χ t is earnings at time t,
Chapter 4 Earnings Persistence
4.2 Hypothesis Development
In this chapter, I define accounting quality as earnings persistence. This is consistent with a large number of prior studies (i.e. Sloan 1996; Francis et al. 2004; Hanlon 2005). Earnings persistence in this study is a predictability test of earnings information. Accounting information is a proxy to represent unobservable underlying economic constructs. Information users demand economic constructs for their decision making. Maines and Wahlen (2006) suggest that accounting information relevance stimulates measurement and reporting reliability. However, they further argue that information reliability is an essential but not adequate characteristic for the efficient use of applicative
Chapter IV Earnings Persistence
information. Because of the different perspective of reliability and relevancy, it is controversial to differentiate as to whether reliability or relevancy is more important. Prior studies have extensively evidenced accounting quality in these two areas (among others, Sloan 1996; Francis and Schipper 1999; Lev and Zarowin 1999; Francis et al. 2004; Richardson et al. 2005; Schmidt 2006). For earnings persistence, prior studies find current earnings can explain future earnings and accruals information is more persistent than cash flow information (i.e. Sloan 1996; Hanlon 2005). Richardson, Sloan, Soliman and Tuna (2005) highlight that less reliable accrual estimates result in lower earnings’ persistence. Schmidt (2006) finds that current pre-tax earnings can explain future earnings. Barth et al. (2008) argue whether IFRS brings a higher quality of earnings than domestic accounting standards due to country-specific factors.
This thesis employs earnings persistence to compare accounting quality between pre- and post-adoption periods based on three reasons. Firstly, the persistence of earnings is a statistical property of earnings (Dechow, Ge and Schrand 2009). The persistence of earnings is the study of the prediction ability of earnings information. The predictability of earnings is an important input for the estimation of usefulness of accounting information (value relevance), suggesting that the higher persistence of earnings induces the higher earnings informativeness. A consequence of earnings persistence is expected to reflect on firms’ share prices. Earnings can be decomposed to accrual and cash flows. According to the principle-based accounting system, firms are required to provide their forecasting information, i.e. growth and future of their business. In addition, future cash flows must be predicted and performed (i.e. through asset valuations). The earnings persistence is one method that allows outsiders (i.e. analysts) to determine this classified forecasting information. For example, Li (2008) finds that firms with readable financial statements have more persistence in earnings. It implies that firms with more volatile in accounting information are less likely to provide useful guidance to users (Dechow et al. 2009). In terms of the Thai accounting system, the old Thai accounting standard does not require firms to
Chapter IV Earnings Persistence
cash flows according to new accounting starndards, it is more likely to observe the persistence of earnings reflected on financial statements. The earnings persistence facilitates the usefulness of accounting incormaiton (value relevance). The usefulness of accounting information is reflected on share prices, representing firm economic values. Because the share price is the market’s expectation on firms’ net cash flows, relative to the accounting information prepared by the old accounting standard, the accounting information prepared by the new accounting standards should correspond to more true firm performance that is more useful for decision making. Secondly, Dechow and Schrand (2004) suggest that the corporate governance system plays a significant role when the principle-based accounting system is employed. In terms of the Thai accounting system, when the new accounting standard is in effective, it will force the corporate governance mechanism to have more role and involvement in financial statements preparation. This mechanism, hence, should induce more predictability of accounting information. Thirdly, ones could contend that the earnings in Thai firms based on old accounting standards are more persistent because the firms are not required to forecast future information and they employ historical cost method. This thesis argues that although Thai firms are not required to provide expected future cash flows or they employ historical cost method, accounting information can be less persistent. It is due to the fact that Thai firms can employ different accounting methods based only on their discretions. For example, they can use different valuation methods for the same assets without concerning about the asset’s ability to generate cash flows that represents firm performances or use different accounting methods to smooth earnings the the smoothed earnings have low quality (Dechow et al. 2009).
In addition to the above reasons, the persistence of earnings permits researchers to calculate the long run equilibrium earnings of the firm or the permanent earnings (Mueller 1990). The persistence estimate signifies the adjustment of excess earnings to the equilibrium value of earnings. The lower persistence indicates the faster adjustment whiles the higher persistence indicates the gradually adjustment. The calculation of earnings is related to the measurement method specified by
Chapter IV Earnings Persistence
accounting standards (Dechow et al 2009). Relative to the old accounting standard, the earnings calculated by the new accounting standard, the IFRS-based system, are more reliable as claimed by the Thai accounting standard setter. Therefore, the current earnings calculated by the new accounting standard are expected to be close to the long run equilibrium earnings comparing with the current earnings calculated by the old accounting standard. As a result, the earnings calculated by the new accounting standard relative to the old accounting standard are of higher persistence, slow adjustment to the equilibrium value.40 I have a prediction that earnings quality is enhanced after the IFRS adoption in Thailand. As a result, I propose the first hypothesis, stated in alternative forms:
H.4.1: Earnings are of higher persistence after the adoption of IFRS in Thailand.
