Economic analysis of law has traditionally assumed that legal rules are or ought to be designed to maximize social welfare taking as given that legal subjects are like Holmes’s “bad man”—rational, self-interested agents who care about complying with the law only insofar as noncompliance exposes them to the risk of sanctions. But while it is plausible to suppose that some legal subjects are, like Holmes’s bad man, “externalizers” of legal rules, it is likely that others regard the law from the “internal point of view”—that is, they are “internalizers” of legal rules who are motivated to conform to legal rules regardless of the consequences of defiance. In this Article, I ask what specific doctrinal implications for contract law follow when we assume that there are both internalizers and externalizers in the subject population while keeping the standard economic framework otherwise intact. I analyze several contract law phenomena: the expectation principle, according to which a victim of a breach of contract is entitled to her expectation, and no more than her expectation; the penalty doctrine and stipulated damages clauses; the doctrine of willful breach; and the doctrine of contract modification. The resulting analysis shows that internalizers respond to these doctrines differently from externalizers such that the prescriptive implications of economic analysis for the design of contract law doctrines may change once we recognize that there are both internalizers and externalizers in the subject population. For example, the presence of internalizers in the subject population provides a justification for punishing some “willful breaches” more harshly than ordinary breaches. It also helps to justify the modern doctrine of contract modification.
- DO PEOPLE INTERNALIZE LEGAL RULES?
- The Disposition to Follow Norms
- From Social Norms to Legal Rules
- A SIMPLE MODEL OF THE INTERNAL POINT OF VIEW
- Externalizers and Internalizers
- Why Internalize Legal Rules?
- The Perform-or-Breach Decision
- Rules, Standards, and Crowding Out
- Stipulated Damages Clauses
- Fault-Based or Strict Liability and Willful Breach
- Duress and Contract Modification
APPENDIX: CONTRACT MODIFICATION UNDER THE CREDIBILITY RULE
At any given moment the life of any society which lives by rules, legal or not, is likely to consist in a tension between those who, on the one hand, accept and voluntarily co-operate in maintaining the rules, and so see their own and other persons’ behavior in terms of the rules, and those who, on the other hand, reject the rules and attend to them only from the external point of view as a sign of possible punishment. One of the difficulties facing any legal theory anxious to do justice to the complexity of the facts is to remember the presence of both points of view and not to define one of them out of existence. 1 H.L.A. Hart, The Concept of Law 90–91 (2d ed. 1994) [hereinafter Hart, Concept of Law].
Much of contemporary contract theory asks what substantive principles or values best explain and justify contract law. Should contract law be designed to maximize social welfare, as economic approaches to contract law suppose? 2 For examples of such approaches, see Richard A. Posner, Economic Analysis of Law 95–155 (9th ed. 2014); Charles J. Goetz & Robert E. Scott, Enforcing Promises: An Examination of the Basis of Contract, 89 Yale L.J. 1261 (1980); Lewis A. Kornhauser, An Introduction to the Economic Analysis of Contract Remedies, 57 U. Colo. L. Rev. 683 (1986). If not, what other values ought it strive to promote? Is it better understood as a vehicle for vindicating party autonomy, 3 For an example of such a theory, see Jody S. Kraus, The Correspondence of Contract and Promise, 109 Colum. L. Rev. 1603 (2009). for instance, or as a set of rules that directly enforce our underlying moral obligations to keep our promises? 4 For an example of such a theory, see Charles Fried, Contract as Promise: A Theory of Contractual Obligation (2d ed. 2015). There are of course other possibilities. For a helpful discussion, see Liam Murphy, The Practice of Promise and Contract, in Philosophical Foundations of Contract Law 151, 153–63 (Gregory Klass, George Letsas & Prince Saprai eds., 2014).
But this focus on ultimate values obscures a second significant question—a question about the conception of the legal subject that lies at the heart of contract law. Is contract law just a system of incentives for rational, self-interested actors, as economists would have it? 5 For an example of such a theory, see Richard Craswell, Two Economic Theories of Enforcing Promises, in The Theory of Contract Law 19 (Peter Benson ed., 2001). Note, however, that a significant body of research in behavioral economics challenges the assumptions that agents are self-interested and adept at discerning and acting on their self-interest. For challenges to the former assumption, see Herbert Gintis, Samuel Bowles, Robert Boyd & Ernst Fehr, Moral Sentiments and Material Interests: Origins, Evidence, and Consequences [hereinafter Gintis et al., Moral Sentiments], in Moral Sentiments and Material Interests 3, 8–22 (Herbert Gintis, Samuel Bowles, Robert Boyd & Ernst Fehr eds., 2005); Christine Jolls, Cass R. Sunstein & Richard Thaler, A Behavioral Approach to Law and Economics, 50 Stan. L. Rev. 1471, 1489–97 (1998); Russell B. Korobkin & Thomas S. Ulen, Law and Behavioral Science: Removing the Rationality Assumption from Law and Economics, 88 Calif. L. Rev. 1051, 1126–43 (2000); Colin Camerer & Richard H. Thaler, Anomalies: Ultimatums, Dictators and Manners, J. Econ. Persp., Spring 1995, at 209. For challenges to the latter assumption, see Jolls et al., supra, at 1477–79, 1548–50; Korobkin & Ulen, supra, at 1075–126; Amos Tversky & Daniel Kahneman, Judgment Under Uncertainty: Heuristics and Biases, in Judgment Under Uncertainty: Heuristics and Biases 3 (Daniel Kahneman, Paul Slovic & Amos Tversky eds., 1982). Or is it designed to tell people how they ought to behave, the view implicitly held by those who think that the law’s purpose is to give legal expression to an underlying promissory morality? 6 See, e.g., Fried, supra note 4, at 17. This Article considers the second, more neglected question from the perspective of economic analysis of law. It asks what happens to the economic understanding of contract law when we relax the assumption that all legal subjects view the law from the external point of view, conforming to legal rules only insofar as they are fearful of legal sanctions, and replace it with the more realistic assumption that many legal subjects are motivated to conform to the law because it is the law.
The Holmesian “bad man” view of the law that economists favor follows from a positive conception of legal subjects as agents who know what is best for themselves and single-mindedly pursue their own self-interest. 7 See Oliver Wendell Holmes, The Path of the Law, 10 Harv. L. Rev. 457, 461 (1897) [hereinafter Holmes, Path of Law] (“But what does [the notion of legal duty] mean to a bad man? Mainly, and in the first place, a prophecy that if he does certain things he will be subjected to disagreeable consequences by way of imprisonment or compulsory payment of money.”). In the language of economics, the preferences that subjects rationally maximize—by which I mean their judgments about how actions should be ranked for the purpose of choosing among them—are determined solely by their beliefs about their self-interest and so are exogenous to the legal regime. 8 Notice that there is no necessary connection between a subject’s preferences, as I define them, and her self-interest. In this sense, I depart from the common, though not universal, assumption made by economists that preferences provide a measure of individual welfare. See Amartya Sen, Choice, Welfare and Measurement 66–67 (1982) (explaining how many central results in economics are based on the assumption that preferences reflect individual welfare). There is nothing incoherent about departing from this assumption, as Professor Amartya Sen’s discussion of “commitment” shows. See Amartya Sen, Rational Fools: A Critique of the Behavioral Foundations of Economic Theory, 6 Phil. & Pub. Aff. 317, 322–24, 326–29 (1977) (explaining that commitment may drive a person to choose “an act that he believes will yield him a lower level of personal welfare . . . than an alternative that is also available to him,” thus creating a “wedge between personal choice and personal welfare”). Legal rules matter to subjects only insofar as they affect their well-being, and the primary way in which the rules do so is by changing the relative prices of available options through the imposition of rewards and sanctions. 9 See Oliver Wendell Holmes, Jr., The Common Law 301 (30th prtg., Little, Brown & Co. 1938) (1881) (“The only universal consequence of a legally binding promise is, that the law makes the promisor pay damages if the promised event does not come to pass.”). In short, subjects view the law as nothing more than an incentive scheme. They maximize their self-interest given their legally altered option sets.
This reductionism might obscure important features of the legal landscape. Legal rules might be directly action guiding, altering subjects’ reasons for action even when their options are held constant. In other words, the rules might speak to subjects who adopt an “internal point of view” toward the law—subjects who are willing to bracket their immediate self-interest (and other goals) in order to conform to the law’s prescriptions. 10 In Professor H.L.A. Hart’s words, they adopt:a critical reflective attitude to certain patterns of behaviour as a common standard, . . . [which] display[s] itself in criticism (including self-criticism), demands for conformity, and in acknowledgments that such criticism and demands are justified, all of which find their characteristic expression in the normative terminology of “ought”, “must”, and “should”, and “right” and “wrong”.Hart, Concept of Law, supra note 1, at 57.
Thus, the following question arises: Should we view legal subjects as rational agents who exhibit no preference to conform to the law for its own sake, agents who I will refer to as externalizers, 11 See id. at 88–91 (contrasting the external and internal points of view). or should we view them as internalizers, agents who adopt legal rules as reasons for action even when their self-interest (and other things they care about) dictates doing otherwise? 12 See id. at 57, 102 (discussing the internal point of view). Economists, as we have seen, have traditionally adopted the former conception alongside a view that the law’s purpose should be social welfare maximization. 13 See supra notes 2, 5 and accompanying text (discussing the assumptions that underlie traditional economic analysis). But a commitment to social welfare maximization does not entail a commitment to a conception of legal subjects as self-interested externalizers. Indeed, if economists are operating with an unrealistic model of legal subjects, their attempts to design rules that maximize social welfare are likely to fail.
This Article asks how economic analysis of contract law changes when we relax the assumption that all legal subjects are externalizers and instead assume that the population consists of both internalizers and externalizers while preserving the assumption that the objective of the law is social welfare maximization. 14 I adopt such a utilitarian standpoint not because I believe that utilitarianism is true, but because I want to consider how relaxing the assumption that agents respond to the law in a narrowly self-interested fashion alters the prescriptions of economic analysis. The Article proceeds in three parts. Part I explores theoretical considerations and empirical evidence that support the suggestion that in a well-functioning society a significant portion of the population obeys the law for reasons other than fear of legal sanctions even in market contexts where we might expect self-interested motivations to be especially prevalent. 15 Hart, Concept of Law, supra note 1, at 116–17. Part II considers the reasons why agents might become internalizers of legal rules and develops a framework for understanding the preferences of internalizers. Part III then develops some contract law applications. These applications illustrate the ways in which the behavioral responses of internalizers to legal rules can differ sharply from the responses of externalizers such that once we relax the assumption that everyone is a self-interested externalizer, the optimal shape of contract law doctrines may change.
Section III.A asks how an internalizer thinks about the decision whether to perform or breach a contract in the light of the legal doctrines that govern that decision. An externalizer focuses only on the likelihood that breach will result in a legal sanction plus, if she is morally motivated, any moral reasons that point in favor of or against performance. An internalizer, by contrast, will in the first instance try to conform to the prevailing legal norms as she perceives them regardless of self-interested and other moral considerations. And so the internalizer’s behavior will depend on how she resolves a prior interpretative question: What exactly is the legal norm that governs the perform–breach decision? In particular, does the law impose on her a duty to perform her promises unless she obtains a release from the promisee, consistent with the maxim pacta sunt servanda (“agreements must be kept”)? 16 See Daniel Friedmann, The Efficient Breach Fallacy, 18 J. Legal Stud. 1 (1989) [hereinafter Friedmann, Efficient Breach Fallacy] (defending such an interpretation). Or does it give her the option to choose between performance and paying the promisee his expectation? 17 See Holmes, Path of Law, supra note 7, at 462 (proposing this perform-or-pay interpretation of a promisor’s legal duties); Richard A. Posner, Let Us Never Blame a Contract Breaker, 107 Mich. L. Rev. 1349, 1349 (2009) [hereinafter Posner, Contract Breaker] (endorsing the Holmesian reading).
Section III.B discusses the possibility that the legal norm, if it takes the perform-or-pay form as economists typically assume that it does, might have the effect of crowding out an internalizer’s moral motivation to keep her promises. That is, the norm might encourage a subject who would otherwise be motivated to perform for moral reasons to conclude that it is perfectly fine if she breaches so long as she pays the price of doing so by paying expectation damages to the promisee. When the norm has such an effect, the law might make morally motivated internalizers less inclined to keep their promises than they would have been in the absence of such a rule.
The empirical literature suggests that crowding out of virtuous motivations is not a unitary phenomenon. 18 See Samuel Bowles, Policies Designed for Self-Interested Citizens May Undermine “The Moral Sentiments”: Evidence from Economic Experiments, 320 Sci. 1605 (2008) [hereinafter Bowles, Policies Designed for Self-Interested Citizens] (reviewing this literature). Thus, section III.B distinguishes the particular form of crowding out that I hypothesize might happen due to the presence of internalizers in the subject population from another kind of crowding out that has been discussed in the contract theory literature—namely, crowding out that can arise when legal enforcement prevents contracting parties from expressing their trust in one another by keeping their relationships outside the legal sphere. Whereas this latter kind of sanction-based crowding out grounds arguments in favor of formalist approaches to contract interpretation, 19 See, e.g., Robert E. Scott, A Theory of Self-Enforcing Indefinite Agreements, 103 Colum. L. Rev. 1641, 1685–92 (2003) [hereinafter Scott, Indefinite Agreements] (arguing that the indefiniteness doctrine may avoid crowding out reciprocal social norms); see also Robert E. Scott, A Relational Theory of Default Rules for Commercial Contracts, 19 J. Legal Stud. 597, 614–15 (1990) [hereinafter Scott, Default Rules] (arguing that concerns about crowding out of extralegal norms may make contractual gap fillers that assign risks on an all-or-nothing basis superior to gap fillers that attempt to judicially enforce those norms). the norm-based type of crowding out that I identify suggests that governance by standards that are hard to contract around has some countervailing advantages over governance by bright-line rules. Thus, concerns about norm-based crowding out may help to explain why much of contemporary contract law, especially Article II of the Uniform Commercial Code (UCC), has tended to favor standards over rules. 20 See infra note 145 and accompanying text (explaining this feature of the UCC).
Section III.C examines how stipulated damages clauses—clauses specifying the damages that are to be paid in the event of breach—affect the behavior of internalizers in the light of the penalty doctrine, which specifies the conditions under which such clauses are legally enforceable. 21 See infra note 150 and accompanying text (discussing the penalty doctrine). Effects of such clauses depend on both how internalizers interpret the legal norm governing the perform–breach decision and how they understand the normative import of stipulated damages clauses given their interpretations. 22 See supra notes 16–17 and accompanying text (explaining that internalizers may interpret the norm governing the perform–breach decision as giving rise to a legal duty either to “perform or pay” or to perform unless released by the promisee). If an internalizer views the underlying performance norm as creating a disjunctive duty to either perform or pay expectation damages, she might view a stipulated damages clause as altering the amount she will be duty-bound to pay if she chooses not to perform. She will then perform more or less often depending on whether the clause specifies damages that are higher or lower than the promisee’s actual expectation. Alternatively, consistent with a common understanding of the point of stipulated damages clauses, the internalizer might view the stipulated damages clause as simply an attempt to estimate the promisee’s expectation. 23 There is doctrinal support for such a view. See infra note 153 and accompanying text. If she views the clause in this way, she might plausibly take the view that the clause has not altered the underlying content of her legal duty, in which case she will continue to view herself as duty-bound to perform or pay the true expectation (rather than the estimate embodied in the clause).
If the internalizer views the norm governing the perform–breach decision as pacta sunt servanda—implying that she has a duty to perform unless released by the promisee—then there are two further possibilities. One is that she views a stipulated damages clause merely as specifying the remedy in the event of breach, leaving the content of her primary duty to perform unless released by the promisee unchanged. The more interesting possibility is that the stipulated damages clause transforms her perception of the applicable primary norm, such that she comes to believe that her duty is no longer to perform unless released by the promisee but rather to perform or pay the stipulated amount of damages. In that case, the clause might make her less willing to perform.
Section III.D examines the doctrine of willful breach, a small fault-based pocket of liability in an otherwise pervasively strict liability regime that awards supercompensatory damages for breaches that a court deems “willful.” 24 There is support in the case law for a limited doctrine of willful breach. See infra notes 159, 176–178 and accompanying text. However, this proposition remains controversial. See Posner, Contract Breaker, supra note 17 (defending a fault-free conception of contract law). It is difficult to explain what distinguishes a willful breach from other kinds of breaches if our operating assumption is that all subjects are self-interested externalizers because under that assumption all breaches will be the result of subjects’ pursuit of their self-interest. Thus, standard economic explanations of the doctrine equate willful breaches with breaches that are simply undesirable from an economic standpoint—that is, inefficient. A possible justification of supercompensatory damages then arises: If some inefficient breaches will not be deterred by the operation of the ordinary regime of expectation damages (e.g., because there is a high chance that the breach will go undetected), then damages should be increased to deter these breaches more effectively. But this explanation guts willful of its ordinary meaning, while also failing to explain why the willful breach doctrine employs a fault-based standard, given that efficiency could also plausibly be realized by a rule of strict liability in which supercompensatory liability would be triggered by any breach (regardless of the breaching party’s fault) that falls within a category of inefficient breaches that will be insufficiently deterred by expectation damages.
