Volume 27 Number 3 2001 65 Investment Opportunity Set Dependence of Dividend Yield and Price Earnings Ratio by Ahmed Riahi-Belkaoui and Ronald D. Picur, University of Illinois at Chicago Abstract This study develops and tests a market valuation model whose main prediction is that equity value is a function of earnings, dividends and book value, where the function depends on the relative level of the investment opportunity set. Using a sample of US multinational firms, the study demonstrated that firms in a high investment opportunity set are PE valued and that firms in a low investment opportunity set are dividend yield valued. I. Introduction Valuation models based on accounting information show equity to be related to: (1) accounting earnings (Ball and Brown, 1968; Barth, 1991; Collins and Kothari, 1992), (2) balance sheet measures of assets and liabilities (Landsman, 1986), or (3) both earnings and book value. The last model is based upon the proposition that in more realistic settings, where market imperfection exists, the accounting system provides information about book value and earnings as complementary rather than redundant components of equity value (Burgstahler and Dichev, 1997). For example, Ohlson expresses price as a linear function of book value and abnormal earnings (Ohlson, 1990, 1995). A linear form, where equity value is an additive function of earnings and book value, is typically assumed in most empirical specifications derived from the model (Easton and Harris, 1991; Sougiannis, 1994). That model can be further refined by decomposing the earnings component into dividends and retained earnings (Rees, 1997). What results is a valuation model that relates equity to dividends, retained earnings and book value. The presence of dividends and retained earnings in the model is compatible with survey and empirical evidence that suggests the dominant valuation model used by analysts is the price-earnings (PE) ratio. However, other approaches, such as the dividend yield, are also important. In fact, there is evidence of industry-dependence in analysts preference between the dividend yield and price-earnings ratio valuation models (Barker, 1999). These results raise the question of the nature and importance of the role of dividend and earnings in a valuation model. This paper argues that the function depends on the relative level of the investment opportunity set. Firms in a high investment opportunity set group are shown to be PEvalued while firms in a low investment opportunity set are shown to be yield valued. II. Hypothesis This study extends the basic valuation model to include both dividends and retained earnings reflecting analysts use of both price-earnings (PE) ratio and dividend yield in determining and evaluating share price. The following null hypotheses are expected to be rejected:
Managerial Finance 66 H 1 : H 2 : There is no correlation between dividends and share prices. There is no correlation between earnings and share prices. This study also tests the investment opportunity set-dependence of dividend yield and PE ratio. The presence of a high investment opportunity set implies profitability and growth opportunities that can translate into even higher profitability. Dividends are expected to be of greater valuation-relevance for firms with a low investment opportunity set level. One may also argue that such firms are valued on the basis of dividends alone; accordingly, retained earnings are not valued-relevant. This implies the rejection of the following null hypotheses H 3 and H 5 and acceptance of H 4 : H 3 : H 4 : H 5 : There is no difference in the value relevance of dividends and retained earnings for firms with a low investment opportunity set level. Retained earnings are not value-relevant for firms with a low investment opportunity set level. There is no difference in the value relevance of dividends between firms with a high investment opportunity set level and firms with a low investment opportunity set level. Similarly, for firms with a high investment opportunity set level, earnings are expected to be value relevant. Therefore, both earnings and dividends will be value relevant for these firms. This implies acceptance of the following hypothesis H 6 and rejection of H 7 : H 6 : H 7 : There is no difference in the value relevance of dividends and retained earnings for firms with a high investment opportunity set level. There is no difference in the value-relevance of retained earnings between firms with a high investment opportunity set level and firms with a low investment set opportunity set level. III. Market-Based Model The study relies on a valuation model rather than an event study. The dependent variable is price rather than returns, and the independent variables are levels rather than expected differences. All the tests are derived from an empirical version expressing the market price (P) as a function of book value per share (BV) and earnings per share (E): P it =a 0 +a 1 BV it +a 2 E it +e it (1) To examine dividend relevance, earnings per share (E) is decomposed into dividend per share (DV) and retained earnings per share (RE). An argument for dividends as a sign of value has been made in both survey and empirical research (Battacharya, 1979; Miller and Rock, 1985). Therefore: P it =b 0 +b 1 BV it +b 2 RE it +b 3 DV it +e it (2) To examine investment opportunity set relevance to the role of dividends and retained earnings the following model is proposed:
