FINANCIAL REPORTING QUALITY FOLLOWING THE CORPORATE GOVERNANCE REFORMS: A CONDITIONAL CONSERVATISM PERSPECTIVE

How to cite this paper: Ismail, W. A. W., Harymawan, I., Agustia, D., & Kamarudin, K. A. (2021). Financial reporting quality following the corporate governance reforms: A conditional conservatism perspective [Special issue]. Journal of Governance & Regulation, 10(2), 216–225. https://doi.org/10.22495/jgrv10i2siart3 Copyright © 2021 The Authors This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). https://creativecommons.org/licenses/by/


INTRODUCTION
Financial reporting quality is vital to users of financial statements and practitioners, regulators, and accounting researchers. High-quality accounting information, such as earnings, is also essential for firms to access equity and debt markets.
The informative function of earnings means that it is often used to describe the firm's financial performance. For example, the earnings numbers and various ratios or metrics have been widely used in compensation agreements and debt agreements. Analysts also use earnings to evaluate firms' previous and current performances and forecast firms' future ability to create additional wealth for shareholders.
The importance of financial reporting quality can also be explained from at least two perspectives, the contracting perspective and the investment perspective (Schipper & Vincent, 2003). From the contracting perspective, low quality of financial reports may result in unintentional wealth transfers. For instance, firms that reward managers based on profit may overcompensate the managers if the profit is overstated. From an investing perspective, poor quality of financial reports is problematic as it can mislead investors, resulting in a misallocation of resources (Myers, Myers, & Omer, 2003;Schipper & Vincent, 2003). High-quality financial reporting would also increase the attractiveness of stocks to outside investors and increase market liquidity (Young & Guenther, 2003), lower cost of debt ( This study aims to assess whether Malaysia's corporate governance reforms significantly affect the financial reporting quality, particularly conditional conservatism. Prior to the corporate governance reforms, Ball, Robin, and Wu (2003) documented no evidence of conditional conservatism in Malaysian listed firms' financial statements. Ball et al. (2003) concluded that although during the period of 1984-1996, Malaysian accounting standards were perceived as being of high quality as they were derived from the United Kingdom and based on the International Accounting Standards, weak institutional structures gave low incentives for preparers of financial statements to produce high-quality financial reports.
Our study is the extension of this empirical strand of research. After the 1997 Asian financial crisis, Malaysia's regulatory bodies have undertaken significant corporate governance reforms; thus, we assessed the financial reporting quality during the period following the reforms. We predicted that corporate governance reforms will lead to increased conditional conservatism. We conjectured that institutional structure is an essential element influencing the quality of financial reports. In this study, we offered both demand and supply arguments for the quality of financial reporting. The changes in Malaysia's regulatory environment, particularly corporate governance reforms, have allowed us to examine whether the period following the reforms has been associated with more conservative reporting.
Our sample comprised all Malaysian public companies for the ten years of 2002-2011. We chose this sample period because the Malaysian Code of Corporate Governance came into effect from the financial year ending June 30, 2001. We posit that corporate governance reforms have provided strong demand and incentives for high-quality reporting. From the analysis, we found robust evidence of conditional conservatism following the corporate governance reforms. The results show that economic losses are recognised in a more timely way than economic gains. Specifically, earnings are seven times more sensitive to negative stock returns than to positive stock returns. These results are also robust to various sensitivity tests.
