Toward Sustainable Accounting Information: Evidence from IFRS Adoption in Korea

Abstract

The harmonization of accounting standards has been an international trend in the past two decades. As of 2018, 144 of 166 profiled jurisdictions require the use of International Financial Reporting Standards (IFRS). Nevertheless, there is mixed evidence on the effect of IFRS on sustainable accounting information. This study examines whether IFRS adoption improves earnings sustainability, focusing on emerging markets. Specifically, it tests the effect of IFRS on earnings quality by comparing earnings management and financial statement comparability of Korean listed firms for the pre- and post-IFRS periods. The results show that firms report less managed earnings in the post-IFRS period than in the pre-IFRS period. Furthermore, the results suggest the enhancement of financial statement comparability in the post-IFRS period compared to the pre-IFRS period. In particular, this paper documents that the effect of IFRS on sustainable accounting information is more pronounced in competitive industries. Moreover, it shows that small firms benefit more from adopting IFRS. Overall, this study finds that IFRS adoption in Korea improves the overall sustainability of accounting information.

1. Introduction

This study examines whether earnings sustainability differs between pre- and post-International Financial Reporting Standards (IFRS) periods. Specifically, it tests the effect of IFRS on earnings quality by comparing the earnings management and financial statement comparability of Korean listed firms between the pre- and post-IFRS periods.

The goal of the IFRS is to establish an internationally acceptable set of financial reporting standards [1]. To achieve this, the International Accounting Standards Board (IASB), formerly the International Accounting Standards Committee (IASC), has been attempting to establish a new accounting standard since 1975. Consequently, as of 2018, 144 of 166 profiled jurisdictions require the use of IFRS. Among the Group of 20 (G20) countries, 15 have adopted the IFRS (www.ifrs.org). Following these global trends in accounting harmonization, Korea too decided to adopt the IFRS in 2007, and all listed firms and financial institutions in Korea have adopted the IFRS since 2011.

The movement toward the adoption of a global accounting standard has generated attention and debate in both academics and practice. The purpose of establishing and adopting a new international standard is in order to enhance the quality of financial reporting and increase its comparability across firms and countries [1]. However, research has provided mixed evidence on whether the new accounting standard exhibits higher earnings quality [2,3,4,5]. Reflecting this, several jurisdictions permit rather than require the IFRS standards (i.e., Japan and Switzerland). Additionally, seven jurisdictions including the United States, Vietnam, and India still use national standards (www.ifrs.org).

This study differs from previous studies in that it focuses on an emerging market. Studies that investigate the effect of IFRS on sustainable earnings generally test firms from developed countries. For instance, most observations in Landsman et al.’s [2] sample are from the United Kingdom and Japan. Similarly, Horton et al. [3] use firms from Australia, France, Singapore, Sweden, Hong Kong, and the United Kingdom as mandatory adopters, while the majority of voluntary adopters are from Germany, Italy, and Switzerland. However, this study considers the Korean stock market, in which firms may benefit more from adopting the IFRS than those in a developed market because of its less transparent investing environment. The reason for adopting IFRS in Korea was to overcome the so-called “Korea discount,” which refers to the phenomenon of the undervaluing of stock prices of Korean firms by foreign investors because of their less transparent accounting information. For instance, both foreign investors and the International Monetary Fund pointed out that the low-quality accounting system was one of the factors that triggered the financial crisis in 1997. Thus, it is valuable to test whether the adoption of IFRS eventually improved the quality of financial statements.

This study examines two attributes of sustainable accounting information: earnings management through accruals and financial statement comparability. This study compares pre- and post-IFRS adoption periods in terms of their relative quality of earnings. The findings show that firms report less managed earnings in the post-IFRS period than in the pre-IFRS period. Furthermore, the results suggest an enhancement in financial statement comparability in the post-IFRS period compared to the pre-IFRS period. In additional tests, this paper documents that the effect of IFRS on sustainable accounting information is more pronounced in competitive industries. Moreover, it shows that small firms that suffer from the lack of internal resources and inferior information environment benefit more from adopting IFRS. Ultimately, this study finds that IFRS adoption improves the overall sustainability of accounting information.

This study contributes to the literature in several ways. First, it provides further evidence on the consequences of IFRS by investigating the several attributes of earnings of Korean firms. Specifically, the results suggest that IFRS enhances the quality of financial statements of Korean listed firms. Second, the results should be of interest to stakeholders who use financial statements. For example, analysts from outside Korea may use the reported earnings of firms in Korea more efficiently, as the adoption of IFRS enhances their sustainability. Investors willing to invest in the Korean stock market could benefit from the results of this study as it shows that IFRS adoption has improved the comparability of financial statements among firms in Korea. Since the primary purpose of adopting IFRS in Korea was to increase transparency in accounting information, the results of this study provide relevant evidence on this. Moreover, although previous studies found a positive association between IFRS adoption and accounting quality mostly by focusing on European Union countries, there is little evidence from the emerging market. This study fills this gap by showing the enhanced accounting quality after adopting IFRS in Korea. Nevertheless, the results of this paper are not generalizable to other countries as it only focuses on one country.

The remainder of this paper is arranged as follows. Section 2 describes the related literature and develops hypotheses. Section 3 provides the research design and the sample. Section 4 presents the empirical results and Section 5 concludes the paper.

