High-Risk Mechanism Design
DescriptionMechanism design concerns setting up (institutional) rules. In general, agents under a mechanism may have hidden information and take hidden actions that are unverifiable. Potential conflicts of interest require a judiciously designed mechanism to recognize those information asymmetries. Inherent to many mechanism design problems is Risk. From the incentive structure of an organization to government policies that affect consumer welfare, firm competition, or even national security, the outcome of a mechanism can be highly uncertain. Despite significant advances in mechanism design theory in the past half-century, important decisions involving high stakes and high risks are still made in practice without satisfactory guidance: the state-of-the-art mechanism design literature typically assumes that the planner and players are of quasi-linear (or risk-neutral) utility. While facilitating the analysis of various complex problems, this assumption can be overly restrictive when risks entail significant consequences for the involved parties.In this proposed research, we aim to develop a comprehensive theoretical framework that incorporates risk, and economic agents’ heterogeneous risk preferences, into considerations. A well-known difficulty in designing mechanisms beyond the quasilinear environment is to establish suitable equilibrium solutions. We will take a novel, dichotomous approach recognizing the fact that the risky outcome of any nontrivial mechanism typically affects some agents favorably and the others unfavorably, and that each agent’s risk preference can be contingent on these events. Characterizing heterogeneous risk preferences by the powerful tool of log-supermodularity, rather than the conventional Arrow-Pratt measures, the proposed research will contribute a tractable, new approach to addressing issues of risk in mechanism design theory.Along with the theoretical analysis, we will investigate the high-risk, high-stakes mechanisms empirically. We select two types of mechanisms frequently adopted in practice to conduct structural estimations, i.e., asset allocation and innovation-motivating mechanisms. Both of these typically involve high risks and high stakes. To overcome the non-identification characteristic of non-quasilinearity, we will inventively construct the identification sources by combining two separate behavior data of the same group.The joint work will provide an in-depth analysis of how risk and heterogeneous risk preferences affect outcomes of high-risk mechanisms, and how optimal risk sharing can make the planner and players both better off. Our anticipated results will have new and important policy implications for mechanism design problems in the real world.
|Effective start/end date||1/11/19 → …|