This study contributes to the literature in the following ways. First, this study extends a study by Ball et al. (2003) by focusing on the period following corporate governance reforms. Our point of departure is that the institutional structure became much stronger following the various initiatives undertaken by the regulators in Malaysia, as immediate responses to the Asian financial crisis. These initiatives included: the Financial Reporting Act 1997; amendment of the Securities Law and Bursa Malaysia Listing Requirements; the introduction of the Code of Corporate Governance; and establishment of the Malaysian Accounting Standards Board (MASB), the Malaysian Institute of Corporate Governance (MICG), and the Minority Shareholders Watch Group (MSWG). Our motivation to examine conditional conservatism following the corporate reforms is to provide insight into the extent of financial reporting quality compared to the period before the reforms. Second, this study provides additional insight into conditional conservatism in an emerging market. By focusing on a single country, this study avoids several concerns relating to cross-country studies, such as the possibility of endogeneity amongst the variables at the country level, noise in the variables, and severe correlated omitted variable problems (Miller, 2004) Our study is closely similar to Li, Wu, Zhang, and Chand's (2018), which examined the effects of the two major changes in the accounting standards of 2001 and 2007 in China on accounting conservatism. Li et al. (2018) found that the 2001 accounting reforms positively impacted conservatism in earnings. They concluded that accounting reforms that restrict managers' ability to conduct earnings management and improve conservatism in earnings serve as a vital substitute mechanism for investor protection institutions. Whilst Li et al.'s (2018) study focused on the effect of changes in accounting practices, our study has investigated the effect of another type of institutional reform: corporate governance practices. This paper's remainder is organised as follows. The second section provides the literature review and hypothesis development. The third section discusses the research methodology, followed by Section 4 that discusses the findings. The discussion of the results is presented in Section 5. Finally, the main findings are summarized in Section 6.

Background of corporate governance reforms in Malaysia
In 1997, several countries in East Asia, including Malaysia, were hit by a severe financial crisis. The crisis originated from international currency speculation, which led to major slumps in exchange rates. It began as a currency crisis in Thailand after the country removed the pegged exchange rate between the Thai baht and the US dollar. The crisis turned into a financial and economic crisis. It spread rapidly to other East Asia countries when foreign investors lost confidence and started to pull out their investments, causing the countries' currencies to fall further. By January 1998, the Malaysian ringgit had depreciated by almost half its value from MYR2.42 in April 1997 to MYR4.88 to the US dollar, and interest rates climb to more than 12% per annum.
The vulnerability of the Malaysian economy to the Thai spill-over was due to several reasons. Some argued that the main contributing factor to the crisis was the large proportion of private sector debt (Suto, 2003). Other than the excessive use of unhedged short-term foreign debt, the vulnerability was caused by corporate governance weaknesses and poor investments. Poor corporate governance had resulted in the poor quality of corporate reporting by publicly listed companies. Ineffective boards of directors, weak internal controls, poor audits, lack of adequate disclosure, and lax legal enforcement are believed to have contributed to the crisis by shielding companies from market discipline and by failing to terminate unprofitable projects, recognise bad loans, restructure or liquidate poorly performing enterprises, and discipline banks (Zhuang, Edwards, Webb, & Capulong, 2000). The majority of Malaysian companies also have concentrated shareholdings, leading to potential problems in protecting minority shareholders (Haniffa & Hudaib, 2006). Furthermore, the lack of an active market for corporate control and little role for hostile takeovers to discipline managers have given managers less incentive to maximise shareholder value (OECD, 1999 Governance, 1998) in the United Kingdom. The code sets out the principles and best practices on corporate governance to improve the directors' monitoring function, audit committee, and external audit. This includes the essential criteria for the structure and operational process of the monitoring units, such as the composition of the board; procedures for recruiting new directors; remuneration of directors; and the use of board committees, their mandates, and their activities.
In January 2001, the MCCG came into full effect when an amendment was made to the Bursa Malaysia Listing Requirements. The revision requires all public listed firms to include in their annual report a statement of corporate governance, a statement of internal control, the composition of the board of directors, and the composition of and quorum for the audit committee. All listed firms must also include in their annual reports a narrative of the application of the principles and best practices set out in the MCCG. The reasons for areas of non-compliance and alternative practices that were adopted must be justified and disclosed. Besides that, directors of publicly listed firms must attend a directors' training programme, known as the mandatory accreditation programme, which includes topics such as the Companies Act 1965, the Listing Requirements of Bursa Malaysia, risk management and internal control, and relevant securities laws.
The efforts to improve corporate governance and protect investors and shareholders continued even after the issuance of the MCCG. Various initiatives to strengthen the existing mechanism were undertaken, including the issuance of the guideline Best Practice in Corporate Disclosure by Bursa Malaysia in August 2004. The guideline's objective is to ensure public listed firms' compliance with disclosure obligations under the listing requirements and securities law.
