25 of 25 DOCUMENTS
Copyright ? 1983 Yale Law Journal Company.
Yale Law Journal
92 Yale L.J. 763
LENGTH: 17784 words
ARTICLE: Paternalism and the Law of Contracts.
NAME: Anthony T. Kronman +
+ Professor of Law, Yale University. An earlier version of this Article was presented at a meeting of the
Society for Ethical and Legal Philosophy; I am indebted to those who were present for their comments and
criticisms. Bruce Ackerman, Jerry Mashaw, Rob Prichard, and the Honorable Ellen Peters also made helpful
suggestions. Peter Swire provided valuable research assistance.
... The prohibition against suicide, the requirement that motorcyclists wear helmets, laws that restrict the use of drugs or make education compulsory are all examples of legal paternalism. ... The first of these two arguments -- the argument from economic efficiency -- rests on the view that the prohibition against waiver of the implied warranty of habitability is, in effect, an efficiency-enhancing adjunct to the fraud remedy a disappointed tenant would have were the warranty freely disclaimable. ... The relevant question, therefore, is not whether the decision to use a nondisclaimable warranty of habitability as an instrument of distributive justice can be frustrated by the private contractual reallocation of its burdens, but whether the general problem of pass-along is more serious here than it would be, say, in a scheme that attempts to achieve the same goal by taxing landlords to subsidize low-income tenants. ... The weaker a person's imagination, the less likely he is even to consider the possibility that the desires motivating him to act may one day seem antithetical to his deepest interests, making it difficult for him to understand why he acted as he did and impossible to forgive himself for doing so; a person with a weak imagination is therefore especially susceptible to regret and its corrosive effects. ...
[*763] Our legal system, like every other, limits the power of individuals to enlist the state in the enforcement of their private agreements. n1 In a broad sense, all limitations of this sort restrict the contractual freedom of those involved by depriving them of the right to decide whether their voluntary arrangements shall be legally binding. Many of these limitations are intended to protect the interests of third parties, including the general interests of society at large:
Two neighbors cannot make an enforceable contract to rob a third, nor can a group of businessmen negotiate a price-fixing agreement that will be binding as a matter of law. Other restraints on contractual freedom, however, are primarily intended to protect those whose freedom they restrict. Restraints of this sort aim to protect people from themselves by limiting their capacity to make enforceable agreements of various kinds.
92 Yale L.J. 763, *
In general, any legal rule that prohibits an action on the ground that it would be contrary to the actor's own welfare is pateralistic. The prohibition against suicide, n2 the requirement that motorcyclists wear helmets, n3 laws that restrict the use of drugs n4 or make education compulsory n5 are all [*764] examples of legal paternalism. In this Article, I shall be concerned with one branch of this wide and heterogeneous family of legal rules -- those that may properly be regarded as belonging to the law of contracts because the liberty they restrict is the liberty to bind oneself by making a legally enforceable promise.
One (relatively new) example of the kind of restriction I have in mind is the rule invalidating any provision in a residential lease that purports to waive the tenant's right to withhold rent if the property fails to meet certain minimum standards of habitability. n6 Even if a tenant believes that a waiver of this sort would be in his own interest and voluntarily agrees to include it in the lease, the law protects him by refusing to enforce his waiver. The invalidity of contracts of peonage or self-enslavement, n7 of agreements purporting to waive the promisor's right to obtain a divorce n8 or sue for relief under the bankruptcy laws, n9 of provisions conferring on either party a right to specifically enforce their agreement (where no right of this sort exists as a matter of law); n10 the voidability of most contracts made by infants; n11 and the nonwaiveable "cooling-off" period imposed by law in many consumer transactions n12 all also have, at least in part, a paternalistic objective. Unquestionably, some of these limitations on the enforceability of private agreements also seek to protect the moral and economic interest of third parties and, to this extent, have a nonpaternalistic objective as well. One central purpose of each, however, is to protect the promisor himself by limiting his power to do what the law judges to be against his own interests; this is paternalism, and there is more of it in our law of contracts than one might suspect.
All paternalistic restrictions on conduct, including those contained in the law of contracts, raise special problems from a moral point of view. It is possible, of course, to deny that any such restrictions exist (by maintaining that every apparent example of paternalism is in reality a disguised prohibition against conduct that violates the rights of other persons). n13 But [*765] by acknowledging that it is ever morally permissible to prevent a person from acting solely because he himself will be harmed by the action, one embraces paternalism (in however limited a form) and has an obligation to explain why such interference is permissible in some instances but not in others. Everyone who has written on the subject of paternalism has wrestled with this problem. n14 Mill, for example, considered the prohibition against self-enslavement paternalistic, but made a considerable effort to demonstrate its moral legitimacy. n15 One who believes, as Mill did and I do, that some paternalistic restrictions on contractual freedom are not only permissible but morally required, must supply a standard or principle for evaluating paternalistic arguments in particular cases; only in this way can the legitimacy of paternalism be established and its limits defined.
