Brokerage-Firm Trading and Profits Around Recommendation Revision Dates

Abstract

We examine a large and comprehensive data set of brokers' daily transactions on the Stockholm Stock Exchange covering a ten-year period (1997-2006). We show that recommending brokers execute, on average, profitable trades around positive recommendation revision dates. These profits are matched by identical negative profits obtained by the rest of the market (including a group of brokers similar to recommending brokers, which we use as a control group). These results support the idea that positive recommendation revisions are a valuable source of investment information, for which brokers should be compensated. The finding that round-trip brokerage commissions triggered by stock recommendations are of the same order of magnitude as abnormal profits suggests that brokers actually capture almost all the surplus created by their research. We also find that approximately half of the profits we identify are generated by transactions that take place before the recorded recommendation date. This finding, coupled with evidence that recommending brokers' market shares and net trades increase prior to the release of recommendations, suggests the existence of information leakages, tipping or postdated recommendations. Our results also indicate that broker clients do not profit from negative recommendations. The results we document go beyond what could have been inferred by an analysis of the abnormal returns associated to recommendation revisions: with poorly dated recommendations and/or in the presence of contemporaneous public news, abnormal return measures face serious identification problems and may in fact be a poor indicator of the actual profits obtained by those who trade on recommendations.

Keywords: stock recommendations, performance evaluation, information leakages

JEL Classification: G14, G24, J44