In October 2007, the MCCG was revised to strengthen boards of directors and audit committees' roles and responsibilities. The revision provides guidelines for the appointment of directors, the role of nominating committees, the eligibility criteria for appointment as an audit committee member, the audit committee composition, the frequency of meetings, and the need for continuous training. The 2007 Revised MCCG Code requires the audit committee to consist of at least three members with a majority of independent directors. Further, all audit committee members should be financially literate, with at least one member being a member of an accounting association or body. The revised code also enhances the nomination committee's role by requesting that when candidates are recommended for directorships, they should have the necessary skills, knowledge, expertise, experience, professionalism, and integrity to strengthen the board and ensure that the board discharges its roles and responsibilities effectively.
Recent revisions of the MCCG were made in 2012 and 2017. The revised code highlights the importance of auditors' independence and its potential impairment caused by non-audit services. The code suggests that the audit committee establishes policies governing the circumstances under which contracts for the non-audit services can be entered into and procedures that must be followed by the external auditors (Securities Commission, 2012, p. 19). Audit committees are also expected to obtain written assurance from the external auditors to confirm that the auditors are, and have been, independent throughout the conduct of the audit engagement following the terms of all relevant professional and regulatory requirements (Securities Commission, 2012, p. 19).

Hypothesis development
Empirical studies examining conditional conservatism in Malaysia are still sparse. The most significant study was carried out by Ball et al. (2003), who examined conditional conservatism in four East Asian countries (Malaysia, Singapore, Thailand, and Hong Kong). Using a sample of 2,726 firm-year observations from 1984 to 1996, they found that the four East Asian countries, including Malaysia, had low levels of conditional conservatism, hence a low quality of financial reporting. Ball et al. (2003) linked this result with the weak institutional structures of these countries, which appeared to provide little incentives to prepare high-quality financial reports, i.e., report earnings conservatively. Since these countries' accounting standards are perceived as of high quality 2 , the evidence suggests that incentives appear to dominate accounting standards as a determinant of financial reporting quality.
Another study, by Bushman and Piotroski (2006), also included Malaysian firms in its sample.
To examine the effect of country-level institutions, including a country's legal/judicial system, securities laws, and political economy, on conservative accounting practices, Bushman (2006) where their sample was only up to 2001. Notably, the results of this study will give a better insight as to whether the period following the corporate governance reforms has exhibited the presence of conditional conservatism. This study predicts that conditional conservatism is a pervasive feature of Malaysian financial reporting following corporate governance reforms. Several reasons are supporting this prediction.
First, after the 1997 Asian financial crisis, there were major reforms in the institutional structure governing Malaysian firms, including matters relating to corporate governance, securities laws, and financial reporting. Amongst the initiatives were Secondly, the period examined in this study is associated with the high demand for conservative reporting due to several high-profile corporate scandals, which include the alleged failure of auditors to detect and persuade their clients to make timely recognition of economic losses. The high-profile corporate scandals in US firms, such as Enron and WorldCom, eroded investors' confidence in the securities markets and caused a credibility crisis in the accounting profession (Wall Street Journal, 2002). In the Malaysian context, the failure of local firms, such as Perwaja Steel, Technology Resources Industries, and Malaysian Airlines System, had also eroded the public's confidence in financial statements and thus created a strong demand for reporting that was more conservative.
From the above arguments, it is predicted that the period following the corporate governance reforms will exhibit conditional conservatism, as shown by the timelier recognition of bad news as compared to good news. We conjecture that institutional structure reforms, i.e., corporate governance, create strong incentives for corporate managers, account preparers, auditors, and investors to favour reporting earnings conservatively. The hypothesis is stated as follows: H1: Ceteris paribus, earnings reported after the corporate governance reforms have incorporated bad news (negative returns) faster than good news (positive returns).

