Accounting Information, Value Relevance, and Investors Behavior in the Egyptian Equity Market

Transcription

1 Accounting Information, Value Relevance, and Investors Behavior in the Egyptian Equity Market Aiman A. Ragab a and Mohammad M. Omran b,1 a College of Management & Technology, Arab Academy for Science & Technology, Alexandria, Egypt, P.O. Box 1029 b College of Management & Technology, Arab Academy for Science & Technology, Alexandria, Egypt, P.O. Box Corresponding Author: momran@aast.edu, Tel: , Fax:

2 Accounting Information, Value Relevance, and Investors Behavior in the Egyptian Equity Market* Abstract This study examines empirically whether national and international investors in the Egyptian stock market perceive accounting information based on the Egyptian accounting standards to be useful in stock valuation. Using a sample of all available listed firms in the Emerging Market Data Base (EMDB) from 1998 to 2002, we obtain evidence of the value relevance of accounting information in Egypt based on both return and price models. More importantly our results suggest that stock prices in Egypt are less informative about the future value of the firm than is accounting information. It is perhaps unreasonable to conclude that accounting information has higher value relevance in Egypt because financial reporting is of higher quality. It might however, imply that competing information sources such as earnings forecasts, firm research by financial analysts, management conference calls, etc. Are far less prevalent in Egypt. A likely policy implication of this study is that the Egyptian stock market needs complementary information sources other than published accounting reports, to become more informationally efficient. Key words: Stock Returns; Return and Price Models; Accounting Earnings; Market Value; Egypt. 2

3 I. Introduction Recent years have witnessed the growth of a global economy, which has been made possible by the advance in communications technology. Financial statements, usually prepared in accordance with national accounting standards, must be studied and interpreted if they are to be used as a basis of comparisons. Given the varied legal, economic and political backgrounds in countries around the world, it is not surprising that national accounting standards vary among countries. It is argued that the growth of international capital markets has focused attention on accounting information as an important source of credibility for national corporations with foreign investors, and as a factor assisting such investors in making international financial comparisons and in the flow of foreign investments, especially to underdeveloped countries. Consequently, a framework of international accounting standards was developed by the International Accounting Standards Committee, (Cooper and Lybrand, 1993). Due to the vital role published financial information plays in capital market, Egypt like many other countries has enacted a law enforcing implementation of these accounting standards by all entities doing business in Egypt (Egyptian Wakayeh 1997). The primary purpose of all accounting standards is to meet the needs of capital market (FASB, 1978; and IASC, 1994). Consequently, it is an empirical question whether accounting information is useful to domestic and foreign investors in the Egyptian market. This can be measured by the contemporaneous association between accounting information, stock returns and market value, in which valuation models link market prices and returns to different accounting measures of financial position and performance. 3

4 Since Ball and Brown (1968), the interest of researchers have been attracted to assessing the usefulness of accounting information, and the relation between accounting earnings and stock returns. Equity value of firms is linearly related to these accounting measures without any differentiation among these firms (Riahi- Belkaoui, 1994). Valuation models based on accounting information show that equity value is related to accounting earnings (e.g. Ball and Brown, 1968; and Collins Kothari, and Rayburn, 1989), and balance sheet measurements or both book value and earnings (e.g. Landsman, 1986; Barth, 1991; and Shevlin, 1991). Ohlson model (1995) led to an expansion of research studies on value relevance of accounting information to include both balance sheet measures of assets and liabilities and income statement measurements. In this context, Chen et al. (1999) empirically examine whether domestic investors in the Chinese stock market perceive accounting information to be useful in stock valuation. Their results confirm that accounting information is value relevant in the Chinese market despite the age of the market and the perception of inadequate accounting and financial reporting in china. Also Balbalyan (2001) proves that Swiss companies who reported their earnings under USA accounting standards provide more informative earning numbers. In view of the recent emphasis of researchers to study the usefulness of accounting information of non-us markets (Amir et al.1993; Harris et al. 1994; Barth and Clinh, 1996; Chan and Seow, 1996; and Graham and King, 1998), this study aims at extending such work by investigating the value relevance of accounting information in the Egyptian equity market. The importance of this study stems from the lack of alternative information sources in Egypt, such as earnings forecast, firm research by financial analysts, management conference calls, etc. Given that most transactions in Egypt are executed by individual investors with somewhat limited access to 4

5 information, reliance on the published accounting numbers. To the best of our knowledge, there is no empirical evidence on the informational efficiency of the Egyptian equity market, so this study adds additional empirical evidence to the literature by examining another part of the world (Middle East and North Africa region). As such, this presents novel evidence on this important issue in Egypt. Using a sample of all Egyptian firms listed in the IFC global index over the period , our empirical findings show that both return and price suggest accounting information is value relevant in the Egyptian equity market. However, our models contradict other previous studies, in that we failed to find any significant relationship between earnings changes and stock returns. The empirical findings provide evidence that both current earnings levels and earnings changes are not simply substitutes, since significantly more of the cross-sectional variation in returns is explained by earnings levels and earnings changes than is explained by either variable considered alone. The price model, on the other hand, incorporates both income statement and balance sheet measurements, and its empirical findings produce similar results. More precisely, we find that both earnings levels and the market value of the firm are positive and significant. In sum, the empirical findings are consistent between the return and price models, where both provide strong evidence that accounting information is value relevant in the Egyptian stock market. Comparing our results to more sophisticated markets accounting information has relatively more value relevance, this might be due to the fact that competing information sources such as earnings forecast, firm research by financial analysts, management conference calls, etc. are far less prevalent in Egypt. A potential policy implication is that the stock market Egypt needs complementary information sources other than published accounting reports, to become more efficient. 5

