Edited by Gerrit De Geest
Chapter 8: Interpretation and implied terms in contract law
[In: Volume 6, Gerrit De Geest (ed) Contract Law and Economics]
1. Introduction and Scope
Questions of how courts interpret, and should interpret, contract terms and when courts imply, and should imply, terms to which the contracting parties have not explicitly agreed loom large in contract disputes and in the legal literature on contract law. In the last decade, these questions have received increased attention from law and economics scholars.
Contract law draws a distinction between interpretation and implied terms, as well as between these doctrines and other contract law doctrines. Interpretation refers to the process by which courts ascertain the meaning of express terms. Implied terms are those that courts deem to be part of the contract even though the parties did not expressly agree to them. Implied terms are sometimes used to ‘fill gaps’ that parties have left in their contracts, such as when parties leave out the time of performance and courts read in a reasonable time term. Implied terms can also refer to terms that limit the application of existing terms, the most notable example being the doctrine of good faith. Interpretation and implied terms are closely related concepts. For example, if the question is whether to read in an exception to an express term, such as a price or quantity term, that could be viewed either as an act of interpreting the express term or of implying an additional term (the exception). Similarly, if a contract contains a ‘best efforts’ clause, determining what that clause requires is a question of interpretation, although the specific content a court reads into such a vague term could easily be viewed as an act of implication. On the other hand, if the contract contains no such clause, a court may have to decide whether to imply a best efforts obligation, and if it does, it has to determine the content of that obligation, which may involve considerations similar to those for interpreting an express best efforts clause. Some questions of interpretation, however, are typically viewed as not involving questions of implication. An example would be a dispute over the meaning of the word ‘chicken’.
Although a contract need not be written to raise questions of interpretation and implied terms, many discussions of these topics assume a written contract, which is perhaps more common than oral contracts. Written p. 126contracts often call into play the parol evidence rule, which renders inoperative certain agreements made prior to, or in some cases contemporaneously with, a final, or ‘integrated’, writing. Although the parol evidence rule is technically neither a doctrine of interpretation nor implied terms, but rather determines which terms are part of an enforceable contract,1 it often is discussed in conjunction with doctrines of interpretation and involves many similar considerations and so will be included here.
In some sense, all contract disputes involve questions of interpretation and implied terms. Most contract law doctrines, such as formation, excuse, and damage rules, can be viewed as implied terms. And express terms that add to or trump background doctrines often require interpretation. But contract law has generally used the labels ‘interpretation’ and ‘implied terms’ more narrowly, to refer to questions of contract performance and breach, rather than questions of formation, excuse, defense, or remedy. That is, the legal issue addressed by these doctrines is whether one or more parties have performed as the contract requires, or have breached. The key exception is the doctrine of indefiniteness, which deems unenforceable certain contracts whose gaps courts decline to fill in, and so is often viewed as a formation doctrine. Even the indefiniteness doctrine, however, is a performance doctrine to the extent that it concerns only missing performance terms. Courts never hold a contract to be indefinite because it lacks a liquidated damages clause, for instance. Other than the indefiniteness exception, I will generally follow the conventional focus on performance and breach here, recognizing that many of the economic arguments have broader application to other doctrinal areas of contract law.
2. Complete or Incomplete Contracts
Economic analyses of contract law tend to start with the idealized concept of a ‘complete’ contract, though this term has perhaps engendered more confusion than clarity. Traditionally, a complete contract refers to one that provides a complete description of a set of possible contingencies and explicit contract terms dictating a performance response for each of these contingencies.2 Contingencies include changes in ‘exogenous’ economic variables, such as a production cost increase. But they also include ‘endogenous’ behavioral responses, such as falsely claiming a cost increase or seeking refuge from a now-disadvantageous bargain behind a contract term intended to serve a different purpose. Economic analyses generally conclude that if a contract is complete, there is no efficiency-enhancing p. 127role for a court other than to enforce the contract according to its terms; that is, incompleteness is a necessary, though not sufficient, condition for an active court role in interpretation and implied terms.
But because no real-world contracts are fully complete in this sense, the concept of completeness does not get us very far. The concept can be rescued in one of three ways. One way is to view completeness as a useful theoretical benchmark, similar to perfect competition. Just as some markets are close enough to being perfectly competitive that the perfect competition model is a useful predictor, so some contracts, or at least the contract terms at issue in a dispute, may be complete enough that no reasonable interpretation or implied term questions arise. Stated this way, however, completeness comes close to being simply a tautology. The question is what criteria can courts use to determine when a contract or term is complete in this sense. Shavell defines a ‘specific term’ as one that identifies a particular action for a given contingency and a ‘fully detailed complete’ contract as one that has a specific term for each contingency.3 But a ‘contingency’ can itself be defined with more or less specificity and we still need a criterion for deciding what the optimal level of specificity is.
A second way to rescue completeness is to recognize that contracting parties can make a contract complete by using general ‘catchall’ clauses that state what happens in all unspecified states of the world.4 If these ‘general terms’ cover all possible contingencies, the contract can be considered ‘obligationally complete’.5 For example, a catchall clause might state: ‘The price term will be x, and will apply regardless of any change in circumstances or conduct by either party.’ Alternatively, the parties could include a catchall term that dictates an interpretive methodology that makes the contract complete. For example, the parties might include a clause that directs courts not to interpret or imply terms, either generally or in circumscribed ways. But although contracting parties often use merger clauses, which direct a court to apply a particular interpretive methodology (that is, do not look beyond the writing), they do not seem to use express catchall clauses that are broad enough to make contracts complete (though there is some question whether broadly stated ‘non-interpretation’ or ‘non-implication’ clauses would be enforceable). Even if parties did write catchall clauses, those clauses themselves would often require interpretation.6 Moreover, contracting parties often use p. 128contracting clauses that are the exact opposites of completeness catchalls: general clauses such as ‘good faith’ or ‘best efforts’ clauses signal contracting incompleteness, as opposed to completeness.7
Of course, clauses that are not expressly stated as catchalls could be – and sometimes are – understood that way. For example, Schwartz and Scott (2003, p. 573) seem to have this notion of completeness in mind when they define it to mean that the ‘writing expresses the parties’ solution to the contracting problem at issue’. Formalism or textualism, discussed below, commonly assumes this form of completeness.8 But reading an express contractual term as a catchall, that is, applicable in all states of the world without exception, is itself an act of interpretation or implication that requires justification, at least in the legal realm.9
A third way to rescue completeness, more common in formal economic modeling, is to tie the concept of completeness to the efficient use of available information. A complete contract is one that makes full use of the private information available to the contracting parties, even if it does not expressly specify the response to every contingency.10 But the fact that parties may in a simplified model be able to write ‘economically complete’ contracts does not answer the question of whether in a given legal dispute they have in fact written one. And the ability and willingness of private parties to write economically complete contracts in the real world is unclear. We do not seem to see, for example, contracts of the type described by Hermalin and Katz, in which the contract leaves the quantity and price unspecified, and then after some period one party names the price and the other names the quantity.