Measurement of conditional conservatism
Following the research design of Basu (1997), we used the asymmetric timeliness of the earnings model as a primary measure for conditional conservatism. This measure has been widely used in many studies, for example, Ball et al. (2003), Ball and Shivakumar (2005), Bushman and Piotroski (2006), Francis, LaFond, Olsson, and Schipper (2004), and Roychowdhury and Watts (2007). In this model, conservatism is defined as the extent to which the current period accounting earnings asymmetrically incorporate economic losses relative to economic gain. Stock return is used as a proxy for economic income as it is assumed that stock prices reflect all information from various sources other than financial statements, consistent with evidence that the stock prices lead to earnings information (Ball & Brown, 1968; Beaver, Lambert, & Morse, 1980; Kothari & Sloan, 1992). In contrast, accounting earnings impose different verification standards for the recognition of different types of economic news. For bad news (negative stock returns), lower verification standards are used, which results in the immediate recognition of losses. However, for good news (positive stock returns), higher verification standards are imposed for gains to be recognised in accounting earnings (Basu, 1997;Ball et al., 2003). This approach results in a stronger positive association between bad news and earnings, suggesting a timelier reporting of bad news relative to good news.
To conduct the tests, we estimate the accounting earnings on stock returns, as shown in equation (1) below: where, E it is the earnings per share for firm i in fiscal year t; P it-1 is the price per share at the beginning of the fiscal year; RET it is the annual return of firm i in fiscal year t; RD it is a dummy variable equal to one if RET it is negative, and zero otherwise, and ε it is the error term. In equation (1), β 1 measures the response of earnings to returns when returns are positive, and β 1 + β 3 measures the response of earnings to negative returns. If β 1 + β 3 > β 1 or β 3 > 0, then earnings reflect bad news more quickly than they reflect good news or conditional conservatism.

Descriptive analysis
Panel A of Table 1 Table 1 presents the comparison of means between two groups in the sample. The groups are companies with positive stock returns (good news) and companies with negative stock returns (bad news). We found that the good news sample had a significantly higher earnings yield, higher stock returns, bigger firm size, higher market-to-book ratio, and lower leverage than the bad news samples. A correlation analysis was performed for testing the relationships amongst the variables used in this study. Table 2 provides Pearson and Spearman correlations amongst all the variables. In general, the analysis shows no problem of collinearity between the variables. As reported in Table 3, earnings yield (E/P) and stock returns (RET) are positively correlated at the 1% level (Pearson = 0.203; Spearman = 0.414), which is consistent with the previous evidence on the earnings-return relationship. Other variables also show significant correlation with other variables, but no serious multi-collinearity exists amongst the independent variables since none exceeds 0.7 (Pallant, 2007).

Main analysis: Conditional conservatism in Malaysia
To provide greater insight into the differences in timeliness of earnings between good news and bad news samples, we partitioned the sample into good news and bad news subsamples. The results reported in Column 1 and Column 2 of Table 3 show that the coefficients for RET in both good news and bad news subsamples are positive and significant at a 1% level. These results show evidence of the timeliness of earnings in response to economic news. The coefficient for RET in the good news sample is 0.036, which is lower than the coefficient for RET in the bad news sample, which is 0.255. These results indicate higher timeliness of earnings in recognising bad news as compared to good news.
We then estimated equation (1) to test for a significant difference in the asymmetric timeliness of earnings between the two subsamples. The results are summarised in Column 3. We also report the regression estimates of Ball et al. (2003) in Column 4 for comparison purposes.  (White, 1980). Asterisks denote statistical significance at the 1% (*), 5% (**), or 10% (***) level, respectively. Variable definitions: E it is the earnings per share for firm i in fiscal year t; P it-1 is the price per share at the beginning of the fiscal year; RET it is the annual return of firm i in fiscal year t; RD it is a dummy variable that equals one if RET it is negative, and zero otherwise, and ε it is the error term.
This result differs from the result documented by Ball et al. (2003) from 768 firm-year observations from 1984 to 1996. They found that the coefficient for RET * RD was 0.00 and insignificant (p > 0.10), showing no evidence for conditional conservatism or no significant difference in the timeliness of earnings between good news and bad news. In an environment of a weak institutional structure, Ball et al. (2003) suggested that preparers of financial reports have less incentive to report high-quality earnings, even though the accounting standards are of high quality. Complementing Ball et al. (2003), this study shows evidence of conditional conservatism in the period following the corporate governance reforms, implying that managers and preparers of financial reports have high incentives to report earnings conservatively in order to meet higher market demand for quality accounting information. In short, our results suggest that strong institutional factors give greater incentives for managers to report earnings conservatively.

