II. Individual Rights in the Minimal State

What, then, are the rights of individuals that the minimal state is bound to protect? I do not believe that Nozick had any direct engagement with the system of common law rules,⁵

but his articulation of the minimal state draws heavily, in practice, from the ordinary system of common law rights and duties.⁶ The initial building block of this system is individual autonomy, construed as the right of self-control free from the interference of others. Clearly, Nozick's mission would abort at the outset if the definition of autonomy were extended beyond self-rule to self-rule plus some minimum level of welfare guaranteed to all (although it is never clear by whom). The second building block or piece of the puzzle allows individuals to acquire ownership of property that is external to themselves, most notably land and chattels, or what Nozick terms “justice in acquisition.” Nozick does not address any of the special problems raised by those resources, such as air and water, that have generally been regarded by legal systems as held in common. Further, he does not deal with the special problems raised by common carriers (e.g., railways, electricity, and power); nor does he address systematically the modern concerns with intellectual property, such as copyrights, trademarks, and patents.⁷ A comprehensive theory of the minimal state would be required to explain whether and, if so, how these various common laws and statutory property-based institutions survive. Next, Nozick notes that the rules of transfer (contract and conveyancing) are critical to the overall operation of the system insofar as they allow for the constant recombination of initial rights, either to labor or services, by the individuals who hold them. This point is made clear when he notes that “[n]o one is entitled to a holding except by (repeated) applications of 1 [justice in acquisition] and 2 [transfer].”⁸ Special attention has to be paid to the word “repeated,” which does not play any role in the singular events that count as original acquisition, but does enormously expand the utility of justice in transfer. The repeated application of this one simple rule means that the inputs acquired in one transaction may, if not consumed, become, in whole or in part, the outputs of the next transaction. The constant repetition of this one strategy makes contract a tool of tremendous flexibility and introduces a versatility in transactions that counts as one of the prime ingredients of any social system. From individual rights can emerge a complex set of consensual arrangements through partnerships, leases, bailments, loans, and the like. All of these are created in ways that improve the utility of the parties to the agreements, without treading on the rights of third parties. Nozick clearly appeals to this repetitive feature of contractual life to explain how real voluntary agreements could solve the problem of creating a protective association, capable of defending its individual members (of full age and competence).

Finally, he defends what should be regarded as a variation on the law of tort. The individual's rights to one's person and to one's property (whether that property is acquired by initial acquisition or subsequent purchase) must be protected from the incursions of others, whether by force (including the threat thereof) or fraud. Nozick does not speak of this function as “protection” but as “rectification,”⁹

where the difference in emphasis matters. Rectification, literally understood, means to make right a (completed) wrong by supplying redress that restores the wronged party to the status quo ante, to the extent that the law can do so. Protection has a somewhat broader meaning, which includes the prevention of wrongs by allowing the interposition of relief before the fact either by state action or by private injunction. This broader meaning is in part a response to the long-standing debate over whether the primary function of the tort law is deterrence, compensation, or some mix of the two. Dealing solely with matters of rectification eliminates all levels of uncertainty in the picture; but the decision of whether or not to allow preventive relief necessarily requires a balance of two kinds of error: injunctions when they are not warranted, and no injunctions when they are clearly needed. That unavoidable uncertainty does not square well with a moral theory whose sole focus is the delineation of both right and wrong human actions, and Nozick does address these questions in the conventional terms, noting that any society, capitalist or socialist, “should permit those polluting activities whose benefits are greater than their costs,”¹⁰ without noting the tension between his analysis and the stark theory of justification that he defends later. For these reasons, I shall use the broader term “protection” where Nozick would speak of “rectification.” But even after these difficulties on terminology and remedies are duly noted, Nozick's basic position resonates quite closely with the general rules in Hume's A Treatise of Human Nature (1739).¹¹ In turn, Hume's position appealed quite consciously to the classification of legal rights articulated in Justinian's Institutes, and has remained, notwithstanding details in application, the mainstay of all common and civil law systems since that time.¹² Nozick cared little for any demonstration of how his own position tied in with the general conception of legal order as it developed down through the ages. But make no mistake, part of the appeal of his system is that it resonates with a set of legal rules, with small variations in detail, that are in common use in primitive and advanced cultures. Nozick's universalist argument benefits from the fact that there is much less variation in basic legal rules across time and space than is commonly acknowledged—at least before the rise of the modern regulatory state. He does not write in a void.