6 II. A Brief Review of Stock Market Development and Accounting Practices in Egypt Capital market activity in Egypt goes as far back as 1888, and during its heyday the Cairo stock exchange was ranked as the fifth most active exchange world-wide. However, as a result of waves of nationalization policy that started in late 1950s, the 93 most active listed firms had their stocks transferred to government bonds at 4% annual interest rate for 15 years. Thus, the stock market remained dormant for the following 20 years. The adoption of the 1974 open-door policy was meant to change things, and included reforms aimed at mobilizing domestic investment through the capital market. As such, the Egyptian Capital Market Authority (CMA) was established in the mid- 1970s in order to manage the stock market. Nevertheless, due to a host of reasons including biases in the tax code against investment in securities, absence of a governing securities law, inadequacy of financial disclosure, lack of protection of small investors, and adverse economic conditions stock market activity remained insignificant until the early 1990s. However, renaissance of capital market was an essential of the economic reform program introduced in In order to simplify the regulatory environment, a new capital market law was enacted in 1992 which aimed at encouraging private investment, increasing investor protection, and enhancing banks role in stimulating capital markets through the establishment of mutual funds. By and large, such developments in the stock market had their impact on accounting practices in Egypt by the mid 1990s. The Minister of Economy issued Decree No.323 in 1996 concerning the formation of a permanent committee to issue accounting and auditing standards. The committee suggested accounting standards and exposed drafts 6

7 to open debate among interested parties to be approved. This resulted in creating an initial set of comprehensive accounting standards. By 1997 the Minister of Economy issued another Decree No.503 which resulted in imposing international accounting standards on all firms listed in the Egyptian stock market (Egyptian Wakayeh, 1997). III. Data Selection Sample firms are selected for a pooled cross-sectional, time series model estimated for the period 1998 to We restrict the time series to 5 years because Egyptian accounting standards were continually revised until 1997, when the Egyptian government implemented its new accounting standards based on the International Accounting Standards. Accordingly, our sample starts from the year 1998 to mitigate the misleading effect of changes in accounting rules on reported numbers. To select the most active firms in the Egyptian stock market, all firms listed in the IFC global index from 1998 to 2002 were targeted. We use monthly stock prices listed in the Emerging Market Data Base (EMDB) of the IFC. The total number of firms listed in EMDB is 66 firms. However, no complete data were available for three firms, and several other firms exhibit potential outlier values. 1 Hence, we believe it appropriate to exclude them from the analysis. Thus, the maximum sample in 2002 contains 59 firms, and the minimum number is 56 firms (1998 and 2001). 2 The IFC indices are widely accepted in the international investment industry, forming the basis for index funds and structured financial instruments. Additionally, stocks included in the IFC indices were selected on the basis of market size, trading activity, and sector 1 It is possible, however, that results could be influenced by outliers. There is a choice of outlier trapping criteria, such as removing firms whose returns values are greater than a certain percentage or are in excess of a certain number of standard deviations from the mean. Due to the small number of firms, we limited our rule to those firms with CR or BHR in excess of 70% in either signs. 2 The number of firms in each year is not similar across years because of outliers. 7

8 representation. For these reasons, we believe they give a better representation of the market, whereas local emerging markets might not be representative since they include a large number of stocks that might be traded infrequently. IV. Methodology and Empirical Models We employ two models to measure the value relevance of accounting information. The first model is the return model introduced by Easton and Harris (1991), which includes both earnings levels and earnings changes as explanatory variables to assess their usefulness in determining stock returns (see for example, Amir et al. 1993; and Cheng et al. 1993). Current earnings can be decomposed as follows: EPS it P EPS = t 1 P + EPS P t 1 t 1 t 1 (1) Where EPS it is earnings for firm, i period t, t 1 EPS is earnings for firm i, period t-1, and EPS is the change in earnings from period t to t-1. Easton and Harris (1991), however, note that all three variables cannot be included in one regression model, since the inclusion of two of the variables preclude the inclusion of the third. Therefore, we consider each variable on its own to examine its impact on stock returns as follows: R i, t = α + β1epsit / Pit 1 + ε it (2) R R i, t = α + β1 EPS / Pit 1 + ε it (3) i, t = α + β1epsit 1 / Pit ε it (4) Where EPS it is level of annual earnings per share for firm i, period t and the change of annual earning per share, EPS it EPSit 1. EPSit is. Since there is no theoretical argument that supports the inclusion of EPSit 1 with either EPSit or EPSit in one model, we regress the annual stock returns in the current 8