Sensitivity analysis
To check the robustness of the main results, we conducted several specification tests. The results are reported in Table 4. First, we estimated the Fama and MacBeth (1973) annual regression to cope with potential cross-sectional dependence problems, particularly since the data are concentrated in a small number of years 4 . Estimating the regression separately for each year partially controls this problem. The coefficients for the parameters are obtained as the simple average from the crosssectional regression. The t-statistics are the ratios of the mean estimated coefficients to the standard deviation of the distribution of the annual estimated slope coefficients, divided by the square root of the number of years. The regression estimates are summarised in Table 4 and show that the coefficient for RET * RD is positive (0.326) and significant at the 1% level, implying that bad news is recognised significantly faster than good news.
Second, we controlled for serial correlation problems of residuals of the panel data by incorporating industry and year dummies in the regression model. Using the four-digit Global Industry Classification Standard (GICS), we created seventeen industry dummies. For year variables, we created nine dummy variables from 2003 to 2011, whilst the year 2002 was used as a base variable. The result reported in Table 4 shows that conditional conservatism is a pervasive feature of Malaysian financial reporting even after controlling for industry and year effects. The coefficient value for RET * RD after controlling both industry and year effects is 0.255, respectively, which is significant at the 1% level. These results imply robust evidence of conditional conservatism in Malaysia in the period following the corporate governance reforms.
Third, we controlled for firm-specific variables, such as firm size (SIZE), firm leverage (LEV), and growth (MTB). According to Khan and Watts (2009), these variables could significantly influence conditional conservatism. Following LaFond and Watts (2008), we included the interaction variables of SIZE, LEV, and MTB with RET, RD, and RET * RD in equation (1). The regression estimates reported in Table 4 show qualitatively similar results to those reported in the main analysis. The coefficients for RET * RD are positive (0.473) and significant (p < 0.01), implying a more timely recognition of economic losses than economic gains in Malaysia even after controlling the effect of SIZE, LEV, and MTB. For the control variables, all the control variables have a significant influence on conditional conservatism. Firm size (SIZE) has a negative relationship with conditional conservatism, supporting the argument that large firms have lower information asymmetry, thus reducing the contracting demands for conservatism (Khan & Watts, 2009). For firm leverage (LEV), the results show a positive relationship between LEV and conditional conservatism, suggesting that highleverage firms have higher agency costs and higher financial distress costs, creating greater demand for conservatism to alleviate the problems. For growth (MTB), the results show a strong negative relationship between MTB and conditional conservatism. This is consistent with the explanation by Roychowdhury and Watts (2007) that the short horizon MTB is negatively correlated with conservatism flows due to prior unrecognised increases in asset values reducing the necessity to recognise asset value losses. Fourth, a study by Morck, Yeung, and Yu (2000) reports that the systematic component of returns variation is large in emerging markets as compared to the United States and other developed countries. In Malaysia and many developing countries, the stock returns variation appeared unrelated to the fundamentals' co-movement, whilst in the United States, high firm-specific returns variation was documented (Li, Morck, Yang, & Yeung, 2004;Morck et al., 2000). To control this problem, timeseries non-stationarity in the returns processes, we employed market-adjusted returns as an alternative proxy for economic news. Also, the use of marketadjusted returns provides a useful robustness check because it can be argued that market-adjusted returns may provide a more reliable indicator for good versus bad news (Pope & Walker, 1999) 5 . The results reported in Table 4 show qualitatively similar results to those reported in the main results, in which conditional conservatism is a pervasive feature of Malaysian financial reporting. The coefficients for RET * RD are positive and significant (p < 0.01) in all the models, implying a more timely recognition of economic losses than economic gains in Malaysia.
Fifth, we used inter-announcement period stock returns as an alternative measure for economic news. This variable is calculated based on the period of three months after the end of the previous fiscal year to three months after the current fiscal year. We re-estimated equation (1), and the results are reported in Table 4. The results show that the coefficients for RET * RD are positive and significant at the 1% level, implying that bad news is recognised in a more timely fashion than good news.