It seems fair to say, however, that many if not most contracts are incomplete, or at least the question of their completeness is itself a legitimate question for judicial interpretation. A more limited claim would be that incomplete contracts make up a large share of disputed contracts. Incomplete contracts may nevertheless be efficient contracts. The costs of contractual completeness would often exceed the benefits, just as the costs of reducing crime or pollution or accidents to zero would exceed the benefits.
Scholars have offered numerous reasons why the costs of contractual completeness are often high, leading parties to write incomplete contracts. Many emphasize the direct costs of negotiating and drafting complete p. 129contracts. The future is unpredictable and identifying and contracting over how parties should respond to remote contingencies may not be worth the time and effort,11 or may simply be beyond the capabilities of parties who are only boundedly rational.12 Even if drafting more detailed terms would be relatively cheap, parties may intentionally make certain contract terms ambiguous to sidestep contentious issues, which could blow up the deal.13 Alternatively, parties may have asymmetric information and one party may strategically withhold information that would facilitate more complete contracting.14 Parties may also find certain contract terms too costly to monitor or enforce ex post.15
Incompleteness and ambiguity may also result from unintended ‘formulation error’, which is costly to avoid.16 In fact, such error may be more common in complex contracts with many terms. Parties may find it difficult or undesirable to keep track of all the terms and ensure their internal consistency. They may not want to revisit terms already agreed on because that may open up new areas of disagreement. Agency cost problems may also lead to error from unintended incompleteness. Lawyers may fail to ‘clean up’ negotiated contracts because they may not want to risk having their mistakes and oversights exposed.17
Not only may the direct costs of achieving contractual completeness be high, but the opportunity costs may be high as well because other ‘governance mechanisms’ might be superior substitutes for ex ante contracting and strict court enforcement.18 One important set of alternative governance mechanisms involves extralegal enforcement, including social sanctions and reputation, ex ante vertical integration, ex post renegotiation, and arbitration. Parties may write incomplete contracts if they think those contracts are likely to be ‘self-enforcing’ so that court enforcement is not necessary and may even be unavailable if the contract is considered too ‘indefinite’.19 Parties may even intentionally draft ambiguous contracts to increase the potential costs of litigation, as a kind of bond against litigating.20 On the flip side, insisting on more complete contract terms may be p. 130taken by a contracting party as a signal that the other party is litigious or untrustworthy.21
The other key alternative governance mechanism is interpretation and implied terms supplied by courts. Courts may be able through interpretation and implied terms to provide the necessary flexibility – efficient adjustments to contingencies – that an incomplete contract otherwise lacks. Courts may be superior to nonlegal institutions such as reputation because reputation effects may be weak due to such things as cognitive dissonance, optimism about the ability of a party with a poor reputation to change, the difficulty of knowing when a contracting partner has behaved badly, and the last period problem.
The question of contract interpretation and implied terms then is really a question of when court use of these devices is a superior governance mechanism for facilitating efficient contracting. From an economic perspective, the nature and scope of court intervention should depend on the extent of and reason for contractual incompleteness, at least if courts can determine these at reasonable cost. In general, the role for courts in interpreting contracts and implying terms expands as contracts become more efficiently incomplete.
3. Incomplete Contracts and Presumed Contractual Intent
As noted above, scholars and courts generally agree that if a contract is complete, courts should enforce the contract according to its express terms, at least barring any concerns with third party interests. But what should a court do if a contract is incomplete, or its completeness is reasonably contested? Economists and courts start from the presumption that courts should follow the intention of the parties. To admit incompleteness, however, is to admit that the intention of the parties is uncertain, or at least disputed. Thus, actual intent will either be costly for courts to ascertain or, as Lipshaw (2005) stresses, nonexistent.
The next best solution is to use presumed or hypothetical intent. Under this approach, courts adopt the term,22 or interpretive methodology,23 the parties would have chosen had they bargained over the matter. A common variant is the majoritarian default approach, under which courts try to identify the terms or methodology most contracting parties would want. But how is presumed or majoritarian intent determined?
There are two main interpretive methodologies on which economic p. 131analyses focus.24 In the first, courts presume that complete contracting is both feasible and desirable. This presumption has both a positive and a negative component. On the positive side, the express terms of the contract, interpreted as catchalls, are presumed to best approximate the parties’ intentions and deemed to create a complete contract. This methodology is usually referred to as textualism or literalism, especially if the presumption is not rebuttable by evidence extrinsic to the express terms. On the negative side, if parties fail to write a complete contract, the incompleteness is presumed to be inefficient, whether unintended or strategic, and the court’s approach should be to deter this behavior and encourage complete contracting.
The second interpretive methodology involves a presumption that contractual incompleteness is unavoidable and/or desirable, for reasons such as those discussed in the previous section. The courts then fill in the gaps or interpret terms by presuming the parties intended to contract with reference to some standard external to the express terms.25 A methodology that relies on these presumptions is usually referred to as contextualist, though contextualism is not a singular methodology. One branch of contextualism focuses on the kind of evidence courts should consider as relevant to proving the parties’ intentions. Sometimes this evidence comes from the parties themselves, whether through testimony about negotiations, or the parties’ conduct under the existing contract (course of performance) or prior contracts (course of dealing). Sometimes, the evidence concerns the conduct and understandings of other similarly situated contracting parties. This evidence includes trade usage or custom, or other norms and fairness conventions.26
Another variation of contextualism focuses not so much on evidence as on economic theory and business practice more generally. Under this approach, courts might presume the parties contracted with the expectation that courts would fill in any gaps with a joint maximizing implied term that would have been written by rational parties under conditions of low transaction costs.27 Similarly, courts may make a ‘best guess’ about which party is the superior risk bearer or what term or interpretation would make ‘commercial sense’.28 The meaning of these concepts can often be contested. Ben-Shahar (2004), for example, argues that if parties intentionally leave gaps, courts should fill these gaps with terms most favorable p. 132to defendants. Geis (2006) challenges this argument on the ground that it may frustrate the intentions of parties who see intentional gaps as ‘embedded options’ either to accept the interpretation of the other party or take a chance that a court will enforce their preferred interpretation.