9 period on both the earning per share (EPS) in the same period relative to the share price in the previous period and changes in the EPS relative to share price in the previous period. This allows us to compare results of the multivariate model with the aforementioned univariate models (equations 2 and 3). R i, t = + β1epsit Pit 1 + β 2 EPS / Pit 1 α / + ε (5) it Where R i, t is the annual return of firm i, period t, and is computed as follows: Pi, t Pi, t 1 + divi, t 1 R i, t =, (6) P i, t 1 Where, P i, t refers to the closing price of firm i at time t, i, t 1 P is the price of firm i at the end time t-1 and div i, t 1 is the dividend received for period t-1. There is no consensus on the appropriate methodology for calculating the annual stock returns (see among other, Barber and Lyon, 1997; Kothari and Warner 1997; Brav and Gompers, 1997; and Lyon, Barber, and Tsai, 1999). In general, researchers utilize two methods to calculate annual stock returns: cumulative return (CR) and buy-and-hold return (BHR). Given the fact that each method answers a different question, and thus yields different results, it would be appropriate to consider both methods in calculating annual returns. The CR for each individual firm is calculated as: CR e i, s, e = t= s R i, t, (7) Where CR i, s, e is the cumulative return for firm i from the event month s to the event month e, where s is the start month, which refers to the month of December or the 9

10 month of June prior to the end of the fiscal year and e refers to the month of December or the month of June after the fiscal year 3. Stock returns using the BHR method is calculated as follows: T = BHRi, ( 1+ Ri, t ) 1 T = 12 (8) T t= 1 Where BHR, is buy-and-hold return for firm i in period T, where T is the trading i T month number 12, and t = 1 indicates the first event month of calculating the return. Since returns on stocks might be affected by general economic factors, which might lead stock market indices to move up or down, we also estimate the earnings model using abnormal returns as follows: AR α β β + ε i, t = + 1EPSit / Pit EPS / Pit 1 it Where AR i, t is the annual abnormal return of firm i for period t, calculated as the raw return for a given firm minus the IFC global index benchmark return AR = R R, (9) i, t i, t IFC, t Where R, is the raw return on the IFC global index portfolio for period t. IFC t To calculate the CAR and BHAR for each individual firm, we utilize the following equations: e e CAR,, = R R, (10) i s e i, t t= s t= s IFC t T T BHARi, = ( 1+ Ri, t ) 1 (1 + RIFC, t ) 1 T = 12 (11) T t= 1 t= 1 3 Since the return of a given stock is based on a period extending from 9 months prior to the fiscal-year end and 3 months after the fiscal year-end, corresponding roughly with the period between announcing the financial statement, the starting month would be September for firms whose fiscal years end on June 30 th and March for firms whose fiscal years end on December 31 st. 10

11 However, since the rate of return using the above equations (10 and 11) is calculated without explicitly adjusting for risk (beta), we utilize the Sharp-Lintner CAPM to calculate the rate of return, to take the risk factor into consideration: CAPMR = R β R R ] (12) i, t f, t i[ IFC, t f, t Where CAPMR, is the annual return for firm i, time t using the CAPM, i t R, is the f t risk-free rate proxied as a short-term one-month rate for bank deposits, β is the risk of security i compared with the corresponding reference portfolio, and it is calculated from the CAPM regression, which is the slope obtained from regressing R R ] on R R ] for the estimated period. [ i, t f, t [ IFC, t f, t With the CAPMR calculated, we apply the same two methods, CRs and BHRs, mentioned previously. Therefore, the abnormal return using the CAPM is calculated as follows: CAPMAR Where = R R β R R ], (13) i, t i, t f, t i[ IFC, t f, t CAPMAR i, t is the CAPM abnormal return for firm i in time t. i The other model used in this study is the price model by Ohlson (1995), which has been extensively used by researchers in examining the value relevance of accounting information (see for example; Landsman, 1986; Barth, 1991; Barth and Clinch, 1996; Eccher et al. 1996; Burgstahler and Dichev, 1997; and Collins et al. 1997). The reason for such extensive use of the price model is that it yields an unbiased earnings coefficient, since the stock price reflects the cumulative effects of earnings information (Kothari and Zimmerman, 1995). In addition, the market value is related to both the book value and the accounting earnings. 11

12 We utilize the price model to assess the usefulness of accounting information in valuing the stock prices, which shows how a firms market value is related to both book value and accounting earnings. The modified Ohlson model is as follows: MVit = α + β1 EPSit + β 2BVit + ε (14) Where MV it is the market value per share of firm i at time t, and BV it is the book value per share of equity of firm i at year t. V. Descriptive Statistics and Results V.1 Descriptive Statistics Table 1 presents descriptive statistics for different variables based on the price and return models. All accounting data are based on the Egyptian generally accepted accounting principles and are subtracted -along with stock prices- from the Emerging Market Data Base (EMDB). (Insert Table 1 near here) The average stock returns over a five-year period from 1998 to 2002 is a negative 30-34% depending on the method of calculation. The median stock return tends to be similar to the mean. Abnormal returns however seem to differ significantly according to whether these returns are adjusted for risk. The abnormal returns calculated from the market-adjusted model, whether based on CR or BHR, are always higher (less negative) than those calculated from the CAPM. Consequently, this indicates in all likelihood a negative risk premium 4 and an average beta is likely to be less than one. There does not, however, seem to be any significant difference in stock returns based on the calculation method (CR and BHR). Nevertheless, the standard deviations of the three calculated stock returns based on CR are higher compared with the BHR 4 Risk premium is the difference between market rate of returns and the risk-free rate. The Egyptian stock market returns (at close to 6%) over performed less than returns on Treasury bills and saving deposits (at close to 10%); see Bolbol and Omran (2005). 12

