This chapter surveys and analyzes the substantial literature on optimal remedies for contractbreach in a variety of settings. It begins with a standard analysis of the behavioral effects ofexpectation and reliance damages, then discusses the application of these damage measures ina world where courts are not perfectly informed about the parties' valuations of the contract.When valuation problems are extreme, courts may turn to alternative remedies such asspecific performance, or parties may attempt to solve the problem themselves throughliquidated damages clauses. The chapter considers whether these solutions to the valuationproblem alleviate or exacerbate opportunistic behavior by the parties. It also highlights therecent contributions that game theory and options theory have made to the understanding ofremedial choices.

Keywords: Contract damages; remedies

JEL classification: K12

1. Introduction

The principal remedy for breach of contract in Anglo-American law is an award of moneydamages. The preferred measure of damages is the expectation measure, under which thepromisee receives a sum sufficient, in theory, to make him indifferent between the award andthe performance. Other damage measures, and other remedies such as specific performanceand rescission, are available in special circumstances. This Chapter discusses the basic designof the remedial system.

A. The General Problem

2. Sanctions and Incentives

A contract is an exchange of promises or an exchange of a promise for a present performance,and the parties enter into it because each values the thing received more than the thingforegone. These values are based on expectations about the future because some or all of thecontractual performance will occur in the future. When the future diverges from what a partyexpected, he may conclude that the performance he will receive under the contract is nolonger more valuable than the performance he must provide. He has, in the terminology ofGoetz & Scott (1980), experienced a "regret contingency" and now would prefer not toperform and not to receive the promised performance from the other party.

Absent a system of contract remedies, a party who regrets entering into a contract will notperform unless he fears that the breach will result in sanctions by the other party (who mighthave required security for the performance) or by third parties (who might revise their opinionof the breacher and reduce their economic and/or social interactions with him accordingly).The economic function of contract remedies, then, is to alter the incentives facing the partywho regrets entering into the contract, which will directly affect the probability ofperformance and indirectly affect the number and type of contracts people make, the level ofdetail with which they identify their mutual obligations, the allocation of risks between theparties, the amount they invest in anticipation of performance once a contract is made, theprecautions they take against the possibility of breach, and the precautions they take againstthe possibility of a regret contingency.

An administratively simple system of remedies would aim to reduce the probability of breachto near zero. That could be achieved by the routine (and speedy) grant of injunctions againstbreach backed by large fines for disobeying the injunction or by imposing a punitively largemonetary sanction for breach. This would give promisees a high degree of confidence that thepromised performance will occur and induce a high level of investment in anticipation ofperformance. In the standard parlance, this would be a "property" rule because it would entitlethe promisee to the performance except to the extent the promisor is able to negotiate amodification on terms acceptable to the promisee.

3. Efficient Contracts and Efficient Nonperformance

Were it possible to enter into complete state-dependent contracts (that is, contracts thatidentified every possible contingency (state) and specified the required actions of the partiesfor each), parties would be willing to be bound to contracts even were the sanction for breachpunitive. Such contracts would require performance in some states but excuse it in others, insuch a way that each party would be willing ex ante to be absolutely bound to perform therequired actions in all states. Shavell (1980) defines a "Pareto efficient complete contingentcontract" as a complete state-dependent contract to which no mutually beneficialmodifications could be made, viewed at the time of contracting. We will call such contracts"efficient". In doing so, we will assume unless otherwise stated that the parties are riskneutral, each party's objective is to maximize his wealth, post-contractual renegotiation isprohibitively costly, performance is all or nothing (that is, partial performance is not possible),and the contracts do not create uncompensated gains or losses for third parties.

Under what conditions would an efficient contract excuse performance? Shavell demonstratesthat the contract would require performance in all circumstances except those in whichnonperformance would result in greater joint wealth. An example will illustrate the point.Imagine that Seller agrees to manufacture and sell to Buyer a machine that Buyer will use inits own manufacturing process. The value of the machine to Buyer is $300; however, Buyerhas an opportunity to make certain alterations to his manufacturing plant, at a cost of $50,which will increase the value of the machine to $375. Such investments by a promisee inanticipation of performance are called "reliance expenditures" or "reliance investments" in theliterature, and we will use the terms interchangeably. Assume for the moment that the futurecan be represented as a set of two possible states; Seller's production cost is $200 in one stateand $400 in the other. An efficient contract would require Seller to make the machine in thelow-cost state but not in the high-cost state. In the high-cost state, the joint wealth of theparties is greater if Seller does not perform than if it does. This can be seen by comparing thecost of performance to Seller ($400) with the benefit to Buyer ($300 or $375, depending onwhether Buyer makes the reliance investment). The contract price is irrelevant as it istransferred from Buyer to Seller and does not affect their joint wealth.

Both parties can be made to prefer this contract to one that requires performance in bothstates. They can allocate between themselves the extra wealth created by the efficient contract,and there will be some allocations under which each party's expected gain exceeds theexpected gain from the contract that always requires performance. By choosing such anallocation, each party will be better off at the time of contracting and willing to be bound toperform or not perform as required. In the literature, a breach that occurs in circumstances inwhich an efficient contract would excuse performance is called an "efficient breach".

4. Barriers to Efficient Contracting; Remedies as a Substitute for Efficient Contracts

To reiterate, faced with an efficient contract, courts would have the simple task of requiringstrict adherence to its terms. Unfortunately, the writing of efficient contracts is no easy task. Itis costly to bargain over remote contingencies and the parties may lack the foresight to dealwith all possible states. Moreover, the parties may not have equal access to the informationnecessary to tell which state occurs. In the above example, Seller may know whether the costof manufacturing the machine is $200 or $400, but Buyer may have no way to verify Seller'sassertion.

Given these barriers to efficient contracting, the law faces a more complex problem than thatof compelling adherence to efficient contracts. Instead, it must take incomplete contracts andaugment them by damage measures that induce behavior that mimics reasonably closely thebehavior that an efficient contract would require. A particular damage measure can be termed"efficient" with respect to a particular decision if it creates an incentive for the relevant partyto make the same decision it would under an efficient contract. Because standard damagemeasures allow a promisor to breach and pay compensatory (rather than punitive) damages,they are called "liability" rules in contrast to property rules as defined above.

5. Other Approaches

An alternative framework for the design of damage measures is offered by Barton (1972). Heposes the problem as one of designing damage measures that would induce the parties to makethe same decisions regarding performance or breach, and reliance prior to performance orbreach, that they would make were the parties divisions of a single, integrated firm and hadthe sole objective of maximizing the value of the firm. Shavell and Barton each show that theobjective of an efficient regime of contract damages is to cause the parties to maximize theirjoint wealth, although one might prefer Shavell's conceptual approach on the grounds thatBarton's assumes away the problem by positing a wealth-maximizing firm.