The choice between textualist and contextualist methodologies often comes down to which presumption better approximates the parties’ intentions, which in incomplete contracts are uncertain and contestable. For example, suppose a buyer rejects goods delivered later than the time of delivery specified in the contract after the market price drops below the contract price, even though the buyer has always allowed late deliveries before. A court might be called upon to decide whether to imply a limitation on the buyer’s ability to reject, perhaps by using the doctrine of good faith. A textualist would argue no on the ground that such implication would be contrary to the parties’ intentions as expressed in the time of delivery term. A contextualist would argue yes on the ground that implying a limitation on the buyer’s right of rejection would best effectuate the parties’ intentions. Economic analysis can help to identify the conditions under which the different interpretive methodologies are more likely to approximate the parties’ intentions, and whether courts are better off pursuing a pure interpretive strategy or a mixed one.
4. Negotiating and Drafting Costs
A key economic argument for an expansive court role in interpreting and implying terms is that court willingness to engage in these practices enables and encourages parties to write less complete contracts than they otherwise would. Writing less complete contracts saves on drafting and negotiating costs so long as the court-supplied interpretations and terms sufficiently approximate the parties’ intentions.29 Thus, if the costs of court interpretation and implication are low, contract law rules that promote such acts facilitate efficient contracting. Even if contracting parties would write incomplete contracts anyway, for reasons independent of the law of interpretation and implied terms, that law may still promote efficiency if it leads parties to view court enforcement as less costly than extralegal enforcement.
Moreover, the higher the ex ante transaction costs of drafting and monitoring, the more likely it will be efficient for a court to adopt broader rules of interpretation and implied terms that encourage parties to contract less explicitly, because it will more likely be cost-effective for the parties to rely on contextual evidence such as trade usages. If courts take too restrictive p. 133a view of interpretation and implied terms, the development of cost-saving interpretive devices might be discouraged in favor of more complete, but costlier, writings.30 Alternatively, too few contracts might be formed ex ante, as the promisor’s costs rise to cover an anticipated remedy that the promisee does not value at this cost. And too much performance might occur ex post, as the promisor performs even when the cost of doing so exceeds the value of performance.31
One approach courts can take is to identify categories of contracts in which ex ante contracting costs appear to be high. A classic example of high-transaction costs contracts is principal-agent contracts usually referred to as ‘fiduciary’. These contracts typically involve complex tasks for which the principal cannot easily measure the agent’s effort or outcome, thus making express contracting difficult.32 Other categories of high contracting cost situations might include contracts between unsophisticated parties or long-term contracts. But problems of incompleteness and ambiguity can arise in any contract, despite the best efforts of sophisticated parties and their lawyers.
5. Litigation Costs
Although the reduction of ex ante contracting costs is a potentially large benefit of an expansive court role in interpretation and implied terms, that benefit must be balanced against the costs. One potentially significant cost is the cost of litigation. Law and economics scholars often argue that contextualism is associated with higher litigation costs than textualism. For example, allowing more contextual evidence may encourage parties to spend more on litigation because the marginal benefit of expenditures to develop such evidence is higher than under a textualist regime.33 Alternatively, allowing contextual evidence may undermine certainty and therefore make settlement less likely.34
A number of scholars have argued that the optimal contract rules of interpretation and implied terms are determined by the tradeoff between ex ante negotiation and drafting costs and ex post litigation costs.35 Posner, for instance, posits a simple model in which as parties spend more on drafting a more complete contract, the likelihood of litigation decreases, p. 134as does the cost of any litigation that occurs and the likelihood of court error. On the other hand, as parties spend less on ex ante contracting and rely more on extrinsic evidence to prove their intent, drafting costs go down but expected litigation costs rise. As a result, allowing more evidence is not always desirable; the question is whether the benefits exceed the costs. In balancing contracting and litigation costs, it is important to keep in mind that contracting costs are certain and incurred across all contracts, while litigation costs, though often much larger than contracting costs, are incurred in only a small fraction of contracts.36
Concern with litigation costs helps explain doctrines such as the parol evidence rule, which limits the role of the jury and prevents parties from introducing evidence of prior negotiations or agreements when they have drafted a contract sufficiently complete as to be deemed ‘integrated’.37 In addition, many courts use the four corners rule to determine whether a writing is integrated for purposes of the parol evidence rule or to determine whether a term is ambiguous for the purpose of applying the plain meaning doctrine of interpretation. The four corners rule is a textualist doctrine that bars the use of contextual evidence extrinsic to the writing to prove integration or ambiguity. Judge Posner argues that the four corners rule is based on the assumption that parties prefer ex ante contracting to the expense and uncertainty of a jury trial.
Schwartz and Scott argue that if courts adopt a contextualist methodology, the higher litigation costs that will ensue under that regime will have an additional, indirect effect: they will encourage parties to write less complete contracts than they otherwise would prefer.38 The benefits of express terms would be discounted by the higher costs of enforcing those terms, as well as the risk that courts would incorrectly refuse to enforce those terms in favor of some implied term or contextualist interpretation. That argument depends on the assumption that the expected costs and risks that contextualism will undermine a writing that correctly expresses the parties’ intentions exceed the expected benefits of contextualism in avoiding litigation over, and enforcement of, a writing that incorrectly or incompletely expresses the parties’ intentions. Moreover, there is a parallel concern under textualism: parties will have an incentive to write more complete contracts than they would otherwise prefer. Greater complexity can in fact lead to more litigation, as the chance that terms will conflict or support alternative conduct increases.
