the relevance of the statement of financial position to equity valuation, this increase could influence the banking industry negatively. The increase in accounting choices complicates verification of compliance with the debt covenants. This provides opportunities for moral hazard for both parties of the debt contract. The reduction of contractibility of accounting information could have a significant impact on capital markets, which necessitates intervention by banking regulators [3,25].

Hypothesis Development
Regarding IFRS adoption and banks' cost of equity, two risks should be considered. The first risk is banks' business risks, which come from operating characteristics; for example, borrowers' credit risks. The second is information risk [19,20]. Both risks increase banks' cost of equity.
Because the minimum capital regulation is applied stringently, the regulatory capital ratio efficiently reduces banks' risk [28]. This risk reduction decreases banks' cost of equity. If banks' risk is already lowered by banking regulations, IFRS adoption has little impact on the disclosure of information about banks' risks. Therefore, IFRS adoption has little impact on the cost of equity if capital regulation is strong. Based on this conjecture, I suggest the following hypotheses: H1: Banks' cost of equity decreases as the minimum capital regulation strengthens.
H2: Strong capital regulation weakens the impact of IFRS adoption on banks' cost of equity.
If banking regulatory agencies have strong power, they can require private information directly from banks or banks' auditors for regulatory actions [1,28]. Therefore, bank regulation strength reduces banks' cost of equity because strong banking regulators can monitor and discipline banks.
Banking regulations have priority over financial reporting in most countries; therefore, these regulations could interfere with IFRS adoption if new accounting standards have a negative effect on the banking system. IFRS adoption increases choice among accounting rules; therefore, using accounting information for debt covenants allows for moral hazard for any one of the contracting parties in debt contracts [1]. Several banks' contracts use accounting information for debt covenants; hence, changes in accounting standards could affect banks' existing contracts. Therefore, bank supervisors have the incentive to intervene in the adoption of new accounting standards to prevent potential turmoil, which would interfere with the faithful implementation of IFRS [3,25]. The intervention of bank supervisors increases information risk of banks, which would increase bank supervisors' power. Based on this conjecture, I suggest the following hypotheses: H3: Banks' cost of equity decreases as the bank supervisors' power strengthens.
H4: IFRS adoption increases banks' cost of equity in the countries with strong banking supervisors.
Market discipline needs a good information environment including high-quality accounting information. Country-level investor protection improves accounting quality by helping faithful financial reporting [9,29], which leads to a reduction in the cost of equity [11,30,31]. Thus, in countries with strong investor protection, IFRS adoption reduces banks' cost of equity. In relation to the institutional aspects of the banking sector, I therefore suggest the following hypothesis:

H5:
The influence of IFRS adoption on banks' cost of equity is weakened when investor protection is strengthened.
The impact of IFRS adoption varies with the extent of changes that occur in IFRS adoption [32]. In most European countries, IFRS adoption requires more disclosure. Thus, the impact of IFRS adoption increases additional disclosure requirements. Moreover, the impact of IFRS adoption varies with the inconsistencies between IFRS and the local accounting standards implemented before IFRS adoption. Accordingly, I propose the following hypotheses.

H6:
The influence of IFRS adoption increases when IFRS adoption requires additional disclosures.

H7:
The influence of IFRS adoption increases when inconsistencies exist between IFRS and the local accounting standards implemented before IFRS adoption.