A more recent perspective on contract damages is to consider money damages as an optionunder which, for example, Seller may purchase Buyer's entitlement to Seller's performance.The option expires on the date fixed for performance and its strike price is the damage award(which may from the parties' perspective be a random variable). The value of the option isreflected in the contract price. See Mahoney (1995); Ayres & Talley (1995). This literaturederives from the more general use of option theory to analyze decision making underuncertainty. See Dixit and Pindyck (1994). We will make occasional reference to the optionsperspective below.

B. The Standard Damage Measures

6. A Taxonomy of Damage Measures

The preferred measure of contract damages is the amount of money that will make thepromisee indifferent between performance and damages. It should be noted at the outset thatthis formulation of the measure of damages is not fully accurate; there are a number oflimiting doctrines, discussed in section 4620, that often reduce money damages below thepromisee's subjective valuation of the performance. There is also some evidence that courtsaward greater damages for breaches that appear opportunistic. See Cohen (1994). As courtsexpress it, however, the preferred measure of damages is the amount necessary to put theaggrieved party in same position as if performance had occurred, which is known as theexpectation measure.

Fuller and Perdue (1932) provide the standard taxonomy of contract damage measures. Theyidentify three different "interests" of the promisee that are affected by a breach-- expectation,reliance, and restitution--and state that the most common damage measures providecompensation for one of the three. The expectation interest is measured by the net benefit thepromisee would receive should performance occur, as described above. The restitution interestconsists of any benefit the promisee has provided the breaching party. For example, if a selleragrees to make monthly deliveries of a commodity in return for fixed payments due 60 daysafter each delivery and the buyer repudiates the contract after receiving and retaining twodeliveries but making no payments, restitutionary damages would restore to the seller thevalue of the delivered goods. The reliance interest is measured by the promisee's wealth in thepre-contractual position. Reliance damages provide compensation both for any benefitconferred on the breaching party and for any other reliance investments made by the promiseein anticipation of performance to the extent such investments cannot be recovered.

In most instances, the restitution measure will provide the lowest recovery and the expectationmeasure the highest. One complication is how to treat other contractual opportunities thatBuyer passed up in order to enter into the contract with Seller. Analytically, these seemsimilar to reliance investments and are often treated as such. In a competitive market, whereBuyer could have entered into another contract at an identical price had he not contracted withSeller, the reliance measure and the expectation measure will converge approximately."Approximately", because the value of the alternative contract is a function of the probabilitythat it will be performed (see Cooter & Eisenberg (1985)) and of the damage remedy if it isnot performed, and thus the problem is somewhat circular. When analyzing the differencebetween expectation damages and reliance damages below, we will assume that they differand that Buyer's expectation interest exceeds his reliance interest. We will also assume thatthe reliance interest equals or exceeds the restitution interest, although we will relax thatassumption in section 12 below.

7. Incentives Within an Existing Contract: The Decision to Perform or Breach

The expectation measure leads to efficient decisions to perform or breach an existing contract,given a fixed level of reliance. (See Barton (1972), Shavell (1980), Kornhauser (1986)). Thiscan be illustrated using the example set out above. Assume that the contract price for themachine is $250 and that Buyer makes an irrevocable decision to invest $50 in reliance, aninvestment that has no value absent the contract. When production costs are $200, Seller willmanufacture the machine and Buyer will pay $250 for it. Buyer then obtains a machine worth$375 to him for a total expenditure (contract price plus reliance expenditure) of $300. Thetransaction increases Buyer's wealth by $75. When production costs are $400, Seller willbreach. The expectation measure seeks to make Buyer as well off as if Seller had performed.Seller's breach relieves Buyer from his obligation to pay the contract price. Accordingly, ifSeller pays Buyer damages of $125, Buyer will be in the same position as if Seller hadperformed, having paid out a non-recoverable $50 in reliance and received $125, for a netincrease in wealth of $75.

So long as Buyer is awarded $125 in the event of breach, Seller will breach only when the costof performance exceeds $375, the value of the performance to Buyer. Compare this result tothat obtained under the reliance measure. Under the reliance measure, Seller must compensateBuyer for his $50 reliance investment. Assume for a moment that there is a third possible stateunder which Seller's cost of production is $350. Performance would be efficient because itsvalue to Buyer exceeds its cost to Seller. Seller will perform given expectation damages,because the damage award of $125 exceeds the net loss from performance ($350 cost minusthe $250 contract price). Given reliance damages, however, Seller will breach and pay $50rather than perform at a loss of $100. More generally, it is obvious that given expectationdamages, only when the production cost reaches $376 will Seller become better off bybreaching and paying damages then by performing and losing the difference between hisproduction cost and the contract price. The expectation measure, unlike the reliance measure,causes Seller to internalize fully the effect on Buyer's wealth of Seller's decision to perform orbreach.

8. Incentives within an Existing Contract: The Decision to Rely

While the expectation measure produces efficient decisions to breach given reliance, it doesnot produce efficient levels of reliance. In general, expectation damages result in excessivereliance expenditures, because they cause Buyer to act as if performance were alwaysforthcoming. In the example, Buyer will always spend $50 to increase the value ofperformance from $300 to $375, because either (1) the performance will be forthcoming or (2)Buyer will be compensated for the lost $375 in value. In the high-cost state, however, theparties' joint wealth would be greater if Buyer refrained from investing. Seller would be liablefor damages of $300 less the $250 contract price, or $50. By contrast, if Buyer relies, hereceives $125 in damages as shown above and increases his wealth by $75 net of the relianceexpenditure. Unlike the no-reliance case, where Buyer gains $50 and Seller loses $50, hereBuyer gains $75 and Seller loses $125. The difference reflects the fact that the $50expenditure is wasteful in the high-cost state. Expectation damages, then, do not cause Buyerto internalize fully the effect on Seller's wealth of Buyer's decision to make a relianceinvestment.

The reliance measure is subject to the same objection. Under the reliance measure, Buyer willrecover $50 if it invests that amount in reliance. Once again, Buyer's investment decision willbe made as if the investment is not risky, even though it is (because performance is inefficientin some states). Indeed, reliance damages create a perverse incentive for Buyer in somecircumstances. Assume for a moment that Seller's production cost is $310. Under the reliancemeasure, Seller will pay damages of $50 rather than perform and suffer a loss of $60 ($310minus the $250 contract price). Breach deprives Buyer of a $75 gain (showing again that anymeasure of damages less than the expectation measure induces inefficient breach decisions).Buyer may be able to avoid breach, however, by making an additional (and we will assumewasteful) reliance expenditure of $11. Now reliance damages amount to $61, and Sellerperforms. Thus the excessive breach problem can be cured in part, but at the cost of excessivereliance. In general, as Shavell (1980) demonstrates, the reliance measure will result in greater(inefficient) reliance expenditures than the expectation measure. There is no measure ofdamages that results both in efficient decisions to perform or breach and efficient decisions tomake or not make reliance expenditures. However, expectation damages do better thanreliance damages at inducing efficient breach decisions, and do no worse than reliancedamages at inducing efficient reliance decisions. Accordingly, given the various assumptionsoutlined above, the expectation measure is preferable on efficiency grounds.

