This dissertation studies the contracting efficiencies among different parties in a firm. The first and second essays explore the impact of agency problems, i.e. the ex-post-settling-up problem and opportunistic overproduction, on CEO compensation contracts. The third essay investigates the effect of nonaudit fees on the auditor-and-client relationship.

Essay I examines the determinants of cross-sectional variation in the asymmetric sensitivity of CEO cash compensation to positive and negative stock returns. Prior research finds that CEO cash pay is more sensitive to bad news than to good news and suggests the cause might be the ex-post-settling-up problem between shareholders and managers, i.e. the difficulty of retrieving the undeserved CEO past cash pay. Consistent with this explanation, I find the asymmetric sensitivity of CEO cash compensation to stock returns is larger when the probability of CEO turnover is higher and when stock returns reflect higher unrealized gains. These results indicate that asymmetric sensitivity of CEO cash compensation is related to the importance of the ex-post-settling-up problem.

Essay II examines whether compensation committees shield CEO compensation from overproduction—a real activity manipulation. Overproduction under absorption costing system enables CEOs to allocate more fixed manufacturing overhead costs to the ending inventory so as to boost current earnings. Measuring overproduction by the difference between earnings under absorption costing and variable costing, I find that in manufacturing firms, compensation committees place a negative incremental weight on earnings in CEO total (i.e. both cash and equity-based) compensation when overproduction is present. This negative weight is larger when overproduction is more likely to be opportunistic, indicating that CEO compensation contracts are efficient in shielding overproduction manipulation.

In Essay III (coauthored with Richard Frankel and Nicole Thorne Jenkins), we investigate whether nonaudit fees are associated with the presence of quai-rents in the auditor-and-client relationship. We use auditor turnover to identify situations of reduced client-specific quasi-rents and examine the relation between auditor turnover and nonaudit fees. We find that higher nonaudit fees are associated with lower auditor turnover, suggesting nonaudit fees are an indicator of quasi-rents. In addition, we show that the magnitude of the negative relation between nonaudit fees and auditor turnover is significantly smaller in the post-SOX period than in the pre-SOX period, consistent with the notion that nonaudit services prohibited by SOX are more closely tied to quasi-rents.