Wanda Mimra

Second Opinions in Markets for Expert Services: Experimental Evidence

We experimentally investigate the role of second opinions in markets where experts such as physicians both diagnose and provide the services. Physicians may exploit their informational advantage and overtreat their patients by providing a more costly and expensive treatment than necessary. We show that introducing costly second opinions significantly reduces the level of overtreatment. Lowering search costs leads to significantly more second opinions, but the overtreatment level does not decrease. Under low but not under high search costs, market efficiency rises with the introduction of second opinions, as the reduction in treatment costs due to less overtreatment exceeds the increase in incurred search costs.

Excluding Ringleaders from Leniency Programs

We analyze the impact of ringleader discrimination in leniency programs on the sustainability of collusion. An important role of leniency is to elicit cartel members’ evidence about collusive activity in order to increase the probability of convicting the cartel. We find that ringleader discrimination does not play a role when the ringleader’s relative amount of evidence is high as the collusive agreement will specify and compensate the ringleader to never blow the whistle in any case whereas ringleader discrimination has a pro-collusive (deterring) effect when the ringleader’s relative evidence is moderate (low). Furthermore, ringleader discrimination may deter collusion if the investigation probability of the antitrust authority is sufficiently high but not too high. In this case, the effect of asymmetry from exclusion — even if profits are redistributed — is stronger than the loss of ringleader evidence for conviction.

Endogenous Insolvency in the Rothschild–Stiglitz Model

Even 30 years after Rothschild and Stiglitzs (1976) seminal work on competitive insurance markets with adverse selection, existence and characterization of the equilibrium outcome are still an open issue. We model a basic extension to the Rothschild and Stiglitz (1976) model: we endogenize up-front capital of insurers. Under limited liability, low up-front capital gives rise to an aggregate endogenous insolvency risk, which introduces an externality among customers of an insurer (Faynzilberg, 2006). It is shown that an equilibrium with the second-best efficient Miyazaki–Wilson–Spence allocation always exists.

A Note on Uniqueness in Game-Theoretic Foundations of the Reactive Equilibrium

Riley (1979)’s reactive equilibrium concept addresses the problem of equilibrium existence in competitive markets with adverse selection. The game-theoretic interpretation of the reactive equilibrium concept in Engers and Fernandez (1987) yields the Rothschild and Stiglitz (1976)/Riley (1979) allocation as an equilibrium allocation, however multiplicity of equilibrium emerges. In this note we imbed the reactive equilibrium’s logic in a dynamic market context with active consumers. We show that the Riley/Rothschild–Stiglitz contracts constitute the unique equilibrium allocation in any subgame perfect Nash equilibrium.

Quality provision and reporting when health care services are multi-dimensional and quality signals imperfect

We model competition for a multi-attribute health service where patients observe attribute quality imprecisely before deciding on a provider. High quality in one attribute, e.g. medical quality, is more important for ex post utility than high quality in the other attribute. Providers can shift resources to increase expected quality in some attribute. Patients rationally focus on attributes depending on signal precision and beliefs about the providers’ resource allocations. When signal precision is such that patients focus on the less important attribute, any Perfect Bayesian Nash Equilibrium is inefficient. Increasing signal precision can reduce welfare, as the positive effect of better provider selection is overcompensated by the negative effect that a shift in patient focusing has on provider quality choice. We discuss the providers’ strategic reporting incentives and reporting policies. Under optimal reporting, signals about the important attribute are always published. However, banning reporting on less important attributes might be necessary.