Evidence that Analysts Are Not Important Information-Intermediaries

Abstract:     
This study examines whether security analyst earnings forecasts are informative. A widely held view supported by several empirical studies is that security analyst earnings forecasts are informative. We present evidence drawn from more detailed analyses of security returns than used in past studies which shows analyst forecasts are not particularly informative. This finding agrees with the recent finding of Altinkilic and Hansen (2009), that analyst recommendations are not informative. We also show that analysts forecast tend to piggyback on the news and recent events. We examine whether our conclusions also apply in the case of bold forecasts, more accurate forecasts, more timely forecasts, and forecasts from analysts at more reputable brokerages. However, in all cases we find forecast revisions to be information-free. We conclude from the combined findings that security analysts are not information agents in securities markets, contrary to the conventional view.

Keywords: Analysts, Analysts' forecasts, Analysts' under/overreaction to information, Brokerage research, Capital markets, Contrarian strategies, Efficient markets, Financial markets, Information production, Market efficiency, Momentum effects, Post-earnings announcement drift, Sell-side analysts

JEL Classifications: D82, G11, G12, G14, G24, G28, K22, M41

Working Paper Series