SCIENCE
Contract Theory
Abstract. Contract theory studies the feasible organization of economic cooperation when participants possess different information, pursue partly conflicting objectives, face uncertainty, and cannot always commit to or enforce every future action. Its fundamental object is the allocation of incentives, information, risk, control, and continuation rights through explicit and implicit agreements. This monograph reconstructs the discipline from transaction-cost economics and asymmetric-information models through moral hazard, screening, mechanism design, incomplete contracts, relational agreements, and empirical contract design. It develops the core optimization and incentive-compatibility conditions, differentiates contractual incompleteness from informational asymmetry and legal nonenforceability, and examines practical applications in labor, finance, regulation, procurement, corporate organization, insurance, digital platforms, and algorithmic agency. Evidence from field experiments, observational studies, and current theoretical research is assessed with attention to identification, sample dependence, and the distinction between welfare and private profit. The final synthesis identifies the conditions under which contractual detail, monitoring, discretion, ownership, or repeated interaction are complementary or competing forms of governance. The contribution is an integrated analytical interpretation of established scholarship; no new theory, model, or priority claim is advanced.
Keywords: contract theory; principal–agent problem; adverse selection; moral hazard; incentive compatibility; participation constraints; risk sharing; multitask incentives; incomplete contracts; residual control rights; hold-up; renegotiation; relational contracts; empirical identification; procurement; corporate governance; algorithmic contracting
1. Subject, Scope, and the Economic Meaning of a Contract
Contract theory is the branch of economics that analyzes how cooperative arrangements are formed and designed when parties differ in preferences, information, capacity, power, and access to enforcement. A contract is an agreement or governance rule that maps at least some observable or verifiable conditions into rights, obligations, transfers, decisions, or continuation outcomes. The economic meaning extends beyond a signed document: a durable relationship may depend on tacit expectations, reputation, organizational authority, and recurring exchanges. The comprehensive approach in Contract Theory (Bolton and Dewatripont, 2004) treats screening, signaling, hidden action, multilayer principal–agent structures, and organizational control as connected problems.
A contract specifies a feasible relationship between conduct and consequences. Its design problem depends on five questions: who knows what, when information arrives, what actions can be observed, what facts a court or other enforcer can verify, and who can credibly commit to which future behavior. The distinction between observation and verification is decisive. An employer may know that care was insufficient yet have no transferable evidence suitable for an adjudicator; a court may verify shipment on time while remaining unable to measure the developmental quality of a service. Moral Hazard and Observability (Holmström, 1979) provides a foundation for analyzing what information belongs in compensation contracts. Contractibility Design (Corrao, Flynn, and Sastry, 2025, revised 2026) adds the cost of constructing an evidentiary environment to the decision about what can be contracted upon.
The central economic problem is the allocation of the gains from trade subject to incentive, information, resource, commitment, and enforcement constraints. The analysis distinguishes allocative efficiency, distribution of surplus, risk-bearing, productive investment, and dynamic stability. These may respond differently to the same contractual clause. A stringent performance bonus may raise measured output and transfer more risk to a worker. A broad discretion clause may permit adaptation but expose one party to opportunistic reinterpretation. Private profitability is not a sufficient statistic for social welfare where externalities or public-service quality are present.
Contract theory relates to microeconomics through optimization, game theory through strategic responses, information economics through private signals, industrial organization through vertical structure, corporate finance through control and claims, labor economics through compensation and employment, and institutional economics through governance. Contract law supplies rules of formation and enforcement; contract economics asks what incentives and allocations those rules make possible. The perspectives interact while preserving distinct objects of analysis.
2. Intellectual Origins and the Development of the Discipline
2.1 Exchange, institutions, and the firm
Economic concern with contractual incentives precedes formal principal–agent theory. Sharecropping, debt, joint-stock corporations, and wage agreements show recurrent tensions between effort, control, financing, and division of returns. The historical narrative in the Academy's scientific background on contract theory (Royal Swedish Academy of Sciences, 2016) connects these earlier questions to the formal contributions of Bengt Holmström and Oliver Hart. The 2016 prize recognized their contributions to contract theory, with Holmström especially associated with incentive design under hidden action and Hart with ownership and incomplete contracting.
Coase's The Nature of the Firm (1937) makes the costs of market transactions central to explaining why economic activity takes place within organizations. An exchange governed by external prices can require search, bargaining, monitoring, adjustment, and enforcement. A firm may substitute authority for repeated market bargaining, but internal governance has its own costs. Williamson's governance of contractual relations (1979) emphasizes asset specificity, uncertainty, and the governance requirements of recurring transactions. In this tradition, the relationship between market and hierarchy is contingent on the characteristics of exchange rather than determined by a universal preference for integration.
2.2 Information, agency, and allocation
Information economics transformed contract analysis by treating unobserved characteristics as economic variables. Akerlof's market for lemons (1970) demonstrates how privately known quality can damage exchange. Spence's Job Market Signaling (1973) analyzes costly signals that inform parties of hidden worker types. Mirrlees's optimum income taxation (1971) shows how redistribution must accommodate privately known earning ability and the behavioral consequences of tax schedules. These contributions establish that an allocation must be evaluated against choices individuals will actually make when they know more than the designer.
In corporate organization, agency costs and ownership structure (Jensen and Meckling, 1976) links delegated control to monitoring, bonding expenditures, residual loss, and financial claims. Holmström's hidden-action incentive framework (1979) formalizes the interaction of imperfect performance measurement and risk. The subsequent literature broadened the unit of analysis from a single worker to many tasks, assets, teams, and repeated decisions.
2.3 Mechanisms, ownership, and relational governance
Mechanism design makes the allocation rule itself an object of choice. Myerson's Optimal Auction Design (1981) demonstrates how allocation and payments can be chosen under private valuations. The bilateral-trade result in Efficient Mechanisms for Bilateral Trading (Myerson and Satterthwaite, 1983) identifies a fundamental limit to simultaneously satisfying key constraints. These papers place the feasibility of truthful revelation at the heart of contract design.
The property-rights approach moves from payment formulas toward residual authority. Grossman and Hart's costs and benefits of ownership (1986) and Hart and Moore's Property Rights and the Nature of the Firm (1990) explain why asset ownership matters when some future uses cannot be specified. Relational contracting then models commitments sustained by valuable future business rather than immediate third-party verification. Relational Contracts and the Theory of the Firm (Baker, Gibbons, and Murphy, 2002) shows why asset integration changes the feasible informal bargain. Historical development thus yields several mutually illuminating research programs rather than a single model that absorbs every institutional setting.
3. Analytical Foundations: Timing, Feasibility, and Welfare
3.1 A general contractual environment
Consider a principal P and an agent A. Before their relationship begins, they may select a transfer rule , an allocation , a division of decision rights , or an ownership arrangement . An agent has information , chooses an action , and faces a stochastic state . Gross output is ; utility and production cost can differ by type. The contract designer chooses from a feasible set , subject to the information available at the contracting date and the consequences an enforcing institution can verify later.