The analysis to this point has assumed risk neutrality. A risk averse Buyer would haveadditional cause to prefer the expectation measure, because it eliminates variability fromBuyer's outcome. At the same time, the expectation measure introduces greater variability intoSeller's outcome than does the reliance measure. It is accordingly possible that where bothparties are risk averse, they may find that a sum of damages greater than the reliance measurebut less than the expectation measure offers the highest joint utility level. The preciseformulation of the damage amount would depend on the parties' comparative levels of riskaversion. See Polinsky (1983). It seems plausible that courts have not tried to alter damagemeasures to accommodate risk aversion (except to the extent specific performance can do so,as discussed in section 10 below) because of the administrative and error costs that wouldresult.

The analysis has also assumed that renegotiation at the time of breach is prohibitively costly.Were negotiation costless, the damage rule would be irrelevant, as the parties would in allcases bargain to Pareto efficient breach/performance and reliance decisions, as per the CoaseTheorem (Coase (1960)). In the more plausible situation where renegotiation is costly but notalways prohibitively so, the choice of remedy is necessarily more difficult. Some critics havetherefore argued that much of the literature on damage remedies is beside the point, as thechoice of remedies should be informed principally by an analysis of transaction costs. (SeeFriedmann (1989); MacNeil (1982)). Friedmann analyzes potential transaction costs in avariety of contractual settings and argues that over compensatory remedies (remedies thatprovide compensation to the promisee in excess of the expectation interest) will generally beefficient.

9. Incentives at the Stage of Contract Formation

Friedmann and MacNeil are surely correct to argue that a better understanding of the costs ofpostcontractual renegotiation is necessary for making efficient remedial choices. It is also,however, worth paying attention to the effect of remedies on precontractual negotiations.

The price Seller will require to enter into a contract is increasing in the damage measure.Returning to our example, when Buyer makes a $50 reliance expenditure and Seller breaches,again assuming no opportunity costs, Buyer's wealth decreases by $50. Buyer can be no worseoff from entering into the contract so long as the remedy for breach is at least $50. Will Buyerbe willing to pay more for the more generous expectation measure, and will Buyer and Sellerprefer the resulting contract to one that provides for reliance damages only? As Friedman(1989) notes, the difference in remedies affects the contract price, the quantity contracted, andthe quantity actually consumed, with effects that vary with market structure and utilityfunctions. In general, however, the range of contract prices for which the contract increasesboth parties' wealth will be greater under a reliance measure than an expectation measure.That is, the reliance measure will create a greater bargaining range, which might increase thenumber of contracts entered into.

The choice of remedies where precontractual as well as postcontractual incentives areanalyzed remains an underdeveloped area. Friedman (1989) provides a formal analysis ofexpectation and reliance for two contexts in which those measures diverge. The first is thecase of a breaching buyer who has contracted to purchase from a monopolist selling at a singleprice. The second is the case of a breaching buyer in a competitive market where the sellerdoes not know its production cost in advance but the buyer does. Friedman demonstrates thatneither damage remedy dominates the other under those conditions. Friedman's analysis islimited, however, by his assumption that reliance is fixed and exogenous. The situations heanalyzes, moreover, have the desired formal characteristics (expectation and reliancemeasures diverge) but are probably not very common.

A possible alternative would be to start by assuming that the expectation and reliancemeasures diverge without specifying market structure in detail. A model could then bedeveloped in which the choice between expectation and reliance damages affects the structureof the contract, the decision to breach, and the decision to rely. The equilibrium andcomparative statics of such a model might shed light on the type of market conditions underwhich expectation or reliance damages would be more nearly optimal. It would also bevaluable to consider carefully whether there are plausible conditions under which the cost ofnegotiating around an inefficient damages measure at the time of contracting is greater or lessthan the cost of renegotiating at the time of performance.

C. Alternative Damage Measures

10. Specific Performance

Disappointed promisees are not in all cases limited to an award of damages; under appropriatecircumstances, they may seek the equitable remedy of specific performance. A decree ofspecific performance requires the breaching party to perform according to the contract. Theprincipal criterion for awarding specific performance is a demonstration that money damagesare insufficient to compensate the promisee for the lost performance. Traditionally, this wasmost often found when the breaching party was a seller who had agreed to sell a "unique"good. Real estate has long been presumed in many jurisdictions to be unique, while othergoods such as artworks and heirlooms are often found to be unique.

Specific performance is analogous to a punitive sanction that seeks to deter breach absolutely.In order for it to have that effect, we must assume that renegotiation is costly. It would thenseem clear that expectation damages are preferable to specific performance, because the latterwould sometimes result in performance even though nonperformance would result in greaterjoint wealth. On the other hand, it should be clear that the assumption that courts canadequately calculate a sum of money sufficient to make the promisee indifferent betweendamages and breach is not always accurate, particularly where cover is not possible. In suchcircumstances we must rely on a lost surplus measure of damages, and the calculation ofBuyer's consumer surplus is necessarily subjective. This is not a fatal objection if we believethat courts will guess correctly on average, but if they systematically underestimate Buyer'ssurplus, the monetary remedy will result in too much breach, just as specific performanceresults in too much performance.

Kronman (1978) started the law and economics debate on specific performance by employinga framework similar to that of the prior paragraph. He notes that specific performance is aproperty rule in the sense defined in section 2 above; it effectively assigns the promisee anabsolute entitlement to the goods from the moment the contract is made. This does not makesense in most instances because renegotiation (meaning a transfer of the property right back tothe promisor) is costly and the result will be an inefficiently high level of performance. Thedanger of under compensation, which would result in an inefficiently low level ofperformance, is normally lower because there is often a substitute price available. When,however, there is no substitute price available (the case of "unique" goods), the danger ofunder compensation likely outweighs the cost of renegotiation. Accordingly, the legal rules, ina rough manner, promote efficiency.

Schwartz (1979) argues that under compensation is not merely an isolated problem limitedprincipally to goods for which there is no obvious substitute, but is built into the structure ofmoney damages. The reluctance of courts to award damages that are uncertain, difficult tomeasure, or unforeseeable (see Chapter 4620), or to provide compensation for emotional harmresulting from a breach, makes money damages systematically under compensatory. Schwartzargues that the resulting inefficiencies are likely greater than those resulting fromrenegotiation costs, and accordingly that specific performance, rather than money damages,should be the default remedy.

