Asymmetric Information, Adverse Selection, and Moral Hazard

When one side knows more

Competitive markets assume that price and contract terms can reflect relevant quality, risk, and behavior. Often one party knows facts the other cannot cheaply verify. Sellers know more about a used car's defects; borrowers know more about their projects; insured people observe more of their daily care.

Asymmetric information exists when participants in a transaction possess different relevant information. The problem is not merely that everyone is uncertain. It is that hidden information or action systematically changes participation, prices, or incentives.

Adverse selection: hidden type before exchange

Adverse selection occurs when privately known characteristics influence who enters a contract. Suppose 70% of riders have a 10% accident probability and 30% have a 35% probability. For a $1,000 covered loss, expected claims are$100 and $350. The population average is$175.

If a pooled premium rises above low-risk customers' willingness to pay, they leave. The remaining pool becomes riskier, so the break-even premium rises again. Some mutually beneficial insurance may disappear even though buyers are risk averse. This is a composition problem, not evidence that insurance caused accidents.

Similar logic can create a “lemons” market. If buyers cannot distinguish good used cars from bad ones, they offer a price based on average quality. Owners of good cars withdraw, lowering average quality and the price buyers will offer.

Moral hazard: hidden action after protection

Moral hazard occurs when a contract changes behavior that is costly to observe or enforce. Full insurance reduces the marginal loss borne by the insured, which may reduce prevention effort. An employment guarantee can alter effort; a lender's downside protection can encourage risk taking.

Moral hazard does not mean people become immoral. It is an incentive effect created by incomplete contracting. Protection remains valuable, so efficient design balances risk sharing against behavior rather than eliminating coverage.

Screening and signaling

Screening is action by the less-informed party to induce types to reveal themselves. An insurer may offer a menu: low premium with high deductible, or high premium with broad coverage. If risk types value these terms differently, self-selection reveals information.

Signaling is action by the informed party to credibly communicate type. A warranty, certification, collateral, or demanding training can be a signal. For separation, the signal must be sufficiently less costly or more valuable for the type making the claim. An expensive signal that conveys no productivity or quality can waste resources.

Worked contract example

Insurance covers a $100 loss with a$20 deductible, so the insurer pays $80 after an accident. Expected claims are$8 for a 10% risk and $28 for a 35At an$18 premium, the low-risk type may decline while the high-risk type enrolls. The insurer then faces expected claims of $28, not the original pooled average.

Raising the deductible to $50 cuts coverage to$50 and expected claims to $5 and$17.50. It can improve sorting and effort incentives but exposes households to more financial risk. The calculation identifies a trade-off; it does not select a fair contract by itself.

Institutions and limits

Reputation, disclosure law, audits, warranties, professional licensing, and third-party certification can support trust. Mandated insurance can prevent low risks from leaving a pool; risk adjustment can compensate plans that attract expensive members. Privacy, discrimination, administrative gaming, and barriers to entry are genuine costs of collecting more information.

Credit scoring illustrates the tension. Prediction can expand lending and price risk, yet biased or opaque data can deny opportunity and make errors hard to appeal. A realistic evaluation asks whether data are relevant, contestable, secure, and better than less intrusive alternatives.

Low risk: 10% accident probability, 70% of riders
High risk: 35% probability, 30% of riders
Covered loss: 1,000
Extra value of risk protection: 40

Contract design can reduce information failures but rarely removes every trade-off. The final section asks the same discipline of public action: policies also operate with limited information, incentives, and administrative capacity.