It would be a mistake, however, to assume that there is a single principle that best explains every paternalistic restriction in our law of contracts. There is considerable diversity among these restrictions, and while it is true that all of them seek to protect the promisor against the damaging consequences of his own agreements, they do so in different ways and for different reasons. Some paternalistic limitations on contractual freedom are best explained by considerations of economic efficiency and distributive fairness, others by the idea of personal integrity, and a third set of limitations by the familiar, though poorly understood, notion of sound judgment. None of these explanations is exclusive, and there is considerable overlap among them, but each provides the most plausible justification for one particular group or class of paternalistic restrictions in our law of contracts. Although these three classes are connected in important ways, the differences among them are real and worth emphasizing.
The principle (or principles) on which each class of restrictions rests establishes a different framework of analysis and criticism, and in this Article I shall attempt to clarify the basic differences among them. More is involved here, however, than the presentation of a classificatory scheme. The variety of paternalistic restrictions in our law of contracts reveals a complexity in its intellectual premises that, though rarely acknowledged, is one of the law's most distinctive features. I hope to bring this significant, but forgotten, complexity more clearly into view.
[*766] I. Economic Efficiency and Distributive Justice
In many jurisdictions, a nondisclaimable warranty of habitability is now implied, as a matter of law, in every lease of residential property. n16 Because the warranty is nondisclaimable, any agreement a tenant makes to waive its benefits will be unenforceable. It is sometimes said that if tenants were given the power to waive the warranty of habitability they might be tricked or forced into doing so -- to their own disadvantage -- and that the warranty has been made nondisclaimable in order to protect tenants from themselves. n17
92 Yale L.J. 763, *
To the extent the rule barring free waiver of the warranty of habitability is acknowledged to have a paternalistic aim, its justification is a mixed one, turning in part on considerations of efficiency and in part on a conception of distributive justice. The first of these two arguments -- the argument from economic efficiency -- rests on the view that the prohibition against waiver of the implied warranty of habitability is, in effect, an efficiency-enhancing adjunct to the fraud remedy a disappointed tenant would have were the warranty freely disclaimable. Making an entitlement inalienable is a draconian, but sometimes efficient, way to protect its possessor against fraud or deception, and the decision to make the warranty of habitability nondisclaimable can be defended on precisely these grounds. At first glance, this view may seem implausible since considerations of efficiency are typically invoked to attack, rather than defend, nondisclaimable warranties. n18 It will be helpful, therefore, to begin by recalling why nondisclaimability is generally thought to be economically objectionable.
In many transactions, various standard terms are implied as a matter of law. These legally-implied provisions establish a set of ready-made contract terms, and whenever the parties would have included similar provisions in their agreement, they are made better off by being spared the time and expense of having to do so. n19 If in most transactions of a particular sort the parties prefer the standard terms the law provides, the reduction in their transaction costs is likely to exceed the increased costs incurred by the few who would prefer other implied terms and must contract out of the regime to which they will otherwise be legally subject. [*767] The legal implication of standard terms is therefore perfectly consistent with the goal of promoting efficiency by reducing transaction costs. What appears indefensible from an economic point of view is the decision to make a standard term (like the warranty of habitability) nondisclaimable, for this can make a difference only where the parties would agree to waive the warranty if they could, and in every case of this sort, the prohibition against waiver seems to reduce the parties' welfare. Such a prohibition might therefore be thought justifiable only on noneconomic grounds.
The inefficiency of nondisclaimable warranties, however, is not as obvious as this overly simple argument suggests. If we regard the warranty of habitability as a device for allocating the risk that an undetected condition will render the premises uninhabitable, it is indeed difficult to understand why, from a strictly economic point of view, the warranty should be made nondisclaimable. The warranty is a form of insurance and the tenant's decision to solicit or refuse its inclusion in the lease will depend, in theory at least, on which of the parties is able, at least expense, to insure against or
take steps to prevent the feared event. n20 This assumes, however, that when he signs the contract, the landlord does not know of any latent condition that will make the property uninhabitable. If he knows that such a condition exists, but denies that it does or fails to disclose it, the tenant's agreement to waive the warranty of habitability can hardly be characterized as an efficient allocation of risk based upon his preference for self-insurance. If the landlord has lied to the tenant, there is no economic justification for enforcing their bargain; doing so would only give others an incentive to spend more on fraud protection (a deadweight loss from society's point of view). n21 It is no answer to say that the tenant "takes the risk" of his landlord's fraud when he agrees to waive the warranty, for this would be like saying that I take the risk of being forced to sign a contract at gunpoint when I go out unarmed and should therefore be held to any promises extorted in this way. No economist would say that it is efficient to enforce an agreement of the latter sort, and the reasons for refusing to do so are just as strong when the agreement has been procured through deliberate misrepresentation. Even if the landlord has done nothing more than fail to disclose his knowledge of the latent condition, there is no economic justification for enforcing the lease, unless his knowledge is the fruit of a deliberate and socially productive search (which seems to me unlikely for reasons I have elaborated elsewhere). n22
[*768] Of course, when the law implies a warranty, but permits the parties to vary or waive it at their discretion, a disappointed promisor has the right to rescind the contract and recover damages if his agreement was induced by misrepresentation or fraudulent concealment. n23 A disclaimable warranty, therefore, is always accompanied by a general remedy that protects the intended beneficiary of the warranty in the event he agrees to waive it after having been defrauded by the other party. The protection this general remedy affords may be inadequate, however, since claims of fraud are often difficult to prove. n24 To establish such a claim, the victim must show that he was intentionally deceived, n25 and proof of the wrongdoer's mental state can be difficult. In theory, a tenant will take these proof problems into account in deciding whether to insist on a warranty of habitability from his landlord, but those who are unfamiliar with the legal system and do not have the benefit of professional counsel may underestimate the magnitude of these problems or be largely unaware of their existence.