Bishop (1985) adopts a similar analytic approach but argues that both Kronman and Schwartzhave overgeneralized their arguments. He breaks down contract breaches into a number ofcategories depending on the identity of the breaching party (buyer or seller), the type ofcontract, and the alternative transactions available to buyer and seller. He also identifiesanother cost of awarding specific performance. Because the value of a specific performanceaward (including the amount the promisor will pay to be released from performance) will insome cases exceed the value of performance to the promisee, the promisee will be tempted tobehave opportunistically in hopes of causing a breach and satisfying the conditions forspecific performance. Bishop argues that in some categories the problem of excessive breachresulting from under compensation will dominate, and in others the problem of excessiveperformance resulting from renegotiation costs and opportunism will dominate.

The relative magnitudes of the inefficiencies generated by costly renegotiation and undercompensation are ultimately empirical questions and to date the literature does not providedata from which we could confidently identify the preferred remedy. Accordingly, Mahoney(1995) takes a different approach to the problem, using the option methodology outlinedabove. The methodology is first employed to confirm the argument made by Craswell (1988)that were renegotiation costless and money damages perfectly compensatory, risk aversecontracting parties would always prefer money damages to specific performance. The intuitionis that entering into a contract with a money damages remedy is analogous to holding ahedged position in a commodity, whereas the identical contract with a specific performanceremedy is analogous to holding an unhedged position. The variance of possible outcomes isgreater for both parties with the unhedged contract and they will accordingly prefer moneydamages. In the face of costly renegotiation and under compensation, we can still make somesense of the case law using the option heuristic. Many contracts involving "unique" goods areprompted by the desire of the buyer to speculate on the future value of the land, artwork, etc.,and speculation involves holding an unhedged position. Thus buyer and seller would likelyprefer specific performance. Other cases in which specific performance has been consistentlyawarded (long-term contracts to supply a fuel input to a public utility or other regulated entity)can be explained by noting that the buyer is likely more risk averse with respect to pricefluctuations than is the seller, and specific performance better accommodate that distinction.

11. Liquidated Damages

We began the analysis of damages by arguing that court-awarded damages function as asubstitute for complete state-dependent contracts. The court's application of an efficientdamages rule creates appropriate incentives to perform or not perform, rely or not rely, and soon, and thereby saves the parties the trouble of drafting their contract to provide for allcontingencies.

Some parties, however, choose to create a tailor-made incentive structure by specifying theamount of damages payable in the event of breach. Courts have adopted a skeptical attitudetoward these so-called liquidated damages clauses. In general, courts will enforce a liquidateddamages clause only if (a) at the time of contracting, the damage that the promisee will sufferin the event of breach (i.e., the promisee's expectation) is uncertain, and (b) the amount ofliquidated damages is both a reasonable estimate of (the mean of) those damages and notdisproportionate to the actual (ex post) damages. A larger amount is called a "penalty" and isunenforceable.

This attitude is puzzling to most law and economics scholars. Absent some reason to believethat one or both parties misunderstood the terms of the agreement, we would normally assumethat they went to the trouble to specify liquidated damages because the resulting contract isPareto superior to a contract that calls for the ordinary court-awarded damages. The courts'approach might increase the net wealth of the parties, but only if the existence of a "penalty" isstrong evidence that the contract does not represent the parties' actual intent, perhaps becauseone defrauded the other. The focus of scholarship in the area, therefore, has been to ask underwhat circumstances rational, well-informed parties would agree to a penalty clause.

Goetz and Scott (1977) note, in terms similar to those of the later specific performance debate,that damages measures do not adequately compensate for the subjective value the promiseeattaches to performance, particularly when close substitutes for the performance areunavailable or the timing of the performance is critical. In those circumstances, normaldamage measures will lead to excessive breach.

The promisee could purchase third-party insurance that would pay off in the event of breach inan amount sufficient to make him whole. In some circumstances, however, the probability ofbreach is determined not just by exogenous variables beyond the parties' control, but also thelevel of care taken by the promisor. For example, a delivery service can affect the probabilityof timely delivery by the level of care it takes with the package. In such circumstances, thepromisor can insure more cheaply than a third party because the promisor can alter its level ofcare in response to the insurance clause. In other circumstances, the promisor may be betterinformed about the probability of breach than a third party insurer. The delivery service, forexample, may be better informed than the promisee or any third party insurer about thebreakdown rate of its trucks. In such cases, the promisor can use its willingness to agree to apenalty clause as a means of credibly signaling a low probability of breach.

It is accordingly not true that the mere existence of a penalty clause is a strong indicator offraud or mistake; there are plausible conditions under which a penalty clause would make bothparties better off. Schwartz (1990) supplements this analysis by considering the effects of apenalty clause on pre-contractual incentives. He notes that the price that the promisor willcharge is increasing in the damage measure. To the extent promisees insist on inefficientlylarge penalties, therefore, they will either pay too much or enter into too few contracts. Thepromisee accordingly has an incentive to demand a penalty clause only when it would increasethe joint wealth of the parties.

Other commentators have argued that penalty clauses can create externalities. Aghion &Bolton (1987), for example, demonstrate that a supply contract containing a penalty clause forbuyer's breach can restrict entry by competing sellers. We might at first conclude that buyerwould have no incentive to agree to a contract that limited competition from other sellers. Awell-designed penalty, however, will merely redistribute wealth from the new entrant to thecontractual buyer and seller. To see why, assume a contract between Buyer and Seller with acontract price of $200, and Seller's cost of performance is $150. At a later time, Entrantappears, who can provide the good for $100. Seller's expectation damages will be $50, soabsent a penalty Buyer can profitably breach if Entrant offers a price of $149 or less.Assuming that Entrant does not face competition, Entrant will demand a price of $149. Nowimagine that Buyer must pay a penalty of $80 upon breach. Now it is not profitable for Buyerto breach unless Entrant offers a price of $119 or less. Entrant can still profitably sell at thatprice, and will do so. Thus the penalty transfers wealth from Entrant to Seller (which Sellercan agree ex ante to share with Buyer).

Aghion & Bolton's analysis works only if Entrant has market power. The externalityarguments for the most part are not sufficiently general to provide a compelling explanationfor the judicial hostility toward penalty clauses. Moreover, while they can provide support forpart of the judicial approach (disallowing liquidated damages that are not a fair ex anteestimate of actual damages), they cannot explain the failure to enforce liquidated damages thatare excessive ex post.