Regression Model
I use the implied cost of equity as my proxy for expected returns because it has fewer errors than realized-return-based proxies [30,31,33] from information shocks. I average four estimates calculated using the models of Easton [34], Gode and Mohanram [35], Gebhardt et al. [36], and Claus and Thomas [37] to mitigate error in each measurement [30,31].
Studies on the effect of IFRS adoption frequently use a difference-in-differences model using voluntary adopters as the control group. This model controls for the influence that occurs simultaneously with IFRS adoption. However, except the treatment, the control group of the difference-in-differences model should be identical to the treatment group. Furthermore, only three countries have banks that adopted IFRS voluntarily. Most European banks adopt IFRS mandatorily, which means that IFRS adoption was an exogenous event for most European banks. Hence, I do not use the difference-in-differences design.
To test H1 to H5, I use the following model (1): Variable definitions are in the Appendix A. POST is the variable of interest. I include measures for the strength of capital regulation (CAPITAL), the power of bank supervisors (OFFICE), and the efficiency of legal enforcement (ENFORCE) in the regression model. CAPITAL and OFFICE are measured by The Bank Regulation and Supervision Survey 2003 conducted by the World Bank [28,38]. I centered CAPITAL, OFFICE, and ENFORCE by the sample mean of each variable to mitigate multicollinearity problems from biases of spurious correlations [39].
I control firm-level risks using proxies of size, return volatility, financial leverage, total capital ratio, and book-to-price ratio. Size, return volatility, and leverage are measured by the decile rank of each variable to mitigate measurement errors. I include variables to control for cross-listing on the US stock market because investor protection in the US market is stronger than it is in most European countries, but it is not affected by mandatory IFRS adoption. I also control for the annual inflation rate and the indicator variable for the adoption of IFRS 7, which could affect banks. I include the bias and dispersion of analyst forecasts to mitigate the effect of biases and the nonlinearity of the models for the implied cost of equity [37,40]. Many bank-year observations have only one one-year-ahead earnings forecast; hence, I include an indicator variable for the observations to control for potential bias and replace the dispersion of analyst forecasts with zero. I adjust the influence of the firm-level serial correlation using a firm-clustered standard error in all of the regression results in this study [41].
To test H6 and H7, I revise model (1) by including additional disclosure requirements (ADD) and inconsistencies between IFRS and the local accounting standards (INC). I use the survey of Nobes [42] to measure ADD and INC. Nobes [42] did not focus on banks; thus, items irrelevant to banks, for example, inventory or plant assets, are included. To avoid potential measurement errors from irrelevant items, I exclude items irrelevant to bank operations from ADD and INC. I centered ADD and INC by their sample means to avoid multicollinearity problems [39]. The following are the models for H6 and H7, respectively. Model (2) and model (3) are models for testing the effects of ADD and ICC, respectively. CoC = α + β 1 POST + β 2 ENFORCE + β 3 OFFICE + β 4 CAPITAL + β 5 POST*ENFORCE + β 6 POST*OFFICE + β 7 POST*CAPITAL + β 8 ADD + β 9 POST*ADD + β 10 POST*ENFORCE*ADD + β 11 POST*OFFICE*ADD + β 12 POST*CAPITAL*ADD + CONTROLS + ε

Sample Selection
Mandatory IFRS adoption by the European Union provides the setting for a natural experiment. Therefore, I use data from listed banks of European countries from 1995 to 2009. The observations are required to have the Standard Industry Code between 6020 and 6099. Analyst forecast data and financial data are obtained from I/B/E/S and Compustat Global, respectively. I match the stock prices and analyst forecasts of seven months after the previous fiscal-year end to make sure that precious accounting information is fully incorporated. Non-positive earnings forecasts were excluded. If three-year-ahead to five-year-ahead analyst forecasts are missing, I fill in missing values using long-term earnings growth rate forecasts. I use the average of a historical three-year payout ratio to calculate the expected dividend payout ratio. If the payout ratio is missing, or smaller (larger) than 0 (1), I use the country-median value instead. I exclude banks that do not have observations both before and after the mandatory IFRS adoption in 2005. I classify years before 2004 as the pre-mandatory adoption period and years from 2005 as the post-mandatory adoption period [6,30,31]. Table 1 presents the composition of the final sample, which has 376 observations from 52 banks in 12 countries having 7 voluntary adopters and 45 mandatory adopters. Among the 376 observations, 52 and 324 observations are obtained from voluntary and mandatory adopters, respectively. Only three countries, namely, Germany, Greece, and Poland, have voluntary adopters. However, voluntary adopters could not provide a good benchmark for difference-in-differences tests, because they are not evenly distributed. The sample selection did not drive this result. By examining the entire Compustat Global database, I confirm that only three countries have banks that voluntarily adopted IFRS. This result implies that IFRS adoption is more like an exogenous event than an endogenous one. Furthermore, this also implies that financial reporting and banking regulations could have conflicting goals.

Descriptive Statistics
Panel A in Table 2 shows the means of the main variables for regression analyses by country. Means of implied cost of equity are from 10% to 14%, whereas means of the regulatory capital ratio are larger than 10% and lower than 14.5%. Only three countries, namely, Germany, Greece, and Poland, have banks that voluntarily adopted IFRS. This implies that, unlike non-financial industries, European banks' IFRS adoption might be regulated by banks [6,25]. Voluntary adopters have a higher regulatory capital ratio than mandatory adopters in the same countries, suggesting the possibility that sound banks choose to adopt IFRS voluntarily to indicate their financial stability. Panel B presents the descriptive statistics for the full sample.  Table 3 presents the differences in bank characteristics before and after IFRS adoption. The variables in Table 3 are chosen differently from those in Panel B of Table 2, because the purpose of Table 3 is to present the changes in bank characteristics intuitively. CoC significantly changes after the mandatory IFRS adoption. However, this univariate test does not confirm that the difference is due to IFRS adoption.  Table 4 documents variables for institutional environments, and Table 5 presents the estimation results of model (1). ENFORCE, OFFICE, and CAPITAL have negative coefficients in both full and partial sample analyses. However, the coefficient on CAPITAL is insignificant in the partial sample analysis. These results imply that investor protection and bank regulation reduce banks' risk in general.  The interaction term of POST and OFFICE has positive coefficients, suggesting that banks in countries with strong regulations experience an increase in the cost of equity. Financial reporting and banking regulations conflict regarding IFRS adoption. In this case, banking regulations have priority over financial reporting [26]. Therefore, bank supervisors intervene in the IFRS adoption to suppress the negative impact of IFRS on the banking system, at least temporarily [3,25]. The intervention in IFRS adoption reduces the quality of accounting information and increases the uncertainty of banks and the cost of equity. This supports H4. However, the interaction terms of ENFORCE or CAPITAL with POST are insignificant; thus, H2 and H5 are not supported. The results are qualitatively consistent with the results for non-financial firms. Listed banks are also exposed to disclosure requirements; therefore, the results should be consistent with prior study [6].