Volume 27 Number 3 2001 67 P it = a+bios it + cbv it + dre it + eios it RE it +fdv it + gios it DV it +e it (3) Where IOS is a dummy variable which takes the value of 1 for a high investment opportunity set and the value of 0 for a low investment opportunity set. Equation (3) will be used to test each of the null hypotheses. IV. Methods Sample The population consists of all the firms in Forbes Most International firms from 1992 to 1998. These firms were classified as being either a high investment opportunity set group or a low investment opportunity set group. Based on a factor score of the investment opportunity set defined below, high investment opportunity set firms were chosen from the top 25% of the distribution scores while low investment opportunity set firms were chosen from the bottom 25% of the distribution factor scores. The final sample included 256 firm-year observations. Measuring the Investment Opportunity Set Three measures of the investment opportunity set were factor analyzed, namely, market to book assets, market to book equity and earnings/price ratio. One common factor appeared to explain the intercorrelations among these three individual measures. The factor score was used to measure the level of the investment opportunity set of each firm. V. Results Descriptive statistics of the variables used in the model are shown in Exhibit 1. The data exhibit a moderate amount of skewness and kurtosis. The matrix for the correlations between all the variables used in the analysis is presented in Exhibit 2. The low intercorrelations among the prediction variables used in the model indicate no reason to suspect multicollinearity; various diagnostic tests run on the derived regression models confirmed it was not a problem. Exhibit 3 presents the results of the regression coefficients for all the independent variables using the pooled sample and the results for individual years. The Breusch and Pagan (1979) test for heteroscedasticity yielded an X 2 with a minimum of 137.65 and a maximum of 183.62 for all the regressions, indicating the heteroscedasticity would be a problem in these regressions. Accordingly, the reported t-statistics were based on White s heteroscedasticity corrected covariance matrix (White, 1990). The results of all the regressions show a high R 2 (0.36 in 1994 and 0.23 in 1996), together with significant coefficients on BV, IOSRE, DV and IOSDV. As expected, the results confirm a strong cross-sectional relationship between price on one hand and book value, retained earnings, and dividends on the other hand. Both H 1 and H 2 are therefore rejected consistent with the use of RE and DV in the valuation model. H 3 is firmly rejected. In all the regression results, the coefficient in RE is much smaller than the coefficient of DV. A stronger test of the relationship is provided by H 4 which stipulates that dividends are not relevant. H 4 is supported as the coefficient of RE is significant in all cases. In addition, when a restriction that the coefficient of RE is equal to
Managerial Finance 68 zero is imposed in the model, a Wald statistic of 6.32 allows a rejection of the null hypothesis that the coefficient of RE is equal to zero. H 5 is also rejected. IOSDV is negative and significant in all cases indicating that there is a difference between firms in both groups, with the rate at which dividends are capitalized being lower for the high investment opportunity set group. H 7 is also rejected since the coefficient on IOSRE is large, positive and significant in all cases, establishing a relationship between price and retained earnings for firms in a high investment opportunity set group. Finally, H 6 is rejected in favor of retained earnings being more important for firms in a high investment opportunity set group. Further testing of H 6 by imposing a restriction on the model than the sum of the coefficients on RE (i.e. d and e) is equal to the sum of those on DV (i.e. f and g) results in a significant Wald statistic of 8.42, and a corresponding rejection of H 6. VI. Conclusions This paper developed and tested a market valuation model whose main prediction is that equity value is a function of earnings, dividends and book value, where the function depends on the relative level of the investment opportunity set. Using a sample of US multinational companies, the study demonstrated that firms in a high investment opportunity set group are PE valued and that firms in a low investment opportunity set group are dividend yield valued. For firms in a low investment opportunity set group, dividends are of greater relevance than earnings, while for firms in a high investment opportunity set group both retained earnings and dividends are relevant, even though retained earnings are more relevant than dividends.
