| This study constructs composite forecasts of annual earnings and evaluates the performance of composite forecasts under different contexts. The composite earnings forecasts are based upon three forecast sources: time-series forecasts, financial analysts' forecasts, and price-based forecasts. The evaluation of performance is based upon the predictive accuracy of composite forecasts versus each of the individual forecast sources used in its construction. The contextual factors examined include (1) smaller vs. larger firms; (2) lower vs. higher analyst-covered firms; and (3) lower vs. higher growth firms. Our results show that composite forecasts are more accurate than each of the individual forecasts in the pooled sample. When the sample is partitioned based upon the contextual factors, we found that the composite forecasts are significantly more accurate than any of the individual forecasts for smaller firms and for firms with lower analyst coverage. Among larger firms and firms with higher analyst coverage, composite forecasts outperform time-series and price-based forecasts, but not analysts' forecasts. |
Updated 07/09/2013