PREDICTORS OF VARIATION IN STOCK RETURNS: EVIDENCE FROM MALAYSIAN COMPANY PANEL DATA

We use Malaysian company panel data (1,729 firms/years observations) to examine the effects of beta, size, book-to-market value (BM) ratio, earnings-price (E/P) ratio, dividend yield, payout and leverage on the expected stock returns. Our results are based on the fixed effects regression models as these perform statistically better than the random effects and pooled OLS models. Results of the one-way fixed effects univariate regressions indicate that beta, size, book-to-market value (BM) ratio, earnings-price (E/P) ratio and dividend yield individually play a significant role in explaining variation in stock returns, and payout and leverage do not have statistically significant effect. The explanatory power of size (natural log of market capitalization) is the highest. The one-way fixed firm effects multivariate regression results show that size is persistently a significant dominant variable together with other variables in explaining stock returns. Beta has consistently a positive relation with stock returns by itself and together with other variables. But its explanatory power is less than size and other variables. Contrary to results of previous research, BM ratio is not persistently a significant variable; its significance disappears when we incorporate size and E/P ratio or dividend yield in regressions. The explanatory power of regressions increases considerably when we use the two-way fixed firm and time effects models for estimation. Further, size remains a dominant predictor of stock returns variation, leverage loses its significance and B/M ratio becomes statistically significant when we control for both the firm and time effects.