Sixth, we reduced the sample to firms with stock prices higher than MYR1.00. This procedure was undertaken to avoid creating artificial scale problems that could even lead to the nonlinearity of the model (Kothari & Zimmerman, 1995). Smaller firms have, on average, less liquid stocks, greater information asymmetry, and higher idiosyncratic uncertainty (Khan & Watts, 2009) 6 . As a result of excluding all observations with a stock price less than MYR1.00, the number of observations was reduced to 4,532 firm-year observations. Consistent with the earlier results, the regression estimates in Table 4 show robust evidence for conditional conservatism in Malaysia. The coefficients for RET * RD, which measures conditional conservatism, are positive and significant at the 1% level, hence providing evidence of conditional conservatism in Malaysia following the corporate governance reforms.  (White, 1980). Asterisks denote statistical significance at the 1% (*), 5% (**), or 10% (***) level, respectively. Variable definitions: E it is the earnings per share for firm i in fiscal year t; P it-1 is the price per share at the beginning of the fiscal year; RET it is the annual return of firm i in fiscal year t; RD it is a dummy variable that equals 1 if RET it is negative, and 0 otherwise; and ε it is the error term.

DISCUSSION OF THE RESULTS
We hypothesised that conditional conservatism is pervasive in Malaysia following the corporate governance reforms because of the high demand and incentives to produce high-quality reporting of earnings. Previously, Ball et al. (2003) found no evidence of conditional conservatism in a sample from 1984 to 1996 and concluded that Malaysian financial reporting was of low quality. This study argues that corporate governance reforms have provided strong demand and incentives for highquality reporting.
The results show that there is robust evidence of conditional conservatism following corporate governance reforms. Economic losses are recognised in a more timely way than economic gains. Specifically, earnings are seven times more sensitive to negative stock returns than to positive stock returns. These results are also robust to various sensitivity tests, including 1) estimation of Fama-Macbeth t-statistics; 2) estimation controlling for industry and year effects; 3) estimation controlling for firm size, leverage, and market to book ratio; 4) controlling the effect of market return; 5) employing an inter-announcement stock return as a proxy for economic news, and 6) reducing the observation to firms with a share price greater than MYR1.00. All the results support the prediction concerning bad news being recognised in a more timely way than good news; hence, the results provide enough evidence to reject the first null hypothesis.
This finding supports the arguments made by Bushman and Piotroski (2006) and Ball et al. (2003) that country-specific factors and preparer's incentives have a significant influence on the financial reporting practice. Evidence of conditional conservatism in the post-reform period shows that firms have greater incentives and demands to report high-quality earnings, in particular by more timely recognition of bad news compared to good news. This is contrary to the period before the economic crisis, when the weak institutional structures resulted in low incentives for preparers to report earnings conservatively (Ball et al., 2003). In short, the results support the prediction that managers and preparers of financial reports have had higher incentives to report earnings conservatively in order to meet higher market demand for quality accounting information following the corporate governance reforms.

CONCLUSION
Our study demonstrates that the institutional reforms have generally led to conservative financial reporting in Malaysian companies. The results imply that policymakers and regulators in a particular country could put more efforts into providing infrastructures to support the reforms of the institutional environment, not only in terms of strengthening the corporate governance of corporations but also in reinforcing a strong economic and legal system to achieve a better quality of financial reporting practices. The results of this study are also useful for investors when assessing the performance of companies and making investment decisions, not only in Malaysia but also in other similar emerging countries. Practically, investors should put extra caution when investing in companies from countries where the institutional factors are weak, compared to those where a strong institutional environment is in place.
Our results should be interpreted with several caveats. Although we have employed a large dataset of the sample of 6,819 firm-year observations of Malaysian listed companies from 2002 to 2011, the dataset is still restricted by the availability of the data. Also, we mainly focused on the Malaysian sample, where future studies can explore the effect of corporate governance reforms in other emerging markets. Going forward, we encourage more research on the specific regulatory and policy efforts as well as firm-level reforms. Other dimensions, such as reforms in board composition, political connection, culture, and expertise, are other potential research avenues.