Measure for Measure: The Relation between Forecast Accuracy and Recommendation Profitability of Analysts

Authors


  • We thank Brian Cadman, Michael Eames, Rick Johnston, Roger Loh, Ulrike Malmendier, Maureen McNichols, Michael Mikhail, Douglas Skinner (editor), Beverly Walther, Kent Womack, an anonymous referee, and seminar participants at University of Chicago, Duke University, New York University, Ohio State University, Stanford University, Santa Clara University, Washington University, the Financial Economics and Accounting Conference 2005, and Financial Accounting and Reporting Section Conference 2006 for helpful comments. We thank I/B/E/S for providing access to data used in the analysis.

ABSTRACT

We examine the contemporaneous relation between earnings forecast accuracy and recommendation profitability to assess the effectiveness with which analysts translate forecasts into profitable recommendations. We find that, after controlling for expertise, more accurate analysts make more profitable recommendations, albeit only for firms with value-relevant earnings. Next, we show that conflicts of interest from investment banking activities affect the relation between accuracy and profitability. In the case of buy recommendations, more accurate forecasts are associated with more profitable recommendations only for the nonconflicted analysts. For hold recommendations, higher levels of accuracy are associated with higher levels of profitability for conflicted analysts, provided these recommendations are treated as sells. Finally, we find that regulatory reforms aimed at mitigating analyst conflicts of interest appear to have improved the relation between accuracy and profitability. Specifically, the integrity of buy and hold recommendations has improved and the change is more pronounced for analysts expected to be most conflicted.

Ancillary