The Economics of Non-State Legal Systems "Perhaps the most distinctive characteristic of the Western legal tradition is the coexistence and competition within the same community of diverse jurisdictions and diverse legal systems. The pluralism of Western law, which has both reflected and reinforced the pluralism of Western political and economic life, has been, or once was, a source of development, or growth - legal growth as well as political and economic growth. It also has been, or once was, a source of freedom. A serf might run to the town court for protection against his master. A vassal might run to the king's court for protection against his lord. A cleric might run to the ecclesiastical court for protection against the king." --Harold Berman, Law and Revolution Bryan Caplan Econ H195AB Cooter Fall '92-Spring '93 TABLE OF CONTENTS 1. Introduction 2. Adjudication of Disputes 2.1. Inefficiency in the public courts: overview 2.2. Inefficiency in the courts: under-pricing, nuisance suits, input waste, and more 2.3. Inefficiency in the courts: fostering wasteful legal conflicts 2.4. Private dispute resolution: an efficient alternative 2.5. The potential and limits of private dispute resolution 3. Formation of Rules 3.1. Preview 3.2. The economic theory of rule-making with no externalities 3.3. Rule-making in the real-world: the problem of externalities 3.4. Non-patentable innovations and externalities problems 3.5. The evolution of custom as a substitute for intellectual property rights 3.6. Ways to internalize the benefits of rule-production 3.7. The variety and flexibility of private rule production 3.8. Public vs. private rule-creation: an exercise in comparative institutions 4. Private Enforcement of Law 4.1. Ostracism and boycott 4.2. The Becker-Stigler model 4.3. Posting rewards and the Posner criticism 4.4. Torts and crimes: incentives for enforcement 4.5. Security guards and private police 4.6. Monopoly on coercion: the source of the limits of arbitration 4.7. Purely private enforcement: a model 4.8. Criticisms of purely private enforcement 5. Conclusion ------------------------------------------------- 1. Introduction Law, even more than national defense, appears to be the perfect example of a public good which simply must be supplied by the government if society is to exist at all. It is non- excludable because everyone enjoys the fruits of law merely by living in society. And it is entirely non-rivalrous -- once the state creates a body of sound legal principles, an unlimited number of people can benefit from them at no additional cost. But these seemingly iron-clad truisms can be undermined by even a cursory glance at history. In primitive societies, law develops gradually from custom in the absence of any sort of government. As Richard Posner explains, "The remaining source of law [in the absence of the state], and the one that dominates primitive law, is custom. Custom (including customary law) resembles language in being a complex, slowly changing, highly decentralized system of exact rules."1 On occasion, this primitive law gradually spreads outside the narrow confines of a single tribe to encompass a broader community. Thus, early tribal Germanic law evolved into a more universal legal code in the absence of a central government. What makes this progress gradually unfold? As legal historian Harold Berman explains, "Violation of the peace of the household by an outsider would lead to retaliation in the form of blood feud, or else to interhousehold or interclan negotiations designed to forestall or compose blood feud."2 (emphasis added) By this process primitive law became more civilized, broadening its vision to include anonymous as well as face-to-face societies. In order to exist and evolve, such systems must have given rule- creating incentives to someone; in other words, at least some of the benefits were not public but exclusive. There is another sense of "non-exclusion" that legal systems must allegedly have: they must govern everyone in order to function at all. If law-breakers could simply drop out of the system, law could hardly protect us from their misdeeds. And yet, history contains many instances of pluralistic legal systems, in which there were multiple sources of law in one and the same geographic region. In the medieval society that Prof. Berman investigates, canon law, royal law, feudal law, manorial law, mercantile law, and urban law co-existed; none was automatically supreme over the others. Naturally, there were jurisdictional conflicts. But this system of concurrent jurisdiction overlapped with a period of economic and political growth (c.1050-1250), not a period of chaos and impoverishment. Apparently these diverse systems did what Thomas Hobbes (along with most modern political thinkers) declared impossible: They created social order and peace in the absence of a distinct, supreme sovereign. These examples might seem to be historical anomalies, fascinating but irrelevant to current economies and legal systems. Yet in modern times parallels have sprung up, albeit in a less dramatic form. Commercial disputes, for example, are handled virtually in toto by means of private arbitration. Accurate figures are very difficult to get; but one expert in alternative dispute resolution, Jerome Auerbach, estimates that businesspeople arbitrate 75% of their commercial disputes.3 In earlier times, one of the central functions of government was business dispute resolution; but now it has largely escaped the state's sphere of influence. A more extreme example is that of the VISA corporation.4 Member banks agree to keep their quarrels within the VISA family when they join the central organization. Anticipating many costly legal disputes between the system's members, the VISA corporation saw the opportunity to invent a cheaper way to resolve disagreements. It created the VISA Arbitration Committee to judge the disputes of the member banks according to VISA's very own legal code. The methods are quick, lawyerless, and unbureaucratic. Compared to the slow and costly justice that the banks receive when they have to settle a conflict with a firm outside the VISA camp (and within the reach of the public courts), the VISA banks get a bargain. It is the economic characteristics of these sorts of legal systems that this thesis investigates. How do they work? To what extent might private legal systems of this kind partition and swallow up the near-monopoly of law that most governments possess? How does customary law give incentives to create and develop a legal framework? It should be noted that I am not merely studying the economics of federalism. While there are some parallels between pluralistic legal systems and federalism, there is a crucial difference. Under federalism, a central government delegates the authority to make laws to two or more sub-polities. In this sense, there is legal choice and legal competition. But federalism gets rid of a national legal monopoly by creating a host of smaller geographic monopolists, each of which has sovereignty in its own territory. Truly plural legal systems such as arbitration compete within the same geographic region. There is more than an analogy between pluralistic legal systems and the economist's conception of competition. With a large number of potential source of law in a given geographical region, plus some method of excluding non-contributors from benefits, economic theory implies the familiar results that hold of competitive markets generally: productive efficiency (a given level of output gets produced at the minimum cost), allocative efficiency, (resources get assigned to their most socially productive uses), and dynamic efficiency (declining cost curves over time). We might also expect to find the greater flexibility and attentiveness to individual differences that typify private supply. This thesis investigates the non-state supply of the three essential aspects of law: dispute resolution, rule formation, and enforcement. The first and least controversial is the non-state resolution of individual disputes. The state could still choose the rules and merely privatize their application. Economists usually see the rule-application aspect as the best candidate for privatization, because the parties to the dispute get all of the benefits. This is one factor that explains the emergence of primitive law: the need to avoid inter-family bloodshed over every wrong gave both sides to a dispute an incentive to voluntarily submit to a peaceful settlement procedure and abide by whatever result emerged. Even the loser benefits, since he trades submission on a single quarrel for long-run peace. In terms of game theory, what looks like a zero-sum game (a tort, say) is actually a cooperative game because refusal to adjudicate leads to feuds between individuals, families, or clans that leave both sides worse off. But there is a second element to law; as Landes and Posner explain, "A court system (public or private) produces two types of service. One is dispute resolution - determining whether a rule has been violated. The other is rule-formation - creating rules of law as a by-product of the dispute-resolution process."5 To many people, the latter is a less likely candidate for private supply, because the production of rules of law is a public good. The economic value of precedents (like other intellectual innovations) is hard to internalize. Section three will examine this issue at length. But there are prima facie reasons to be skeptical of this criticism of private rule creation. The VISA corporation has its own legal code; so do other professional associations, clubs, and cooperatives. More startlingly, most primitive law and a great deal of commercial law were developed by bodies other than the state. Thus, the law merchant (later adopted, not created, by most governments) evolved during the Middle Ages (c.1000-1200), while national governments ignored the demand for well-defined rules of international trade. Merchant courts gradually developed contract and tort law, defining rules of incorporation, credit instruments, and damages.6 Legal growth on this level would seem to be possible only if private incentives existed somewhere; section three will seek out the incentives' sources. Last, there is the most controversial function of law that might be executed privately: enforcement. The right to use force seems to be a necessary monopoly of the state; otherwise wouldn't chaos and random violence be inevitable? Still there are many types of non-violent private enforcement. Commercial boycott and ostracism enforce most forms of arbitration. Within the business community these sanctions are quite effective. Becker and Stigler note that termination of employment (rather than legal action) can and often does deter malfeasance. "The fundamental answer is to raise the salaries of enforcers above what they could get elsewhere. A difference in salaries imposes a cost of dismissal equal to the present value of the difference between the future earnings stream in enforcement and in other occupations."7 Section four explores the enforcement tools of private legal systems and how they work. It also raises questions about what might happen if we relaxed the state's monopoly on force and let more drastic forms of enforcement fall into private hands. In sum, this thesis examines how non-state legal systems work in their three distinct aspects: dispute resolution, rule- creation, and enforcement. It will draw on many conclusions from economic theory; most frequently, the prima facie superiority of private competitive supply to public monopoly. There are many conceivable market failures that might exist; but the strategy of this paper is to at least explore overlooked legal alternatives. Before the body of the paper can begin, however, it is necessary to point out a a major misconception that might prevent one from seriously considering the private supply of law. The most deeply rooted obstacle to the appreciation of non- state legal systems is the theory, eloquently stated by Thomas Hobbes, that the law-making function indivisibly and necessarily belongs to