Insurance as a lemons market: Coverage denials and pooling

Abstract

The standard monopoly insurance model with adverse selection implies that there are always gains to trade, that only the best (unobservable) risks can go uninsured, and that a profit-maximizing menu cannot pool all types. We show that insurance-provision costs can explain both coverage denials only to those likely to be the worst risks and complete pooling. Specifically, we prove a general comparative statics theorem formalizing coverage denials only to those deemed to be the worst risks; and two theorems showing that the insurer offers a single contract (complete pooling), with either zero or positive coverage. We point out some implications of these results for empirical work on insurance. Our results expand upon a point made by Hendren (2013), that the main effect of adverse selection on insurance might not be misallocation in active markets – the traditional emphasis after Rothschild and Stiglitz (1976) – but simply in shutting down markets, as in Akerlof (1970) classic lemons model.