Once, however, we recognize that there are both externalizers and internalizers in the subject population, a superior explanation of the doctrine emerges. This explanation begins with a more intuitive definition of willful as acting for the wrong reasons. More specifically, a willful breach is one that arises when the promisor fails to give the requisite weight to the expectation interest of her contracting partner. Fault-based supercompensatory liability then becomes a mechanism for discriminating between internalizers and self-interested externalizers. Because they are guided only by their self-interest, these externalizers would be inclined to willfully breach but for the threat of supercompensatory damages. Internalizers, by contrast, never willfully breach, because they take seriously their duty to perform or fully compensate the promisee by paying the promisee expectation damages. Thus, internalizers will always choose the efficient course of action even though they will never be held liable to higher damages. Strict liability, by contrast, may send internalizers the message that it is their duty to perform or pay supercompensatory (rather than mere expectation) damages. So while strict liability could induce externalizers to behave efficiently, it would do so only at the cost of distorting the behavior of internalizers, who would end up performing too often from a social welfare maximizing standpoint.
Finally, section III.E examines the doctrine of duress as it applies to contract modifications by specifying the criteria such a modification must satisfy in order to be enforceable when it was procured by a threat of breach. Traditional economic models that assume that everyone is a self-interested externalizer suggest that modifications should be enforced whenever the party seeking the modification has a credible threat to breach. This is because if the threat to breach is credible and the threatened party agrees to the modification in order to avoid the threatened breach, both parties will be better off if the modification is enforced. If the modification is not enforced, the threatening party will follow through on its threat, and the threatened party will suffer the consequences of breach.
Such a proposal, however, is counterintuitive and not fully reflected in the doctrine. 25 See infra notes 192–194 and accompanying text (discussing the doctrine). The presence of internalizers in the subject population may help to explain why. When the legal rule makes the credibility of the threat the criterion, it provides indeterminate guidance to internalizers. Internalizers want to conform to the rules, but the credibility rule tells them to threaten breach to obtain a modification whenever they would like to do so. But this prescription has indeterminate content given that their preferences about what to do depend on the legal rule. In effect, the rule tells them that they are permitted to threaten breach whenever they would have a preference to breach if the other party does not acquiesce to their demand. But whether internalizers have such a preference depends on the rule. By contrast, a rule that tells subjects that they may “threaten” only inefficient breaches instructs internalizers to refrain from breaching (and threatening breach) even if it would be in their narrow self-interest to do so. Thus, the law can do better than the credibility rule from an efficiency standpoint when there are many internalizers in the subject population.
Together these applications illustrate a more general problem that a legal system faces when there are both internalizers and externalizers in the subject population: Rules that are optimal for internalizers are not necessarily optimal for externalizers. 26 The dominance of the “bad man” approach to questions of legal design means that there has been a tendency to ignore implications of cognitive and motivational diversity when evaluating questions of legal design. But there are notable exceptions. Professors Colin Camerer et al. consider how policies should be designed when the population comprises both rational agents and boundedly rational agents, arguing for “asymmetric paternalism”—policies that create “large benefits for those people who are boundedly rational . . . while imposing little or no harm on those who are fully rational.” Colin Camerer, Samuel Issacharoff, George Loewenstein, Ted O’Donoghue & Matthew Rabin, Regulation for Conservatives: Behavioral Economics and the Case for “Asymmetric Paternalism,” 151 U. Pa. L. Rev. 1211, 1219 (2003) [hereinafter Camerer et al., Regulation for Conservatives]. And Professors Yuval Feldman and Henry Smith develop a model to examine the effects of law versus equity on good-faith and bad-faith people. Yuval Feldman & Henry E. Smith, Behavioral Equity, 170 J. Institutional & Theoretical Econ. 137 (2014). Thus, for example, the doctrine of contract modification provides the right guidance for internalizers, but under many conditions it will fail to create optimal incentives for externalizers. By contrast, in the case of willful breach, the law is able to speak to both constituencies simultaneously: The fault-based standard resolves the potential conflict between the two objectives.
I. Do People Internalize Legal Rules?
A common justification for the economist’s assumption that people are single-minded, self-interested utility maximizers is methodological. Self-interest is undoubtedly an important driver of behavior, and it is simple to model. By contrast, non-self-interested motivations range widely, creating the risk that we end up explaining everything and nothing simply by changing the utility function to fit observed behavior. 27 See Richard A. Posner, Rational Choice, Behavioral Economics, and the Law, 50 Stan. L. Rev. 1551, 1559–60 (1998) [hereinafter Posner, Rational Choice] (arguing that the “descriptive accuracy” of behavioral law and economics is “purchased at . . . the price [of] loss of predictive power”). But at least one kind of non-self-interested motivation is empirically plausible, seems likely to affect behavior in predictable ways, and is relatively easy to pin down ex ante: the motivation to conform to the law. Many people appear to comply with legal rules (and/or nonlegal norms) not out of fear of the sanctions that may result from noncompliance on a particular occasion but because, rightly or wrongly, they simply believe that they ought to comply. 28 On determinants of compliance, see generally Tom R. Tyler, Why People Obey the Law (2d ed. 2006). The economics literature finds that levels of compliance with the tax laws cannot be fully accounted for using the standard economic model of choice under uncertainty, expected utility theory, which focuses exclusively on the expected self-interested costs and benefits of compliance. See, e.g., James Andreoni, Brian Erard & Jonathan Feinstein, Tax Compliance, 36 J. Econ. Literature 818, 850–52 (1998); Joel Slemrod, Cheating Ourselves: The Economics of Tax Evasion, 21 J. Econ. Persp. 25, 38–41 (2007). For theory and evidence that people comply with the tax laws out of a sense of duty, see Philipp Doerrenberg & Andreas Peichl, Progressive Taxation and Tax Morale, 155 Pub. Choice 293 (2013); James P.F. Gordon, Individual Morality and Reputation Costs as Deterrents to Tax Evasion, 33 Eur. Econ. Rev. 797 (1989).
Still, there is room for skepticism about the empirical relevance of such motivations, particularly in market contexts where it is natural to suppose that only those who ruthlessly pursue their own interests survive. 29 See Milton Friedman, The Methodology of Positive Economics, in Essays in Positive Economics 3, 21–23 (1953) (arguing that firms that fail to maximize their profits will be driven out of the market); Armen A. Alchian, Uncertainty, Evolution, and Economic Theory, 58 J. Pol. Econ. 211, 213 (1950) (same); Posner, Rational Choice, supra note 27, at 1570 (arguing that fair-minded people will either avoid or “be forced out of” highly competitive environments). This Part argues that such skepticism is not warranted. Considerable evidence suggests that people are internally motivated to follow norms in general. And there are good reasons to believe that many people are internally motivated to follow legal norms.
A. The Disposition to Follow Norms
To begin with, at least some members of the legal system have to view the law as directly action guiding for the legal system to have a stable existence. If ordinary legal subjects only comply out of fear of sanctions, then there must exist people who enforce the law, and it seems implausible to suppose that all the officials who enforce the law do so only because they are fearful of the negative consequences of noncompliance with their official duties. If that were the case, then there would have to be others to enforce the norms of the official realm, and so on. The chain cannot go on indefinitely. And so it seems that there must be some officials who are internally motivated to conform to the rules. 30 As David Hume put it: “The soldan of Egypt, or the emperor of Rome, might drive his harmless subjects like brute beasts, against their sentiments and inclination: But he must, at least, have led his mamelukes or praetorian bands, like men, by their opinion.” 1 David Hume, Essays and Treatises on Several Subjects 27 (Edinburgh, Bell & Bradfute 1817).
It is conceptually possible that an all-powerful tyrant might instill such fear in her officials that they enforce her edicts against one another out of fear of sanctions. 31 As Professor Frederick Schauer suggests, it may be more than a conceptual possibility. Frederick Schauer, The Force of Law 82 (2015). But to the extent that she must rely on the actions of subordinates in order to govern, her situation is a precarious one. For once it becomes common knowledge that everyone dislikes her, it will cease to be rational for her officials to force others to comply with her commands. It is also not clear that the tyrant’s commands count as genuine laws. Since obedience is instilled by fear, there is no sense in which her edicts exist when not backed up by the threat of sanctions. There is no accepted standard by which she is authorized to rule. Observations like these lead Professor H.L.A. Hart to argue that the existence of a legal system requires, at minimum, that most of the officials of the system adopt an internal point of view toward the secondary rules of the system—the rules that determine what has to be done in order to produce, change, interpret, or enforce the primary rules that govern the conduct of ordinary subjects. 32 Hart, Concept of Law, supra note 1, at 93.
It is possible to put pressure on Hart’s claim. Imagine a society in which all officials obey the tyrant out of fear because they erroneously believe that other officials believe that she has the right to rule, and all the other subjects obey out of fear. 33 See Gregory S. Kavka, Hobbesian Moral and Political Theory 260–66 (1986) (making a suggestion along these lines). In such a regime, officials believe that the secondary rules exist even though no one actually accepts them from the internal point of view. Arguably this might be enough for the system to function like an ordinary legal system. But, resting as it does on a kind of confidence trick, it looks like a pathological case and seems unlikely to be stable in the long run.
Thus, at least in nonpathological cases, it is plausible to suppose that many people regard legal rules as directly action guiding. At the very least, a significant number of officials must adopt an internal point of view toward the secondary rules, which determine how the primary legal rules get altered, interpreted, and enforced. And it would seem like a strange coincidence if this propensity to internalize prevailing norms was confined entirely to the official realm (though later I consider one reason why this might be so 34 See infra notes 53–59 and accompanying text. ).
Indeed, analogous observations suggest that the tendency to conform to prevailing norms must be a more general one, for moral and social norms exist even though no centralized enforcement apparatus exists to enforce them. And in the absence of a centralized enforcement mechanism, it must be the case that at least some people are willing to enforce those norms by applying social sanctions and moral pressure to those who fail to conform to them. 35 See Joseph Henrich & Robert Boyd, Why People Punish Defectors, 208 J. Theoretical Biology 79, 80 (2001) [hereinafter Henrich & Boyd, Punish Defectors] (discussing punishment of defectors in the absence of a centralized authority). The general point here is that enforcement of socially beneficial norms is a costly activity that benefits all. And so if everyone is motivated purely by self-interest, collective-action problems will inhibit their enforcement. 36 Id.
The challenge, then, for the standard economic model is to explain how the large-scale cooperation between distantly related people that is characteristic of the human species arises. 37 See Robert Boyd & Peter J. Richerson, The Evolution of Reciprocity in Sizable Groups, 132 J. Theoretical Biology 337 (1988) [hereinafter Boyd & Richerson, Evolution of Reciprocity] (demonstrating that the likelihood that selection will favor reciprocating strategies decreases as the size of the social group increases). For a set of stylized facts about human cooperation, see Peter J. Richerson, Robert T. Boyd & Joseph Henrich, Cultural Evolution of Human Cooperation [hereinafter Richerson et al., Cultural Evolution], in Genetic and Cultural Evolution of Cooperation 357, 359–60 (Peter Hammerstein ed., 2003). Explaining large-scale cooperation between distantly related people is a challenge, even if a central enforcement apparatus exists, to the extent that such an apparatus itself depends on cooperation between distantly related people. While under certain conditions self-interested agents will rationally behave in a cooperative fashion contrary to their short-run self-interest because of the long-run benefits they derive from avoiding a breakdown of cooperative relations, it is implausible to suppose that these dynamics can sustain social cooperation on a large scale. 38 On the incentive to cooperate in repeated prisoners’ dilemmas, see, for example, Drew Fudenberg & Eric Maskin, The Folk Theorem in Repeated Games with Discounting or with Incomplete Information, 54 Econometrica 533 (1986). Cooperative equilibria become harder to sustain as the number of participants becomes large, and they break down whenever the end of the game comes into sight. 39 Eric Rasmusen, Games and Information 122–24 (2d ed. 1989). And while reputational forces that arise from the gains that can be made from creating the impression that one is a cooperative type expand the potential for cooperation, 40 David M. Kreps & Robert Wilson, Reputation and Imperfect Information, 27 J. Econ. Theory 253 (1982); see also Eric A. Posner, The Strategic Basis of Principled Behavior: A Critique of the Incommensurability Thesis, 146 U. Pa. L. Rev. 1185, 1186–87, 1189–91 (1998). those reputational forces depend on people being well-informed about the past actions of others—a condition that is hard to fulfill as groups become large. 41 David Kreps, A Course in Microeconomic Theory 526–29 (1990). Orthodox evolutionary explanations premised on weakly reciprocal strategies (like tit-for-tat,) which instruct persons to cooperate with those who cooperate and refuse to cooperate with those who do not, and strategies of indirect reciprocity, pursuant to which individuals cooperate only with those who have cooperated with others in the past, suffer from similar limitations. 42 This is perhaps unsurprising in the light of the parallels between orthodox evolutionary theory and orthodox economic theory: “Selfish” genes replace selfish individuals as the fundamental units of analysis, and the instrumental rationality of individuals is replaced by the forces of natural selection. On weak reciprocity, see Robert Axelrod, The Evolution of Cooperation (1984); see also Boyd & Richerson, Evolution of Reciprocity, supra note 37, at 338–40; Rajiv Sethi & E. Somanathan, Understanding Reciprocity, 50 J. Econ. Behav. & Org. 1, 11 (2003). On indirect reciprocity, see Robert Boyd & Peter J. Richerson, The Evolution of Indirect Reciprocity, 11 Soc. Networks 213, 232 (1989).
It is therefore unsurprising that the predictions of the standard economic model are directly contradicted by empirical evidence. Experimental evidence indicates that many people are not only willing to cooperate in one-shot, anonymous interactions in which there is no potential for future rewards or punishment; they are also willing to incur personal costs to punish those who violate cooperative norms. They exhibit, in other words, a disposition toward strong reciprocity. 43 Weak reciprocity involves simply declining to cooperate with defectors. Strong reciprocity involves altruistic punishment of defectors. For an overview of the evidence on strong reciprocity, see Gintis et al., Moral Sentiments, supra note 5, at 8–18. Examples of experiments revealing subjects’ propensities for strong reciprocity include Joyce Berg, John Dickhaut & Kevin McCabe, Trust, Reciprocity, and Social History, 10 Games & Econ. Behav. 122 (1995); Ernst Fehr, Simon Gächter & Georg Kirchsteiger, Reciprocity as a Contract Enforcement Device: Experimental Evidence, 65 Econometrica 833 (1997).
Because strongly reciprocal strategies involve altruistic punishment of defectors while weakly reciprocal strategies do not, the former are weakly dominated by the latter. Nonetheless, strongly reciprocal dispositions may well be evolutionarily stable under the conditions of human evolution. This is because the presence of enough strong reciprocators in the population discourages defection even when there is a low likelihood of future interactions, and the payoff disadvantage experienced by strong reciprocators relative to weak reciprocators declines to zero as the number of defectors declines to zero. 44 Henrich & Boyd, Punish Defectors, supra note 35, at 81; Rajiv Sethi & E. Somanathan, The Evolution of Social Norms in Common Property Resource Use, 86 Am. Econ. Rev. 766, 774 (1996) [hereinafter Sethi & Somanathan, Evolution of Social Norms]. And since cooperation rates are higher when enough strong reciprocators are present, groups with many strong reciprocators tend to do better than groups with few strong reciprocators. 45 Herbert Gintis, Strong Reciprocity and Human Sociality, 206 J. Theoretical Biology 169, 178 (2000) [hereinafter Gintis, Strong Reciprocity]. If, furthermore, cultural forces encouraging conformity (operating in conjunction with slower processes of genetic change) preserve variation between groups, group selection pressures cause strongly reciprocal strategies to thrive in the long run. 46 Robert Boyd, Herbert Gintis, Samuel Bowles & Peter J. Richerson, The Evolution of Altruistic Punishment, in Moral Sentiments and Material Interests, supra note 5, at 219; Gintis, Strong Reciprocity, supra note 45; Henrich & Boyd, Punish Defectors, supra note 35; Sethi & Somanathan, Evolution of Social Norms, supra note 44; see also Richerson et al., Cultural Evolution, supra note 37, at 367 (arguing that it is likely that early man experienced “much genetic change . . . as a result of humans living in groups with social institutions heavily influenced by culture”).
