Norms: Social and Legal

by Richard A. Epstein

How to Enforce Norms?

        The subject of norms is once again hot. Although norms as a topic never quite disappeared from view, today it receives ceaseless discussion across disciplines: law, economics, philosophy, political theory and the like. The norm resurgence depends on a rich confluence of factors, of which the dominant one perhaps is the recognition of the critical role that norms play in overcoming the ubiquitous collective action problems that crop up in all sorts of social settings. The destructive logic of the prisoner's dilemma (PD) game is by now so well known that it is pointless to demonstrate anew why in many simple two-party games both players defect from the cooperative solution to their dual ruin. That simple model explains not only why individuals confess to crimes when they and their confederates should remain silent, but also why individual sellers cheat on cartels and why individual drillers overexploit productive oilfields. So central is this problem of noncooperative behavior that much of social science fixates on one single question: what social institutions and strategies can overcome this familiar, if deadly, collective action problem?

        The standard answer to this question is often encapsulated in a single word: norms. The norms in question set the standards of cooperative behavior for individuals living and working in groups. When all individuals comply with the applicable norm, the group is able to achieve its ideal cooperative solution, and, most likely will (if only by other norms) find some acceptable way to distribute fairly the cooperative surplus generated by norm compliance.

        Stating the question in this fashion, however, only pushes back the inquiry. What conditions allow for the emergence and stability of norms? That inquiry in turn is plagued by a further ambiguity that is the chief subject of this short essay. Norms come in all forms. In the simplest case a norm could result from an express agreement among group members that norm violations count as breaches of contract for which the violator must answer in a court of law. Yet this equation of norm with contract ignores vital social norms whose emergence cannot be traced to an agreement of any sort. Usually, the discussion of social norms is steeped in the language of custom and usage, common practice, social standards, implied rules of conduct. All of these conventions are widely shared and respected by group members, but do not rest on formal or explicit agreement. Yet in turn this reference to implicit social norms misses the role played by legal norms, the violation of which is met with a coercive response from the state: damages, fines, imprisonment and the like. Any complete theory of norms must deal with all three variations on the common theme: private agreements; social conventions; and public enforcement. How is that best done?

        A complete answer must approach this question from at least two angles. First, there is the descriptive question: different norms are enforced in different ways. What accounts for the distribution of norms across these three categories? Why historically are some norms generated by explicit contract, others by social convention, and still others by legislation or common law adjudication? The second question, which closely dovetails with the first, is normative. Which substantive norms should be placed in which class and why? Even if we assume that all norms are designed to overcome PD games or other collective action problems (which I take to be almost a truism) it is necessary to select the proper enforcement mechanism to counteract the breach. Which norms should be defined and enforced by explicit agreement, which by custom and common usage, and which by legislation?

        The standard literature does not appear to have a good deal to say about this process. Rather, it confines itself to the question of why enforceable norms are needed, without asking the question which type of enforcement mechanism is appropriate for what norm. In answering that question, the lines between the positive and normative tend to blur, for the customary distribution of norm enforcement often reflects the desirable distribution, perhaps for the same reason that customs are efficient guides in other contexts. In this short paper, I cannot comment at great length on the positive question, but I hope to give some clues on how best to answer the normative question swirling about the issue of norm enforcement.

        This normative problem asks what should be the role of the state in creating and enforcing social norms. To look at earlier legal systems, it seems that a good many social norms were just that: norms that were enforced exclusively by social sanctions. Yet to the modern mind, a system that relies exclusively on social enforcement of norms is often said to suffer from some kind of a gap or defect. If social enforcement is good, then why isn't legal enforcement even better? The expressive power of condemnation and approval is made stronger if the entire state stands behind the norm. Individuals have been known to hold out against social pressures and sanctions: indeed some social practices can be undone by persistent behavior that stands in opposition to norms. If there is a social norm against integration, then why not institutionalize the norm with Jim Crow laws? If there is a social norm for integration, why not enforce that norm with an antidiscrimination law? If there is a social norm that calls for survival of the fittest, then why not a social norm that prohibits the giving of charity to the weak and needy? If there is a social norm that the rich must give of their wealth to assist the poor, then why not a legal statute that requires compulsory contributions to a social welfare scheme?