12. Rescission/Restitution

Courts divide contract breaches into "partial" and "total" breach. A partial breach gives thepromisee the right to seek a remedy but not to refuse his own performance. The classicexample is when a builder constructs a house that contains a minor deviation from the agreedarchitectural plan. The builder must compensate the owner for the difference in value (intheory, the difference in subjective value to buyer, but it will usually be difficult to convince acourt that this differs substantially from the difference in market value). The owner may not,however, refuse to accept delivery of the house and to pay the agreed price. A total breach, bycontrast, permits the promisee to refuse to render his own performance. In effect, a totalbreach permits the promisee to rescind the contract.

As courts express it, a promisee can respond to a total breach by seeking expectation damagesor by rescinding and seeking recovery of any value he has provided to the breaching promisor.The latter alternative is equivalent to the restitution measure of damages (although in somecircumstances the promisee may seek return of the performance in specie rather than itsmonetary equivalent). Restitution is also a remedy in quasi-contractual situations, such aswhen parties partly performed a contract that is voidable for mutual mistake, but the followingdiscussion will be limited to restitution damages as a remedy for breach.

In the typical case, expectation damages will exceed restitutionary damages and the promiseewill seek the former. There are two instances, however, in which we would expect thepromisee to seek the latter. The first is when the promisee is risk averse and prefers thecertainty of the return of money or property that he has given the promisor to the uncertaintiesof a jury's assessment of his expectation and the additional litigation costs that would beincurred in the attempt. The second is when the contract was a losing deal for the promisee, sothat his expectation is negative. Where the promisee has provided something of value to thepromisor that cannot be easily returned, but can be valued in a judicial proceeding, thepromisee may be better off receiving that value in cash than receiving the promisedperformance.

This might be thought a remote possibility, but it occurs in a number of reported cases. Thetextbook example is one in which a builder agrees to build a house for an owner and thebuilder's costs turn out to be greater than expected, making the contract a losing one for thebuilder. The owner, however, later decides it does not want the house and repudiates thecontract when the house is partly completed. The builder's expectation is negative because ofthe unexpectedly high costs of construction, so the builder seeks restitution. Restitution in thisinstance is measured by the value the builder has conferred on the owner, or the market valueof the nearly-completed house. By hypothesis, this exceeds the contract price.

When promisees have attempted to recover reliance damages for a losing contract, courts haveconcluded that the expectation measure puts an upper bound on the recovery (see L. Albert &Son v. Armstrong Rubber Co.). By contrast, some courts have permitted a promisee to recoverrestitution damages in excess of expectation. (See Boomer v. Muir) This seeminginconsistency has been largely ignored in the law and economics literature. The most usefuldiscussions appear in a symposium issue of the Southern California Law Review in 1994. Init, Kull (1994) provides an analysis of restitution that is similar in many respects to Bishop'sanalysis of specific performance. Money damages are not always an adequate substitute forperformance and the damage calculation is in any event uncertain. Thus where the promiseehas provided something of value to the promisor that can easily be returned, the promisee mayprefer to rescind the transaction, putting both parties back in the pre-contractual position. Forexample, the promisee may have paid in advance for a good or service that the promisor failsto provide. Taking litigation costs into account, the promisee may prefer to rescind thetransaction and retrieve the advance payment.

One example of a situation in which it seems likely that rescission and restitution willminimize the costs associated with breach is where a seller delivers goods that do not conformto the contractual specifications. The perfect tender rule, recognized under the common lawand the Uniform Commercial Code, permits a buyer to reject nonconforming goods even ifthe variation is minor. As noted by Priest (1978), the administrative costs involved incalculating the difference in value between the goods as delivered and as promised will likelyexceed the cost of returning the goods to Seller and money to Buyer. The costs associated withsalvaging the nonconforming goods might also be minimized by the perfect tender rule, as inmany instances it will be cheaper for Seller to find another purchaser for the goods than it willbe for Buyer to adapt the goods to Buyer's own use.

On the other hand, where the contract is a losing one for the promisee and the promisee hasconferred a benefit on the promisor that cannot easily be returned, the remedy of rescissionand restitution is potentially over compensatory. Kull argues that the threat of opportunisticbehavior (that is, socially wasteful efforts to exploit an inefficient remedy to obtain anunbargained-for benefit) will be substantial for such contracts. The promisee can turn a lossinto a gain by inducing breach by the promisor (or convincing a court that mutualuncooperativeness constituted or resulted from such a breach). By contrast, the perfect tenderrule permits a buyer to behave opportunistically by unreasonably claiming that goods aredefective when their market value has declined, but because the goods can be returned toSeller, the parties are spared the additional cost of a court proceeding to determine their value.

D. Calculation of Expectation and Reliance Damages

13. A Categorization of Approaches to Calculating Damages

There is consensus that the expectation measure is in most circumstances superior to relianceor restitution damages. A separate but no less important question is how expectation andreliance are to be defined and measured in typical contractual settings. Parties' valuations areoften unknown to one another and to the court, and promisees have an incentive to overstatetheir valuations, making the calculation of expectation damages difficult in some settings.Cooter and Eisenberg (1985) present a very helpful categorization and analysis of alternativecalculation methods. They identify five broad categories and note that the calculation ofmoney damages in reported cases usually falls into one of these categories. They are:

i) Substitute Price. Often there is a spot market for the contractual performance at the time andplace that performance was due, most obviously if the performance consists of the delivery ofa marketable commodity. In such an event, Buyer can respond to Seller's breach by cover, orthe purchase of the commodity on the spot market. (Seller can respond to a breach by Buyerby selling on the spot market.) The difference between the contract price and the price atwhich cover occurred or could have occurred is then a measure of the cost of making Buyer(or Seller) indifferent between the contract and the substitute performance. We should note,however, that the Substitute Price measure can be over compensatory when a promiseechooses not to cover but instead to sue for the difference between the contract price and thespot price. That choice itself suggests that the promisee may value the commodity at less thanits market price.

ii) Lost Surplus. When cover is unavailable, Buyer's expectation can be thought of as the lostconsumer surplus from the contract. In our ongoing example, if Buyer cannot cover, he losesthe difference between his valuation of the machine ($300 or $375, depending on reliance)and the $250 contract price. The analysis of Seller's lost producer surplus from Buyer's breachis analogous. The Lost Surplus measure is feasible only when a court can obtain credibleevidence of the promisee's valuation.

iii) Opportunity Cost. If a market exists for the performance, Buyer could have entered into acontract to buy the machine from any one of a number of competing sellers. The value toBuyer of the best alternative contract available at the time of the contract with Seller is animportant component of his reliance. This value cannot be measured objectively becauseBuyer did not enter into this hypothetical contract and we don't know whether the hypotheticalcontractual party would have performed. Assuming that the probability of performance of thealternative contract is high, however, then the difference between the spot price at the timeand place of breach and the price of the foregone contract is a good measure of reliance(augmented by any out-of-pocket expenditures in reliance on the contract with Seller). In acompetitive market, the next-best price and the contract price should be the same, and theOpportunity Cost measure will equal the Substitute Price measure (a conclusion consistentwith Fuller and Perdue's conclusion that expectation and reliance damages are equal in acompetitive market).

iv) Out-of-Pocket Cost. This is the amount of reliance investment, less any salvage value ofthat investment. Out-of-Pocket Cost is the most common measure of reliance damages; amore complete measure of reliance damages is Out-of-Pocket Cost plus Opportunity Cost.

v) Diminished Value. So far we have ignored partial performance. In the real world, however,performance is often rendered but is defective or incomplete. In such cases an appropriatemeasure of Buyer's lost expectation is the difference between Buyer's valuation of thepromised performance and his valuation of the actual performance.

