ECO.5 - Choose Exchange Terms for Private Information and Effort
Type: Method pattern Status: Stable
ECO.5:1 - Problem frame
Use this Method when the terms of exchange depend on quality, circumstances or later action that one party knows or controls and another cannot cheaply verify. Identify each party’s available choices. Assess the existing terms or compare proposals to establish when the desired contribution and participation would be worthwhile.
Return the incentive or selection result the terms support, including a condition that prevents the intended agreement. If different terms are needed, propose them with their mechanism and limits. Use an adequate established purchase directly.
ECO.5:2 - Problem
A price may attract low-quality sellers and exclude more reliable ones. A payment rule can reward a measurable proxy while the desired work is neglected. An unenforceable promise to take care may leave the provider’s returns unchanged. Terms that shift all risk to the provider can also exclude someone capable of the work.
ECO.5:3 - Forces
Information has acquisition and verification costs. Incentives and risk sharing can conflict. A strong warranty or stake can make an offer credible while excluding a good provider without enough capital. Observable outcomes can depend on both effort and circumstances outside the performer’s control.
ECO.5:4 - Solution
ECO.5:4.1 - Locate the private fact or action
Identify the information or action that matters and who can know or control it. Existing quality, intended use or financial condition may be privately known before agreement. Later maintenance, care or effort may be privately chosen. The same exchange can involve both. Evaluate separately whether changed payment terms affect an action and whether a choice reveals an existing fact.
Identify the choices each party can make under the actual terms. They may disclose or withhold information, select an offer, decline, change effort, manipulate a measure, withdraw or seek new terms. Include the alternatives available to the participant, including ones the designer would prefer them not to take.
ECO.5:4.2 - Construct terms from the choices they must change
Start with terms the parties could carry out. These might include payment after an interpretable test, a refund for a defined failure, a deposit at risk, or a choice of services. Establish how the specified event can be observed and who has the means and authority to deliver the consequence. Relate each available choice or event to its payment or service. Use an existing rule directly when its effects meet the need.
For a private action, compare the participant’s returns under each relevant choice. Include ways to improve the measure while neglecting the result. To design different incentives, change a term that alters these relative returns. Check participation against the outside alternative, including the participant’s exposure. A design that is to induce the intended action needs both participation and the relevant choice after entry.
In a risk-neutral model, let b be the payment received regardless of the test, w the extra payment for passing, p the probability of passing under an action and c its cost. Expected return is b + p × w − c. Compare returns for the intended action and its relevant alternatives. The difference in expected bonus equals w times the difference in passing probabilities. Increasing b by the same amount for every action raises the return to entering without changing their order in this model. If risk aversion, wealth or payment timing matters, compare the relevant distributions and financing needs or use a qualified specialist result.
For a private fact, examine the choices of each relevant type of participant. A seller may know the item’s quality, or a buyer its own need. A model can stipulate possible types without identifying the actual participant’s type. For each type, compare the proposed offer, the other available offers and nonparticipation using its costs, benefits and exposure.
A refund that will be paid costs the more failure-prone seller more in expectation. Different service packages can also have different values to buyers. Use a difference of this kind to construct terms when separation is wanted. Check each type’s available choices. If different types would accept the same offer, acceptance alone leaves their types unresolved. The exchange can still be worthwhile under a supported mixed or conditional account. If later actions can alter the result, compare the incentives for those actions as well.
Find terms that meet the participation and choice conditions together. If designing terms, vary the price, contingent payment or service within actual observation and resource limits. Check the result for the other party as well. Indifference establishes a tie. To predict or attribute one of the tied choices, use additional grounds. When no feasible terms satisfy the requirements, report the limit. A further design may change the service, observation, risk allocation or participant. Raising all payments can leave relative incentives unchanged.
Use supported probabilities and valuations for claims about actual participants. For a conditional calculation, state the assumptions. Where the evidence supports only cases or ranges, report what holds across them and what could change the conclusion. Include verification, enforcement, tied-up capital and unpriced effects. A deposit can discourage an unwanted choice while excluding a capable participant unable to bear the exposure.