Volume 27 Number 3 2001 69 References Ball, R. and P. Brown. An Empirical Evaluation of Accounting Income Numbers, Journal of Accounting Research (Autumn 1968), pp.159-178. Barker, R.G. Survey and Market-Based Evidence of Industry-Dependence in Analysts Preference Between the Dividend Yield and Price-Earnings Ratio Valuation Models, Journal of Business Finance and Accounting (96, 1999), pp.393-418. Barth, M. Relative Measurement Errors Among Alternative Pension Asset and Liability Measures, The Accounting Review (July 1991), pp.433-463. Battacharya, O. Imperfect Information, Dividend Policy and the Bind in the Hard Fallacy, Bell Journal of Economics (10, 1979), pp.259-270. Burgstahler, D.C. and I.D. Dichev. Earnings, Adaptation and Equity Value, The Accounting Review (April 1997), pp.187-216. Collins, D. and S.P. Kothari. An Analysis of Intertemporal and Cross-Sectional Determinants of ERCs, Journal of Accounting and Economics (March 1992), pp.148-183. Easton, P. and T. Harris. Earnings as an Explanatory Variable for Returns, Journal of Accounting Research (Spring 1991), pp.19-36. Landsman, W. An Empirical Investigation of Pension Fund Property Rights, The Accounting Review (October 1986), pp.44-64. Miller, M. and K. Rock. Dividend Policy Under Asymmetric Information, The Journal of Finance, 94, 1985, pp.1039-1051. Ohlson, J.A. A Synthesis of Security Valuation Theory and the Role of Dividends, Cash Flows and Earnings, Contemporary Accounting Research (Spring, 1990), pp.648-647. Ohlson, J.A. Earnings, Book Values and Dividends in Security Valuation, Contemporary Accounting Research (Spring, 1995), pp.661-687. Rees, William P. The Impact of Dividends, Debt and Investment on Valuation Models, Journal of Business Finance and Accounting (24, 1997), pp.1111-1140. Sougiannis, T. The Accounting Based Valuation of Corporate R&D, The Accounting Review (January 1994), pp.44-68. White, H. A Heteroscedasticity-Consistent Covariance Matrix Estimator and a Direct Test of Heteroscedasticity, Econometrica (48, 1990), pp.17-38.
Managerial Finance 70 Exhibit 1 Descriptive Statistics for the Basic Model Variables Mean Standard Skewness Kurtosis Minimum Maximum Deviation P it 47.288 23.777 1.093 2.226 2 170 BV it 25.109 16.249 1.319 2.726-3 113 RE it 3.495 4.638 0.883 23.113-26.252 52.279 DV it 0.066 0.113 0.553 5.012 0.000 1.788 Notes: P is the price per share for the firm at the end of financial year t, BV is the book value of equity per share, RE is the retained earning per share and DV is the dividend per share. The full sample of 356 firm-years consist of all the firms in the Forbes 100 Most Multinational Firms with data available for all the variables in the basic model. Outliers were defined as the top and bottom of 0.5% of cases for each of the basic model variables (BV, RE, and DV). The sample is drawn from the years 1992-1998. P 1.000 Exhibit 2 Correlation Between Explanatory Variables P BV RE DV 0.3899 BV 1.000 0.3563 0.1766 RE 1.000 0.4272 0.3276 0.6995 DV 1.000 Note: Variables are defined in Exhibit 1.
Volume 27 Number 3 2001 71 Total Sample Exhibit 3 Basic Valuation Model Results a b c d e f g F Adjusted R 2 23.322 (17.325)* 1992 25.262 (7.702)* 1993 17.325 (8.725)* 1994 29.368 (9.274)* 1995 32.628 (8.678)* 1996 30.262 (11.324)* 1997 34.526 (11.234)* 1998 32.342 (10.862)* 0.586 (8.625)* 0.325 (5.625)* 0.256 (3.252)* 0.095 (4.262)* 0.325 (3.023)* 0.632 (6.325)* 0.526 (8.324)* 0.542 (7.283)* 0.657 (0.304) 2.608 (0.520) 5.624 (1.068) 1.625 (0.321) 0.385 (0.062) 0.285 (0.023) 0.386 (0.321) 0.292 (0.295) 0.352 (0.258) 1.256 (0.028) 1.862 (0.528) 1.321 (0.625) 0.529 (0.302) 0.632 (0.502) 0.635 (0.208) 0.636 (0.218) 10.562 (3.265)* 11.552 (2.932)* 9.189 (3.812)* 12.054 (2.979)* 9.862 (2.816)* 10.232 (2.916)* 11.231 (2.854)* 12.325 (2.625)* Notes: a. The estimated equation is: P it = a+bios it + CBV it + dre it + eios it RE it + fdv it + gios it DV it +e it b. The variables are defined in Exhibit 1. c. *Significant at 0.01 d. ** Significant at 0.05. e. t-statistics are between parentheses. 62.523 (5.263)* 23.182 (1.878)** 48.023 (3.568)* 68.325 (4.685)* 53.262 (2.812)* 51.321 (2.812)* 48.325 (2.979)* 43.525 (3.682)* -4.701 (-1.93)** -7.582 (-2.856)* -41.703 (-2.386)* -31.526 (-3.705)* -2.392 (-2.479)* -30.625 (-3.859)* -29.682 (-3.256)* -28.523 (-4.262)* 35.728 0.28 18.232* 0.31 18.425* 0.35 18.425* 0.36 12.062* 0.26 13.256* 0.23 11.282* 0.25 15.623* 0.28