Search

This study examines the reaction of security returns to revisions in security analysts' earnings forecast distributions. Two issues are addressed. The first issue is whether changes in distributions of forecasts published by security analysts convey information that is relevant to the valuation of firms' shares of stock. The second issue is the relationship between unsystematic security returns and two information variables that are functions of the parameters of analysts' earnings forecast distributions. To address the first issue, the study examines unsystematic returns, obtained from the market model, that are conditional upon revisions in analysts' earnings forecast distributions. Parametric and nonparametric tests are conducted to examine the relationship between unsystematic returns and revisions in analysts' earnings forecast distributions. Results of the tests conducted to examine the first issue indicate that revisions in security analysts' earnings forecast distributions, on the average, convey information that is relevant to valuing firms in the stock market. Upward revisions in the mean forecast from one month to the next are accompanied by upward changes in stock returns. Downward revisions in the mean forecast are accompanied by downward changes in stock returns. Two information variables are used in the study. The first variable is defined as the percentage change in the mean earnings forecast from one month to the next. The second variable is the change in the mean forecast from one month to the next divided by the st and ard deviation of the first month's earnings forecast. The nonparametric tests indicate that there is a significant positive rank-order correlation between unsystematic returns and each of the information variables. This correlation was observed for downward changes in forecasts but not for upward changes. The results of the parametric tests indicate that there is a significant positive correlation between unsystematic returns and the second information variable. No such relationship was observed with the first information variable. Once again, the significant positive relationship was observed for downward forecast revisions but not for upward forecast revisions.