Contrast Effects: The Phantom of The Latest Forecasts

Abstract

Psychology research on contrast effects suggests that the information from a previous decision may be compared to the information provided for the current judgement task. We document a new stylized fact that an analyst’s current earnings forecast of one firm tends to bias against her latest forecast revision of another firm. Such evidence of contrast effects is further supported by another finding that the current forecast tends to be less accurate when the discrepancy between the current revision and the latest revision is larger. In addition, consistent with previous findings in the psychology literature, the contrast effects are stronger when the two successive forecasts are closer in forecasting time, when the two forecasted firms are more related (e.g., from the same industry), and when the analyst has more experiences. We also test stock market reaction and show that investors discount for contrast effects. Finally, additional robustness tests confirm that our results cannot be explained by alternative effects documented in the prior literature.

Keywords: Contrast Effects, Analysts, Earnings Forecasts, Cognitive Biases, Behavioral Finance.

JEL Classification: G02, G14, G24, G41