The Changes in Disclosure Requirements by IFRS Adoption on Cost of Equity
Panel A of Table 6 documents the regression result of model (2). In column (A), the sign of the three-way interaction term of ENFORCE shows that legal enforcement facilitates the implementation of additional disclosure requirements by mandatory IFRS adoption, resulting in the decrease in banks' cost of equity. However, the coefficients on the three-way interaction terms of CAPITAL and OFFICE are insignificant. The result of the subsample period test presented in column (B) is qualitatively the same, except that the significance and magnitude are weaker. The results support the conjecture that the institutional environment for investor protection supports the implementation of IFRS adoption because it improves the relevance of accounting information on the equity valuation.  Panel B of Table 6 shows the influence of the improved comparability on banks' cost of equity. I use model (3) for this test. INC indicates the differences between IFRS and the pre-IFRS local accounting standards. Therefore, INC also proxies for the improved comparability across countries. Unlike ADD, INC indicates disclosure requirement changes to the pre-existing accounting standards. Therefore, from the banking regulators' point of view, INC could be a threat to the debt market because it relates to compliance with debt covenants of the pre-existing debt contracts. By contrast, ADD is likely unrelated to the compliance with debt covenants because ADD indicates new disclosure requirements. The items related to ADD were not in the previous accounting standards; hence, those items have little impact on debt covenants. Thus, the effect of bank supervisors' intervention is related to INC, not to ADD.
The three-way interaction term of CAPITAL and OFFICE is positive. This means that mandatory IFRS adoption increases the cost of equity in countries with stringent banking regulations and where the pre-existing accounting standards change significantly. As INC increases, bank supervisors' incentive to intervene in the implementation of IFRS increases because the adoption decreases the contractibility of accounting information [1,3,25]. Moreover, changes in accounting standards impact the regulatory capital ratio, which potentially impacts the stability of the banking system. As a result of the intervention, accounting standards are implemented to minimize the potential negative influence on the debt markets, which increases information risks.

Conclusions
This study examines the effect of mandatory IFRS adoption on European banks' cost of equity. The empirical results of this study show that the impact of IFRS adoption on banks' cost of equity varies depending on institutional aspects. Strong investor protection is helpful in decreasing the cost of equity following IFRS adoption. However, banking regulation increases banks' cost of capital, especially when IFRS adoption has a strong impact on debt contracts. These results show that market monitoring and bank regulation are potentially at odds because of differences in policy objectives. Consequently, the cost of capital is affected differently by IFRS adoption in two institutional aspects.
The results of this study have policy implications. Unlike other industry sectors, the banking sector has a strong regulatory environment. Therefore, the incentives of banking regulators must be considered when designing a disclosure policy for the banking sector. If these incentives are ignored, a disclosure policy can be distorted; hence, the intended results cannot be obtained. Furthermore, this policy can yield results opposite to the intended ones. In addition, although market discipline is an important part of the banking regulatory system, factors that enhance market discipline can easily be weakened by bank supervisors. As market discipline has become more important because of rapid innovations in the finance sector, policy makers should carefully design policies related to the banking system.
This study also has several limitations. First, due to availability, some countries are not included in the analysis, which could cause a selection bias. Second, I incorporate only two aggregate measures of bank regulation, which are not enough to explain every detail of banking regulation. Third, the effects of specific regulatory events that occurred during my sample period are not totally addressed in this study. Fourth, this study focuses only on listed banks. Several banks are unlisted; hence, market discipline on unlisted banks should be addressed in future research settings. Finally, this study does not address the impact of IFRS adoption on several aspects other than stock price; for example, credit allocation activities. These could be examined separately in other studies.