Print
E-MailThe following post comes to us from Zhonglan Dai of the Accounting and Information Management Department at the University of Texas at Dallas, Li Jin of the Finance Unit at Harvard Business School, and Weining Zhang of the Department of Accounting at the National University of Singapore.
In the paper, Risk and Incentive: An Event Study Approach, which was recently made publicly available on SSRN, we take an event study approach to reexamine the standard principal-agent model prediction with respect to executives who have likely experienced an exogenous risk shock. Existing empirical studies of the relationship between risk and incentives provide mixed results, some positive, some negative, and some showing no relationship. We analyze not only the relation between level of pay-performance-sensitivity and firm risk subsequent to a litigation event, but also incremental incentives (changes in pay-performance-sensitivity embedded in executives’ annual compensation) occasioned by a litigation event.
After carefully addressing the endogeneity problem related to firm risk, we find executives’ pay-performance-sensitivity to be strongly negatively related to firm risk, consistent with the standard principal-agent model prediction. Our examination of executives’ annual compensation provides further evidence that boards reduce executives’ incentives by decreasing equity and increasing cash compensation. Taken together, this evidence supports the prediction of efficient contracting theory, that everything else held equal, incentives decrease with risk.
Dividing the event firms into well- and poorly-governed firms, we find the negative relation between incentives and firm risk to be strong for well-governed and much weaker for poorly-governed firms. We further find that poorly-governed firms appear to try, temporarily, to increase the incremental incentives in executives’ annual compensation, possibly to avoid public scrutiny after litigation, but this change is reversed after the lawsuit is dismissed.
The full paper is available for download here.