        Over and over again individuals who are confident in the soundness of social norms are tempted to argue that the legal enforcement of their vision is an unmixed blessing, and that the separation of law and morals should be regarded as a regrettable lapse to be overcome by prompt legislative or judicial action. The set of purely social norms is often regarded as falling in an awkward no-man's land between the world of purely subjective preferences (vanilla against chocolate ice cream) and the law of fully enforceable legal norms. The older term, "imperfect obligation," refers to obligations enforced by conscience and social pressures but not law, and was thought in classical natural law theory to represent the correct societal response on the implementation of norms of benevolence. It was never taken as a verbal dodge or a pious evasion of social obligations, but as the preferred mode of their enforcement. Consistent with the dominant pressure for legalization of norms in our time, that phrase has fallen out of common use. In principle, there is no a priori reason to think that legal sanctions should back all social norms. The separation of law from morals is sometimes a good thing, and sometimes a bad one. How do we draw the distinction?

One Failed Attempt

        One notable early effort to draw the line between legal and social norms (or, as is sometimes said, between law and morals) is that offered by H.L.A. Hart in his extraordinarily influential book, The Concept of Law. Hart's major intellectual mission was to defend some version of legal positivism ?- the claim that law should be understood only as the enforceable command of the sovereign, wholly without regard to the substantive content or moral worth of the rule in question. In an effort to respond to the natural law theorists such as Lon Fuller, who believed in some sense that unjust laws were no law at all, Hart sought to give four characteristics that separated legal from moral rules, with the clear understanding that both were subject to some form of sanction. The four characteristics that he specified were (i) importance, (ii) immunity from legal change, (iii) voluntary character of legal offenses, and (iv) the form of moral pressure. A moment's reflection indicates, however, that these distinctions do not carry the day. The initial point is that any distinction between law and morals, between legal and social sanctions, is not strictly an either/or matter. In some cases both systems of control will be invoked, and in some cases either one or the other. That question of distribution does not respond well to Hart's fourfold distinction. As to his first point, moral violations are always regarded as important, while legal ones are often important but sometimes technical. Yet the most vital of the social norms, those which involve the prevention of aggression and theft, are subject to both kinds of sanction. Other practices, such as the keeping of promises, are often subject to both legal and social sanctions.

        But sometimes only social and not legal sanctions apply, as with invitations of hospitality, on the simple ground that the parties do not intend to create legal relations. In still other cases, serious commercial promises are not enforced, because they are not "supported by consideration" (that is, for nonlawyers, part of a bargain relationship), such as a promise to forgive part of a debt. Similarly, the classical common law refused to allow A and B by mutual agreement to impose duties on C ?- a rule vital for the stability of social relationships. Yet they also refused to allow A by promise to B to confer benefits on C for which C could sue in his own right, a dubious legal judgment largely reversed with the passage of time.

        A quick look at the wise and the foolish in legal norms leads to only one conclusion. The importance of a norm has little do with the question of whether it is backed by legal or moral sanctions, or by both. The heterogeneity within both classes of norms makes it impossible to form any rank ordering that correlates importance with method of enforcement. And I dare say that the sentiment runs deeper than this: even if asked to organize their own perfect legal world, most people would not use importance, however defined as the test for separation: some norms are too important, and others not important enough, to be left to the law.

        Similarly, Hart's second line of distinction ?- the immunity from legal change of moral rules ?- is also insufficient to carry the day. The basic insight behind Hart's position is the common-sense observation that social norms differ from legislation in that they are created by an accumulation of practices and not by some distinct enactment of which public notice is then given. It would be quite odd to say that a new custom will take effect as of January 1, 2000, however much the "norm managers" of today might wish it so. But common law (that is, judge-made) rules assume exactly the same form as customary rules, and these Hart's account tends to slight. It would be a foolhardy legal historian that would venture to find one, first clear announcement of the legal prohibition against aggression, or the legal enforcement of (most) promises. As its name suggests, the common law involves those rules that are common to the realm, as articulated by judges who themselves were heavily influenced by customary practices as to what forms of conduct should be regarded as illegal.

        Even if the distinction were sound, however, it would hardly be responsive to the major task. Self-conscious change could become the mark of a legal norm, but it would hardly offer any justification as to what norms should be subject to legal and enforcement and what not. Any principled account of the distribution between legal and social norms should offer more than taxonomy. It should offer a logic why given substantive rules should be placed in either category or both.