When a fraud has been committed, but cannot be proven, the agreement will be enforced. This is an inefficient result. Of course, if this happens only rarely -- if fraud can be established in almost every case in which it occurs -- the inefficiency of enforcing a few fraudulent bargains may be justified by some other, more desirable, consequence of a strict proof system (for example, its tendency to discourage vexatious lawsuits brought only to harass or blackmail the
92 Yale L.J. 763, *
defendant). But if a large number of fraudulent bargains are enforced, the efficiencies of a strict proof system may be outweighed by its inefficiencies. One way of remedying this situation, of course, is to lower the proof requirements. But if fraud is widespread, if it can be concealed with sufficient ease, and if the victims of fraud typically lack the resources to prosecute their legal claims, lowering the proof requirement may not be enough. A more radical solution -- but one that is nevertheless justifiable from an economic point of view, at least under certain conditions -- is to give the victims an inalienable entitlement they cannot waive and therefore cannot be fraudulently induced to abandon. This solution has the obvious defect of preventing the parties from modifying the entitlement, even when it would be efficient for them to do so and neither has practiced fraud upon the other. But if the entitlement can be waived, if most of the waivers that are given are procured through fraud, and if the fraud can rarely be proven, the inefficiencies of a nowaiver rule may be outweighed by the Greater inefficiency of enforcing too many fraudulent bargains.
[*769] The argument I have just sketched is not a defense of the implied warranty of habitability in particular, but only an outline of the general form such a defense might take. For the argument to have merit in this particular case, there must be some basis for believing that fraudulent deception is more likely and less easily provable here than in general. One factor that may make the danger of unprovable fraud especially great in certain situations, including this one, is the existence of a significant asymmetry in the parties' access to relevant information. Whether there is a latent condition that will render the premises uninhabitable often cannot be determined by a brief inspection, even if it is conducted with care; typically, conditions of this sort come to light only after a period of occupancy. In this respect, nonhabitability resembles other defects that appear only with use. n26 Even if a landlord does not himself occupy the premises he is renting, he is likely to have the benefit of whatever information previous tenants have acquired and there are a variety of ways he can prevent them from disclosing this information to prospective renters, who may find it difficult to acquire such information on their own. n27 In addition, the landlord's superior knowledge and continuing access to the property increase his ability to conceal latent defects in ways that are not obvious and, therefore, not easily demonstrable in a subsequent lawsuit. For these reasons, it is arguable that the landlord's informational advantage increases the likelihood of unprovable fraud in just those cases in which he has persuaded his tenant to relinquish the protection afforded by a warranty of habitability; if so, it may be economically efficient to make the warranty nondisclaimable. n28
[*770] Since the same information that reveals the existence of a latent defect often suggests a method for concealing it, informational asymmetries not only increase the likelhood of fraud but also make such fraud more difficult to prove. The case for implying a nondisclaimable warranty is therefore stronger when such asymmetries exist. But this is only one factor bearing upon the frequency and legal proof of fraudulent deception. The relative wealth and sophistication of the parties is another. The more a person is willing to spend on concealment, the harder he can make it to establish that the concealment was fraudulent; the most subtle deceptions are often distinguished by their seeming innocence.
Undoubtedly, other factors bear on the question of fraud, and their complete description would represent an important step toward an economic theory of fraud, a subject curiously neglected in the law and economics literature. n29 The formulation of a comprehensive theory of this sort is a task beyond the scope of this Article; until we have such a theory, however, any conclusion regarding the efficiency or inefficiency of nondisclaimable warranties must remain tentative.
A second and quite different justification for refusing to permit the voluntary waiver of certain warranties, such as the warranty of habitability, follows from the idea that in some circumstances a prohibition of this sort may be an essential part of a program of distributive justice. Many, including myself, have defended the view that private law norms may legitimately be used to redistribute wealth when alternative methods of doing so are likely to be more costly or intrusive. n30 The distributive justification for making certain contractual entitlements inalienable by prohibiting their waiver is simply an extension of this idea and is implicit in the familiar and widely accepted notion of an adhesive contract. n31
Where there is a striking imbalance in the bargaining power of the parties to a contract, so that one is able to dictate terms to the other -- to insist that the exchange be on his terms or not at all -- the contract is said to be one of adhesion.