2. Prior Research and Hypotheses Development

2.1. Prior Research on Sustainable Earnings

When discussing sustainable earnings, most authors usually presume the persistence of earnings [6]. A firm can use deferred revenues or accrued expenses to manipulate earnings temporarily (a phenomenon known as a “cookie jar”). These temporarily increased or decreased accruals reverse in subsequent years, reducing earnings sustainability. Baber et al. [7] document that such reversals depend on both the magnitude and the reversal speed of past discretionary accruals. Prior research uses several proxies of earnings sustainability to capture these short-period reversal effects of managed earnings. Following previous studies, this study uses two attributes of earnings: earnings management through accruals and financial statement comparability.

First, prior research on earnings management documents that earnings that increase through discretionary accruals may reverse in subsequent periods and deteriorate earnings sustainability. For instance, Teoh and Wong [8] provided evidence that highly managed earnings through accruals in the Initial Public Offerings (IPO) year experience poor stock returns in the three years after the IPO. Their findings show that aggressive managers experience 20 percent lower stock returns than conservative managers. Recently, Lizińska and Czapiewski [9] found similar results, showing that positive and significant discretionary accruals in the IPO year are negatively correlated with subsequent long-term market value. Rangan [10] used seasoned equity offerings and examines earnings management around offerings. Specifically, investors temporarily overvalue the earnings sustainability of issuing firms, and subsequently, are disappointed by the decline in earnings due to earnings management.

Second, financial statement comparability is used as a proxy of sustainable earnings. Comparability refers to “the quality of information that enables users to identify similarities and differences between two sets of economic phenomena” [11]. Practically, it refers to the accounting systems of two firms reporting similar accounting information for an identical economic event. Firms with similar earning predictability and smoothness will have a higher comparability of financial statements. Comparable accounting information is more sustainable because it has little room for intentional manipulation of financial statements. De Franco et al. [12] introduced a proxy to measure financial statement comparability. The study finds that their measure of financial statement comparability is positively related to analyst following and forecast accuracy, and negatively related to the forecast dispersion of analysts. Recent studies document the relationship between financial statement comparability and other attributes of accounting information. For instance, Sohn [13] found that financial statement comparability decreases accruals-based earnings management, whereas it increases real activity-based earnings management. Choi [14] documented the positive association between earnings persistence and comparability. Several studies investigate the economic consequences of financial statement comparability. For instance, Kim et al. [15] found evidence that expected stock price crash risk decreases with financial statement comparability. Imhof et al. [16] suggested that greater financial statement comparability is associated with a lower cost of equity capital. To summarize, previous studies have generally found evidence that financial statement comparability results in favorable economic consequences for the capital market.

2.2. Prior Research on the Effect of IFRS on Sustainable Earnings

The mission of IFRS is “to develop Standards that bring transparency, accountability and efficiency to financial markets around the world” (www.ifrs.org). However, research has revealed mixed evidence on whether the new accounting standards exhibit a higher earnings quality.

Studies that document positive consequences with regard to earnings quality insist that firms from IFRS adoption countries experience a greater increase in the information content of earnings than firms from non-adoption countries. For instance, Barth et al. [1] documented that firms applying a new accounting standard show a smaller earnings management, more timely loss recognition, and more value relevance of accounting information than those who do not adopt the new standard. Landsman et al. [2] found that firms from IFRS adopting countries experienced a greater increase in abnormal return volatility and abnormal trading volume. Horton et al. [3] found that the mandatory IFRS adoption affects the forecast accuracy positively compared to non-adopters and voluntary adopters. The study shows that the findings are driven by comparability benefits. Similar to these findings, Yip and Young [17] suggested that mandatory IFRS adoption improves cross-country information comparability. For instance, the paper documented that the similarity of accounting functions, the degree of information transfer, and the similarity of the information content of earnings and of the book value of equity are improved after adopting IFRS.

Other studies have documented the negative consequences of IFRS regarding the quality of earnings. Atwood et al. [4] found that losses reported under IFRS are less persistent than under US-GAAP. Moreover, their study shows that the association between future cash flows and current earnings is lower for firms that report financial statement under IFRS than under US-GAAP. Using 22 European countries between 2000 and 2010, Doukakis [5] suggested that mandatory IFRS adoption had no impact on either real- or accrual-based earnings management.

Next, although not many, few studies have investigated the change in accounting quality under IFRS focusing on emerging markets. For instance, by using Abu Dhabi Stock Exchange data, Alali and Foote [18] documented that IFRS contributes to the value relevance of accounting information. Similarly, using firms listed in China, Liu et al. [19] reported decreases in earnings management and increases in value relevance of accounting information after the mandatory adoption of IFRS. Different from the above two studies, Cho et al. [20] who used Korean listed firms, failed to document the assumption that IFRS adoption had strengthened the negative relationship between earnings quality and information asymmetry.

Collectively, although its primary mission is to establish high-quality financial statements through worldwide harmonization of accounting standards, the impact of IFRS on accounting quality is still under debate.

2.3. IFRS Adoption in Korea and Hypotheses Development

Both foreign investors and the International Monetary Fund noted that one factor that triggered the Asian financial crisis in 1997 was the low-quality accounting system. The debate to improve the quality of financial statements has been raised since then. Another purpose of this debate, raised in Korea, was to address the so-called “Korea discount,” which refers to the phenomenon of undervaluing of stock prices of Korean firms by foreign investors because of their less transparent accounting information. Consequently, Korea decided to adopt the IFRS in 2007, and since 2011, all listed firms and financial institutions in Korea have adopted the IFRS.

This study investigates whether the initial purpose of adopting IFRS has been successfully implemented in the Korean stock market. Specifically, it predicts that firms report more sustainable accounting information proxied by less earnings management and greater financial statement comparability due to IFRS adoption for the following reasons.