Notice that this evidence is not simply a challenge to the economist’s self-interest assumption. It also challenges the more basic assumption that agents have a fixed set of goals that define the preferences that they rationally maximize. 47 See Samuel Bowles, Endogenous Preferences: The Cultural Consequences of Markets and Other Economic Institutions, 36 J. Econ. Literature 75, 75 (1998) (pointing out that economists have long assumed that preferences are exogenous). This is because it suggests that agents may exhibit a preference for “cooperative” over “uncooperative” behavior, which in turn means that their preferences will be shaped by norms that define what counts as appropriate behavior in their group—norms that may vary group by group. In line with this prediction, the experimental evidence suggests that the specific content of agents’ other-regarding preferences is highly context dependent. For instance, context is an important determinant of how people play ultimatum games. 48 Ultimatum games are games in which two subjects get to split a sum of money in accordance with the division proposed by the subject assigned the role of the “Proposer” so long as the subject assigned to the role of “Responder” accepts the Proposer’s offer (if she rejects it, both receive nothing). An equal division norm appears to be salient for many in unadorned games played in Western societies, with fifty percent being the modal offer and insufficiently generous offers getting punished with rejection. Gintis et al., Moral Sentiments, supra note 5, at 11–13. But Proposers tend to offer less when they have “earned the right” to be Proposers, for example, or when an “exchange frame” is elicited by assigning subjects roles as sellers or buyers. Cristina Bicchieri, The Grammar of Society 123–25 (2006). And people appear to interpret the ultimatum game differently, and so view different kinds of behavior as normatively appropriate, depending on the norms of their culture. For instance, the Au and Gnau view it as a gift-giving game rather than a cake-splitting game. In their society the acceptance of gifts commits a person to reciprocate at a future point in time and establishes that person in a subordinate position. Thus, Proposers regularly make, while Responders often reject, hyper-fair offers. The Aché, by contrast, view the ultimatum game as a sharing game that is governed by their highly cooperative sharing norms that are not punitively enforced. Hence they make generous offers but also tend not to reject low offers. Joseph Henrich, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, Herbert Gintis & Richard McElreath, In Search of Homo Economicus: Behavioral Experiments in 15 Small-Scale Societies, 91 Am. Econ. Rev. 73 (2001).
Even if people are norm guided in general, however, we might question the relevance of such dispositions in the contract arena where competitive pressures reward those who pursue selfish ends. Those who opportunistically try to evade the rules of the marketplace, we might suppose, do better and in the long run drive out those who follow the rules. 49 See supra note 29 (citing sources making such an argument). But a market is a social institution that depends on social cooperation like any other. It is, of course, an institution that creates pressure to maximize profits. But there is no reason to doubt that many market participants nonetheless internalize the rules of the marketplace that constrain the pursuit of those profits. Perhaps temptations to break the rules are more commonplace in market settings, but at most this is a difference of degree rather than kind. 50 Social theorists have long taken this point for granted: “[A]ny conceptual scheme which utilizes only the motivational elements of rational instrumental goal-orientation can be an adequate theory only of certain relatively specialized processes within the framework of an institutionally structured social system.” Talcott Parsons, The Social System 43 (2d ed. 1991). Indeed, the results of an experimental trust game conducted on CEOs from Costa Rica indicated that “reciprocal preferences are even stronger among . . . CEOs [than among] student populations.” 51 Ernst Fehr & Armin Falk, Psychological Foundations of Incentives, 46 Eur. Econ. Rev. 687, 698 (2002).
B. From Social Norms to Legal Rules
I suggested earlier that it would be a strange coincidence if the internal point of view was confined entirely to the official realm. 52 See supra text accompanying note 34. But there is at least one important reason why ordinary legal subjects’ attitudes toward the primary rules of the system might differ from the attitudes of legal officials toward the secondary rules and persons’ attitudes toward the rules of positive morality. The reason is that institutional primary rules differ from moral and social norms in that they arise from a set of institutional procedures—procedures constituted by the secondary rules—rather than in a decentralized fashion. While the institutional format greatly expands the potential for social cooperation by giving society the ability to deliberately create and change many of the prevailing rules, 53 Hart, Concept of Law, supra note 1, at 95. rapid change of the primary rules distances legal subjects from those rules so that it becomes “an open question whether ground-level practice is able to keep up.” 54 Jeremy Waldron, All We like Sheep, 12 Canadian J.L. & Juris. 169, 178–79 (1999). The rules of positive morality and the social norms that constitute the secondary rules, by contrast, have “no presence in society apart from their being practiced and their having a shared normativity—their ‘internal aspect’—in the minds and actions of those who practice them.” 55 Id. at 177. Accordingly, those rules “change through the slow modification of existing practice,” reducing the chance that those who are subject to those rules feel alienated from them. 56 Id. at 178.
But a legal system in which the internal point of view is confined to the official realm is likely to be a highly dysfunctional one. Most likely, those in control of the means of coercion have appropriated the machinery of law for their own benefit, ruling in such a way that ordinary legal subjects feel alienated from the rules. 57 Hart, Concept of Law, supra note 1, at 202. If, by contrast, the legal system is at least minimally decent, it is less likely that subjects will feel alienated from the primary rules in this way, and so we should expect more people to voluntarily comply with the law. There is empirical evidence that supports this conjecture. 58 See supra note 28. And, as we will now see, theoretical considerations combined with additional empirical evidence suggest that many ordinary legal subjects are internally motivated to comply with legal rules. 59 See infra Part II.
II. A Simple Model of the Internal Point of View
This Part develops a simple model of the preferences of internalizers, according to which internalizers treat legal rules as constraints on their decisionmaking. Section II.A describes the model and explains how an internalizer’s attitude toward legal rules differs from that of an externalizer. Section II.B outlines some theoretical considerations that lend support to this model by examining the reasons why a person might internalize legal rules.
A. Externalizers and Internalizers
An externalizer simply weighs the costs and benefits of likely rewards and sanctions against the ordinary, nonlegal reasons—self-interested or otherwise—that count in favor of or against her available options. 60 Economic analysis of law assumes that agents are self-interested, see supra note 5 and accompanying text, but the defining feature of an externalizer is not that she is self-interested but that her preferences are exogenous to the legal regime. Thus, legal rules influence her behavior only by changing the nature of the options in her choice set—typically, by exposing her to the risk of legal sanctions if she fails to comply. The externalizer’s preferences do not directly depend on the rules. And so no modification of the standard economic model is required. 61 See supra notes 7–9 and accompanying text (discussing the traditional rational-actor model).
An internalizer, by contrast, believes that legal rules matter because they are the rules, rather than because of the consequences that result from defiance of them. Thus, an internalizer will exhibit a preference to comply with the rules, even if the options in her choice set are held constant. In particular, she will exhibit a preference to comply with the rules regardless of the risk of legal sanctions. Moreover, her preference to comply is driven by the rules themselves rather than by considerations that would motivate her in the absence of law. A subject who performs her contractual obligations only because there is a moral reason to keep her promises is an internalizer of morality but not of the law. 62 See Schauer, supra note 31, at 6, 48–50 (explaining that a person who conforms to the law because of law-independent moral reasons is not obeying the law qua law).
How exactly do legal rules change an internalizer’s preferences? Economic models of internally motivated actors typically suppose that internalizers simply balance a desire to conform to the norms that they have internalized against the other reasons that motivate them. 63 See, e.g., Robert Cooter, Do Good Laws Make Good Citizens? An Economic Analysis of Internalized Norms, 86 Va. L. Rev. 1577, 1581–88 (2000); Robert Cooter, Expressive Law and Economics, 27 J. Legal Stud. 585, 588–89 (1998); Brian Erard & Jonathan S. Feinstein, The Role of Moral Sentiments and Audit Perceptions in Tax Compliance, 49 Pub. Fin. 70, 74–77 (1994); Louis Kaplow & Steven Shavell, Moral Rules, the Moral Sentiments, and Behavior: Toward a Theory of an Optimal Moral System, 115 J. Pol. Econ. 494, 497–98 (2007); Raúl López-Pérez, Aversion to Norm-Breaking: A Model, 64 Games & Econ. Behav. 237, 243 (2008). But legal norms purport to tell subjects what to do regardless of their inclinations to do otherwise. 64 In Hart’s terms, they are peremptory reasons. See H.L.A. Hart, Essays on Bentham: Studies in Jurisprudence and Political Theory 253–55 (1982). In Professor Joseph Raz’s terms, they are exclusionary reasons—second-order reasons (i.e., reasons “to act for a reason or to refrain from acting for a reason”) that exclude and replace (a subset of) the underlying first-order reasons that the agent would otherwise balance in determining what to do. Joseph Raz, Practical Reason and Norms 39 (1999). Thus, if an agent simply weighs a desire to conform to a legal rule against the nonlegal reasons that favor noncompliance, she has not perfectly internalized the rule: She will end up defying the rule whenever she perceives that the nonlegal reasons favoring noncompliance are sufficiently weighty. A perfect internalizer, by contrast, will treat legal rules as constraints on her decisionmaking—not simply as considerations to be weighed against others.
B. Why Internalize Legal Rules?
It is ultimately an empirical question whether agents are perfect or imperfect internalizers. But when we reflect on the reasons why agents might be internally motivated to conform to legal rules, we see that there are some reasons to suppose that internalizers rationally will regard legal rules as constraints on their pursuit of other objectives.
There are broadly two kinds of reasons why someone might internalize legal rules. 65 Elsewhere, I explore in much greater depth how the particular grounds of an agent’s disposition to internalize legal rules affect her behavior. See Rebecca Stone, Legal Design for the “Good Man,” 102 Va. L. Rev. (forthcoming 2016) (manuscript at 16–25) (on file with the Columbia Law Review). First, she might believe that there are epistemic reasons to conform to legal rules. Such an agent does not believe that there is anything intrinsically valuable about conforming to legal rules as such. But she believes that following the rules is a way of furthering the nonlegal reasons that she cares about. Second, an agent might internalize rules because she believes that conforming to legal rules is intrinsically valuable.
An agent who internalizes legal rules for epistemic reasons might sound like an externalizer, for an externalizer might give some weight to the expertise that is embodied in a legal rule in a way that makes her more likely to conform to it. But if an externalizer gives weight to this expertise, she does so for purely information-based reasons. Her conformity to the law becomes more likely only because she believes that legal norms embody useful information about what she has most reason to do. She has not internalized the rule: The rule has altered her view of the options in her choice set, but it has not changed her fundamental preferences. 66 Schauer, supra note 31, at 146–47.
But suppose that our agent is also boundedly rational. That is, suppose that she operates under cognitive and motivational constraints that make her decisionmaking error prone and costly. 67 For discussions of bounded rationality, see John Conlisk, Why Bounded Rationality?, 34 J. Econ. Literature 669 (1996); Jolls et al., supra note 5, at 1477 (1998); Herbert A. Simon, A Behavioral Model of Rational Choice, 69 Q.J. Econ. 99 (1955). If, in addition, she is aware that she is boundedly rational, she might believe that the simpler strategy of following the law (i.e., internalizing legal rules) is a better way of satisfying the balance of the nonlegal reasons that she cares about than trying to figure out how to satisfy those reasons herself. 68 Cf. Conlisk, supra note 67, at 671 (“For a boundedly rational individual, heuristics often provide an adequate solution cheaply whereas more elaborate approaches would be unduly expensive.”).
Our boundedly rational agent might believe, for example, that following the rules is a way of ensuring that she does what she has most reason to do anyway, in line with Professor Joseph Raz’s theory of authority. That is, she might believe that Raz’s normal justification thesis is satisfied—that she is “likely better to comply with reasons which apply to [her] . . . if [she] accepts the [rules] . . . as authoritatively binding and tries to follow them, rather than by trying to follow the reasons which apply to [her] directly.” 69 Joseph Raz, The Morality of Freedom 53 (1988) (emphasis omitted). Although it is unlikely that the legal system satisfies the normal justification thesis in general, 70 See Joseph Raz, The Obligation to Obey: Revision and Tradition, 1 Notre Dame J.L. Ethics & Pub. Pol’y 139, 147–48 (1984) [hereinafter Raz, Obligation to Obey]. the thesis has plausibility in some regulatory domains, especially those that regulate cooperative activity on a large scale. In such domains, the problems that regulation seeks to solve are often complex and subjects’ underlying obligations highly interdependent such that a centralized authority may be better at structuring those obligations fairly and efficiently than subjects reasoning in a decentralized manner for themselves. 71 See id. at 146, 148–49; see also Joseph Raz, Introduction, in Authority 1, 9–10 (Joseph Raz ed., 1990). There is evidence that perceptions that others are complying are determinants of taxpayers’ willingness to comply with tax laws. See James Alm, Betty R. Jackson & Michael McKee, Estimating the Determinants of Taxpayer Compliance with Experimental Data, 45 Nat’l Tax J. 107, 111 (1992); John T. Scholz & Mark Lubell, Trust and Taxpaying: Testing the Heuristic Approach to Collective Action, 42 Am. J. Pol. Sci. 398 (1998); Slemrod, supra note 28, at 40. Think about the decision of a well-meaning factory owner as to how much of some polluting activity she should engage in. How much of the activity she ought to engage in depends on, among other things, the number of other polluters in the area, the ways in which that pollution builds up to affect the surrounding community, and any offsetting benefits of the polluting activity both to the industry and to the community at large, as well as principles that determine the normative significance of such facts. A lawmaker is likely to be better positioned than any individual factory owner to figure out how much of the polluting activity she ought to engage in in the light of this complex set of considerations.
A boundedly rational agent who internalizes legal rules for epistemic reasons such as these will not be inclined to balance the rules against the nonlegal considerations that she cares about, precisely because her bounded rationality makes it hard for her to evaluate the balance of those nonlegal considerations herself. Instead, she will view the legal rules as standing in the place of the nonlegal reasons that she ultimately cares about when she is deciding whether or not to conform to the rules. 72 The legal rules operate as “exclusionary” or “peremptory” reasons. See supra note 64 and accompanying text (defining exclusionary and peremptory reasons). In other words, she will ignore those nonlegal reasons to the extent that they conflict with the rules. By simply following the rules, she serves those nonlegal reasons better than she would were she to attempt to satisfy them directly. 73 Of course, the preferences that describe what such an agent ultimately cares about remain unchanged by the legal rules that she has internalized. She conforms to the rules as an indirect way of serving the reasons that she ultimately cares about. But the preferences that actually govern how she makes choices will diverge from those ultimate preferences by prioritizing conforming to the rules over those reasons. It is as if she cares about conforming to the rules for their own sake. See Stone, supra note 65 (manuscript at 16–18) (explaining that legal norms transform an epistemic internalizer’s “actual preferences” but not her “true preferences”).
Of course, an agent’s bounded rationality does not imply that she should respect the rules come what may. On a particular occasion, she might be able to do better by weighing the reasons herself rather than by following the rules. She might find herself with more time than usual to make the decision, for example. Or she might believe that she is particularly adept at making the type of decision that she has been presented with. But on such an occasion, it is not that she balances the legal rule against nonlegal considerations. Rather, her reasons for deference to the rules disappear, so that the rules drop out of her deliberations entirely. In other words, she behaves like an externalizer with respect to the rules.
Even if our boundedly rational agent does not regard legal rules as authoritative pronouncements in Raz’s sense, she might nonetheless internalize legal rules because she believes that conforming to the law is a way of supporting institutions that tend to promote the good or justice better than alternatives that might emerge in their absence. 74 See Liam Murphy, What Makes Law 129–30 (Brian H. Bix & William A. Edmundson eds., 2014) [hereinafter Murphy, What Makes Law]; Raz, Obligation to Obey, supra note 70, at 147–48. Reasons to support justice-promoting institutions do not give agents a reason to treat the legal norms as directly action guiding. On their own, they simply direct people to weigh the benefits and costs of compliance in order to figure out whether following the law on any particular occasion makes sense from the perspective of justice or the good. 75 Murphy, What Makes Law, supra note 74, at 130. But once we add the plausible assumptions that subjects are boundedly rational and view conforming to the law as a sensible rule of thumb because of the tendency of defiance of legal rules to undermine reasonably just institutions, then they have reason to internalize legal rules as a simpler means of avoiding these consequences than undertaking the cost-benefit analysis themselves. 76 See supra notes 67–68 and accompanying text (explaining that the simple strategy of following the law may be a superior way for a boundedly rational agent to pursue her objectives than trying to pursue those objectives directly).
