Predicting and detecting the equity premium in Malaysia

Date of Award


Degree Type


Degree Name

Doctor of Philosophy


Faculty of Business and Law

First Advisor

David E. Allen

Second Advisor

Marilyn Clark-Murphy


The empirical findings of Goyal and Welch (2003) and Cochrane (2006) suggest that dividend yields and dividend ratios are robust predictors of annual stock returns and annual equity premia. However, Goyal and Welch (2003) asserted that many researchers considered dividend yields to be a good predictor of the equity premium up to the 1990s. Fama and French (1988) explored equity returns and the premium utilising dividend yields and dividend price ratios as the predictor variables. Currently, more than 200 publications quote this paper. Therefore, this study addresses two questions: (1) Are stock returns and the equity premium predictable using dividend yields and price ratios in Malaysia? (2) And if so, how can investors or policy markers detect and predict the equity premium?

There are numerous studies assessing stock market returns and the equity premium in developed markets but fewer studies for emerging markets. It is interesting to investigate whether the stock market returns and equity premiums are predictable and detectable using dividend price ratios and dividend yields in the Malaysian Stock Market. This will be attempted using data stretching from 1990 to 2007. Furthermore, the research will also investigate whether the equity premium embodied excess returns using the rational bubbles concept.

This study will extend the work by Goyal and Welch (2003) to incorporate the idea of investigating whether rational bubbles exist in stock market returns and the equity premium. This study will use the Duration Dependence Tests through the Log Logistic Hazard Model and Weibull Hazard Model as developed by McQueen and Thorley (1994) and Harman and Zuelkhe (2004). Also, in an extension to the study I employ two asset pricing models including a conditional beta CAPM in applying the duration dependence test.

The general results suggest that the prediction of stock returns and the equity premium shows poor forecasting performance even though they were significant for the insample performances. The Mincer-Zarnowitz regressions suggested that there was no evidence of the predictability of stock returns and the equity premium for two different economic conditions as well as in the overall market using both dividend price ratios and dividend yield models. A treatment for correcting small sample bias as suggested by Lewellen (2004) also has shown an improvement in the results especially the explanatory power but overall the performances are still relatively weak in terms of prediction. In comparison to a study conducted by Cochrane (2008), the results do not show any difference except the dividend price ratios can significantly explain the growth of dividends at a 95% confidence level.

The findings for the duration dependence test on the KLCI show that the dividend yield is a better predictor than dividend price ratios. The results suggest that, the pricing models of Fama and French (1998) and Ferson, Sarkissian and Simin (2008) produce similar results in data from the KLCI. The general results suggest that there is negative duration dependence consistent with the results by McQueen and Thorley (1994).

Access Note

Access to this thesis is restricted to current ECU staff and students only. Email request to library@ecu.edu.au

Access to this thesis is restricted. Please see the Access Note below for access details.


Paper Location