Earnings Forecast-Based Returns Predictions: Risk Proxies in Disguise?

This paper develops a novel approach to evaluate the likelihood that omitted risk factors or market mispricing underlies the anomalous security returns to analyst earnings forecast-based attributes, reported in Frankel and Lee (1998) and Elgers, Lo and Pfeiffer (2001). My approach is to incorporate predictable errors in analyst earnings forecasts and use the adjusted forecasts to evaluate the underlying phenomenon. The results show that the adjustments of analyst forecasts improve both the predictive accuracy of the forecasts and the ability of the forecasts to represent investor earnings expectations. These adjusted forecasts, however, do not enable improvements in the abnormal security returns resulting from the trading strategies in the two studies. The inability to show improvements is consistent with the interpretation that the predictable security returns documented are more likely due to omitted risk factors rather than to security mis-pricing.