Indeed, considerations of long-term self-interest, in conjunction with an agent’s bounded rationality, might make her willing to internalize legal rules for purely self-interested reasons. 77 This might have been what Hart meant when he noted that “allegiance to the [legal] system may be based on . . . calculations of long-term interest.” Hart, Concept of Law, supra note 1, at 203; see also Schauer, supra note 31, at 137 (noting that in some legal regimes, officials may “internalize rules more out of fear than substantive acceptance”). Given the risk of legal sanctions and, perhaps, social sanctions that attend nonconformity to legal norms, it could be rational for a boundedly rational agent simply to comply with legal norms rather than weighing the self-interested costs and benefits of conforming whenever she must make a decision. The strategy of complying with legal rules would operate as a heuristic for avoiding certain negative consequences of defiance. 78 In Raz’s terms, “secondary prudential considerations” may lead agents to comply. Joseph Raz, The Authority of Law 243 (2d ed. 2009). And should an agent follow such a heuristic, she will exhibit a preference to follow the rules regardless of her inclinations to do otherwise, thereby treating the underlying self-interested considerations that count in favor of or against complying on a particular occasion as irrelevant to her decisionmaking. 79 See Conlisk, supra note 67, at 671 (explaining that following heuristics is simpler than case-by-case decisionmaking). Thus, considerations of long-term self-interest can ground an agent’s general preference to conform to the rules, even on occasions when the balance of self-interested considerations favors noncompliance. 80 An intermediate possibility is that people voluntarily conform to the law because they seek the esteem of others and following the law is generally considered the right thing to do. On esteem, see Geoffrey Brennan & Philip Pettit, The Economy of Esteem (2004). On its role in explaining norms, see id. at 267–88. On the role of esteem in explaining conformity to the law, see Richard H. McAdams, An Attitudinal Theory of Expressive Law, 79 Or. L. Rev. 339 (2000).
Of course, the significance of these observations depends on how likely it is that actual subjects internalize legal rules for self-interested reasons. This is ultimately an empirical question, but it is plausible to suppose that there are significant numbers of self-interested internalizers. It seems likely, for example, that many corporations will behave like self-interested internalizers of legal rules. Coordination and agency problems make corporate decisionmaking especially complex, such that even a corporation that is ultimately interested only in maximizing its own profits may often find it advantageous to adopt a standing policy of conformity to legal rules. 81 See Stone, supra note 65 (manuscript at 16); supra text accompanying note 65.
Instead of internalizing rules for epistemic reasons, an agent might believe that conforming to legal rules is intrinsically valuable and so internalize legal rules, even if she is not boundedly rational, simply because she believes that following the rules is the right thing to do. For instance, she might believe that rules promulgated pursuant to fair and democratic procedures are intrinsically worthy of her respect. 82 See Thomas Christiano, The Authority of Democracy, 12 J. Pol. Phil. 266, 286 (2004) (“Citizens who skirt democratically made law act contrary to the equal right of all citizens to have a say in making laws when there is substantial and informed disagreement.”); Dale A. Nance, Rules, Standards, and the Internal Point of View, 75 Fordham L. Rev. 1287, 1293 (2006) (“A sense of authorship becomes a wellspring of the internal point of view.”). Relatedly, she might believe that the rules of a legal system that adheres to norms of procedural justice ought to be obeyed. 83 See generally Tyler, supra note 28. Alternatively, she might believe that because she benefits from the rules, she has reasons of fair play to respect them. 84 As a normative matter, it is not clear why the receipt of unchosen benefits generates an obligation to comply. See Raz, Obligation to Obey, supra note 70, at 152. But our question here is a positive one. And powerful norms of reciprocity might give rise to a feeling of obligation to obey. See supra notes 43–51 and accompanying text (discussing evidence suggesting that humans have a disposition toward strong reciprocity). Or she might believe that she ought to conform to the rules because she has consented to them in some way—a plausible belief when we are in the realm of contract law, 85 See Zev J. Eigen, The Devil in the Details: The Interrelationship Among Citizenship, Rule of Law and Form-Adhesive Contracts, 41 Conn. L. Rev. 381, 398–99 (2008) (suggesting that persons feel more bound by an agreement that they have had a part in negotiating); Zev J. Eigen, When and Why Individuals Obey Contracts: Experimental Evidence of Consent, Compliance, Promise, and Performance, 41 J. Legal Stud. 67 (2012) (providing experimental evidence that participation in the negotiation of terms increases subjects’ willingness to comply with the resulting agreement). given that the parties to a contract have considerable scope to define the rules that govern their relationship. 86 The extent to which the rules are a result of party choice is, of course, limited. “Agreement rules,” which “specify the conditions and procedures the parties must satisfy in order to change an otherwise applicable background rule,” are not chosen by the parties. Richard Craswell, Contract Law, Default Rules, and the Philosophy of Promising, 88 Mich. L. Rev. 489, 503 (1989). American courts do not condition enforcement of a contract on a showing that the parties intended to be legally bound. Restatement (Second) of Contracts § 21 (Am. Law Inst. 1981) (“Neither real nor apparent intention that a promise be legally binding is essential to the formation of a contract . . . .”). And many rules of contract, like the duty of good faith, are difficult to contract around. See U.C.C. § 1-302(b) (Am. Law Inst. & Unif. Law Comm’n 2014); E. Allan Farnsworth, Contracts § 7.17, at 489 (4th ed. 2004).
Whether agents who believe that there are intrinsic reasons to respect the rules also believe that the rules should be given preeminent weight over other considerations will depend on the weight that they give the underlying values—procedural fairness, fair play, consent—that they understand the rules to be serving. They might believe that these values can be outweighed by other considerations. But, since these are moral values, many will at a minimum view them as trumping self-interested considerations and so will at least prioritize conforming to the rules over the pursuit of such considerations.
The arguments in this section suggest that it makes sense to think of an internalizer of a legal norm as one who maximizes her nonlegal preferences—preferences that reflect the considerations that motivate her in the absence of a legal rule—subject to the constraint of conforming to the norm. At the minimum, it makes sense to suppose that she will subordinate self-interested considerations to conforming to the rules.
Thus, exactly why people internalize norms may be less important here than the fact that many do. This is not to say that the reasons why people are prepared to treat legal rules as action guiding are irrelevant from a consequentialist standpoint. Complete agnosticism about the reasons why people internalize legal rules is unlikely to be justified when we are making predictions about their behavior. But in many cases the “why” question will be detachable from the “what” question, greatly simplifying the analysis. Thus, for the most part, I bracket the former question and treat internalizers as a relatively homogeneous category. 87 Elsewhere, I depart from this agnosticism by systematically exploring the ways in which the particular grounds of an agent’s disposition to internalize legal rules matter for her behavior. Stone, supra note 65.
With this simple picture of an internalizer in mind, we can explore some contract law applications. These applications illustrate the ways in which the economic analysis of contract law changes when we assume that there are internalizers in the subject population.
A. The Perform-or-Breach Decision
Whether an externalizer of legal rules will perform or breach a contract depends on the remedy for breach and her nonlegal preferences, which, recall, are the preferences that would completely determine her choices in the absence of an applicable legal rule. 88 See supra section II.B (defining nonlegal preferences). If, as is usually the case, the remedy for breach is expectation damages, 89 U.C.C. § 1-305; Restatement (Second) of Contracts § 347. then, assuming perfect enforcement, a self-interested externalizer will breach whenever the value of breaching exceeds the value to the promisee of performance—that is, whenever breach would be the efficient outcome. 90 Charles J. Goetz & Robert E. Scott, Liquidated Damages, Penalties and the Just Compensation Principle: Some Notes on an Enforcement Model and a Theory of Efficient Breach, 77 Colum. L. Rev. 554, 558 (1977) [hereinafter Goetz & Scott, Liquidated Damages]. Because expectation damages force her to compensate the promisee for the value of the lost performance, 91 Restatement (Second) of Contracts §§ 344(a), 347. thus forcing her to internalize the cost that she imposes on the promisee by breaching, she will always make a social welfare maximizing choice. Thus, from an economic standpoint, her decision is optimal. This is the central claim of efficient breach theory. 92 Goetz & Scott, Liquidated Damages, supra note 90, at 558.
Of course, in practice, enforcement is not perfect. The promisee might not realize that there has been a breach. Even if he realizes that there has been a breach, he might not sue the promisor. Even if he sues the promisor, he might not be able to prove that there was a breach, or if he can prove a breach, he might not be able to prove the entirety of his lost expectation with sufficient certainty. Anticipating that enforcement will be imperfect for reasons such as these, the self-interested externalizer will sometimes breach even when performing is the efficient option. 93 Daniel A. Farber, Reassessing the Economic Efficiency of Compensatory Damages for Breach of Contract, 66 Va. L. Rev. 1443, 1444–45 (1980); see also Friedmann, Efficient Breach Fallacy, supra note 16, at 6–7 (explaining that the efficient breach hypothesis makes the unrealistic assumption that enforcement is costless).
If the remedy is specific performance, the externalizer’s calculus will be different. So long as the promisee realizes that the externalizer has breached the contract and he is motivated to bring suit and able to prove that the breach occurred, a court will order the promisor to perform anyway. 94 Thomas S. Ulen, The Efficiency of Specific Performance: Toward a Unified Theory of Contract Remedies, 83 Mich. L. Rev. 341, 364 (1984). If such a result is sufficiently likely, and the court can easily enforce its order, then the self-interested externalizer might as well perform or negotiate with the promisee to release her from her obligation to perform. 95 See Daniel Markovits & Alan Schwartz, The Myth of Efficient Breach: New Defenses of the Expectation Interest, 97 Va. L. Rev. 1939, 1944–45 (2011) (explaining that when transaction costs are low enough the parties will rationally renegotiate around an order of specific performance when performance would be inefficient).
I have been assuming that our externalizer is self-interested. An externalizer might instead be morally motivated. Such an agent might perceive that she has moral reasons to keep her promises. 96 For empirical evidence that people are so motivated, see Gary Charness & Martin Dufwenberg, Promises and Partnership, 74 Econometrica 1579 (2006); Tore Ellingsen & Magnus Johannesson, Promises, Threats and Fairness, 114 Econ. J. 397 (2004); Christoph Vanberg, Why Do People Keep Their Promises? An Experimental Test of Two Explanations, 76 Econometrica 1467 (2008); Tess Wilkinson-Ryan & Jonathan Baron, Moral Judgment and Moral Heuristics in Breach of Contract, 6 J. Empirical Legal Stud. 405 (2009); Florian Ederer & Alexander Stremitzer, Promises and Expectations (Cowles Found., Discussion Paper No. 1931, 2016), http://ssrn.com/abstract=2368941 (on file with the Columbia Law Review). In that case, she may well perform even absent a sufficient self-interested reason to do so. Notice, though, that if she does so, it is her moral beliefs, not the legal rule, that are driving her to perform. If her moral beliefs were otherwise, she would no longer be motivated to perform. 97 See supra note 62 and accompanying text (differentiating internalizers of morality from internalizers of law).
What about internalizers? An internalizer is motivated to conform to legal norms regardless of her moral beliefs. 98 See supra text accompanying note 62. Thus, she must first figure out what the applicable legal norm is. She must, in other words, engage in an interpretative exercise to figure out the content of the law.
Liability for breach of contract is generally strict: A breaching promisor is liable regardless of whether or not she is at fault for breaching. 99 See Restatement (Second) of Contracts ch. 11, intro. note (Am. Law Inst. 1981) (“Contract liability is strict liability . . . . The obligor is therefore liable in damages for breach of contract even if he is without fault and even if circumstances have made the contract more burdensome or less desirable than he had anticipated.”). And if, as is ordinarily the case, the rule of expectation damages applies, then the promisor must compensate the promisee for the value of the performance of which he has been deprived so as to make him indifferent between breach and performance. 100 Restatement (Second) of Contracts § 347; see also U.C.C. § 1-305(a) (Am. Law Inst. & Unif. Law Comm’n 2014).
How, then, will an internalizer understand her legal duties if the remedy is expectation damages? Her legal duty in the event that she breaches is clear: She must compensate the promisee for the value of the performance of which he has been deprived. 101 See Restatement (Second) of Contracts §§ 344(a), 347. Thus, if she breaches, she will pay the promisee his full expectation.
It is less clear how an internalizer will understand the primary legal norm that governs her decision whether to perform or breach the contract in the first place. She might understand herself as simply duty-bound to perform absent a release from the promisee. Such an interpretation is consistent with the moral reading of contract according to which pacta sunt servanda—agreements must be kept. 102 See Friedmann, Efficient Breach Fallacy, supra note 16, at 2–13 (defending such an interpretation). On this view, her duty to pay expectation damages in the event of breach is a purely remedial duty—a second-best substitute for performance that “exists to serve, so far as may still be done, the reasons for the primary obligation that was not performed when its performance was due.” 103 John Gardner, What Is Tort Law For? Part 1–The Place of Corrective Justice, 30 Law & Phil. 1, 33 (2011). If this is the internalizer’s view of her legal duties, then she will perform regardless of the benefits of breach, unless she can persuade the promisee to release her from her obligation to perform.
Alternatively, the internalizer might see the damages “remedy” as part of the promisee’s right and so understand herself to be under a disjunctive duty to perform or pay expectation damages. 104 See supra note 17 (citing sources that defend this interpretation of a promisor’s legal duties). As Oliver Wendell Holmes put it: “The duty to keep a contract at common law means a prediction that you must pay damages if you do not keep it,—and nothing else.” 105 Holmes, Path of Law, supra note 7, at 462. This Holmesian interpretation of a promisor’s duties is favored by economists because it encourages efficient breaches if actors are self-interested and enforcement is perfect. 106 See Markovits & Schwartz, supra note 95 (defending the view that contracts typically impose duties to either perform or pay). The fact that contractual liability is strict rather than fault based lends support to this interpretation because liability in a strict liability regime does not depend on a showing that the defendant acted wrongfully. 107 See John C.P. Goldberg & Benjamin C. Zipursky, Torts as Wrongs, 88 Tex. L. Rev. 917, 951–52 (2010) (noting that strict liability in tort for abnormally dangerous activities—that is, liability for conduct that is ordinarily considered permissible—is thought to pose a challenge to the view that tort law is about wrongs). This is not a decisive argument. The fact that liability may be imposed without a showing of fault is consistent with the defendant having acted wrongfully in the law’s eyes. And the Second Restatement appears to endorse this interpretation by refusing to distinguish among breaches according to the culpability of the promisor. 108 See Restatement (Second) of Contracts ch. 16, intro. note (Am. Law Inst. 1981) (“‘Willful’ breaches have not been distinguished from other breaches, punitive damages have not been awarded for breach of contract, and specific performance has not been granted where compensation in damages is an adequate substitute for the injured party.”); Posner, Contract Breaker, supra note 17, at 1350 (“As long as you pay the damages awarded by the court in the promisee’s suit for breach of contract . . . no blame can attach to your not performing even if it was deliberate . . . .”). If the law disapproved of a promisor’s decision to breach and pay damages, we might expect intentional breaches to be punished more harshly than accidental breaches, but, for the most part, all breaches are treated the same. 109 In fact, as explained in section III.D, the law does occasionally punish certain breaches more harshly than others, but I will argue that this can be reconciled with an interpretation of the law as creating a duty to preserve a promisee’s expectation rather than to perform. We might also expect promisors to be under a general duty to disgorge gains that they have accrued from breach on unjust enrichment grounds, but the law creates no such general duty. 110 A promisee is entitled to seek restitution for benefits he has conferred on the promisor as a result of his own performance in lieu of expectation damages. But, with a couple of exceptions, a promisee may not recoup additional gains that the promisor makes from breach on the grounds that the promisor has been unjustly enriched by the breach. For example, if the promisor sold a good she had promised to the promisee to a higher bidder, courts are not supposed to order the promisor to disgorge the extra profit she received to the promisee. See Restatement (Second) of Contracts §§ 370–371; Farnsworth, supra note 86, § 12.20, at 823–24; Andrew Kull, Disgorgement for Breach, the “Restitution Interest,” and the Restatement of Contracts, 79 Tex. L. Rev. 2021, 2027 (2001); Posner, Contract Breaker, supra note 17, at 1355. Professor Andrew Kull points out that some courts have recently departed from this view, but he persuasively argues that this is based on a misunderstanding of the Second Restatement. Kull, supra, at 2021–44. Finally, specific performance is the exception, not the rule. 111 Farnsworth, supra note 86, § 12.6, at 746. The preference for damages over specific performance is suggestive of, but is not a decisive reason to favor, the disjunctive interpretation of a promisor’s duties, for there are reasons that are orthogonal to the content of the primary duty, such as the avoidance of administrative costs, why courts might not want to force performance. See Anthony T. Kronman, Specific Performance, 45 U. Chi. L. Rev. 351, 373 (1978) (“There is another common explanation for the reluctance of courts to enforce private injunctive agreements: the specific enforcement of contracts . . . entails special administrative costs which normally can be avoided under a money damage rule . . . .”). Although none of these considerations is decisive on its own, together they lend support to a perform-or-pay interpretation of a promisor’s duties. 112 For a contrary view, see Friedmann, Efficient Breach Fallacy, supra note 16.