As these alternative methods of calculation should make clear, the measure of damages isusually straightforward and uncontroversial where cover is possible. The accepted measure ofdamages in such cases is the difference between the cover price and the contract price, whichis easy to apply and provides appropriate incentives regarding the decision to perform orbreach. The difficult questions arise when there is no perfect substitute for the performance(or there is room for debate about whether the substitute is adequate) or where the manner ortiming of the breach causes harm that cannot be remedied by cover. We will provide twoexamples of cases that arise frequently and that have been the much discussed in the literature,in which there is debate over the appropriate means of measuring the non-breaching party'sexpectation.

14. Example 1: Anticipatory Repudiation

Common law judges and scholars initially found anticipatory repudiation--a definitivestatement by a promisor, made prior to the time for performance, that he intended tobreach--extraordinarily vexing. Some concluded that any such statement must be without legaleffect; the performance was due on a particular date and breach could therefore only occur onthat date (Williston (1901)). Courts eventually came to the view that the promisee could treatthe repudiation as a breach (Hochster v. De La Tour), but found it more difficult to decidehow damages should be measured. The most famous early case, Missouri Furnace v.Cochrane, held that the appropriate measure was the difference between the contract price andthe spot price at the time specified for performance. The Uniform Commercial Code, bycontrast, encourages prompt cover, presumably in the futures market. As noted by Jackson(1978), the legal literature on anticipatory repudiation from the early part of this century isvoluminous.

Jackson argued that in applying the Uniform Commercial Code's provisions on cover toanticipatory repudiation, courts should fix damages at the difference between the contractprice and the futures price at the time of repudiation. He noted that the Missouri Furnacemethod is systematically over compensatory. Imagine, for example, that Seller breaches acontract to supply a commodity in the future and that the spot and futures prices at the time ofrepudiation are higher than the contract price. Over a large number of contract breaches,however, the spot price at the time of performance will sometimes be higher, and sometimeslower, than the contract price (in present value terms). Whenever it is lower, Buyer will notbring a damages action because he has been made better off by the breach. He is under noobligation to share this gain with Seller. When the spot price is higher than the contract price,Buyer will recover the difference between the two. Averaged over a large number of contracts,buyers in the aggregate receive more than would be required to make them as well off as theywere under the contract. Awarding the difference between the contract price and the futuresprice, by contrast, puts each buyer in the position he occupied prior to the repudiation and at alower average cost to sellers.

We might simplify Jackson's argument by noting that the Missouri Furnace rule replaces aforward contract by an option with a strike price equal to the forward price. Because the value(prior to expiration) of an option with a strike price of X is always greater than the value of aforward contract with a contract price of X, the Missouri Furnace damage measure is overcompensatory.

15. Example 2: The Lost-Volume Seller

Sellers in a competitive market have often argued that the Substitute Price measure ofdamages, which awards them the difference between the contract price and the spot price, isunder compensatory. In many instances, there is little or no difference between the contractprice and the spot price, and accordingly the damage award is trivial. Sellers contend,however, that they are not "made whole" by selling in the spot market; the seller had thecapacity to sell to both the substitute buyer and the original buyer at the market price, and thebreach reduced their sales volume by one unit. Thus in place of two sales and two profits, theyhave received only one sale and one profit. Courts have often awarded the so-called"lost-volume seller" an amount of damages equal to its ordinary profit on one sale. In thewell-known case of Neri v. Retail Marine Corp., Retail Marine, a dealer in boats, agreed tosell a boat to Neri at a fixed price. Retail Marine ordered the boat from the manufacturer butNeri repudiated the contract. Retail Marine sold the boat to another customer for the sameprice and successfully sued Neri for the profit it would have made on the sale to him. Thecourt concluded that Retail Marine, as a dealer, had an "inexhaustible" supply of boats, andNeri's breach deprived it of a profitable sale.

There is a substantial law-and-economics literature on the lost-volume seller. An earlycontribution appeared in an anonymous student-written comment (Anonymous (1973)). Thecomment noted that in a perfectly competitive market, each seller would choose output byequating marginal cost with demand and the demand curve would be presumed horizontal. Atthe chosen output, the firm's marginal cost would be rising and therefore any additional salewould be at a cost in excess of the price. Because the seller could not, in fact, satisfyadditional buyers at the market price, the breach and resale would create no "lost volume" in aperfectly competitive market. A seller with market power (that is, one facing adownward-sloping demand curve) might be able to make additional sales at a profit. However,by hypothesis, such a seller could eliminate the "lost volume" by reducing its price andmaking an additional sale. Thus the standard contract price minus cover price measure wouldfully compensate such a seller.

Goetz and Scott (1979) provide an additional argument against awarding lost profits to theretailer who has market power. They note that the breach removes the breaching buyer as acompeting seller. The buyer presumably breaches because it no longer wants the good at thecontract price. In lieu of breaching, however, the buyer could complete the purchase and thenresell the good. This resale, if made in the same market in which the retailer operates, shiftsthe demand curve facing the retailer to the left by one unit. Once again, if we compare theretailer's position after the breach to its position assuming no breach but resale by buyer, thereis no "lost volume".

Goldberg (1984) disputes Goetz and Scott's analysis. He first argues that the observation that anon-breaching buyer could sell in competition with the retailer is unrealistic. In fact, heargues, the buyer, lacking expertise, would have to engage the services of a retailer. Theretailer's usual markup is a reasonable estimate of the fee the retailer would charge for hisservices. Accordingly, the award of lost profit to the retailer approximates the result thatwould obtain if the buyer purchased and resold.

