I provide evidence that analysts whose earnings forecast revisions showed signs of greater exaggeration in the past make recommendation changes that lead to lower abnormal returns than their peers. Interpreting stock recommendations as a forecast of future abnormal returns, I show that this...