|dc.description.abstract||This thesis investigates the behavior of investors and analysts participating in the financial market. In the first essay, we examine the effects of local stock returns on antidepressant usage using the Truven Health MarketScan® individual prescription drug data. There are three main findings. First, a one standard deviation decrease in local stock return increases local investors’ antidepressant usage by approximately 0.42 percent (an economic cost of approximately 19 million dollars) in the subsequent weeks than what would have been in the absence of the decrease in stock return; in contrast, a stock price increase has no impact on antidepressant usage. Second, the effect of stock returns on antidepressant usage depends on an array of local socioeconomic characteristics including demographic structure, religiosity, political affiliation, political ideology, and personality traits. Third, local stock return fluctuations have significant effects on certain illnesses including insomnia, peptic ulcers, abdominal pains, and substance abuse which often result from depression. The results are consistent across a variety of robustness checks.
In the second essay, we examine the behavior of analysts. In July 2009, the Global Research Settlement (GRS), which was implemented to mitigate the conflicts of interests between analysts and investors, expired. The GRS mandated that sanctioned banks contract with Independent Research Firms (IRFs) to make independent research available to the banks’ customers. We find that after the GRS expiration, the probability that analysts employed at sanctioned banks to issue positive recommendations increased by 3.3 percent compared to a control group. Our findings show that, after the GRS expiration, sanctioned banks might have become more optimistic and conflicts of interests seem to again threatened the credibility of the research by sanctioned banks. Our paper calls into question the SEC's decision not codify to the GRS into permanent rules.
In the third essay, we examine the performance of analysts from Independent Research Firms (IRFs) and investment banks that cover firms in the financial sector. In particular, we evaluate six aspects of analyst performance: recommendation optimism, recommendation informativeness, earnings forecast optimism, earnings forecast accuracy, target price forecast optimism, and target price forecast accuracy. Using a sample of analyst recommendations and forecasts from 1994 to 2013, we document two important findings. First, compared to investment bank analysts, IRF analysts generally provide less biased, more informative, and more accurate recommendations and forecasts when covering firms in the financial sector. The only exception is the market reaction to pessimistic recommendations. The pessimistic recommendations issued by investment banks outperform those by IRFs. Second, conflicts of interests appear to play a significant role when investment bank analysts cover other bulge bracket investment banks. These results suggest that compared to their counterparts from IRFs, investment bank analysts are susceptible to institutional pressure related to underwriting business when covering firms in the financial sector.||