Paying for coverage: Conflict of interest among company-sponsored research firms

The Securities and Exchange Commission has recommended company-sponsored research to address the need for greater analyst following of smaller companies. Anecdotal evidence suggests company-sponsored research is more optimistic than research from traditional brokerage firms (Kalra 2004; Lee and Metaxas 2004; Richardson 2006). Using a sample of companies that are followed by both company-sponsored analysts and traditional analysts, I examine whether the conflict of interest between the company-sponsored research firm and the client company leads to more optimistic recommendations and forecasts.;In the first part of my dissertation, I find that analysts at company-sponsored research firms are more likely to issue favorable recommendations than analysts at traditional brokerage firms. This optimism appears to be a result of company-sponsored analysts upgrading their recommendation more quickly and downgrading their recommendations more slowly than analysts at traditional brokerage firms. Furthermore, recommendations issued by company-sponsored analysts appear to be overly optimistic since portfolios of stock based on their recommendations underperform relative to similar recommendations issued by traditional analysts. Finally, I find that the market does not appear to differentiate between recommendations issued by company-sponsored analysts and recommendations by traditional analysts.;In the second part of my dissertation, I find that forecasts issued by company-sponsored analysts are not significantly different from forecasts issued by traditional analysts. I find neither a significant difference in bias or accuracy between forecasts issued by company-sponsored analysts and forecasts issued by traditional analysts. However, company-sponsored analysts appear to revise their forecasts less often then analysts at traditional brokerage firms. Finally, I find that the market reaction to forecast revisions issued by company-sponsored analyst is not significantly different from zero, but I cannot reject the hypothesis that the market reaction to forecast revisions issued by company-sponsored analysts is the same as the market reaction to forecast revisions issued by traditional analysts.;The results of my dissertation suggest that the relationship between the company-sponsored research firm and the client company influences analysts' recommendations but not forecasts. These findings are consistent with Dechow, Hutton, and Sloan (2000) and Malmendier and Shanthikumar (2007) who suggest that analyst affiliation does not affect near-term EPS forecasts. The different results for company-sponsored recommendations and forecasts may be due to the fact that recommendations are more difficult to objectively evaluate than forecasts, and EPS forecasts tend to be geared toward sophisticated institutional investors who are more likely to recognize the conflict of interest (Malmendier and Shanthikumar 2007a; Mikhail, Walther, and Willis 2007).;I also find that company-sponsored analysts revise their forecasts less often than analysts at traditional brokerage firms. This is consistent with Agrawal and Chen's (2005) theory that commissions from trading volume play an important role in shaping analysts' forecasting behavior. Analysts who are compensated by commission revenue appear to be more willing than company-sponsored analysts who are paid a fixed fee to issue timely revisions that reflect changing expectations about earnings. These findings suggest that company-sponsored research may not be an adequate substitute for research from analysts at traditional brokerage firms.;Keywords. Company-sponsored research, paid-for research, analysts' recommendations, analysts' forecasts, meet or beat earning, forecast revisions...