6. p. 135Default versus Mandatory Rules
Even if contracts are generally incomplete and court interpretation and implication are appropriate, economists generally agree that the rules governing interpretation and implication, like other contract rules, should be default rules rather than mandatory rules. Default rules are rules that parties can contract around, whereas mandatory rules apply regardless of the parties’ intentions. Implied terms that serve as ‘gap fillers’, such as a reasonable price term, are usually viewed as paradigmatic examples of default rules in the sense that all the parties need to do to contract out of the implication is to specify the missing term, such as price. Other implied terms, such as the duty of good faith and the duty of loyalty in fiduciary contracts, are usually considered mandatory in the sense that parties cannot write contract terms broadly disclaiming these duties. The usual critique of mandatory terms is that because they disregard the intentions of the parties, they make worse off parties who prefer terms other than those mandated. For example, if a court imposes a stronger performance obligation on an obligor than the parties intended, then future obligors will extract a higher price, which is more than the obligee wanted to pay (else he would have paid for it originally).39 Economists sometimes defend mandatory terms if there are third party concerns, or asymmetric information between the contracting parties.40
The distinction between default and mandatory terms is not always so clear, however, once one allows for the possibility of efficiently incomplete contracts and unclear intent. For example, whether one views the duty of good faith or the duty of loyalty as mandatory depends on how well one thinks those doctrines track contractual intent. If parties intend to write incomplete contracts for which they expect courts to fill in the gaps, the duties of good faith and loyalty might easily be viewed as defaults. That view is further supported to the extent that if the parties want a particular obligation that conflicts with what courts ordinarily view as good faith or loyalty, and they specify that obligation, courts will generally enforce it.41 On the other hand, if one believes that courts use the duties of good faith and loyalty to fill in gaps that the parties did not want to be filled (for example, to preserve private discretion rather than court discretion), or to reject or ignore obligations the parties thought they had fully specified, then the duties look more like mandatory rules.42
p. 136Another example is the doctrine of indefiniteness, under which courts will sometimes decline to fill gaps the parties have left in contracts. The doctrine could be viewed as a default rule if one is willing to presume that when the parties have left ‘too many’ gaps for the courts to fill, they do not have contractual intent, and if they do have such intent they will override the default by filling in the terms themselves. Alternatively, the doctrine could be viewed as a mandatory rule if one assumes that the courts use it to refuse to fill in gaps when the parties wanted them to.
On the other hand, implied terms, intended to serve as gap-filling defaults, could instead become de facto mandatory rules if it is difficult for courts to determine whether parties, when they use express terms, intend to contract out of implied terms or merely to supplement them.43 That is, it may be difficult for courts to tell in a particular case whether the parties intended to incorporate implied terms by writing an incomplete contract, or whether they intended the express terms they used to create a complete contract. The more courts favor and encourage implied terms and common usages, the more costly it becomes for the parties who want to contract out of those terms to do so. Ostensible default rules begin to look more like mandatory rules. The courts’ choice of interpretive strategy, therefore, may affect not only the parties’ incentives to contract more expressly, but also their ability to contract around the implied default rule.
The default rule concept can be applied not only to implied terms but also to interpretive methodologies such as textualism and contextualism. Although scholars generally agree that these methodologies should be defaults, some argue that interpretive rules themselves are at least to some extent mandatory.44 Others take the position that interpretive rules are in effect defaults, because parties have many ways to contract for their preferred interpretive methodologies, including choice of law and choice of forum clauses, as well as merger and no-oral modification clauses.45 Schwartz and Scott argue further that it is easier for parties to contract out of a textualist regime (for example, by writing trade usages into their contracts) than it is to contract out of a contextualist regime.46 It is not clear, however, why it is easier to write a contract that says that the usages of the widget trade will apply than to write one that says that no trade usages will apply, or the usages of the widget trade will not apply, or no usages will apply except for the usages of the widget trade. It is true that if courts p. 137presume most contracts are written in ‘majority talk’, then by definition fewer parties would have to contract around than if courts did not make that presumption,47 but it is not clear why trade usage presumption would not be ‘majority talk’ for parties in a particular trade.
7. Superior Risk Bearer
Law and economics analyses of contract doctrine often make use of the superior risk bearer concept, which views contractual nonperformance as analogous to a tort accident, and assigns contract risks to the party better able to bear those risks on the assumption that the parties would have wanted this result. The application of the superior risk bearer concept to implied terms and interpretation depends in part on the methodological presumptions one adopts. A contextualist is more likely to presume that a contract is efficiently incomplete, and so more likely to view the parties as having reasonably left a gap with respect to a given contingency, a gap that a court should fill. The contextualist would then apply the superior risk bearer concept at the gap-filling stage by asking which party is better situated to bear the risk they failed to contract over.
A textualist is more likely to presume that if an incomplete contract occurs, the incompleteness is the accident, not the underlying contingency. Thus, the party who fails to protect himself against a contingency in the contract is the superior risk bearer, because that party has failed to take cost-effective ‘contract-based precautions’.48 On this view, courts should use the doctrines of interpretation and implied terms to encourage the parties to ‘facilitate improvements in contractual formulation’,49 which in this case means encouraging more complete contracts, that is, the greater use of express written terms or more precise language. If a court is willing to ‘insure’ parties through flexible interpretations and implied terms it creates a classic moral hazard problem: the parties have less incentive to write good contracts themselves.
The textualist approach would favor strict interpretation of doctrines such as the parol evidence rule and the indefiniteness doctrine to encourage parties to write more complete contracts. By giving more weight to the written document and limiting the extrinsic evidence courts can consider, the parol evidence rule encourages parties to put all aspects of their agreement into their written contracts.50 By requiring that contracting parties p. 138include key terms to make the contract enforceable, the indefiniteness doctrine encourages parties to include these terms.51 The question, however, is whether such contract-based precautions are invariably cost-effective.
An alternative approach to the superior risk bearer question is to ask which party is the cheaper contract drafter, namely the party in a better position to clarify a term or to identify what should happen in the event of some contingency. This approach explains such interpretive rules as contra proferentem, under which ambiguities are construed against the drafter. This rule encourages the party in the better position to draft a more complete contract to do so. Similarly, if one of the parties is a repeat contractor or is assisted by legal counsel and the other is not (as in many consumer contracts), imposing liability on the repeat and represented contractor in cases of contractual ambiguity or incompleteness will encourage that party to improve the terms of its contracts. In addition, if one of the parties has an informational advantage (for example, a party with idiosyncratic preferences), imposing liability on that party could encourage similarly situated parties in the future to reveal the information. Because the party with the informational advantage is not always the contract drafter, there may be a tension between this criterion and the contra proferentem doctrine.52 In some cases, the doctrine of mistake may excuse a drafter who makes a drafting error, especially if the other party should have noticed it. Of course, there may not always be a ‘cheaper contract drafter’, or if there is, the necessary precautions might not be cost-justified.