III. Enter Forced Exchanges

This close concordance between common practice and high philosophical theory is sufficient in my mind to show that Nozick is on to something that matters in his appeal to individual rights. But it is one thing to say that these rules of autonomy, acquisition, transfer, and protection (which includes both injunctions and rectification) are indispensable constituents of any legal order, and quite another thing to say that they alone offer the needed ingredients to constitute a full-fledged state, let alone one that uniquely satisfies the constraints of strong moral theory. As I have noted, I think that in the end Nozick is instructively and famously wrong on both counts. His austere account of the minimal state fails because it offers no credible explanation as to how the minimal state is formed. Yet once we add even one additional premise to the initial argument, it no longer follows that the minimal state is all that political morality allows. For example, Nozick's view would prevent the state from supplying what have been commonly termed public goods—namely, those goods (such as streetlights and national defense) that must be supplied to all members of a community if they are to be supplied to any individual member. This position can easily result in situations in which everyone is worse off in a world without taxation than they would be in a world with a well-designed system. In his classical exposition of the public goods issue, the economist Mancur Olson demonstrated that voluntary provision of these goods would result in their systematic underproduction, as no rational self-interested person would agree to shoulder the burdens of providing goods that he would receive even without his contribution. Instead, each member would want to “free ride” off the efforts of others. That strategy of nonsupport could be adopted by all individuals simultaneously, so that goods that everyone wanted no one would supply—a perfect rendition of the standard prisoner's dilemma game.¹³

To be sure, this model is overdrawn since some individuals do act out of a sense of moral obligation even in collective situations.¹⁴ But in large and anonymous settings, we can be confident that any level of compliance will, given individual self-interest, be far lower than optimal.

Nozick is aware of the difficulties of coordination, as he demonstrates in his discussion of the possibility of class action lawsuits against polluters,¹⁵

but he overestimates the ability of clever voluntary solutions to overcome the standard public goods problem. The rules of property, contract, and tort, applied in combination, do not solve the problem. What then is the missing piece of the puzzle? Here my tack, as it has been for many years,¹⁶ is to insist that the one missing piece of the puzzle is the practice of forced exchanges, initiated either by private parties or by the state. These are exchanges in which someone is required to surrender either liberty or (most commonly) property in exchange for compensation in cash or in kind. They are often described as the use of the state's power of eminent domain or condemnation, but in fact the issue goes far beyond the simple case where the government takes land for a road or a military installation. It also includes what are now termed “regulatory takings,” which cover all government rules that limit the right of an owner to use or dispose of his property as he pleases, without dispossessing him from the land in question.¹⁷ In some instances, these regulations are regarded as permissible without payment of compensation, but in others they are not. Deciding which forms of regulation are permissible without compensation and which are not presents some of the most devilishly difficult questions that arise in the operation of any mature legal system.

It is perhaps useful at the outset to give a simple catalogue of some of the common types of forced exchanges, as understood in opposition to an outright taking, where property is taken by the state which supplies nothing at all in exchange. The simplest version of a forced exchange is outright dispossession in which the state condemns a parcel of land for which it then pays the fair market value in cash. The taking of the land is forced on its owner, but it does not count as a form of theft because cash compensation is provided in exchange. The theory of the practice is that the landowner will be left (if compensation is calculated correctly) as well off as he was before the taking took place, while all the members of the society (including the individual whose property is taken) are left better off because the value to them of the land in its new use (say, as part of a road instead of a farm) is greater than the cash that they have been taxed to acquire it.

On this analysis, the taxes themselves are a form of taking (which is apparent to anyone whose property is seized if the taxes are not paid), which standing in isolation leaves all individuals who are taxed worse off than they were before. But if the benefits of good order or good roads acquired by taxes leave these taxpayers better off (by their own subjective lights) than before, this taking of cash is compensated in kind by the goods and services the government provides. In other takings situations, cash need not enter into the equation at all: the state could impose restrictions on how one person uses his land for the benefit of all his neighbors, and could impose parallel restrictions on the neighbors that give him benefits in exchange. In these regulatory takings, all persons could be required to suffer low-level nuisances (e.g., hearing voices from a neighbor's house) in exchange for being allowed to engage in similar activities that create similar low-level nuisances for others. If this live-and-let-live pattern is adopted generally, the alteration of the rule of strict liability for invasive behavior improves the situation of all against all.