Of course, subjects who have had limited contact with the legal system may not have a sophisticated understanding of their legal duties. They may instead simply guess that their contractual duties mirror their moral duties. 113 Cf. Robert E. Goodin, An Epistemic Case for Legal Moralism, 30 Oxford J. Legal Stud. 615, 623–24 (2010) (arguing that “our first recourse, in trying to decide what the criminal law requires, is to reflect upon ‘ordinary standards of moral behaviour’”). In other words, they may suppose that they are simply duty-bound to perform. 114 For evidence that this is how people understand their moral duties, see supra note 96. If, however, the considerations listed above lend support to the perform-or-pay interpretation of a promisor’s duties, then contact with the legal system will tend to encourage such subjects to adopt such a view of their legal duties. Moreover, it is likely that a social welfare maximizing lawmaker will want to encourage promisors to adopt this disjunctive view. This is because, at least so long as the parties will not be investing in the contract, efficiency is more likely to be realized when the rule is “perform or pay” rather than “perform unless released by the promisee” whenever, as is likely to be the case, the transaction costs created by parties’ attempts to renegotiate around orders of specific performance will be high and the ex ante costs of contracting will be lower when the rule is “perform or pay.” 115 For an argument to this effect, see Markovits & Schwartz, supra note 95, at 1973–77. If such transaction costs are zero, we should be indifferent among contract law remedies from an ex post efficiency standpoint. See id. at 1973. Inducing efficient investment may require a remedy other than expectation damages. See id. at 1956 n.34, 1963 n.44.
A caveat is that in circumstances in which it will be difficult to quantify the value of the lost performance to the promisee, a promisor may feel that the only way to ensure that she discharges her duty to perform or pay is to perform. Performing guarantees the promisee his expectation. By contrast, when the value of performance to the promisee is difficult to estimate, making a good-faith effort to compensate the promisee for the value of that performance runs the risk that the promisee will be undercompensated. 116 Consider, for example, an employment contract in which an employee who possesses a special skill agrees not to work for a rival. It will be inherently difficult to quantify the value of this exclusivity agreement to the employer, and so the employee may regard performing the contract as the most certain way of discharging her disjunctive obligation to her employer. The promisor may therefore feel that she ought to compensate a risk-averse promisee for the risk of undercompensation if she chooses to pay instead of perform, which would incline her to perform more often. 117 This might not be a straightforward calculation. To calculate the risk premium, the promisor needs to know the promisee’s attitudes toward risk, how the risk interacts with other uncertain prospects in the promisee’s portfolio, and the probability distribution that captures the promisee’s uncertainty about her expectation. Uncertainty about the risk premium will put pressure on the promisor to err further on the side of caution. This is because the law adopts an asymmetrical attitude toward mistakes of over- and underestimation. If the promisor underestimates the promisee’s expectation, then she fails to satisfy her duty, a failure that is worse from the legal point of view the larger is her error. By contrast, if the promisor overestimates the promisee’s expectation, she more than satisfies her duty, but the law cares only that she satisfies her duty; the law does not count it as a bonus that she has paid the promisee more than the promisee is owed. Or she might believe that she should just eliminate the risk of undercompensation entirely by performing. Indeed, sometimes the law orders injunctive relief under these circumstances, in which case it no longer makes sense for the internalizer to understand her duties in disjunctive terms. 118 See Farnsworth, supra note 86, § 12.5, at 744–45 (explaining that courts enforce exclusivity clauses with injunctions when a damages remedy would be inadequate).
Setting this caveat aside, what are the behavioral implications of an internalizer’s belief that she is under a disjunctive legal duty to perform or pay? Since internalizers view legal norms as constraints, they will regard themselves as duty-bound to either perform or fully compensate the promisee for breach. 119 See supra section II.B (defending the view that internalizers likely view legal norms as constraints). Thus, unlike externalizers, the behavior of internalizers will not be influenced by departures from perfect enforcement: An internalizer will try to pay the promisee his full expectation in the event that she opts to breach and pay damages.
Because both performance and breach are legally permissible options, however, whether the internalizer will choose performance or breach plus compensation of the promisee’s expectation depends on her nonlegal preferences—the preferences that would completely determine her choices in the absence of the law. 120 See supra section II.B (defining those preferences). If she is a self-interested internalizer, she will choose whichever option best furthers her self-interest. And so she will choose the social welfare maximizing action, just as a self-interested externalizer would in a legal regime in which enforcement is perfect. 121 For a discussion of the self-interested internalizer, see supra notes 77–80 and accompanying text. Unlike the self-interested externalizer, however, in the event that enforcement is imperfect, a self-interested internalizer will not attempt to shirk her duty to fully compensate the promisee should she choose to pay instead of perform. Thus, she will make the efficient perform–breach decision even if enforcement is imperfect.
If the agent is a morally motivated internalizer, then she will perform so long as she believes that her moral duty to perform her promises trumps her self-interest. 122 See Fried, supra note 4, at 8 (“By promising we transform a choice that was morally neutral into one that is morally compelled.”). For evidence that people are internally motivated to keep their promises, see Vanberg, supra note 96. Even if she does not believe that there is a moral duty to perform, she might be motivated to conform to any prevailing social norm that requires performance. 123 See Lisa Bernstein, Social Norms and Default Rules Analysis, 3 S. Cal. Interdisc. L.J. 59, 67–69 (1993) (arguing that social norms may induce contracting parties to perform).
There is a further possibility if the internalizer is morally motivated. Suppose that she internalizes the rules of contract law for epistemic reasons. More specifically, suppose that she internalizes the rules because she believes that the rules embody authoritative judgments about what she must do to satisfy the nonlegal moral reasons that she cares about and that, given her cognitive limitations, she is more likely to conform to those nonlegal moral reasons by conforming to the rules than by trying to conform to those reasons directly. 124 See supra notes 67–71 and accompanying text (discussing epistemic internalizers). Suppose further that she believes that the judgments embodied in contract law norms are comprehensive. That is, the norms, as she understands them, do not simply set minimum standards. They are not, in her view, like speed limits, which clearly do not embody all the safety considerations a driver should attend to when deciding how fast to drive. Instead, she believes that the legal norms reflect all the moral reasons that are relevant to her decision. In that case, she will believe that compliance with the law makes it unnecessary for her to consider those moral reasons when selecting among the actions that conform to the norm. And so the law will have the effect of crowding out moral considerations from her deliberations when she is choosing between conforming actions. In this way, a perform-or-pay norm will make our internalizer more willing to break a promise than she would have been in the law’s absence. She will regard paying damages for breach as sufficient to discharge any moral duty to the promisee, leading her to breach the contract (and pay damages) in situations in which, in the absence of the rule, she would have felt morally duty-bound to perform.
B. Rules, Standards, and Crowding Out
How plausible is it that crowding out of this kind will occur, and what are its normative implications for the design of contract law rules? Contract theorists have suggested that crowding out of moral and social norms is a concern in the contract law arena, 125 See Lisa Bernstein, Merchant Law in a Merchant Court: Rethinking the Code’s Search for Immanent Business Norms, 144 U. Pa. L. Rev. 1765, 1789–90 (1996) (discussing the relational costs of legal enforcement) [hereinafter Bernstein, Merchant Law]; Scott, Default Rules, supra note 19, at 614–15 (arguing that efforts to legally enforce social norms may undermine the effectiveness of social sanctions); Scott, Indefinite Agreements, supra note 19, at 1685–92 (arguing that concerns about crowding out of reciprocal social norms justifies a formalist approach to contract law). but they have not disentangled different possible psychological mechanisms that may underpin the larger empirical phenomenon. This is the task that I undertake in this section.
Empirical findings and theoretical considerations suggest that crowding out can take at least three different forms. First, bad motives may simply drive out more virtuous motives. Considerable psychological evidence suggests that when agents are offered extrinsic rewards (e.g., money) to work on a task, their intrinsic motivation to work on the task declines. 126 For a review of evidence showing that extrinsic rewards tend to undermine intrinsic motivation, see Edward L. Deci, Richard Koestner & Richard M. Ryan, A Meta-Analytic Review of Experiments Examining the Effects of Extrinsic Rewards on Intrinsic Motivation, 125 Psychol. Bull. 627 (1999). The immediate prospect of extrinsic rewards appears to undermine the motivational salience of other kinds of reasons. But this is not a form of crowding out that is closely associated with norm-guided behavior, and so it is of minimal interest here.
The other two forms of crowding out, however, depend on agents being motivated to follow prevailing norms. Both suggest that the law can undermine agents’ willingness to conform to nonlegal (social and moral) norms. The first arises from legal sanctions; the second arises from legal norms themselves. I will refer to the former as sanction-based crowding out and to the latter as norm-based crowding out. As we will see, they have different sources and potentially different normative implications for the design of contract law.
Much of the empirical evidence on crowding out of social norms focuses (implicitly) on sanction-based crowding out. 127 For a review of the literature on crowding out, see Bowles, Policies Designed for Self-Interested Citizens, supra note 18. But it does not appear from this evidence that the threat of sanctions per se causes the crowding out. Rather, the crowding out seems to arise from the message that sanctions (or their absence) communicate. Threats of material sanctions that are implicit rather than explicit, because they arise endogenously from reciprocity-driven behavior in repeated interactions, do not crowd out other-regarding behavior. 128 Fehr & Falk, supra note 51, at 701–04. That is, when a person simply anticipates that bad behavior will be punished, but such punishment was not promised in advance, her motivation to behave well is not crowded out. But an express, advance commitment to punish bad behavior often does crowd out other-regarding behavior. 129 Id. at 703. In other words, it is the explicit, ex ante nature of the threat of sanctions that seems to matter.
A plausible explanation is that the explicit threat communicates something to the threatened subject that the implicit threat does not, namely that others do not trust her and expect her to behave in a self-interested fashion. 130 Id. This signal, in turn, induces the subject to anticipate that those others will behave self-interestedly toward her, making it rational for her to do the same, even if she would otherwise be willing to reciprocate kindness with kindness. 131 See id. at 695 (discussing evidence that suggests that agents will be less willing to put in effort in response to explicit incentives because those incentives are perceived as evincing hostile intentions). Conversely, the absence of an advance threat of sanctions may signal that others trust her. And so, for example, if courts are too willing to fill gaps in incomplete contracts, thus depriving contracting parties of the ability to leave matters unenforceable in order to signal that they trust their contracting partners, courts may crowd out the parties’ willingness to conform to reciprocal social and moral norms. 132 See Scott, Indefinite Agreements, supra note 19, at 1685–92 (arguing that the indefiniteness doctrine can be defended in these terms).
The final kind of crowding out—norm-based crowding out—does not depend on the signals that are sent by the choices of the parties. Rather, it arises when a legal norm communicates to parties that self-interested behavior is appropriate. Although much of the pertinent empirical evidence does not clearly disentangle the effects of sanctions from the effects of their associated norms, 133 An exception is Roberto Galbiati & Peotro Vertova, Obligations and Cooperative Behaviour in Public Good Games, 64 Games & Econ. Behav. 146 (2008) (finding that obligations have a positive effect on cooperation in a public goods game, holding marginal incentives constant). it is plausible to suppose that crowding out can be a consequence of norms that are perceived to regulate behavior in a comprehensive fashion. 134 Crowding out of this kind may be exhibited by those who internalize legal norms for epistemic reasons and regard legal norms as authoritative expressions of what they have most reason to do. But it will not be exhibited by those who internalize legal rules for intrinsic reasons or those who internalize legal norms for epistemic reasons but do not regard legal norms as expressing authoritative judgments about what they have most reason to do. Stone, supra note 65 (manuscript at 32–36). Norm-based crowding out might, for example, have been driving the results of a famous field experiment that studied the impact of a daycare center’s introduction of a small fine for parents who arrived late to pick up their children. 135 Uri Gneezy & Aldo Rustichini, A Fine Is a Price, 29 J. Legal Stud. 1 (2000). The introduction of the fine perversely increased the numbers of parents arriving late. 136 Id. at 8. A plausible explanation is that the introduction of the fine signaled to parents that it was okay to behave self-interestedly subject to paying the fine. 137 An alternative possibility that is consistent with traditional economic theory is that the fine resolved uncertainty about how the daycare center would react to parents’ lateness such that parents were no longer motivated to be on time out of fear of more serious consequences. Id. at 10–11. Prior to the introduction of the fine, the parents felt that they had a moral duty to the center not to be late because the daycare workers were generously providing their after-hours services for free. Once the fine was introduced, they perceived the center to be telling them it was okay to be late so long as they paid the price (the “fine”), leading them to behave in a self-interested fashion subject to paying the price for any lateness. 138 The tone of the announcement that the center posted on the bulletin board announcing the “fine” made the fine sound more like a price or tax than a “fine” intended to prohibit lateness: “As you all know, the official closing time . . . is 1600 every day . . . . [W]e . . . have decided to impose a fine [of NIS 10] on parents who come late to pick up their children . . . . This fine will be calculated monthly, and it is to be paid together with the regular monthly payment.” Id. at 16. Although the announcement uses the word “fine,” it otherwise suggests mild rather than strong condemnation of parents coming late. An alternative explanation is that parents adhere to a social norm that states: “‘When help is offered for no compensation in a moment of need, accept it with restraint. When a service is offered for a price, buy as much as you find convenient.’” Id. at 14. A final possibility is that the fine changed perceptions of the applicable official norm. Before the fine was introduced, there was a perception that the center required parents not to be late. After the fine was introduced, parents perceived that they were permitted to be late subject to paying the fine. Neither of these explanations, nor the explanation I propose in the main text, easily explains the fact that parents continued to be late when the fine was removed. Id. at 8.
Consistent with this interpretation, many people believe that more permissive legal norms crowd out more prohibitive moral norms, even when both could be complied with simultaneously. For instance, opponents of tradable permits have argued that permits improperly convey the message that pollution is like an ordinary “commodity to be bought and sold,” thus removing “the moral stigma that is properly associated with it.” 139 Michael J. Sandel, Opinion, It’s Immoral to Buy the Right to Pollute, N.Y. Times (Dec. 15, 1997), http://www.nytimes.com/1997/12/15/opinion/it-s-immoral-to-buy-the-right-to-pollute.html (on file with the Columbia Law Review); see also Robert E. Goodin, Selling Environmental Indulgences, 47 Kyklos 573, 582 (1994) (likening the trading of emissions rights to the sale of religious indulgences); cf. Bowles, Policies Designed for Self-Interested Citizens, supra note 18, at 1609 (discussing evidence suggesting that paying people to donate blood makes people less willing to donate); Friedmann, Efficient Breach Fallacy, supra note 16, at 1 (arguing that Holmes’s analysis of contract “mistakenly converts the remedy into a kind of indulgence that the wrongdoer is unilaterally always entitled to purchase”).
What could be going on here, psychologically speaking? There are at least two possibilities. The first is that some legal rules signal to subjects that lawmakers expect them to behave self-interestedly, which in turn encourages them to view self-interested behavior as normatively appropriate all the way down. In other words, laws designed for knaves may produce knaves. 140 See Bruno S. Frey, A Constitution for Knaves Crowds Out Civic Virtues, 107 Econ. J. 1043 (1997). Notice that this explanation implies not just that morally motivated behavior will be crowded out in the interstices of the legal rules, but that compliance with the legal rules themselves will be undermined. This is because self-interested agents will also avoid complying with the rules when those rules are imperfectly enforced if the costs of compliance exceed the expected costs of noncompliance. In the terms of this Article, internalizers of legal norms will be turned into self-interested externalizers.
The second possibility is less extreme and can account for the crowding out described at the end of section III.A. Certain legal rules may simply signal to internalizers that they may behave self-interestedly in the interstices of the rules. If legal rules have this effect, then internalizers will continue to comply with the rules, while acting self-interestedly when deciding how to comply. The rules, in other words, will have the effect of crowding out internalizers’ willingness to attend to moral considerations that are relevant to the choice among conforming actions. Thus, for example, contract law’s perform-or-pay norm might signal to a morally motivated internalizer that she need not worry about moral considerations when selecting between the legally permissible options of performing or breaching and paying full expectation damages. 141 See supra text accompanying note 124. Such an internalizer would then view the legal rules as substituting for the moral norms that she would otherwise feel bound by. She would assume, in effect, that the legal rules have done the relevant moral work for her.
The latter possibility is more plausible than the former. It seems incongruous to suppose that persons who are inclined to look to the legal system for normative guidance would also view the legal system as endorsing noncompliance with its own norms. It is more plausible to suppose that such persons draw the more limited conclusion that their compliance with the legal rules simply renders certain moral considerations irrelevant when choosing among conforming actions.
Notice that norm-based crowding out is not necessarily undesirable from a social welfare maximizing standpoint. Under certain conditions, social welfare will be maximized if, for example, subjects view the disjunctive duty to perform or pay expectation damages as fully exhausting their other-regarding obligations so that the norm crowds out a more restrictive understanding of their moral duties. This is because such subjects will breach if and only if it is efficient to do so. 142 See supra text accompanying note 121. By contrast, a legal subject who, for example, understands herself to be under a residual moral obligation to perform no matter what will sometimes perform when it is inefficient to do so. Thus, if the rule does in fact reflect the balance of all relevant moral considerations, and if the sole objective of the law is social welfare maximization, then it may be optimal for the agent to dispense with her own consideration of those reasons when deciding which conforming action to choose.