A variant of the superior risk bearer argument that does not involve encouraging more complete contracting is giving the party who has the ability to cheaply discover a particular meaning the incentive to do so. A contextualist rule that can be justified under this argument is the rule that contracting parties ‘in the trade’ are bound by trade usages, even if they did not know about them. This rule not only encourages the parties in a trade to develop such usages but also to familiarize themselves rapidly with these usages, hence reducing the need for heavily lawyered documents.53 By the same token, the textualist ‘plain meaning rule’ could be viewed as a way of encouraging contracting parties to learn the common (one might even say implied) meaning of words, thus reducing the need for and costs of elaborate definition and explanation. In addition, the Restatement has a rule under which if the first party has reason to know the meaning attached by the second party and the second party has no p. 139reason to know of the first party’s meaning, the second party’s meaning prevails. The first party’s negligence makes it the superior risk bearer of the misunderstanding.
8. Opportunistic Behavior
Putting the risk on the superior risk bearer, whether the risk is the underlying contingency causing contractual disruption or the risk of an imperfectly drafted contract, is one fault-based approach to questions of interpretation and implied terms. A second fault-based approach focuses not on encouraging efficient contracting, but on deterring opportunistic contractual behavior (though obviously the two overlap). Opportunism can be broadly defined as deliberate contractual conduct by one party contrary to the other party’s reasonable expectations based on the parties’ agreement, contractual norms, or conventional morality.54 Alternatively, opportunism is an attempted redistribution of an already allocated contractual pie, that is, a mere wealth transfer.55
Opportunistic behavior is costly. It ‘increases transaction costs because potential opportunists and victims expend resources perpetrating and protecting against opportunism’. Muris (1981, p. 524). Moreover, opportunistic behavior makes complete contracting extremely difficult. Even if contracting parties could anticipate all of the possible changes in economic variables, they would have a much harder time anticipating and protecting against opportunistic behavior by the other party. At the extreme, the greater the concern one contracting party has with the possible opportunistic behavior of his contracting partner, the less likely he will be to want to contract with that partner at all. Because contracting parties cannot solve all problems of opportunism on their own, courts can potentially reduce transaction costs by imposing liability on the ‘most likely opportunist’.
But there are difficulties with using courts to deter opportunism. In particular, opportunism is often ‘subtle’, that is, difficult to detect or easily masked as legitimate conduct.56 Whether a contracting party is legitimately relying on an express term of the contract, for example, or opportunistically exploiting it to justify conduct the parties did not originally expect depends on the parties’ intentions, which, as we have seen, are often uncertain and contested. Lipshaw (2005) contends that the concept of opportunism will rarely be helpful in litigated cases because it requires the identification of an ex ante intent that may not exist.
p. 140Other scholars are more optimistic that courts can identify opportunistic behavior and in fact can use the concept to help determine uncertain intent. Thus, Judge Posner argues that courts should hesitate to interpret a contract in such a way as to permit conduct that would ordinarily be understood as opportunistic.57 Similarly, Muris argues that courts should hesitate to attribute to contracting parties an intention not to have courts police against opportunistic behavior.58 Moreover, courts may be able to identify ‘objectively verifiable circumstances that act as surrogates for the existence of opportunism’ Muris (1981, p. 530). For example, the risk of opportunism is greater whenever one party has an asymmetric information advantage over the other. Thus, in the fiduciary context, courts adopt, via the duty of loyalty, a strong presumption of wrongful misappropriation by an agent when that agent has a conflict of interest, engages in self-dealing, or withholds information from the principal.59
In addition, courts may be able to identify risks that the parties allocated and risks that they did not contemplate. If the court finds that the parties intended that the contract assign a particular risk to one of the parties, such as a change in market price, and that risk materializes, the court should be skeptical of attempts by the disadvantaged party to escape his obligations via a different contract term. For example, a buyer in a requirements contract may suddenly experience a large drop in ‘requirements’ after the market price has fallen below the contract price, or a large increase in requirements after the market price has risen above the contract price. In these cases, the court should suspect opportunism, or in legal terms, a violation of the implied obligation of good faith. On the other hand, if the court finds that a change in economic circumstances was not contemplated by the contract, the court may infer a lack of opportunism. For example, if the buyer’s requirements decrease or increase because of a change in costs or technology subsequent to the contract, the buyer’s behavior is likely not opportunistic (is in good faith) because the very purpose of the requirements contract is to assign some risk of variation in the buyer’s needs to the seller. Goldberg (2006, chapter 5), however, criticizes this kind of analysis because he believes the purpose of requirements contracts is to provide discretion to one party, and implicitly assumes that the parties’ preference for preserving private discretion takes priority over their interest in deterring opportunistic abuses of discretion.
Another example of objectively verifiable circumstances on which courts p. 141can focus is ex post transaction costs. For example, if parties are likely to undertake ‘interpretation-specific investments’ or investments that ‘are especially vulnerable to changes in contractual interpretation’, they may favor a more expansive approach to interpretation to reduce the risk of opportunism.60 On the other hand, if the market for substitute performance is thick, opportunism is less likely and contextualism less necessary.61 Although this distinction may be useful in establishing general presumptions, it is of limited help in deciding specific cases. Litigated cases tend to be precisely those in which ex post transaction costs are likely to be high; otherwise, the cases would be settled.
Although opportunism is often discussed as an ex post problem, opportunism can occur ex ante as well. For example, the drafter of a standard-form contract might try to sneak in one-sided but inefficient terms into the fine print.62 By the same token, under a strict parol evidence rule, a party might intentionally make oral statements that the other party relies on as part of the contract, and then leave the provision out of (or put or leave a contradictory provision in) the writing. On the other hand, under a more flexible parol evidence rule, a party might intentionally ‘pad’ the negotiation record with statements that the party knows will be rejected by the other party both orally and in writing, in the hopes that the first party can later convince the court that these statements were in fact part of the contract (a common practice in legislative history).63 Katz (2004, p. 531) questions how often this problem will really occur because in most cases parties will have equal access to negotiating history and can observe and counter blatant, self-serving attempts to manipulate parol evidence.
Still, opportunism, whether of the ex ante or ex post variety, may help explain why courts tend to be much more skeptical of evidence that the parties can easily manipulate (especially prior negotiations) than evidence over which the parties have less control (such as trade usages).64 Concern with opportunism may also explain the motivation behind the use of merger clauses as well as one reason why they should not be interpreted too broadly. Parties may find it difficult to predict in advance which negotiation tidbit the other side might seize on later,65 so they may find it necessary to write a broad clause that excludes them and not p. 142worthwhile to expend resources on identifying exceptions that may be jointly maximizing.