There are in fact many situations, both voluntary and coercive, where all the interests exchanged are property, not cash. The usual planned unit development (condominium or gated community) is organized by a set of (repeated) voluntary transactions that satisfy Nozick's most exacting tests for justice in acquisition and transfer, and yet no cash is involved; each purchaser of a unit is subject to restrictions that cost him, say, the equivalent of $100, for which he receives in exchange the benefit, worth perhaps $150, of like restrictions imposed on others. Various zoning ordinances, when properly constructed (which is usually not the case), could replicate that pattern of distribution when high transaction costs prevent people from being able to deal with these situations voluntarily.

Nozick touches on these complex situations in chapter nine of Anarchy, State, and Utopia, “Demoktesis,” which asks how democratic institutions might justify more than the minimal state. But here his lack of legal knowledge leads him astray. At one point, he notes that it is possible for neighbors to buy the right to determine the color of a neighbor's house,¹⁸

but he does so in the guise of an isolated transaction and not as a part of some comprehensive plan. Accordingly, he fails to see how these transactions typically govern all plots of land, reciprocally and simultaneously, and are often imposed by operation of law, and not by voluntary consent. He then carries the discussion further by speculating how it might be possible for each individual to acquire fractional interests in the labor of other persons. Such transactions do in fact take place in voluntary markets, as when a young golfer agrees to divide his future earnings with a set of backers who provide him with a stake that allows him to get established on the professional tour. But these arrangements are neither universal nor reciprocal. Everyone does not take a part-interest in the future earnings of all the others in the group. Nozick's effort to generalize on this theme is instructive, however, because it helps explain a scheme of general income taxation, whereby the state allows individuals to pursue whatever occupations they choose, and then taxes their earnings in accordance with a fixed (preferably flat) tax,¹⁹ which is justified, as noted above, by the return benefits that individuals receive in a well-functioning state. But even here Nozick does not quite want to allow taxation into the system, even on a proportional basis,²⁰ which distinguishes his position from that of other classical thinkers such as John Locke,²¹ Adam Smith,²² and Friedrich von Hayek,²³ all of whom gravitate in that direction.

Nozick is mistaken in his treatment of taxation, which looks solely at the cost and never at the benefit side. Here, and in other contexts, he fails to see how allowing forced exchanges introduces an extra degree of (analytical) freedom into the overall analysis, which requires close attention. It is now possible to have, as it were, justice in forced transactions so long as the party who is subject to coercion is left better off than before. This is not an effort to introduce into the law some Rousseauian notion that individuals may be forced to be free, for the use of forced exchanges is not designed to alter or override the preference structures that ordinary individuals hold. It is only intended to allow them to move to higher levels of utility (albeit with lower levels of political freedom) than they could achieve through voluntary transactions in light of the well-known coordination problems that arise in the provision of public goods. The full range of examples should make it clear how risky matters become once forced exchanges are allowed into the system. Clearly the permissible domain of forced exchanges should not be infinite, and one advantage of Nozick's more limited rules of acquisition, transfer, and protection is that they preclude petty and tyrannical abuses from the outset, abuses that often take place when the state uses its eminent domain power. But the price of his parsimonious assumptions is to block any coherent account for the legitimacy of the state at all. The key challenge, therefore, is to develop a set of rules that permit some forced exchanges while guarding against the potential for systematic abuses, including the forms of redistribution that were the subject of Nozick's attack. Within the American constitutional tradition (which speaks to concerns that transcend America's borders), this effort is dominated by two conditions that require some attention.²⁴

The forced exchanges in question should be made with “just compensation,” and should only be allowed in those cases where they are for “public use.” This formulation is suggestive of and consistent with the following set of conditions for allowing forced exchanges under the eminent domain power.

First, the transaction costs of voluntary arrangements must be high if not prohibitive. This condition alone suggests that it is not appropriate for me to take your watch while leaving its replacement cost in your mailbox. There are many individuals who can supply watches through ordinary market exchanges; why take the risk of abuse when the voluntarist solution is so attractive? It is clear that the public use requirement precludes this transaction, even if it must allow many others. Rightly understood, the function of this requirement is to allow the state to supply public goods when the transaction costs of voluntary arrangements among multiple parties are prohibitive.

