2005 Reports

Forecasting skewness in stock returns: Evidence from firm-level data in Tokyo markets

Fujii, Mariko; Takaoka, Makoto

Although it is well known that the market rate of return tends to show negative skewness, we find that the return distribution of individual stocks has shown positive skewness in the two principal Tokyo markets from 1980 to 2001. This is consistent with the results reported for the US market. Positive skewness in the returns is more evident for smaller firms. From our analysis of pooled cross-section data of individual stocks from 1990 to 2001, we find that there is clear evidence of the predictive power in the information up until the current period for the skewness in the following period. Specifically, a higher volatility and skewness in the return distribution of the current period is followed by a higher value of skewness in the following period. Also, the up trends in own returns in the preceding periods may explain the rather negatively skewed return distribution in the following period. Hong and Stein's (2003) model predicts a relation between turnover and skewness. For 1990 to 2001, negative relation to the skewness has been found to hold for smaller cap firms with past de-trended turnover. Furthermore, after the regulatory change in short-selling enacted in February 2002, we find more supportive evidence that high trend-adjusted turnover predicts more negative skewness in returns. Overall, our empirical evidence suggests that conditional skewness is partly explained by previous de-trended trading proxies, however, the results are not conclusive for the data in Tokyo markets.