Scholars disagree about whether the problem of opportunistic behavior supports textualism or contextualism. Schwartz and Scott (2003, pp. 585–6) argue that contextualism creates a greater likelihood of opportunism because parties can always falsely allege some contextual evidence, such as a ‘private language’, that courts may incorrectly find to be true. Kostritsky (2007), by contrast, argues that either textualism or contextualism can lead to opportunistic behavior, so courts should focus on deterring opportunism rather than on definitively choosing one or the other interpretive methodology. Because contextualism and textualism are both useful for deterring different types of opportunism, we should therefore expect – and we find – that courts are never completely committed to one or the other. Even the most contextualist courts may reject or restrict evidence they deem to be too self-serving or backed up merely by the say-so of one of the litigants rather than objectively demonstrable evidence. And even the most textualist courts may blink when a narrowly literalist reading of a contract flies in the face of the parties’ evident intentions.
9. Joint Fault and Multiple Contingencies
A further difficulty in applying fault concepts such as superior risk bearer and most likely opportunist to questions of interpretation and implied terms is that in many contractual disputes, both parties can be viewed as at fault to some extent. In these cases, courts may have to make judgments about the relative fault of both parties to decide whose behavior it is more important to deter in a particular case. In particular, if one party is the least cost avoider of some contingency while the other party regrets the contract for other reasons and is opportunistically seeking to avoid its obligations, courts face a ‘negligence-opportunism tradeoff’.66 To take a classic example, suppose a builder promises to use a particular brand of pipe in building a house but inadvertently substitutes a different, but functionally equivalent brand, a fact not discovered by the owner until the house is nearly completed. The owner refuses to make the final payment on the house. The court must choose between placing liability on the negligent builder or the potentially opportunistic owner. There is an economic case to be made that opportunism – if sufficiently proved – is more costly behavior and deterrence of that behavior should take priority.67 On the other hand, the more likely it is that the builder ‘built first and asked questions p. 143later’ (Goldberg 1985, p. 71), the more willing courts should be to find for the owner by implying a condition.
The negligence-opportunism tradeoff can also be created by the concurrent occurrence of two contingencies. Suppose that a buyer rejects goods delivered late after a market price drop and the seller sues. There are two contingencies here: the price drop and the late delivery. The contract assigns the risk of the price drop to the buyer and the late delivery to the seller. Textualism will not resolve this dispute: either the price term or the time of delivery term cannot be read absolutely. It is not sufficient to say that only the seller has breached, because what constitutes a breach and the consequences of that breach are precisely what is at issue. Nor can it be said that only the seller could take precautions here because neither party could do anything about the price drop and the buyer did not cause the delay in delivery. If the buyer’s rejection is viewed as opportunistic behavior, then refraining from such behavior could be viewed as a ‘precaution’. Depending on the circumstances, there is an economic argument to be made for implying a good faith ‘limitation’ on the buyer’s ability to escape its obligations.
10. Characteristics of Contracting Parties
Party characteristics unrelated to fault may also affect the optimal approach to interpretation and implied terms. One such characteristic is the parties’ attitudes toward risk. Katz argues that relatively risk-averse parties will prefer contextualism because that reduces the variance of results among judges by equalizing the ‘information sets’ of more and less experienced judges (2004, pp. 527–8). By contrast, Schwartz and Scott argue that risk-neutral contracting parties, such as many businesses, would support textualism. They discuss a model that assumes that parties’ respective payoffs increase under favorable interpretations and decrease under unfavorable interpretations, court interpretations are unbiased, and there is some positive probability that textualism will yield the answer the parties intended (Schwartz and Scott 2003, pp. 573–7). They argue that the parties will invest resources in drafting until each term represents the mean possible judicial interpretation. Contextual information would therefore produce no benefits because it could only reduce variance, and risk-neutral parties care only that interpretations are right on average and do not care about reducing variance.
Bowers (2005) criticizes the Schwartz and Scott risk neutrality argument on a number of grounds. First, large interpretive variance could lead to more ‘chiseling’ behavior by the parties, so that even risk-neutral parties would care about reducing interpretive variance to discourage such behavior (Bowers 2005, p. 601). Further, the parties’ asymmetric abilities p. 144to ‘manipulate the context’ and mislead the court may call into question the assumption that the mean of the distribution of error is zero (2005, p. 602). More generally, if allowing contextual evidence would enable courts to reach the mean interpretive result at lower total cost, including the costs and uncertainty associated with opportunism, contextualism would be superior. In addition, the very assumption of risk neutrality is to some degree in tension with contracting itself, which is at least in part about reducing risk (2005, pp. 596–7). If parties are willing to expend resources to reduce risk by contracting, it is hard to rule out a priori the possibility that parties might prefer court adjustments ex post that further reduce risk in a cost-effective way. Contracting parties do not demand enforceable coin flips to resolve their disputes.
Another potentially relevant characteristic is the degree of homogeneity among contracting parties and transactions. Some scholars argue that the more likely contracting parties are heterogeneous, the more inefficient an expansive approach to implied terms and interpretation will likely be, because contracting parties will more likely be unhappy with the court’s implied terms and interpretations and want to contract out of them. By contrast, the more likely contracting parties are homogeneous and engage in repetitive transactions with low transactional variance, the more likely a contextual approach will be efficient because it will foster the development of more standardized terms by trade groups, lawyers, and the parties themselves.68 Other scholars, however, argue that heterogeneous parties, such as one-shot contractors and parties with asymmetric information, have higher renegotiation costs, and so are more likely to favor broader court intervention through interpretation and implied terms, whereas more homogeneous parties in commercial subcommunities are able to rely more on renegotiation as well as extralegal sanctions and so prefer more formalistic approaches by courts.69
The relative bargaining power of the parties may also be important, though scholars again draw different inferences from this characteristic. Schwartz and Scott (2003, p. 580) argue that if the seller has bargaining power, it will not lose as much from adverse interpretation and so would not be willing to pay for a contextual interpretation even if that party would benefit from that methodology. Ben-Shahar (2009), by contrast, argues that courts should set default rules in a way that benefits parties with bargaining power, because they would have been likely to obtain those benefits had the parties bargained over them explicitly.