Second, the individual whom the state coerces must receive compensation that leaves him at least as well off by his own lights as he was before the forced exchange was undertaken. The point here is that coercion is justified not by some abstract appeal to the good of the community at large, but only by a specific showing that it causes no harm (or perhaps even works some net benefit) to the very individual who bears the brunt of that coercion. The requirement here raises obvious questions of valuation, but the reference to “his own lights” is intended to highlight the role of subjective value in the context of coerced transactions as another barrier against the abuse of state power. This requirement raises, in turn, some serious issues of measurement that cannot be escaped once any forced exchanges are allowed as legitimate.

Third, the transaction should work to the net benefit of those individuals who acquire the property in question. This condition will be easy to satisfy when ordinary individuals are able to use the coercive power of the state for their own advantage. They will only initiate transactions from which they hope to benefit. But the requirement is much more demanding in those cases when government agents act for the benefit of the community (against one of its members), using revenues that they have obtained through coercive means, for example, taxation. In one sense this position seems to be in tension with the view that all government coercion should be exercised for the benefit of those to whom it is directed. But in fact that condition is not violated here, because the willingness to provide full compensation means that the state power is only used to prevent a single individual or group from making a disproportionate exaction from the population as a whole. So understood, this third condition reveals the two-sided nature of the question of political legitimacy. Taken in connection with the second condition, it changes the view of the world from one of dominant reliance on the ideas of property and consent to reliance on the idea of making Pareto improvements, that is, bringing about new states of the world in which at least one person is better off and no one is worse off than before. This test is a “social” test insofar as it requires the welfare of each individual within the group to be taken into account before comparing two social states of affairs with each other. Voluntary transactions routinely generate Pareto improvements because no one will enter into them if they expect to be left worse off than before. Indeed, in virtually all voluntary, or win-win, situations, the transaction satisfies an even stronger condition in that all players regard themselves as better off with the deal than without it. Although the drafters of the Fifth Amendment adopted the “just compensation” standard before the formulation of the Pareto principle more than a century later, their views mirrored what that principle requires. We count as a social improvement any change that benefits each and every member of the society against whom coercion has been exercised. For these purposes, it does not matter whether that coercion has been exercised by general taxation and regulation or by the specific taking of particular pieces of property.

Fourth, in many cases it is desirable and feasible to have not only gains shared by all individuals in the group, but also a pro rata division of gain, so that no individual gains from public action more than any other. Stated otherwise, the ideal is that each person should get the same rate of return on his investment in collective activities as anyone else. Thus, in a simple three-person society in which A, B, and C are forced to contribute respectively (say by a tax on real property) 100, 200, and 300, then ideally, if A receives a 10 return (for a total of 110) on his contribution, B should receive 20 (for a total of 220), and C should receive 30 (for a total of 330). This is the same distribution that they would receive in any voluntary joint venture, and the coercive venture should follow the voluntary one in choosing this stable baseline in order to prevent a dissipation of surplus by factional intrigue. This fourth condition is needed because the simple just compensation requirement under conditions two and three is intended only to ensure that no individual loses from the use of state coercion. It does not speak about the distribution of the gain across other members of society. The Pareto test is satisfied if A ends up with property and public benefit worth 101 or more, B ends up with 201 or more, and C with 301 or more. The full 60 units of gain in the hypothetical example could be the focal point of factional struggles, or rent-seeking, that could wholly dissipate its amount. The highly stringent test of pro rata division is hard to satisfy in practice, but the point of this requirement is to prevent any two members of this three-party group from ganging up on the third in order to gobble up the lion's share of the surplus, even if the just compensation test has been met. The principles here, moreover, are easily expandable to any number of individuals. If the pro rata standard can be met, the level of political intrigue can be kept low. This is one argument for the flat or proportionate tax, whose form eliminates partisan battles over the steepness of any progressive tax. In making this argument, I do not mean to gloss over the serious measurement problems that are raised when voluntary transactions are not available to determine the value that each person attaches to some collective good such as law and order. There can be no doubt that these are highly variable in different contexts, and in practice we often assume that the benefit achieved from any given transaction is proportionate to the stake that individuals have in society, even when it is unlikely that this is the case in reality. Stated otherwise, there is nothing in this (or any other) model of governance that constrains public debate over whether to wage war or to make peace with foreign nations. But this is a problem that dogs any theory of collective action, even for the devotees of the minimal state.