Goldberg also argues that it is inaccurate to say that the retailer "saves" the marginal cost of asale when the original buyer breaches and then incurs that marginal cost when the substitutebuyer appears. He contends that the retailer's cost of servicing an additional buyer consistsprincipally of the cost of "fishing" for a buyer, or convincing the marginal buyer to purchase(represented, perhaps, by costs of advertising, wages paid to salespeople, etc.). That cost isirretrievably lost once a contract is concluded with the original buyer and must be incurredagain in order to induce another buyer to purchase. More recently, Scott (1990) argues thatGoldberg's equation of marginal cost with the cost of "fishing" is inaccurate; for some goods,the cost of delivery and preparation for delivery are significant, and those costs are notincurred twice when a buyer defaults. Cooter and Eisenberg (1985) provide an analysis similarto Goldberg's, but focus on the seller with market power. They argue that many sellers holdprice at a constant level reflecting expected demand and marginal cost over some period,rather than constantly adjusting price to reflect realized demand. Such sellers can lose volumein a particular period.

Goldberg also notes that consumer demand is decreasing in the damage measure.Accordingly, were the legal rule to shift suddenly from a substitute price damages measure toone awarding lost profits, the demand curve facing the retailer would shift downward,offsetting the benefit of the higher damage awards. Whether consumers and producers wouldprefer the resulting contract to one that provides only substitute price damages again dependson comparative levels of risk aversion.

It appears that the literature on the lost-volume seller is at an impasse. The identification ofthe best damages rule turns on complex and contestable claims about market structure. Abetter avenue of inquiry would be to pay attention to actual contractual practice. Many sellersof custom goods require non-refundable deposits, which in effect contracts for a lost-profitsmeasure. Other sellers (such as many computer retailers) permit a buyer to return an item for afull refund for some period after delivery, which in effect contracts for an even more lenientapproach than the substitute price measure. It seems likely that greater ground will be gainedby empirical analysis of the characteristics of markets in which varying cancellation/returnpolicies are used than by further refinements of the theoretical arguments.

E. Conclusions

16. The Puzzle of Over Compensatory Remedies and Some Suggestions for FurtherResearch

Most of the prior analysis could be summed up as follows: when courts and contractingparties are well-informed about each party's valuation of the contract, money damagesmeasured by the promisee's valuation (or expectation) provide reasonably good incentives forefficient pre- and post-contractual behavior. Problems arise, however, when there aresignificant informational asymmetries between the parties and/or between each party and thecourt. Such asymmetries raise two pervasive issues in contract law. The first is subjectivevalue. The existence of potentially over compensatory remedies such as specific performance,liquidated damages and restitution can be attributed to judicial recognition that moneydamages measured by the promisee's expectation will sometimes under compensate, becausecourts use objective indicators of value that may diverge from the promisee's subjectivevaluation. Only a few brief attempts have been made, however, to explore subjective value asa unifying theme in contract remedies. (See De Alessi & Staaf (1989), Muris (1983)).

The second issue is opportunism. The possibility that a remedy, although designed to beperfectly compensatory, will in fact under compensate (over compensate) may encourage thebreaching party (non-breaching party) to use the defect in the remedy to gain bargainingleverage over the other party. The risk of opportunism is the likely reason why courts have notresponded to the problem of subjective valuation by instituting over compensatory remediesacross the board.

A worthwhile avenue for additional work would be a careful comparison of the ways in whichcourts have or have not managed to reduce the risk of opportunism across a range of remedialchoices. A promising approach to this question appears in the liability rule versus propertyrule literature. When neither party knows the other's true valuation of the contract, each has anincentive to over- or understate his valuation in an attempt to capture as much as possible ofthe gains from contract modification or cancellation. The result is to make agreement morecostly. The costs imposed by asymmetric information, which we will call "bargaining costs",are a subset of the cost of reaching a deal. The key question is whether the choice of remedyaffects bargaining costs.

A specific application to liquidated damages is offered by Talley (1994). He uses themechanism design branch of game theory to analyze the effects of different enforcement ruleson bargaining costs, concluding that enforcement of liquidated damages that exceed actualdamages ex post creates significant bargaining costs. By refusing to enforce penalty clauses,courts may make it more likely that the parties will bargain to an efficient outcome. Theargument is unique in offering a plausible economic justification of the ex post component ofthe liquidated damages rule.

Ayres & Talley (1995), employ game theory to argue that bargaining costs are generally lowerunder liability rules than under property rules. The intuition is as follows. Going back to ourcontract between Seller and Buyer, imagine that Seller wishes to breach, and believes Buyer'svaluation of the contract to be uniformly distributed on the interval [$300, $400]. Consider arule that provides for damages of $500 in response to Seller's breach. Buyer's offer to rescindthe contract for a payment of $400 would provide Seller with no new information--Selleralready knows that Buyer's valuation is no greater than $400. Now consider a rule providingfor damages of $350. Buyer might now conceivably offer to cancel the contract in return for apayment from Seller (if Buyer's valuation is less than $350), or it might offer Seller a paymentto forego breach (if Buyer's valuation is more than $350). Thus the type of offer that Buyermakes conveys information about its valuation and ameliorates the bargaining costs resultingfrom asymmetric information. Johnston (1994) offers an analogous argument to show thatbargaining costs can be lower under a "standard", in which an entitlement is dependent on adiscretionary judicial determination, than under a "rule", in which the entitlement is moreprecisely defined.

Kaplow and Shavell (1996) criticize Ayres and Talley's analysis on the grounds that it is not amarginal analysis. They argue that in most contexts in which bargaining is impossible orprohibitively costly, liability rules will dominate property rules for the reasons outlined in ourdiscussion of expectation damages above. Thus for liability rules to dominate property ruleswhere bargaining is possible does not prove that they generate lower bargaining costs; thelatter point would be proved conclusively only if liability rules dominate property rules to aneven greater extent where bargaining is possible than where it is impossible.

It is perhaps unfortunate that the game-theoretic analysis of bargaining costs has been usedprincipally to analyze the relative efficiency of liability and property rules. The more generalquestion is the design of remedies that will create optimal incentives for the parties to revealtheir actual valuations or other private information about the state of the world. A liability rule(that is, a rule under which the damage award may be greater or less than the promisee'svaluation) might create superior incentives compared to a property rule (that is, one underwhich the damage award is at or beyond the endpoint of the promisee's valuation), but weshould be able to make similar comparisons between different liability rules. The discussionto date has covered liability and property rules generally, whether located within the law ofproperty, torts and contract. There is accordingly room for a more focused look at bargainingcosts in contractual settings, with an additional emphasis on ex ante mechanisms other thanjudicially-crafted damage rules that might help to reduce ex post bargaining costs.

Acknowledgements

Paul G. Mahoney thanks Eric Talley and two anonymous referees for helpful comments.

Bibliography on Contract Remedies: General (4600)

Anonymous (1973), 'Comment: Microeconomics and the Lost-Volume Seller', 24 CaseWestern Reserve Law Review, 712 ff. Reprinted in Kronman, Anthony T. and Posner, RichardA. (eds.), The Economics of Contract Law, Boston, Little Brown, 1979, 213-220.

Aghion, Philippe and Bolton, Patrick Bolton (1987), 'Contracts as Barriers to Entry', 77American Economic Review, 388-401.