the sovereign. As Hobbes famously wrote, "The only way. to defend them [mankind] from the invasion of foreigners, and the injuries of one another. is, to confer all their power and strength upon one man, or upon one assembly of men, that may reduce all their wills, by plurality of voices, unto one will."8 On this view, the presence of more than one law-giver in a region has to degenerate into violent battle until only one remains. Imagine, the argument might go, that two councils issue law in one city. Their laws will inevitably conflict on occasion. And since by assumption there is no higher body to resolve their dispute, they will have to resolve it violently. Two things can be said here. First, history gives examples of concurrent jurisdictions that operated peacefully. Harold Berman, for example, describes multiple competing sources of law in the Middle Ages, with only a handful of serious conflicts between them. So Hobbes must be wrong at least sometimes. Second, Hobbes' game theoretic deductions are inconsistent. Supposedly, driven by the desire for food, money, and power, rational egoists turn to violence to get what they want when there is no stronger entity to preserve peace. And yet, as Hobbes says, the fear of death is even stronger than the drive for food, money, and power. Given this, isn't the rational strategy the defensive one of live-and-let-live? Given the choice between two evils - a small chance of death vs. slightly less consumption - a Hobbesian man would surely prefer the latter .9 This preliminary observation gives a plausible explanation for why peaceful cooperation was frequently the dominant strategy. Throughout this paper I imagine that profit-making firms, rather than non-profit organizations, would be the major alternative suppliers of law. This assumption overlooks the fact that much, perhaps most, arbitration is not run for profit. I do this for two reasons. First, I find it difficult to believe that non-profit firms could compete with for-profit firms if they were on equal legal footing. The second reason is simply that the economic theory of profit-making organizations is better-developed than the theory of non-profit groups. I trust that this assumption will not seriously affect the analysis. A final caveat. This paper considers only the non-state supply of private, as opposed to public law. That is, it considers systems of alternative resolution of conflict between individuals whose rights are defined by the common law categories of property, contract, tort, and crime. Public law must be left for another time. 2. Adjudication of Disputes 2.1 Inefficiency in the public courts: overview One of our most ridiculed institutions is the public court system. Interestingly, economic analysis gives us clues about many of its central weaknesses. Some of these problems could be solved with fairly minor reforms. For example, court services are under- priced (usually, free). These leads to serious excess demand, permanent shortages, and strategic delays. Presumably, if the public courts were so inclined they could charge court fees for civil cases to ration demand. A second problem is nuisance suits. Frequently, there are lawsuits where one party is clearly in the wrong, but drags the battle into court anyway in the hope that the other side will simply give up. The public courts could solve this problem by changing the indemnification rule - for example, by making the loser in a nuisance suit pay the other side's court costs. Other inefficiencies in the public courts would be hard to eliminate with minor reforms. The public courts are supported by taxes; they therefore have little incentive to control costs. Trials take too long, appeals are too frequent, and labor discipline is lax. Since juries are conscripted, courts treat their labor as a free good; consequently, they use juries even when the value of their contribution to justice is small. And perhaps most seriously, the courts foster wasteful legal battles. Instead of encouraging litigants to limit their joint legal expenditures, they give them incentives them to race to out-spend each other. But since the expenditures usually cancel each other out, this legal competition is rather futile. It is hard to see how the public court system could solve the second group of problems even if it wanted to. Indeed, it probably won't remedy the first set of difficulties either. In order to understand the benefits of turning disputes over to private alternatives, we must first understand why and how the public courts fail. This done, we can investigate the ways that private bodies overcome the inefficiencies that the public courts cannot. 2.2. Inefficiency in the courts: under-pricing, nuisance suits, input waste, and more An excellent work detailing the public courts' failures is Judge Richard Neely's Why Courts Don't Work. As judge with training in economics, he is particularly qualified to point out the failures of our court system. Neely has a long list of complaints. First, the courts grossly underprice their services, leading to excess demand and non-price rationing (usually, waiting in line). It is illegal to sell one's place in line. The most urgent cases must wait as long as trivial ones, leading to protracted legal conflict and higher legal costs. If courts charged user fees, people with insignificant disputes would be more likely to drop them. Or they might try a cheaper resolution method. Yet the courts provide subsidized (free) services, often complete with juries. The whole process is expensive, but litigants only need to consider their lawyers' fees. They ignore the cost to taxpayers of extra trials. They also ignore the cost of justice denied and delayed to everyone waiting behind them for their day in court. Underpricing also helps the legally well-endowed wear out their opponents. As Neely puts it, "Often the attractive products that the court delivers free are delay itself or a forum that provides the stronger litigant with an opportunity to wear out or outgun the opposition."10 (With perfectly functioning capital markets this could not happen: banks would happily loan money to litigants with good cases. But strategic delay does seem to be a real problem.) Since a longer delay gives both sides a greater opportunity to out- spend each other, delay usually favors the richer litigant. Delay also deadens the deterrent effect of damages: future damages, like other future income streams, will be discounted by the interest rate. There is a simple economic explanation for all of these problems: since court services are free, there is no way to ration them other than waiting in line. And this probably is uniquely severe for dispute resolution, because the litigants can ferociously struggle with one another while they wait to go before the judge. Neely points to a second problem that plagues the justice system: the courts give incentives to litigate non-disputes. For instance, in landlord-tenant cases or creditor-debtor cases, there is rarely any "legal issue." Instead, as Neely explains, one side normally just refuses to fulfill its half of the bargain. "In the universe of all the routine cases that go to court, most of the time one party will be flat wrong, and he or she will know that from the beginning."11 If legal costs exceed the expected value of the judgment, then aggrieved parties may well drop legitimate cases. When nuisance cases do get a full trial, they crowd out more substantive disputes. One solution for insincere cases (though Neely rejects it on distributive grounds) would be to make the loser pay both sides' costs -- but this is a reform that courts are loath to try. At the very least, in areas of the law filled with nuisance suits, this idea has potential. Profs. Landes and Posner argue that the public courts waste and misallocate their resources; after all, it is taxpayers, not judges or litigants, who pick up the tab. Landes and Posner then reasoned that the arbitrators are likely to make more efficient use of inputs than the public courts. Since arbitrators do survive on user fees, disputants and arbitration firms alike will want to contain costs. From these assumptions, these two legal economists proposed the following test of the efficiency of the public courts' civil trials: the public courts are efficient if they match the practices of arbitration. Since arbitration does not use juries or lawyers, and the public courts do, Landes and Posner conclude that these may be inefficient, as least in civil cases.12 (Since criminal cases are not arbitrated, we can't know what criminal arbitration would be like.) This is input waste on a huge scale - a majority of the inputs in civil cases may well be unwarranted. The misallocation of jury time is especially egregious, since they are conscript labor - virtually a free good. While juries cost a great deal to society (including the jury members themselves, who lose work time), courts and litigants have the incentive to use them even when the benefit is negligible. Landes and Posner use the same efficiency test for other public court practices. For example, controversy exists over the merits of the loser-pays rule for legal expenses. Arguments cut both ways: If trials occur because of over-optimism, then the loser- pays rule leads to more suits; if many suits are "nuisance" suits in which the party in the wrong strategically delays, then the loser- pays rule would lead to fewer suits.13 It is hard to do an empirical test to everyone's satisfaction. However, we could reasonably predict that profit-maximizing private courts would use the more efficient rule, especially if the parties pre-contract to arbitrate with a specific firm with a set indemnification rule. How does the test turn out? The American Arbitration Association requires the defendent to pay all legal costs if the plaintiff wins, but splits the difference if the defendant wins.14 Posner suggests that this vindicates the American rule over the English; but actually this rule implies that defendants are often clearly guilty, whereas malicious suits by plaintiffs are infrequent. Or in other words, nuisance suits by defendents are more common than nuisance suits by plaintiffs (in disputes currently open to arbitration). The parties can also voluntarily change the indemnification rule: some contracts change the standard AAA rules, stipulating that the plaintiff pays the legal costs if he loses. Landes and Posner bring up another interesting issue: appeals. They argue that private arbitration (excluding trade associations) lack appellate courts because the sole function of such courts is to formulate rules of law, not resolve disputes; and the former, unlike the latter, is a public good. Landes and Posner view the production of rules as a pure public good: society at large benefits when someone refines a legal principle, but (as is often the case with intellectual creations), it is hard to claim a property right in a precedent. This claim may well be true; it will be examined in section three. But perhaps there is another explanation: Private courts do not permit appeals simply because the extra costs (in time, legal fees, court services, and so on) are not worth the social benefits. Both parties gain if they agree ex ante to limit each other to a single hearing. But in public courts there is no way to credibly commit to limit appeals. On this theory, the lack of appeals is a benefit to both parties because it keeps dispute resolution costs low. Posner and Landes point out that trade associations do permit appeals; and these appeals sometimes produce precedents. They argue that this happens because a trade association can internalize the benefit of a precedent. True, but they also concede that appellate tribunals are not universal. In all likelihood, trade associations rarely permit appeals, strictly limit their expense, or both. Furthermore, the VISA corporation does not permit appeals, even though the corporation's unique structure and secrecy enable it to fully internalize the benefit of precedents. Quite possibly the permission of appellate review in criminal cases makes economic sense; for as Posner suggests, the high error costs of convicting the innocent may justify the reasonable doubt rule for evidence in criminal proceedings.15 On the other hand, the behavior of many arbitration firms and trade tribunals suggests that the appeals process in civil cases has excessive costs. 