The Role of Management Forecast Precision in Predicting Management Forecast Error

Abstract

By studying a large sample of annual management earnings forecasts during 1996 to 2006, I document a significant and negative relation between management forecast precision (measured by the width of forecast range) and signed management forecast error (difference between management forecast level and reported EPS). This finding is consistent with the hypothesis that optimistic forecasts are bundled with a low level of precision ex ante. This relation is statistically and economically significant and remains robust to including firm-fixed effects, endogenizing forecast precision, and controlling for self-selection bias. This relation is strongest for long-horizon forecasts and weakest for preannouncements. Finally, this relation appears to be partially understood and incorporated into analyst revision post management forecast. Taken together, my results enrich the literature on management forecast precision and further the understanding of how managers choose forecast characteristics.