THE RISK-RETURN RELATIONS OF PROPERTY STOCKS: EVIDENCE FROM FIVE ASIA-PACIFIC COUNTRIES

Abstract

This paper examines the risk-return relations of property stock portfolios in five Asia-Pacific countries (Australia, Hong Kong, Japan, Malaysia and Singapore), from September 1992 to August 2004, using the traditional beta and conditional (Pettegngill et al 1995) beta CAPMs. Given CAPMs popular usage by practitioners, this research seeks to find out if the model is relevant for use in practice, for the general Asia-Pacific region. CAPM's main proposition, that returns are priced only by beta is specifically questioned. The main finding of this research is that though there seems to be no systematic relationship between beta and returns for the unconditional model, when data for each country is tested separately, there exists a strong systematic relationship when the conditional criteria is met; that there is a positive relationship between beta and returns only when the excess market returns are positive and vice versa. Besides, beta is also significantly priced when the datasets of various countries are combined for a pooled regression for both the conditional and unconditional models. Other risk variables (unsystematic risk, skewness, and kurtosis) are also priced in returns, refuting CAPM's proposition that beta is the only risk factor concerned. The conditional test results offer a higher explanatory power on returns, for all significant risk variables, justifying its superiority over the unconditional model. In general, this study did not manage to validate CAPM's various propositions (linear SML, Sharpe-Lintner's unique risk free rate hypothesis etc), but shows at least that beta is justified as an adequate measure of returns in the marketplace, though its predictability powers improve when other risk factors are considered.