The implications of this model are quite simple. No matter what his protestations to the contrary, Nozick, as I shall show, relies implicitly on this model of forced exchanges to justify the state. Yet once that assumption is laid bare, it is possible to identify extensions of state power that go beyond the minimal state without violating any of these same restrictive conditions. By the same token, these restrictions do argue strongly against, and thus limit if not preclude, the one form of government action that Nozick wished to ban: to wit, the forced redistribution of wealth through state coercion. Such forced redistribution flunks all four of the conditions set out above, and can be allowed, if at all, only if one introduces notions of diminishing utility of wealth that are so profound that they justify the very large administrative expenses needed to work a greater equalization of wealth with its consequent losses in productive capacity—a challenge that it is difficult to meet, to say the least.²⁵

Indeed, the last dollar lost would have to be worth nothing in order for this redistribution to work a Pareto improvement, a highly unlikely proposition.

IV. Forced Exchanges in Private Transactions

As I have noted, much of Nozick's account draws its intuitive force from its close connection to ordinary common law conceptions of right and wrong, as embodied in the law of property, contract, and tort. His rule of initial acquisition is like the first-possession rule of common law, which awards ownership of a thing to the party who captures it first. The law of contract is intended to secure voluntary transfer of human labor or the property acquired with it, and the rules of tort deal with the rectification of past wrongs and the prevention of future ones. The combination of contract and tort rules is of special importance because contract rules stress the role of consent in the creation of private obligations while tort rules emphasize the prohibition against the use of force and fraud to achieve private ends. But it would be a serious mistake to assume that the full range of common law rules shows any invariable preference for consent over coercion. Rather, the more complete statement of the applicable principles introduces a necessary note of complexity into the overall picture, which shows the irreducible role of coercion in private relations.

Looking at the rules for the acquisition of property, it is evident that they cannot rest on a system of universal consent, as that phrase is ordinarily understood. The individual who takes something from the commons necessarily reduces the amount that is available to others, which is said to give rise to the endless fretting over the famous Lockean proviso: leave enough and as good for others. Yet the moment the first person takes the best, then, by definition, this condition fails. The blunt truth is that the removal of things from the commons necessarily imposes losses on other individuals. It is only with the dubious kind of “implied” consent that we can assume that others have agreed to these losses, which is a desperate effort to make the stubborn facts fit a contractual analysis that simply cannot carry the day. In truth, however, consent does not explain the rule; only necessity does. Locke himself justified the right to remove resources from the commons by noting that we would all starve if we had to wait for uniform consent to act,²⁶

which was his way of noting that the coordination problem from unanimous consent was intractable: any one individual, and hence every individual, could hold out against a solution that is desired by everyone else in order to claim an ever-larger share of the pie. It was, therefore, preferable to have a legal order in which people could take unilaterally in order to avoid mass starvation. To that general position, the Lockean proviso offers only a crude guess at how we could prevent the first taker from consuming everything that is available; yet this proviso, too, operates as an external norm, not as the outgrowth of consent.

Locke's explanation of these arrangements is incorrect not only because he does not quite hit on the right rules, but also because he fails to couch the analysis in the language of forced exchanges. That language refers to discrete bundles of rights: each of us, without individual consent, gives up any right of blockade over others in exchange for the right to remove resources from the commons. The Lockean side constraint is stated too categorically when it insists that enough and as good be left behind. A better way to put the point is to note that privatization, without the consent of others, should be allowed only to the extent that its marginal benefits to the acquirers (and those who benefit through them) exceed the marginal costs to outsiders, taken over the society as a whole.

At this point, it is clear why Locke is correct to link the legal constraints on the removal of property from the commons to the existence of trade. So long as the trade of things is possible after they are taken from the commons, it is relatively unimportant who gets what: others can exchange property or labor in order to acquire those things that they value more greatly.²⁷

It is only when the possibility of trade is blocked that it becomes a social imperative to leave some resources of value behind for others who would starve if they were unable to provide for themselves. Without some such restraint, in a state of nature each individual in good Hobbesian fashion would take as much as he could, without regard to the welfare of others. The legal rules are intended to counteract the destructive consequences of that universal practice.

