Uncertainty, Skill, and Analysts' Dynamic Forecasting Behavior

We study the risk-taking behavior of stock analysts under varying market conditions. We examine how the risk analysts take by providing bold forecasts that deviate from consensus depends on the degree of uncertainty in the environment as well as the analysts' skill level. We provide evidence that low-skill analysts become significantly bolder, hence taking more risk, and significantly less accurate when market uncertainty increases. These findings are consistent with previous experimental results that show that skill levels moderate differential attitudes towards uncertainty. We further provide evidence that the difference in an analyst's boldness between times of low and high uncertainty is a signal of the analyst's skill that predicts future forecasting accuracy over and above standard skill measures such as past forecasting accuracy