I develop a simple model of optimal contract vagueness that allows for renegotiation and builds on the literatures on authority delegation and transaction costs in contracting. The main predictions of the model are that optimal contract vagueness and renegotiation rates depend on the complexity of underlying business relationships and the characteristics of the adjudicators who enforce the contract. In particular, contracts will be vaguer when relationships are more complex and when enforcement institutions are of higher quality. Meanwhile, firms that enforce contracts in higher quality institutions may renegotiate their contracts less often while those with more complex projects always renegotiate them more often. Using new data on material contracts available through the SEC, the joint comparative statics of the model are used to perform an evaluation of public versus private contract enforcement. In particular, the empirical analysis finds that contracts committed to be enforced through private enforcement (i.e., arbitration) are both vaguer and renegotiated less often than those committed to be enforced through public enforcement (i.e., courts). This evidence is consistent with private enforcement having a quality advantage over public enforcement.

This paper investigates whether allocating more control rights to industry groups in customer-firm enforcement proceedings (i.e., the self-regulation of enforcement) leads to greater industry bias and expertise in enforcement. The approach used in this paper focuses on a particular way that bias and expertise can arise in enforcement: through the selection of adjudicators to enforcement panels. Using novel data on customer-firm enforcement run by a specific self-regulatory organization (the National Association of Securities Dealers (NASD)), I document that pro-industry arbitrators are selected to arbitration panels more often than pro-investor ones (selection on bias) and that experts are selected more frequently to cases (selection on expertise). To assess whether the NASD is responsible for these patterns, I examine the impact of a change in regulation that moved much of the control rights over arbitrator selection from the NASD to investors and brokers jointly. Following this change, the allocation of expertise to cases declined dramatically while selection on bias increased. These findings suggest that the NASD is not responsible for selection on bias but that it increases selection on expertise. Thus, I do not find support of industry favoritism at the NASD and find evidence that it adds expertise to enforcement. Moreover, these results suggest that alternative control structures commonly used in enforcement have important drawbacks (e.g., jury selection mechanisms). The extent to which these findings are robust and generalizable is discussed in this paper as well.