Contract Remedies: Foreseeability,Precaution, Causation and Mitigation
Eric A. Posner
Assistent Professor of Law, University of Pennsylvania Law School
© Copyright 1997 Eric A. Posner
This section discusses undercompensatory damages rules in contract law.These are rules that reduce the award to the promisee to an amount below thatto which he would be entitled under the expectations rule. The reason forreducing compensation is to encourage the promisee to behave in a value-maximizing manner. The Hadley rule, which denies recovery for losses that thepromisor could not ordinarily foresee, is thought to provide promiseesincentives to reveal how much they value performance, in order to give thepromisor optimal incentives to take care. The mitigation rule, which deniesrecovery for losses that the promisee could have avoided by mitigating damagesafter the breach, is thought to provide promisees with the proper incentives toreduce losses caused by the promisor's breach. Several theoretical and empiricalambiguities, however, weaken these arguments. Two other undercompensatorydamages rules -- the "reasonable certainty" doctrine and the emotional distressdoctrine -- are also discussed.
The section also deals with precautionary behavior by both the promisor andthe promisee. The expectations rule encourages the promisor to take optimalprecautions, but it also discourages the promisee from taking precautions. Toencourage the promisee also to take optimal precautions, damages must beinvariant with respect to the amount of reliance, and, ideally, should be set at theefficient level of reliance. Causation is not an important concept in contract lawand scholarship.
JEL classification: K12
Keywords: Contract Remedies: Foreseeability, Precaution, Causation, andMitigation
The expectation measure of contract damages, which requires the breachingpromisor to pay the promisee an amount of money sufficient to put the promiseein the position it would have been in if the promisor had performed, is thestandard measure of damages in Anglo-American law, but it is subject to severalexceptions. First, the award is reduced when a portion of the promisee's loss isattributable to circumstances that the promisor could not ordinarily foresee. Thisresult is due to the rule of consequential damages, also known as the Hadleyrule, after Hadley v. Baxendale (1854). Second, the award is reduced when aportion of the promisee's loss is attributable to the promisee's failure to takeactions that would have minimized its losses subsequent to the promisor'sbreach. This result is due to the mitigation rule. Two less important doctrineshold that the award is reduced when the promisee cannot show that the loss was"reasonably certain," and when the loss consists of emotional distress. Becausethese four rules on average produce an award lower than that produced by theexpectation rule if applied unconditionally, they can be thought of as composinga larger category of "undercompensatory" contract damages rules.
The value of the Hadley, mitigation, reasonable certainty, and emotionaldistress rules, like that of all other rules of contract law, depends on their effectson the various forms of contract-related behavior. This behavior includes theselection of contracting partners, the choice of contracting rather than usingother forms of commitment, the care in drafting the contract, the disclosure byeach party of private information, investment in anticipation of performance, thetaking of precautions to minimize the expected loss from breach, the allocationof risk, performance rather than breach, and the readiness to renegotiatecontractual obligations in light of changed circumstances. We concentrate onthe promisee's incentives to disclose information, to take precautions, and tomitigate damages. The Hadley rule and the mitigation rule have their greatestimpact on these forms of behavior, and they will be discussed in the first andsecond sections of this chapter. The third section discusses the effect ofdamages rules on the parties' incentives to take precautions. The fourth sectiondiscusses causation, which, however, is not an important concept in contractlaw. The fifth section briefly discusses the reasonable certainty and emotionaldistress rules; however, these rules either have not received much attention inthe literature or are discussed in other sections of this Encyclopedia.
The Hadley rule states that promisors are not liable for losses that are not"foreseeable." This term has attracted a great deal of criticism. Somecommentators charge that courts simply call a loss "unforeseeable" wheneverthey believe, for an unarticulated reason of policy, that the loss should not fallon the promisor. See Fuller and Perdue (1936). Whether or not this skepticism isjustified, the term does not provide much guidance, because sophisticatedparties can foresee and often do foresee the range of losses consideredunforeseeable by courts. For example, if a carrier fails to deliver goods or deliversthem late, and this breach causes large losses for the shipper, perhaps becausethe goods were to perform a crucial role in the production process, the carrier willusually be liable for a limited amount, such as the purchase price, rather than theamount necessary to satisfy the shipper's expectation. In the traditionaldoctrine, the shipper receives full expectation damages only if the losses arepredictable or if the shipper informs the carrier of the size of the anticipated lossprior to agreement. But surely the uninformed carrier can "foresee" that someshippers place a greater value on performance than other shippers. When courtssay that a loss is unforeseeable, the best interpretation of this conclusion is thatthe loss is considerably greater than the average of the losses that result fromthe kind of breach in question.
The Hadley rule is a default rule: parties are free to bargain around it andfrequently do. Carriers, for example, often limit their liability and separately sellinsurance against losses from misdelivery of expensive goods. The questionasked by the Hadley literature is why the default rule limits liability toforeseeable damages, rather than imposing full liability on the promisor. It is thustypical to compare the Hadley (or "limited liability" rule) to a full expectation rule(or "unlimited liability").
