نوسانات وابسته به طرز فکری ویژه و حقوق صاحبان سهام بازده : شواهد انگلستان

The proposition that idiosyncratic volatility may matter in asset pricing is currently a topic of research and controversy. Using data from the UK market we examine the predictive ability of various measures of idiosyncratic risk and provide evidence which suggests that: (a) it is the idiosyncratic volatility of small capitalization stocks that matters for asset pricing and (b) that small stocks idiosyncratic volatility predicts the small capitalization premium component of market returns and is unrelated to either the market or the value premium. The predictive power of the aggregate idiosyncratic volatility of small stocks remains intact even after we control for the possible proxying effects of business cycle fluctuations and liquidity and is robust across time and different econometric specifications.

Standard asset pricing models predict that only systematic risk is priced in equilibrium. Accordingly most of the empirical work on the validity of asset pricing models was focused on whether one or multiple systematic factors are incorporated in asset prices and command a risk premium. The possibility that idiosyncratic risk (which in theory can be eliminated through diversification) maybe priced in equilibrium and therefore investors might demand a systematic risk premium to bear it, has been largely neglected2. The recent paper by Campbell, Lettau, Malkiel and Xu (2001) on the statistical properties of idiosyncratic volatility has rekindled interest in the role of idiosyncratic volatility in asset pricing and stock return prediction. Campbell et al. (2001), using monthly data over the 1962–1997 period, show that average idiosyncratic risk is the most important component of total volatility, has increased noticeably over the period (while market volatility shows no significant trend), is countercyclical and helps forecast future economic activity. The possibility that idiosyncratic risk might be priced in equilibrium has been studied in two recent papers. Xu and Malkiel (2001) build an equilibrium model, on the assumption that some investors might not be able to hold the market portfolio, which includes idiosyncratic risk as one of the systematic determinants of future stock returns. They study empirically the cross sectional relation between idiosyncratic volatility and firm returns and find a positive relation between idiosyncratic risk and future returns. Ang, Hodrick, Xing, and Zhang (2006) reach the opposite conclusion; stocks with high idiosyncratic risk deliver abysmally low returns. The other strand of the literature, more relevant for this paper, studies the inter-temporal relationship between lagged aggregate idiosyncratic volatility and market stock returns. Goyal and Santa-Clara (2003) find that the equally weighted stock volatility is a significant predictor of subsequent returns of the value-weighted market portfolio. This result persists even after controlling for other variables that are known to predict future equity returns. Bali, Cakici, Yan and Zhang (2005) argue that the positive relation uncovered by Goyal and Santa-Clara (2003) is not robust across different stock portfolios, disappears if the sample includes the more recent history of returns3 and is partially driven by a liquidity premium. Guo and Savickas (2003a), using quarterly data to measure volatility, reach the conclusion that value-weighted idiosyncratic stock volatility is negatively related to future stock returns. Finally, Brown and Ferreira (2004) create two measures of idiosyncratic risk: one based on large capitalization stocks and another based on small capitalization stocks. Their evidence suggest that only the small capitalization based measure of idiosyncratic risk is significantly and positively related to future returns of the market portfolio as well as portfolios of large and small stocks. The existing evidence on the relationship between idiosyncratic volatility and future stock returns based on US data are conflicting and confusing. On the one hand the evidence by Goyal and Santa-Clara (2003) point to a positive relation between market returns and the lagged equally weighted idiosyncratic volatility. On the other hand, Wei and Zhang (2005) and Bali et al. (2005) argue that the results are driven by small stocks traded on the Nasdaq and disappear in the extended sample, while Guo and Savickas (2003a) find a negative relation between idiosyncratic risk and returns. The evidence in Brown and Ferreira (2004) point to a special role for idiosyncratic volatility based only on small capitalization stocks. The purpose of this paper is to examine the properties and forecasting ability of idiosyncratic volatility of the UK stock market. Given that most of the research on idiosyncratic volatility is based on USA data, using data from another major stock market minimizes the biases that arise due to data snooping (Lo & MacKinlay, 1990) and offers an independent assessment of the empirical findings. In addition to the value-weighted measures of idiosyncratic volatility, we also study idiosyncratic volatility measures based on large and small capitalization stocks. Consistent with the evidence presented in Brown and Ferreira (2004) we also find that idiosyncratic risk based on small capitalization stocks is different from either the value-weighted volatility used in the previous studies or from the volatility measures based on large cap stocks. The idiosyncratic volatility of small capitalization stocks is highly correlated with the aggregate equally weighted idiosyncratic measure used by Goyal and Santa-Clara (2003). The predictive power of the aggregate idiosyncratic volatility of small and large stocks remains intact even after we control for the possible proxying effects of business cycle fluctuations and liquidity. The evidence on the predictive power of idiosyncratic risk (and especially the idiosyncratic volatility of small capitalization stocks) is robust across time and remains significant after we control for possible persistence in idiosyncratic volatilities. The structure of the paper is as follows. The second section describes the measures of idiosyncratic risk, while the third presents the data and summary statistics. Section 4 investigates the forecasting ability of the idiosyncratic risk. Section 5 examines whether the relation between idiosyncratic risk and expected returns is attributed to either business cycle or illiquidity variables. Section 6 performs several robustness tests, while the last section concludes the paper.

The behaviour, properties and pricing of idiosyncratic volatility has become a hot issue in the literature. This is perhaps not surprising given the importance of idiosyncratic risk for portfolio management. Evidence that idiosyncratic volatility is a priced factor in asset pricing and that it can be used to forecast future stock returns are currently a topic of research and debate. This paper uses data fromtheUKmarket to study the properties of average idiosyncratic volatility and its predictive power for future stock market returns. We report convincing evidence that the idiosyncratic volatility of small stocks predicts the small capitalization premium but has no forecasting power for market risk or the value/growth spread. The predictive power of idiosyncratic volatility is unrelated to either business cycle or liquidity variables. These conclusions are robust across time and remain intact after we take into account the possibility that volatility is persistent. The possibility that small stock idiosyncratic volatility is a proxy for non-traded risk like human and entrepreneurial capital or that it reflects the inability of investors to hold undiversified portfolios due to transaction costs, taxes or other institutional restrictions, should be a topic for further research. The evidence presented in this paper suggests that the relevant variable of interest is the idiosyncratic volatility of small capitalization stocks and the research question is why it has predictive power for the small capitalization premium and not the other risk factors