But crowding out is undesirable if internalizers treat their compliance with legal norms as sufficient to discharge their obligations to others when the legal norms do not reflect all the relevant moral considerations. In the contract arena it will often make sense for the law to delegate some of the regulatory work to prevailing social and moral norms, particularly in commercial settings in which industry-specific, course-of-dealing-specific, and transaction-specific norms emerge that likely better reflect prevailing marketplace conditions than a distant lawmaker’s attempt to develop an alternative set of norms. 143 An express purpose of the UCC is to “permit the continued expansion of commercial practices through custom, usage, and agreement of the parties.” U.C.C. § 1-103(a)(2) (Am. Law Inst. & Unif. Law Comm’n 2014). But such a delegation will be undermined if internalizers believe that compliance with legal norms obviates the need for them to consider applicable commercial norms when deciding among legally permissible actions. And internalizers may come to believe this if the law fails to communicate transparently to subjects that the legal rules are intended to set minimum standards rather than regulate in a comprehensive fashion. 144 This danger is exacerbated to the extent that subjects suffer from self-serving biases that lead them to resolve moral ambiguity in a way that justifies acting in a self-interested way. For evidence that people do this, see Linda Babcock, George Loewenstein, Samuel Issacharoff & Colin Camerer, Biased Judgments of Fairness in Bargaining, 85 Am. Econ. Rev. 1337 (1995); George Loewenstein, Samuel Issacharoff, Colin Camerer & Linda Babcock, Self-Serving Assessments of Fairness and Pretrial Bargaining, 22 J. Legal Stud. 135 (1993); Alvin E. Roth & J. Keith Murnighan, The Role of Information in Bargaining: An Experimental Study, 50 Econometrica 1123 (1982); Leigh Thompson & George Loewenstein, Egocentric Interpretations of Fairness and Interpersonal Conflict, 51 Organizational Behav. & Hum. Decision Processes 176 (1992).
A concern about norm-based crowding out of this kind may provide a justification for much of contemporary contract law’s emphasis on standards over rules. In particular, it may help to justify the UCC’s instructions to courts to seek out and enforce “immanent business norms”—to liberally fill contractual gaps in line with prevailing usages of trade and parties’ courses of dealing. 145 Bernstein, Merchant Law, supra note 125, at 1766 (internal quotation marks omitted); see also U.C.C. § 1-303(d); Robert E. Scott, The Rise and Fall of Article 2, 62 La. L. Rev. 1009, 1038–39 (2002) [hereinafter Scott, Rise and Fall of Article 2]. A standard wears its justification on its face and so requires a legal subject who is motivated to conform to the standard to exercise her moral judgment in figuring out how to do so. 146 See Seana Valentine Shiffrin, Inducing Moral Deliberation: On the Occasional Virtues of Fog, 123 Harv. L. Rev. 1214, 1222–23 (2010) (“Where standards incorporating moral terms regulate conduct, citizens may themselves have to deliberate about what is morally proper and should be expected of them.”); Jeremy Waldron, Vagueness and the Guidance of Action 9 (N.Y.U. Sch. of Law Pub. Law & Legal Theory Research Paper Series, Working Paper No. 10-81, 2010), http://ssrn.com/abstract=1699963 (on file with the Columbia Law Review) (“[A] standard is a norm that requires some evaluative judgment of the person who applies it, whereas a rule is a norm presented as the end-product of evaluative judgments already made by the law-maker.”). Thus, norm-based crowding out of moral considerations is less likely to occur when an internalizer is trying to conform to a standard than when she is trying to conform to a rule.
The normative implications of sanction-based and norm-based crowding out can therefore point in opposite directions. On the one hand, if the goal is to avoid sanction-based crowding out, then the law should maximize the ability of contracting parties to signal that they trust one another by avoiding legal enforcement of their agreements. 147 For arguments along these lines, see Bernstein, Merchant Law, supra note 125, at 1789–90; Scott, Default Rules, supra note 19, at 614–15. This may be best achieved by filling contractual gaps in a more formalistic fashion with clear rules that are easy for parties to contract around, rather than by liberally filling gaps in accordance with amorphous and harder-to-contract-around norms of good faith and fair dealing. 148 See Bernstein, Merchant Law, supra note 125, at 1766–69; Scott, Rise and Fall of Article 2, supra note 145, at 1038–39. On the other hand, amorphous standards may be preferable to formalistic rules when the goal is to avoid norm-based crowding out.
C. Stipulated Damages Clauses
Contracting parties sometimes specify the monetary remedy for breach in advance by incorporating a stipulated damages clause into their contract. They are not free, however, to stipulate any amount of damages. A stipulated damages clause is enforceable only if the specified amount represents a reasonable estimate of the promisee’s expectation. 149 See Restatement (Second) of Contracts § 356 cmt. b (Am. Law Inst. 1981) (stating that “the amount fixed is reasonable to the extent that it approximates the loss anticipated at the time of the making of the contract”). Stipulated damages in excess of such an amount will be struck pursuant to the penalty doctrine, and the promisor will be forced to pay the promisee the court’s estimate of the promisee’s expectation instead. 150 Farnsworth, supra note 86, § 12.18, at 814–15.
It is clear how such clauses will affect the behavior of externalizers. A stipulated damages clause can alter the average price of breaching a contract. If it increases the price, the externalizer will breach less often. If it decreases it, she will breach more often.
How will such clauses affect the behavior of internalizers? Suppose that an internalizer perceives that, in the absence of a stipulated damages clause, the norm governing the breach–performance decision is a disjunctive one—that she has a duty to perform or pay the promisee expectation damages. 151 See supra notes 104–112 and accompanying text (defending such an interpretation). In that case, it would be natural for the internalizer to suppose that a stipulated damages clause alters the content of the norm such that her legal duty becomes a duty to perform or pay the stipulated damages instead of expectation damages. Such an interpretation makes sense if the purpose of such clauses is simply to reduce risk in order to provide the parties with certainty about the remedy for breach. 152 See Goetz & Scott, Liquidated Damages, supra note 90, at 558–59 (noting that a primary reason for negotiating a stipulated damages clause is that “[t]he expected damages are uncertain or difficult to establish and the parties wish to allocate anticipated risks”); see also Restatement (Second) of Contracts § 356 cmt. b (“The second factor is the difficulty of proof of loss. The greater the difficulty either of proving that loss has occurred or of establishing its amount with the requisite certainty . . . the easier it is to show that the amount fixed is reasonable.”).
But an internalizer might not view the norm as transformed in this way. Pursuant to the penalty doctrine, stipulated damages are supposed to provide a measure of the promisee’s true expectation when it is likely that the promisee’s expectation will be difficult for a court to estimate ex post. 153 See XCO Int’l Inc. v. Pac. Sci. Co., 369 F.3d 998, 1001, 1004–05 (7th Cir. 2004) (“When damages for breach of contract would be difficult for a court to determine after the breach occurs, it makes sense for the parties to specify in the contract itself what the damages for a breach shall be . . . .”); Lake River Corp. v. Carborundum Co., 769 F.2d 1284, 1289–90 (7th Cir. 1985) (“[A] liquidation of damages must be a reasonable estimate at the time of contracting of the likely damages . . . and the need for estimation at that time must be shown by reference to the likely difficulty of measuring the actual damages from a breach of contract after the breach occurs.”). An internalizer who understands this might therefore believe that her duty remains one of performing or paying the promisee his true expectation, and so view the stipulated damages as specifying the amount that she has to pay only if a more precise estimate of the expectation is not available.
What if the internalizer perceives that the prevailing legal norm in the absence of a stipulated damages clause is pacta sunt servanda rather than perform or pay? 154 See supra notes 102–103 and accompanying text (discussing this possibility). Again, there are two possibilities. The first is that the agent continues to perceive that the norm is pacta sunt servanda despite the existence of the stipulated damages clause, so that the effect of the clause is simply to alter what she believes she ought to do in the event that she breaches the contract. Notice that the behavior of an internalizer who interprets a stipulated damages clause in this way will be insensitive to the amount of stipulated damages. This is because such an internalizer will always be motivated to perform, unless she can persuade the promisee to release her from her obligations.
The second possibility is that the stipulated damages clause transforms the promisor’s perception of the legal norm that governs the perform–breach decision. The promisor might infer from the fact that the promisee agreed to the clause that the stipulated damages represent the value of performance to the promisee and so understand herself to be under a disjunctive duty to perform or pay the stipulated amount, even though, absent an advance agreement on damages, she would have believed that she was duty-bound to perform unless released by the promisee. The promisor would then believe that the law deems it perfectly acceptable for her to breach the contract so long as she pays the stipulated damages to the promisee.
Indeed, if this is how the promisor interprets the stipulated damages clause, then even if ordinarily she would be morally motivated to perform her promises, the clause might render her willing to breach the contract and pay damages for purely self-interested reasons if legal norms can crowd out her willingness to adhere to applicable moral norms. 155 See supra section III.B (discussing norm-based crowding out). This is because if such norm-based crowding out occurs, then the promisor will believe that by paying the stipulated damages, she fully discharges her moral obligations to the promisee, thus obviating the need for her to attend to moral considerations when she is deciding how to conform to the law. 156 See Tess Wilkinson-Ryan, Do Liquidated Damages Encourage Breach? A Psychological Experiment, 108 Mich. L. Rev. 633 (2010) (providing some evidence that lay people view liquidated damages clauses in this way).
D. Fault-Based or Strict Liability and Willful Breach
Consistent with the economist’s injunction that efficiency is promoted by giving a promisor the choice between performing and paying damages, 157 See supra notes 88–92 and accompanying text (discussing efficient breach theory). liability for breach of contract is generally strict, as we have already seen. 158 See supra note 101 and accompanying text. But there are exceptions: Some so-called “willful” breaches—breaches that are both profitable and opportunistic—are singled out from the rest by holding perpetrators liable for higher damages. 159 Kull, supra note 110, at 2044–45.
Several scholars have proposed standard economic explanations of these pockets of willfulness. 160 E.g., Oren Bar-Gill & Omri Ben-Shahar, An Information Theory of Willful Breach, 107 Mich. L. Rev. 1479 (2009) [hereinafter Bar-Gill & Ben-Shahar, Information Theory]; Richard Craswell, When is a Willful Breach “Willful”? The Link Between Definitions and Damages, 107 Mich. L. Rev. 1501 (2009) [hereinafter Craswell, Willful Breach]; Steve Thel & Peter Siegelman, Willfulness Versus Expectation: A Promisor-Based Defense of Willful Breach Doctrine, 107 Mich. L. Rev. 1517 (2009). These explanations assume that contracting parties always act in a self-interested fashion. And so they naturally deny that promisors who willfully breach are more culpable than promisors who commit ordinary breaches. The willfulness of a breach ends up being nothing more than the conclusion of the analysis, demarcating a category of breaches that ought to trigger supercompensatory damages when the aim is to induce efficient breach. An evaluation of the promisor’s culpability for the breach is not an independent determinant of liability. 161 Omri Ben-Shahar & Ariel Porat, Foreword: Fault in American Contract Law, 107 Mich. L. Rev. 1341, 1344 (2009). Professors Oren Bar-Gill and Omri Ben-Shahar give some more content to the idea of willfulness, though on their account it is the fact that breach is of a certain type that triggers the higher damages; a finding of culpability is not an independent requirement. According to their model, some kinds of breaches reveal information about the prior undetected conduct of the breaching party and therefore should be punished more harshly in order to deter the undetected conduct. Bar-Gill & Ben-Shahar, Information Theory, supra note 160.
But such reconciliations of the doctrinal tension do not seem entirely satisfactory. We lose something when we remove any notion of fault from the doctrinal landscape. As the name suggests, there seems to be something about so-called willful breachers that makes them more culpable than run-of-the-mill breachers. And in some circumstances courts appear to independently evaluate the culpability of the breach when determining the measure of damages. 162 See cases cited infra note 176 and accompanying text (treating a finding of willfulness as a basis for awarding higher damages).
Professor Richard Craswell argues that it is vacuous to apply the label willful to a breach, if willful is understood as a reference to the defendant’s mental state—to the knowing or intentional nature of his breach. 163 Craswell, Willful Breach, supra note 160, at 1502–04. This is because “‘knowing’ and ‘intentional’ (and their adverb forms, ‘knowingly’ and ‘intentionally’) are most easily applied to specific actions,” while “[a] breach . . . is not an action but a state of affairs . . . . Thus, before we can apply tests like ‘knowingly’ or ‘intentionally’, we need to know the individual actions in [the sequence of actions that caused the breach] to which those terms should be applied.” 164 Id. at 1502. But when we look at any such sequence of actions, we will typically see some actions that were intentional, others that were responses to unforeseen eventualities, and others that were simply mistakes. 165 Id. at 1502–04. For example, the contractor in Jacob & Youngs, Inc. v. Kent apparently used the wrong brand of pipe by mistake, suggesting that we should characterize the breach as unintentional. 166 129 N.E. 889, 890 (N.Y. 1921). Yet, once he realized his mistake, he made a deliberate decision not to rectify the mistake, demolishing the house and starting over, which suggests that the breach was intentional. 167 Id. Or consider a promisor who plans to breach a contract from the very outset. This appears to be a clear-cut case of intentional breach, but our confidence in this conclusion might be shaken by the observation that the promisor “might have lived up to the contract if, after the contract was signed, an eccentric millionaire had unexpectedly offered . . . a reward” for performing. 168 Craswell, Willful Breach, supra note 160, at 1504. These kinds of observations lead Craswell to suggest that the law should dispense with any mens rea inquiry when determining whether to subject a promisor to supercompensatory damages. 169 Id. at 1507. There are reasons grounded in ordinary deterrence considerations to award supercompensatory damages for certain kinds of breaches, and those breaches should be deemed willful regardless of the promisor’s state of mind. 170 Id. at 1507–14.
But Craswell may be focused on too narrow an understanding of the relevant mental state. There is no need to break down the sequence of actions and ask of each action in the sequence whether the agent committed the action on purpose or not. Instead, we can ask a question about her reasons for acting that is pitched at a higher level of generality: What was her motivation for acting as she did? In short, we can try to determine the content of her preferences. Although an agent’s preferences may change over time, generally they will remain fairly stable. Thus, in Jacob & Youngs, the contractor’s behavior is consistent with him having internalized the perform-or-pay norm that requires him to give the same weight to the interests of the promisee as he gives to his own when taking actions that bear on the likelihood of breach. His accidental use of the wrong pipe is consistent, that is, with him having accorded appropriate weight to his client’s interests, for his intention was to use Reading pipe, and he gained no advantage from using the wrong pipe. 171 See Jacob & Youngs, 129 N.E. at 890 (“The evidence sustains a finding that the omission of the prescribed brand of pipe was . . . the result of the oversight and inattention of the plaintiff’s subcontractor.”). Likewise, his subsequent decision not to rectify his mistake at great cost to himself is consistent with him having given equal weight to his own interests and his client’s interests, given that the costs he would have had to incur in order to rectify the mistake were large and the stakes for his client small due to the negligible difference in quality. 172 See id. (explaining that the evidence suggested that the “defect was insignificant in relation to the project”); id. at 891 (explaining that the cost of replacement “would be great”). Thus, there is nothing peculiar or difficult to operationalize about a court taking into account the mental state of a contract breaker, so long as the inquiry is pitched at the right level of generality.
Accordingly, I suggest that we should define a willful breach as one where the contract breaker acts for the wrong reasons under conditions in which the promisee is unlikely to be able to use the legal system to protect his expectation interest. Such a definition is intuitively plausible and coheres with the language used by the courts. Consider, for instance, the following dicta from Judge Benjamin Cardozo in Jacob & Youngs: “The willful transgressor must accept the penalty of his transgression. For him there is no occasion to mitigate the rigor of implied conditions. The transgressor whose default is unintentional and trivial may hope for mercy if he will offer atonement for his wrong.” 173 Id. at 891 (citations omitted). But, more importantly, as we will now see, distinguishing among contract breakers in this way allows the law to speak to externalizers and internalizers at the same time. Once we take culpability out of the equation, then, while the law may provide the right incentives to externalizers, it sends the wrong message to internalizers.