Default rules have well-recognized functions: by providing background termsthey spare the parties the costs of negotiating "complete" contracts and theyprovide a focus to their negotiations. Assuming for the moment that the onlypurpose of default rules is to minimize transaction costs, the optimal default ruleis the one that on average minimizes the necessity of bargaining, or, what is thesame thing, that maximizes the ex ante value of the contract.
The relative advantages of the Hadley rule and expectation damages dependon which of the forms of behavior mentioned above has the greatest impact onthe ex ante value of the contract. Inefficient breach, for example, is not fullydeterred by the Hadley rule. To see why, consider a promisor (for concreteness,the "seller") who would gain more from breaching than keeping its promise tothe promisee (the "buyer"). Suppose that when deciding whether to breach, theseller accurately anticipates the buyer's loss from breach. If the seller expects tobe required to pay only average damages, it will breach even when the loss tothe buyer from breach exceeds the seller's gains from breach. In contrast, theexpectations measure deters inefficient breach. The reason for the differentoutcomes is that the Hadley rule does not permit compensation for the loss ofthe buyer's private, idiosyncratic value if not disclosed to the seller; but thisloss is real, and if the seller is permitted to ignore it, then it will breach evenwhen the buyer's loss exceeds the seller's gain. The effects of the rules onprecaution -- to take another example -- are more complex. By forcing the buyerto incur part of the loss caused by the breach, any undercompensatory damagesrule, including the Hadley rule, gives the buyer an incentive to take precautionsagainst the loss. There is, however, no necessary relation between the Hadleyrule and the optimal level of buyer precaution. (More on this later.)
Unifying the analysis of the different effects of the Hadley and expectationrules is difficult. It is possible that for some broad class of transactions the exante value of contracts depends mostly on deterring inefficient breach, whereasfor another class of transactions the ex ante value of contracts depends mostlyon deterring insufficient precaution-taking by the promisee. If so, the Hadley ruleshould be enforced more vigorously in the second class than in the first. Thisapproach has not, however, been taken in the literature, and it is not clearwhether it would be fruitful.
The literature has focused instead on the problem of asymmetric information.Suppose that the carrier deals with two kinds of customers: high-value shippers,who suffer a large loss in case of misdelivery, and low-value shippers, whosuffer a small loss in case of misdelivery. The carrier does not know whether aparticular customer is high-value or low-value, but does know the proportion ofeach type in the population. Shippers know their own type. All parties areneutral with respect to risk. Because the loss avoided by investing inprecautionary measures increases from the low type to the high type, the optimalcontract with the high-value shipper would require the carrier to take costlyprecautions, whereas the optimal contract with the low-value shipper wouldrequire the carrier to take few precautions. Suppose that the expectation ruleprevails. With full information the carrier would simply charge a higher fee forthe high-precaution shipments than for the low-precaution shipments, and takethe optimal level of precaution in each case. Without full information, the carrierwould take an intermediate level of precaution and charge an intermediate price,producing two sorts of inefficiency. First, the carrier would take a suboptimallevel of care when making high-value shipments. Second, too few contracts withlow types will occur, because the intermediate price charged for unnecessarycare will drive some of the low types from the market; and too many contractswith high types will occur, because their price is subsidized by the low typeswho remain in the market.
The Hadley literature originated with an influential argument that the Hadleyrule cures these two inefficiencies. The argument asserts that under theexpectation rule the high-value shipper has no incentive to reveal information.Whether or not the high-value shipper reveals its type, the expectation ruleawards it damages sufficient to cover its loss. Because the shipper thereforeexpects full compensation whether the carrier breaches or not, it is indifferent asto whether the carrier breaches or not; and so it does not care whether the carrierminimizes the chance of breach by taking the optimal (high) level of precaution.The carrier can take the optimal level of precaution only if it knows the shipper'svaluation; but because the shipper does not care about whether the carrier takesprecautions, the shipper does not disclose its high valuation. Under the Hadleyrule, in contrast, the high-value shipper is undercompensated if it does notdisclose its type, because the average level of damages is intermediate betweenhigh and low. To ensure adequate compensation in case of breach, the high-value shipper discloses its type. Now "foreseeing" the large loss should itbreach, the carrier will be liable under the Hadley rule for high damages if itbreaches. To limit its expected liability the carrier invests in precautions untiltheir marginal cost equals the marginal gain from avoided loss. Thus does thecarrier invest in optimal precaution. Meanwhile, the carrier can distinguish thelow-value shippers as those who do not state that they belong to the high type.The carrier can therefore charge the low-value shippers a lower fee, also takingthe fewer precautions that are optimal for low-value shippers. Thus are theoptimal number of low-value contracts entered. The Hadley rule cures bothforms of inefficiency. See Barton (1972); more recent treatments include Perloff(1981); Bishop (1983); Quillen (1988); Cooter and Ulen (1988); Ayres and Gertner(1989); Bebchuk and Shavell (1991); Eisenberg (1992); and Posner (1992).