For an incentive linked to an observed result, establish what the observation reveals about the desired work. If the provider can improve the measure while neglecting the result, reassess the terms or observation. A signal that mainly adds uncontrollable risk can discourage participation. Staged payment can limit exposure to nondelivery while leaving quality unknown. State which consequence the mechanism supports.
ECO.5:4.3 - Make the terms usable
To establish that the terms are usable, check whether the warranty provider can pay, how the acceptance condition is interpreted, whether the stake remains exposed and which parties can agree. Obtain a qualified legal or specialist mechanism-design result where the claim depends on enforceability or a particular guarantee. A comparison under stipulated terms can be reported with that condition open.
Report what the terms imply for the relevant choices and what remains uncertain. If proposing a change, explain how the new terms would alter the comparison. Keep effective existing terms when they meet the need. Seek additional information when its likely value justifies the burden.
ECO.5:5 - Archetypal Grounding
ECO.5:5.1 - Care under an outcome-based payment
Suppose a maintenance provider can choose low care, costing €4 with probability 0.5 of passing the test, or high care, costing €12 with probability 0.9. The provider is risk-neutral, has no other relevant costs and could earn €32 from its next best feasible alternative. The test is functionally valid and cannot be cheaply manipulated under these assumptions.
Let the contract pay a nonnegative base b and a nonnegative bonus w for passing. High care strictly beats low care when (0.9 − 0.5) × w > €12 − €4, so w > €20. It strictly beats the outside option when b + 0.9 × w − €12 > €32. With an available €30 bonus this requires b > €17. A €20 base satisfies both. To assess the whole exchange, also compare the result for the buyer with its feasible alternative.
With b = €20 and w = €30, expected payment less effort cost is €31 for low care and €35 for high care. The €12 gain in expected bonus exceeds the €8 extra cost. High care also beats the €32 outside return. Under a flat €35 payment the provider would net €31 for low care and €23 for high care, both below the outside return.
Suppose the risk-neutral buyer values a pass at €80 and a failure at zero. The test and other costs are the same for both offers. Under the conditional contract, expected benefit is €72 and expected payment €47, leaving €25. A second provider offers the same 0.9 pass probability for a fixed €50, leaving €22. The buyer therefore proposes the conditional contract, which also gives the first provider more than its outside option. If a pass were worth only €40 and idle equipment carried no further loss, either contract would give the buyer negative expected net benefit, so the buyer would decline both.
The comparisons establish the modeled direction of the incentives under the given assumptions. An optimality or general behavioral claim needs further grounds. If the provider can pass the test while leaving the equipment unreliable, the terms reward the wrong result. If variable receipts exceed the provider’s tolerance for risk, reassess participation.
With a €15 cap on the bonus, the largest additional expected reward for high care is 0.4 × €15 = €6. That is less than the €8 additional cost. Raising the common base leaves this difference unchanged under the assumptions. If the buyer still needs the work, it can use the other provider if the complete offer remains acceptable or reconsider the observation, the task or the terms.
ECO.5:5.2 - A warranty whose issuer cannot pay
Two sellers offer the same warranty, but one has no resources to honour a claim and can disappear after sale. The same promise therefore has different expected consequences. A deposit, credible guarantor, continuing relationship or different offer may change the comparison. Assess what would make the payment or service available.
ECO.5:5.3 - Construct a quality offer and test who would accept
Suppose a buyer needs one refurbished component. Sellers know whether their own component has a 0.1 or 0.5 chance of failure during a fixed warranty period. The buyer cannot tell which applies. The model specifies the same verifiable failure, fixed probabilities unaffected by later care, a €40 supply cost, and risk-neutral buyer and sellers. The more reliable seller’s alternative yields €50 net and the less reliable seller’s yields €30. Supply, administration and cash are available with no other relevant costs.
Set a price P and a failure refund R. A trusted administrator can hold R from the price throughout the warranty and refund it if the verifiable failure occurs. The seller can finance the amount held. For a seller with failure probability q, expected net return is P − €40 − q × R. Compare that return with the outside option for that seller.
| Required choice | Comparison | Bound on price |
|---|---|---|
| More reliable seller prefers this offer | P − €40 − 0.1 × R > €50 | P > €90 + 0.1 × R |
| Less reliable seller prefers its outside alternative | P − €40 − 0.5 × R < €30 | P < €70 + 0.5 × R |
Both strict comparisons require R > €50. At R = €50 the two price bounds meet at €95, where both types are indifferent between accepting and their outside option. With a feasible €80 refund, strict separation requires €98 < P < €110. At P = €100 the more reliable seller nets €52 against €50 outside. The less reliable seller would net €20 against €30 outside and declines. Under these assumptions the choices separate the types.