First, a feature of the IFRS is that it enhances the transparency of financial statements, which lowers the monitoring costs of stakeholders. This may act as a disincentive for managers who attempt to engage in earnings management [1]. Furthermore, internationally accepted accounting standards may improve the comparability of financial statements, which leads to stricter monitoring by foreign investors. Therefore, this study expects a negative relationship between IFRS adoption and earnings management.

Second, this study expects an improvement in the comparability of financial statements due to increased managerial discretion and the requirement of reporting consolidated financial statements. Since a higher level of financial statement comparability leads to market efficiency by improving the investing decisions of financial statements users, managers have incentives to increase financial comparability [12]. A previous study also demonstrates that an increase in financial statement comparability prevents stock price crash risk [15]. Thus, as the IFRS allows a certain level of managerial discretion in financial reporting, managers can engage in activities that lead to higher comparability. Moreover, in the pre-IFRS era, firms in Korea emphasized separate financial statements to hide their complicated ownership [21]. The IFRS requires firms to report financial statements based on consolidated financial statements, which provides more relevant and comparable accounting information.

To distinguish between the IFRS adopted in Korea and other countries, this study uses the Korean International Financial Reporting Standards (K-IFRS) in the hypotheses. In addition, as the GAAP was the domestic accounting standard used before the IFRS, the hypotheses use the Korean Generally Accepted Accounting Principles (K-GAAP) to distinguish them from the GAAP used in other countries. The study sets forth the following hypotheses.

Hypothesis 1:

Earnings management is lower under K-IFRS than under K-GAAP.

Hypothesis 2:

Financial statement comparability is higher under K-IFRS than under K-GAAP.

3. Research Design

3.1. Proxies of Sustainable Earnings—Earnings Management

This study uses two proxies of sustainable earnings. First, following prior studies on earnings management, this study uses discretionary accruals [22,23]. Discretionary accruals are the residuals from the following cross-sectional Modified Jones model.

Accruals_it/TA_it-1 = α0 + α1(1/TA_it-1) + α2(ΔSales_it − ΔREC_it)/ TA_it-1 + α3PPE_it /TA_it-1 + ε_it

(1)

Here, i and t denote the firm and fiscal year, respectively. Accruals refers to the total accruals defined as net income minus cash flow from operations. TA refers to the total assets. All variables are scaled based on the lagged total assets. ΔSales and ΔREC are the change in sales and accounts receivables, respectively. PPE is the net property, plant, and equipment. Equation (1) is estimated for each industry-year and the residuals of the model are discretionary accruals (DA). This study uses both the raw values of residuals (DA) and the absolute value of residuals (AbsoluteDA). DA captures income-increasing earnings management while AbsoluteDA indicates accruals management through both income-increasing and -decreasing management. This study uses DA along with AbsoluteDA because a firm can also manipulate earnings downwards. For instance, accumulating reserves are examples of income-decreasing earnings management—a case of a “cookie jar.”

3.2. Proxies of Sustainable Earnings—Financial Statement Comparability

Next, this study uses financial statement comparability as another proxy of sustainable earnings. De Franco et al. [12] first introduced a measure to capture financial statement comparability. If a firm i reports similar financial statements to firm j, which faces identical economic events, then the two firms’ financial statements are comparable. Their study uses stock returns as a proxy of economic events. Thus, a firm’s financial statements are a function of the stock return. Using accounting earnings as a summary of financial statements, De Franco et al. [10] developed the following equation.

Earnings_it = α_i + β_i Return_it + ε_it

(2)-1

Earnings_it = α_i + β_i Return_it + γ_iLagged Return_it + ε_it

(2)-2

For each firm-year, the above equations are estimated using the 16 previous quarters. In Equation (2)-1, the estimated coefficients (α_i and β_i) are the proxies for the accounting system of firm i. In Equation (2)-2, Lagged Return_it is included to reflect the tendency that stock price leads earnings. Here, α_i, β_i, and γ_i reflect the accounting system of firm i.

Next, using firm i’s accounting system and returns, the predicted earnings of E(Earnings)_iit are generated. Similarly, using firm j’s estimated coefficients and firm i’s returns, the predicted earnings based on firm i’s economic events and firm j’s accounting system are generated (E(Earnings)_ijt). The same procedure for Equation (2)-1 is conducted in Equation (2)-2, but by including γ_i as another factor in the firm’s accounting system.

Lastly, the financial statement comparability between firm i and j is defined as the negative value of the average absolute difference between the predicted earnings using i and j’s accounting system, as follows:

$$F/S {\mathit{Comparability}}_{\mathit{ijt}}=(-1/16)\times {{\displaystyle \sum }}_{t-15}^t\left|E\left(Earnings\right)iit-E\left(Earnings\right)ijt\right|$$

(3)

In Equation (3), a negative value of one is multiplied to describe greater value indicates higher accounting comparability. To estimate a firm-year measure of comparability, following De Franco et al. [10], this study measures the financial statement comparability as the average of the highest four firms with j being among the firm i-j combinations with respect to its level of comparability. Comp1 and Comp2 refer to the final measures of financial statement comparability derived from Equations (2)-1 and (2)-2, respectively.