A problem with some of the earlier discussions in the literature results fromtheir focus on the behavior of the high-value shippers to the exclusion of thebehavior of the low-value shippers. The argument tacitly assumes that althoughhigh-value shippers can and do contract out of the Hadley rule, low-valueshippers would not contract out of the expectation rule. If the low-value shipperscannot contract around the expectation rule, then, of course, inefficiencies result.If the low-value shippers can contract around the rule, however, they will. Thereason is that the low-value shippers prefer a package offering a low price, fewprecautions, and low damages, to the package offering intermediate price,intermediate precautions and intermediate damages. Therefore, the low-valueshippers will identify themselves as low-value shippers and offer to agree tolimited liquidated damages in case of breach. Because their liability in case ofbreach of contract with low-value shippers is thus limited, the carriers take fewerprecautions. The carriers also charge the low-value shippers a reduced fee,which attracts back into the market all low-value shippers who value shipmentmore than its marginal cost to the carrier. This cures the problem of too fewcontracts with low-value shippers. Since the carriers know that only high-valueshippers would decline to identify themselves as low-value shippers, the carriersknow to charge a high fee to any shipper that fails to identify itself as a low-value shipper. Because the carriers remain subject to full expectation damageswith respect to the high-value shippers, and because they can accurately predictthe extent of their liability (now that the high-value shippers are no longerpooled with the low-value shippers), they will increase their level of precautionto the optimal level with respect to the high types. This cures the problem ofinefficient precaution. See Perloff (1981); Ayres and Gertner (1989); Bebchuk andShavell (1991).
Barring further analysis or empirical work, the arguments are in equipoise. Tosee more clearly why, observe that to compare default rules, one sums (1) theexpected costs of contracting around the rules (also sometimes called the"communication costs"), and (2) the losses resulting from the failure to contractaround when doing so is not individually cost-justified. Consider one elementof the expected costs of contracting around each rule: the proportion of high andlow types in the population. If low types form a majority, then more contractsrequire disclosure of information under the expectation rule, where low typesbargain around, than under the Hadley rule, where high types bargain around.If it is costly to bargain around contracts by disclosing information, then,everything else being equal, the Hadley rule generates fewer costs. The problemwith the argument, however, is that there is no reason for believing that lowtypes outnumber high types. More likely, the valuations are approximatelynormally distributed, and if a model requires bifurcation of the distribution, thenit should assume that half the types are high and half are low, an assumptionwhich, unfortunately, renders the analysis indeterminate. These considerationsappear to account for the skepticism toward the Hadley rule in Perloff (1981).Ayres and Gertner (1989) and Bebchuk and Shavell (1991) argue informally thatthe Hadley rule is superior when communication is socially optimal, but it is clearfrom their models, as the authors appear to recognize, that this conclusiondepends on the assumption that high types form a minority.
Consider now the costs of bargaining around a default rule -- the costs ofcommunicating one's type and of negotiating and drafting the contracts.Suppose that the low types and the high types are equal in number. The Hadleyrule is favored if the cost to low types is systematically higher than the cost tohigh types. There is no reason, however, to assume that the different typeswould typically incur different bargaining costs. As an aside, note that if thebargaining costs are trivial, never preventing parties from bargaining around,then the entire default analysis becomes irrelevant, since the parties can alwayschoose the optimal terms.
Finally, when the cost of communication exceeds the gains produced whenthe carrier takes the optimal level of precaution rather than the intermediate levelof precaution, it is socially suboptimal for communication to occur, but undercertain conditions the Hadley rule, but not the expectation rule, will cause thehigh types to communicate. See Bebchuk and Shavell (1991). To see why,imagine that the equilibrium obtains in which only the high types communicate.If a high type were to deviate, then it would save the costs of communication,but it would incur an expected loss as a result of the carrier's move from high tolow precautions. The high type would stop communicating its type only if theformer exceeded the latter. But from the perspective of social welfare, therelevant comparison is between the savings in communication costs, on the onehand, and the loss resulting from the move from the high level of precaution tothe intermediate level of precaution. The reason is that if no high typescommunicated, then the carrier would offer the intermediate package ofprecaution, price, and damages, not the low package. It is thus possible thatcommunication costs for high types are high enough that no communication issocially optimal, but also low enough that no high type would decline tocommunicate, given that all other high types are communicating. To be sure, thelow types are injured by a move from a communication to a no-communicationequilibrium, since they communicate in neither case and prefer the package ofprice, precaution, and damages in the first case. But under the right conditionstheir injury is smaller than the gain to the high types.
In sum, although a cursory reading of the literature may suggest otherwise,the information revelation arguments do not clearly favor the Hadley rule or theexpectation rule. The Hadley rule dominates the expectation rule only underspecial conditions, and there is no reason to believe that these conditionsgenerally prevail or that, when they do prevail, courts can reliably identify themfor the purpose of applying the rules. The indeterminacy is not surprising, giventhat the Hadley articles so far discussed rely on signaling models, and thedifficulty of ranking equilibria produced in signaling models is well-known.
