The Relation between Investor Uncertainty and Market Reactions to Earnings Announcements: Evidence from the Property-Casualty Insurance Industry

Abstract

This paper examines the association between pre-disclosure uncertainty and investor reliance on reported earnings. Our study is unique because it incorporates both context-specific and forecast-based measures of uncertainty. The analysis is based on data for 118 publicly-traded property and casualty insurers for the period 1989 through 1998. We find significantly positive correlations among loss exposure, our context-based measure of uncertainty, and more general measures of uncertainty based on analysts' forecasts. Further, each uncertainty measure considered individually is significantly associated with the magnitude of abnormal returns around subsequent earnings announcements. However, the relation between loss exposure and abnormal returns is nonlinear. Although the market response to earnings generally increases with the firm's exposure to losses in the prior period, that effect is attenuated when the loss is very large. Simultaneous consideration of contextual and forecast-based measures of uncertainty indicates that, once forecast-based measures are taken into account, our contextual measure of uncertainty provides little incremental explanatory power for abnormal returns. The fact that generalized uncertainty measures based on analysts' forecasts apparently outperform a context-specific variable is good news for empirical researchers confronted with samples drawn from diverse industry settings. Taken together, our results provide compelling support for the view that investor reliance on reported earnings increases in periods of heightened uncertainty about the firm's prospects.