ارتباط ارزش اطلاعات حسابداری محافظه کار و غیر محافظه کار

The present paper examines effects of reporting conservatism on the value relevance of accounting earnings of a sample of Greek firms over the period from 1989 to 2003. The results of the paper indicate that conservatism is a salient feature of the Greek Accounting System. Moreover, the results depict that the level of conservatism has increased after the market crisis of 1999, potentially as a result of the additional regulation, imposed by the market authorities during the post-crisis period. Finally, the results show that there is a non-linear association between conservative reporting and value relevance of earnings. In particular, value relevance increases when moving from low-conservative firms to medium-conservative firms and decreases when moving further to high-conservative firms. Overall, the results of the paper lend empirical support to the theoretical underpinnings of Watts (2003a) who, on the one hand, report a number of arguments in favor of conservatism but, on the other hand, questions the practice of excessive conservative reporting as being a potential cause of the distortion of the earnings-returns relation.

The interest in the association between the value relevance of accounting earnings and the degree of reporting conservatism has increased in line with two distinct but not unrelated research paths: (a) the research on the value relevance of accounting earnings and earnings yields; and (b) the research on the level of reporting conservatism and its prevailing effects. In particular, the first path of research has revealed, among other things, that over the past forty years the information content of accounting earnings (or yields) for stock prices (or returns) exhibits a declining trend pattern (Collins, et al., 1997 and Lev and Zarowin, 1999). The second path of research has revealed that, over the same period of time, reporting in the United States and several other countries displays a gradual increase in the level of conservatism (Givoly and Hayn, 2000 and Holthausen and Watts, 2001 for the U.S.; Grambovas, Giner, & Christodoulou, 2006 for EU countries; Bushman & Piotroski, 2006 for several countries all over the world). Although an observable implication of these results is that conservative reporting is a potential cause for the decline in the value relevance of earnings, the existing research on a direct link between conservatism and value relevance is scarce and offers contradictory results. For example, Balachandran and Mohanram (2006) observe that the information content of conservatively reported earnings for stock returns exceeds that of non-conservatively reported earnings and conclude that reporting conservatism is unlikely to be the primary cause for the declining value relevance of accounting earnings. On the other hand, Watts (2003b) lends conceptual support to such findings in the area, by arguing that conservatism in accounting alleviates potential measurement errors which cause lack of reliability of accounting information. The lack of reliability limits the usefulness of decisions made by investors and therefore the effect of conservatism is to increase rather than to decrease the value relevance of earnings. The present paper uses data from an unbalanced-panel sample of Greek firms listed on the Athens Stock Exchange (ASE) over the period from 1989 to 2003 and expands on prior research in a threefold manner. First, in contrast to prior studies in the area, the present paper uses an improved methodological framework, which accommodates the effects of cross-sectional dependence in the data and attempts to provide an unbiased answer to the question of whether the information content of earnings for stock returns in Greece has declined over time. Second, under a similar methodological framework, the paper gauges the level of reporting conservatism in Greece over the period 1989 to 2003. Finally, it combines the results of the aforementioned research questions and sheds light on the effects of conservatism on the value relevance of accounting information over the same period. For the purposes of the present paper, conservatism is defined as the asymmetric way in which good and bad news are recognized in earnings (Basu, 1997 and Pope and Walker, 1999) and is quantified using a measure of conservatism that stems from the Basu (1997) model of conservatism. On the other hand, the value relevance of accounting earnings is defined as the ability of earnings to explain market returns (Francis & Schipper, 1999) and is assessed using the two methods proposed by these authors; (a) the magnitude of the adjusted R2 of the Easton and Harris (1991) model; and (b) the magnitude and sign of the returns of a hedge portfolio, which is formed by ranking sample firms according to their conservatism measure and then by taking a long position on the top 30% and a short position on the lower 30% of the ranked sample firms. In a final illustration, the present paper investigates whether new regulation that induces conservatism in financial reporting practices has a bearing on the results of the present paper. While the Greek Accounting System is considered to be conservative by nature (Ballas, 1994), after the market crisis of 1999, the Hellenic Capital Market Committee (HCMC henceforth) has introduced new laws on corporate governance and financial reporting, which provide for strict penalties to the non-obeying firms. Therefore, a reasonable assumption that can be made is that after fiscal year 1999 Greek firms have adopted more conservative reporting practices in order to avoid litigation and penalties from the HCMC1. This assumption is tested in the paper in two ways. First, by partitioning the research horizon into two sub-periods, one ending in year 1999 and the other starting from year 2000. Second, by performing stepwise regressions to test for the incremental effect of every year on reporting conservatism. The results of the paper indicate that over the period examined, the information content of accounting earnings for stock returns in Greece displays, on average, a decreasing pattern. On the other hand, the level of reporting conservatism displays no obvious pattern, although it appears that after the market crisis of 1999, Greek firms became more conservative. This result is consistent with the assumption that, in post-crisis periods, the increased legislation and the fear of litigation forces firms to more conservative accounting practices i.e., immediate write-off of all security losses. Finally, the paper classifies sample firms into three portfolios according to their level of reporting conservatism (low, medium and high conservatism portfolios) and reports a non-linear association between the value relevance of earnings and the degree of reporting conservatism. In particular, it is found that the value relevance of earnings increases when moving from a portfolio of low-conservatism firms to a portfolio of medium-conservatism firms, but reverts and decreases when moving further to a portfolio of high conservatism firms. This association is found to be significant for each one of the two sub-periods and for both sub-periods taken together. Moreover, this association is justified both by the values of the adjusted R2 of the Easton and Harris model and by the returns of the hedge portfolio. Interestingly, however, firms failing in the high-conservatism portfolio appear to deliver slightly higher mean returns than firms included in the low conservatism portfolio and to offer better protection to investors in the post-market crash period. The take away from these findings is that the benefits of conservative reporting are not unlimited. Extreme conservatism distorts rather than enhances the value relevance of accounting information. However, it appears that in certain cases, such as a persisting market crisis, increased levels of conservatism emerge as a natural mechanism for investor protection. The results of the present paper are potentially interesting for researchers, regulators and investors for a number of reasons. First, the present study consists of the pioneering attempt to examine the relation between the value relevance of earnings and reporting conservatism in Greece. Some previous studies by Bushman and Piotroski (2006) and Grambovas et al. (2006) assess the level of conservatism in Greece but focus their research interests on cross-country comparisons without considering any value relevance effects of conservatism. Also, their research setting does not allow them to consider country specific circumstances that may have affected the level of conservatism, such as the 1999 crisis of the Greek Capital market. Moreover, the present study provides the first out-of-sample evidence on the issue. Previous research, i.e., Balachandran and Mohanram (2006), usually employs samples of U.S. firms which operate in a highly developed, common law institutional framework. The results of the present paper come from a less-developed, code-law institutional framework, which shares many commonalities with the institutional framework of other Balkan countries. In this context, the results of the present paper could be generalized to apply for many Balkan countries and could be interesting for regulators and for investors who wish to place investments in the Balkans. Finally, in terms of methodological issues, the present paper advances application of research tools by using Pesaran's (2006) Pooled Common Correlated Effects (CCE henceforth) estimator to control for two problems arising in panel-estimation with a large number of members but short time-span of data: cross-sectional dependence and heterogeneity. In so doing, it offers two modified versions of the Easton and Harris (1991) and of the Basu (1997) models which can be used in small sample panel settings. The remainder of the paper is organized as follows: Section 2 expands on methodological issues and defines the hypotheses to be tested. Section 3 describes the data and offers some descriptive statistics. Section 4 presents and analyzes the empirical results, and finally, section 5 concludes the paper and offers implications for further research.