3.3. Research Model

The study tests the two hypotheses by examining the change in the levels of discretionary accruals and financial statement comparability between the pre- and post-IFRS period. Specifically, it estimates the following two equations:

DA (AbsoluteDA) = α₀ + α₁IFRS + α₂SIZE + α₃LEV + α₄Quick + α₅CF + α₆INVREC + α₇SalesGRW + α₈LOSS + Industry Fixed Effects + ε

(4)

Comp1 (Comp2) = α₀ + α₁IFRS + α₂SIZE + α₃LEV + α₄Quick + α₅CF + α₆INVREC + α₇SalesGRW + α₈LOSS + Industry Fixed Effects + ε

(5)

Here, DA refers to the discretionary accruals; AbsoluteDA refers to the absolute value of DA; IFRS equals one for the post-IFRS period and zero for the pre-IFRS period; SIZE refers to the natural logarithm of the total assets; LEV refers to the total liabilities divided by total assets; Quick refers to the current assets divided by the current liabilities; CF refers to the operating cash flows divided by total assets; INVREC refers to the sum of inventories and accounts receivables divided by the total assets; SalesGRW refers to the sales minus the lagged sales divided by lagged sales; Loss equals one for loss reporting firms and zero otherwise; Industry FE refers to the industry fixed effects.

Equation (4) is examined to test hypothesis 1. The dependent variables are either DA or the absolute value of DA. The main coefficient of interest is α₁, which measures the change in the levels of earnings management after the adoption of IFRS in comparison to the pre-IFRS period. Under the assumption of hypothesis 1, α₁ is expected to be negative if the adoption of IFRS decreases the level of earnings management.

Next, Equation (5) shows the model for testing hypothesis 2. The dependent variables are Comp1 and Comp2, which are derived by equations in Section 3.2. The main coefficient of interest is α₁, which measures the change in the levels of financial statement comparability after the adoption of IFRS in comparison to the pre-IFRS period. A positive and significant α₁ supports hypothesis 2. The same control variables from Equation (4) are used in this model.

Control variables known to be factors that affect financial reporting practices are included. Firm size (SIZE) is included as a prior study documents that larger firms are more likely to engage in income-increasing activities, as their complex operations make it more difficult to detect earnings management [24]. SIZE is defined as a natural logarithm of the total assets, which is most commonly used in the related literature [25]. The financial leverage ratio (LEV) is controlled to pick up the leveraged firms’ higher incentives for managing firm profits. Quick and CF are included to control liquidity. INVREC is included as large inventories and accounts receivables indicate a more complex operating environment. Change in sales (SalesGRW) is included to mitigate the effect of firm growth rate on earnings management. Reporting losses (Loss) are included to control the effect of the tendency to avoid losses by engaging in earnings management. Lastly, industry fixed effect dummies are included to mitigate industry effects on the dependent variable. The models do not include year fixed dummies as they are correlated with IFRS variable causing a multicollinearity problem. All the statistics in this study are corrected for firm-level clustering [26].

3.4. The Sample

All financial statement data are obtained from the Korea Listed Companies Association’s TS2000 database. This study uses public firms only because all listed firms must adopt IFRS mandatorily while unlisted firms can adopt IFRS voluntarily.

Panel A of

Table 1. The sample.

Panel A. Sample Selection Procedure
Initial observations from 2009 to 2016				14,498
Less:
		Missing discretionary accruals and control variables	(1285)	13,213
		Missing financial statement comparability variables	(5997)	7216
Final observations for main tests		Pre-IFRS	Post-IFRS	Total
		1440	5776	7216
Final observations for PSM tests		1401	1401	2802
Panel B. Yearly Distribution
Year	2009	2010	2011
N	672	768	830
Year	2012	2013	2014
N	883	942	983
Year	2015	2016	Total
N	1049	1089	7216

describes the sample selection procedure. The initial sample consists of all public firms that belong to non-financial industries. The number of observations in this initial sample, from 2009 to 2016, is 14,498. The study excludes missing observations necessary to compute discretionary accruals and control variables. After deleting firms that do not satisfy these criteria, 13,213 observations remain. The study also excludes missing observations necessary to compute financial statement comparability variables. Since this measure requires the earnings and stock returns data of the previous 16 quarters, a large number of observations (5997) are lost. Imposing these data requirements yields a final sample of 7216 firm-years. Among those observations, 1440 and 5776 observations refer to the Non-IFRS and IFRS samples, respectively.

The study also matches Non-IFRS firm-years with IFRS firm-years using the propensity score. It conducts one-to-one matching. The final samples used in this matched sample tests are 1401 and 1401 for the Non-IFRS and IFRS sample, respectively. The additional test section contains further details.

Panel B of Table 1 shows the yearly distribution of the sample. The pre-IFRS period refers to 2009 and 2010. Observations for the two years are 672 and 768. As Korea mandatorily adopted the IFRS since 2011, the remaining years refer to the post-IFRS period. The number of firms increases by year, reflecting the economic growth of the Korean stock market.

4. Empirical Results

4.1. Descriptive Statistics

Panel A of

Table 2. The descriptive statistics and correlations.