13 method, in which this reflects that the former method appears to have more volatile return than the latter one. 5 Furthermore, the current earnings, the previous earnings, and the earning changes variables, all exhibit similar trends as the differences between the mean and the median of each variable are very low and, hence, the standard deviations are very low as well. Moving to the price model variables, the data show some dispersion among firms. Our sample firms are diversified in terms of size (market value, and book value), and profitability (EPS). Notwithstanding the abovementioned differences, from the relationship between the mean and the median of each variable, we conclude that most variables skewed toward the right. V.2 Results All regression models are estimated for the pooled cross-section and time-series sample as well as for each year of available data. (Insert Tables 2 and 3 near here) The results from the return model (level model; equation 2), which incorporates the deflated EPS as explanatory variable, are reported in Table 2. For pooled data, the results show that the deflated EPS has a positive and significant impact on stock returns at the 1% level; whether we employ CR (Panel A) or BHR (Panel B). The F- ratios are highly significant at the 1% level in both panels. The results obtained from the year-by-year regressions are pervasive and corroborate our previous findings of the pooled data. In each year, the coefficients of the deflated EPS are positive and significant at the 1% and the 5% levels, except for the year 2001 in Panel B, in which the coefficient is significant at the 10% level. As for the changes model, the reported results given in Table 3 indicate that none of the coefficients of the changes in EPS is significant at any level. Furthermore, F- 5 CRs have larger difference between the mean and median and between the minimum and maximum returns compared with BHRs, so that they are more volatile. 13

14 statistics are not significant and the values of the R 2 are very low. These results are applied to both year-by-year and pooled data, using both CR and BHR methods. Although our first findings (the positive and significant coefficients of the deflated EPS) are consistent with previous research studies, the insignificant coefficients of the changes in EPS are inconsistent with the literature (see e.g., Easton and Harris; 1991 and Chen et.al. 2001). This might be because investors in the Egyptian stock market might have a very short-term horizon. Consequently, investors might be more concerned with earnings levels in valuing stocks (rather than earnings), changes, which leads them to concentrate on the contemporaneous earnings. In sum, the association between stock returns and earnings levels is significant, while it is not with changes in earnings. Recall from equation (1), that the current earnings levels equal to previous period levels plus changes in earnings, we also consider the association between deflated prior EPS and stock returns. (Insert Table 4 near here) We regress the stock returns on the deflated earnings in the previous period and report the results in Table 4, in which Panel A considers CR method and Panel B considers BHR method. The results show that prior EPS is positive and significant, both for the pooled and year-by-year regressions and regardless of the return method of calculation. To a large extent, these findings are qualitatively similar to our previous findings on the association between current EPS and stock returns. Also, the values of R 2 in any year are higher than the reported results for the pooled data; just as similar to the results on current EPS. We conclude that deflated EPS, whether current or previous period, is correlated with stock returns, while earnings changes have no significant relationship with stock returns. 14

15 Here we empirically examine a multivariate return model that incorporates both the earnings levels and earnings changes as previously stated is equation 5. In our multivariate model, we regress the annual stock returns in the current period on both EPS in the same period relative to share price in the previous period and changes in EPS relative to share price in the previous period. This allows us to compare the results of the multivariate model with the aforementioned univariate models (equations 2 and 3). (Insert Table 5 near here) The multivariate regression estimates are reported in Table 5, Panels A and B. The results we find mirror those obtained in Tables 2 and 3. Qualitatively the results are substantially identical. More precisely, the empirical findings suggest that EPS is positive and significantly associated with stock returns, while the earnings changes have no relationship with stock returns. In all years, as well as the pooled data, when we add changes in earnings to the regression model (equation 2) the changes in R 2 are not significant. 6 Overall, the empirical findings so far provides us with evidence that only earnings levels are relevant in explaining stock returns, while earnings changes are independent of stock returns. Notwithstanding these findings, the results suggest that both current earnings levels and earnings changes are not just substitutes. They rather complement each other in the sense that for the pooled, as well as year-by-year results, significantly more of the cross-sectional variation in returns is explained by earnings levels and earnings changes than is explained by either variable considered alone. 6 This means that the addition of changes in earnings to equation 2 yields insignificant partial F- statistics. 15