2.3 Inefficiency in the courts: fostering wasteful legal conflicts Even the most harmonious society has disputes. In the interest of peace, these disputes must somehow get resolved. The upshot is that every dispute comes with resolution costs tied to it. There is some additional burden on top of the costs of the dispute that the disputants must eventually split. This is often a substantial sum. From this is follows that to a limited extent the plaintiff and defendent have a common interest: minimizing their respective dispute resolution costs. We may infer that the parties share an incentive to cheaply resolve their conflict. In the government's court system we see this principle at work when defendants settle, or alleged criminals plea bargain. Yet public courts have clear inefficiencies that both increase the costs of dispute resolution and make settlement more difficult. Judge Neely's strongest criticism of the public court system is that it promotes futile but expensive strategic behavior. The outcome of a case depends not merely on the facts of the dispute; it is also a function of the respective legal expenses of the two sides. Since the disputants cannot reach a cooperative solution to the dispute itself, it is likewise difficult to agree to limit joint legal expenditures. The result is that both plaintiff and defendant rush to outspend each other, but ultimately the probability of success remains unaltered because the competitive expenditures cancel one another out. Neely's proposed solution would likely arouse skepticism from economists: "What is needed is a court version of the Strategic Arms Limitation Treaty - a method for determining in advance what a reasonable investment in a particular lawsuit is, and a court order forbidding both sides from spending money for a competitive advantage that in the nature of things will be illusory."16 Like statutory caps set on excessive punitive damages, this is probably a band-aid measure destined to create bureaucratic inefficiencies of its own. Hard questions present themselves. First, judicial determination of maximum legal expenditures would itself use up valuable court time. Maximum permissible expenditures might itself be an issue of legal contention. Second, there would surely be active interest groups who would struggle to adjust the ceiling in their preferred direction. Frequent defendents might very well want the cap pushed down as far as possible to remove the incentives for plaintiff's lawyers to bring suits against them. Habitual plaintiffs and their attorneys would in turn lobby in the opposite direction. (Or perhaps they would favor a cap that applied solely to defendents!) Third, it seems difficult to see how lawyers working on contingency could be dealt with under Neely's rule. Would their awards be capped? Would the courts set a ceiling for the permissible total lawyer-hours per case? These problems seem serious -- for public courts. Politics would probably prevail over economics. Private courts, on the other hand, already limit legal expenses with marked success. The AAA does not permit lawyers -- a quite drastic limitation. VISA does not even allow the parties to attend their own hearing. But these are efficient results. On the average, these practices are unlikely to harm the legal prospects of either side. Yet these practices limit the joint legal costs per case -- a big plus. Since both sides typically agree to arbitrate disputes in advance, before cooperation has broken down, it is easy to pre- commit to mutually limit legal costs in future disputes. Why has private dispute resolution worked so well in this area? The public courts would find it hard to administer such a rule. Is there a critical difference between the limitations that private and public courts place on legal expenditures? I think there is. To understand it, we should turn to Ronald Coase's classic article, "The Nature of the Firm."17 Managers of a firm, Coase explained, usually run it by the "command-and-control" methods that economists deplore on the economy-wide level. Once a worker gets hired, he is expected to do what he is told; similarly, office supplies are likely to be centrally allocated to each department rather than sold to them. What this shows, said Coase, is that command-and-control methods (since they survive and thrive within competitive firms) must be useful to some extent; most notably, command-and-control reduces transactions costs. The problem with centrally-planned economies is that they extend command-and-control techniques far beyond their efficient point -- and then eliminate the competitive pressure that checks this inefficiency. In competitive markets, a firm that grows so large that it cannot effectively manage itself starts to lose market share and profits. This gives an incentive to scale back to a less cumbersome size. Central planning by a government does not face this disincentive. For private courts, limiting the use of lawyers, expert witnesses, and so on would be akin to any other business decision. All firms use command-and-control to some extent; what prevents inefficient command-and-control is the pressure of competition. When private courts experiment with restrictions on legal expenditures, they need merely judge the needs of their own clientele, not those of the whole society. And when they judge incorrectly, competition can straighten them out. If a firm decides to limit legal expenditures, market share and profits can indicate whether or not the practice is efficient in that particular case. No one ever needs to make the much more difficult judgment about what is efficient for all firms. In contrast, the public court system is a centrally-planned industry -- when it uses command-and-control there is little or no feedback to show whether its actions are sensible. And if it chooses wrongly, consumers often have no close substitute. More critically, the public courts choose not merely for a single firm; they choose for the whole society. Judgments of this kind are likely to be wrong because, first, there is no market feedback, and second, because individual preferences and circumstances differ too much for one set of rules and procedures to suit them all. Judge Neely is surely right that litigants would often be better off if they could mutually "disarm" by jointly slashing expenditures. Unfortunately, given the public courts' structure and political constraints, this task may be impossible. Yet private arbitration firms could cheaply explore these possibilities; they do so already (for example, by prohibiting the use of lawyers and appeals). One way to cut back on legal costs without massive administration problems would be to open up a wider range of disputes to private resolution and increase the autonomy of private alternatives from public control. This list of inefficiencies in the public courts is hardly exhaustive. We might also note that judges make many economically unwise rulings. Juries in civil cases, guided more by emotion than by consideration of long-run consequences, do the same. In any case, the public courts are sufficiently unattractive that we should seriously consider alternatives. And I think that Landes and Posner point in the right direction: if we can measure the efficiency of public courts by comparing them to arbitration, then private dispute resolution, if possible, is a welcome option. Fortunately, this option is not only possible but real and growing. The next two sections will explore its benefits, its solutions to the problems of the public courts, and the its feasible scope. 2.4 Private dispute resolution: an efficient alternative Many of the faults of the public courts would not exist (or would be less severe) under private arbitration. Private courts could raise fees to efficiently ration judicial services. They could experiment with indemnification rules to reduce their clients' expected legal costs. Firms might offer various methods to restrain joint legal costs (by, for example, prohibiting or limiting the use of lawyers). They could limit or eliminate appeals. Each of these problems seems difficult for the public courts to manage: partly for political reasons, but also because public monopolies have little ability to recognize entrepreneurial opportunities. Consider some further advantages. One characteristic of private supply is that it recognizes that consumer have different needs; and since many suppliers can survive in an industry simultaneously, it is possible for them to sell a wide variety of services side by side. Some parties prefer swift decisions at the cost of lower accuracy -- the member banks of the VISA corporation, for example, realize that errors will even out in the end, but that adjudication costs increase with each dispute. VISA consequently has a system of rough but swift justice. Other conflicts -- for example, over isolated contracts between strangers -- require a more thorough investigation. Public courts have a systematic bias toward excessively slow resolution; but even if the courts were right on average, they would still ignore the fact that litigants' preferences vary. The most impressive arguments for privatizing dispute resolution have little to do with the unique attributes of the adjudication industry; rather, they are the standard arguments for the prima facie superiority of private to public supply. Namely: (1) Public bodies have no incentive to be efficient, and private ones do; and (2) Public bodies usually don't know what is efficient, while private bodies, though not omniscient, know better. Why don't courts have any incentive to be efficient? First, there is no residual claimant with an interest in cutting costs and increasing consumer satisfaction. In profit-making firms, the owners have an incentive to keep costs low and make them fall over time. And the incentives of the employees are different. Judges are typically either elected, or appointed for life. Elections are a a bad way to monitor work effort -- informing oneself about each judge's attributes is a pure public good: society at large benefits from intelligent selection of judges, but individual diligent voters bear the costs. Life appointments take away even the meager incentive effects of voting. If we want the public courts to work, we must rely on the self-monitoring of the judges themselves. This might work sometimes. But is it a good incentive structure? The incentive structure of private labor markets is more sensible: while they have imperfections, private labor markets leave employment decisions up to a concerned manager or entrepreneur, not the public at large. These managers reward their employees if they work well and fire them if they don't. Surely this spurs work effort better than voting or life appointments. Second, as Hayek and others suggest, private markets use knowledge more effectively than public monopolies.18 They are more able to calculate costs and benefits. In markets there are explicit prices that measure costs and benefits. But public bodies must estimate social costs and benefits by using (at best) surveys or (at worst) guessing. The judicial industry needs the low-cost experimentation that private firms can provide. It is cheap and safe if one firm decides to restrict the use of lawyers, or get rid of appeals, or change the indemnification rule; even if these experiments flop, the losses to society are small. Often an experiment proves useful, at least for one section of the consuming public. Private adjudication services would be free to experiment and see what their clientele thinks. Public courts, in contrast, rarely try new ideas. But there is perhaps a justification for this -- namely, their error costs are terribly high because public experiments involve everyone. Perhaps public courts hold to the status quo because they fear that their experiments will fail miserably. What we need is to permit experimentation, but keep it decentralized so that mistakes can be abandoned before they become disasters. And private firms are the ideal arena for low- cost experimentation. For years, academics in law and economics have speculated about the relative efficiency of different rules and procedures for trials. Rather than have the public courts try each out in succession, we could expand the scope of private courts and see what innovations evolve. 