The complete-information, full-commitment benchmark chooses allocations to maximize joint expected surplus, often represented as
subject to physical feasibility and any distributional or legal restrictions. Under transferable utility and absent third-party effects, transfers redistribute surplus without altering its total. The benchmark becomes incomplete when allocation or action choices cannot be controlled directly. Its purpose is analytical: it isolates the welfare cost introduced by particular frictions.
The principal's problem with strategic response is instead
where is an equilibrium action conditional on the contract and the agent's information. A profitable optimum for P can be socially inefficient if the principal internalizes only a fraction of the relationship's costs and benefits. The distinction matters in employment, public regulation, and platform contracts.
3.2 Participation and incentive compatibility
A participation or individual-rationality condition requires an agent to prefer acceptance to a relevant outside option:
An incentive-compatibility condition requires the action, message, or type report intended by the designer to be optimal for the informed participant. For a direct mechanism where type is reported as , Bayesian incentive compatibility is written
Truthful reporting is a behavioral implication of an equilibrium under specified assumptions, not a psychological expectation that people generally tell the truth. Different informational structures produce dominant-strategy or Bayesian constraints, and these are not interchangeable. The ex ante, interim, and ex post interpretations of participation also impose different requirements. The comprehensive principal–agent treatment (Laffont and Martimort, 2002) makes the dependence of solutions on constraints explicit.
3.3 Commitment, contractibility, and state dependence
Contract completeness has several senses. A complete contingent specification describes every relevant future event and prescribed response. A verifiable contract attaches enforceable consequences to information an adjudicator can assess. A fully committed contract binds the parties against future opportunistic renegotiation. An agreement may be detailed while enforcement is weak, or sparse while adaptation remains reliable through governance. Imperfect foresight, expensive drafting, costly verification, limited liability, and deliberate discretion all contribute to observed incompleteness. Incomplete contracts under uncertainty (Wang, 2024) illustrates how differences between partners, private information, and investment incentives can change the efficient degree of contractual detail in a particular formal environment.
A temporal sequence clarifies causation: nature draws types; contracting parties offer and accept agreements; one or more agents make investments or exert effort; information and uncertainty resolve; verifiable events are presented; enforcement, transfers, or renegotiation occur; and future interaction may continue. Moving the information arrival, investment, or bargaining dates can reverse optimal ownership and contract terms. Every comparative-static conclusion is therefore conditional on the declared timing.
4. Private Information, Adverse Selection, Screening, and Signaling
4.1 The economic mechanism of adverse selection
Adverse selection arises when a party privately knows relevant characteristics before contracting and the agreement changes who selects into the exchange. In Akerlof's quality-uncertainty analysis (1970), buyers unable to identify quality may offer prices that make high-quality sellers withdraw, changing the average quality remaining in the market. This mechanism depends on the distribution of types, reservation values, and informational asymmetry; market collapse is one possible outcome, not a universal theorem for every market with private information.
Screening reverses the direction of the informational task: an uninformed designer offers a menu of terms, allowing informed agents to reveal their types by their selections. Consider two suppliers with marginal costs . A contract menu lists allocation-transfer pairs and . If supplier utility is , the two incentive constraints are
Adding them implies : under this single-crossing structure, the low-cost supplier receives at least as much output. When the high-cost participation constraint and low-cost incentive constraint bind, the low-cost supplier receives an information rent tied to the cost difference and the high-cost quantity. The exact rent depends on the model's quasilinear utility, two-type assumption, outside options, and other constraints. These are illustrative screening calculations, not a claim that all procurement contracts exhibit this monotonic ordering.
4.2 Information rents and distortion
If a principal offered a separate complete-information contract to each type, the type's private knowledge would become irrelevant to resource allocation. Hidden information introduces rents because a high-productivity or low-cost agent can imitate a less desirable type. One common second-best response is to distort allocations for lower-value or higher-cost types to reduce attractive mimicking opportunities. The presence, direction, and magnitude of distortion depend on the objective and type distribution. An allocation that maximizes profit for a monopolistic principal need not maximize weighted total welfare.
The public-policy analogue is the Mirrlees income-tax problem (1971). Tax schedules depend on realized earnings rather than inherent earning capacity, and higher-ability agents can adjust their labor supply and reported earnings in response. Although taxation is not a voluntary bilateral contract in the everyday legal sense, its self-selection constraints share the logic of incentive-compatible design. Regulation similarly uses cost reports and performance observables. Using Cost Observation to Regulate Firms (Laffont and Tirole, 1986) characterizes incentive arrangements involving private cost information and investment in cost reduction.
4.3 Signaling and credible information
Signaling concerns actions taken by informed participants to alter others' beliefs before a transaction or allocation. In Job Market Signaling (Spence, 1973), an educational credential can serve as a signal when acquiring it is differentially costly across worker types. Separation requires that the incentive to obtain the signal differs sufficiently by type. A costly signal can improve informational sorting while also consuming resources and potentially producing privately valuable but socially wasteful expenditure. Its welfare significance depends on whether education also raises productivity, an issue outside the pure signaling benchmark.
Screening and signaling can coexist. A lender posts a menu of contracts, and borrowers choose terms and provide information; a prospective employee acquires credentials and chooses whether to apply to particular wages. The analyst must identify which side moves first, which information can be verified, and whether the signal conveys information beyond other observables.
5. Moral Hazard, Monitoring, and Optimal Risk Sharing
5.1 Hidden action versus hidden information
Moral hazard in the economic agency sense arises when actions taken after an agreement affect outcomes but are not perfectly contractible. An insurer can observe a claim without perfectly observing precautions; an investor can observe quarterly revenue but only imperfectly reconstruct managerial effort. Incentives then rely on outcomes correlated with actions. An observable noisy result is not identical to an observable action, and the statistical informativeness of an outcome matters for contract choice. Holmström's observability theorem and agency analysis (1979) locates this problem in the structure of performance information.
The first-best arrangement can condition payment on effort or directly prescribe effort when it is enforceable. Under hidden effort, the wage-output rule induces a best response. Higher sensitivity to output can increase effort while imposing greater risk on a risk-averse agent. Monitoring improves incentives if it supplies information sufficiently informative about effort and can be used at reasonable cost. A new metric can be informationally redundant when conditional on existing data; collecting more data is therefore not automatically valuable.
5.2 A linear-normal illustration
Consider output , where and . The agent bears effort cost , receives wage , and has constant absolute risk aversion with coefficient . With normally distributed noise, the certainty-equivalent utility is
The agent chooses for an interior optimum. If the principal is risk neutral, cannot observe effort, and makes a take-it-or-leave-it offer satisfying participation at reservation certainty equivalent , expected profit after the participation constraint binds reduces to
The interior optimum is
The first-best effort level with contractible effort is when gross output has unit marginal product. Thus the optimal incentive coefficient falls as risk aversion, noise variance, or the effort-cost curvature rises, under the assumptions just stated. If the agent is risk neutral and can bear losses without limited liability, approaches one. Where wage floors, limited liability, wealth constraints, nonnormal shocks, or nonlinear mechanisms matter, the solution changes. This example reconstructs standard agency intuition explained in Incentives in Principal-Agent Relationships (Sappington, 1991) and is an illustrative derivation rather than a new model.