Consumer contracts are often characterized as adhesive, [*771] since the consumer has little or no control over the terms of the agreement. In recent years, courts and legislatures have intervened in the exchange process, with increasing frequency, to correct the imbalance of bargaining power that contracts of this sort appear to involve. Typically, the first step has been the judicial or statutory implication of warranty terms that increase the consumer's rights under the contract, giving him what he wants but has no power to demand. n32 But so long as the party with the greater bargaining power can force the other to waive whatever liability these implied terms create, he can easily
92 Yale L.J. 763, *
restore the original imbalance the warranty is meant to correct. At this point, a court or legislature determined to achieve greater equality in bargaining power may be tempted to make the implied warranty nondisclaimable. n33
The attack on contracts of adhesion rests upon an unstated conception of distributive fairness; though often overlooked, it is this conception that gives the attack its appeal. Many contracts are contracts of adhesion in the general sense that one party is able to dictate terms to the other, but this alone does not make an agreement objectionable. Suppose, for example, that may neighbor owns a painting I happen to covet. I offer him $ 5000 for it. He responds, "$ 10,000 and no warranties regarding its authenticity. Take it or leave it." Clearly, the fact that I lack bargaining power and must adhere to the terms he proposes does not by itself justify a judicial or legislative effort to tip the balance in my favor. The imbalance in this case, which stems from the fact that he owns the painting and I do not, is unobjectionable because we do not care how control over the painting is distributed. n34
We feel differently about the distribution of control over society's available housing stock, and inequalities of bargaining power in this context therefore seem a more appropriate target for judicial or legislative attack. The distribution of housing matters more to us than the distribution of paintings: Only the first is likely to seem important from the standpoint of most theories of distributive justice. Those contracts of adhesion that disturb us do so, then, because they reflect an underlying distribution of power or resources that offends our conception of distributive fairness; when distributive concerns are weak or nonexistent, contracts of adhesion [*772] are less troubling and the concept of adhesion itself loses meaning.
It is therefore misleading to describe the nondisclaimable warranty of habitability as simply a device for correcting an imbalance in bargaining power. More accurately, it is an instrument of redistribution that seeks to shift control over housing from one group (landlords) to another (tenants) in a way that furthers the widely shared goal of insuring everyone shelter of at least a minimally decent sort. To achieve this goal, the warranty must be made nondisclaimable, for if it is not, tenants -- poor tenants in particular -- will routinely be required to waive their rights to habitable premises,
thereby restoring whatever distributional inequities exist at the outset.
This argument is subject to two familiar criticisms. First, even if a nondisclaimable warranty of habitability does redistribute wealth or power from landlords to tenants, it need not necessarily improve the overall position of the poor. If enough landlords are themselves poor, the warranty may conceivably have the opposite effect, and it will in any event also work to the benefit of those who can afford housing of better quality. n35 Consequently, from the standpoint of a comprehensive program of distributive justice, which seeks to achieve a fair distribution of wealth in some overall sense, a limited measure of this sort may seem ineffective or even perverse.
This first criticism is unpersuasive. In the United States, wealth and the ownership of land are correlated to a sufficiently high degree that one may be used as an imperfect, but convenient, proxy for the other; n36 a legal rule that redistributes wealth from landlords to tenants is thus likely to result in a more equal distribution of wealth overall. More importantly, partial measures of this sort, which affect only certain groups in society, can play a distinctive and needed function in a comprehensive program of distributive justice. Even if the total wealth of society were distributed more evenly, it would still be necessary to insure that no group retained control of any vitally important resource for which it could extract a monopoly rent. To prevent the accumulation of monopoly power, redistributive schemes must take into account not only how much people own, but what they own as well. Partial measures that affect the distribution of only a single good, like housing, are therefore likely to be a necessary component of any redistributive program.
A second objection to using the nondisclaimable warranty of habitability as an instrument of distributive justice has greater force. Even if landlords [*773] are barred from disclaiming responsibility for the habitability of property they rent, as long as they are not similarly prevented from altering other aspects of their contractual relationship with their tenants, they can easily pass along -- in the form of an increased rent charge -- whatever additional insurance or compliance consts they incur as a result of their expanded warranty liability. n37 To the extent this is true, the tenant must pay for the increased protection the warranty gives him -- whether he wants it or not. Landlords, one could argue, are likely to be unaffected by a nondisclaimable warranty of habitability, since they will not bear its cost; by contrast, tenants will have to purchase a form of compulsory insurance and can only be made worse off as a result.
There are two ways of meeting this criticism. The first is to deny the unstated premise on which it rests -- that the rule in question must be evaluated from the standpoint of the tenants' own preferences. Even if the costs of complying with the warranty are fully passed along, this only means that some tenants will have to pay for protection they do not want, and this is objectionable only in case the wishes of the tenants themselves should be controlling. But there may be nothing wrong with forcing tenants, including poor tenants, to spend their money on better housing (or more exactly,
92 Yale L.J. 763, *
on insurance against the risk of inadequate housing). We recognize the legitimacy of compulsory insurance in other areas; social security is one example, n38 and the inalienable right to a discharge in bankruptcy is another. n39 Whether we should also recognize it in the area of housing will depend upon the relative importance we attach to this good and our confidence that poor tenants will not discount too sharply the value of housing insurance.