2.5 The potential and limits of private dispute resolution As Posner and Landes point out, there are two cases where private dispute resolution works best: "(1) those where a preexisting contract between the parties requires submission to arbitration according to specified rules for selecting an arbitrator, and (2) those where the disputants belong to an association which provides both arbitration machinery for its members and a set of effective private sanctions for refusal to submit to arbitration in good faith or to abide by its results."19 They go on to suggest that (1) works only because the government courts enforce the contract. This seems partly correct. But reflection suggests that preexisting contracts to arbitrate could work if the public courts simply refused to overrule them. Consider video rentals -- before we can get a rental card, we must authorize the rentor to use our credit card if we do not return the video or pay our fees. Our credit card company in effect guarantees our trustworthiness; and if we break our agreement, it pays the video rental firm. If we refuse to pay our credit card bill, our company could of course take us to court; but this is usually so ineffective that it simply ruins our credit rating if we renege. If we extend this model further, we can imagine an effective way to enforce all sorts of contracts -- including arbitration contracts -- non-violently and without the help of the public courts. Put simply, it is not actually necessary that we have repeated interaction with all of our fellow contractors in order to make non-violent enforcement necessary; we must merely repeatedly interact with one firm whose job it is to guarantee our payments. Repeated interaction, game theory teaches us, can substitute for enforceability. So long as parties develop some kind of bonding relationship, they do not need the public courts to enforce their agreements. However, it is necessary that the public courts and legislature refuse to overrule them. Arbitration is a way to escape from the public courts; but if the public courts regulate arbitration, the "escape" is less effective. What I am suggesting is that contra Posner and Landes, arbitration can work even if the public courts don't enforce it, but that it can't work if the public courts positively disallow it. Nathan Isaacs, a professor of business law at Harvard during the 1920's, noticed that when the government began to enforce arbitration (in 1920 in New York) it also began to hamper it: "There is irony in the fate of one who takes precautions to avoid litigation by submitting to arbitration, and who, as a reward for his pains, finds himself eventually in court fighting not on the merits of his case but on the merits of the arbitration."20 What then, should public courts and legislatures do (or rather, not do) to make arbitration work as well as possible? First, courts must refuse to review any arbitration clause in a contract; i.e., make a clear rule about what an arbitration clause must say to be court- proof, and then rigidly stick by that rule. Second, courts must refuse to review the content of arbitration, leaving the efficiency and justice of arbitration to the parties' judgment. Third, legislatures must refrain from legally hampering the ostracism and boycott efforts of arbitration firms, professional associations, and credit card companies (for example, by banning credit ratings as an invasion of privacy). Private firms do not have violence at their disposal, so they must use subtler methods of enforcing agreements, like credit ratings and reputation. To put restrictions on these comparatively mild enforcement techniques makes it difficult for arbitration to work at all. Fourth (and least importantly), the government should refrain from making any antitrust charges against professional associations that publicize the untrustworthiness of members who refuse to arbitrate or submit to sanctions. How come only preexisting agreements to arbitrate work? First, it is difficult to enforce a claim non-violently unless both parties can credibly pre-commit to comply. Especially if one party is sure to lose in any fair trial, he has no incentive to cooperate unless he previously made himself vulnerable to sanctions (for example, by authorizing his credit card company to pay for any damages). Second, it might be difficult to get a fair trial even if two common law strangers agreed to arbitrate. Each side might try to get arbitrators, rules, procedures, etc., beneficial to its case; and if the other declined, he might be accused of "foot-dragging," of willfully delaying and stifling the resolution process. These problems might make us skeptical of purely private resolution of quarrels between strangers. Nevertheless, many disputes do not involve strangers; and in principle all of these might be turned over to the private sector. All commercial disputes, employment quarrels, creditor- debtor complaints, landlord-tenant problems, and perhaps divorce and products liability fit the mold well. The parties need merely record their mutual decision to arbitrate future quarrels (including perhaps their preferred arbitrator), plus some sort of assurance or guarantee of compliance. Commercial disputes might rely on reputation effects; employment quarrels on the threat to fire on the one hand (to exact compliance from employees) and the harm to worker morale on the other (to get employers to accept decisions); landlord- tenant relations on a security deposit. In other situations the parties might use credit cards or the like to pre-commit themselves to pay up. (More on this in the section on enforcement.) Consider, for example, sexual harassment. Instead of having court-enforced anti-harassment laws, we might leave firms to develop their own policies, or "law" on the matter. When a male manager and a female employee have such a dispute, either the firm itself or a sub-contractor might conduct an internal investigation with its own procedures, rules of evidence, etc. The firm has the power to enforce the arbitrator's ruling: it might fire an offending employee, attach his wages in order to compensate the injured party (coupled with a threat to enter the harasser's name with a sort of "employee rating firm" if he should opt to quit), or any number of other remedies. In order to attract a satisfied workforce, each firm might compete to develop more efficient rules and enforcement techniques. There are at least two important factors that they might weigh: first, the harm done to harassed employees; second, the harm done to those falsely or mistakenly accused. Competing employers would probably better balance these two costs than the courts. If the government were involved at all, it might restrict itself to summarily enforcing the result of outside arbitration. If, for example, a landlord-tenant arbitration case rules that the tenant must leave at once, the government might restrict itself to enforcing the order. Or if a products liability arbitrator rules in favor of an injured customer and the firm refuses to pay up, the courts might merely summarily order the firm to pay. What about the hard cases, such as torts (between strangers, anyway) and crimes? Courts might summarily enforce arbitration to which both parties agree. In a way, they do this now with plea bargains and settlements. The only difference would be that instead of enforcing a determinate outcome, the courts would agree to enforce the outcome of a process, however it turns out. It might be difficult to get two strangers to agree to a common arbitrator. But, as Posner and Landes point out, arbitration entrepreneurs have thought about this problem for a long time and developed some workable solutions. The AAA sends both parties a list of arbitrators; each side then crosses off unacceptable arbitrators and ranks the remainders according to his or her preferences. Whichever arbitrator gets the highest joint ranking gets the job. And this system aligns both sides' incentives skillfully. As Posner and Landes explain, "A party who crosses everybody off the list hurts only himself, by guaranteeing that the arbitrator will be selected from among the names not deleted by his opponent."21 It might be unfair for the courts to impose arbitration on unwilling parties; but perhaps arbitrators would strive to develop reputations for fairness to lure parties to try it voluntarily. After all, the parties to a court case rarely have any choice about their judge unless he is plainly biased. (The side that likes its assigned public judge might refuse to arbitrate. To prevent this problem, one might delay announcing the judge until after the parties learn of their option to arbitrate.) If this section has an overall conclusion, it is that competitive arbitration could successfully take over a large array of disputes currently handled in the public courts. There are no obvious market failures for individual dispute resolution. Arbitration firms' decisions and procedures would be kept fair and efficient by means of free competition, rather than by self-monitoring or voting. Since there are no glaring market failures, the results of competitive arbitration would approximate the high expectations that we have of freely competitive industries in general. But for arbitration to work, we must meet two conditions: (1) The public courts must adopt a simple rule about what arbitration clauses must say to bind the signatories, and then stick to that rule; (2) The legislature must refrain from passing laws that hinder non-violent private enforcement like ostracism and boycott. Resolution of individual disputes could easily be privatized. There are no apparent market failures, such as externalities, imperfect information, or monopoly. At least these problems do not seem unusually serious. (We might also note that these "market" failures have parallels in the public courts.) The next function of private legal systems that this thesis will investigate has, in contrast, been accused of a quite serious market failure. Landes and Posner allege that private organizations and arbitrators have no incentive to make rules because of serious positive externalities. The next section examines private "law-making" and investigates whether the market failure is as serious as Landes and Posner think. 