Yet even when trade is allowed, there may still be reasons to limit the ability to remove property from the commons. With water, some constraints are needed even (indeed especially) if trade is allowed. Trade expands the demand for water, and its excessive removal from a small river of the sort found in England destroys its “going concern” value for fishing, navigation, recreation, and the like.²⁸

Of course, valuation questions abound in this setting, but to treat them as insurmountable means that no one can remove any water from the river or, alternatively, that anyone and everyone acting in a noncooperative fashion can dam it or drain it dry. Limited use is what takes place, where the effort, however crude, is to allow those private uses whose value exceeds that of the collective uses that remain. One appropriate forced exchange, easily stated, limits the removal of water for use on the riparian property and prevents its resale, which might otherwise seem allowable by Nozick's rules of justice in transfer. That rule is not ideal because it does not allow for the comparisons of the utility that rival riparians receive from their water. But for my present purposes, the fine-tuning of a system of water rights is not to the point.

All that need be recognized here is that the creation of property rights in their multiple configurations does not involve some simple relationship of an individual to an object in the fashion that Nozick supposes. The creation of correlative duties that fall on nonowners is an indispensable feature of any regime of justice in acquisition for any and all forms of property rights. Look again at the four conditions enumerated in Section III and it seems clear that property arrangements (including resource-specific property arrangements that differ between land and water) satisfy these four constraints on the libertarian system. The coordination costs are high; there is a clear social improvement from the change in rights that is shared by all and all alike in common proportion. It is for just this reason that customary rules of this sort start to verge on contract and make sense from a consequentialist perspective in which they are not usually articulated.

It may be said, however, that these constraints on the removal of property from the commons depend on the choice of some pre-political or common law baseline, but in reality they do not. Let us assume a baseline of a state of nature in which the rule is that no one can take anything from the commons. At this point, an alternative rule that allows a free-for-all on possession is likely to do far better. The life expectancy of all individuals under the first rule is near zero, given the certainty of starvation. The second rule marks an improvement that is shared across the board, where it is difficult to figure out what restraints could guarantee equal access—the last and most stringent of our conditions. But the “free-for-all” rule does not exhaust the gains with water, for example, so that the additional constraint that holds that each riparian (and only a riparian) can take water for reasonable personal uses marks an improvement over that rule. The key point is that the process of forced exchanges for social improvements shares this feature with voluntary contracts. Both create situations where sequential application of the same rule of forced exchanges can move the overall situation to a higher social level, even when judged by the exacting conditions set out above.

It is critical to note here that the process does not run in reverse. It is not the case that we could start (arbitrarily) with a rule that followed the customary system of water rights and justify a return to the state of nature. The losses suffered by all would block movement in that direction. The overall system, therefore, should be capable of reaching a unique equilibrium no matter what our original baseline. Where we start should make a difference solely in a high-transaction-costs environment when it might be less possible to shift across regimes. But conceptually, this analysis shows that Nozick's rules of acquisition through first possession are cut from a different cloth than the rules for contract (transfer) or for tort (protection/rectification) as he conceives them. The libertarian approach seeks to avoid consequentialist arguments, but in fact it only buries them out of view. Once individuals sense the justice of a certain practice, they choose to follow the practice even if they cannot articulate its reasons. Whatever the justificatory weaknesses of the categorical rule of acquisition through possession, that rule gives good and universal guidance to individuals who live within the legal framework—which is no small virtue if the effort is to develop a coherent approach that respects the rule of law.²⁹

Most people are happy if the instrumental logic of rules they accept is kept securely in the background.

The implicit breakdown of the libertarian approach is also evident when we press harder on the rules of contract and tort, which are not as categorical as Nozick would have them. Here two forms of correction are needed. First, there are many voluntary agreements that are not enforced because they are said to be against public policy. These include contracts to commit crimes or to fix prices. The argument for this nonenforcement is that the losses to third parties exceed the gains to the parties to the transactions, so that the refusal to enforce is a Pareto improvement over a rule that enforces all contractual undertakings indiscriminately. Once again, if we move back to a time before any contracts are formed, this rule also seems to provide social improvements in a high-transaction-costs environment. The moral intuition that promises are generally enforceable is best explained by the observation that these produce gains for the parties involved and, in the ordinary case, also for strangers by increasing the opportunities for further transactions. The latter condition is false in criminal contract cases. The consequentialist approach gives a certain coherence to both the rule and the ordinary exceptions to it. No intuitive moral theory can do the same.