Panel A. Descriptive statistics
Variable		Mean			STD			P25				Median			P75
IFRS		0.800			0.400			1.000				1.000			1.000
DA		−0.012			0.095			−0.051				−0.008			0.032
AbsoluteDA		0.067			0.076			0.018				0.042			0.085
Comp1		−0.010			0.012			−0.011				−0.006			−0.004
Comp2		−0.012			0.013			−0.013				−0.007			−0.005
SIZE		19.120			1.415			18.167				18.833			19.819
LEV		0.419			0.200			0.258				0.418			0.565
Quick		1.737			1.969			0.671				1.080			1.902
CF		0.041			0.080			−0.001				0.042			0.088
INVREC		0.266			0.148			0.154				0.252			0.364
SalesGRW		0.058			0.291			−0.078				0.028			0.140
LOSS		0.271			0.444			0.000				0.000			1.000
Panel B. Pearson correlations
Variable	1.		2.	3.		4.	5.		6.	7.	8.		9.	10.		11.	12.
1. IFRS	1.00
2. DA	−0.07		1.00
3. AbsoluteDA	−0.07		−0.23	1.00
4. Comp1	0.05		0.12	−0.28		1.00
5. Comp2	0.05		0.12	−0.30		0.98	1.00
6. SIZE	0.00		0.10	−0.18		0.23	0.25		1.00
7. LEV	−0.03		−0.12	0.17		0.00	−0.01		0.17	1.00
8. Quick	0.03		0.04	−0.03		−0.08	−0.07		−0.18	−0.63	1.00
9. CF	−0.01		−0.35	−0.18		0.15	0.17		0.17	−0.18	0.06		1.00
10. INVREC	−0.03		0.12	0.01		0.15	0.15		−0.15	0.19	−0.19		−0.04	1.00
11. SalesGRW	−0.12		0.18	0.07		−0.08	−0.09		−0.03	0.00	0.01		0.06	0.06		1.00
12. LOSS	0.07		−0.38	0.26		−0.22	−0.24		−0.17	0.28	−0.10		−0.38	−0.08		−0.18	1.00

Bolds indicate significant at 1% or higher level. See Section 3.3 for variable definition.

presents the descriptive statistics of variables. All the continuous variables are winsorized at the top and bottom of 1%. The mean value of IFRS is 0.800, indicating that 80% of the sample belongs to the post-IFRS period. The mean values of DA and AbsoluteDA are −0.012 and 0.067, respectively. Since financial statement comparability measures are the values generated from multiplying negative one with the absolute differences of the two firms’ expected earnings; the mean values of these variables are below zero. Specifically, the mean values of Comp1 and Comp2 are −0.010 and −0.012, respectively. Other control variables show comparable descriptive statistics to previous studies. For instance, the mean value of SIZE is 19.120. The mean values of LEV, Quick, and CF are 0.419, 1.737, and 0.041, respectively. The firm growth rate based on sales is 5.8% on average. About 27.1% of firm-years report losses.

Panel B of Table 2 shows Pearson correlations among the variables. The IFRS is negatively related to both DA and AbsoluteDA (correlation coefficient = −0.07). In addition, the IFRS is positively related to both Comp1 and Comp2 (correlation coefficient = 0.05), supporting the hypotheses, at least on these univariate correlations. The IFRS is negatively related to INVREC and SalesGRW while it is positively related to LOSS.

Next,

Table 3. The univariate mean differences tests.

Variable	Post-IFRS	Pre-IFRS	Differences
DA	−0.016	0.001	−5.53	***
AbsoluteDA	0.064	0.077	−5.69	***
Comp1	−0.010	−0.011	3.52	***
Comp2	−0.012	−0.013	3.81	***
SIZE	19.123	19.109	0.32
LEV	0.417	0.431	−2.57	**
Quick	1.764	1.630	2.43	**
CF	0.041	0.042	−0.43
INVREC	0.264	0.275	−2.69	***
SalesGRW	0.041	0.128	−9.16	***
LOSS	0.286	0.209	6.29	***

***, **, and * denote the significance level at the 1, 5, and 10% level. T-statistics are shown in the differences. See Section 3.3 for variable definition.

shows the univariate mean comparisons across the pre- and post-IFRS adoption periods of variables used in the analysis. Most of the variables are significantly different across the pre- and post-IFRS adoption periods with respect to their mean values. DA and AbsoluteDA are lower in the post-IFRS period than in the pre-IFRS period. In contrast, Comp1 and Comp2 are higher in the post-IFRS period than the pre-IFRS period. Thus, the results in Table 3 are consistent with the hypotheses. Figure 1 visualizes the same statistics.

4.2. Hypotheses Tests

Table 4. The effect of International Financial Reporting Standards (IFRS) on earnings management.

Independent Variables	(1) Dependent Variable = DA		(2) Dependent Variable = AbsoluteDA
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.143 ***	(−4.38)	0.208 ***	(8.18)
IFRS	−0.007 ***	(−3.01)	−0.012 ***	(−5.44)
SIZE	0.010 ***	(11.69)	−0.008 ***	(−8.98)
LEV	−0.072 ***	(−8.95)	0.074 ***	(7.76)
Quick	−0.000	(−0.48)	0.003 ***	(3.95)
CF	−0.728 ***	(−37.15)	−0.050 **	(−2.51)
INVREC	0.067 ***	(7.29)	−0.005	(−0.60)
SalesGRW	0.040 ***	(8.38)	0.024 ***	(5.28)
LOSS	−0.108 ***	(−35.70)	0.028 ***	(9.41)
Industry	Included		Included
Adj. R–sq.	48.48%		14.98%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering.

presents the empirical findings for hypothesis 1. All the t-statistics are corrected for firm-level clustering. Model (1) reports the results where the discretionary accruals (DA) is included as a dependent variable, while model (2) includes the absolute value of discretionary accruals (AbsoluteDA). The coefficients of IFRS in both models are −0.007 and −0.012, respectively, and all are significant at the 1% level. These significant and negative coefficients suggest that the firms experience a decrease in accrual earnings management after IFRS adoption. This evidence supports hypothesis 1.

In addition, this result supports previous studies that document the positive associations between IFRS adoption and earnings quality as well as the information environment [1-3, 18,19]. Control variables show significant coefficients, meaning that the model is well specified and mitigates the omitted variable problems. For instance, SIZE, INVREC, and SalesGRW are positively associated with DA, while LEV, CF, and Loss are negatively related to DA. Similarly, most of the control variables in model (2) are significant at the 5% level or higher. Collectively, as predicted in hypothesis 1, earnings management is lower under the IFRS than the pre-IFRS accounting standards (GAAP).

