Earnings Surprises that Motivate Analysts to Reduce Average Forecast Error
研究发现，大的或负的盈利意外会促使分析师更努力获取私有信息，从而降低个人预测误差，并因误差的异质性而减少平均预测误差。
Large earnings surprises and negative earnings surprises represent more egregious errors in analysts' earnings forecasts. We find evidence consistent with our expectation that egregious forecast errors motivate analysts to work harder to develop or acquire relatively more private information in an effort to avoid future forecasting failures. Specifically, we find that after large or negative earnings surprises there is a greater reduction in the error in individual analysts' forecasts of future earnings, and these individual forecasts are based more heavily on individual analysts' private information. This increased reliance on private information reduces the error in the mean forecast of upcoming earnings (even after controlling for the effect of reduced error in individual forecasts). As reliance on private information increases, more of each individual forecast error is idiosyncratic, and thus averaged out in the computation of the mean forecast.