        Hart's third test, that moral condemnation requires the capacity to avoid certain forms of harm, is again not quite accurate. For one thing, the criminal law (unless altered by statute) usually adopts the moral concern with mental states, even though the tort law typically does not when it adopts either an objective standard of care that cannot be met by impaired persons or when it more forthrightly adopts of a strict liability rule that attaches liability to defendants who have neither malicious intention nor want of reasonable care. Here the overlap between law and morals seems quite strong, for it is hardly immoral to insist that individuals compensate others for the pains they have caused them unless some wrong of the victim excuses or justifies the injury. The criminal law may have luxury of acquitting a defendant without convicting a victim, but the tort law that exonerates the defendant leaves the injured plaintiff without compensation.

        Strict liability therefore has as many moral defenders (those who think it is immoral to leave the innocent without redress) as it does detractors. The variation within the class of legal and moral judgment once again undercuts this effort at a litmus test tied to mental states. The voluntariness of one person's action is relevant to both moral and legal theory (as it is to tortious and criminal liability) but decisive to neither. Once again too much depends on specific context.

        Finally, it would be a mistake to assume that moral or social sanctions rely solely by appeals to conscience, that is arguments that seek to alter conduct by pointing out the error of one's ways. Social sanctions for persons who have deviated from moral standards take far more concrete forms. When divorce was regarded as a socially unacceptable, divorcees found it difficult to get jobs, join clubs, or run for public office. And in ordinary social life various forms of criticism, hostility, ostracism, expulsion, ridicule, snubs, and boycotts were, and are, are used against individuals who violate rules of dress and decorum, even if no legal sanctions are imposed. These social sanctions work well in cohesive groups, and while they may be intended to educate the offender, they usually carry out an implicit threat to punish certain actions as well. The traditional hostility toward illegitimacy showed the power of these social sanctions even when no legal enforcement was involved.
 

A Tentative Approach

        So the question is how does one decide which rules are to be subject to which form of sanctions? Here it is very difficult to give a clear answer, but a couple of relevant considerations could be pointed out. One constant of the system is that legal sanctions are more expensive than social ones. They require the use of police, lawyers, judges, administrators of all different kinds and sorts, whose behavior must be coordinated up and down the line. One way to put the question therefore is to ask whether the use of these additional resources is justified in terms of improving the outcome that is achieved by the application of ordinary social sanctions by an amount which justifies the extra cost. That question of marginal impact in turn depends on the cost to society of the defection of a single individual from the operation of the basic norm. A few examples might illustrate the point.

        Start with the critical case of aggression and theft. Social sanctions against these practices should be very strong owing to the enormous losses that they impose on others. These individual losses are, moreover, typically social losses in that they are offset only in small part by the gains to the aggressor or the thief. Thus with physical aggression: does anyone think that the momentary pleasure of the rapist is larger than the permanent physical and psychological scars of his victim? And in most cases the fence will pay only a fraction of the value of property to its owner. (And where he values it more, he can do the unthinkable: buy it.)

        The size of the social dislocations are likely to be huge, given the repetitive nature of the wrongs. But social sanctions will not be sufficient to deter the harm. The single individual who is willing to bear the scorn and ostracism of his neighbors could kill (and threaten to kill) with impunity. The ability of the wrongdoer to increase unilaterally the size of the gain is not stopped by any broad social consensus that things should be otherwise. The social equilibrium is massively destabilized even if 99 percent of the population responds to the social sanctions by avoiding the bad conduct. Sooner or later force must be met with force, lest the behavior of the outsider set the norms for the group. Self-defense is one option that is routinely available, but some collective public force is often necessary to deter offenders before they attack an isolated victim. While serious disputes might arise as to how many resources should be directed toward the prevention of aggression, no one will dispute the initial collective decision to impose legal sanctions on force and aggression. The attack on libertarian theory always comes from those who want government to have a greater role. It never comes from those who think that private aggression and fraud are a good thing.

        The enforcement of promises raises much more complicated questions, as the cloudier discussion of the law set out above indicates. There are in fact many situations where parties who are entitled to seek legal enforcement of promises decide that they are better off in keeping their disputes out of the legal system. One reason for doing this is that they have no confidence that judges and juries will sensibly apply the correct legal rules to their disputes. They prefer to rely on the informal sanctions imposed by the operation of the trade. Typically, however, that response arises in those cases where the parties have repeat relationships in close-knit communities, for now the reputational sanctions that deter future transactions will usually be large enough to prevent even the unscrupulous trader from taking advantage of a situation, as was the case in the earlier days when the diamond industry was largely the province of Hasidic Jews trading in close proximity in New York City.

