No abstract

No abstract

No abstract

We develop a simple measure of volatility based on extreme-day returns and apply it to market returns from 1885 to 2002. Because returns are not normally distributed, the extreme-day measure, which is distribution free, might provide a better measure of stock market risk than the traditional standard deviation. The extreme-day measure more accurately explains investor behavior relative to standard deviation as shown by equity fund flows, and we find evidence that large negative changes appear to influence investor behavior more than large positive changes. 2004 The Southern Finance Association and the Southwestern Finance Association.

This paper reexamines the existence of seasonal anomalies in daily stock prices by integrating seasonal patterns into a single comprehensive model that captures the joint effects of seasonal variations for each of the three major markets. This model incorporates serial correlation and corrects for non‐normality by using robust regression techniques. Serial correlation is found to be important, as is the day of the week and the January variable. Furthermore, the Tuesday after a Monday holiday is significant for two markets using the robust technique (but not ordinary least squares). Finally, the day‐preceding‐a‐holiday effect is strongly significant.