CEO Overconfidence, Managerial Earnings Forecast and Feedback

Prior research suggests that many managers are overconfident. Overconfident managers typically attribute positive feedback to their own superior ability but negative feedback to external factors. In this paper I examine how this attribution bias affects managerial forecasting accuracy. I expect overconfident managers to issue optimistic earnings forecasts and to be slow to improve their forecasting accuracy because of the attribution bias. To test these conjectures, I analyze how managers respond to feedback in the form of forecasting errors and associated market reactions. My results show that overconfident managers do learn from feedback, but they do so more slowly than their less confident peers. Further, I document that overconfident managers learn to correct their biased forecasts only when the feedback is less ambiguous in the form of forecasting errors. In contrast, managers who are less confident respond to both forecasting errors and market responses to prior forecasts.