Idiosyncratic risk and expected returns : an investigation in the context of real estate investment in China

Abstract

In the asset-pricing framework, idiosyncratic risk is the risk that is independent of systematic risk and peculiar to one specific asset or company, it is left with no role in expected returns according to the classic finance theory since it could be completely diversified away. However, in the case investors holding under-diversified portfolios, previous theoretical studies generally demonstrate a positive relationship between idiosyncratic risk and expected returns. However, negative empirical evidences regarding the idiosyncratic risk-return tradeoff have been reported recently in the stock market of the U.S. and China, as well as in several real estate literatures. To reconcile the conflict, this thesis is dedicated to investigate the role of idiosyncratic risk in the context of real estate investment. In the theoretical exploration, an asset-pricing model with short-sales restrictions in the market and heterogeneous beliefs among investors is established. Specifically, a simplified version with only three risky assets, in which two of them are direct and indirect real estate investments, demonstrates when investors endowed with incomplete information setting and under-diversified holdings, idiosyncratic risk would play an important role in the expected returns in equilibrium. Furthermore, the comparative static analysis reveals a positive cross-sectional relationship between idiosyncratic risk and expected returns. In the empirical study, this thesis employs the Fama and French (1992) three-factor model to estimate monthly idiosyncratic volatilities of the Listed Property Companies (LPCs) in the A-share market of China, based on the daily data from May 1999 to Aug 2011. Specifically, for each LPC in each month, its idiosyncratic risk is computed as the standard deviation of the three-factor model’s daily residuals. The estimation outputs show that idiosyncratic volatility dominates the LPCs’ overall volatility during the study period, and it is features with a distinct pattern when compared to that of the U.S. REITs: the LPCs’ idiosyncratic volatilities are significantly higher and more persistent; they are less irrelevant to the firm’s market capitalization and present an evident co-movement with the broad market. Hence, this scenario reveals a special interest to further study on the cross-sectional relationship between the LPCs’ idiosyncratic risk and their expected returns. In the cross-sectional test, conditional idiosyncratic volatility forecasted by the EGARCH-GED model is employed as the proxy for expected idiosyncratic risk, as the LPCs’ lagged idiosyncratic risk is shown to be not a good estimate. Over the study period, a firm positive cross-sectional relationship between idiosyncratic risk and expected returns is documented, after controlling for various pricing factors such as firm size and book-to-market equity ratio, indicators of liquidity and momentum as well as returns reversal effect. This evidence not only confirms the prediction of previous theoretical studies and the model in this thesis, it also suggests a profitable trading strategy based on the idiosyncratic risk of the LPCs.