---
title: Asymmetric Information and Contracts
url: https://doc.liz6.com/en/general-education/economics-foundations/12-information-and-contracts
locale: en
area: general-education
tags:
- General Education
- Economics Foundations
- General education
- Economics foundations
date: 2026-09-12
modified: 2026-09-12
description: You want to buy a used piece of equipment. The seller knows whether it is prone to frequent breakdowns, but it is difficult for you to confirm this before the purchase. A low price might mean a bargain, or it might mean the seller is eager to offload it; a high price does not automatically prove good quality. The reason trade is hindered is not that there are no potential gains from trade, but that key facts are held by different people.
---
# Asymmetric Information and Contracts
Prerequisites: [Surplus, Taxes, and Subsidies](/en/general-education/economics-foundations/07-surplus-efficiency-and-distribution) · [Strategic Interaction and Cooperation](/en/general-education/economics-foundations/11-strategic-interaction)
You want to buy a used piece of equipment. The seller knows whether it is prone to frequent breakdowns, but it is difficult for you to confirm this before the purchase. A low price might mean a bargain, or it might mean the seller is eager to offload it; a high price does not automatically prove good quality. The reason trade is hindered is not that there are no potential gains from trade, but that key facts are held by different people.
This article treats two problems separately: not knowing the type of the counterparty or the good before the transaction, and being unable to observe the counterparty’s behavior after the transaction. Both fall under information asymmetry, but they require different contracts and evidence.
## Why Good Equipment Might Exit the Market
The market consists of half good equipment and half bad equipment. Buyers value good equipment at 100 and bad equipment at 40; the minimum acceptable price for good equipment sellers is 80, and for bad equipment sellers, it is 30. These numbers are set for teaching purposes; buyers are risk-neutral, and appraisal and transportation costs are temporarily ignored.
If quality could be seen clearly, good equipment could trade between 80 and 100, and bad equipment between 30 and 40, leaving room for gains for both parties. Now, buyers can only see the label "used equipment" and cannot distinguish types. Believing that sellers are evenly split, they are willing to pay at most the average value: `0.5×100+0.5×40=70`.
| Uniform Offer of 70 | Buyer's Expectation | Seller's Actual Choice |
| --- | --- | --- |
| Good Equipment | Believes there is still a 50% chance of buying one | Minimum price is 80, so exits the market |
| Bad Equipment | Believes there is only a 50% chance of buying one | Minimum price is 30, so is willing to sell |
Once only bad equipment remains, the original belief of "half good, half bad" no longer holds. Rational buyers, if they understand the exit process, will be willing to pay at most 40 for the remaining goods. The market may end up trading only bad equipment, and the gains from trade that existed for good equipment are not realized.
This is called adverse selection: trading conditions change who is willing to participate, and the composition of participants in turn changes value and price. This does not mean that low prices inevitably attract bad actors, nor that all markets will disappear completely; the outcome depends on the quality distribution, reservation prices, identification methods, and buyer beliefs.
## What Cost Differences Are Needed to Prove Quality
A statement like "quality guaranteed" cannot distinguish types if anyone can write it up at zero cost. Consider an enforceable repair promise: for the same selling price of 90, the expected warranty cost for a good equipment seller is 5, while for a bad equipment seller, it is 65.
The good equipment seller nets 85, which is above the reservation price of 80; the bad equipment seller nets 25, which is below the reservation price of 30. Under these conditions, good equipment is willing to provide the promise, while bad equipment is unwilling to imitate; the promise can then serve as a credible signal of quality.
Note the conditions: the seller cannot take the money and disappear, the scope of repair is clear, buyers know that the costs borne by the two types of equipment differ, and there are no cheaper ways to fake it. If warranties are not enforced, the actual costs for both types approach zero, and the differentiation mechanism described above immediately fails.
Third-party inspection is another path. Suppose inspection is perfectly accurate and costs 8. For a good equipment unit, the maximum gain for buyers and sellers originally was `100−80=20`; after deducting the fee, there is still 12, which might be worth the inspection; if the inspection fee were 25, it would exceed this achievable gain. Real-world inspections are usually not perfect; one must also compare missed detections, false positives, liability, and re-inspection rules, rather than just looking for a certificate.
