Abstract

This thesis systematically explores the relationship between investor holding behavior and stock as well as stockholder characteristics. It formulates a new estimator of the holding period. It examines the data for stocks listed in the Toronto Stock Exchange 300 index over the 1986-1996 period by using two methodologies; namely, regression analysis and screen-sorted portfolios. The holding period is found to be related to four families of characteristics of stock returns based on the regression results. These characteristics are relative price level (value and growth, or in-favor and out-of-favor stocks), past return performance, trade costs and liquidity, and risk. The holding period is strongly affected by stockholder characteristics, especially the proportion of trades effected by large institutional investors. The empirical results support the overreaction hypothesis which assumes that some investors are overly optimistic about firms which have done well in the past, and are overly pessimistic about those that have done poorly. These findings add to the behavioral finance literature. The results support the "trend-chaser" hypothesis which suggests that momentum investors make their investment decisions only on stock return movements, and not on fundamental factors. Stocks with larger spreads or smaller sizes or less risk have longer holding periods. Stocks with a larger percent of institutional trading have shorter holding periods since investors in these stocks trade more frequently. This study makes a strong case for examining stock market and investor behavior together. It suggests that this may best be done by using both the paradigms of modern finance theory and of behavioral finance.