An analysis of new buy and sell recommendations of stocks by security analysts at major U.S. brokerage firms shows significant, systematic discrepancies between prerecommendation prices and eventual values. The initial return at the time of the recommendations is large, even though few recommendations coincide with new public news or provide previously unavailable facts. However, these initial price reactions are incomplete. For buy recommendations, the mean postevent drift is modest (+2.4 percent) and short-lived, but for sell recommendations, the drift is larger (-9.1 percent) and extends for six months. Analysts appear to have market timing and stock picking abilities.

An analysis of new buy and sell recommendations of stocks by security analysts at major U.S. brokerage firms shows significant, systematic discrepancies between prerecommendation prices and eventual values. The initial return at the time of the recommendations is large, even though few recommendations coincide with new public news or provide previously unavailable facts. However, these initial price reactions are incomplete. For buy recommendations, the mean postevent drift is modest (+2.4%) and short‐lived, but for sell recommendations, the drift is larger (−9.1%) and extends for six months. Analysts appear to have market timing and stock picking abilities.

An analysis of stock and sell recommendations by security analysts at fourteen major U.S. brokerage firms shows significant, systematic discrepancies between prices and eventual values, in contrast to the conclusions of many previous studies. The average initial price change at the time of the recommendations is large, even though few recommendations coincide with other public information releases or provide previously unavailable facts. These price changes which are then ostensibly due primarily to information processing and dissemination are consistent with the expanded Grossman and Stiglitz (1980) definition of market efficiency where informed investors earn a return on their information gathering and processing efforts. Furthermore, these initial price reactions are incomplete. Post-recommendation risk-adjusted price drift implies that analysts' recommendations provide tradable value for investors. For recommendations, the mean, post-event drift is modest (+2.4%) and short-lived (one month), but for sell recommendations, the drift is larger (-9.1%), accruing over a longer six-month interval. Analysts appear to have market timing as well as stock picking abilities. Also, accurate industry predictions appear to be an important component for pessimistic recommendations (sell and removed from buy ) but not for optimistic ones (buy and removed from sell).