## Who Acts in Screening and Signaling
The party with information actively provides verifiable commitments or evidence, often called signaling; the party lacking information designs a menu or conditions, allowing different types to make different choices, often called screening. For example, a buyer offers two contracts: "higher price with strict warranty" and "lower price, sold as-is," observing the seller's choice.
The menu must satisfy the incentive compatibility for each type. If both good and bad sellers prefer the same contract, it has not achieved differentiation. Signals may also waste resources: some costs incurred to prove ability may not increase actual output; analysis must look at both information improvement and resource consumption.
Calling certificates, degrees, or deposits "signals" merely proposes a possible mechanism; it does not prove that it is indeed effective in a specific real-world market. One needs to observe whether different types have different costs, whether information is verifiable, and whether there are other explanations.
## Post-Transaction Behavior: How Insurance Affects Maintenance
Now that the equipment quality is determined, the problem shifts to whether the buyer maintains the equipment carefully after buying insurance. A single equipment breakdown causes a loss of 100; without maintenance, the probability of breakdown is 20%, while maintenance costs 4 and reduces the probability to 5%.
In terms of social resource costs, the expected loss without maintenance is 20; the expected total cost of maintenance is `4+0.05×100=9`. Maintenance saves 11, so in this model, it is worth doing.
However, if insurance fully compensates for damage and premiums are fixed in advance, the buyer does not bear repair losses and cannot be observed as to whether they maintain the equipment, maintenance only costs them an extra 4. Thus, the individual might choose not to maintain it. This is moral hazard: incentives for unobservable behavior change after being insured or entrusted. The term does not constitute a judgment on the moral character of the parties involved.
## How Deductibles Change Behavior and What Costs They Bring
Set a deductible of 40 for damage, with other conditions unchanged. The buyer's expected out-of-pocket cost without maintenance is `0.2×40=8`; with maintenance, it is `4+0.05×40=6`, so they are now willing to maintain it. If the deductible were only 10, the two values would be 2 and 4.5, respectively, and they would still be unwilling to maintain it.
| Rule | Personal Relevant Cost Without Maintenance | Personal Relevant Cost With Maintenance | Personal Choice |
| --- | ---: | ---: | --- |
| Full Compensation | 0 | 4 | No Maintenance |
| Deductible 10 | 2 | 4.5 | No Maintenance |
| Deductible 40 | 8 | 6 | Maintenance |
Fixed premiums are the same under both actions, so they cancel out in this marginal comparison; they still affect whether to buy insurance. Deductibles make individuals bear more of the consequences of their actions, but they also leave risk with the household. If a single breakdown would make it difficult for a low-income family to pay, one cannot claim that a contract is better for everyone just because the average cost is lower.
One can also provide verifiable maintenance services, maintenance discounts, or safety equipment; each method requires monitoring, enforcement, or privacy costs. There is no contract button that makes all actions fully visible and incurs no costs.
## Migrating from Equipment Trade to Work and Platforms
It is difficult to confirm ability before hiring, which is a type information problem; it is difficult to observe effort and collaboration after onboarding, which is an action information problem. Probation periods, portfolios, recommendations, and task assignments provide different kinds of evidence. One cannot claim that all problems are solved just because one mechanism improves one piece of information.
Platform ratings also affect both information and incentives. If only completed counts are rewarded, there may be cherry-picking of simple tasks, ignoring quality, or manufacturing reviews; one needs to ask whether the metrics reflect the goals and whether they can be manipulated at low cost. This connects to the previous lesson on joint collaboration: contracts change the structure of returns, and observation institutions determine whether these returns can be realized.
## Judging Whether a Mechanism Truly Holds
A seller offers a one-year warranty, but the company can deregister immediately after the transaction and bear no responsibility. Can it still distinguish between good and bad equipment based on the calculations above?
Expand Reasoning
It cannot directly use the warranty costs of 5 and 65, as those are the costs when the commitment is enforceable. If everyone can evade responsibility, bad equipment can also imitate, and the signal loses its differentiation power. One should supplement with information on the performing entity, financial guarantees, recourse liability, and actual execution evidence, and then judge why buyers should believe the commitment.
Expand Reasoning
It cannot. High-risk individuals may have been more willing to buy insurance in the first place; this is a selection difference; changes in behavior after insurance is another mechanism. One needs to distinguish between pre-insurance risk and post-insurance actions, specifying the control group and observation period. Both mechanisms can exist simultaneously; one cannot attribute a correlation directly to one of them.