Consider, first, the standard economic rationale for more harshly penalizing a certain subset of breaches. A perfect regime of expectation damages guarantees that self-interested parties will breach if and only if it is efficient to do so. Thus, the main efficiency-based reason to punish a subset of breaches more harshly arises when the regime does not work as it should because, say, breaches will not be detected by the promisee with certainty, or the promisee is otherwise unlikely to bring suit. 174 Another possibility is that the remedy of expectation damages does not adequately protect the promisee’s expectation interest because his expectation is difficult to verify, but this alone cannot explain why the law should care directly about the promisor’s culpability. The promisor’s culpability could have evidentiary significance if opportunistic promisors are more likely to exploit the inadequacy of remedies by inefficiently breaching. But if damages are increased to ensure that remedies are adequate, promisors will cease to have an incentive to opportunistically breach. Under these conditions, a standard economic justification for supercompensatory damages exists as it does in tort: Higher damages offset the negative deterrence effects created by a low probability of suit. 175 See Steven Shavell, Foundations of Economic Analysis of Law 244 (2006) (discussing the standard economic justification of supercompensatory damages in tort); Richard Craswell, Contract Remedies, Renegotiation, and the Theory of Efficient Breach, 61 S. Cal. L. Rev. 629, 664–65 (1988) (offering an analagous justification of supercompensatory damages in contract). Bar-Gill and Ben-Shahar offer a gloss on this explanation for willful breach, arguing that certain types of breaches should be punished more harshly because they signal that it is likely that other kinds of difficult-to-verify breaches have been committed and gone unpunished. Bar-Gill & Ben-Shahar, Information Theory, supra note 160, at 1479–99.
In line with this rationale, courts do appear to punish breaches more harshly when there is a relatively low ex ante probability of suit. When considering breaches of building specifications in construction contracts, for example, courts sometimes allow promisees to recover damages equal to the cost of repairing a nonconforming building rather than the difference in market value between a conforming and nonconforming building, when they find a breach to be willful. 176 In Jacob & Youngs, Judge Cardozo found the breach not to be willful and instead awarded damages equal to the diminution in value, which was negligible. Jacob & Youngs, 129 N.E. at 890–91; see also Kangas v. Trust, 441 N.E.2d 1271, 1275–77 (Ill. App. Ct. 1982) (awarding cost of repair damages upon a finding of willfulness); H.P. Droher & Sons v. Toushin, 85 N.W.2d 273, 279–80 (Minn. 1957) (emphasizing the importance of good faith in determining the measure of damages); Am. Standard, Inc. v. Schectman, 439 N.Y.S.2d 529, 534 (N.Y. App. Div. 1981) (awarding cost of completion damages upon finding that contractor had attempted to evade his contractual obligations); Robert A. Hillman, Contract Lore, 27 J. Corp. L. 505, 509 (2002) (“[I]n construction contracts, the degree of willfulness of a contractor’s breach helps courts determine whether to grant expectancy damages measured by the cost of repair or the diminution in value caused by the breach, the latter often a smaller measure.”). Such breaches are likely to be hard to detect because it is easy for builders to use substitute materials surreptitiously. 177 See Thel & Siegelman, supra note 160, at 1527. Likewise, insured persons can sue in tort to recover more than contract damages if their insurers intentionally deny them benefits without reasonable grounds. 178 See Anderson v. Cont’l Ins., 271 N.W.2d 368, 371 (Wis. 1978); Bar-Gill & Ben-Shahar, Information Theory, supra note 160, at 1496. Given the asymmetry of sophistication between an insurance company and many of its customers and the costs and uncertainty associated with bringing suit, the insurance company is likely to get away with many such denials, making supercompensatory damages warranted from an efficiency standpoint. 179 See Bar-Gill & Ben-Shahar, Information Theory, supra note 160, at 1496–98 (arguing that punitive damages may be appropriate in the insurance context because breach may be hard for consumers to detect).
But the standard economic explanation for supercompensatory damages does not explain why courts sometimes make a finding of willfulness an additional requirement for supercompensatory liability. In a world populated entirely by self-interested externalizers, optimal incentives can be provided by holding the promisor strictly liable to pay additional damages for any breach that will tend to be inadequately deterred by expectation damages alone. But the regime we actually see seems to be a fault-based regime in which liability for supercompensatory damages is triggered by a finding that the promisor breached willfully 180 See supra note 176 (citing cases in which damages were awarded upon a finding of willful breach); cf. Kull, supra note 110, at 2044–45 (arguing that the remedy in cases of opportunistic breach should render the breach unprofitable). (though not, of course, a pure fault-based regime, given that every breaching promisor must pay at least expectation damages). Why subject promisors to additional liability only when we judge that their reasons for breaching the contract were opportunistic? 181 There are a number of standard economic reasons why negligence is sometimes preferable to strict liability and vice versa, which may be able to account for this anomalous pocket of fault-based liability. See, e.g., Craswell, Willful Breach, supra note 160, at 1507–15. But these are explanations of why conditioning liability on conduct that falls short of an efficient standard of care can be superior to strict liability. The doctrine of willful breach, by contrast, conditions supercompensatory liability not on conduct as such but rather on a determination that the breach was motivated by opportunistic reasons. And so it is not obvious that the standard economic rationales for the superiority of negligence over strict liability are applicable here.
A simple explanation for such fault-based liability is available, however, once we recognize that the law is speaking to internalizers and externalizers at the same time. Suppose that instead of a fault-based regime the regime were one of strict liability. That is, suppose that the law required a promisor who committed any breach that was likely to be insufficiently deterred by expectation damages to pay supercompensatory damages regardless of her fault. Then it is plausible to suppose that an internalizer would view herself as duty-bound to perform or else pay those supercompensatory damages, 182 See supra note 107 and accompanying text (arguing that strict liability supports a perform-or-pay interpretation of a promisor’s duties). in which case, were she to opt to breach and pay damages, she would feel duty-bound to pay the supercompensatory amount. Unlike a self-interested externalizer, she would not discount the supercompensatory damages by the probability that she might escape liability for breach when deciding whether to perform or breach and pay damages. 183 See supra note 93 and accompanying text (explaining that when enforcement is imperfect a self-interested externalizer will breach too often from an efficiency standpoint because expected damages for breach will fall short of the promisee’s expectation). Accordingly, under such a strict liability regime, she would perform too often from an efficiency standpoint because the damages that she would feel duty-bound to pay would, by definition, exceed the promisee’s valuation of performance. 184 This depends on the assumption that the promisor internalizes the rule rather than the rule’s underlying rationale. An internalizer who is sensitive to the underlying rationale of the rule rather than the rule itself will understand that the rule was designed for self-interested externalizers, not for people like her.
By contrast, by conditioning supercompensatory liability on a finding that the promisor breached the contract for opportunistic reasons, the law conveys to the internalizer the message that her duty is simply to perform or pay ordinary expectation damages. Supercompensatory liability is directed only at those who try to evade this basic obligation to perform or pay ordinary expectation damages. And because an internalizer, by definition, is not motivated to evade her legal duties, she will understand that the threat of supercompensatory damages is not directed at her.
In short, the law’s inquiry into a contract breaker’s motivations enables it to send the right messages to internalizers and externalizers at the same time. 185 Here, I am assuming that internalizers adopt a “perform-or-pay” interpretation of their duties. But the explanation may work in a modified form even if they adopt a “perform-come-what-may” interpretation or, perhaps more realistically, a “perform-unless-there-are-significant-attenuating-circumstances” interpretation. Internalizers who adopt either of the latter interpretations of their duties will be inclined to perform their contracts, regardless of the level of damages. However, awarding higher damages for breaches of contracts in a particular class regardless of fault still sends the message to such internalizers that those breaches are worse than others (breaches that trigger ordinary compensatory damages). And this will alter the behavior of an internalizer who faces conflicting duties (making her inclined to give preference to some kinds of contractual duties over others) or attenuating circumstances (making her less inclined to breach such contracts in the face of such circumstances). Such an explanation is complicated by the fact that if internalizers adopt these more absolute interpretations of their legal duties, then their behavior may be suboptimal from the social welfare maximizing perspective anyway, and so we are already in a second-best world. Still, it is plausible to suppose that imposing strict supercompensatory liability for certain breaches tends to distort such internalizers’ behavior even further away from the social optimum. Supercompensatory liability is imposed only on those who opportunistically try to evade the obligation to perform or pay damages, leading to efficient behavior all round. Internalizers perform their contracts when, and only when, it is efficient for them to do so because they believe themselves to be duty-bound to perform their contracts or fully compensate their promisees by paying them expectation damages. At the same time, the threat of higher damages deters externalizers from acting opportunistically. In Professor Meir Dan-Cohen’s terms, a fault-based rule achieves “acoustic separation” between internalizers and externalizers. 186 Meir Dan-Cohen, Decision Rules and Conduct Rules: On Acoustic Separation in Criminal Law, 97 Harv. L. Rev. 625, 630–34 (1984). Dan-Cohen uses the term “acoustic separation” to refer to the insulation of decision rules, which guide official decisionmaking, from conduct rules, which guide the behavior of ordinary citizens. The more general idea is that the law might be designed to speak to multiple constituencies at once. For other applications of this general idea, see Camerer et al., Regulation for Conservatives, supra note 26; Feldman & Smith, supra note 26.
E. Duress and Contract Modification
When should a court enforce a modification of the terms of a contract that one party forced another to accept by threatening to breach the original agreement? On the one hand, we might think that social welfare will be increased if courts enforce any modification that both parties have agreed to because parties will not rationally agree to modifications that make them worse off. On the other hand, the fact that the modification worsens the terms of the contract for the threatened party—a buyer, say—raises the concern that he would not have rationally agreed to it unless coerced into doing so by the seller. 187 This is what Professors Varouj Aivazian, Michael Trebilcock, and Michael Penny call the “apparent analytical paradox presented by contract modifications.” Varouj A. Aivazian, Michael J. Trebilcock & Michael Penny, The Law of Contract Modifications: The Uncertain Quest for a Bench Mark of Enforceability, 22 Osgoode Hall L.J. 173, 174 (1984). This suggests that some kind of duress analysis is appropriate: Was the modification entered into voluntarily by both sides? But this simply raises a further question: What does it mean to voluntarily agree to a modification? Since voluntariness is not demonstrated merely by showing that the acquiescing party had his rational capacities intact, 188 Everyone should agree that a modification should be unenforceable if one of the parties lacked the capacity to rationally assent to it. See Restatement (Second) of Contracts § 12(1) (Am. Law Inst. 1981) (“No one can be bound by contract who has not legal capacity to incur at least voidable contractual duties.”). we are forced to ask whether his consent was “induced by pressure ‘which the law does not regard as legitimate.’” 189 Aivazian et al., supra note 187, at 184 (quoting Universe Tankships Inc. of Monrovia v. Int’l Transp. Workers Fed’n  1 AC 366 (HL) 384 (appeal taken from Eng.)). And so the doctrine must tell us what constitutes “legitimate pressure.” 190 Aivazian et al., supra note 187, at 184; see also Selmer Co. v. Blakeslee-Midwest Co., 704 F.2d 924, 927 (7th Cir. 1983) (Posner, J.) (“The fundamental issue in a duress case is therefore not the victim’s state of mind but whether the statement that induced the promise is the kind of offer to deal that we want to discourage, and hence that we call a ‘threat.’”).
Standard economic analysis, which assumes that subjects are all self-interested externalizers, suggests that courts should enforce modifications when doing so will make both parties better off. This means that a modification should be enforceable whenever the party seeking it would in fact breach the original contract if it were not so modified—that is, when her threat to breach is credible—and her contracting partner prefers the modified contract to the breach of the original contract.
The first implication of this view is uncontroversial. A party threatening breach should not be rewarded with an enforceable modification if she would never have carried out her threat to breach. 191 See Oren Bar-Gill & Omri Ben-Shahar, The Law of Duress and the Economics of Credible Threats, 33 J. Legal Stud. 391, 404–05 (2004) [hereinafter Bar-Gill & Ben-Shahar, Law of Duress] (arguing that courts should “verify that the seller had a credible modification demand”); Jason Scott Johnston, Default Rules/Mandatory Principles: A Game Theoretic Analysis of Good Faith and the Contract Modification Problem, 3 S. Cal. Interdisc. L.J. 335, 369 (1993) (arguing that a good-faith inquiry into whether circumstances justified a modification is necessary to dissuade parties from strategically misrepresenting the likelihood that they will breach if an advantageous modification is not procured); Alan Schwartz, Relational Contracts in the Courts: An Analysis of Incomplete Agreements and Judicial Strategies, 21 J. Legal Stud. 271, 310 (1992) (“[T]he performing party will prefer ex ante that a court not enforce a modification when no new information is revealed ex post: she has no incentive to exploit under this rule, and, so, the rule functions as her commitment to behave in good faith.”); Daniel A. Graham & Ellen R. Peirce, Contract Modification: An Economic Analysis of the Hold-Up Game, Law & Contemp. Probs., Winter 1989, at 9, 11 (“The law is . . . disinclined to enforce a modification obtained by [an empty threat]. By hypothesis, such a modification has no effect on the action that the party making the empty threat would have actually taken—the result is . . . a transfer payment from one party to the other not unlike robbery . . . .”). Such a modification produces no net gain to the parties and simply redistributes the contractual surplus away from the victim of the threat to the threatening party. Ex ante, both parties would like to be able to commit to refrain from making such incredible threats so that the bargains they make are not unraveled by this kind of opportunism.
Consistent with this analysis, courts generally will not enforce a modification procured by a threat to breach that was nothing more than a party’s strategic ploy to exploit her contracting partner’s uncertainty about her behavior. While a threat to breach is not, of itself, improper, courts care about the threatening party’s reasons for making the threat. 192 See Farnsworth, supra note 86, § 4.17, at 259. The UCC makes “good faith” the general test for enforceability, which “may in some situations require an objectively demonstrable reason for seeking a modification,” such as a “market shift which makes performance come to involve a loss . . . even though there is no unforeseen difficulty as would make out a legal excuse from performance.” U.C.C. § 2-209 cmt. 2 (Am. Law Inst. & Unif. Law Comm’n 1990); see also T & S Brass & Bronze Works, Inc. v. Pic-Air, Inc., 790 F.2d 1098, 1105 (4th Cir. 1986) (applying the UCC’s “good faith” requirement to find that a delay in production was not a “legitimate commercial reason” for a requested modification); Roth Steel Prods. v. Sharon Steel Corp., 705 F.2d 134, 146 (6th Cir. 1983) (stating that courts should ask “whether the parties were in fact motivated to seek modification by an honest desire to compensate for commercial exigencies”). Likewise, the Restatement asks courts to consider whether “the modification is fair and equitable in view of circumstances not anticipated by the parties when the contract was made.” Restatement (Second) of Contracts § 89(a). And so they look for evidence that the modification was sought for a “legitimate commercial reason” grounded in changed circumstances, at least so long as those changed circumstances were outside of the control of the threatening party. 193 See T & S Brass & Bronze Works, 790 F.2d at 1105 (“Courts have determined that a legitimate commercial reason is one outside the control of the party seeking the modification.”); Aivazian et al., supra note 187, at 194 (asserting that “to allow risk reallocation through modification away from the superior risk bearer attenuates incentives to take efficient risk reduction or risk insurance precautions”); Bar-Gill & Ben-Shahar, Law of Duress, supra note 191, at 415–17 (suggesting that courts should refuse to enforce modifications induced by credible threats to breach when the threatening party manipulated the credibility of her threat); Schwartz, supra note 191, at 310 n.88 (“There is concern that freely enforcing modifications may create moral hazard. A performing party will take insufficient precautions against factors that can make performance more costly because she can exact a modification if such factors materialize.”). But the doctrine diverges from the standard economic account after this point.
Standard economic analysis makes the inadequacy of the threatened party’s legal remedies a reason that counts in favor of enforcing the modification. This is because the threatened party will prefer a modification that results in a worsening of the terms of the original contract to a breach only if his remedies would fail to adequately compensate him if his contracting partner were to breach. If the threatened party would receive full expectation damages were he to sue his contracting partner for breach of the original contract, then he would prefer to do that than to accede to the worsened terms. Conversely, the threatening party is less likely to have a credible threat to breach if the threatened party’s remedies are adequate. In short, in order for both parties to have something to gain from a modification, remedies must be inadequate. 194 Thus, Professor Alan Schwartz argues that courts should enforce modifications only if the defendant (the threatened party) would be cut off from the market. This is because in the absence of a freely available market substitute, he would find it hard to prove his damages (lost profits) in court (i.e., his remedies will be inadequate). Schwartz, supra note 191, at 309, 312.
