Analysts' Forecasts and Brokerage-Firm Trading

Abstract

Using unique data on brokerage-firm trading volume, I examine whether analysts' earnings forecasts and recommendations generate trading volume for their brokerage firms. I find that analysts' forecasts differing from the consensus forecast generate signif icant brokerage-firm volume in the forecast stocks for two weeks after the forecast release date, consistent with analysts' forecasts affecting their brokers' commission revenue. I find that buy recommendations generate relatively more trading, both buyin g and selling, through the analyst's brokerage firm. I find no evidence that analysts' forecast errors, the difference between forecast earnings and actual earnings, increase brokerage-firm volume. Adding error to their forecasts does not seem to be an ef ficient way for analysts to generate trading. I conclude that analysts are more likely to respond to trading incentives through their recommendations rather than through biasing their forecasts.

Keywords: earnings forecasts, trading incentives, brokerage-firm volume

JEL Classification: G14, G24, G29, M41