5.3 Sufficient performance statistics and evaluation
If two performance measures are available, a contract should reward the information they contribute about the hidden action, not their superficial correlation with gross revenue. A relative-performance metric can filter common shocks when coworkers face shared disturbances and their actions are sufficiently independent. Such metrics may conversely undermine cooperation if an agent's reward falls as a colleague performs better. Good governance assesses informational content, controllability, manipulation costs, and effects on unmeasured tasks. In Social Preferences and the Response to Incentives (Bandiera, Barankay, and Rasul, 2005), workers' observed behavior differs under relative incentives and piece rates, revealing that strategic social interaction changes the response to payment systems.
6. Multitask Agency, Teams, and Performance Measurement
6.1 Why one metric can distort an organization
Most productive jobs comprise multiple activities with unequal measurability. A teacher develops reasoning alongside test performance; an insurer rewards legitimate service and controls unwarranted claims; a procurement contractor balances cost, durability, safety, and future maintenance. If compensation responds strongly to only one visible dimension, effort may shift away from valuable but poorly measured components. Holmström and Milgrom's Multitask Principal–Agent Analyses (1991) establishes this organizational consequence under a formal model with interacting tasks. An apparently weak bonus may be optimal when stronger incentives would degrade unmeasured output.
Represent tasks by an effort vector , a gross-value vector , a measured-performance vector , and a convex effort-cost function . Under a linear reward , an interior risk-neutral agent satisfies
If an economically important component of lies outside the span of the columns that performance signals can detect, no vector of measured-output rewards can directly implement the full first-best effort mix. For a risk-averse agent, the covariance matrix of measurement noise also enters incentive design. Job design can separate tasks or combine their supervision, and fixed compensation may protect effort allocation when measurement is systematically incomplete.
6.2 Teams and interdependence
A team produces joint output through several agents whose individual contributions are costly to observe. Shared-output bonuses risk free riding, while forced relative ranking may encourage sabotage or reduced cooperation. Peer monitoring, joint ownership, repeated interaction, and team reputations can supplement direct pay incentives. The strength of each mechanism depends on observability between peers, the enforceability of sanctions, the production technology, and workers' outside opportunities. Analysis of teams belongs with the governance of authority and not solely with the algebra of bonus rates.
The managerial layer creates further agency problems: managers allocate tasks and attention, decide whom to hire, and evaluate subordinate performance. Incentives for Managers and Inequality Among Workers (Bandiera, Barankay, and Rasul, 2007) finds in a firm-level experiment that incentives to managers changed both mean and dispersion of workers' productivity. A theoretically successful contract for managers can therefore have distributional implications across employees; average output alone cannot characterize its effects.
7. Mechanism Design and the Limits of Contractual Implementation
7.1 The revelation approach
Mechanism design generalizes a contract into rules that elicit reports and determine allocations and transfers. Under appropriate conditions, an equilibrium outcome of a mechanism can be represented through a direct-revelation mechanism that is truthful in the corresponding equilibrium sense. The revelation principle simplifies search across mechanisms; it does not remove incentive constraints or eliminate the cost of collecting and verifying messages. Auction design offers a clear example. Optimal Auction Design (Myerson, 1981) characterizes revenue-maximizing mechanisms given a seller's informational disadvantage about buyers' values, subject to appropriate assumptions about valuation distributions and feasibility.
Consider a bidder with value choosing report , allocation probability , and expected payment . Truthful dominant-strategy incentives require for all reports . Under the regular single-parameter envelope assumptions, implementability imposes monotonic allocation and a corresponding payment identity. The model shows how an apparently simple allocation rule can encode strategic constraints.
7.2 Bilateral trade and impossibility
An informed seller with cost and an informed buyer with value would exchange whenever under full information and zero transaction costs. The bilateral-trade impossibility (Myerson and Satterthwaite, 1983) shows that, with independent private values having overlapping supports and the paper's regularity conditions, an incentive-compatible, individually rational, budget-balanced mechanism cannot generally implement ex post efficient trade in every state. The result identifies a binding incompatibility among constraints rather than proof that all trade under asymmetric information is inefficient. Outside subsidies, changed participation requirements, restricted type distributions, or stronger information can alter the feasible set.
7.3 Regulation and delegated procurement
A regulator frequently designs an agreement for a firm that knows more about its own cost. Fixed-price payments create strong cost incentives but expose firms to risk and informational rent; cost-plus arrangements insure suppliers against realized expenses while weakening their incentives to economize. Using Cost Observation to Regulate Firms (Laffont and Tirole, 1986) shows how accounting observations and cost information can be incorporated into optimal incentive contracts in their model. The general lesson is conditional: a best schedule varies with the cost distribution, risk-bearing, investment, regulatory objectives, and the credibility of audits and sanctions.
8. Incomplete Contracts, Ownership, and Residual Control
8.1 The economic role of incompleteness
Future transactions often contain decisions too costly or difficult to describe and verify in advance. An asset can support uses not foreseen when the relationship begins; technological change can alter the meaning of performance; and quality can be intelligible to the parties without being provable to a court. The incomplete-contract perspective asks who will decide in these contingencies and how that authority affects investments made beforehand. Grossman and Hart's residual-rights theory of ownership (1986) treats ownership as control of uses of assets not specifically allocated by contract. It highlights a genuine trade-off: concentrating control raises one party's incentives while reducing those of another.
In the property-rights account of firm boundaries (Hart and Moore, 1990), ownership affects who can access productive assets and make personnel decisions after bargaining deteriorates. The institutional identity of the firm matters because bargaining positions and investment incentives depend on which assets can be withheld, reorganized, or redeployed. The framework assumes particular noncontractible investments and a bargaining structure. It should not be reduced to a universal statement that more ownership concentration produces better performance.
8.2 A hold-up illustration
Suppose one party must invest at cost before the partner can obtain a relationship-specific surplus . First-best investment maximizes and satisfies at an interior optimum. If the investor expects only a share of the marginal surplus after renegotiation, private choice instead satisfies . Under standard concavity and convexity conditions this yields underinvestment relative to first best. The location of bargaining power and the outside options determine , and the result may change when the investment directly benefits the investor or when contracts affect disagreement payoffs.
The choice of remedy can matter. Holdups and Standard Breach Remedies (Edlin and Reichelstein, 1996) analyzes noncontingent contracts under renegotiation and finds cases where specific performance or expectation damages support efficient investment, with results dependent on what can be contracted and the separability conditions. Conversely Cooperative Investments and the Value of Contracting (Che and Hausch, 1999) demonstrates a setting where cooperative investments that chiefly benefit the partner can eliminate the value of ex ante contracting when renegotiation cannot be prohibited. These results are complementary warnings against treating the hold-up effect as a mechanically universal prediction.