Another branch of the Hadley literature also assumes asymmetric informationbut assumes further that the party without the private information has marketpower. Suppose the carrier has market power and the shippers do not. Theexistence of market power changes the high-value shippers' calculations underthe Hadley rule. If they reveal their type, they will expect the carrier to respondby charging them a supracompetitive price. If they do not reveal their type,however, they will not gain the benefit of full compensation in case of breach.In weighing these costs, shippers will sometimes choose not to revealinformation, resulting in the precaution and number-of-transactionsinefficiencies discussed above. Wolcher (1989); Johnston (1991). In contrast,under the expectation rule, the low-value shippers lose nothing from revealingtheir type. Indeed, they gain. Because the carrier can price discriminate againstthe high types only if it can identify them, and it can identify them only if thelow-value shippers disclose their type, the carrier will offer the low types a lowerprice in order to encourage such disclosure. The parties thus distinguished, thecarrier will offer each type a different package of price, compensation, and care.On these grounds Johnston (1990) prefers the expectation rule to the Hadleyrule.
The argument does not, however, show that the expectation rule dominatesthe Hadley rule under general conditions. To see why, observe that the carriercan force the parties to reveal their types only by offering alternative contracts,one of which (the "high" contract) the high types prefer to the other (the "low"contract), but the other of which the low types prefer to no contract at all. At thesame time, the carrier will maximize profits by exploiting its monopoly power, andthis means charging prices that are as close to the types' valuations as possible.To encourage the low types to accept the low contract, the carrier must givethem a price below their valuation; but if this price is too low, the high types willchoose the low contract over the high contract, frustrating the carrier's attemptto maximize profits. To make the low contract sufficiently unattractive to the hightypes, the carrier will set the damages level at a low level. Obviously, the carrierwould not set the damages level above the low type's valuation; but it can alsobe shown mathematically (see Ayres and Gertner (1992), pp. 768-69) that thedamages level in the low contract will be lower than the low type's valuation.(The intuition for this result is that a high damages level for the low contractwould drive the price for the low contract down, leading to a greater differencein prices for the high and low contracts, which would attract the high types tothe low contract, but this would prevent the carrier from price discriminatingagainst them.) But if the damages level in the low contract is less than the lowtype's valuation, then the carrier will take an inefficiently low level ofprecautions. This efficiency loss occurs under both the expectations and theHadley rule. The relative size of the efficiency loss depends on a variety offactors, such as the parties' relative valuations, the proportions of types in thepopulation, and so on, so that one cannot describe abstract conditions underwhich each rule produces better outcomes.
The analysis of Hadley thus plummets into theoretical indeterminacy, but afew routes may lead out of this impasse. First, greater attention to historical andcurrent commercial practices may reveal that some common assumptions areimplausible and can safely be ignored in analysis. Legal and commercialconstraints, for example, may inhibit the exercise of market power by carriers andother promisors. The empirical efforts of Danzig (1975), Landa (1987), Epstein(1989), and Johnston (1990) provide a foundation for such work. Second, greaterattention to institutional issues may put helpful constraints on the analysis. Ifcourts have general knowledge about valuations in an industry, if they canreliably distinguish between cases in which transaction costs are high and casesin which they are low, and if they can detect market power, then they shouldapply a rule that makes the extent of liability turn on all these factors; but if theycannot, then perhaps a bright-line rule should be used, such as unlimitedliability, on the grounds that it would put the least demands on their abilities.
The mitigation doctrine requires the promisee to take steps to reduce the lossfrom breach after it learns of the breach or acquires reason to know of it. If thepromisee fails to take such steps, it will not recover full expectation damages.Instead, it will receive damages sufficient to compensate it for the loss that itwould have incurred if it had taken the proper steps to mitigate. In this way, themitigation doctrine, like the Hadley rule, is an undercompensatory damagesremedy. Also like the Hadley rule, the mitigation doctrine is a default rule.
The buyer (that is, the promisee) often has a chance to minimize the lossresulting from breach. For example, upon learning that the seller will not make atimely delivery of components that the buyer plans to use in producing goods,the buyer can consider buying replacements on the market and using theminstead, or it can suffer a delay in production. If it does not purchasereplacements, then it will lose profits from the delay in production. If it doespurchase replacements and their price exceeds the contract price of the promisedgoods, then it will incur a loss equal to the difference in price. Mitigation refersto the buyer's choice of the response that minimizes the loss resulting frombreach.
The most thorough analysis of the mitigation doctrine is that of Goetz andScott (1983). See also Wittman (1981), and the comments in Bebchuk and Shavell(1991). Suppose that renegotiation is prohibitively expensive and that nomitigation rule exists, and that after the seller breaches, the buyer has theopportunity to reduce its expected loss from $100 to $50 by taking some action,like purchasing a substitute on the spot market. (For simplicity, the cost ofmitigation is built into the loss; this could also be understood as the buyerincurring a $10 expense to reduce the expected loss from breach to $40.) We alsoassume throughout that the buyer can more cheaply reduce the expected lossthan the seller can; otherwise, there is no problem, since the seller has everyincentive to reduce its own liability. Under the rule of expectation damages, thebuyer expects to be fully compensated for its loss regardless of whether itmitigates. If it mitigates properly, then it will incur only a $50 loss, but it willreceive only the $50 damages necessary to compensate it. If it fails to mitigateand so loses profits, then it will incur a $100 loss, but it will receive the $100 indamages necessary to compensate it. Thus, the buyer has no incentive toengage in mitigation in the absence of the mitigation rule. But clearly mitigationis efficient: it reduces net losses by $50. The reason the buyer does not mitigateis that the seller enjoys the gains while the buyer incurs the costs. To give thebuyer the incentive to mitigate, it must be penalized when it fails to do so. Themitigation rule produces this penalty by imposing on the buyer the loss that theseller would otherwise incur from the buyer's failure to mitigate. If the buyer failsto mitigate, it would incur a loss of $100 but receive damages of only $50; if itmitigates, the $50 damages would fully compensate it for the $50 loss. Therefore,the buyer would mitigate.
