Abstract: We investigate the potential of conflict management to settle disputes that otherwise escalate to a costly Bayesian game. Players possess private information relevant in the escalation game only. The threat of escalation serves both as an endogenous outside option and as a screening device. We provide an economically intuitive characterisation of optimal conflict management for a general class of two-player Bayesian escalation games. Two features are essential. First, belief management post-escalation is sufficient to solve the conflict-management problem. Second, optimal conflict management maximizes the sum of two simple functions of the information structure in the escalation game, measuring welfare and discrimination. Our results link the classical mechanism design problem of eliciting information to the information design problem of distributing that information. We illustrate our findings by comparing two common escalation games: simple lotteries that call for sorting mechanisms, and contests that advocate type-independence.
Abstract: We study the optimal design of alternative dispute resolution (ADR) mechanism by a third-party mediator. ADR takes place before two litigants face each other in court. Litigation is a legal contest with players who are privately informed about the cost of collecting admissible evidence. Players update their beliefs after the mediation process, but before they decide on evidence collection. Different from standard mechanism design problems, the belief-system post-ADR is important for the outcome of the continuation game: within litigation, choice variables are similar to strategic comple ments and the evidence supplied is driven by the belief system. There is an incentive for parties to misreport in ADR to profit from this deviation in litigation should ADR fail to resolve the conflict.We show that optimal ADR has to break down on-path in some cases to screen the players with respect to their costs. Furthermore, ADR induces truthful reporting by creating post-breakdown beliefs which are independent of type-reports during ADR. To reduce inefficiency vis-à-vis symmetric litigation, optimal ADR induces asymmetric breakdown beliefs even for ex-ante symmetric types to increase the settlement rate compared to symmetric mechanisms. Independent of the set of parameters, ADR achieves settlement for the majority of cases.
Abstract: I study a model in which an informed sender can propose a project to an uninformed receiver. The receiver can accept or reject the project's implementation. If the receiver rejects, the sender can propose a different project to the receiver, which, in turn, may be accepted or rejected. Overall, only one project can be implemented. Both players share preferences about the features of the project. Across projects, preferences are not aligned.
There are two classes of equilibria. There always is a mixed-strategy equilibrium in which the sender panders towards the receiver-preferred project and the sender-preferred project is only implemented with a probability less than one. If the time horizon is sufficiently long a second type of equilibrium exists in which the sender persuades by waiting. In principle, many waiting equilibria exist. The shortest waiting equilibrium is ex-ante Pareto-dominated by the mixed strategy equilibrium, but a waiting equilibrium with intermediate waiting time is preferred by the receiver to all other equilibria.
As an application, I consider a firm that needs clearance of a proposed merger by an anti-trust authority. Merger realizations are private to the firm. Both players prefer higher synergies. The authority prefers high competition, the firm prefers low.