        Modern regulators typically underestimate the importance of these reputational sanctions and often insist on direct legal sanctions where the parties to a transaction might have agreed otherwise. One important illustration of this modern practice is the question of whether employees (but never employers) should be afforded legal remedies against unjust dismissal on the ground of employers greater power. A very large percentage of private contracts take the opposite tack, by allowing firing and quitting both to take place "at will," that is, without showing any "just cause" for terminating the relationship. But many states, both by statute and common law rule, today seek to regulate the practice on dismissal on the ground that these reputational sanctions are inadequate to deter the large institutional employer from serious misconduct. But that conclusion misunderstands the size of the firm is both a liability and an asset. The firm that fires a single worker unjustly faces the risk of demoralizing its entire workforce. The costs of hiring and training and socializing new workers to a firm is immense and is known on both sides of the relationship. The employer that pushes too hard in one case risks retribution in many others, and thus has power incentives to stay its hand. In contrast, the worker who quits and leaves the employer in a lurch will be subject to much weaker reputational sanctions, especially today when employers are afraid that they will expose themselves to defamation actions by giving candid employee evaluations to prospective employers.

        In this social setting the cost and unreliability of the legal system could make the system of social sanctions standing alone far more efficient than the combined legal and social sanctions routinely adopted under the modern law. After all, dismissal still leaves the worker free to seek employment elsewhere, and thus has far different consequences than the killing and maiming that are universally subject to legal sanction. The older distribution of power between the legal and social system had much more sense than the modern critics understand. The very size of the employer creates powerful levers against misbehavior that could outperform legal sanctions.

        The employment relationship is illustrative of a common pattern that can be found in many social settings. The informal norms of the workplace, like the informal norms an industry, club, and school, may well contain detailed specifications of what are and are not proper forms of behavior. It may well be understood that acts taken in violation of these norms are indeed wrongs for which social sanctions are appropriate. The social norm may be one of termination or separation only for cause while the legal norm is one for termination or separation at will. Yet there is no irrationality in that disjunction. Owing to the cost and unreliability of the legal sanctions, it may well be that the parties all accept the basic proposition that they are better off using the less coercive set of social sanctions than subjecting themselves to legal rules. The best run apartment buildings reserve the right to evict tenants at the expiration of leases; yet they are thrilled to renew good tenants at reasonable rates, which they routinely do. The best run firms may reserve the right to fire capriciously; yet the firm spends thousands of dollars working to build good employment relationships with their staffs. Traditional charities reserved the right to turn away the needy at the door, and yet routinely took them in. In each of these cases the social norms are often used when the sole sanction available to a private party is the refusal to deal with other people.

        The older legal position allowed persons to protect themselves by contract against dismissal, eviction and rejection, and indeed set the background legal rule in ways that precluded the creation of legal liability unless otherwise assumed. That system could result in certain miscarriages of justice which galvanize courts and legislatures into action. It is easy to rally around the tenant dispossessed in the dead of winter or the sick patient turned away at the hospital door. Yet all systems are subject to individual cases of failure. The acid test comes in the overall evaluation of the system at large taking into account the melancholy fact that excesses committed in the name of enforcing community norms are often far greater than those of private parties. It is too easy to miss the tenants who leave a public housing project because no one will evict the one tenant who terrorizes the others. It is too easy to forget that the lavish treatment demanded for the hopeless addict fills the emergency rooms that could be used for more worthy recipient.

        Given the complex nature of these social systems, it is not possible to construct any list of factors, such as the factors proposed by Hart, to decide which norms are legally enforceable and which ones are not. It is necessary to give a fuller explanation of the efficiency of the social system under the alternative legal arrangements. But when that task is finished, we should not routinely and complacently regard the separation between social practices and legal enforcement as a gap in immediate need of a cure. In some settings at least, we should regard it as a stable and sensible feature of social life. When legal norms cause more mischief than they cure, they should be avoided. Social norms without legal enforcement do an enormous good. They should not be disparaged simply because they are not perfect. No system is perfect, least of all the law. The excesses of big government today often stem from a systematic misevaluation of the relative value of social and legal norms.
 

Professor Epstein is the James Parker Hall Distinguished Service Professor of Law at the University of Chicago School of Law.