3. Formation of Rules 3.1 Preview What does economic theory have to say about the private supply of rules of law? First, a rule-production market with no externalties would be economically efficient. If it were possible for rule-inventors to charge judges and arbitrators who use their rules (with some costless copyright system, for example), then private supply would be quite feasible. In fact, such an idealized system would be a marked improvement over rule-creation by the public courts: there would be stronger incentives for efficient law- production, and product differentiation would make it possible to satisfy many sub-markets simultaneously. Alas, the rule-production market has serious externality problems because it is difficult to establish property rights in a precedent. Posner and Landes take this as proof that private rule formation is an industry with little or no realistic potential. But things are not so simple. For one thing, patents are not the sole or even primary incentive for innovation. Surveyed businessmen in R&D-intensive industries rate several other factors more highly, as we shall see below. Unsubsidized firms continue to produce non- patentable innovations, so other incentives must exist somewhere. On top of this, it is sometimes possible for firms and parties to pre- contract to internalize any possible externalities of rule production. In other words, there are several mitigating forces that could reduce the externality problem for private rule-creators. Posner himself (in The Economics of Justice) points to another source of private rule provision: the evolution of custom. Cultural evolution gradually produced customary law in the absence of government supply; and despite some disadvantages, customary law and other "grown" orders have many attractive features. They draw on centuries of legal wisdom and experience; they are extremely stable; and they easily adapt to marginal social changes. The common law originated in the customary law of early tribes. Cultural evolution leads most societies down parallel tracks: divergent societies' customary laws of property, contracts, and torts are remarkably similar. And these evolved customs of property, contract, and tort are precisely the kind of legal rules that law and economics scholars consider efficient. All this points to a tentative conclusion: Leaving the production of legal rules to the market may be feasible after all. Markets supply many other non-patentable innovations. Posner and Landes are probably too swift to dismiss the possibility of non- governmental supply; after all, many industries overcome the problem of non-patentability. The potential for private supply turns out to be greater than commonly assumed. The following sections challenge this common assumption in depth. 3.2 The economic theory of private rule-making with no externalties As is often the case in economics, it is convenient to begin the analysis of private rule-making with a counter-factual assumption. Let us assume for the moment that all legal innovations' benefits can be internalized by their producer (whether an individual judge or the firm that employs him). Imagine a costless system of rigidly enforced copyrights. Any time someone uses a precedent, procedure, ruling, or law pioneered by another person, the user must pay royalties to the creator. What would happen under these circumstances? First, every judge or private adjudication firm would strive to develop the most efficient laws that it could find in order to maximize royalty revenue. They would funnel money into research and development, search out weaknesses in the law, and advertise the advantages of their respective systems. Just as with an ordinary copyright, the marginal cost curve shifts upwards in order to eliminate supernormal profits, because firms increase their spending in research and development and educate customers about the latest advances. This is perhaps the dreamland of law and economics scholars -- the efficiency of every conceivable legal rule could be tested, not in the arid world of econometrics, but in the practical realm of business. This market might have a host of possible structures. If Posner and Landes are correct, there are economies of scope in dispute resolution and rule-formation, so we would see "vertically integrated" adjudication firms performing both functions simultaneously. If they mistake an accidental feature of public courts in the U.S. for an economic advantage, then there might be two interrelated industries -- one resolving disputes for the public, the other selling rules to the dispute resolution firms. Just as in normal patent law, the discoveries of the distant past would belong in the public domain; only recent innovations would be rewarded by monopoly grants. All firms would probably embrace the basic categories of property, contract, tort, and crime -- innovations on par with Newton's work in physics or Darwin's in biology. Legal rules at this level of generality, being part of our accumulated cultural heritage, would naturally be free. But specific developments within these categories might be patented just like any other intellectual innovation. To be sure, there would be the usual problems in patent law concerning the breadth of the patent and its duration; but the core intuition is that if there were perfect exclusion it would be possible to have a flourishing law production market. Posner and Landes assert that private judges might have an incentive to promulgate vague rules that maximize the public's need for court services. It is difficult to imagine that any system could be more vague that our current public court system, so the argument seems misplaced. More to the point, this could not happen under free competition. Customers would patronize the arbitrators that offered the cheapest and highest quality service. Arbitrators would have to select the proper inputs -- most importantly, clear and simple legal rules. The incentives to promulgate vague rules and thereby increase demand might exist if the government sub- contracted its rule-making function to a single monopolist. But the genuine market in rule-formation that this section explores would be fully competitive. In such a situation, a firm that tried to "increase demand" by adopting vague rules would be much more likely to lose all of its clients than reap a bounty from an explosion of litigation. A competitive market in the formation of rules would be different from our current system of rule-formation by judges and legislatures in one crucial respect: there would be legal diversity within one and the same geographic region -- or if you prefer, product differentiation. Courts and legislatures (leaving the mitigating effects of federalism aside) usually make the same rules for everyone. Posner and Landes offer a plausible economic rationale: "there would appear to be tremendous economies of standardization in the precedent market, akin to those that have given us standard dimensions for electric sockets and railroad gauges."22 True enough; but the existence of some economies of standardization hardly shows that total standardization is economically desirable. Only when the production of legal rules became freely competitive could we correctly balance the preferences for both diversity and uniformity. Properly functioning markets do this without us even trying. Restaurants, for example, have many things in common: they are clean, have waiters and waitresses, serve non-poisoned food, and have bathrooms; but they have many differences: quality of service, type of food, decor. Presumably, restaurant entrepreneurs balance the costs and benefits of standardization and diversity when they make a final decision on their business practices, realizing that customers want both uniformity and choice. A competitive rule-formation market that strictly protected intellectual property rights would do the same. In fact, private adjudication (still assuming no externalities) might be better able to standardize law than governments; for it is a characteristic of state law that it applies merely within national borders. Yet as the world's economies become more interdependent, the need of international businesses for a common legal code increases. Private rule-formation could gradually amalgamate diverse business norms into a common legal code; for just as inventions from one nation can be exported to another, so too might legal practices. So long as law-formation remains a monopoly of governments, standardization requires international treaties, years of negotiations, and so on. And while such agreements claim to bring productive efficiencies like a shared commercial code, there is also the danger (as some critics of the European Community argue) that international standardization might eliminate legal competition on tax and regulatory policy. The economically preferable outcome would be legal standardization without a legal cartel. Prof. Bruce Benson sums up the "legal standardization" controversy aptly: "Where the tremendous economies of standardization that Landes and Posner alluded to exist, the private sector will take advantage of them. Government typically cannot respond because of the artificial constraints of political boundaries. There is no reason to believe that any national government is of the ideal size to take full advantage of the economies of standardization in law. In some areas of law (e.g., commercial law), these economies appear to be greater than any existing nation can encompass. In other areas, such economies may be considerably more limited so that existing political entities are too large. A private system of law would generate efficiently sized 'market areas'. "23 (emphasis added). These "efficiently sized market areas," I might add, need not be geographic in character -- they could just as easily be broken down by industry-type, by cultural norms, by risk-preferences, or any other factor. Since by assumption law would be produced in competitive markets without externalties, we could be confident that whatever market sizes and levels of standarization emerged would be optimal. 3.3 Rule-formation in the real world: the problem of externalties We have seen that a law-production market with no externalities generates rather attractive results. What happens if we relax the assumption of perfect patent and copyright protection for law-producing individuals and firms? As Posner and Landes first suggested, all legal innovation could disappear. Everyone might try to free ride on the precedent and procedure R&D of their competitors; but this individually rational strategy is not, as the saying goes, "socially rational." If there were no public courts, the probable result would be a steadfast adherence to tradition, because tradition would be the only source of law left. Put simply, we would have nothing but a static stock of legal knowledge inherited from the past. It is not clear how serious a defect this would be; as Posner hints, both the rule of stare decisis and the low rate of depreciation of precedents (4-5% according to his research) indicate that rapid innovation is not even desirable.24 Richard Epstein's mistrust of legal change is still more pronounced: "the merits of freedom of contract in no way depend upon the accidents of time and place"25. Other basic legal principles, he says, are similarly timeless. Unfortunately, legal innovation can alter good laws as well as bad ones. (One of Epstein's favorite examples is the doctrine of assumption of risk, which modern public courts seriously weakened.) Perhaps legal innovation is not as important as some people think. It is at least possible that the harmful changes made by the public courts in this past century outweigh all of the beneficial ones. Nevertheless, Posner and Landes' case against purely private production of rules seems strong. It does appear that patent and copyright protection would be difficult to bestow upon legal innovators. And normally economists assume that intellectual property rights of this kind are the only effective way to get the market to invest in research and development. Without effective ways to internalize the benefits of innovation, government funding seems like the only way to assure us of the supply of this valuable good. 3.4 Non-patentable innovations and externalties problems The central problem with the Posner-Landes analysis is that it proves far too much. It is a historical fact that private systems of law began, evolved, and thrived. The medieval law merchant, for example, was so effective that national governments adopted the market's rules rather than the other way around. Bruce Benson points out that during the 19th century private commercial arbitration competed fiercely with the common law courts.26 The former relied extensively on business custom and the law merchant, while the latter did not. One of the main ways that the private sector competed with the public was by adopting a different legal code -- one that the business community found more efficient and fair. Gradually, competition drove the common law courts to adopt rules similar to those of private adjudication. Competition persisted even though the common law courts were partly subsidized by the crown. Even Posner and Landes admit that, "the doctrines developed by the [private] merchant courts to deal with contract and commercial matters were absorbed into the common law and the official courts began winning business from the merchant courts. In a