Next,

Table 5. The effect of IFRS on financial statement comparability.

Independent Variables	(1) Dependent Variable = Comp1		(2) Dependent Variable = Comp2
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.057 ***	(−10.00)	−0.065 ***	(−10.12)
IFRS	0.001 ***	(4.35)	0.002 ***	(4.75)
SIZE	0.002 ***	(8.70)	0.002 ***	(9.05)
LEV	−0.004 **	(−2.29)	−0.005 ***	(−2.82)
Quick	−0.000 **	(−2.52)	−0.000 **	(−2.37)
CF	0.010 ***	(3.14)	0.012 ***	(3.55)
INVREC	0.008 ***	(4.79)	0.010 ***	(5.35)
SalesGRW	−0.004 ***	(−5.83)	−0.005 ***	(−5.76)
LOSS	−0.004 ***	(−7.83)	−0.004 ***	(−7.69)
Industry	Included		Included
Adj. R-sq.	42.96%		40.80%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering.

presents the evidence from testing hypothesis 2. Similar to the previous table, all the t-statistics are corrected for firm-level clustering. Model (1) reports results when Comp1 is included as a dependent variable, while model (2) includes Comp2 as a dependent variable. The coefficients of IFRS in both models are 0.001 and 0.002, respectively, and all are significant at the 1% level. These significant and positive coefficients suggest that after IFRS adoption, firms experience an increase in financial statement comparability. This evidence supports hypothesis 2 and previous studies (e.g., Reference [17]) that document a positive association between IFRS adoption and financial statement comparability. All the control variables show significant coefficients. To summarize, as predicted in hypothesis 2, financial statement comparability is higher under IFRS than under GAAP.

4.3. Additional Tests

As reported in Table 3, firms differ between pre- and post-IFRS periods. This may cause the endogeneity problem because the results may be derived by different firm characteristics and not by the IFRS adoption itself. To mitigate this concern, this study utilizes the propensity score matching (PSM) technique. The PSM generates propensity scores by matching pre-IFRS firms to post-IFRS firms. Scores are generated using all control variables used in the previous tests. This study includes all controls to reduce self-selection bias and uses a matched sample within 3% caliper distance, which is the conventional distance in this research stream. As the observations are based on one-to-one matching, the final sample for pre-IFRS and post-IFRS periods is 1401 each.

Panel A of

Table 6. The additional test: using a propensity score matched sample.

Panel A. Univariate mean differences of matched sample
Variable	Post-IFRS		Pre-IFRS			Differences
SIZE	19.079		19.119			−0.74
LEV	0.437		0.429			1.08
Quick	1.603		1.642			−0.56
CF	0.042		0.042			−0.04
INVREC	0.275		0.276			−0.15
SalesGRW	0.103		0.111			−0.71
LOSS	0.211		0.212			−0.58
Panel B. Earnings management and IFRS
Independent Variables		(1)Dependent variable = DA			(2) Dependent variable = AbsoluteDA
		Coefficient		(T-stat)	Coefficient			(T-stat)
Intercept		−0.137 ***		(−2.85)	0.201 ***			(6.46)
IFRS		−0.010 ***		(−3.51)	−0.013 ***			(−4.21)
Controls		Included			Included
Industry		Included			Included
Adj. R-sq.		54.38%			15.95%
N. of obs.		2802			2802
Panel C. F/S comparability and IFRS
Independent Variables		(1) Dependent variable = Comp1			(2) Dependent Variable = Comp2
		Coefficient		(T-stat)	Coefficient			(T-stat)
Intercept		−0.055 ***		(−10.32)	−0.063 ***			(−10.30)
IFRS		0.001 ***		(3.87)	0.002 ***			(4.09)
Controls		Included			Included
Industry		Included			Included
Adj. R-sq.		43.57%			41.89%
N. of obs.		2802			2802

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering. Controls are included but omitted for brevity.

presents the univariate mean comparison between pre-IFRS and post-IFRS using the matched sample. All the differences of control variables are now insignificant, showing that the matching is well constructed.

Panel B and C of Table 6 present the findings of using the PSM sample. Panel B reports the testing results of hypothesis 1. The coefficients of IFRS remain negative and significant. Panel C presents the results of hypothesis 2, and the coefficients of IFRS remain positive and significant. Overall, the PSM results are consistent with the previous findings of this study.

To check the robustness of the results, this study uses alternative definitions of earnings management and financial statement comparability in its analysis. First, as Kothari et al. [27] suggested, controlling extreme performance is important when using DA, this study follows their method and uses performance-matched DA (PMDA). Panel A of

Table 7. Using other proxies of earnings management and financial statement comparability.

Panel A. Earnings management and IFRS
Independent Variables	(1) Dependent Variable = PMDA		(2) Dependent Variable = AbsolutePMDA
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.155 ***	(−5.36)	0.184 ***	(8.16)
IFRS	−0.001	(−0.66)	−0.012 ***	(−5.96)
Controls	Included		Included
Industry	Included		Included
Adj. R-sq.	42.55%		12.45%
N. of obs.	7216		7216
Panel B. F/S comparability and IFRS
Independent Variables	(1) Dependent Variable = Comp1_median		(2) Dependent Variable = Comp2_median
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.071 ***	(−9.43)	−0.079 ***	(−9.55)
IFRS	0.001	(1.56)	0.001 **	(2.21)
Controls	Included		Included
Industry	Included		Included
Adj. R-sq.	29.94%		29.72%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering. Controls are included but omitted for brevity.

presents the results. The coefficients of IFRS remain negative and significant in the case of AbsolutePMDA, as for the dependent variable.