8.3 Authority and the economic boundaries of firms
Corporate integration can transfer residual rights but can also change relational bargaining, information flows, internal incentives, and access to markets. This is why transaction-cost governance and property-rights models explain different aspects of the same organizational problem. The former focuses on comparative governance costs and maladaptation; the latter formalizes bargaining and investment consequences of ownership. A disciplined empirical analysis specifies which observable variable measures a proposed mechanism—asset specificity, renegotiation, authority transfer, or another channel—and what counterfactual organizational form is being compared.
9. Renegotiation, Commitment, and Dynamic Incentives
9.1 Why renegotiation changes ex ante behavior
The value of a promise rests partly on what happens after new information becomes available. The parties may freely renegotiate an earlier agreement because a later allocation could make both better off. This ex post efficiency can undermine incentives that relied on the original agreement remaining in force. A performance penalty that both parties will later prefer to waive has limited deterrent power if the agent anticipates renegotiation. A promised reward may be similarly unreliable if an employer with unilateral control can retrospectively change the criteria. Consequently, renegotiation-proof contracting is a stronger requirement than writing clauses that are optimal at the initial date.
The results depend on which terms are transferable and whether a credible commitment technology exists. Long-term debt covenants, ownership of collateral, staged financing, termination rights, independent dispute resolution, and escalation clauses can change later bargaining positions. In incomplete financial contracting (Aghion and Bolton, 1992), state-contingent control rights help align investor and entrepreneur decisions when monetary incentives alone cannot coordinate objectives. A choice of control regime is thus an ex ante response to anticipated ex post conflict.
9.2 Repetition and promised continuation values
If the same parties anticipate a continuing relationship, future gains can support promises that cannot be directly enforced in court. The simplest repeated-game condition compares today's deviation benefit with the discounted loss of future cooperation. If deviation yields an immediate gain and sustaining the relationship creates a continuation-value advantage measured in a common discounted payoff unit, self-enforcement requires . If instead denotes a constant per-period payoff advantage that begins in the following period, the condition becomes
where is the common discount factor and the continuation benchmark is credible. A more patient party, a larger relationship-specific surplus, or a greater reputational loss expands the set of enforceable informal arrangements. Market exit, short horizons, changes in identity, and financial distress shrink it. Dynamic self-enforcement therefore requires an institutional environment in which threatened responses are themselves sequentially rational.
9.3 Relational contracts and formal governance
Relational Incentive Contracts (Levin, 2003) characterizes important self-enforcing compensation structures under hidden action and information. relational contracts within and between firms (Baker, Gibbons, and Murphy, 2002) shows that vertical integration affects temptation to renege, thereby changing the set of agreements that can be sustained informally. Formal ownership and informal cooperation are jointly determined: the strength of one mechanism can increase or diminish the viability of the other, depending on the pattern of temptations and outside options.
Delegation is also relational. Informal Authority in Organizations (Baker, Gibbons, and Murphy, 1999) explains how a superior who retains formal veto power may sustain effective subordinates' discretion through implicit promises. Authority is thus both a legal allocation of decision rights and an equilibrium behavior. Measured formal decision rights alone can miss the working constitution of an organization.
10. Behavioral Foundations, Fairness, and Experimental Contracting
10.1 Preferences beyond monetary payoff
Many baseline incentive models use expected monetary utility and stable outside options. Actual workers and trading partners may care about recognition, equitable treatment, reciprocity, identity, and the meaning of contractual promises. These preferences can reshape participation and equilibrium behavior. The economic task is to identify how they enter utility or beliefs rather than to presume that every departure from a money-only benchmark is an error. The experimental results in Contracts as Reference Points—Experimental Evidence (Fehr, Hart, and Zehnder, 2011) provide evidence that flexible terms can generate aggrievement and reduced performance compared with rigid agreements in the studied laboratory environment.
Hart and Moore's Contracts as Reference Points (2008) interprets agreements as shaping parties' entitlements; performance shading can occur when an outcome is permitted by a flexible contract but experienced as unfair. This approach illuminates why flexibility, normally valued as a response to uncertainty, may have a hidden relational cost. It does not imply that rigid terms always dominate flexible ones. Rigidity reduces legitimate adaptation, and both costs must be compared under the relevant distribution of shocks and contractual expectations.
10.2 Experimental tests and the boundaries of inference
A laboratory study creates controlled variation and permits close observation of behavior. It offers high internal clarity for a defined mechanism while leaving external validity dependent on stakes, experience, market structure, and social context. Putting Relational Contract Theory to the Test (Erkal, Wu, and Roe, 2025) reports experimental support for self-enforcement and some predicted preferences for informal agreements when verification is coarse, alongside departures from predicted discount-factor responses and punishment choices. Models must be tested at the level of specific comparative statics instead of treated as all-or-nothing propositions.
A cross-country online experiment in Material Incentives and Effort Choice (Davies and Fafchamps, 2025) finds that individual self-interest, conditional reciprocity, and intrinsic motivation help explain different participants' choices under stylized employment contracts. The study measures behavior in an online task and its recruited samples; it does not justify treating nationalities as fixed preference types. The general analytical implication is that contract design can depend on the social interpretation of work as well as its monetary consequences.
10.3 Endogenous perceptions of breach
A party's sense that a promise has been violated can depend on which expectations the agreement created. A contract with open-ended performance standards may distribute interpretive discretion as well as monetary rights. Explicit review procedures, prior negotiations, communication, and third-party adjudication can reduce conflicts about meaning, while persistent ambiguity may sometimes preserve flexibility. The best policy balances evidence and adaptability in the particular transaction, a distinction taken up by recent research on designing contractibility itself.
11. Labor Contracts, Personnel Economics, and Organizations
11.1 Wages as incentive and risk-sharing devices
The structure of a wage contract determines how income responds to output, time, team performance, and employment continuation. A fixed salary insures the worker against some output volatility but may weaken direct effort incentives. Piece rates reward measured units and transmit production risk. Deferred compensation and promotion incentives shift rewards over time, but they depend on enforceable or credible expectations about evaluation and retention. Labor markets add worker mobility, career concerns, human-capital accumulation, monitoring, and workplace norms to the basic agency problem.
The organizational consequences of measured-task incentives require special attention. A bonus based solely on sales can encourage high volume at the expense of suitability or customer service. A fixed salary may improve cooperation when product quality is difficult to measure, as shown analytically by multitask job-design theory (Holmström and Milgrom, 1991). Incentive pay should therefore be assessed against a multidimensional objective rather than by its effect on one indicator.
11.2 Field evidence: output and selection
In the Safelite automobile-glass study, Performance Pay and Productivity (Lazear, 2000) reports a 44 percent increase in output per worker after a transition to piece rates. The observed total includes both behavioral responses and workforce-composition effects; the precise causal decomposition depends on the study's design, timing, and econometric assumptions. It is therefore misleading to report the entire 44 percent as effort induced for unchanged workers. The result establishes that substantial effects are possible in a setting where individual production can be measured.