similar vein English procedural reform in the nineteenth century has been attributed in part to the competition from private arbitration."27 Notice how the private and the public courts competed by offering different rules of law, which would not be possible if public courts were the only possible source of law in the absence of patent protection for legal inventors. Prof. Benson raises a still more challenging point: Who free rides off of whom? Posner and Landes observe that public and private courts typically have similar legal codes; they then conclude that the private courts free ride off of the public courts by "borrowing" their precedents. But Benson argues for the opposite view: "[C]ommercial arbitration should enforce virtually the same laws recognized by the public sector, but the causation actually flows in the opposite direction. Public courts enforce virtually the same laws as commericial arbitrators do. If they did not, the public court bureaucracy would lose its commercial business because businessmen would use their own courts (as they are)."28 I doubt that this is wholely true today, because the private courts cannot compete with the public courts on equal terms. The public courts can trump the decisions of the private courts. Before we can test these alternative hypotheses, we would first have to give private courts the autonomy that they need to compete effectively. There is a second and more serious reason to doubt the seriousness of the externalties problem. Posner and Landes suggest that private rule-production would not work because of the difficulty of claiming a property right in a precedent. But they overlook the large sectors of our economy in which (a) Patent and copyright protection is unavailable, (b) There is no government R&D, and (c) Innovation still occurs. For example, the person who "invented" the modern super-market could not get royalties from other people who wanted to open up super-markets of their own, even though it was a remarkable innovation. General style and model changes, whether in autos, architecture, fashion, etc., cannot be patented. In order to be patentable, something must be relatively specific, but many useful innovations are not. Similarly, many business practices are not patentable: If, for example, I successfully experiment with a new form of vertical integration, my competitors are likely to copy me; but I cannot charge them for the privilege. And still such innovation survives. How is this possible? Part of the explanation is that economists overrate the importance of copyrights and patents. Businessmen surveyed typically rank the legal protection of their innovations as one of the least important determinants of the rate of return on research and development. The most recent survey of R&D executives yielded startling results: on a scale of 1 ("not effective at all") to 7 ("very effective"), 650 R&D executives rated the effectiveness of different ways to reap the rewards of innovation thusly: Method Ave. Score, Product Innov. Ave. Score, Process Innov. Patents to 4.33 3.52 protect duplication Patents to secure 3.75 3.31 royalty income Secrecy 3.57 4.31 Being first 5.41 5.11 w/innovation Move quickly down 5.09 5.02 learning curve Superior sales/ 5.59 4.5529 service efforts In a similar vein, F.M. Sherer, an economist who has written extensively about patent protection, notes several non-patent sources of protection for inventions: "natural imitation lags, the advantages of competitive product leadership, and the existence of non-patent barriers to the emergence of a competitive market structure."30 In the first category, natural imitation lags, Sherer includes several factors. First, there is secrecy on the part of the innovator. It may secretly prepare to introduce a new product, then spring it on the market; even with perfect imitation, it would still have a lead on its competitors equal to the time from the firm's decision to market the product to the firm's announcement of the product's availability. Second, notes Sherer, "It takes time for entrepreneurs to learn about a new and promising invention, even after its existence has been publicized, and it takes even longer for them to decide it is worth imitating."31 Many innovative processes, he notes, take years to catch on completely in an industry. A high price gives short-run profits and swift imitation; a low price gives long- run profits and slow imitation. Finally, imitation is often very difficult unless one duplicates the first-mover's R&D, or hires away former employees of the innovating firm. As Scherer and Ross explain, "free riding on an innovator's technical contribution is often far from free. At one extreme in this respect are airliners. One can inspect a rival's design, but to build a similar aircraft, one must generate detailed engineering drawings for each part, program machine tools to produce the parts, build prototypes, andr subject them to static and dynamic testing to ensure that unnoticed design flaws do not lead to catastrophe."32 Next, says Sherer, the first firm on the market frequently wins an excellent reputation. Image and consumer trust matter in business; innovation enhances both. In a parallel fashion, uninnovative firms' image and consumer confidence suffer. Sherer's final point is that the lack of atomistic market structures enhances the gains from innovation in the absence of patents. Supposedly, oligopolies with their strict price discipline manage to prevent imitation and profit erosion. I find this third point difficult to take seriously in the light of later work in industrial organization by Demsetz, Bork, and other critics of the structure-conduct- performance paradigm in industrial organization.33 Anyway, there are strong reasons to think that the market structure of private rule-suppliers would be fairly unconcentrated; even now, while the market is fairly thin and undeveloped, many suppliers co-exist. But Sherer's first two points are quite pertinent. Patent protection is not the only spur to innovation; if it were, we would see no innovation in areas with no patent protection. Apparently, one of the reasons why unprotected innovations continue to appear is that legal protection is not the primary incentive source. Given this, why couldn't we spur the production of rules with the same incentives that inspire other non-patentable innovations? 3.5 The evolution of custom as a substitute for intellectual property rights F.A. Hayek observes that there are many sorts of "grown orders" in our society. "Although there was a time when men believed that even language and morals had been 'invented' by some genius of the past, everybody recognizes now that they are the outcome of a process of evolution whose results nobody foresaw or designed."34 The most obvious example is the market economy, but other valuable cultural features like language, custom, and law fit the same description. In fact, some law is a subset of custom. Basic private law categories like property, contracts, torts, and crimes exist even in primitive societies without any permanent government. One interesting challenge for economic analysis is to explain what incentives cause the development of language and custom, both of which look like public goods. If we could figure out what incentives there are to produce language and custom, we would understand how customary law, one important branch of custom, arises and evolves. If we could do that, then we might see if these incentives still exist today. Richard Posner, the very scholar who has criticized private production of legal rules, has coincidentally also studied the legal systems of primitive societies. He does not deny the plain fact that they have legal rules without government supply. But at the same time he gives no clear account of the primitives' incentives to create rules. "Two of the common sources of legal norms, legislation and executive decree, are ruled out by the assumption introduced in the last chapter that there is no state. Since the arbitrators, though private, are a sort of judge, it may seem that the third common source of law - judicial decisions viewed as precedents guiding future conduct - could operate in a primitive society. But. one has still to ask what incentive the arbitrator has to issue opinions that will stand as precedents."35 Posner concludes that custom is the only alternative source of law; but this merely pushes the question one step back. What incentive there is to create customs? This question is especially pressing for Posner, who argues that even the bizarre laws of primitive societies have an economic justification. Who has the incentive to produce these detailed legal customs, especially customs of such remarkable efficiency? If we looked at customary law as a final product, the puzzle is difficult. Who would have the incentive (much less the ability) to design a body of efficient customs for his entire clan to use? And how would these customs gain general acceptance in the absence of a government? So long as we look as customary law as a final product, it is difficult to see any solution. I do not have a fully satisfactory explanation, but I think that the most promising route begins with the observation that customs arises piece-meal by a process of cultural evolution.36 When an individual conceives of a radical change in social norms and tries to promote it, his effort is indeed a public good. But suppose that he thinks up a tiny innovation, useful mainly to himself. He then has an incentive to initiate the change. Since the change is so minute, people are less likely to object to it. The better the innovation is, the more people imitate and spread it. The broader order is a sum of these piece-meal innovations. Each tiny improvement benefits the innovator at least somewhat -- though there may be externalities, the innovator gets part of the payoff. But the final product of these tiny improvements benefits everyone, even though it would not reward any individual to develop an entire system from scratch. This process is evident in the birth and growth of language. No individual would have an incentive to create a language from scratch. But it would pay to (a) Learn whatever language exists already, and (b) Make marginal improvements and innovations in one's language, which others may in turn adopt. Apparently, this simple process led speechless primitives to slowly evolve common languages, which in turn became modern languages. Only by understanding that language develops piece-meal and gradually could we see how it could arise without government supply or subsidization. What can we say about the efficiency characteristics of these evolved practices? No rigorous theorem, analogous to the results of welfare economics, proves conclusively that they are maximally efficient. But they nevertheless have many attractive features. First, spontaneous orders create the central features of society, such as language and custom; and it is difficult to imagine that these features could have arisen by design. Second, spontaneous orders have a cheap feedback mechanism. If something is inefficient (always or sometimes) then cultural evolution helps weed it out; if efficient, then cultural evolution helps it spread. Designed orders, in contrast, often lead to disaster before their inefficiency becomes clear. Third, spontaneous orders are a cheap route to social coordination: complete strangers can interact so long as they share a similar cultural background. Fourth, they are extremely stable. I suspect that this is one of their chief drawbacks in the eyes of critics; but perhaps they overlook the dangers of hasty and radical social changes and see only the costs imposed by stagnant traditions. Studying the deadly history of social experiments in the 20th-century yields a more balanced perspective .37 The most reasonable explanation for the birth and evolution of customary law is that it too is a spontaneous order. As with language and custom generally, there were no patents or copyrights to give incentives to develop customary law. But because customary law accretes piece-meal, patent protection was not necessary. The detailed legal codes established purely by custom that Richard Posner discusses indicate that gradual accretion works at least tolerably well. Similarly, the somewhat more civilized Germanic tribes had a sophisticated body of evolved customary law.38 While there was not vigorous R&D in precedents that perfect patents would foster, neither was there the pathological free-riding that Posner and Landes fear. The development of rules via cultural evolution is hardly limited to ancient history. Robert Ellickson, in his recent book Order without Law, discusses the informal norms that govern the ranchers and farmers of Shasta County, California.39 Due perhaps to the costliness of legal disputes, the people of Shasta County are fiercely opposed to litigation. They prefer to resolve disputes informally, in accordance with a simple and unbureaucratic code of rules. These rules frequently differ from the official law -- but no one seems to know or care. Here is a list of the basic rules of law that people in Shasta County live by, all of which were the product of gradual accretion via cultural evolution rather than a single law- giver or body of law-givers. 