In like fashion, the entire law of restitution shows that forced exchanges are even allowed when initiated by private parties. The standard case of private necessity allows one person to take water from the well of another if that is the only way to prevent dying of thirst. The same rule has been applied in legal cases to permit one person to use the dock of another without permission in order to escape the ravages of a storm. Indeed, the legal system is so insistent on the maxim that “necessity suspends property rights” that it treats the owner of the dock as a wrongdoer if he seeks to use force to put the intruder off his premises. The argument in all these cases is that just compensation in the form of rental for the use of the dock or payment for damages caused is a preferred remedy to a voluntary situation in which the owner could hold out for a huge fraction of the wealth of the intruder if the usual rules of consent applied. Quite simply, consent becomes a viable requirement when lots of choices are available to a person in need. But it looks considerably less attractive in those cases where there is only one person who has the power of life and death, which is what the necessity rule is intended to forestall. Once again, the point here is that forced exchanges are part and parcel of the private law out of which Nozick attempts to construct his principles of justice in holdings. It remains to show how this paradigm is necessary to explain the emergence of the minimal state.

V. Toward the Minimal State

The previous discussion sets the stage for determining Nozick's error in thinking that voluntary transactions alone are sufficient to account for the creation of both the ultraminimal and the minimal state. The ultraminimal state arises when, within a given territory, one protective association emerges as the dominant market player. The minimal state arises in recognizable form when that dominant association becomes the only association within the community to which all persons, now citizens, subscribe. The simplest objection to this program is that it cannot happen on any realistic assumptions about individual self-interest.

Start with the formation of the ultraminimal state. Here it is clear that two or more people are entitled to band together for mutual aid and protection so long as their combined rights do not exceed the rights that they held separately. You and I can join together for self-defense, but not for attack. We can then admit additional members into our group so as to create ever-larger alliances. Other individuals have, of course, the same option to form or join groups in competition with our own, so the question arises: When entry is free, how can any organization assume a dominant position while hewing strictly to its mission of defense only, never aggression? One possible explanation of how dominance could be achieved comes from the creation of the kind of network externality that is common in telecommunications. Let us assume that the value of telephone service is directly proportional to the number of individuals who are connected to any given network. Here the network that gets a leg up on its rivals will be able, ceteris paribus, to offer the better service, and thus win over individuals who might otherwise choose to join some other network. This sequence confers a first-mover advantage, and opens up the prospect that all individuals will find themselves with a supplier that none of them really wants. Political philosopher David Miller, in his thoughtful critique of Nozick, identifies this as one weakness in the Nozickian system, because the protective association that comes first might not be the one that a majority of the individuals would choose to have.³⁰

The vagaries of path dependence could override the preference for majority rule.

There is good reason to question whether path dependence will, or should be, critical with either networks or political order. In practice, a first network may have an initial advantage, but several complications intrude. First, there is no clear delineation of territory before the fact that will allow anyone to decide which organization has the dominant territorial control. Only after a state is cobbled together can its territory be defined, and then it is too late for the analysis to work. Second, there is no clear lexical or temporal order in these cases. Over large populations, many different networks could overlap in the time of their formation or in the rate of their spread. The question here is a bit like asking which riparian is first to reach the river. In practice that question has proved so hard to answer that the first-possession rule used with respect to land has always given way to a proration rule among riparians with respect to water rights: no matter when you arrive, your use-rights are a function of the total frontage along the river, without any temporal priority. Third, even if one network did get a jump outside the starting gate, others may prefer to join a second network that better fits their personal needs, counting on some master agreement to link the two networks together. This often happens in telecommunications under “bill and keep” arrangements, where each side agrees to carry the traffic of the other while both keep the fees for those transactions that were initiated on their side of the network.³¹

These deals involve no transfer payments, and they prove stable because they have a strong focal point due to their reciprocal nature. Fourth, the first mover always faces the risk of a mass switch of customers to a rival association, brought on by an attractive group offer. Exactly how this works out in the end is dependent on the transaction costs needed to bring on any particular switch: in the limiting case, as transaction costs approach zero, the path advantages disappear as well. In the intermediate cases, all that one can say is that first-mover advantages would count for something if political institutions behaved like network industries, although it is uncertain by how much. But overall dominant positions should be hard to come by in light of the counterstrategies available to other parties.