The present paper examines the relationship between conservatism in financial reporting and value relevance of accounting earnings in Greece. The sample used consists of an unbalanced panel sample of commercial and industrial Greek firms listed on the Athens Stock Exchange over the period 1989–2003. The models of Basu (1997) and of Easton and Harris (1991) are used to assess the level of conservatism in financial reporting and the value relevance of earnings respectively. Because the panel sample used in the paper consists of a large number of firms over a short period of time, both models are adjusted by the Pesaran (2006) estimator to mitigate the effects of cross-sectional dependence and heterogeneity. Sample firms are classified into three portfolios according to their level of conservatism and the horizon of the study is divided into two sub-periods setting as the point of split the market crash of year 1999. The results reveal that conservatism is a salient feature of the Greek financial reporting system. Moreover, in the post-crash period, the increased level of conservatism in earnings implies that regulation and perceived fear of litigation has lead management to adopt more conservative practices. However, this finding is contingent upon estimation that mitigates the effects of cross-sectional dependence and heterogeneity. The results clearly indicate an increase in conservatism in the post-crash period. In addition the results suggest a non-linear relationship between conservatism and value relevance. Both in the periods preceding and following the market crash of 1999, a portfolio of medium conservative firms is found to be more value relevant (i.e. higher R2 values) than either portfolio of low or high conservatism portfolio in the post crash period. The observed increase in conservatism is accompanied by a decrease in the explanatory power of accounting information for the returns of all three conservatism portfolios. However, the results continue to display a non-linear relationship between conservatism and value relevance. Therefore, the results of the paper may not be construed as providing a causal connection between the value relevance of earnings and the level of reporting conservatism.