Piece Rates, Fixed Wages and Incentives (Shearer, 2004) uses a field experiment in tree planting, finding approximately 20 percent higher productivity under piece rates in experimental conditions. The structural extrapolation is separately model-dependent; it should not be confused with the experimental estimate. The contrast between these two studies illustrates why a quasi-experimental firm transition and randomized pay assignment offer different evidence about incentives, selection, and heterogeneous effects.
Social Preferences and Personnel Data (Bandiera, Barankay, and Rasul, 2005) finds that the nature of pay externalities and coworker monitoring affects productivity. The companion managerial incentive experiment (2007) shows that incentivizing supervisors can alter which workers receive management attention and who remains employed. Thus a policy that improves firm output may redistribute benefits or opportunities unevenly. Contract evaluation should track the full relevant outcome vector, including quality, retention, worker well-being, selection, and distribution.
11.3 Organizational authority and career governance
Employment arrangements combine formal job descriptions with a residual domain of managerial instructions. The firm can reassign tasks when technologies or demand change, while workers protect themselves through exit, professional norms, union arrangements, and external legal rights. The optimal degree of discretion depends on the productivity of local knowledge, risk of abuse, the costs of detailed specification, and the durability of the relationship. The employment relation is therefore a governance institution, not merely a price paid for one unit of labor.
12. Corporate Finance, Credit, and the Allocation of Control
12.1 Financial claims as incentive arrangements
An entrepreneur seeking external funds often controls operational decisions that investors cannot observe perfectly. A lender's fixed claim can encourage repayment but create risk shifting or underinvestment; an equity holder bears residual profit variability and may seek voting rights or monitoring access. The classic theory of agency costs (Jensen and Meckling, 1976) links outside financing to the costs generated by divergence between managers and claimants. Ownership and financial structure affect both incentives and the allocation of uncertainty.
A contract can specify interest, maturity, collateral, covenants, reporting, events of default, voting rights, conversion rights, and control shifts. Their economic roles are distinct: collateral affects recovery and bargaining leverage; covenants condition allowable decisions; repayment priority allocates downside exposure; control transfer can alter future investment strategy. In financial contracting with incomplete contracts (Aghion and Bolton, 1992), the allocation of control plays a role when the parties have distinct pecuniary and nonpecuniary objectives.
12.2 Commitment and dynamic lending
Credit agreements are dynamic: borrower type, repayment prospects, and economic conditions evolve. Information about risk can be asymmetrically distributed between incumbents and entrants. Data Sharing with Dynamic Contracts (Dovis and Martellini, 2024, revised 2026) studies disclosure choices when the incumbent learns persistent consumer type and competitors have different information. Its theoretical conclusions differ sharply according to whether an incumbent can commit to long-term contracts, illustrating the joint importance of information flow and credible promises. The work remains a working paper; its conclusions should be interpreted as model results rather than measured average market effects.
12.3 Risk sharing and insurance
Insurance pools losses across states, but coverage can modify insured behavior and selection. A deductible reduces the insured's consumption protection while preserving a marginal personal cost of certain claims. A coinsurance rate changes how risk and costs are distributed between buyer and insurer. The contracting problem includes actuarial information, claims verification, exclusions, waiting periods, underwriting, and repeated renewal decisions. The insurer may have superior statistical data even as the policyholder has superior private information about individual precautions. In such bilateral informational environments, simple labels of informed principal and uninformed agent can conceal the complexity of the contract.
Insurance also illustrates that the best measurable outcome may be an imperfect proxy for desirable action. Lower observed claims may reflect prevention, underreporting, selection, or risk differences. A credible evaluation separates these mechanisms through randomization, exogenous changes, panel evidence, or structurally justified identification. The Academy's contract-theory review (2016) treats insurance alongside labor and credit as a domain where risk sharing must be reconciled with behavioral incentives.
13. Public Procurement, Regulation, and Service Provision
13.1 Designing a procurement agreement
Government and corporate buyers must specify deliverables while suppliers have superior knowledge about costs, methods, and possible adaptations. A fixed-price contract can transfer cost overrun risk and sharpen incentives to economize; a cost-reimbursement agreement can adapt to uncertain scope at the price of weaker cost discipline. The appropriate choice depends on whether quantity and quality are verifiable, whether contractors can price uncertainty, how many bidders participate, and how often work orders change.
In Auctions Versus Negotiations in Procurement (Bajari, McMillan, and Tadelis, 2009), data from private-sector construction in Northern California are used to analyze how complexity and incomplete design relate to procurement choice. The authors identify possible limits of auctions when design information requires communication with suppliers. The evidence is observational: procurement method is selected in response to project attributes, and causal interpretation requires acknowledging that endogeneity.
13.2 Adaptation costs and renegotiation
A low initial bid need not minimize the final resource cost of a project. Where a contractor expects incomplete specifications, its initial bid can anticipate later changes, bargaining, or compensation adjustments. Bidding for Incomplete Contracts (Bajari, Houghton, and Tadelis, 2014) studies highway paving agreements and estimates adaptation costs corresponding to 7.5–14 percent of the winning bid under the authors' structural model. This is a result from a particular sample and estimation approach, not a universal percentage for public procurement. Its importance is conceptual and empirical: adaptation can be economically large enough that ignoring it biases the comparison of procurement systems.
The legal enforceability of change orders, the definition of acceptable quality, and independent documentation requirements affect supplier incentives. Contractibility Design (Corrao, Flynn, and Sastry, 2025, revised 2026) studies the costs of establishing proof that a promise was breached. The paper distinguishes front-end design costs from back-end evidence costs and obtains different implications for the coarseness of contracts. A procurement manager must therefore evaluate not only how many clauses to write but also how evidence about compliance will be produced.
13.3 Outsourcing, public values, and noncontractible quality
The division between public and private service provision adds citizen welfare and political accountability. The Proper Scope of Government (Hart, Shleifer, and Vishny, 1997) demonstrates in a model applied to prisons that private operators can have strong incentives to reduce cost but may reduce quality when its socially important components are difficult to contract. The paper is a theory with an application, not a randomized evaluation of prison outcomes. It gives a conditional argument about privatization when contracting on quality is incomplete; actual policy evaluation requires institution-specific evidence and value judgments about the relevant outcomes.
Procurement efficiency should be defined through total life-cycle costs and quality-adjusted social outcomes. The budgeted payment is only one component. Procurement contracts can influence the allocation of safety risks, environmental costs, maintenance obligations, innovation incentives, and dispute-resolution costs. Their welfare appraisal necessarily draws on public economics, engineering, administrative institutions, and empirical causal analysis.