Following Guenther et al. (1997), Mills (1998) and Hanlon (2005), book-tax differences (BT) reflect the management’s incentive but it is difficult for firms to increase book income without increasing taxable income or increase tax deductions without increasing book expenses. This thesis considers that the cross- section variation in using a tax strategy is lower than the cross-section variation in managing accruals because tax rules are more restrictive than accounting standards. Based on Hanlon (2005), earnings are of lower persistence for firms with large book-tax differences relative to firms with small book-tax differences. In this study, I trace book-tax differences over 5-years before and 5-years after adopting the new accounting standards and partition firms to two groups. Group 1 includes firms whose book-tax differences during post-adoption periods are larger than those during pre-adoption periods (BT – Pre-adoption < BT – Post-adoption). Group 2 includes firms whose book-tax differences during post-adoption periods are smaller than those during pre-adoption periods (BT – Pre-adoption > BT – Post-adoption). I use the two groupings of firms to track overall enhancement of earnings quality. In addition, the book-tax difference in this analysis is used as a
40
Chapter IV Earnings Persistence
proxy to differentiate firms’ incentives of using accruals and tax activities. If we hold tax activity constant, the larger book-tax difference reflects over-reported earnings, resulting in lower accounting quality. If we hold earnings management unchanged, the larger book-tax difference reflects high tax planning, resulting in lower accounting quality (Ayers et al. 2009). On the other hand, if we hold tax activities constant, the large book-tax difference reflects high earnings manipulation, resulting in low accounting quality.
Generally speaking, although the sample firms are partitioned according to book- tax differences, the enhancement of earnings persistence should be observed for all firms after IFRS adoptions in Thailand. However, there are two potential explanations about the larger book-tax difference. Firstly, the firm might engage in aggressive tax activities and use accruals for earnings manipulations. Secondly, Dechow and Schrand (2004) suggest that corporate governance is essential when using IFRS, principles-based accounting system. Thus, the corporate governance for firms in group 1 might be weak and probably unable to detect the use of accruals for earnings manipulations, especially after IFRS adoptions. It is possible to anticipate the improvement of earnings persistence for firms in group 2 because the smaller book-tax difference reflects lower manipulation in accounting numbers (Hanlon 2005). Therefore, following the prior study (Hanlon 2005), the expectation is that earnings during pre-adoption periods are of higher persistence than earnings during post-adoption periods for firms with larger book-tax differences (group 1), whereas earnings during pre- adoption periods are of lower persistence than earnings during post-adoption periods for firms with smaller book-tax differences (group 2). However, this anticipation must be interpreted with caution because firms could manage their figures to obtain small book-tax differences. In addition, Hanlon (2005) suggests that it is not clear that a large book-tax difference indicates low earnings persistence. I propose the second hypothesis as follows, stated in the alternative form:
Chapter IV Earnings Persistence
H.4.2: When comparing between pre- and post-adoption periods, the earnings persistence is related to the book-tax difference.
A number of research studies have investigated the association between corporate governance structure and earnings informativeness. For example, Vafeas (2000) concludes that earnings-return relation is found in firms with 5-9 board members. Bushman et al. (2004) conclude that the relationship between earnings timeliness and ownership concentration, equity benefits to executives and reputation of outside directors varies inversely. Fan and Wong (2002) examine ownership structure and earnings informativeness in East Asia and find that low earnings information value-relevance is found in concentrated ownership. Defond et al. (2005) investigate the market reaction after appointing new financial experts as audit committee members and find that the market reacts positively to accounting financial experts but not to general financial experts. Xie et al. (2003) report that the characteristics of audit committee and audit committee activity discourage the management to be involved in earnings management. Chhaochharia and Grinstein (2007) find a significant impact of new corporate governance rules, Sarbanes-Oxley Act, on firm values. Gompers, Ishii and Metrick (2003) report that higher firm value is observed in firms with stronger shareholder rights. Those studies suggest that corporate governance plays an important role in accounting information quality. As stated by Dechow and Schrand (2004, pg.114), “Clearly, the quality of corporate governance and monitoring of financial reporting will have a significant impact on whether principles-based standards result in higher- quality earnings.” Building from prior studies, I perform the analysis to determine whether there is a relationship between accounting quality and the firm governance system in Thailand. If accounting quality relates to firm governance, I expect that accounting quality should be varied with the firm governance system. I, thus, hypothesise in an alternative form as follows:
H.4.3: Earnings persistence varies according to the firm governance system in Thailand.
Chapter IV Earnings Persistence