Gender Differences Among Analyst Recommendations

Using a large and unique real-life dataset we study gender dierences in the recommendation issuing process of security analysts. We observe gender heterogeneity in the probability to issue a particular type of recommen- dation. We document that the dierences are most pronounced when the dispersion in existing recommendations is low; male analyst have a larger probability to issue extreme positive recommendations and to deviate from the consensus recommendation, exactly at the time the market could in- terpret this behavior as being skilled. The dierences in opinion between Strong Buy recommendations of male analysts and conservative Hold rec- ommendations of female analysts are almost 30% before 2002, while they decrease to 9% after 2001.