Market for Lemons

The market for lemons is an economic model, published by George Akerlof in 1970, which shows how a market can shrink or collapse when sellers know the quality of what they sell and buyers do not. In American slang a "lemon" is a defective product, especially a car. Because buyers cannot tell good items from bad ones, they will pay only an average price; owners of good items then withdraw, the average quality of what remains falls, and the price falls with it.

Akerlof's paper, "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism", appeared in the Quarterly Journal of Economics after being turned down by several other journals, and is now one of the most cited papers in economics. It is the standard illustration of adverse selection caused by asymmetric information, and it helped earn Akerlof a share of the 2001 Nobel Prize in economics.

Imagine a market stall selling sealed boxes of fruit. Some boxes hold fresh fruit and some hold bruised fruit, but only the sellers know which is which. Shoppers, unable to look inside, offer a price somewhere in the middle. The growers with the fresh boxes feel that price is too low and take their fruit elsewhere.

Now the stall holds more bruised boxes than before, shoppers lower their offers further, and more of the good sellers leave. Step by step, the good fruit disappears, not because nobody wants it, but because nobody can prove it is good.

A simple numerical example, using second-hand phones, shows the mechanism. Suppose half the phones on offer are good and half are faulty, and only the sellers know which is which.

Buyers value every phone more than its owner does, so in a world of full information every phone would be sold. Without information, a risk-neutral buyer will pay at most the expected value:

• $E[V]$ is the expected value of a randomly chosen phone to the buyer.
• $q$ is the share of good phones in the market.
• $V_{G}$ and $V_{F}$ are the buyer's values of a good and a faulty phone.

With $q = 0.5$, the buyer pays $0.5 \times 400 + 0.5 \times 200 = 300$ dollars. But owners of good phones will not sell below $350, so they withdraw. Buyers who understand this realise that only faulty phones remain, so $q$ falls to zero and the price drops to $200. The only equilibrium is one in which nothing but faulty phones trade. Every good phone that could have been sold at a price both sides would accept goes unsold. The loss is a pure inefficiency, a form of market failure.

Akerlof himself applied the model to health insurance for people over 65, to hiring decisions in which employers use crude proxies such as race when they cannot observe skill, and to the high interest rates charged by informal moneylenders in developing countries. Economists have since used it to explain:

• Insurance markets, where people who expect to claim are the most eager buyers, so premiums rise and healthier customers drop out.
• Credit markets, where lenders who cannot tell safe borrowers from risky ones charge a higher average rate, discouraging the safe borrowers, one reason for credit scores and underwriting.
• Securities issuance, where investors may assume that firms choose to sell new shares when insiders believe the stock is overpriced, so an IPO or new share issue is often priced at a discount.
• Financial crises, where uncertainty about which banks or securitised assets are sound can freeze trading altogether, as happened in parts of the asset-backed securities market in 2007–2008, contributing to a credit crunch.

• Warranties and guarantees. A seller who offers a meaningful guarantee signals confidence in quality.
• Brands and reputation. Repeated sales give firms a reason to protect their name; online marketplaces use ratings for the same purpose.
• Third-party certification. Inspections, vehicle history reports, audits and an independent auditor's opinion give buyers verifiable information.
• Law. US states have "lemon laws" for new vehicles that fail repeatedly, and the federal Magnuson-Moss Warranty Act of 1975 sets rules for written consumer warranties. Disclosure rules in securities law serve a similar purpose.
• Escrow and returns. Holding payment in escrow or allowing a chargeback shifts risk back to the seller.

Empirical work on used-car markets has produced mixed results. Some studies find little quality difference between used and new vehicles of similar age, suggesting that warranties, dealers and inspections already limit the problem; others find evidence of adverse selection in specific segments. Critics note that full collapse is rare in practice, because the remedies above are widespread. Supporters reply that this is the point: the institutions that dominate used-goods, insurance and financial markets exist precisely because the lemons problem is real, and the model explains why they look the way they do.

• The Nobel Prize: Economic Sciences 2001