Second, even if we evaluate the warranty of habitability on the basis of what tenants actually want, the pass-along argument sketched above is less persuasive than its initial formulation suggests. The extent to which landlords are able to pass along the increased costs of a nondisclaimable warranty of habitability will depend upon characteristics of the rental market that are contingent and variable and cannot be determined in an a priori fashion. n40 Under centain empirically possible conditions, landlords will be able to pass along only a small portion of these costs and will have to absorb the rest; n41 moreover, if alternative uses of the property are economically unattractive, the added cost to landlords of a nondisclaimable [*774] warranty of habitability may cause the number of available rental units to decline only slightly, if at all. When these conditions are satisfied, tenants will receive the benefit of the warranty for less than its full cost and will not be hurt through disinvestment in the rental market. In the limiting case, tenants will get something for nothing, and, short of that, they may get it for less than what they would be willing to pay. If so, they are clearly better off, from their own point of view, under a legal regime that makes the warranty of habitability nondisclaimable.
The pass-along argument used to support the claim that nonwaivable warranties are inherently self-defeating is only one example of a more general criticism applicable to all redistributive schemes that leave any contractual feedom to those whose wealth is being redistributed. Whether one elects to redistribute through the tax system or by modifying the rules of private law that fix the terms on which individuals are permitted to exchange their resources, private arrangements that shift the incidence of the tax or pass along the cost of the regulatory rule may seriously compromise redistributive objectives. This problem can be avoided only by imposing even more dramatic restrictions on the contractual freedom of the individuals involved. If we are unwilling to do this, we have no choice but to evaluate the problem from a comparative point of view and ask ourselves whether it is likely to be more serious under one scheme than another. n42 The relevant question, therefore, is not whether the decision to use a nondisclaimable warranty of habitability as an instrument of distributive justice can be frustrated by the private contractual reallocation of its burdens,
but whether the general problem of pass-along is more serious here than it would be, say, in a scheme that attempts to achieve the same goal by taxing landlords to subsidize low-income tenants. Unless one abandons the goal of providing minimally decent housing for everyone, this comparative question cannot be avoided, and it is by no means obvious that it must be decided against the implied warranty of habitability and in favor of some other redistributive technique (especially when costs of administration and enforcement are taken into account). Which technique we adopt will depend, in large part, upon empirical characteristics of the housing market; the choice, therefore, cannot be made on theoretical grounds alone.
II. Personal Integrity
The nondisclaimable warranty of habitability seems best explained by considerations of economic efficiency and distributive justice. There are, however, many paternalistic restrictions in our law of contracts that cannot [*775] be wholly, or most convincingly, explained on these same grounds. One important example is the varied group of restrictions intended to prevent an individual from contracting away too large a part of his personal liberty. The most obvious and elementary restriction of this sort is the prohibition against contracts of peonage or self-enslavement. Similar, though less dramatic, restrictions include the bar against agreements purporting to waive the promisor's right to engage in a particular profession, obtain a discharge in bankruptcy, initiate a divorce action, or breach a contract of employment and substitute money damages for the promised performance. Each of these limitations restricts a person's contractual powers by creating an inalienable entitlement of some sort, and in each case the justification for the restriction is the same: A person who would give away too much of his own liberty must be protected from himself, no matter how rational his decision or compelling the circumstances.
To justify the prohibition against self-enslavement and its various corollaries on distributive grounds, the concept of distributive justice must be expanded in ways many will think improper. These restrictions are not meant to insure the fair distribution of some scarce material resource like housing; to the extent they are concerned with distribution at all, it is with the distribution of a personal right or liberty -- the liberty to use one's own self in whatever way seems best. It is possible, of course, to treat the distribution of this right in the same way one treats the distribution of rights to nonhuman resources, by assuming that, within certain very broad limits, a person's own talents and capacities represent assets belonging to a common fund, the rights to which must be allocated in accordance with the same principles of
92 Yale L.J. 763, *
fairness that govern the distribution of material goods. n43 Pushed to its limits, this view would empty the notion of personal identity of all meaning and undermine the concept of independent individuality on which all programs of distributive justice rest. But more important for present purposes, even if we accept the idea that a right to dispose of one's own self must be justified on the same distributive grounds as any other entitlement, it is still unclear why this right should be limited by a prohibition against self-enslavement. One can argue, as Mill does, n44 that without such a prohibition the right to self-control may be destroyed through its alienation and conclude that a bar against self-enslavement is necessary to preserve the desired distributional pattern of personal liberties. But this begs the question: Why does a person's inability to enslave himself increase his self-control rather than diminish it? Any theory of distributive justice that purports to explain the prohibition against self-enslavement [*776] assumes the answer to this question, an answer that cannot be supplied by the theory itself.
It is even more difficult to explain the bar against self-enslavement and other analogous legal restrictions in purely economic terms. Consider, for example, the inability of debtors to contract away their statutory right to a discharge in bankruptcy. n45 In certain circumstances, it may be rational for a debtor to waive his right to a discharge in return for a reduction in the cost of credit. An agreement of this sort can effectively communicate information both about the debtor's creditworthiness and the likelihood he will need to invoke the bankruptcy laws. n46 In some cases, it may also be more efficient for the debtor to self-insure against the risk of insolvency than for him to purchase insurance from his creditor (by retaining the right to avoid the debt in bankruptcy). Barring the debtor from waiving his right to a discharge forces him to employ a more expensive method for communicating such information or forego the exchange entirely (if the creditor refuses to lend on any other terms). Both results seem inefficient and are difficult to justify from an economic point of view.