16 In our multivariate model, however, we do not consider only stock returns as dependent variables; we extend that by looking at the abnormal returns using CAR and BAHR, based on the market-adjusted returns and the CAPM. Employing abnormal returns is based on the idea that much of the literature on the relationship between earnings and returns has focused on unexpected earnings and unexpected or abnormal returns. Consequently, we repeat equation 5 by replacing abnormal returns instead of stock returns and regress that on both earnings levels and earnings changes. 7 (Insert Tables 6 and 7 near here) The results presented in Tables 6 and 7 support our previous findings, in which only earnings levels are positive and significant. However, comparing the values of R 2 in reported in both A Panels of Tables 6 and 7 (market-adjusted model) with those reported in B Panels of Tables 6 and 7 (CAPM) we notice that the latter explains more of the cross-sectional variation in abnormal returns. This might indicate adjusting abnormal returns for risk provides a more accurate measure than market-adjusted returns. The return models discussed above allows us to assess the value relevance of accounting information. There are, however, the price models that can show us how a firm s market value is related to both book values (BV) and accounting earnings (EPS). The advantage of price models is that the two accounting information they incorporate play different roles in stock pricing, hence the use of these models expands the scope of assessing value relevance into both balance sheet and income statement. 7 Brown et al. (1987) and Easton and Harris (1991) suggest that given that the change in earnings is a dominant measure of unexpected earnings, it is better to incorporate both earnings changes and earnings levels as measures of unexpected earnings. 16

17 (Insert Table 8 near here) Table 8 shows the multiple regression results of the relationship between market value and both earnings levels and book value. The pooled cross-section and time-series regression, as well as, the year-by-year regressions produce similar results. Both EPS and BV are positive and significant at the 1% level in most cases. In any event, the empirical findings are consistent between the return and price models, in which we have a strong evidence that accounting information is value relevant in the Egyptian stock market. However, when we compare our results with the literature, we observe that the accounting information in Egypt has relatively more value relevance with those reported in sophisticated and large emerging markets (see for example; Landsman, 1986; Barth, 1991; and Shevlin, 1991; Amir et al.1993; Harris et al. 1994; Barth and Clinh, 1996; Chan and Seow, 1996; Graham and King, 1998; Chen et al. 1999; and Balbalyan 2001). This might be due to the fact that competing information sources such as earnings forecast, firm research by financial analysts, management conference calls, etc. are far less prevalent in Egypt. VI. Conclusion This study presents an empirical examination of whether financial statements in the Egyptian equity market, prepared according to the Egyptian accounting standards, are useful for stock valuation. We employ commonly-used models in the value relevance literature, the Easton and Harris return model and the Ohlson price model. Using a sample of all Egyptian firms listed in the IFC global index over the period , we obtained evidence of the value relevance of the accounting information in the Egyptian market. More precisely, based on both pooled cross-section and time-series regressions and the year-by-year 17

18 regressions, the return models indicate that earnings levels are significantly associated with stock returns, which is consistent with the literature. However, our results contradict other previous studies, in that no significant relationship between earnings changes and stock returns was found. This might imply that investors in the Egyptian stock market have a very short-term horizon, hence, they concentrate on earnings levels (contemporaneous earnings) in valuing stocks; not earnings changes. Despite these results, the empirical findings provide evidence that both current earnings levels and earnings changes are not just substitutes because significantly more of the crosssectional variation in returns is explained by earnings levels and earnings changes than is explained by either variable considered alone. The price model, on the other hand, incorporates both income statement and balance sheet measurements, and its empirical findings produce similar results, where we find that both earnings levels and the market value of the firm are positive and significant. In sum, both the return and the price models provide consistent evidence and the results are in general are consistent with the literature, in that we can confirm that accounting information is value relevant in the Egyptian equity market. An important result in this study is that the accounting information has relatively more value relevance in Egypt compared to more mature markets, which might reflect that competing information sources such as earnings forecast, firm research by financial analysts, management conference calls, etc. are far less prevalent in Egypt. A potential policy implication is that stock market in Egypt needs complementary information sources other than published accounting reports to become more informationally-efficient. 18

19 References 1. Amir, E., Harris, T., and E. Venuti A Comparison of the Value- Relevance of U.S. Versus Non-U.S. GAAP Accounting Measures Using Form 20-F Reconciliation. Journal of Accounting Research, 31, Ball, R., and Brown, P An Empirical Evaluation of Accounting Income Numbers. Journal of Accounting Research, 6, Baraber, B., and Lyon, J., Detecting Long-Run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics. Journal of Financial Economics, 43, (3), Barth, M Relative Measurement Errors among Alternative Pension Assets and Liability Measures. The Accounting Review, 66 (3), Barth, M. and Clinch, G International Accounting Difference and their Relation to Share Prices: Evidence from U.K., Australian, and Canadian Firms. Contemporary Accounting Research, 13, Babalyan, L Association Between Accounting Earnings and Stock Returns as a Measure of Value Relevance of Accounting Standards: Empirical Evidence from the Swiss Market. EFA Berlin Meetings, Discussion Paper. 7. Bolbo, A., and Omran, M Investment and the Stock Market: Evidence from Arab Firm-Level Panel Data. Emerging Market Review, 6, Brav, A., and Gompers, P Myth or Reality? The Long-Run Underperformance of Initial Public Offerings: Evidence from Venture Capital and non-venture Capital-Backed Companies. Journal of Finance, 52 (5), Brown, L., Griffiin, P., Hagerman, R., and M. Zmijewski An Evaluation of Alternative Proxies for the Market s Assessment of unexpected Earnings. Journal of Accounting and Economics, 9,

