Do Analysts Sacrifice Forecast Accuracy for Informativeness?
Pennsylvania State University - Smeal College of Business
Amy X. Sun
University of Houston
University of Illinois at Urbana-Champaign
September 15, 2012
Management Science, Forthcoming
We analyze whether analysts sacrifice forecast accuracy for informativeness by examining: (1) the association between analysts’ deviations from management guidance and earnings management; (2) the effect of the deviations on analyst forecast accuracy; and (3) the effect of the deviations on prices. The evidence indicates that analysts apparently deviate from management guidance to correct for perceived earnings management. Although the deviations reduce the analysts’ forecast accuracy, they improve the informativeness of their earnings estimates. More specifically, they bring the analysts’ estimates closer to the true (unmanaged) earnings number and reduce mispricing. An implicit assumption in the literature is that more accurate analyst forecasts (i.e., earnings estimates that are closer to the actual reported earnings) are better for investors, and that analysts’ primary objective is to forecast the reported (managed) earnings number accurately. Our analysis suggests that this is not necessarily the case and that an inaccurate forecast can actually be more informative than an accurate one. Prior studies on analyst deviations from management guidance focus on analysts’ incentives to provide earnings estimates that managers can beat. These studies implicitly assume that analysts side with management against the interests of their clients. Our analysis indicates that analysts could also deviate from management guidance to provide useful valuation information to their clients.
Keywords: Analyst forecast informativeness, Forecast accuracy, Earnings management, Management forecasts
JEL Classification: G14, M41Accepted Paper Series
Date posted: September 16, 2012
© 2014 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo8 in 0.313 seconds