The same is true of the rule barring antenuptial agreements that deprive one or both parties of the right to sue for divorce. n47 By waiving the right to bring a divorce action, an anxious suitor can convincingly convey the depth of his or her commitment (something a checkered marital record, for example, might otherwise leave in doubt). Again, the power to make an enforceable promise of this sort facilitates the communication of information regarding the promisor's sincerity and future intentions; a ban on such agreements makes this communication more difficult and may even exclude some from the marriage market altogether.
The economic argument in favor of allowing parties to make agreements of this kind is similar to the argument in favor of recognizing the validity of so-called "penal" clauses in commercial contracts. In some cases, it is claimed, a buyer or seller can persuade the other party of his trustworthiness only by promising to pay a penalty in the event he fails to perform. n48 The prohibition against penal clauses has been vigorously attacked in the law and economics literature, n49 and it is difficult to see why the same argument does not also apply to the bar against antenuptial [*777] agreements depriving one or both parties of the right to terminate their relationship by divorce.
Even the basic prohibition against self-enslavement is suspect from a purely economic point of view. If Spartacus agrees to become the slave of Claudius in return for a guarantee of food, shelter, and education for his children, there is no a prior basis for thinking the exchange inefficient. The welfare of both parties may be increased by an arrangement that gives one irrevocable control over the labor of the other (subject only to a few broad restrictions designed to insure the slave's safety and physical health) in return for certain contractually specified benefits. If the slave lacks the managerial skills needed to exploit his own labor most productively, it is especially likely that an arrangement of this sort will be efficient. n50
There is, however, at least one plausible economic argument for the prohibition against self-enslavement. Like the nondisclaimable warranty of habitability, the bar against self-enslavement can be viewed as a second-best device for preventing certain forms of deception and duress that cannot be attacked more directly. If we assume that in most cases a person would not contract into slavery unless he were illegitimately compelled to do so, but that such compulsion is difficult to detect and cannot easily be brought under existing rules regarding duress and unconscionability, a flat prohibition against such agreements may be the only administratively feasible way of preventing illicit coercion. n51 There is, however, something unsatisfying about this argument: Even if we had an inexpensive mechanism that enabled us to discriminate, with complete confidence, between coerced contracts of enslavement and uncoerced ones, many, including myself, would still favor the prohibition of all such agreements. This suggests that our intuitive opposition of self-enslavement rests upon considerations that the argument just sketched does not capture or adequately express.
The effort to justify the prohibition against self-enslavement and its legal corollaries is made more difficult by the fact that many enforceable agreements have consequences similar to those the estrictions in question forbid. For
92 Yale L.J. 763, *
example, although an antenuptial agreement waiving either party's right to sue for divorce is unenforceable as a matter of law, an agreement that penalizes the party initiating a divorce action by depriving him or her of any interest in the other's estate may be honored, depending [*778] upon the circumstances and the exact terms of the agreement. n52 Beyond a certain point, of course, there is little difference between simply waiving the right to sue for divorce and penalizing its exercise, and it is puzzling that the law so clearly forbids one, but is increasingly tolerant of the other. Similarly, although a loan contract waiving the debtor's right to a discharge in bankruptcy is legally invalid, a secured transaction in which the debtor uses exempt property as collateral is not. n53 The purpose of a bankruptcy discharge is to give the debtor a fresh start in his financial affairs, but the exemption of certain property from the bankruptcy's estate -- historically, a much older feature of bankruptcy law than the debtor's right to a discharge -- has the same purpose. It is not at all clear why a debtor should be able to compromise his fresh start by contracting away his exemption privileges, but not his right to a discharge. Even the prohibition against peonage or self-enslavement, as well entrenched as it is, permits a variety of employment relationships that exhibit many of the same characteristics. n54 In short, the limits each of these "absolute" prohibitions imposes on contractual freedom may be approached in varying degrees; the result is a gray zone of increasing restrictiveness, rather than a bright line crisply demarcating the permissible from the impermissible. Still, one wonders, why are there any restrictions at all?
The first step in answering this question is to clarify the meaning of self-enslavement. Every executory contract limits the freedom of the parties by creating an enforceable obligation, on both sides, to perform or pay damages: Once an individual has made a contractually binding commitment, his alternatives are limited to these two (assuming the other party is not himself in breach). The distinguishing mark of a contract of self-enslavement is that it purports to take away the latter alternative. From a legal point of view, it is not the length of service that makes a contract of employment self-enslaving, nor is it the nature of the services to be performed; [*779] even a contract of short duration that calls for the performance of routine and unobjectionable tasks is a contract of self-enslavement and therefore legally unenforceable if it bars the employee from substituting money damages for his promised performance. n55 The law will not permit an employee to contract away his right to "depersonalize" a relationship by paying damages in the event he chooses to breach. Whatever its other terms, an employment contract is enslaving if it gives the employer a right to compel specific performance of the agreement. n56
An antenuptial agreement waiving the right to sue for divorce can be described as a contract of self-enslavement in just this sense. The parties to a marital contract have considerable freedom to define in advance the nature and extent of their financial responsibilities in the event of a divorce or separation, but neither can give the other the power to compel specific performance by waiving the right to terminate the relationship through divorce. n57 Here, as in the employment context, the right to substitute damages for the actual performance of the contract is inalienable, and any agreement purporting to forfeit this entitlement is invalid as a matter of law.
