تفاوت در ارزش گذاری درآمد و ارزش دفتری : اثرات مقررات و یا اثرات صنعت؟

This paper uses a valuation framework on a sample of firms from four European countries (France, Germany, Netherlands, and United Kingdom) to examine how income, accruals, and book value of equity are perceived by the respective capital markets. Our model includes adjustments for industry effects and taking into account the linear information dynamics of the accounting variables posited in the Ohlson model. Consistent with previous researchers, we find that both earnings and book value of equity have valuation implications and that there is significant dispersion in the country-specific and industry-specific valuation multiples. However, when using accounting variables to forecast market values we find that industry-specific valuation multiples reduce forecasting error more than country-specific ones.

In 1980, F. Black issued a provocative statement: accounting policies must be chosen so that earnings are useful in the valuation of a firm and, more specifically, that the earnings–price multiple should be constant across firms. It is well known however, that there is substantial variation in the average earnings–price multiple across countries although Land and Lang (2002) found that earnings multiples across countries became more similar over the years. In the present study, we investigate whether the systematic differences in the value relevance of the book value of equity and earnings found among European countries and reported in King and Langli (1998) and Arce and Mora (2002) may also be explained by economic factors such as industry effects. We go even further and ask whether one should regulate accounting at the national level (or, even the supra-national one, e.g., I.F.R.S.) or at the industry level by adopting industry-specific standards as the FASB or even European country regulators have done on a number of occasions in the past. We focus on four European Union (EU) countries. The EU offers an interesting context for such studies because a series of Directives were enacted in the 1980s to ensure that member states would harmonize their accounting policies. Still, generally accepted accounting principles vary substantially across countries. Consequently, a number of recent publications (inter alia, Arce & Mora, 2002 and King & Langli, 1998) have examined the valuation implications of these differences. The modeling approach used in this paper is based on Ohlson (1995) and Feltham & Ohlson, 1995 and Feltham & Ohlson, 1996. This class of models “can best be understood as an attempt to restate economic theories of income measurement in the light of advances in the economics of asset pricing under uncertainty” (Walker, 1997, p. 341). The fundamental characteristics of Ohlson type models are characterized by the clean surplus assumption and linear information dynamics (Walker, 1997). Indeed, estimating variants of the Ohlson model imposes three types of constraints on the estimated equation: “The first constraint is that all components of earnings have the same earnings forecasting coefficient and valuation multiple. The second constraint adds the additional restriction that earnings valuation multiples are identical across industries. The third constraint is that the theoretical structural relation underlying the valuation model is appropriate.” (Barth, Beaver, Hand, & Lansdman, 2002, p. 1). The purpose of this study is to provide evidence on the influence of these constraints in a cross-national context. Several studies provide evidence that the first constraint is binding with regard to the cash flow and accrual components of income (e.g., Barth et al., 1999, Dechow, 1994 and Sloan, 1996). In particular, with relevance to the current study, Barth et al., 1999 and Barth et al., 2002 show that total accruals and cash flow earnings components have different implications for forecasting abnormal earnings and for estimating equity market value in the context of Ohlson (1999). That is, disaggregating earnings into its accrual and cash flow components aids in forecasting abnormal earnings and explaining equity market value. We argue that different institutional environments for accounting regulation are likely to influence the accruals component only. Therefore, any systematic differences in accounting policies will be reflected in the valuation of accruals and not cash flows because cash flows should be valued at approximately the same rate based on economic conditions. Barth et al., 1999 and Barth et al., 2002 also provide evidence that the valuation of the accrual and cash flow components of earnings vary across industries. In the case where the researcher aims to identify differences in the valuation of earnings and book value across countries, constraining the valuation coefficients of earnings components (and book value) to be the same is even more important because economic conditions facing European countries are more likely to be reflected at the industry level rather than the macro level. Finally, the third constraint regarding the structure of the model implies that the linear information model (LIM) describing the time-series evolution of earnings and other variables is descriptively valid. As Myers (1999, p. 26) observes, “linear models of the link between current and future information ensure consistency and are an integral component of accounting based valuation.” Therefore, ignoring the LIM in the estimation of the valuation equation is likely to lead to biased estimates. In addition, time-series properties of earnings are likely to differ between countries. For example, in countries with higher conservatism of accounting measures, accounting earnings will exhibit higher persistence. Accordingly, we use data for a sample of firms from four European countries (France, Germany, Netherlands, and the United Kingdom) for the 9-year period from 1995 to 2003 to examine how their accounting income and book value of equity are perceived by the respective capital markets. This paper extends previous studies of the effect of different regulatory regimes on the valuation implications of accounting income and book value of equity by taking into account the time-series properties of the key accounting variables posited in the Ohlson model, industry effects, and by distinguishing between the accruals and cash flows components of earnings. To evaluate the influence of these two variables we investigate their explanatory power for observed market values and also the forecasting power of the estimated models for out-of-sample companies. Our focus on the predictive ability of alternative definitions of the model is consistent with the expectation that accounting data should help predict future cash flows and consecutively market prices. At the same time, using predictive ability as a criterion is consistent with the requirement that valuation parameters be constant cross-sectionally. Finally, scale induced bias in the estimated coefficients causes heteroscedasticity and this suggests that R2 cannot be used as the primary criterion for evaluating differences in the value relevance of accounting data. Consistent with previous studies, we find that earnings and book value of equity in both models studied, as well as accruals in the second one, have valuation implications, that there are significant country-specific differences in the valuation of these variables, but that industry effects are a more important source of variation in the capitalization rates. The remainder of the paper is organized as follows: In the next section, we describe the clean surplus valuation models used in our study and develop the testable hypotheses. Subsequently we describe the sample and data used in our tests, and present our findings. The final section summarizes and concludes the study and presents ideas for further research.

This study extends previous work on cross-national differences in the valuation of earnings and book value of equity. Using a sample of companies from four European countries, we jointly estimate Ohlson’s valuation equation with linear information dynamics equations and we evaluate the influence of industry sectors in the forecasting performance of the model. We estimate two linear information valuation models (LIM) employing two levels of earnings disaggregation. The first LIM is based on aggregate residual earnings while the second includes total accruals as a separate variable. We initially estimated pooled versions of both LIMs and found earnings and book value of equity to be significant explanatory variables of the cross-sectional variation in market values while accruals in the second is not. Our first set of testable hypotheses was about cross-sectional variation in the valuation multiples of earnings and book value of equity in the two LIMs and accruals in LIM2 when these are estimated by country and by industry. Our findings were that earnings and book value are significant and with the expected sign in almost all country and industry contexts in both LIMs. Earnings valuation multiples differ significantly across both countries and industries while the valuation multiples on book value of equity vary significantly only when the models are estimated by industry. Accruals which is used as a separate variable in LIM2 is a significant variable only in a limited number of country and industry contexts but, nevertheless, there are statistically significant differences in the valuation multiples across both countries and industries. To test whether basing predictions on separate country and industry estimations of valuation model parameters affects equity market value predictions, we compare prediction errors from pooled and separate country and industry estimations for each LIM. Our results indicate that when estimating each LIM separately by industry, prediction errors are substantially smaller than when estimating a pooled model or when estimating the models by country. Finally, our results support disaggregation of earnings into accruals and cash flows if the LIM equity valuation models are estimated by country but not in the case of sector-specific estimation and prediction. In the latter case, prediction errors are larger on average than if we did not include accruals. Overall, our results indicate that there is convergence in financial-reporting practices within sectors. The size of the errors, however, suggests the need for consistency, which implies that there is a need for more sector-specific standards.