TY - JOUR AU1 - Elton, Edwin J. AU2 - Gruber, Martin J. AU3 - Gultekin, Mustafa AB - Edwin J. Elton, New York University Martin J. Gruber, New York University Mustafa Gultekin, New York University In this paper we explore the characteristics of analysts estimates of earnings per share. We have shown that on average over a wide variety of error measures that analysts errors decline monotonicaly as they prepare successive forecasts of fiscal year earnings. When we decomposed analysts' error we found that analysts were extremely accurate in estimating the average level of earnings for all stocks in our sample. When decomposing by level of aggregation the error in estimating company earnings (with industry error removed) was about 4 times the size of the error due to misestimating the level of industry earnings. Two other decomposition of errors by forecast characteristics were presented in the paper. We next showed that there is persistence in companies that analysts have difficulty in forecasting. If analysts on average have large errors when forecasting the earnings of a company in one year they are likely to have difficulty in the next year. Finally we examined some characteristics of the divergence across analysts in their estimate of a firm's earnings. Analysts tend to have greater divergence of opinion for the first TI - PROFESSIONAL EXPECTATIONS: ACCURACY AND DIAGNOSIS OF ERRORS JF - The Financial Review DO - 10.1111/j.1540-6288.1982.tb00070.x DA - 1982-07-01 UR - https://www.deepdyve.com/lp/wiley/professional-expectations-accuracy-and-diagnosis-of-errors-nMoCAsdDRY SP - 42 EP - 42 VL - 17 IS - 2 DP - DeepDyve ER -