Analysts’ GAAP earnings forecast quality

Abstract

We examine the quality of analysts’ GAAP earnings forecasts and consider implications for research that uses these forecasts as inputs. We first exploit a setting that allows for a clean identification of GAAP earnings forecast quality. We find that GAAP earnings forecasts generally fail to incorporate a known event with an estimable GAAP earnings impact, especially forecasts issued by analysts who lack GAAP forecasting effort. We also find that GAAP earnings forecasts are not a good proxy for investor expectations of GAAP earnings. Both findings in this setting cast doubt on the quality of GAAP earnings forecasts. We assess the implications of these findings for two research applications. Using data spanning 2004 to 2020, we find that the presence of GAAP earnings forecasts that lack forecasting effort affects research inferences regarding GAAP earnings response coefficients and the prevalence of managers’ meet-or-beat behavior. We illustrate two simple remedies to mitigate the adverse effects of low-quality GAAP earnings forecasts on research inferences.