Consider a simple example. A seller agrees to provide a buyer with a service that is worth V to the buyer at price P1, but before the seller performs, her costs of performance rise unexpectedly from C1 to C2. Remedies are inadequate. That is, the damages, D, that the seller will be forced to pay in the event that she breaches will fail to give the buyer his full expectation: D < V – P1. If the seller is a self-interested externalizer, she will breach in the absence of a modification if P1 – C2 < -D—that is, if the profits she would make from performing fall short of the losses she would incur (in the form of damages that she would be forced to pay to the buyer) if she breaches. The buyer will prefer performance of the contract at a higher price, P2, to breach if V – P2 > D—that is, if the net value he would receive from performance at the higher price exceeds the damages he would receive if the seller breaches the contract. This suggests that so long as D < C2 – P1, rendering the seller’s threat to breach the original contract credible, a modified contract with a new price in the range P1 < C2 – D < P2 < V – D should be enforced, since doing so will make both parties better off. The seller will be better off because she will receive a higher price than she would have received under the original contract such that she will prefer to perform rather than breach. The buyer will be better off because performance at the higher price is better for him than breach, and breach would result were the modification not enforced because the seller has a credible threat to breach.
This is not to say that the simple economic account necessarily requires a separate analysis of the adequacy of available remedies. The credibility rule that is proposed by Professors Oren Bar-Gill and Omri Ben-Shahar makes the credibility of the threat to breach a necessary and sufficient condition for enforceability. 195 Bar-Gill & Ben-Shahar, Law of Duress, supra note 191, at 391–417; see also Oren Bar-Gill & Omri Ben-Shahar, Credible Coercion, 83 Tex. L. Rev. 717, 721 (2005). But this credibility rule exploits gains from trade that exist only if remedies are inadequate: The more inadequate are the threatened party’s remedies, the more likely it is that the threatening party has a credible threat to breach, and the more the threatened party has to gain from an enforceable modification that would prevent breach. 196 Bar-Gill & Ben-Shahar, Law of Duress, supra note 191, at 398–99.
The cases, however, take the opposite position, viewing a threatened party’s inadequate legal remedies as grounds not to enforce a modification. 197 Farnsworth, supra note 86, § 4.18, at 262; see also Rissman v. Rissman, 213 F.3d 381, 387 (7th Cir. 2000) (Easterbrook, J.) (discussing cases in which a finding of duress hinges on the feasibility of litigation as an alternative to acquiescence); Selmer Co. v. Blakeslee-Midwest Co., 704 F.2d 924, 927 (7th Cir. 1983) (Posner, J.) (“It undermines the institution of contract to allow a contract party to use the threat of breach to get the contract modified in his favor . . . because the other party, unless he knuckles under to the threat, will incur costs for which he will have no adequate legal remedy.”); Gross Valentino Printing Co. v. Clarke, 458 N.E.2d 1027, 1031 (Ill. App. Ct. 1983) (“To show compulsion, one must indicate how legal redress would be inadequate.”); Rose v. Vulcan Materials Co., 194 S.E.2d 521, 536 (N.C. 1973) (finding economic duress when the plaintiff had no practical alternative source for the stone he had contracted for and the only alternative to continuing to do business with the defendant was to go out of business); Tri-State Roofing Co. of Uniontown v. Simon, 142 A.2d 333, 336 (Pa. Super. Ct. 1958) (concluding there was no duress where the remedy at law appeared to be “entirely adequate to save defendant from irreparable damage”); Restatement (Second) of Contracts § 175(1) (“If a party’s manifestation of assent is induced by an improper threat by the other party that leaves the victim no reasonable alternative, the contract is voidable by the victim.”). Commentators have concluded that this feature of the doctrine cannot be easily explained using standard economic analysis. 198 Bar-Gill & Ben-Shahar, Law of Duress, supra note 191, at 399 (“If a court were to use the inadequacy of remedies as grounds not to enforce the modification, it would undermine the buyer’s preference and force her to settle for inadequate remedies . . . . Unfortunately . . . this is precisely what the majority of courts do when applying the doctrine of duress in modification cases.”); see also Schwartz, supra note 191, at 308–13 (arguing that modifications should be enforced when the acquiescing party is cut off from the market and so remedies are likely to be inadequate).
We might be able to rationalize what courts do by enriching the standard economic model while preserving its central motivational assumption that everyone is a self-interested externalizer. For example, refusing to enforce modifications in the face of the acquiescing party’s inadequate remedies eliminates the other party’s incentive to make incredible threats. 199 Kevin Davis, The Demand for Immutable Contracts: Another Look at the Law and Economics of Contract Modifications, 81 N.Y.U. L. Rev. 487, 496 (2006). But courts could eliminate this incentive simply by refusing to enforce incredible threats, at least so long as the costs of litigating the question of the credibility of the threat are not too high. 200 See Bar-Gill & Ben-Shahar, Law of Duress, supra note 191, at 405–09. Refusing to enforce modifications in the face of inadequate remedies also eliminates a seller’s incentive to make lowball offers, knowing all along that it will never make sense for her to perform at such a low price. 201 Davis, supra note 199, at 497. But this justification of refusals to enforce such modifications only works if the buyer is ignorant of the seller’s cost at the time the contract is made.
An alternative way to justify the law’s emphasis on the threatened party’s inadequate remedies is to assume that many subjects are internalizers rather than externalizers. Let us assume for simplicity that many subjects are self-interested internalizers. That is, they act in a self-interested fashion subject to complying with the rules. This is not an unrealistic behavioral assumption. It describes the behavior of those who internalize legal rules for self-interested reasons. 202 See supra notes 77–80 and accompanying text (explaining that even a purely self-interested agent may internalize legal rules for epistemic reasons). It also describes the behavior of morally motivated epistemic internalizers who are subject to the crowding out described earlier and so behave in a self-interested fashion subject to conforming to the rules. 203 See supra section III.B (describing norm-based crowding out). Suppose further that these internalizers believe that they have a legal duty to perform or breach and pay expectation damages. 204 See supra note 107 and accompanying text (arguing that strict liability supports a perform-or-pay interpretation of a promisor’s duties). As we have already seen, such internalizers will breach only when it is efficient to do so, even if remedies are inadequate (so that a self-interested externalizer will breach too often from an efficiency standpoint). This is because such an internalizer will be motivated to pay the promisee his full expectation in the event of breach (even if D is less than the full expectation). 205 See supra note 101 and accompanying text (explaining that internalizers are motivated to pay the promisee the full expectation because that is what they are legally duty-bound to do in the event of breach). Thus, she will only ever “threaten” to breach if breach would be the efficient outcome. 206 It is misleading to call this a “threat” insofar as it would not be interpreted as such by an internalizer who thinks that promisors have the option to perform or breach and pay expectation damages.
Now suppose that, as the standard economic model recommends, the rule is that contract modifications will be enforced so long as P1 < C2 – D < P2 < V – D where D is less than the buyer’s expectation (because, by assumption, remedies are inadequate). Such a rule rewards a seller who threatens breach with an enforceable modification whenever it would be in her rational self-interest to breach without such a modification. Thus, it communicates to internalizers that it is permissible to leave their contracting partners undercompensated even when damages are inadequate. This contradicts the message that is conveyed by the ordinary rule of expectation damages, according to which a promisor should fully compensate the promisee for the promisee’s lost expectation. It tells internalizers, who would otherwise be willing to perform or pay full expectation damages regardless of whether or not legal remedies are adequate, that it is permissible for them to instead exploit their contracting partner’s inadequate remedies in order to extract better terms for themselves. 207 If an internalizer’s default interpretation of her duties is that she must “perform unless released by the promisee” rather than “perform or pay,” the economist’s rule will also undermine that interpretation of her duties, by encouraging her to believe that she can opportunistically (threaten) breach. By contrast, a rule that makes the buyer’s inadequate remedies a reason not to enforce a modification that was procured by a threat of breach reinforces the basic principle that the buyer is entitled to his expectation in the event of breach and so does not encourage such self-interested behavior on the part of internalizers.
What about Bar-Gill and Ben-Shahar’s proposal? Their credibility rule stipulates that modifications should be enforced if and only if the seller’s threat to breach is credible (so long as the circumstances that made her threat credible were outside her control). 208 See supra notes 195–196 and accompanying text. Thus, their rule does not actually assume that parties are self-interested. 209 In a companion paper, they consider what happens when the party threatening the breach has a preference for fairness that makes her more reluctant than a purely self-interested agent to perform what she perceives to be an unfair contract and therefore render an otherwise incredible threat to breach credible. Oren Bar-Gill & Omri Ben-Shahar, Threatening an “Irrational” Breach of Contract, 11 Sup. Ct. Econ. Rev. 143 (2004). It does, however, implicitly assume that parties are externalizers—that their preferences are independent of the legal rule—for it directs courts to consider what parties would in fact do in the absence of a modification given their (assumedly fixed) preferences. Instead of enforcing a modification whenever a self-interested seller would breach the contract in the absence of a modification (i.e., whenever D < C2 – P1), the credibility rule enforces a modification whenever the seller would breach the contract in the absence of a modification given her exogenous preferences.
The problem with the credibility rule is that it provides indeterminate guidance to subjects whose preferences are not exogenous because they depend directly on the legal rules. If the seller is an internalizer, she will look to the legal rule to determine what she ought to do. Under the credibility rule, she can procure an enforceable modification through the threat of breach whenever she would in fact breach in the absence of a modification, even if the buyer would be inadequately compensated as a result. But what she would in fact do in this scenario depends on her preferences, which, because she is an internalizer, depend on the legal rule. We are caught in a circle: The credibility rule tells her nothing more than that she may do what she would in fact do. Thus, multiple outcomes are possible. 210 For a more formal analysis of this indeterminacy, see infra Appendix. It follows that even if we preserve the credibility of the threat as a necessary condition for enforcement of a contract modification, the law may be able to improve social welfare, given the presence of internalizers, by enforcing a substantive criterion of its own rather than simply deferring to the threatening party’s preferences.
Perhaps, however, the credibility rule is not indeterminate for a subject who internalizes it together with the rule that she must perform or pay expectation damages. Since an internalizer of the latter rule is prepared to pay her contracting party’s expectation when she breaches a contract, surely she would only credibly threaten breach when it would be efficient to do so? Unfortunately this is not necessarily the case. This is because the credibility rule, like the simpler rule derived from the assumption of self-interest, 211 See supra notes 207–209 and accompanying text (describing the simple rule and the credibility rule). undermines the message that is otherwise conveyed by the rule of expectation damages by rewarding a promisor who credibly threatens breach with an enforceable modification whatever her reasons for breaching. Thus, it rewards credible threats to breach inefficiently for self-interested reasons as well as credible “threats” to efficiently breach. Accordingly, an internalizing subject who thinks that she would like to breach for self-interested reasons if the contract were not modified can appeal to the credibility rule in order to justify doing so.
Ultimately, an internalizer’s behavior will depend on how she interprets her duties in the light of the credibility rule. The internalizer might infer that, in the light of her duty to perform or pay full expectation damages, not all enforceable credible threats are legitimate and thus condoned by the legal regime. But this interpretative move is strained given that, pursuant to the credibility rule, all credible threats to breach will be enforced. In particular, it implausibly supposes that the law adopts a more permissive attitude toward those who are not prepared to follow all of its edicts—that is, toward those who do not internalize the rule that a promisor must perform or pay expectation damages—than toward those who are prepared to follow them. The law would in effect be saying: Do not threaten an inefficient breach, but if you do, and your threat is credible, and the other party acquiesces in your threat, we will treat the resulting contract as perfectly legitimate. Such a mixed message would fail to clearly communicate to an internalizer that she has a duty not to threaten inefficient breaches.
Thus, in the presence of internalizers, courts that are interested in maximizing social welfare should take a party’s lack of adequate remedies as a reason not to enforce a modification procured by a threat of breach, which is what they in fact do. 212 See supra note 197 and accompanying text (explaining that courts view a threatened party’s inadequate legal remedies as a reason to not enforce a modification). Indeed, language in judicial opinions suggests that judges may have internalizers in mind when formulating the rules. As the Texas Court of Appeals once put it: “To stamp this [contract modification] with legal approval would violate, in our judgment, every principle of honesty and fair dealing, and place a premium upon repudiation of binding legal obligations, where to do so would be to the pecuniary advantage of the party bound.” 213 Barreda v. Craig, Thompson & Jeffries, 222 S.W. 177, 179 (Tex. App. 1920).
Courts enforce contract modifications when such modifications appear to have been motivated by the right reasons. Modifications that respond to changed circumstances that were unanticipated by the parties at the time of contracting are enforced, 214 See supra notes 192–193 and accompanying text. while those that appear to have been prompted by one party’s desire to exploit another party’s inadequate remedies are not. 215 See supra notes 197–198 and accompanying text. Distinguishing between the two kinds of modification is not always a straightforward matter. When changed circumstances and inadequate remedies are both present, it may be difficult to discern what was motivating the threatening party. But in principle, conditioning enforcement on a finding that the threatening party was responding to unanticipated changed circumstances rather than threatening an opportunistic breach is sensible when parties are internalizers.
The foregoing analysis shows that externalizers and internalizers respond to the different possible rules of contract modification in different ways and suggests that the law of contract modification may be well designed for internalizers but not for externalizers. It is, however, not obvious how contract-modification doctrine could be designed to induce both internalizers and externalizers to behave optimally. If the law attempted to create a different set of criteria for enforcing modifications procured by externalizers, it would reward the behavior of those willing to self-interestedly threaten inefficient breaches. And rewarding such self-interested behavior would likely disrupt the message that the law sends to internalizers, for internalizers would reasonably conclude that the law condones self-interested threats to breach. Whereas the doctrine of willful breach appeals to externalizers and internalizers simultaneously by singling out those with bad motives for harsher treatment, the doctrine of contract modification must choose between giving good guidance to internalizers and the right incentives to externalizers. In other words, “acoustic separation” does not seem possible in this context. 216 See supra note 186 and accompanying text (defining “acoustic separation”).
This Article has considered several ways in which refining our conception of the legal subject can alter the economic analysis of contract law. Whereas traditional economic analysis assumes a population consisting exclusively of self-interested actors who view legal rules as simply altering the relative prices of actions, it is more realistic to suppose that a significant proportion of the population views legal rules as laying down standards of behavior that should guide them, regardless of the consequences of defiance.
My analysis has shown that internalizers’ responses to legal rules differ significantly from those of externalizers, such that the economic analysis of contract law changes when we assume that there are internalizers in the subject population. In particular, while some rules can be designed to induce both internalizers and externalizers to take social welfare maximizing actions, as the analysis of the doctrine of willful breach shows, sometimes, as in the contract modification setting, rules that would be optimal for externalizers diverge from those that would be optimal for internalizers, so that the lawmaker will be forced to navigate complex tradeoffs.
As we saw in section III.E, if the seller has a credible threat to breach, then the buyer will agree to a modification with a price in the range P2 < V – D. If the seller is an internalizer, she has a preference to act in accordance with the legal rule. But if the legal rule is the credibility rule, then what she is permitted to do depends on what she will in fact do, which depends on her preferences.
We can illustrate this circularity by modeling the seller’s plight as an intrapersonal game—a game between multiple incarnations (or “selves”) of the same agent. The self that forms the prediction about her own future behavior plays a game with the self that actually makes the decision. At stage 1, the seller forms a conjecture about her own behavior, thus fixing her preferences, in the light of the credibility rule. At stage 2, she makes a choice to maximize those preferences. The agent is in “personal equilibrium” when her conjecture correctly predicts her subsequent behavior, and her behavior maximizes her preferences given her conjecture. 217 See Botond Koszegi & Matthew Rabin, A Model of Reference-Dependent Preferences, 121 Q.J. Econ. 1133 (2006) (using the idea of personal equilibrium to derive implications of loss aversion).
Suppose that the seller predicts that she will breach the contract in the event that the buyer rejects the modification. The credibility rule tells her that it is permissible for her to breach, notwithstanding the fact that the buyer will be inadequately compensated, so long as she would in fact do so. This is because the law rewards the seller for her threat by enforcing her modification solely because she would in fact breach the contract as she threatened to. Thus, it disrupts the message that is conveyed by the rule that she must perform or pay expectation damages. So the seller reasons that the law is indifferent between performance and breach plus payment of the inadequate remedies and that therefore she may use her nonlegal preferences to break the tie. Assuming that she is a self-interested internalizer, the seller will in fact breach as she predicted whenever:
P1 – C2 + D < 0. 
Thus, there is an intrapersonal equilibrium (her actual behavior conforms to her predicted behavior) in which she breaches whenever  holds.
Now suppose that the seller believes that she will not breach the contract. Then the credibility rule does not give its stamp of approval to an inefficient breach of the contract. So we can assume that the seller thinks that her ordinary legal duty to perform or breach and pay full expectation damages remains intact. Therefore, she will not “threaten” a breach to secure a modification so long as performance is efficient:
V – C2 > 0. 
Thus, whenever  holds, there is an equilibrium in which the seller performs and therefore does not threaten breach to procure a modification.
Accordingly, if both  and  hold (i.e., P1 + D < C2 < V), then the seller’s behavior is unpredictable if the operative legal rule is the credibility rule: Either she has a credible threat to breach, justifying an enforceable modification, or she does not and therefore would not be able to get a modification enforced, depending on her own, self-confirming beliefs about her behavior.