1. Owners are held strictly liable for all damage done by their livestock.40 2. Ranchers who build or improve cost-justified fences are entitled to bill a share of the cost to noncontributing neighbors.41 3. Aggrieved parties must try each of the following remedies in succession: (a) Give notice to the accused; (b) Circulate truthful negative gossip regarding the deviant person; (c) Physically seize or destroy an appropriate amount of the offender's assets.42 4. Parties are expected to forgive trivial injuries, unless persistent.43 5. Parties should not take their disputes to the government courts except under extreme circumstances. As Norman Wagoner of the Shasta County Board of Supervisors colorfully puts it, "Being good neighbors means no lawsuits."44 Ellickson thinks that the relative efficiency of this informal legal code explains its persistence and vitality. It is cheaper to obey simple customs than to learn the law. And the norms themselves are cheap to enforce on account of their simplicity and the informality of the remedies.45 Ellickson, it should be noted, also sees problems in systems of informal norms -- most notably, they may fail to protect the interests and rights of those outside a given community. That is certainly a danger. Still, many norms that arose within close-knit communities later spread to advanced societies. Why are critics of private rule creation via custom so sure that it doesn't suit the modern world? Posner argues that customary law works for static, primitive societies but is inadequate in dynamic, modern societies: "There is little danger that legal change will lag behind social change in such a society and produce the sort of anachronisms which in the case of English common law and Roman law created the demand for legal fictions, equity, and legislation to keep the law up to date. Evidently Roman and English society were changing faster than a system of purely customary law (no fictions, no equity) could keep up with - which means faster than the typical primitive society."46 But it is not so obvious that modern societies have more need of legal change than primitive ones. Richard Epstein, for one, argues in a forthcoming book that it is precisely complex, modern societies that need a small number of simple and constant rules.47 In primitive societies, life is static enough to create new rules each time a change springs up. But modern societies are so complex that the best that we can do is provide a few clear and stable rules to build a framework for rapid non-legal change. As society grows more intricate and vast, it becomes less, not more, feasible to modify the law each time society changes. Robert Ellickson brings up another startling point: public courts often enforce customs which were the product of cultural evolution rather than the state. "In both its original broad outlines and its current substantive details, the Uniform Commercial Code frequently gives legal status to the usages of merchants."48 The public courts free-ride on the market's rule-creation! Evolved norms aren't perfect, but they have marked advantages over other feasible methods of rule-creation. Hence, judges and law-makers, despite their traditional hostility to private rule-creation, give custom a big place in the law. Posner's own work on the legal systems of primitive societies partially answers his critique of private rule-creation. Primitive legal systems grow out of custom in the absence of government, as Posner points out. So if intellectual property rights won't work, custom might. Is custom only suitable for primitive societies? Hayek's theory of spontaneous order and Ellickson's research on informal norms suggest the opposite: A single planner could adequately give rules to a simple society, but not a complex one. Cultural evolution of law is another possible route for the market to build its own rules of law. 3.6 Ways to internalize the benefits of rule-production It is surprising that Posner, who writes at length on entrepreneurs' clever solutions to difficult incentive problems49, dismisses private rule-creation so swiftly. Prof. Benson points out a plausible way that judges might internalize the benefits of rule creation: "[M]aximizing profit does not involve maximizing the number of cases decided. A judge who provided clear rules and opinions would command a relatively high price for contracts with various organizations. Once under contract, the judge would actually have the incentive to minimize the number of disputes that go to trial by making his rules clear - that is, by setting precedents. Under this scenario, private judges have precisely the opposite incentives to those predicted by Landes and Posner."50 As mentioned previously, parties with a prior contract to arbitrate can easily opt out of the public court system. Hence, a large majority of arbitration consumers could indirectly pay for efficient private legal codes by hiring the firm most noted for keeping peace among its clients. This solution, notes Benson, closely parallels the structure of HMOs. "In the area of medical care, some have argued that physicians have incentives to 'create demand' for their services by advising uninformed consumers that they need more medical care than they actually do. This is quite similar to the argument Landes and Posner made regarding private judges. But new institutional arrangements have been created during recent years in the face of rapidly rising costs of medical care. Consumers can now play a flat fee for services at HMOs and PPOs, reversing the physicians incentives. Physicians have incentives to keep their patients healthy with preventive medicine that avoids more costly treatments after illness arises."51 It is the incentive to supply "preventive law," -- law which keep disputes from happening in the first place -- that public courts and legislatures rarely supply. Competing arbitrators would have a strong incentive to select rules that do precisely that. But what if firms waited for their competitors to invest in research and development, see which firm is most successful, and then copied its rules? Perhaps problems of this sort lead the VISA corporation to keep its legal code secret. A firm might keep its code secret and merely advertise its clients' low dispute rate; after all, customers care about the bottom line, not the production process. This is an imperfect solution, but despite difficulties, private arbitrators have some incentive to adopt and devise prophylactic rules to prevent disputes. The same cannot be said for public courts or legislatures. As mentioned before, innovation can survive without patents. And if firms were to compile their legal rules in books and copyright them, their innovations would enjoy moderate protection. Small competitors could still free ride, but large firms would fear copyright infringement lawsuits. Just as publishers ignore isolated acts of unauthorized reproduction but pursue large-scale violations vigorously, arbitration firms might ignore small-scale abuse but prosecute plagiarism of their major competitors. Since arbitration firms are in the business of preventing and resolving disputes, they might even have the good sense to strike bargains to compensate each other if they borrow from each other's rulebooks. The effectiveness of this depends on the concentration of the arbitration industry; but as Ellickson's work hints, cooperation often blossoms even in large communities so long as members repeatedly interact. 3.7 The variety and flexibility of private rule production As mentioned in the introduction, we usually think of law as a non-exclusive good in two senses. The first is the trivial one that we all benefit from law whether we pay for it or not. That is hardly true of private adjudication, as we saw. But there is a second sense of non-excludability: Opting out of the legal system could not be permitted even if it were possible, because the whole point of law is to encompass everyone, not just people who want to live lawfully. If we could unilaterally renounce legal responsibility, who could resist the option? Since law appears non-exclusive in this sense, it is simply not possible to have the variety and flexibility of normal markets. When parties have the option, however, to jointly opt out of the state's legal system, variety and flexibility are possible. Suppose that you employ me; in my employment contract, we can specify any arbitrator we like. If the market is sufficiently thick, we should be able to open up the phone book or a professional journal and find hundreds of different sorts of legal rules advertised: no- fault or strict liability, damage or injunction remedies, complex appeals systems and one-hour binding arbitration. There might be firms that offer all sorts of options: arbitrators who make house- calls, over-the-phone arbitration, whatever you can imagine. All of these options could co-exist in one and the same geographic region -- just like normal competition. Law and economics scholars write as if one rule, in principle, were the best for a given society. And if law equals public law, it is difficult to get outside this assumption. But once we let people opt out of the legal system, we see the error. There are as many efficient rules as there are possible sub-markets. Rather than doing econometric analysis of the efficiency of rules, we might simply free up the market and see what happens. It would be miraculous if it turned out that one set of rules were right for everyone. Instead we should expect to see the variety and flexibility characteristic of other markets. Indeed, what are various econometric tests but highly imperfect attempts to find out what the market would have chosen? Isn't there a false presupposition here -- that the market would have chosen one and only one legal code? One of the chief efficiencies of private rule-production would be the opportunity for consumers to pick and choose from a host of possibilities. At the same time, law and economics scholars would have a new field to study; instead of theorizing about what is efficient, they could go out and observe what is efficient in each sub-market -- then explain why it is so. 3.8 Public vs. private rule-creation: an exercise in comparative institutions Unlike some economists, Posner and Landes realize that one must do more than point out that an industry has a few market failures in order to justify government supply. It is equally necessary to investigate government failures and see whether they are comparable or worse. In "Information and Efficiency: Another Viewpoint," Harold Demsetz makes this point explicit by contrasting the "comparative institution" to the "Nirvana" approach to policy.52 The former, said Demsetz, considers the real-world advantages and drawbacks of both markets and government; the latter focuses exclusively on the failures of the market and assumes that the government can correct them at no cost. With this in mind, let us draw some