Analysts’ GAAP earnings forecast quality: Implications for research

Abstract

We examine the quality of analysts’ GAAP earnings forecasts and consider implications of GAAP forecast quality for research. We first exploit a tax law change that generates an estimable and material impact on GAAP earnings but not on street earnings in 2017. Analysts generally fail to incorporate this event into their GAAP forecasts yet investors promptly respond to the event, indicating that GAAP forecasts are of low quality and not a good proxy for investor expectations in this setting. Next, using GAAP forecasts spanning 2004 through 2019, we find that low quality GAAP forecasts dampen researchers’ estimates of GAAP ERCs, increase the percentage of firm-quarters classified as meeting-or-beating via exclusions and dampen the market premium for meeting-or-beating, and understate researcher estimates of analysts’ incorporation of exclusions components into GAAP forecasts. We offer two simple empirical strategies to mitigate the adverse effects of low quality GAAP earnings forecasts on research inferences.

Keywords: GAAP earnings forecasts; GAAP earnings response coefficients; meet-or-beat behavior

JEL Classification: G14, M40, M41