Next, as the previous comparability proxy uses the mean value of the higher four firms, in the additional test, this study alternatively uses the median value for all firms (firm j) in the same industry to check the robustness of the results. Panel B of Table 7 shows the findings are weaker, but generally robust, to this alternative definition of financial statement comparability.

Overall, although the results of this tests are somewhat weak, it remains consistent in the use of several alternative proxies of earnings management and financial statement comparability.

Furthermore, this study examines the channels of how IFRS enhances accounting quality. For instance, external and internal governance mechanisms might affect the previous findings.

First, as for the external governance mechanism, this study examines whether market competition facilitates the effect of IFRS on accounting quality. Giroud and Mueller [28] documented that market competition works as external governance by showing that the firms in non-competitive industries benefit more from good internal governance.

By following Giroud and Mueller, the results of using the Herfindahl–Hirschman index (HHI) as a proxy of market competition are shown in

Table 8. The effect of market competition on the relationship between IFRS adoption and sustainable accounting information.

Panel A. Earnings management and IFRS
Independent Variables	(1) Dependent Variable = DA		(2) Dependent Variable = AbsoluteDA
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.133 ***	(−7.22)	0.198 ***	(11.11)
IFRS	−0.006 *	(−1.77)	−0.017 ***	(−4.98)
Competition	−0.022 *	(−1.73)	−0.002	(−0.17)
IFRS*Competition	−0.004	(−0.28)	−0.020 *	(−1.71)
Controls	Included		Included
Industry	Not Included		Not Included
Adj. R–sq.	47.11%		12.84%
N. of obs.	7216		7216
Panel B. F/S comparability and IFRS
Independent Variables	(1) Dependent Variable = Comp1		(2) Dependent Variable = Comp2
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.042 ***	(−10.16)	−0.050 ***	(−11.19)
IFRS	0.002 ***	(4.56)	0.002 ***	(3.94)
Competition	0.017 ***	(5.61)	0.019 ***	(5.83)
IFRS*Competition	0.007 ***	(2.90)	0.005 **	(2.05)
Controls	Included		Included
Industry	Not Included		Not Included
Adj. R-sq.	23.12%		23.49%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering. Controls are included but omitted for brevity.

. HHI is calculated based on sales. As a decrease in HHI refers to higher competition, market competition is measured by multiplying negative one to HHI (Competition). The mean value of raw HHI variable is 0.160 and the standard deviation is 0.167. Industry dummies are excluded due to a presence in multicollinearity with Competition.

In Panel A of Table 8, the interaction term of IFRS*Competition is negatively significant when the dependent variable is AbsoluteDA. This shows that the negative impact of IFRS on earnings management is more pronounced for firms that belong to competitive industries. Moreover, Panel B of Table 8 shows even stronger evidence that IFRS*Competition are positive and significant in both Comp1 and Comp2 models. These results suggest that higher market competition boosts the positive effect of IFRS on financial statement comparability. Collectively, Table 8 shows that the increase in market competition enhances the impact of IFRS on sustainable accounting information.

Second, as for the proxy of internal governance mechanism, this paper investigates the effect of managerial ownership on previous findings. The mean value of CEO ownership is 0.079 and its standard deviation is 0.115. The untabulated results show an insignificant impact regarding managerial ownership. As the interaction terms of IFRS*CEO_Ownership are insignificant on both earnings management and financial statement comparability, these results could be interpreted as CEO ownership not being associated with IFRS and accounting quality. It is plausible because as most of the listed firms in Korea are controlled by the largest shareholder and its related family parties (so-called Chaebol), the hired external CEO has limited discretion on financial reporting in practice.

Although the current paper does not directly test the effects of other governance mechanisms on sustainable accounting information, it is plausible that good governance might positively work when a firm first applies IFRS into its financial reporting. Previous studies have suggested that effective governance mechanisms help firms to report high-quality financial statements after the adoption of IFRS [29,30]. For instance, Verriest et al. [29] suggested that strong governances proxied by board independence, board functioning and audit committees have the potential to affect the convergence of IFRS. There is, however, still much to be learned about the consequences of good governance on accounting quality along with IFRS adoption. Past research on IFRS has generally focused on the IFRS itself when they investigated its impact on sustainable accounting information. Only a few studies examined the moderating effect of country-specific legal/regulatory enforcement (e.g., Reference [2]). Thus, it will be interesting for future studies to consider the effect of firm-specific governance on the accounting quality around IFRS adoption.

Next, this study investigates the effect of firm size on the previous results. As Dang et al. [25] suggested, firm size is commonly used as a control variable in empirical tests and is sensitive to the findings; the results of this paper could differ for large and small firm groups. As firm size captures a firm’s information environment, it is plausible that small firms having an inferior environment benefit more from adopting IFRS. This paper reconciles this prediction into the test by interacting the firm size variable with the IFRS variable.

The results are shown in

Table 9. The effect of firm size on the relationship between IFRS adoption and sustainable accounting information.

