In June 1967, a young Berkeley economist named George Akerlof sent a paper about used cars to The American Economic Review. The editor rejected it because the journal “did not publish papers on subjects of such triviality.” The Review of Economic Studies said the same. The Journal of Political Economy sent two referee reports explaining why he was wrong. In 2001, the work it started won a Nobel Prize.
Information asymmetry exists when one side of a deal knows something relevant that the other side cannot cheaply check. A seller may know a product’s flaws; a buyer may know how they will use it. The gap can distort prices, push good options out of a market, and change behavior after a contract is signed.
The “trivial” paper explained a market problem
Akerlof’s own account of the rejections ends with the Quarterly Journal of Economics accepting “The Market for ‘Lemons’” in 1970. The idea fits in a paragraph. Buyers cannot tell a good used car from a lemon, so they offer a price based on average quality. Owners of good cars find that price too low and keep their cars. The average quality of cars for sale falls, the price falls with it, and more good cars drop out.
The Nobel committee summarized it this way: a market where sellers know more than buyers about quality “can contract into an adverse selection of low-quality products.”
The JPE referees had a fair objection. Eggs of different grades were sorted and sold, so why couldn’t cars be? Their point is the second half of the story. Where quality is cheap to observe, grading solves the problem. Where it is not, markets need other tools. Akerlof’s paper had a section on exactly that, opening: “Numerous institutions arise to counteract the effects of quality uncertainty. One obvious institution is guarantees.”
A gap in knowledge does not mean anyone is lying. A seller can know more simply because they own the car.
Adverse selection: the gap before the deal
Adverse selection concerns hidden information relevant to who or what enters an agreement. The lemons market is the classic case, and the same logic applies when you hire someone.
In a hypothetical service contract, a provider knows whether they have the staff and experience to meet a difficult deadline. The customer sees a polished proposal. Choosing mainly on price can select the provider whose hidden limits make the low bid unrealistic.
The response is to look for evidence linked to the specific requirement: comparable completed work, a bounded demonstration, references that can be checked, or a trial with a clear acceptance condition. A test also has a cost and can exclude suitable participants. A large unpaid demonstration, for example, favors firms that can absorb the expense.
Moral hazard: the gap after the deal
Moral hazard concerns hard-to-observe behavior after an agreement changes who bears the risk. Once the contract is signed, the provider might assign fewer staff than promised. The customer sees only a monthly progress percentage. The question has moved from hidden quality to hidden effort.
| Stage | Relevant hidden item | Possible check |
|---|---|---|
| Before selection | Capability for this task | Comparable evidence or a bounded trial |
| At agreement | What counts as completion | Shared acceptance criteria |
| During execution | Actual progress and changes | Demonstrable milestones and exception reporting |
| At completion | Quality that is costly to inspect | A relevant test, warranty, or independent review |
Demanding more reports is a weak fix. A percentage-complete chart can improve while usable output does not, which is Goodhart’s law at work. Incentives drive much of this, so follow the incentives before writing the contract.
Signals and screens close the gap from opposite sides
Akerlof shared the 2001 prize with Michael Spence and Joseph Stiglitz. Their work names the two main fixes.
A signal comes from the better-informed side. Spence showed when a costly action can credibly carry information; his own example was education as a signal of productivity in job markets. A screen comes from the less-informed side. Stiglitz described insurers offering a menu of contracts, where customers who accept higher deductibles get much lower premiums, so their choices reveal their risk.
| Fix | Who acts | Real example | Fails when |
|---|---|---|---|
| Guarantee | Better-informed seller | Akerlof’s guarantees on consumer durables | The seller cannot or will not honor it |
| Brand name | Seller, over time | Akerlof: brands give buyers “a means of retaliation” | The brand is new or easily bought |
| Signal | Better-informed side | Spence: education in job markets | Low-quality types can fake it cheaply |
| Screen | Less-informed side | Stiglitz: insurance deductible menus | The options do not separate the types |
The test for any signal is whether it is cheaper to produce for someone with the quality than for someone without it. A guarantee passes that test, since only a seller confident in the car can afford to cover repairs. An expensive office usually fails it.
Use this information audit before a deal:
- Name the decision. Which choice will the information change?
- List what each participant knows. Include constraints and intentions on both sides.
- Separate observable facts from claims. Note who can verify each item and at what cost.
- Identify the timing. Is the gap about selection, agreement, or subsequent action?
- Choose the smallest useful check. Define what result would alter the decision.
- Keep residual uncertainty visible. A successful trial does not prove performance under every future condition.
When the fixes backfire
Costly signals that prove nothing. A glossy pitch costs a weak provider as much as a strong one, so it separates nobody.
Verification that drives good people away. If the check costs more than the deal is worth, the best candidates may skip it.
Assuming the market will sort it out. The JPE referees were right about eggs. Grading works when quality is easy to see, and fails when it is hidden until after you buy.
Treating the gap as an accusation. Asking for evidence works better than implying bad faith. The ethical decision-making guide covers the other direction: what the better-informed side owes the other.
Questions
Is information asymmetry the same as lying?
No. Someone can honestly know more because of their role or experience. Deception can exploit the gap, but the gap alone does not prove deception.
Are adverse selection and moral hazard interchangeable?
No. Adverse selection usually concerns hidden characteristics before selection. Moral hazard concerns hard-to-observe actions afterward. Name the mechanism instead of relying on the label.
Can transparency remove the problem completely?
Usually not. Information can be hard to interpret, costly to verify, or unavailable when it matters. Aim to resolve the uncertainties that change the choice, and say what remains unknown. Power dynamics explains why whoever controls information often controls the deal.
Your 5-minute information audit
Deal or decision:
What do they know that I can't cheaply check?
What do I know that they can't?
Gap is about: who I pick / what we agree / what they do after
Signal they could offer that a weak option couldn't fake:
Screen I could set (a choice that reveals their type):
Smallest check worth its cost:Before your next negotiation or hire, paste the filled card into a prompt from Save Effort and ask it to list what the other side probably knows that you do not.