Notice that the buyer has an incentive to mitigate as long as it does notrecover actual damages and that the amount it does receive is unrelated to actualloss. While the mitigation doctrine performs this function, so would anydamages rule that provides an amount that is invariant with respect to the actualloss -- for example, a rule of zero damages.
The argument is less straightforward if renegotiation costs are low. In theabsence of the mitigation rule, the seller could, at the time of breach or inanticipation of breach, offer to pay the buyer to mitigate. Since the seller bearsthe full costs of the failure to mitigate, it will pay the buyer to the extent that thegains from mitigation exceed its costs; and as long as this is the case, the buyerwould be willing to accept the payment and to engage in mitigation. In theexample, the seller would offer to pay the buyer up to $50 to mitigate, sincemitigation would reduce seller's liability by $50. The efficient outcome isobtained. The only difference between this result and the result under themitigation rule is that the cost of mitigation is fully born by the buyer under themitigation rule, whereas in the absence of the rule the cost of mitigation is fullyor partly born by the seller, depending on its bargaining power.
This argument, at first glance, suggests that the value of the mitigation ruledepends on its effect on incentives to engage in ex post renegotiation. Onemight argue that the no-mitigation rule creates worse incentives to renegotiate.Under the no-mitigation rule the buyer has strong incentives to hold out for alarge payment from the seller and the seller has strong incentives to pay as littleas possible, so that the gains from mitigation would be dissipated as the partieshaggle. This problem, however, is offset by an advantage of the no-mitigationrule, namely, that it forces the seller to disclose an anticipated breach at theearliest possible moment. In contrast, under the mitigation rule the buyer mustguess the likelihood and time of breach, and take steps to mitigate in light of itsincomplete information. For discussions of the complex problems of the timingof breach, and of repudiation and cure, see Jackson (1978), Goetz and Scott(1983), and Craswell (1990).
Even supposing low renegotiation costs, the mitigation and the no-mitigationrules have different effects on behavior. Consider, for example, the parties'incentives to take precautions against loss. The seller has higher expected costsunder the no-mitigation rule than under the mitigation rule, because under theformer it must pay the buyer to mitigate in case of breach. These costs willencourage the seller to take precautions against the loss from breach; but at thesame time the expected gain to the buyer in mitigation payments in case ofbreach will deter the buyer from taking precautions. Under the mitigation rule,the buyer no longer has this disincentive to take precautions, but now the seller,because it no longer expects to have to pay the buyer to mitigate, has lessreason to take precautions. Another example of the different effects the ruleshave on the parties' behavior concerns the allocation of risk. Under the no-mitigation rule, the seller charges the buyer a premium to make up for theexpected cost of paying the buyer to mitigate. The premium can be thought ofas the cost of a lottery ticket that gives the buyer the chance, equal to theprobability of breach, to gain the value of the premium. There is no reason tothink that a buyer would be willing to purchase this lottery ticket. Risk-aversebuyers do not want to take on risk, and risk-preferring buyers can gamble moreprofitably by visiting a casino. Cf. Craswell (1990). The no-mitigation rule createsan added cost to contracting, and this cost will deter people from enteringotherwise value-maximizing contracts.
If a mitigation rule seems superior to a no-mitigation rule, some authorsdoubt that courts apply the mitigation rule in a proper way. For example, somecourts award the buyer the difference between the contract price and the marketprice of the substitutes without taking into account the possibility that the buyershould mitigate also by purchasing fewer substitutes or substitutes of lowerquality. But when prices of goods rise, users of them generally should reduceconsumption. See Fenn (1981). Another example is the practice of refusing tocompensate buyers who engage in high-risk mitigation strategies that result inno actual loss reduction, or allowing full compensation of buyers who declineto engage in mitigation strategies that require high capital outlays. AsMacIntosh and Frydenlund (1987) point out, the net present value of amitigation strategy, not its riskiness or costliness, determines its suitability.These criticisms, however, do not undermine the mitigation rule so much as raisethe issue of whether it should be enforced as a series of strict liability rules, asreflected in the doctrines criticized, or as a negligence-like standard as theauthors implicitly argue. A standard taxes judicial competence, but whether thiscost outweighs the systematic distortions caused by rules is not known.
