Adverse Selection

Definition

Adverse selection (also called hidden information or hidden knowledge) occurs when an agent privately observes their own type—such as production cost, productivity, or risk class—and can choose whether and how to reveal it. The agent may benefit from misrepresenting their type to the principal. The challenge for the principal is to design contracts that induce truthful revelation without destroying the benefits of trade. Unlike moral hazard (which involves hidden actions), adverse selection involves the agent's information about themselves existing before the contract is signed.

In the Book

The book develops adverse selection as a central case starting in Chapter 2. The baseline model assumes the principal offers a menu of contracts before learning the agent's type: a contract for a high-cost agent and a separate contract for a low-cost agent. Under complete information, the principal could extract all gains by offering each type a contract just meeting their reservation utility. Under adverse selection, the principal cannot distinguish types, so both agents face the same menu. The low-cost agent—being more efficient—will prefer the contract designed for the high-cost agent if that contract pays more; this violates incentive compatibility.

The solution requires the principal to downward distort the inefficient agent's output below its first-best level. This "separates" the types: the efficient agent no longer wants to mimic the inefficient one. The book illustrates this across many domains: in regulation (Baron-Myerson), where a regulator must induce a monopoly to reveal its true cost; in insurance (Ch. 1.6), where an insurer cannot distinguish low-risk from high-risk customers; in labor contracts (Ch. 2.15.5), where a firm may not know a worker's productivity shock; and in financial lending, where lenders must screen borrowers by loan size.

Why It Matters

Adverse selection explains why screening mechanisms exist: loan application processes, medical underwriting, employee testing, and regulatory audits all attempt to separate good types from bad without destroying the value of the relationship. It also explains why efficient types may leave markets entirely—if forced to cross-subsidize inefficient types too heavily, they may not participate. This mechanism reveals a source of inefficiency that is purely informational and structural, not due to moral failing or market friction.