Panel A. Earnings management and IFRS
Independent Variables	(1) Dependent Variable = DA		(2) Dependent Variable = AbsoluteDA
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.145 ***	(−4.06)	0.237 ***	(7.96)
IFRS	−0.009 ***	(−3.42)	−0.015 ***	(−4.84)
Small_Firm	−0.003	(−0.70)	−0.009 *	(−1.86)
IFRS*Small_Firm	0.005	(1.11)	0.005	(1.06)
Controls	Included		Included
Industry	Included		Included
Adj. R-sq.	48.48%		15.02%
N. of obs.	7216		7216
Panel B. F/S comparability and IFRS
Independent Variables	(1) Dependent Variable = Comp1		(2) Dependent Variable = Comp2
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.053 ***	(−8.47)	−0.059 ***	(−8.35)
IFRS	0.000	(0.97)	0.000	(0.89)
Small_Firm	−0.002 ***	(−3.08)	−0.003 ***	(−3.70)
IFRS*Small_Firm	0.002 ***	(3.01)	0.002 ***	(3.50)
Controls	Included		Included
Industry	Included		Included
Adj. R-sq.	43.07%		40.97%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering. Controls are included but omitted for brevity.

. Small_Firm equals one if the SIZE is less than the median value (which is 18.833). The interaction terms (IFRS*Small_Firm) are not significant in Panel A, however, they are positively significant in Panel B. These results show that the association between IFRS and financial statement comparability is more pronounced in small firms than large firms. Thus, the results document weak evidence that the small firm sub-sample whose information environment is inferior to that of the large firm sub-sample demands and benefits more from the IFRS adoption.

Lastly, given that the sample period of the paper starts right after the worldwide economic downturn, the reported results are possibly the reflection of the market discipline mechanism, instead of the IFRS adoption effect. It can be directly controlled by including year indicators in the regression models. However, since the paper uses the IFRS indicator (which equals to one for from 2011 to 2016, and zero otherwise), it is correlated with year indicators causing a multicollinearity problem.

As an alternative, to overcome this concern, the study uses the unemployment rate as a proxy of the economic downturn. The unemployment rate is a conventional measure of external shocks such as financial crises and economic downturns (e.g., Reference [31]). It is collected from the Korean Statistical Information Service (kosis.kr). The distribution of unemployment rate by year is between 3.1% and 3.7%. It is lowest in 2013 and highest in 2010 and 2016. There is no systematic trend in the unemployment rate on yearly basis. This study includes the unemployment rate into the regression models to control the effect of market condition on the previous findings.

The results of testing hypotheses after including unemployment rate (UER) are shown in

Table 10. Controlling the external shock by including the unemployment rate.

Panel A. Earnings management and IFRS
Independent Variables	(1) Dependent Variable = DA		(2) Dependent Variable = AbsoluteDA
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.144 ***	(−4.00)	0.166 ***	(5.77)
IFRS	−0.007 ***	(−2.69)	−0.010 ***	(−3.89)
UER	0.019	(0.05)	1.143 ***	(2.71)
Controls	Included		Included
Industry	Included		Included
Adj. R–sq.	48.47%		15.05%
N. of obs.	7216		7216
Panel B. F/S comparability and IFRS
Independent Variables	(1) Dependent variable = Comp1		(2) Dependent variable = Comp2
	Coefficient	(T-stat)	Coefficient	(T-stat)
Intercept	−0.068 ***	(−10.67)	−0.075 ***	(−10.55)
IFRS	0.002 ***	(5.58)	0.002 ***	(5.62)
UER	0.302 ***	(5.82)	0.294 ***	(4.91)
Controls	Included		Included
Industry	Included		Included
Adj. R-sq.	43.20%		40.98%
N. of obs.	7216		7216

***, **, and * denote significance level at 1, 5, and 10% level. See Section 3.3 for variable definition. T-stats are corrected for firm-level clustering. Controls are included but omitted for brevity.

. In Panel A and B, the coefficients of IFRS remain negative and positive, respectively; all significant at the 1% level even after controlling for the unemployment rate. Collectively, this shows that the results of this paper are robust to external shock.

5. Conclusions

In the past two decades, the adoption of the IFRS has become an international trend. Consequently, 144 of 166 profiled jurisdictions now require the use of IFRS. Nevertheless, there is mixed evidence for the effect of IFRS on sustainable accounting information.

This study examines whether the adoption of IFRS improves earnings sustainability. Specifically, it tests the effect of IFRS on earnings quality by comparing earnings management and financial statement comparability of Korean firms between the pre- and post-IFRS periods. The reason for adopting IFRS in Korea was to overcome the so-called “Korea discount,” which refers to the phenomenon of undervaluing of stock prices of Korean firms by foreign investors because of their less transparent accounting information. Thus, it is valuable to test whether the adoption of IFRS improved the quality of financial statements in Korean listed firms.

The findings show that firms report less managed earnings in the post-IFRS period than in the pre-IFRS period. Furthermore, the results suggest an enhancement of financial statement comparability in the post-IFRS period compared to the pre-IFRS period. In particular, this paper documents that the effect of IFRS on sustainable accounting information is more pronounced in competitive industries. Moreover, it shows that small firms with inferior resources and information environment benefit more from adopting IFRS. Overall, this study finds that IFRS adoption in Korea improves the overall sustainability of accounting information.

This study contributes to the literature in several ways. First, it provides evidence supporting the positive consequences of the IFRS. With the continuing debate on the impact of IFRS on the quality of financial statements, the results of this study extend the prior findings. Second, the results should also be of interest to users of financial statements. For example, foreign investors and analysts from outside Korea may use the reported earnings of firms in Korea more efficiently, as the adoption of the IFRS enhances their sustainability. Lastly, the purpose of adopting IFRS in Korea was to improve the transparency of accounting information. The findings of this study provide sufficient evidence to support this primary purpose. Finally, future research can extend the findings of this study by investigating other proxies of sustainable accounting information and by examining other emerging markets where IFRS has been adopted.