11. p. 145Court Competence and Error
An alternative approach to interpreting and implying contractual terms is to focus not on the contracting parties, but on the court or other decisionmaker, such as a jury or arbitrator. The degree of court competence, error, and independence can help determine the optimal interpretive strategy. At one extreme, if courts are highly competent and honest, and can accurately interpret contextualist evidence at relatively low cost, then such evidence should always be allowed to interpret or imply terms, at least if the cost of producing such evidence and related litigation costs are not too high. At the other extreme, if courts are incompetent and corrupt, or make too many mistakes in interpreting or implying terms (or can reduce those mistakes only at high cost), then textualism becomes superior if the transaction costs of contracting are lower than the expected savings resulting from fewer court errors.70
Scholars disagree, however, over whether strict approaches to interpretation and implied terms, such as textualism, lead to more court error than broader approaches, such as contextualism. Hadfield (1994) develops a formal model of good faith clauses in intentionally incomplete contracts with probabilistic court error, from which she deduces that courts of low competence should not follow bright line rules, but instead should use more flexible standards. Bright line rules correspond to textualism, whereas standards correspond to contextualism. Bright line rules may compound rather than ameliorate court error by a court of limited competence, because a bright line rule setting forth a required action will so often be ‘wrong’. Thus, parties may respond to a bright line rule by engaging in inefficient behavior encouraged by the rule. Standards, by contrast, are more likely to encompass the ‘efficient’ response, and so the parties will be more willing to engage in optimal behavior.
Other scholars, however, conclude that contextualism is more likely to lead courts astray. Schwartz and Scott (2003, p. 587) posit that contextualism creates more opportunity for court error, because there are many possible ‘private languages’ that parties could falsely assert, and contextualism does not lead courts to reach correct interpretations on average. Goldberg studies a number of cases in depth and concludes that ‘courts seem rather oblivious to the economic context when interpreting contracts’, which might be facilitated or exacerbated by the consideration of contextual evidence (2006, p. 163). These conclusions result from optimism about the likelihood of a textualist approach to yield the result the parties intended, as well as skepticism about the ability of courts to p. 146understand the economic context and police manipulated contextual evidence. One might suppose, however, that if the economic context points clearly toward a particular interpretation, it would not be so hard for good lawyers to enlighten judges about that context. Moreover, manipulation of contextual evidence can be reduced by such means as insisting on objectively verifiable evidence such as trade usage,71 imposing a higher burden of proof, restricting the role of the jury,72 or seeking indicia of opportunism.
Another argument for textualism based on court error draws inferences from an alternative governance mechanism: arbitration. Arbitrators are generally thought to have greater expertise in particular business contexts than courts. This greater expertise enables arbitrators to evaluate contextual evidence such as trade usages at lower cost and with fewer errors than courts. Thus, because parties can easily contract for arbitration, one can infer that their failure to do so suggests a majoritarian preference for textualism.73 Parties may choose arbitration for reasons other than expertise and interpretive methodology, however, including lower costs, greater privacy, and the tendency of arbitrators to make middle-of-the-road rulings.74 The last feature of arbitration may arise out of a concern with arbitrator error; middle-of-the-road damage awards may substitute for judicial review.75 Bernstein (1996, 1999, 2001) takes a somewhat different tack, arguing that contracting parties who deal in specialized markets often prefer private dispute rules and procedures because those rules and procedures are more formalistic, which she argues supports the conclusion that contracting parties do not favor contextualist interpretation. But the formalist approaches of specialized decisionmakers may simply reflect the fact that they already have internalized much of the relevant context, so the marginal benefit of additional contextual information is relatively low. Thus, the preference of parties for formalistic decisionmaking in this setting does not imply a similar preference in more generalist courts, where the variance of expertise, and hence the marginal benefit of contextual evidence, is greater.76
Court error may affect not only the parties’ substantive intentions under the contract, but also their preferred interpretive methodology, namely the choice between textualism and contextualism itself. The p. 147contracting parties may prefer textualism and express that preference through merger clauses, though in fact these clauses do not really endorse textualism but rather exclude certain types of contextual evidence, most typically prior negotiations.77 If courts err in determining the parties’ methodological preference (by declining to enforce merger clauses strictly, for example), they may choose contextualism too often.78 But once again, this conclusion depends on the assumption that if courts use textualism (this time to decide the parties’ methodological preference), they will err less often because the costs to the parties of accurately expressing their methodological intentions are low. One would think, however, that the costs to the parties of drafting a particularized methodological term are quite high. Not only is methodological preference merely a second-order concern for the parties, it is difficult for the parties to predict how court error will likely impact the wide variety of possible disputes they might have, and methodological preference terms (unlike substantive contract terms) have no contractual use to the parties outside of litigation. As a result, it is not obvious a priori that choosing a textualist approach to determine the parties’ methodological preference minimizes court error.
Suppose, for example, that the parties generally prefer a textualist approach and expect interpretation x of some term. There are actually three ways the court could err. The court could take a contextual approach but reach interpretation x. Or the court could take a textual approach and reach interpretation y. Or the court could take a contextual approach and reach interpretation y. Although the parties might prefer the textualist approach, it might be more important to them that the court gets the term right, however the court does it. If courts are more likely to choose the desired interpretation x using a contextual approach and interpretation y using a textual approach – either because the parties underestimate the courts’ competence with respect to that term or because their expressed preference for textualism inaccurately conveys the parties’ correct estimate of the courts’ competence in this instance – then error costs could be reduced if the court ‘mistakenly’ used a contextual approach.
To give a simple numerical example, suppose the court can choose either a textualist or contextualist methodology. If it chooses textualism, the likelihood of interpreting the term the way p. 148the parties want is 0.4; if it chooses contextualism, the likelihood of interpreting the term the way the parties want is 0.6. Suppose further that ex ante the parties would value the court’s using textualism and choosing x at 100; would value the court’s using contextualism and choosing x at 80; would value the court’s using textualism and choosing y at 50; and would value the court’s using contextualism and choosing y at 10. Thus, the parties prefer textualism to contextualism, but prefer the right outcome more. The expected value if the court uses a textualist approach is (0.4 x 100) + (0.6 x 50) = 40 + 30 = 70. The expected value if the court uses a contextualist approach is (0.6 x 80) + (0.4 x 10) = 48 + 40 = 88 > 70.
The point is that the possibility of court error does not always argue in favor of textualism. Both textualist and contextualist methodologies lead to court error. The real question is which methodology has the lowest error rate and at what cost. It is hard to answer this question in the abstract. This may help to explain why courts do not use pure interpretive methodologies, but tend to switch back and forth depending on the circumstances – choices that themselves are subject to error.