20 10. Burgstahler, D., and Dichev, I Earnings, Adaptations and Equity Value. The Accounting Review, 72, Chan, J.P., and G.S. Seow The Association between Stock Returns and Foreign GAAP Earnings versus Adjusted to U.S. GAAP. Journal of Accounting and Economics, 21, Chen, J.P.C., Gul, F. and X. Su, A comparison of Reported Earnings under Chinese GAAP versus IAS: Evidence from the Shanghai Stock Exchange, Accounting and Business Research (Summer): ,, and, Is Accounting Information Value Relevant in the Emerging Chinese Stock Market?. Journal of International Accounting, Auditing and Taxation 10, Collins, D., Kothari, S., and J. Rayburn An Analysis of Intertemporal and Cross-Sectional Determination of Earning Response Coefficient. Journal of Accounting and Economics, 11, Cooper and Lybrand International Accounting Summaries: A Guide for Interpretation and Comparison, (New York: John Wiley & Sons, Inc.). 16. Easton, P., and Harris, T Earnings as an Explanatory Variable for Returns. Journal of Accounting Research, 29, Eccher, E., Ramesh, K., and S. Thiagarajan Fair Value Disclosures by Bank Holding Companies. Journal of Accounting and Economics, 22, Egyptian, Wakayeh Ministerial Decree No.503/1997 concerning the Egyptian Accounting Standards, Issue No.203, 12 October, (Cairo: Al- America Printing Corporation). 19. FASB Statement of Financial Accounting Concepts No. 1, Objectives of Financial Reporting by Business Enterprises. (Norwalk, CT: FASB). 20

21 20. Graham, R., and King, R The Relation of Firm Market Value with Book Values and Residual Accounting Earnings in Six Asian Countries. Working Paper, Oregon State University. 21. Harris, T., Lang, M., and H. Moller The Value Relevance of German Accounting Measures: An Empirical Analysis. Journal of Accounting, 32, IASC International Accounting Standards. (Basingstoke, Hants: Burgess Science Press). 23. Kothari, S., and Zimmerman, J Price and Return Models. Journal of Accounting and Economics, 20, Kothari, S., and Warner, J Measuring Long-Run Horizon Security Price Performance, Journal of Financial Economics, 43 (3), Landsman, W An Empirical Investigation of Pension Fund Property Rights. The Accounting Review October, 61 (4) Lyon, J., Barber, B., and C. Tsai Improved Methods for Tests of Long- Run Abnormal Stock Returns. Journal of Finance, 54 (1), Ohlson, J. A Earnings, Book Value and Dividends in Security Valuation. Contemporary Accounting Research, 11, Riahi-Belkaoui, A Net Value Added as an Explanatory Variable for Returns. Managerial Finance, 20 (9), Shevlin, T The Valuation of R & D Firms with R & D Limited Partnerships. The Accounting Review, 66 (1),

22 Table 1 Descriptive Statistics The table shows some basic descriptive statistics for the variables we used in our analysis. It includes measures of central tendency and measures of variability. We provide the mean, the median, the minimum, the maximum, and the standard deviation values of each variable over the period CR is cumulative return, BHR is buy-and-hold return, CAR is cumulative abnormal return using the market-adjusted model, BHAR is buy-and-hold abnormal return using the market-adjusted model, CAR (CAPM) is cumulative abnormal return using the capital asset pricing model, BHAR (CAPM) is buy-and-hold abnormal return using the capital asset pricing model, EPS t /P t-1 is the earning per share at time t divided by the price per share at the time t- 1, EPS t /P t-1 is the change in earning per share at time t divided by the price per share at the time t-1, EPS t-1 /P t-1 is the earning per share at time t-1 divided by the price per share at the time t-1, MV is the market value per share, EPS is the earning per share, and BV is the book value per share. Variables Mean Median Minimum Maximum Std. Dev. CR BHR CAR BHAR CAR (CAPM) BHAR (CAPM) EPS t /P t EPS t /P t EPS t-1 /P t MV EPS BV

23 Table 2 Simple Regressions of Annual Stock Returns on Deflated Earnings Levels The table shows the results from simple regressions of the relationship between stock returns and deflated earnings levels. The level model is based on the following equation: R i,t = α + β 1 EPS i,t /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of CR and BHR, and EPS i,t /P i,t-1 is the earning per share of firm i at time t divided by the price per share of firm i at time t-1, which refers to the price at the beginning of 9 months prior to the fiscal-year end. CR is cumulative return, BHR is buy-and-hold return, and No. is the number of observations in the regression. Panel A- Dependent Variable: R i,t (CR) Pool Constant EPS i,t-1 /P i,t (-1.23) (-4.76)*** (-6.25)*** (-6.73)*** (-4.95)*** (-7.31)*** (2.69)** (3.24)*** (2.06)** (2.31)** (2.81)*** (2.89)*** No R 2 ( % ) F - Ratio 7.21 *** *** 4.77** 5.33** 7.91 *** 8.37 *** Panel B- Dependent Variable: R i,t ( BHR) Pool Constant EPS i,t-1 /P i,t (-0.71) (-4.47)*** (-5.77)*** (-5.52)*** (-5.42)*** (-9.10)*** (2.42)** (2.85)*** (2.49)** (1.74)* (2.42)** (3.59)*** No R 2 ( % ) F - Ratio 5.84 ** 8.11 *** 6.21** 3.04* 5.84 ** *** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 23