14. Contract Enforcement, Law, and Institutional Capacity
14.1 Incentives depend on credible enforcement
A promise affects behavior when parties anticipate meaningful consequences for performance and breach. Legal enforceability depends on applicable law, admissible evidence, available remedies, procedure, insolvency, and the effectiveness of implementation. A clause may be legally valid yet costly to litigate; an unwritten promise may be impossible to adjudicate while strongly sustained by reputation. Economics therefore models enforcement as a real resource and information problem. It is necessary to distinguish the substantive legal rule from the expected practical probability and cost of applying it.
Remedies can alter investment and renegotiation. breach remedies and investment (Edlin and Reichelstein, 1996) show that specific performance and expectation damages have different implications under modeled trade, investment, and renegotiation conditions. A legal rule that improves compensation after breach may weaken precontractual incentives if its economic effects on effort, verification, and bargaining are ignored. These claims are analytical and depend on jurisdiction, the contract's available remedies, and the paper's formal setting; they do not prescribe a legal outcome for a particular dispute.
14.2 Institutional differences across jurisdictions
Contract theory is global in scope, but legal institutions differ in procedure, rights, mandatory terms, and treatment of good faith or damages. Formal models often abstract from these differences by assuming a verifiable signal leads automatically to a transfer. Comparative institutional research must restore adjudication costs, delay, corruption risks, state capacity, and access to effective enforcement. The presence of arbitration or digital signatures does not by itself establish that all desired contingencies are verifiable or enforceable.
Employment, financial, consumer, and public procurement contracts may be subject to mandatory safeguards. A model of unrestricted transfers can be inappropriate when wages cannot be negative, insurance terms must be disclosed, or public officials must follow bidding rules. These are substantive feasibility constraints. A sound empirical application states the legal regime and the direction of the constraint before generalizing a mathematical solution.
14.3 Remedies as technologies for governing adaptation
Parties sometimes benefit from standard forms, option clauses, staged renegotiation, third-party expert determinations, or arbitration, because such institutions lower the cost of adapting to unforeseen circumstances. Yet standardization may reduce bargaining expenses and simultaneously produce a mismatch in highly specialized exchanges. Contract design therefore involves a comparison of institutional technologies for resolution rather than simply counting clauses or preferring maximal discretion. The appropriate choice is contingent on asset specificity, dispute frequency, evidentiary cost, and the predictability of change.
15. Identification, Measurement, and the Empirical Science of Contracts
15.1 Why observing a contract is not observing its effect
Empirical contracts are endogenous. A firm offering large commissions may employ different workers, sell different products, or operate under different management conditions than one paying salaries. A comparison of output across those two firms can confound incentives with worker ability, selection, product complexity, market shocks, and other factors. The scientific review of contract theory (Royal Swedish Academy of Sciences, 2016) specifically emphasizes the difficulty of causal inference when contracts are selected for reasons related to performance.
Randomization can assign incentive regimes independently of potential outcomes within the study's population. A natural experiment may use a policy change, contractual threshold, or external shock that changes agreements without simultaneously changing unobserved outcomes through another channel. A difference-in-differences design requires credible parallel trends in the absence of treatment; an instrumental variable requires exclusion and relevance conditions; a structural approach identifies parameters only under its behavioral, distributional, and equilibrium assumptions. These methods answer different questions and face different threats.
15.2 Labor experiments and external validity
The studies of tree-planting piece rates (Shearer, 2004), automotive-glass performance pay (Lazear, 2000), and managerial incentive experiments (Bandiera, Barankay, and Rasul, 2007) use different designs and evaluate distinct margins. Their outcomes cannot be averaged mechanically into one general effect of performance pay. A consistent interpretation first identifies treatment, comparison group, outcome, selection process, and whether measured quality and longer-run consequences are included. A firm-specific estimate describes the causal or model-based effect of that change in its setting, not an invariant human productivity parameter.
15.3 Contract evidence beyond payroll data
Research on organizational boundaries commonly observes firm ownership and productivity but not the counterfactual quality of hypothetical alternative contracts. Procurement data contain bids, modifications, completion dates, and cost overruns but often omit private bargaining expectations. Credit data record interest rates, defaults, and covenants but may not measure unobserved borrower effort. Researchers can improve identification by combining institutional change with granular contract terms, transaction timing, quality data, and matched information about both parties. Reliable interpretation requires distinguishing what is measured from what the theory says is causally important.
15.4 Beliefs, misforecasting, and heterogeneous contract choice
The October 2026 working paper Beliefs Over Contracts (Annan and Raymond, 2026) combines elicited managerial beliefs with a large-scale market-level field experiment comparing expenditure-equivalent incentive designs. The authors report that their best-performing treatment increases weekly firm profit by more than 55 percent relative to the status quo while managers systematically underestimate the best-performing option. The result concerns the study's contracts, firms, and randomization unit; it is not evidence that a similar gain follows from contract redesign generally. Its conceptual significance is the distinction between the objectively measured effectiveness of a contract and a decision-maker's forecast of its consequences. Contract optimization may be constrained by managerial beliefs even where better information is obtainable.
16. Digital Contracts, Platforms, and Algorithmic Agency
16.1 Automation changes the information and commitment environment
Digital infrastructure can lower the cost of recording transactions, measuring events, publishing rules, and executing conditional payments. It can also increase the speed of response, the scale of monitoring, and the asymmetry of information between platforms and participants. The economic question is what becomes observable, verifiable, or commit-able at the margin. Technical automation does not ensure that every relevant performance dimension is measurable or that a digital record has the same meaning in different institutional contexts.
The model of Data, Competition, and Digital Platforms (Bergemann and Bonatti, 2024) studies how platform information and targeted matching affect seller competition, quality, consumer choices, and bargaining advantages. Data access is an allocation variable with distributional implications. A platform may improve matching efficiency and simultaneously obtain strategic power through informational advantages. Consequently, data-sharing policy can affect both contractual performance and market structure.
16.2 Contractibility as a designed production process
The research in Contractibility Design (Corrao, Flynn, and Sastry, 2025, revised 2026) provides a contemporary extension of the classic observability-versus-verifiability problem. Organizations can invest in recordkeeping, quality assurance, audit systems, procedural protocols, and evidentiary standards. These systems create a production function for enforceability: costly evidence changes the feasible menu of contractual clauses. The model's result that front-end and back-end costs can produce different contractual structures underscores the need to separate drafting expenses from future proof-production costs.
Automated monitoring produces another challenge: metrics can be accurate conditional on data while misaligned with the principal's actual objectives. A machine can optimize a measurable quantity without knowing the institutional significance of unrecorded harms or exceptional cases. Contract theory provides tools to identify that incentive problem but does not substitute for the design of accountability, appeals, and legal duties.
16.3 Software agents, information, and delegated choice
Autonomous software agents may search for products, recommend terms, negotiate, and execute transactions on a user's behalf. The NBER chapter The Coasean Singularity? (Shahidi et al., 2025) explores how these capabilities may lower transaction costs while creating new frictions such as congestion and price obfuscation. Its argument is a research-oriented examination of future economic arrangements, not an established historical measurement of universal cost savings.
