Contracting in the Presence of Uncertainty

Abstract

This thesis concerns the enforcement of contracts in the presence of uncertainty. Sometimes uncertainty is exogenously given and the agents cannot influence its existence. Frequently, however, there is strategic uncertainty created by the behavior of other agents. Both kinds of uncertainty have important impacts on contracting.<br /> Chapter 1 shows that insurers use uncertainty about auditing strategies to fight insurance fraud. For this purpose, we study a costly state verification model with ambiguity. The insurers abstain from commitment to an auditing strategy, even if commitment is possible without incurring any costs. This contrasts with conventional wisdom, which claims that it is optimal to commit, as the credible announcement of thoroughly auditing claim reports might act as a powerful deterrent to insurance fraud. Yet, empirically it is very unusual for insurers to try to overcome the credibility issue. We prove that it can be optimal for the insurers to maintain the ambiguity and forgo commitment. Thus, strategic ambiguity, i.e., the strategic choice to withhold information about auditing costs and strategies, is an equilibrium outcome. <br /> The second chapter considers legal uncertainty in competition law. I show that legal uncertainty can be welfare-enhancing, if the uncertainty is not too large. As an example, consider Article 101 (TFEU) prohibiting vertical restraints with a block exemption excluding companies with market shares below 30%. There are guidelines available how the relevant market shares are to be calculated. Nevertheless, it is extremely difficult to predict correctly the market share that the competition authorities will determine. In addition, there is uncertainty about the size of the fine that firms have to pay in case of a conviction. This exemplifies legal uncertainty as scrutinized in the chapter. <br /> The third chapter analyzes a principal-agent model, in which the performance measure of the principal is non-verifiable and unobservable by the agent. Instead, the principal has the possibility to communicate with the agent. The communication occurs at the very end of the interaction and there is no repeated interaction. Nevertheless, it is crucial for the agent’s motivation that the principal gives feedback and justifies her evaluation, in particular, in case of bad outcomes. In addition, it is optimal to pool evaluations and to compress wages at the top. These results fit well with empirical observations, like the leniency bias and the centrality bias. Hence, this pattern of evaluations can be understood as a feature of the optimal contract instead of biased behavior. <br /> Corresponding to the distinction between first-order and second-order risk aversion, the fourth chapter defines first-order and second-order ambiguity aversion. With second-order ambiguity aversion, for every ambiguity-averse agent there is an ambiguity-neutral agent so that the set of all improvement directions at an unambiguous endowment is the same for both agents. With first-order ambiguity aversion, in contrast, the set of improvement directions is a strict subset of the improvement directions of an ambiguity-neutral agent. Chapter 4 provides three equivalent definitions for the distinction. For this purpose, I introduce a general ambiguity premium and a notion of reference beliefs of an ambiguity-averse agent. This distinction has direct implications for settings in finance, insurance, and contracting. In particular, I consider the validity of an adapted version of Holmström’s informativeness principle under ambiguity aversion.