24 Table 3 Simple Regressions of Annual Stock Returns on Deflated Earnings Changes The table shows the results from simple regressions of the relationship between stock returns and deflated earnings changes. The change model is based on the following equation: R i,t = α + β 1 EPS i,t /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of CR and BHR, and EPS i,t /P i,t-1 is the change in earning per share of firm i at time t divided by the price per share of firm i at time t-1. CR is cumulative return, BHR is buy-and-hold return, and No. is the number of observations in the regression. Panel A- Dependent Variable: R i,t (CR) Pool Constant EPS i,t-1 /P i,t (2.18)** (-3.73)*** (-7.03)*** (-8.15)*** (-4.37) (-7.35)*** (1.23) (0.017) (0.98) (-1.03) (0.32) No R 2 ( % ) F - Ratio Panel B- Dependent Variable: R i,t (BHR) Pool Constant EPS i,t-1 /P i,t (2.76)** (-3.98)*** (-5.47)*** (-7.10)*** (-5.66)*** (-8.98)*** (0.05) (1.14) (0.22) (0.96) (-0.78) (1.22) No R 2 ( % ) F - Ratio ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 24

25 Table 4 Simple Regressions of Annual Stock Returns on Deflated Prior-Period Earnings The table shows the results from simple regressions of the relationship between stock returns and deflated prior-period earnings based on the following equation: R i,t = α + β 1 EPS i,t-1 /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of CR and BHR, and EPS i,t-1 /P i,t-1 is the prior-period earning per share of firm i,, i.e., the earning per share at time t-1, divided by the price per share of firm i at time t-1, which refers to the price at the beginning of 9 months prior to the fiscal-year end. CR is cumulative return, BHR is buy-and-hold return, and No. is the number of observations in the regression. Panel A- Dependent Variable: R i,t (CR) Pool Constant EPS i,t-1 /P i,t (-0.86) (-3.69)*** (-5.55)*** (-6.48)*** (-5.19)*** (-6.74)*** (2.29)** (2.06)** (1.90)** (2.27)** (3.13)*** (2.30)** No R 2 ( % ) F - Ratio 5.24 *** 4.26 ** 3.60** 5.13** 9.77 *** 5.30 ** Panel B- Dependent Variable: R i,t (BHR) Pool Constant EPS i,t-1 /P i,t (-0.32) (-3.42)*** (-4.76)*** (-5.22)*** (-5.50)*** (-8.03)*** (1.99)** (1.69)* (1.96)** (1.69)* (2.76)*** (2.45)** No R 2 ( % ) F - Ratio 3.97 ** 2.92 * 3.86** 2.76* 6.86 *** 6.02 ** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 25

26 Table 5 Multiple Regressions of Annual Stock Returns on Deflated Earnings Levels and Earnings Changes The table shows the results from multiple regressions of stock returns and abnormal returns on deflated earnings levels and changes. The model is based on the following equation: R i,t = α + β 1 EPS i,t /P i,t-1 + β 2 EPSi,t /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of CR and BHR. EPS i,t /P i,t-1 is the earning per share of firm i at time t divided by the price per share of firm i at time t-1, and EPS i,t /P i,t-1 is the change in earning per share of firm i at time t divided by the price per share of firm i at time t-1, which refers to the price at the beginning of 9 months prior to the fiscal-year end. CR is cumulative return and BHR is buy-and-hold return. No. is the number of observations in the regression. Panel A - Dependent Variable: R i,t (CR) Pool Constant EPS i,t /P i,t-1 EPS i,t /P i,t (-1.53) (-4.65)*** (-6.09)*** (-6.50)*** (-5.19)*** (-7.16)*** (2.92)*** (3.08)*** (2.88)*** (2.22)** (2.97)*** (2.89)*** (-1.11) (0.94) (-1.03) (0.84) (-1.12) (-0.34) No R 2 ( % ) F - Ratio 4.25** 5.68 *** 4.13** 2.99* 4.86 ** 4.22 ** Panel B- Dependent Variable: R i,t (BHR) Pool Constant EPS i,t /P i,t-1 EPS i,t /P i,t (-0.97) (-4.36)*** (-6.01)*** (-5.31)*** (-5.53)*** (-8.7)*** (2.58)** (2.68)** (3.51)*** (1.86)* (2.51)** (3.39)*** (-0.92) (1.15) (-1.12) (0.83) (-1.06) (0.45) No R 2 ( % ) F - Ratio 3.33** 4.75 ** 3.21** 2.85* 3.5 ** 6.53 *** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 26