Economic principal–agent concepts remain relevant when a person delegates tasks to software, but a software system's objective function, training, control rights, error modes, and institutional liability are not automatically equivalent to a human agent's preferences. Contractual governance should specify authority to transact, permissible data use, limits on irreversible actions, logging, recourse, and responsibility. A delegated artificial system may affect outcomes across several firms, making the principal–agent relationship multilateral and technologically mediated. Such systems also introduce questions of market design and technical security outside the canonical bilateral model.
16.4 Algorithmic management and workplace institutions
The OECD's algorithmic management survey (2025) reports adoption and management perceptions across France, Germany, Italy, Japan, Spain, and the United States. It also identifies accountability, transparency, and worker-health concerns. Survey responses describe organizational practices and beliefs; their descriptive prevalence does not establish the causal effect of algorithmic management on welfare. Combined with multitask theory, they suggest that improved measurement can make pay more responsive while leaving questions about metric quality, surveillance burdens, dispute procedures, and the distribution of authority unresolved.
17. Contemporary Research Frontiers and Outstanding Questions
17.1 Optimal contractual detail is endogenous
In many baseline models the set of verifiable variables is given. Contemporary research asks how an organization selects its measurement and documentation infrastructure before writing the agreement. contractibility as a choice variable (Corrao, Flynn, and Sastry, 2025, revised 2026) and contractual incompleteness under disparity (Wang, 2024) provide distinct explanations of why detail may vary across otherwise similar transactions. Their assumptions and objective functions differ; they support a research agenda rather than a universal rule that incomplete contracts are efficient.
17.2 Dynamic data rights and market structure
Data Sharing with Dynamic Contracts (Dovis and Martellini, 2024, revised 2026) adds intertemporal commitment to the design of information disclosure. A market participant may benefit from hiding or releasing information depending on whether it can make credible long-term commitments and how competitors screen types. platform information and competition (Bergemann and Bonatti, 2024) similarly connects data governance with surplus division and market access. Digital contracting increasingly requires joint analysis of private information, privacy, dynamic competition, and enforceability.
17.3 Experimentation and practical contract choice
The Beliefs Over Contracts study (Annan and Raymond, 2026) suggests an empirical distinction between feasible gains and organizations' capacity to recognize them. This creates questions about whether managers use experiments, whether performance metrics are trusted, and whether adoption costs or organizational politics prevent implementation of successful contracts. Research should examine repeated deployment, long-term outcomes, quality measures, and responses from participants who were not represented in the initial experiment. The study remains a working paper, and its preliminary conclusions should be tested by independent replications and further settings.
17.4 Persistent theoretical limits
Several research tensions remain open. How should contracts allocate climate and systemic risks when shocks are correlated and private markets cannot fully insure them? What mechanism implements socially valuable quality when no single observable measure captures it? How can competition limit the power generated by platform control over transaction data? How do relational agreements survive identity changes and transnational digital exchange? What forms of governance work when a principal is itself an organization with multiple constituencies? These are distinct families of research questions. Their resolution requires specifying information and institutional constraints before advancing generalized predictions.
18. An Integrated Analytical Synthesis
The preceding literature supports an integrated interpretation organized around the actual sequence of contractual governance. First, parties choose whether to transact and whom to engage; selection reflects available information and outside options. Second, they allocate tasks, ownership, discretion, payments, and monitoring effort. Third, agents respond with actions and investments, subject to private information and risk. Fourth, the relationship confronts unforeseen developments and evidentiary limits; the parties use courts, renegotiation, authority, or future business to settle outcomes. Fifth, the consequences feed into subsequent choices, organizational learning, reputation, and market structure. This sequence is a synthesis of established literature for exposition, not a newly established theory or formal model.
The first analytical conclusion is that more detailed contracts are valuable only when the specified contingencies are worth their preparation and enforcement costs. Complete contracting is a benchmark, while observed contracting selects among drafting, evidence production, discretion, and adaptation technologies. transaction-cost reasoning (Coase, 1937), residual-rights analysis (Grossman and Hart, 1986), and evidentiary design (Corrao, Flynn, and Sastry, 2025, revised 2026) illuminate different stages of this choice.
The second conclusion is that informational improvement can have competing effects. Better measurement may strengthen incentives and improve the allocation of risk; it may also distort attention toward measured tasks, permit exclusionary screening, or amplify platform bargaining power. informative performance contracts (Holmström, 1979), multitask agency (Holmström and Milgrom, 1991), and data-mediated platform markets (Bergemann and Bonatti, 2024) expose different consequences of greater informational control. The direction of welfare change depends on who receives the information, how it is used, and which outcomes are excluded from the metric.
The third conclusion concerns commitment and adaptation. Long-term rules may protect investments and support cooperation; they can also obstruct appropriate revision after a shock. Flexibility can preserve valuable options, but it may weaken incentives or create uncertainty about expected treatment. cooperative investment contracting (Che and Hausch, 1999), reference-point contracts (Hart and Moore, 2008), and relational governance (Baker, Gibbons, and Murphy, 2002) reveal conditions under which the same contractual flexibility has different economic consequences.
The fourth conclusion is methodological. Theoretical feasibility, optimal design within a model, and causal success in an institution are different epistemic objects. A mathematical optimum can fail when participation restrictions are misdescribed; an experiment can estimate a strong local effect while leaving broader adoption uncertain; an observational association may measure selection into contracts instead of their effect. The distinction is visible in bilateral trading limitations (Myerson and Satterthwaite, 1983), randomized compensation evidence (Shearer, 2004), and structural procurement estimates (Bajari, Houghton, and Tadelis, 2014). The evidence must be interpreted at the level actually warranted by its method.
The fifth conclusion concerns distribution. Contract design assigns bargaining power, risk, decision authority, and information rents. These allocations are integral to institutional outcomes. A profit-maximizing contract may transfer risk to a party less able to bear it, and a cost-saving arrangement may impose costs on third parties. public-service contracting (Hart, Shleifer, and Vishny, 1997) and managerial incentives and inequality (Bandiera, Barankay, and Rasul, 2007) show why measurement of aggregate efficiency and distributional welfare must be separated. Institutional evaluation is complete only when the relevant affected parties and outcomes are included.
Taken together, these conclusions define a disciplined decision procedure for research. Identify the parties and timing; state private information and observable events; specify enforceability and commitment; solve incentive and participation constraints; describe surplus and risk allocation; account for noncontractible margins; identify the empirical counterfactual; and assess external validity and distribution. This procedure is an organizational synthesis of research practice, not a new doctrine. Its value lies in preventing conclusions from one contractual environment from being carried into another with incompatible assumptions.
Conclusion
Contract theory provides a rigorous framework for understanding cooperation under divided interests, private information, imperfect observation, and limited commitment. Its core results explain why theoretically efficient trades may be blocked by informational constraints; why incentive pay trades effort against risk and can redirect attention away from essential but unmeasured activities; why ownership and decision rights influence investment when future states are difficult to contract upon; and why reputation, renegotiation, legal remedies, and relational governance change what promises can achieve.