To return to more general issues, one might ask why the mitigation analysisand the Hadley analysis follow such different paths. The disclosure ofinformation, which is the focus of the Hadley analysis, and the mitigation oflosses, which is the focus of the mitigation analysis, are but aspects of promiseebehavior that minimizes the expected loss from breach. Indeed, the Hadley ruleencourages the promisee to mitigate post-breach whenever the"unforeseeability" of the magnitude of its prospective loss threatens it with anundercompensatory remedy. And the mitigation rule encourages the promiseeto disclose information prior to contracting whenever the disclosure ofinformation would enable the promisor to take precautionary steps that are morecost-effective than post-breach mitigation by the promisee. Both the Hadley ruleand the mitigation rule threaten the promisee with undercompensatory damagesin order to encourage it to engage in loss-minimizing behavior: in this way, theyaddress the same problem in a similar manner. The convenient but artificialdistinction in the doctrine has created an unjustified divergence in scholarlyanalysis.
A "precaution" can be defined as any act by a person that reduces the expectedloss from breach of contract. A person can take precautions by using care inselecting a contracting partner, by disclosing information to that partner, bydrafting the contract in clear and simple language, by purchasing insuranceagainst loss from breach, by modifying its capital investments to minimize theloss from breach, by mitigating, by renegotiating, and so on -- in short, by takingany of the value-maximizing steps that were discussed in prior sections. Oneshould take care in reading the literature, because commentators use the word"precaution" in different ways. Goetz and Scott (1980) limit their discussion ofprecaution to the promisor's use of clear contracting language that encompassesan adequate quantity of contingencies. Cooter (1985) limits his discussion ofprecaution to efforts made by the promisor and the promisee to minimize theexpected loss from breach of contract between the signing of the contract, onthe one hand, and performance or breach, on the other. Because we have alreadydiscussed disclosure of information and mitigation, and because discussion ofthe other aspects of precaution can be found in section 4600, we follow Cooter.
The only sustained discussion of precautionary behavior as an aspect ofcontract law (it appears routinely, however, in commentary on the law ofproducts liability and, of course, on the law of torts generally) is that of Cooter(1985). See also Cooter and Ulen (1988). In Cooter's model, a promisor agrees todeliver goods at a certain time. The promisor and the promisee choose whetherto take precautions, and the jointly maximizing behavior of each party is to takeprecautions. For example, if the promisor agrees to rent warehouse space to thepromisee at a future date, the promisee can take a precaution against breach byreserving storage space at another site or by letting its inventory run down. Theprecaution is costly to the promisee, but we suppose that it reduces the expectedloss from breach by more than it costs the promisee. It is therefore efficient forthe promisee to take the precaution, in the sense that the gains from theprecaution exceeds the cost. Indeed, a complete contract would require thepromisee to take the precautions. When transaction costs prevent the partiesfrom describing the promisee's duty to take a particular precaution, however, thepromisee would not take the precaution in the absence of a legal rule compellingit to do so. The reason is that it incurs the costs of the precaution but thepromisor enjoys the gains. A legal rule that forced the promisee to take theprecaution might therefore seem desirable.
Similarly, the promisor can incur costs to lower the expected loss frombreach. For example, the promisor can take a precaution against breach by hiringa guard to protect the premises, so that the warehouse will not be destroyed byfire. If the cost of hiring the guard is less than the expected loss, the promisorwould maximize the value of the contract by hiring the guard. A legal rule shouldforce the promisor to take such a precaution.
What kind of rules would serve these purposes? For the promisor,expectation damages would ensure the optimal level of precaution. Sinceexpectation damages force the promisor to bear the promisee's loss, the promisorwould take precautions to the point where their marginal cost equals themarginal expected reduction in the promisee's loss. The problem withexpectation damages, however, is that they would discourage the promisee fromtaking precautions. The reason is that if the promisee is fully compensatedwhether or not it takes precautions, it has no incentive to take them. Anotherrule is necessary for encouraging the promisee to take precautions.
One possibility is a rule that awards zero damages. To see why, imagine thatthe breach would result in an expected loss of $100 if the promisee fails to takea precaution, and in an expected loss of $50 if the promisee takes a precautionthat costs it $10. Under the zero damages rule, the promisee expects to bear thefull loss, and so would take the $10 precaution in order to reduce the loss from$100 to $50. The problem with zero damages, however, is that it gives thepromisor no incentive to take precautions. Thus, there is a tension betweencreating incentives for the promisor to behave properly and creating incentivesfor the promisee to behave properly.
One possible resolution is a rule, analogous to the mitigation rule, that wouldrequire the promisee to take precautions whenever they produce gains greaterthan their costs. Such a rule might hold that the promisee must take "reasonable"precautions against breach, meaning that the promisee suffers a reduction indamages equal to the amount of loss it could have avoided by takingprecautions minus the cost of those precautions. Under this rule the promiseewill take cost-justified precautions in order to avoid the reduction in damages.Then the promisor would bear any residual losses caused by its (the promisor's)failure to take cost-justified precautions, and to avoid these costs the promisorwould take precautions as well.