The buyer values a working component at €150 and a failed one at zero. Under the separating offer the buyer’s expected value plus refund less price is 0.9 × €150 + 0.1 × €80 − €100 = €43. A verified alternative with the same 0.1 failure probability costs €120 and carries no refund, giving €15 expected net value. These terms make the proposed exchange preferable to the available alternatives for both the buyer and the more reliable seller. The refund changes who bears the loss. In this model it leaves the selected component’s failure probability at 0.1.
Suppose a refund above €20 cannot be secured. With R = €20, strict acceptance by the reliable seller requires P > €92, while strict rejection by the other requires P < €80. No price meets both. A price that strictly attracts the reliable seller also attracts the other, so acceptance no longer supports the old quality inference. The buyer can use the verified alternative, obtain another credible arrangement or assess an offer for both qualities if their mix can be supported. If the mix is unknown, leave the resulting failure probability open.
Reopen the comparison when a seller’s outside option, ability to finance the retained amount, failure observation or competing offer changes. If only one kind can finance the arrangement, acceptance can reflect access to capital as well as quality. Reconsider the claimed inference before using it to procure the component.
ECO.5:6 - Bias-Annotation
A precise-looking calculation may conceal disputed probabilities, power or a badly chosen proxy. Before attributing a poor outcome to dishonesty, examine which actions the terms made worthwhile and whether the observation tracks the intended result.
ECO.5:7 - Conformance Checklist
Which private fact or later action matters, and what do the terms imply for the participant’s available choices, including nonparticipation? Establish what the specified observation or enforcement arrangement can support. Identify any tie, infeasible term or unresolved condition and the assumptions on which a numerical comparison depends.
ECO.5:8 - Common Anti-Patterns and How to Avoid Them
- Paying for a proxy is treated as paying for the result. Trace how the performer can improve the proxy and whether that supplies the result.
- A written promise is treated as a credible consequence. Recover the means and conditions that make its consequence effective.
- Only incentives are optimized. Retain participation, risk bearing, verification and exclusion costs.
ECO.5:9 - Consequences
Some exchanges become feasible while private facts remain undisclosed. Others may require simpler terms or another participant, or may remain infeasible. Complex terms can increase the work of interpretation and create opportunities to game the rule. Compare that burden with the contribution the terms make possible.
ECO.5:10 - Architectural Rationale
Terms can change a participant’s feasible choices and their relative attractiveness. Acquiring information and changing the consequences of choices are different ways to improve an exchange. Assess their costs in the case at hand. Check the decision to participate and the subsequent choice of action separately.
ECO.5:11 - SoTA-Echoing
Akerlof (1970), §IV identifies guarantees as one possible response to hidden quality. Rothschild and Stiglitz (1976), §I.1, I.3 and I.6 examine how participants select from offered contracts. A contract intended for one type is also available to another, so evaluate both types’ choices. We adapt that comparison to the private-fact case. An insurance-market equilibrium result would require its own applicability grounds in procurement.
Holmström (1979), §2 and concluding remarks examines contract choice subject to participation and the participant’s choice of action, including information and risk sharing. The examples here use finite choices and risk-neutral participants to construct feasible terms. The paper’s risk-averse optimization problem has different assumptions.
An ordinary contract with direct observation can suffice when observation is informative and affordable. When observation is costly or incomplete, compare what feasible terms would induce under the stated conditions. Establish legal enforceability or field performance on the grounds those further claims require. Reassess the affected conclusion when participation, available actions, signal quality or enforcement changes.
ECO.5:12 - Relations
Use ECO.4 for complementary venture commitments and ECO.6 for dependence after investment. Use FPF E.13 if a proxy replaces the intended result and MMP for a needed mathematical model. Applicable law and corporate governance determine who can take a legal or corporate action when the work calls for it.