The discipline has a cumulative analytical architecture. Transaction-cost reasoning identifies why governance arrangements are needed. Screening and mechanism design determine the allocations compatible with private information. Moral-hazard models connect compensation to effort and risk. Multitask theory integrates the neglected margins of production. Property-rights and incomplete-contract approaches explain how assets and control shape investment. Relational contracting shows how future cooperation can sustain commitments beyond immediate legal enforcement. Empirical research reveals that measured effects depend on institutional context, treatment selection, and the structure of outcomes. Recent research extends these questions to evidence production, data governance, managerial beliefs, and technologically mediated delegation.
An economically sound contract is evaluated against the feasible alternatives available to its participants and the society affected by its operation. Its design depends on information, measurement, control, adaptation, and the institutions that give commitments practical force. Contract theory remains a foundational component of economics precisely because these constraints persist while the forms of production, finance, public governance, and digital exchange continue to evolve.
References
Aghion, P., & Bolton, P. (1992). An incomplete contracts approach to financial contracting. The Review of Economic Studies, 59(3), 473–494.
Akerlof, G. A. (1970). The market for “lemons”: Quality uncertainty and the market mechanism. The Quarterly Journal of Economics, 84(3), 488–500.
Annan, F., & Raymond, C. B. (2026). Beliefs over contracts (NBER Working Paper No. 35788). National Bureau of Economic Research.
Bajari, P., Houghton, S., & Tadelis, S. (2014). Bidding for incomplete contracts: An empirical analysis of adaptation costs. American Economic Review, 104(4), 1288–1319.
Bajari, P., McMillan, R., & Tadelis, S. (2009). Auctions versus negotiations in procurement: An empirical analysis. The Journal of Law, Economics, & Organization, 25(2), 372–399.
Baker, G., Gibbons, R., & Murphy, K. J. (1999). Informal authority in organizations. The Journal of Law, Economics, & Organization, 15(1), 56–73.
Baker, G., Gibbons, R., & Murphy, K. J. (2002). Relational contracts and the theory of the firm. The Quarterly Journal of Economics, 117(1), 39–84.
Bandiera, O., Barankay, I., & Rasul, I. (2005). Social preferences and the response to incentives: Evidence from personnel data. The Quarterly Journal of Economics, 120(3), 917–962.
Bandiera, O., Barankay, I., & Rasul, I. (2007). Incentives for managers and inequality among workers: Evidence from a firm-level experiment. The Quarterly Journal of Economics, 122(2), 729–773.
Bergemann, D., & Bonatti, A. (2024). Data, competition, and digital platforms. American Economic Review, 114(8), 2553–2595.
Bolton, P., & Dewatripont, M. (2004). Contract theory. MIT Press.
Che, Y.-K., & Hausch, D. B. (1999). Cooperative investments and the value of contracting. American Economic Review, 89(1), 125–147.
Coase, R. H. (1937). The nature of the firm. Economica, 4(16), 386–405.
Corrao, R., Flynn, J. P., & Sastry, K. (2025; revised 2026). Contractibility design (NBER Working Paper No. 34379). National Bureau of Economic Research.
Davies, E., & Fafchamps, M. (2025). Material incentives and effort choice: Evidence from an online experiment across countries. Games and Economic Behavior, 154, 175–199.
Dovis, A., & Martellini, P. (2024; revised 2026). Data sharing with dynamic contracts (NBER Working Paper No. 33051). National Bureau of Economic Research.
Edlin, A. S., & Reichelstein, S. (1996). Holdups, standard breach remedies, and optimal investment. American Economic Review, 86(3), 478–501.
Erkal, N., Wu, S. Y., & Roe, B. E. (2025). Putting relational contract theory to the test: Experimental evidence. Experimental Economics.
Fehr, E., Hart, O., & Zehnder, C. (2011). Contracts as reference points—Experimental evidence. American Economic Review, 101(2), 493–525.
Grossman, S. J., & Hart, O. D. (1986). The costs and benefits of ownership: A theory of vertical and lateral integration. Journal of Political Economy, 94(4), 691–719.
Hart, O., & Moore, J. (1990). Property rights and the nature of the firm. Journal of Political Economy, 98(6), 1119–1158.
Hart, O., & Moore, J. (2008). Contracts as reference points. The Quarterly Journal of Economics, 123(1), 1–48.
Hart, O., Shleifer, A., & Vishny, R. W. (1997). The proper scope of government: Theory and an application to prisons. The Quarterly Journal of Economics, 112(4), 1127–1161.
Holmström, B. (1979). Moral hazard and observability. The Bell Journal of Economics, 10(1), 74–91.
Holmström, B., & Milgrom, P. (1991). Multitask principal–agent analyses: Incentive contracts, asset ownership, and job design. The Journal of Law, Economics, & Organization, 7(special issue), 24–52.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.
Laffont, J.-J., & Martimort, D. (2002). The theory of incentives: The principal-agent model. Princeton University Press.
Laffont, J.-J., & Tirole, J. (1986). Using cost observation to regulate firms. Journal of Political Economy, 94(3), 614–641.
Lazear, E. P. (2000). Performance pay and productivity. American Economic Review, 90(5), 1346–1361.
Levin, J. (2003). Relational incentive contracts. American Economic Review, 93(3), 835–857.
Mirrlees, J. A. (1971). An exploration in the theory of optimum income taxation. The Review of Economic Studies, 38(2), 175–208.
Myerson, R. B. (1981). Optimal auction design. Mathematics of Operations Research, 6(1), 58–73.
Myerson, R. B., & Satterthwaite, M. A. (1983). Efficient mechanisms for bilateral trading. Journal of Economic Theory, 29(2), 265–281.
Organisation for Economic Co-operation and Development. (2025). How widespread is algorithmic management in workplaces? OECD Publishing.
Royal Swedish Academy of Sciences. (2016). Contract theory: Scientific background on the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2016. Nobel Prize.
Sappington, D. E. M. (1991). Incentives in principal-agent relationships. Journal of Economic Perspectives, 5(2), 45–66.
Shahidi, P., Rusak, G., Manning, B. S., Fradkin, A., & Horton, J. J. (2025). The Coasean singularity? Demand, supply, and market design with AI agents (NBER Working Paper No. 34468). National Bureau of Economic Research.
Shearer, B. (2004). Piece rates, fixed wages and incentives: Evidence from a field experiment. The Review of Economic Studies, 71(2), 513–534.
Spence, M. (1973). Job market signaling. The Quarterly Journal of Economics, 87(3), 355–374.
Wang, S. (2024). Incomplete contracts with disparity, uncertainty, information and incentives. Theory and Decision, 97, 347–389.
Williamson, O. E. (1979). Transaction-cost economics: The governance of contractual relations. The Journal of Law and Economics, 22(2), 233–261.