12. Agency Costs and Third Party Interests
Many analyses of interpretation and implied terms assume a simple contracting situation in which the same two monolithic parties negotiate, draft, perform, and litigate the contract and no third parties have any rights or interests in the contract. Real world contracting often diverges from these assumptions. For one thing, many contracting parties are entities, which act only through agents. Different agents may be involved at different stages of the contract. Entity parties may use various contract terms to protect against undesired conduct by agents. For example, a sales or purchasing agent may make extravagant promises during negotiations that the entity does not want to be bound by. Entities may use merger clauses, and may prefer textualism, to solve this problem. Thus, one can argue that courts should be sensitive to the likelihood of agency cost problems in deciding on an interpretive methodology.79
As with many of the other factors, however, agency cost problems do not invariably favor textualism. For example, although sales and purchasing agents may act contrary to the interests of the entity, in-house counsel who review the contracts and draft the final terms are also agents who may act contrary to the interests of the entity by drafting unnecessarily complex and overly formalistic documents that serve their own self-interest in avoiding liability. If lawyer agency problems are greater than the agency problems associated with negotiating agents, contextualism may be the p. 149optimal interpretive strategy.80 Another problem with the agency cost argument for textualism is that it considers only one side of the contract and ignores the interests and perhaps reasonable expectations of the other side, which may legitimately fear being lured into the contract on false pretenses and so may favor a contextualist approach. Moreover, the agency cost problem may implicate some kinds of contextual evidence, such as prior negotiations, more than others, such as trade usage.
A related problem is that third parties other than the original parties who agreed to the contract often have an interest in the contract. For example, the contract may be assigned, one of the original parties may merge or go bankrupt or die, a non-party may guarantee performance, or a lender or investor may base decisions on contractual obligations.81 These third parties may not know all the contextual evidence the drafting parties know and may not be able to discover such evidence at low cost. It may therefore be in the interest of the original contracting parties to use clauses such as merger clauses and to prefer textualism more generally to make third parties more willing to deal with the original contracting parties on favorable terms.82 Again, however, this conclusion does not invariably hold. For example, the third party may be privy to the original negotiations, may be aware of relevant trade usages, or may be able to find out the relevant contextual information from the original parties at low cost. Moreover, the third party’s interests might not be implicated in a given dispute or the third party might not have relied on the specific text of the contract.
To a large degree, the economic approach to interpretation and applied terms parallels the approach in other areas of contract law. Court intervention, which in this case means a greater willingness to imply terms and use a contextualist methodology for discerning contractual intent, is most justifiable when contracts are efficiently incomplete and extralegal enforcement is relatively ineffective, when courts can accurately assess contextual evidence and police against opportunistic behavior at relatively low cost, and when parties can easily contract for stricter rules. Many of the economic factors, however, can cut in either direction, depending on the circumstances. Therefore the institutional and contractual context matters greatly in deciding what approach efficiency-minded courts should take.
Shavell (2006, p. 295).
Shavell (2006, p. 295).
Hadfield (1994, p. 163).
Richard A. Posner (2005, p. 1590).
Hadfield (1994, p. 160).
Hermalin and Katz (1993, pp. 235, 242).
Eggleston et al. (2000, pp. 122–5).
Eggleston et al. (2000, p. 109).
Eggleston et al. (2000, pp. 110–12, 119–22).
Goetz and Scott (1985, p. 267, n. 11).
Hill (2009, p. 204).
Hill (2009, p. 208).
Hill (2009, pp. 209–10).
For example, Richard A. Posner (2005, pp. 1585, 1586, 1590).
Schwartz and Scott (2003, p. 569).
Hadfield (1992, p. 259).
Eggleston et al. (2000, pp. 114–15).
Goetz and Scott (1981).
Richard A. Posner (2005, pp. 1603–7).
Burton (1980, p. 373).
Easterbrook and Fischel (1993, p. 445).
Katz (2004, pp. 530–31).
For example, Goldberg (2006, p. 163).
Richard A. Posner (2005, pp. 1602–3).
Schwartz and Scott (2003, pp. 587–9).
For example, Easterbrook and Fischel (1993, p. 431).
For example, Shavell (2006, pp. 310–11).
Easterbrook and Fischel (1993).
Goetz and Scott (1985).
Katz (2004, pp. 506–12).
Schwartz and Scott (2003, pp. 584–5).
Cohen (1992, p. 949).
Goetz and Scott (1985, p. 264).
Eric Posner (1998).
Richard A. Posner (2005, p. 1588).
Richard A. Posner (2005, p. 1608).
Cohen (1992, p. 957).
Muris (1981, p. 525).
Richard A. Posner (2005, p. 1604).
Cf. Muris (1981, p. 573, n. 138).
Cooter and Freedman (1991, p. 1054).
For example, Katz (2004, p. 530).
Goetz and Scott (1983).
Katz (2004, p. 531).
Eric Posner (1998, pp. 564–5).
Katz (2004, p. 532).
Eric Posner (1998, p. 572).
Cohen (1992, pp. 983–90).
Goetz and Scott (1985).
Katz (2004, pp. 528–9).
Bowers (2005, p. 610).
Richard A. Posner (2005, pp. 1593–6).
Richard A. Posner (2005, p. 1594).
Richard A. Posner (2005, p. 1609).
Katz (2004, pp. 526–7).
Bowers (2005, pp. 603–4, n. 63).
Katz (2004, p. 533).
Katz (2004, p. 534).
Burton (2009, p. 200).
Katz (2004, pp. 534–5).
Cooter, Robert and Freedman, Bradley J. (1991), ‘The Fiduciary Relationship: Its Economic Character and Legal Consequences’, New York University Law Review, 66(4), 1045–75, reprinted in Roberto Pardolesi and Roger Van den Bergh (eds) (1991), Law and Economics, Some Further Insights: 7th European Law and Economics Association Conference, Rome, Sept. 3–5, 1990, Milan: Sutte, 17–50.
Goldberg, Victor P. (1985), ‘Relational Exchange, Contract Law, and the Boomer Problem’, Journal of Institutional & Theoretical Economics, 141(4), 570–75, reprinted in Victor P. Goldberg (ed.) (1989), Readings in the Economics of Contract Law, Cambridge: Cambridge University Press, 67–71, 126–7.