Ayres, Ian and Talley, Eric L. (1995), 'Solomonic Bargaining: Dividing a Legal EntitlementTo Faciliate Coasean Trade', 104 Yale Law Journal, 1027-1117.

Barton, John H. (1972), 'The Economic Basis of Damages for Breach of Contract, 1 Journalof Legal Studies, 277-304.

Bishop, William (1985), 'The Choice of Remedy for Breach of Contract', 14 Journal of LegalStudies, 299-320. Reprinted in Goldberg, Victor P. (ed.), Readings in the Economics ofContract Law, Cambridge, Cambridge University Press, 1989, 122-125.

Coase, Ronald H. (1960), 'The Problem of Social Cost', 3 The Journal of Law and Economics1-44.

Cohen, George M. (1994), 'The Fault Lines in Contract Damages', 80 Virginia Law Review,1225-1349.

Cooter, Robert D. and Eisenberg, Melvin Aron (1985), 'Damages for Breach of Contract', 73California Law Review, 1432-1481.

Craswell, Richard (1988), 'Contract Remedies, Renegotiation, and the Theory of EfficientBreach', 61 Southern California Law Review, 629-670.

De Alessi, Louis and Staaf, Robert J. (1989), 'Subjective Value in Contract Law', 145 Journalof Institutional and Theoretical Economics, 561-577.

Dixit, Avinash K. and Pindyck, Robert J. (1994), Investment Under Uncertainty, Princeton,New Jersey, Princeton University Press.

Friedman, David D. (1989), 'An Economic Analysis of Alternative Damage Rules for Breachof Contract', 32 Journal of Law and Economics, 281-310.

Friedmann, Daniel (1989), 'The Efficient Breach Fallacy', 18 Journal of Legal Studies, 1-24.

Fuller, Lon L. and Perdue, William (1936), 'The Reliance Interest in Contract Damages', 46Yale Law Journal, 52-98. Reprinted in Goldberg, Victor P. (ed.), Readings in the Economicsof Contract Law, Cambridge, Cambridge University Press, 1989, 77-79.

Goetz, Charles J. and Scott, Robert E. (1980), 'Enforcing Promises: An Examination of theBasis of Contract', 89 Yale Law Journal, 1261-1300.

Goetz, Charles J. and Scott, Robert E. (1977), 'Liquidated Damages, Penalties and the JustCompensation Principle: Some Notes on an Enforcement Model and a Theory of EfficientBreach', 77 Columbia Law Review, 554-594. Reprinted in Kronman, Anthony T. and Posner,Richard A. (eds.), The Economics of Contract Law, Boston, Little Brown, 1979, 194-207.

Goetz, Charles J. and Scott, Robert E. (1979), 'Measuring Sellers' Damages: The Lost-ProfitsPuzzle', 31 Stanford Law Review, 323-373.

Goldberg, Victor P. (1984), 'An Economic Analysis of the Lost-Volume Retail Seller', 57Southern California Law Review, 283-298. Reprinted in Goldberg, Victor P. (ed.), Readingsin the Economics of Contract Law, Cambridge, Cambridge University Press, 1989, 106-113.

Jackson, Thomas H. (1978), '"Anticipatory Repudiation" and the Temporal Element inContract Law: An Economic Inquiry into Contract Damages in Cases of ProspectiveNonperformance', 31 Stanford Law Review, 69-119.

Johnston, Jason S. (1995), 'Bargaining Under Rules Versus Standards', 11 Journal of Law,Economics and Organization 256-281.

Kaplow, Louis and Shavell, Steven (1995), 'Do Liability Rules Facilitate Bargaining? A Replyto Ayres and Talley', 105 Yale Law Journal 221-233.

Kornhauser, Lewis A. (1986), 'An Introduction to the Economic Analysis of ContractRemedies', 57 Colorado Law Review, 683-725.

Kronman, Anthony T., 'Specific Performance' (1978), 45 University of Chicago Law Review,351-382. Reprinted in Kronman, Anthony T. and Posner, Richard A. (eds.), The Economics ofContract Law, Boston, Little Brown, 1979, 181-194.

Kull, Andrew (1994), 'Restitution as a Remedy for Breach of Contract', 67 SouthernCalifornia Law Review, 1465-1518.

Macneil, Ian R. (1982), 'Efficient Breach of Contract: Circles in the Sky, 68 Virginia LawReview 947-969.

Mahoney, Paul G. (1995), "Contract Remedies and Options Pricing", 24 Journal of LegalStudies; 139-63.

Muris, Timothy J. (1983), 'Cost of Completion or Diminution in Market Value: TheRelevance of Subjective Value', 12 Journal of Legal Studies, 379-400. Reprinted in Goldberg,Victor P. (ed.), Readings in the Economics of Contract Law, Cambridge, CambridgeUniversity Press, 1989, 128-132.

Polinsky, A. Mitchell (1983), 'Risk Sharing through Breach of Contract Remedies', 12 Journalof Legal Studies, 427-444.

Priest, George L. (1978), 'Breach and Remedy for the Tender of Nonconforming Goods Underthe Uniform Commercial Code: An Economic Approach', 91 Harvard Law Review, 960-1001.Reprinted in KRONMAN, Anthony T. and Posner, Richard A. (eds.), The Economics ofContract Law, Boston, Little Brown, 1979, 167-175.

Schwartz, Alan (1990), 'The Myth That Promisees Prefer Supracompensatory Remedies: AnAnalysis of Contracting for Damage Measures', 100 Yale Law Journal, 369-407.

Schwartz, Alan (1979), 'The Case for Specific Performance', 89 Yale Law Journal, 271-306.

Scott, Robert E. (1990). 'The Case for Market Damages: Revisiting the Lost Profits Puzzle',57 The University of Chicago Law Review, 1155-1202.

Shavell, Steven (1980), 'Damage Measures for Breach of Contract', 11 Bell Journal ofEconomics, 466-490.

Talley, Eric, L. (1994), 'Contract Renegotiation, Mechanism Design, and the LiquidatedDamages Rule', 46 Stanford Law Review, 1195-1243.

Williston, Samuel (1901), 'Repudiation of Contracts', 14 Harvard Law Review, 317-331 (PartI), 421-441 (Part II).

Cases

Hochster v. De La Tour, 118 Eng. Rep. 922 (Q.B. 1853).

Missouri Furnace v. Cochrane, 8 F. 463 (C.C.W.D. Pa. 1881).

Neri v. Retail Marine Corp., 30 N.Y.2d 393 (1972).

L. Albert & Son v. Armstrong Rubber Co., 178 F.2d 182 (2d Cir. 1949).

Boomer v. Muir, 24 P.2d 570 (Cal. App. 1933).