Only the prohibition against waiving one's right to a discharge in banruptcy cannot be described in similar terms; since the performance called for in this case is the satisfaction of a monetary obligation, the relationship between the parties is depersonalized from the outset. However, with this one exception (to which I shall return), n58 the prohibition against self-enslavement [*780] and its analogues all appear to have the same goal -- to prevent the promisor from contracting away the right to depersonalize his relationship with the other party by substituting damages for the performance he originally promised.
The prohibition against such contracts is best explained, in my view, by the special threat they pose to the promisor's integrity or self-respect. The nature of this threat is revealed with particular clarity when the promisor's own values have changed dramatically since he entered the contract. To see why this is so, it is useful to draw a distinction between what I shall call disappointment and regret.
Every contractual obligation is undertaken on the basis of certain assumptions about the world (that the fall wheat harvest will be a poor one, that Israel and Egypt will remain at peace, and so forth). If an assumption proves to have been mistaken and one of the parties to the contract is deprived of the benefit he expected, we say he is disappointed: Things have not worked out as he wished. It does not follow, however, that a disappointed promisor will regard his contract as an irrational venture; given what he knew at the time he made the contract, it may have been perfectly rational for him to accept the risks associated with it. A mistaken assumption need not give the promisor grounds for questioning the rationality of his initial decision.
Mistakes of this sort, however, are not the only reason a person may subsequently wish he had refrained from making a particular contract. n59 When someone assumes a contractual obligation, he generally has certain goals he believes will be promoted by the arrangement. If his goals change, the contract may lose its original attractiveness and
92 Yale L.J. 763, *
become pointless (or even reprehensible) from the promisor's point of view, despite the fact that all his assumptions about the world have proven accurate. Even if the fall wheat corp is as bountiful as I predicted, making my futures contract hugely profitable, I may wish that I had not made the agreement if, in the meantime, I have experienced a religious conversion that causes me to look on every worldly endeavor as a polluting distraction. When a person wishes that he had not made a particular contract because his goals have changed, he feels regret, rather than disappointment.
There is an important difference between regret and disappointment. Disappointment does not by itself undermine a person's confidence in the rationality of his own choices; regret can and often does. Whether it is rational for a person to pursue a particular course of action depends upon the goals he happens to have; these form the framework within which the [*781] instrumental efficacy of his actions must be assessed. If a person is merely disappointed by the way things turn out (he expected the wheat crop to be significantly larger), but his goals remain unchanged, the framework within which he retrospectively assesses the rationality of his contract will be the same as the one within which he concluded it was rational to make the contract in the first place. A person may of course wish that he had not made a contract that has become burdensome as a result of events he believed would not occur. But if there is continuity in the goals that shape his practical deliberations, his disappointment will not lead him to doubt the rationality of the original decision, given what he then knew; under these circumstances, a person is likely to feel that he did the best he could; at most, he may blame himself for having terminated his search for information short of the point that was economically justifiable and ascribe responsibility for his misfortune to the uncertainties that plague all human conduct.
If, however, a person's goals have changed significantly, his earlier decision may now appear irrational, for his original aims no longer provide the framework for his deliberations. To assess the rationality of his own actions within the context of his earlier (now abandoned) goals, a person must not only forget what he has learned about the world in the meantime; he must also suspend his present values and ask himself what course of action would be rational for a person with the values he once had. This is a much more difficult task, for it requires the imaginative suspension of the evaluative framework that presently determines the ends to which his knowledge of the world shall be put. Some people are able to understand empathetically the system of goals that guided them at an earlier point in their own lives -- much as a sociologist or anthropologist is able to empathize with the subjects of his inquiry, who often belong to a very different culture -- and to understand the meaningfulness of their experience from within. Even in the latter case, however, our powers of empathy are limited, n60 and they seem even more limited when we attempt to empathize with the selves we once were. Perhaps we feel our own personal commitments more threatened in the one case than in the other; we easily accept the fact that some people lead lives directed toward goals different from our own, but we often find it difficult to acknowledge that we ourselves once had significantly different values. Whatever the reason, a dramatic change in goals weakens our ability to review sympathetically our past decisions within a now altered or abandoned framework of ends; the more radical the change, the more difficult such sympathy will be.
