Doctor of Philosophy, The Ohio State University, 2008, Business Administration

I examine the role of sell-side security analysts in financial markets. The first essay addresses the stylized fact that investors' reaction to stock recommendations is often incomplete so that there is a predictable post-recommendation drift. I investigate whether investor inattention contributes to this drift by using turnover as a proxy for attention. I find that the recommendation drift of firms with low prior turnover is more than double in magnitude compared to that of firms with high prior turnover. Volume reactions around the recommendation show that investors fail to react promptly to recommendations issued on low attention stocks. Together, the evidence suggests that investor inattention is a plausible explanation for investors' underreaction to stock recommendations.The second essay studies conflict of interests in analyst research. Analyst research is alleged to be biased because analysts' employers underwrite securities for the firms covered. I argue that this analyst affiliation bias should be strongest for firms with a desire to over-inflate stock prices. Using stock recommendations data, I find that the analyst affiliation bias is on average pervasive across all firms in the bull-market years of 1994-2000. In the regulatory reform years of 2001-2006, only poorly governed firms, firms whose CEO wealth is highly sensitive to stock price, and negative prior return firms continue to exhibit the affiliation bias while the bias mostly disappears for all other firms. Examining the market's reaction around stock recommendations shows that the market does not sufficiently discount the fact that affiliated analyst optimism is more serious for some firms. The third essay investigates the market's response to trends and reversals in earnings surprises where earnings surprises are defined as firms' reported earnings less analysts' consensus forecasts. Trends are defined as consecutive same-signed earnings surprises while reversals occur when the sign of the mos (open full item for complete abstract)

Committee: René Stulz (Committee Chair); Andrew Karolyi (Committee Member); Kewei Hou (Committee Member); Karl Diether (Committee Member) Subjects: Finance

Doctor of Philosophy, The Ohio State University, 2004, Business Administration

This dissertation studies how psychological and reputational considerations affect the behavior of individual investors and security analysts. The first essay examines investors' preference for framing their gains and losses using trading records of individual investors at a large discount brokerage firm. I find that investors tend to bundle sales of losers on the same day and separate sales of winners over different days. The result is consistent with the principles of mental accounting (Thaler (1985)), according to which individuals attain higher utility by integrating losses and segregating gains. Alternative explanations based on tax-loss selling strategies, margin calls, the number of winners and losers in a portfolio, the difference in the potential proceeds from selling winners and losers, and correlations among winners and losers in a portfolio do not fully account for the observed behavior. Logistic analyses show that investors are more likely to sell multiple stocks when they realize losses, after controlling for various factors including market and portfolio returns, overall sales activity during the day, and investor characteristics. The second essay provides a theoretical and empirical analysis of analysts' incentives to incorporate public information in their earnings forecasts. The model show that analysts may underreact to public news due to their reputational concerns, and that an analyst's incentive to underreact to public information 1) decreases with the size of unexpected news; 2) decreases with the uncertainty of earnings; 3) increases with the analyst's initial reputation; and 4) increases with how much the analyst values his/her current reputation relative to forecast accuracy. I test the implications of the model and find that analysts underreact to earnings news less when the size of unexpected earnings is large, when there is more uncertainty about the earnings, and when they have long track records. The model also implies that the strateg (open full item for complete abstract)

Committee: David Hirshleifer (Advisor) Subjects: Business Administration, General