27 Table 6 Multiple Regressions of Cumulative Abnormal Returns on Deflated Earnings Levels and Earnings Changes The table shows the results from multiple regressions of stock returns and abnormal returns on deflated earnings levels and changes. The model is based on the following equation: R i,t = α + β 1 EPS i,t /P i,t-1 + β 2 EPSi,t /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of CAR, EPS i,t /P i,t-1 is the earning per share of firm i at time t divided by the price per share of firm i at time t-1, and EPS i,t /P i,t-1 is the change in earning per share of firm i at time t divided by the price per share of firm i at time t-1, which refers to the price at the beginning of 9 months prior to the fiscal-year end. CAR is cumulative abnormal return using the market-adjusted model, and CAR-CAPM is cumulative abnormal return using the capital asset pricing model. No. is the number of observations in the regression. Panel C- Dependent Variable: R i,t (CAR) Pool Constant EPS i,t /P i,t (-2.06)** (-3.78)*** (-3.99)*** (-4.57)*** (-0.81) (-4.52)*** (2.63)** (2.51)** (2.63)** (2.76)*** (1.68)* (2.74)*** EPS i,t /P i,t (-1.24) (-0.49) (-1.13) (-0.64) (-1.2) (-1.18) No R 2 ( % ) F - Ratio 3.49** 3.47 ** 3.46** 3.79** ** Panel E- Dependent Variable: R i,t (CAR-CAPM) Pool Constant EPS i,t /P i,t-1 EPS i,t /P i,t (-2.01)** (-6.8)*** (-5.32)*** (-7.01)*** (-5.31)*** (-9.75)*** (2.82)*** (4.05)*** (1.96)* (2.12)** (2.85)*** (3.33)*** (-1.12) (1.03) (-1.26) (0.25) (-0.88) (-0.98) No R 2 ( % ) F - Ratio 3.99** 9.46 *** * 4.25 ** 5.56 *** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 27

28 Table 7 Multiple Regressions of Buy-and-Hold Abnormal Returns on Deflated Earnings Levels and Earnings Changes The table shows the results from multiple regressions of stock returns and abnormal returns on deflated earnings levels and changes. The model is based on the following equation: R i,t = α + β 1 EPS i,t /P i,t-1 + β 2 EPSi,t /P i,t-1 + ε i,t, where R i,t is the return on firm i at time t, which refers to the period over 12 months extending from 9 months prior to the fiscal-year end to 3 months after the fiscal-year end, and it takes the form of BHAR, EPS i,t /P i,t-1 is the earning per share of firm i at time t divided by the price per share of firm i at time t-1, and EPS i,t /P i,t-1 is the change in earning per share of firm i at time t divided by the price per share of firm i at time t-1, which refers to the price at the beginning of 9 months prior to the fiscal-year end. BHAR is buy-and-hold abnormal return using the market-adjusted model, and BHAR-CAPM is buy-and-hold abnormal return using the capital asset pricing model. No. is the number of observations in the regression. Panel D- Dependent Variable: R i,t (BHAR) Pool Constant EPS i,t /P i,t-1 EPS i,t /P i,t (-1.27) (-4.03)*** (-3.24)*** (-4.48)*** (-0.93) (-5.34)*** (2.16)** (2.55)** (1.86)* (1.91)* (2.26)** (2.85)*** (-0.83) (-0.25) (-0.97) (0.35) (-1.06) (-0.77) No R 2 ( % ) F - Ratio 2.63* 3.36 ** ** 4.07 ** Panel F- Dependent Variable: R i,t (BHAR-CAPM) Pool Constant EPS i,t /P i,t-1 EPS i,t /P i,t (-1.38) (-6.59)*** (-5.57)*** (-8.42)*** (-6.26)*** (-11.53)*** (2.57)** (3.6)*** (1.69)* (1.85)* (2.94)*** (3.81)*** (-0.86) (1.21) (-1.16) (0.99) (-0.61) (-0.15) No R 2 ( % ) F - Ratio 3.31** 7.94 *** * 4.38 ** 7.51 *** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 28

29 Table 8 Multiple Regressions of Annual Market Values on Earnings Book Values The table shows the results from multiple regressions of the relationship between market values and earnings levels and book values. The model is based on the following equation: MV i,t = α + β 1 EPS i,t + β 2 BV i,t + ε i,t, where MV i,t is the market value of firm i at time t, which refers to fiscal-year end plus 3 months, EPS i,t is the earning per share of firm i at time t and BV i,t is the book value per share of firm i at time t. No. is the number of observations in the regression. Dependent Variable: MV i,t Pool Constant EPS i,t BV i,t (-0.11) (1.15) (2.07)** (-1.61) (-1.81)** (1.26) (2.07)** (4.94)*** (4.7)*** (6.64)*** (7.64)*** (8.38)*** (4.22)*** (2.38)** (1.86)* (3.04)*** (6.08)*** (3.86)*** No R 2 ( % ) F - Ratio 19.93*** *** 38.89*** 36.82*** 58.7 *** *** ***, **, and * Significant at the 1, 5, and 10 % level, respectively. Figures between parentheses are t statistics. 29