[*782] When a person cannot overcome the distance he feels from his own earlier goals, when he cannot view them in the detached but empathetic way a sociologist might, actions he once thought rational may no longer appear so. From the standpoint of his present values, which he cannot shake off or suspend, his past actions may seem pointless or evil; in this case, he is likely to regard his earlier decisions as a foreign element whose continuing influence appears senseless from the standpoint of his present goals. This can be especially demoralizing. n61 Although there are countless ways in which a person's aspirations can be defeated by the senselessness of the world, if he himself, is somehow responsible for the defeat, he may feel not only that he has been overborne by reality, but that he has, in Aristotle's words, failed to be a friend to himself. n62 Self-betrayal of this sort weakens a person's confidence in his ability to make lasting commitments and guard the things he cares for, n63 and this, in turn, strikes at his self-respect. When a person's goals change, his past decisions can have this effect because these decisions remain his own, even though the normative standpoint from which they once seemed rational has become inaccessible. However strongly he now regards some earlier decision as a foreign element, lodged in his life and exerting an irrational influence over it, a person can never entirely disown the decision (without a considerable amount of bad faith) as an arbitrary piece of fate, like cancer at an early age. n64
[*783] Normally, when a person makes a contract he later comes to regret, he is free to abandon the agreement and simply compensate the other party for his loss. The damages the promisor must pay represent the cost of his own decision, and the obligation to pay them is a continuing reminder of former goals he has since modified or abandoned. A reminder of this sort may intensify the promisor's regret and make it more difficult for him to forget what now seems
92 Yale L.J. 763, *
like an irrational decision, but this cannot be avoided entirely as long as we adhere to the basic rule that a contracting party must pay for the losses caused by his own unexcused breach.
If, however, the promisor is required to perform as he had originally agreed -- if he is barred from substituting damages for the specific performance of his obligations -- his feelings of regret are likely to be intensified, particularly when performance entails some ongoing personal cooperation with the other party or subjection to his personal supervision. If the breaching promisor must continue to work or live with the other party and abide by the terms of a cooperative arrangement he now regrets, he will almost certainly find it more difficult to distance himself from his original values. He is likely, as a result, to feel more directly tied to the goals he has repudiated and to be more painfully reminded of their continuing influence in his life. By substituting damages for performance, the promisor gives his original commitment an abstract form less closely linked to the specific goals that led him to make the commitment in the first place; the edge of his regret is dulled and its disabling consequences ameliorated. If he cannot distance himself from the contract by depersonalizing his relationship with the other party, the promisor's regret is likely to be more intense and its effects more serious; the right to depersonalize a contractual relationship is an aid to forgetfulness, which -- within proper limits -- is a condition of moral health.
When the promisor's own values have changed dramatically, the compulsory performance of a contract requiring his personal cooperation with the other party may pose a special threat to his integrity or self-respect. This is not the only situation, however, in which a moral danger of this sort may exist. Suppose, for example, that an employee working for a pharmaceutical concern discovers that his company is selling drugs used to produce a lethal chemical weapon, a fact he did not know when he entered the contract. If the employee believes that continuing to work for the company is inconsistent with certain deeply-held moral convictions, the violation of which would be a serious blow to his self-respect, he should [*784] be permitted to quit and pay his employer damages for whatever loss the company suffers. Perhaps the employee will regard even the payment of damages as a morally compromising act, but he may well consider his continued employment a more personal and therefore more debasing contribution to a program he thinks immoral. The law assures the employee's right to choose between these alternatives by refusing to enforce any provision in his contract that purports to confer a right of specific enforcement on the employer. We might, of course, protect the employee's moral integrity by relieving him of any duty either to perform or pay damages, but a remedy with such extreme consequences would lend itself to exploitation and tend to undermine the stability of contractual relations generally. n65 To this extent, the bar against self-enslavement may be viewed as a compromise the protects employees from the special debasement sometimes associated with specific performance of an obligation, but leaves them exposed to the moral risks entailed by any contractual commitment made on the basis of incomplete information about the world.
Some contracts, of course, are specifically enforceable, even though the parties have not included a provision to this effect. Contracts for the sale of unique goods are the most notable example. n66 This may seem puzzling, however, for here, too, actual performance can be demoralizing in a way that compensation is not. Suppose, for example, that a seller of machine parts, unobtainable from any other source, discovers that his buyer intends to use them for a purpose the seller considers immoral. If the machine parts are unique, the seller may be ordered to perform despite his moral misgivings and a strong, ethically based preference for monetary compensation. To be sure, the seller can (in theory) buy his way out of the contract if his financial resources are adequate, n67 but so can an employee who has voluntarily enslaved himself; the power to buy one's way out of an obligation is therefore insufficient to explain why specific performance is granted in one case but not the other.
Part of the explanation for this difference in treatment is that contracts of employment are more difficult to specifically enforce because they require greater cooperation by the defendant and therefore involve an element of moral hazard that contracts for the sale of goods do not. n68 There is, however, another reason why courts are more willing to compel performance [*785] in the latter case. Long personal service under the direction of another and the payment of money are in reality poles of a continuum; one is the least, and the other the most impersonal way of meeting a contractual obligation. The transfer of property falls somewhere in between. Many sales contracts do not require any personal cooperation or even contact between the parties and are therefore closer to the pole of maximum impersonality represented by money damages. The more impersonal an act, the less likely it is to threaten an actor's integrity or self-respect; taking this into account, the specific enforcement of contracts for the sale of goods seems more defensible from a moral point of view than its traditional justification on grounds of administrative convenience might suggest.
Finally, how is the bar against waiving one's right to a discharge in bankruptcy to be explained? Since a waiver of this sort merely creates a continuing obligation on the part of the promisor to pay his pre-bankruptcy debts, it might not