Are Analyst Recommendations Biased? Evidence from Corporate Bankruptcies

We test whether a bias exists in analyst recommendations for firms that file for bankruptcy during 1995–2001. We fail to find overoptimism in analyst recommendations, including those of affiliated analysts. Our multivariate analysis of the market reaction to changes in analyst recommendations indicates that prior affiliation exerts no impact on either returns or trading volume. We find that the market does not view recommendation upgrades by affiliated analysts as biased since there is no price reversal following these recommendation changes. Overall, our results suggest that recently passed legislation to reduce analysts' conflicts of interest might be an overreaction.