tentative conclusions about the economic desirability of the private supply of rules of law. The Posner-Landes criticism of private supply of rules was that the production of rules is a positive externality; hence they doubted that markets could do the job. This was their justification for the rule-making function of the public courts. In light of the previous sections, what may we conclude about this complaint? Primarily, Landes and Posner overstate the point. There are often ways to internalize the benefits of private rule-creation. (See 3.6) History gives ample evidence that extensive legal systems can arise despite externality problems, largely by gradual accretion. (See 3.5) And while intellectual property rights are one spur to innovation, they is not the most important one; indeed, they they were, all unpatentable innovation would grind to a halt. (See 3.4) Nevertheless, public courts could in principle solve the externalities problem better than private firms could. Since they get their funding from taxes, they could produce rules of law whenever their social benefits exceeded their social costs, even if it were impossible to charge the beneficiaries. Under private rule creation, rules are unlikely to get produced unless it is feasible to collect payments from users. Yet the argument doesn't stop here. In part two of this paper, I pointed out the public courts' inefficient methods of dispute resolution. Their ability to produce good rules of law is no greater. Public judges have bad incentives; since they are either elected or appointed, they are likely to have a different attitude toward work than employees in a profit-making firm. Posner suggests that public judges derive utility from issuing precedents purely from the pleasure of influencing the course of the law.53 (Why mightn't this austere intellectual pleasure not be found in private sector judges -- perhaps on their off-time?) He says that judges clearly don't maximize leisure. Judge Neely's assessment of the judicial work- ethic is less sanguine than Judge Posner's ("It is in the nature of the judiciary, with its life tenure or long elected terms, that it can encourage arrogance and indolence, just as the occupation of salesperson tends to mask them."54) but that is beside the point. The question is not whether public judges' incentives get them to produce precedents; clearly they do. The question is whether their incentives lead them to efficient rules; and that is far less likely. Posner himself suggests that judges advance their own political agendas, heedless of the possible negative consequences. If this is the case, then each successive precedent might be a source of external harm rather than external benefit. Since private judges have better incentives than public judges, we should expect their decisions to be more efficient. Public employees can afford to pursue their own agenda; they are hardly likely to lose customers, their salary is fixed, and their jobs are among the most secure in the world. What motivation do public judges have to create rules that benefit everyone, when they could promote their own political views instead? Private judges might have equally strong convictions, but they face the constraints of competition. Their clientele is not stuck with them, and is unlikely to suffer from inefficient rulings when they can switch their arbitrator freely. And suppose that public judges did want to select efficient rules. How would they know when they had? Law and economics scholars can hardly agree among themselves except in the most egregious cases of legal ineptitude. Knowing whether a legal rule is efficient must be difficult. The problem of public judges is that they have no effective feedback mechanism to tell them if they are doing a good job. Even if they wanted to pick efficient rules, they have no good method to pick them out. Compare this to market-produced rules. The knowledge is much easier to get, because consumers can express their preferences by switching suppliers. This gives a simple and pragmatic test for efficiency. Imagine randomly creating a hundred different legal codes for a hundred different arbitrators. You could easily find out which codes were efficient -- just see what thrives in the market. That would give far better evidence for a rule's efficiency than volumes of scholarly speculation. On top of this, as section 3.7 pointed out, the quest for the efficient legal code is vain because there is no uniquely efficient code for everyone. Instead there are a host of efficient codes, each of which works best for a given sub-market. Public judges' quest for efficient law fails because they can't take this into account. Perhaps the way to put the criticism is this: While public courts could potentially correct the allocative inefficiencies of private rule creation, they are almost sure to be less productively efficient. Moreover, the allocative inefficiencies of private rule creation are partially correctable. History supports this, since many legal rules originated in the private sector. Yet it is unclear how the public sector could ever attain the productive efficiency of the private sector. While some quasi-market reforms could be imagined (e.g., Set scarcity prices for court services), the fundamental incentive and knowledge problems of the public sector remain. Richard Posner is famous for his view that common law decisions by judges are much more likely to be in the general interest than those of legislatures. Legislators' chief activity is redistributing wealth; they do what interest group pressure and expediency dictate. Common law, in contrast, provides efficient rules that encourage production and trade, both of which enhance social wealth rather than merely redistributing it. Posner was not the first to endorse this view; one may find a similar theory in Bruno Leoni's Freedom and the Law: "[T]here is more than an analogy between the market economy and a judiciary or lawyers' law, just as there is much more than an analogy between a planned economy and legislation."55 "[A] legal system centered on legislation resembles in its turn. a centralized economy in which all the relevant decisions are made by a handful of directors, whose knowledge of the whole situation is fatally limited and whose respect, if any, for the people's wishes are subject to that limitation."56 Though his analysis was less detailed than Posner's, in one respect Leoni is more sophisticated. Namely, he notes that legislators and public judges both have the same incentive to act against the commonweal. Both have a public monopoly whose power isn't checked by market competition. As Leoni explains, "it cannot be denied that the lawyers' law or the judiciary law may tend to acquire the characteristics of legislation, including its undesirable ones, whenever jurists or judges are entitled to decide ultimately on a case."57 If Leoni's point is valid, then it Posner's enthusiasm for the common law is premature. It may yield better decisions than legislatures, but retains its basic flaws. A chief point in favor of market creation of rules is that it is genuinely competitive. The arguments that favor judge-made law over legislation also commend private rule creation over the common law. For example, the argument might be made that legislation is too erratic and unstable, while judge-made law has long-run certainty. Yet political motives in the public courts sometimes produce erratic and unstable case law. Products liability, anti-trust, and civil rights law come to mind. Private rule-producers would never a make sudden change unless they were convinced that most of their clients wanted it. Another argument for common law is that it enforces what the parties themselves would have accepted if bargaining were feasible; critics of legislation, in contrast, allege that it is usually the tool for the politically powerful and well-organized to grow rich at the expense of the disenfranchised. Yet public courts also bend the rules to help their favorite groups -- again, products liability, anti-trust, and civil rights law come to mind. Private rule producers would be more impartial, because they clients must jointly choose them. Posner and other scholars also argue that the common law tends to be efficient, but have been unable to persuasively explain why. In The Economic Analysis of Law, Posner suggested that judges simply have a utilitarian instinct.58 Casual observation of judges' philosophies strongly disconfirms this. Other thinkers have pointed out different market-like forces that push common law rules toward efficiency, but Posner and Landes persuasively criticize these.59 The preceding analysis of the private production of rules has two interesting implications for this debate. First, it seems that only genuine market competition systematically yields efficient rules, because private firms escape the incentive and knowledge problems of the public courts. The discipline of competition plus the feedback mechanisms of profit and loss drive arbitrators to pick the rules that consumers want. Second, perhaps one explanation for the efficiency of public court rules is precisely that potential market competition checks them; if public court rules become too inefficient, firms may shift en masse to the private sector. Posner and Landes themselves note this process in the history of English law, and suggest that this might explain the public courts' hostility to private alternatives. Consider a final argument for a greater role for private sector rule-creation: It insures us against dangerous social experiments by diffusing power among competitive bodies. It is risky to let one well-organized group -- like public judges -- choose for a whole society. When most public judges decide to adopt a new rule, all of us bear the costs of their errors. Experimentation by the public sector is costly because its power is centralized and its feedback mechanisms weak. The private sector would give us valuable insurance against the public courts' mistakes. Private firms are innovative, but they sensibly introduce changes gradually. And there are typically many suppliers in a given industry -- so even if one firm unwisely leaps into a new practice, consumers have alternatives. While this benefit is political rather than economic, we should keep it in mind. What policy implications does this analysis have? As mentioned in the section on dispute resolution, private arbitration works mainly if the parties have the opportunity to pre-contract with each other to opt out of the public system. And for opting-out to work well, private alternatives must be independent of the public courts. The same principle applies to private creation of rules. So long as both parties pre-contracted, the courts should categorically refuse to overrule an arbitrator. The public courts might also quit writing precedents for cases that private arbitrators dominate. Tax-subsidized competition dulls private initiative. Consider commercial disputes -- an excellent candidate for privatization at the dispute-resolution level. Since private arbitrators handle such cases already, the courts could spur private production of rules by ceasing to produce them themselves. Courts could encourage private settlement of disputes between strangers by renouncing jurisdiction over any case that both parties agree to arbitrate. If almost everyone waives their right to a court trial in a given area of the law, the court might cut back on its own rule- production in that field. By a process of trial and error, the courts might come to see what types of legal rules the market will supply. Why does private arbitration break down when pre-contracting is not feasible? The answer is that the right of violent enforcement rest with the state, so private bodies must rely on subtle enforcement techniques such as ostracism and boycott. The next chapter discusses the strengths and weaknesses of private enforcement of law. Since the right to use force is what limits the universal extension of non-state legal systems, it also considers what would happen if this final limit disappeared. [To continue, go to Part II.]