Another possibility is a rule that limits the promisee's damages to an amountthat is invariant with respect to its level of reliance. We already saw that a rulethat gave the promisee zero damages would cause it to use the precaution,because by doing so it incurs a $10 cost in order to reduce its losses by $50.Similarly, a rule that gave the promisee $50 of damages would cause it to use theprecaution, because the $10 cost of the precaution is less than the reduction inthe uncompensated loss from $50 (i.e., $100 loss minus $50 damages) to $0 (i.e.,$50 loss minus $50 damages). Even a rule that gave the promisee $200 indamages would cause it to use the precaution. The promisee expects to gainmore from taking the precaution (-$50 - $10 + $200 = $140) than by failing to takethe precaution (-$100 + $200 = $100). The crucial distinction between this ruleand expectation damages is that under the former the promisee benefits fromtaking precautions, whereas under the latter the promisor benefits when thepromisee takes precautions. The promisee can be forced to internalize the costof failing to take precautions only by a rule that makes its damages awardinvariant with respect to the amount which it invests in precautions. To be sure,the $200 rule, like the $0 rule, would create improper incentives for the promisor.It would cause the promisor to take too many precautions and to perform whenit should breach. The optimal damages rule with respect to both promisor andpromisee incentives would require an amount equal to what would be necessaryto put the promisee in the position it would have been in if the promisor hadperformed and the promisee had taken efficient precautions.
The reader might recall from section 4600 that a similar conclusion wasreached with respect to the question of promisee "reliance." The connectionbetween precaution and reliance is indeed very close. Scholars often talk of thepromisee engaging in too much reliance or engaging in too little reliance onpromises, by which they mean that the promisee invests more or less than thejointly value-maximizing amount between the signing of the contract and thecompletion of performance. The notions of promisee reliance and promiseeprecaution capture similar phenomena but from opposite directions: a promiseewho takes too few precautions relies too much; a promisee who takes too manyprecautions relies too little. Commentary on reliance has focused on the behaviorof the promisee in relation to the promisor's incentive to breach, whereasanalysis of precaution has dealt with the precautionary behavior of both thepromisor and the promisee. But the conclusions of the two analyses are thesame: the remedy of expectation damages is suboptimal because it allows thepromisee to externalize costs on the promisor. The optimal damages rule wouldgive the promisee the value of the expectation it would have had if it hadengaged in the optimal level of reliance or -- what is the same thing -- the optimallevel of precaution.
This rule, however, does not exist. No rule straightforwardly encouragespromisees to take efficient precautions. To be sure, promisees may fear that theHadley rule and the mitigation rule will result in undercompensatory damages,and thus take precautions in order to limit their expected losses. Cf. Cooter(1985). But neither the Hadley rule nor the mitigation rule directs courts to takeaccount of a promisee's precautions other than those of disclosing informationpre-contract and mitigating damages post-breach. Cooter (1985) notes thatparties can contractually limit damages to encourage the promisee to rely, butparties are free (usually) to contract around any damages rule. The question iswhy there is no default rule that, on an analogy to the Hadley rule and themitigation rule, encourages precautionary behavior when transaction costsprevent the parties from stipulating damages in advance. There is not anobvious answer, but one possible direction of research would inquire intowhether courts are more competent at making some kinds of evaluations, suchas the reasonableness of mitigation, than others, such as the propriety ofprecautionary behavior.
Causation is not an important concept in Anglo-American contract law. This ispartly because the concept of causation is captured by other contract doctrines.If the promisee fails to mitigate, one might say that the promisor did not causethe entire loss, and therefore should not be fully liable; but this result isproduced through application of the mitigation doctrine, and a separate"causation doctrine" is not necessary, as it is in tort law. In only a handful ofcontract cases do courts discuss causation as a distinct issue. These casesinvolve a promisee who has separate contracts with two different breachingpromisors, both of which breaches are sufficient to cause some or all of the loss.These cases have not attracted the attention of scholars, probably because oftheir rarity. The tort doctrine of proximate causation has an analogy in theHadley doctrine and the doctrine of "reasonable certainty." All of thesedoctrines release the wrongdoer from some or all liability when the victim's lossis not entirely attributable to the promisor's wrongdoing.
Two other doctrines are related to the themes that have been discussed. Thefirst limits expectation damages to an amount sufficient to compensate thepromisee for a loss that is "reasonably certain." The second denies promiseescompensation for emotional distress arising from a breach of contract. Bothdoctrines, like the Hadley rule and the mitigation rule, are undercompensatory.
The "reasonable certainty" doctrine requires the promisee to prove its loss witha greater degree of certainty than that required by the preponderance ofevidence test that governs the other elements of civil actions. Somecommentators argue that, in practice, the courts use the doctrine in a highlyflexible way to prevent recovery of damages whenever they seem toovercompensate the promisee. Possibly, courts use the doctrine to punishpromisees who engage in too much reliance, take too few precautions, ormisbehave in other ways. This would support the analysis of Cooter (1985),described above.
Another possibility is that the "reasonable certainty" doctrine actually doeshave analytic bite, and serves the purpose of preventing parties fromexternalizing the cost of resolving disputes on the courts. Unlike the injurers andvictims under tort law, which has no analogous doctrine, the parties to a contractcan anticipate a legal dispute and take steps in advance of performance tofacilitate adjudication. For example, if they anticipate that a loss, such as the lossof good will or the loss of personal enjoyment of the goods or services, will behard for a court to measure, they can arrange for a liquidated damages provisionto specify that loss in advance. To deter parties from externalizing the cost ofdetermining loss onto the courts, and to encourage them instead to supply aliquidated damages provision, a doctrine denying compensation for uncertainloss might be justified. See section 4600 for discussion of the closely relatedissue of subjective loss.
