| dc.description.abstract | This dissertation studies whether hedge funds are proficient at avoiding investing in firms that conduct financial fraud. Using 13F quarterly holdings data from 1980 to 2012, we find that hedge funds have significantly lower ownership (34% less) in fraud firms over the course of violation period, relative to non-fraud firms and non-fraud quarters. Furthermore, less ownership by hedge funds for fraud firms also occurs one year before the fraud is revealed to the public, and hedge funds continue to divest from fraud firms until the pre-revelation quarter. In contrast, non-hedge fund institutional investors do not show any significant pattern in their holdings in fraud firms around violation period.
A panel logistic regression modeling the probability of fraud indicates that both the level and change in hedge fund ownership can predict fraudulent activities. One standard deviation increase in hedge fund ownership will decrease the odds of conducting fraud by 21.7%.
Next, we test the relationship between fraudulent severity and hedge fund holdings. Using total fine as a proxy for fraud severity, we find that, in the cross section, the more severe the fraud is, the less hedge fund ownership is in the firm.
We also develop three alternative measures for hedge fund holdings that represent better monitoring incentive. We demonstrate that hedge funds with greater incentives to monitor the firms show a larger ownership decrease in fraud firms around violation period.
Finally, we show that the implementation of Regulation Fair Disclosure does not affect
hedge funds’ ability to divest from fraud firms around the violation period.
Taken together, our findings suggest that hedge funds are proficient at avoiding investing in fraud firms, while non-hedge fund institutions mostly react to public news. | en |
