Scherbina sau đó mở rộng trên ý tưởng mà các nhà phân tích với dự báo thu nhập cho người nghèo đơn giản là dừng bảo hiểm chứ không phải là đưa ra xấu Cô trình bày bằng chứng thực nghiệm rằng điều này xảy ra. Ngạc nhiên thu nhập trung bình (báo cáo thu nhập quý trừ đi trung bình ước tính các nhà phân tích) | Implications of Short Selling and Divergence of Opinion for Investment Strategy 155 Scherbina later expanded on the idea that analysts with poor forecasts for earnings simply stop coverage rather than put out bad She presents empirical evidence that this happens. The average earnings surprise reported quarterly earnings minus average of the analysts estimates is negative and correlated with the dispersion of opinion. She estimates a measure of bias in earnings for the case where the number of analysts following a stock declines by assuming they would have estimated earnings one cent lower than the lowest estimate. This estimate of bias turns out to highly significant in predicting the earnings surprise. As in previous studies the earnings surprise is related to the past quarter revision in earnings forecasts . analysts do not adjust their estimates as much as they should probably to minimize sharp changes . High market equity-to-book equity . a measure of growth stocks status predicts negative earnings surprise. This means that the analysts overestimated earnings for growth stocks to a larger degree. When examining the abnormal earnings around the announcements of earnings . whether they did better than the average stock during the three days around the announcement she showed that when used in isolation dispersion of earnings had a statistically significant negative effect. This meant that stocks with a high dispersion of opinion about earnings tended to decline in price when the earnings were announced presumably because the announcement reduced some of the dispersion of opinion. As might be expected from the above finding that analysts do not adequately adjust their estimates the previous quarter s revision has a powerful effect on the abnormal returns. If analysts have been revising returns upwards the abnormal returns will be larger. Interestingly when this variable is in the equation the measure of dispersion of opinion remains negative but .