The "emotional distress" doctrine raises issues similar to those raised by the"reasonable certainty" doctrine, but it raises one additional issue. The tortliterature suggests that the case for compensation for emotional distress isuneasy. On the one hand, it is not clear that people would purchase insuranceagainst nonpecuniary loss, because the marginal utility of money in the accidentstate of the world, once pecuniary losses are compensated, is not necessarilyhigher than the marginal utility of money in the non-accident state of the world.On the other hand, if the promisor is not forced to pay damages for emotionaldistress, it will have an incentive to take too few precautions against breach. Fordiscussion of these issues, see section 5140.
The author thanks Richard Craswell, Steven Shavell, and two referees for helpfulcomments.
Bibliography on Contract Remedies: Foreseeability, Precaution, Causation, andMitigation (4620)
Ayres, Ian and Gertner, Robert (1989), 'Filling Gaps in Incomplete Contracts: An EconomicTheory of Default Rules', 99 Yale Law Journal, 87-130.
Ayres, Ian and Gertner, Robert (1992), 'Strategic Contractual Inefficiency and the OptimalChoice of Legal Rules', 101 Yale Law Journal, 729-773.
Bebchuk, Lucian Ayre and Shavell, Steven (1991), 'Information and the Scope of Liabilityfor Breach of Contract: The rule of Hadley vs. Baxendale', 7 Journal of Law,Economics and Organization, 284-312.
Bishop, William (1983), 'The Contract-Tort Boundary and the Economics of Insurance',12 Journal of Legal Studies, 241-266.
Cooter, Robert D. (1985), 'Unity in Tort, Contract, and Property: The Model ofPrecaution', 73 California Law Review, 1-51.
Danzig, Richard (1975), 'Hadley v. Baxendale: A Study in the Industrialization of the Law',4 Journal of Legal Studies, 249-284.
Eisenberg, Melvin Aron (1992), 'The Principle of Hadley v. Baxendale,' 80 California LawReview, 563-613.
Epstein, Richard A. (1989), 'Beyond Foreseeability: Consequential Damages in the Law ofContract', 18 Journal of Legal Studies, 105-138.
Fenn, Paul (1981), 'Mitigation and the "Correct" Measure of Damage', 1 InternationalReview of Law and Economics, 223-226.
Goetz, Charles J. and Scott, Robert E. (1983), 'The Mitigation Principle: Toward a GeneralTheory of Contractual Obligation', 69 Virginia Law Review, 967-1024.
Johnston, Jason S. (1990), 'Strategic Bargaining and the Economic Theory of ContractDefault Rules', 100 Yale Law Journal, 615-664.
Landa, Janet T. (1987), 'Hadley v. Baxendale and the Expansion of the MiddlemanEconomy', 16 Journal of Legal Studies, 455-470.
MacIntosh, Jeffrey G. and Frydenlund, David C. (1987), 'An Investment Approach to aTheory of Contract Mitigation', 37 University of Toronto Law Journal, 113-182.
Perloff, Jeffrey M. (1981), 'Breach of Contract and the Foreseeability Doctrine of Hadleyv. Baxendale,' 10 Journal of Legal Studies, 39-63.
Quillen, Gwyn D. (1988), 'Contract Damages and Cross-Subsidization', 61 SouthernCalifornia Law Review, 1125-1141.
Wittman, Donald (1981), 'Optimal Pricing of Sequential Inputs: Last Clear Chance,Mitigation of Damages, and Related Doctrines in the Law', 10 Journal of Legal Studies,65-91.
Wolcher, Louis E. (1989), 'Price Discrimination and Inefficient Risk Allocation Under theRule of Hadley v. Baxendale', 12 Research in Law and Economics, 9-31.
Barton, John H. (1972), 'The Economic Basis of Damages for Breach of Contract', 1Journal of Legal Studies, 277-304.
Cooter, Robert and Ulen, Thomas (1988), Law and Economics, Glenview, Scott, Foresman,644 p.
Craswell, Richard (1990), 'Insecurity, Repudiation, and Cure', 19 Journal of Legal Studies,399-434.
Fuller, Lon L. and Perdue, William (1936), 'The Reliance Interest in Contract Damages', 46Yale Law Journal, 52-98.
Goetz, Charles J. and Scott, Robert E. (1980), 'Enforcing Promises: An Examination of theBasis of Contract', 89 Yale Law Journal, 1261-1300.
Jackson, Thomas H. (1978), '"Anticipatory Repudiation" and the Temporal Element ofContract Law: An Economic Inquiry into Contract Damages in Cases of ProspectiveNonperformance', 31 Stanford Law Review, 69-119.
Posner, Richard A. (1992), Economic Analysis of Law (4th ed.), Boston, Little Brown, 722p.
Hadley v. Baxendale (1854), 9 Ex. 341, 156 Eng. Rep. 145.
© Copyright 1997 Eric A. Posner