Nor is it apparent that the analogy to network externalities applies to questions of political organization. On the unlikely assumption that each of these associations will respect the rights of other associations in their own activities, what motivation does anyone have to join any association except the one that promises the best mix of services and payments? No network externalities are at work so long as any assumption of mutual respect holds. Yet once that assumption is relaxed, membership in a small organization could easily confer advantages on groups that can hold out against their dominant rival. The apt analogy to these protective associations is not the network industry, but the small swing-party that holds the balance of power in forming coalition governments under a system of proportional representation. The new entrants enjoy powerful holdout potential disproportionate to their numbers, which should give them a comparative advantage against any larger established party.

Additional adjustments have to be made when members of a smaller association are brought as defendants before a larger one. Further modifications are required to deal with disputes in which, as often happens in cases of alleged self-defense, each party claims that the other is in the wrong. Still further adjustments are required in order to deal with situations that involve individuals from three or more groups, or with cases in which it is uncertain who is the wrongdoer or, for that matter, the victim. The transactional complexities will quickly outstrip the ability to have ad hoc voluntary arrangements, whether at the individual or associational level, to deal with the complexities across a myriad of protective associations. Indeed, the problem of enforcement against the various kinds of wrongdoers is a bit like the problem of autonomy and property in the first instance. The very reason why we start with an institution of property (which requires forbearance but not support from the world at large) is because we know (in a pretty strong sense) that the transaction costs of getting universal agreement on the rules against trespass and misappropriation are prohibitive. That same set of difficulties arises in the related context of enforcement agreements.

Once it is no longer possible to postulate one dominant agency, there is no way to see how any particular agency will be able to gain the total support of all individuals when the gains from holding out could prove so large. Nozick tries to deal with these possibilities by noting that in the intermediate stages it is possible (and necessary) to formulate rules that govern conflicts between individuals who have signed up with different protective organizations. Here the operative language is no longer consent. Rather, the leading trope is that the dominant association will allow its people to be hauled before another organization so long as it judges the procedures of that association to be as good as its own. But this requires some judgment as to the relative value of alternative regimes, which turns the inquiry back to the just-compensation rules associated with forced exchanges. Nozick himself recognizes the importance of the compensation element when he writes: “What is the explanation of how a minimal state arises? The dominant protective association with the monopoly element is morally required to compensate for the disadvantages it imposes upon those it prohibits from self-help activities against its clients.”³²

In effect, the system does not operate by rules of justice in acquisition and transfer. His critical move necessarily represents an explicit deviation from the voluntaristic ideal. The introduction of compensation marks the move to a system of forced exchanges, which in international (and interstate) relations cashes out to a nondiscrimination requirement, whereby the outsiders have no grounds to protest if they are governed by the same rules that bind the locals.

At this point, it is only necessary to establish the strong link between a nondiscrimination rule and the system of forced exchanges mentioned above. This test plays out by noting that each person surrenders his right to be judged by his own sovereign in exchange for the right to be governed by rules that others apply to themselves. Using these standard-issue rules offers a powerful line of defense against abuse, bias, and favoritism that are everywhere the concerns of judicial proceedings. There is today a powerful need for a nondiscrimination provision for individuals who live or work outside their home territories. But this hardly establishes that it makes sense to have multiple authorities within any given territory. The rule of forced exchanges does not, indeed cannot, tell us which organization should have pride of place within a given territory, but it does say that whatever organization or group does emerge must apply its own rules to outsiders who are brought within the fold once territories are established. The use of this “like justice” requirement is meant to finesse strict proof of just compensation, which could not be established in any event in these highly specialized markets. The entire tale, therefore, speaks of a set of arrangements in which the nondiscrimination rule is an imperfect proxy for just compensation. In some cases the law relies on formal equality in the treatment of individuals as a proxy for nondiscriminatory treatment. That test works well in one direction, for explicit distinctions between individuals should be justified before they are tolerated. But the converse is less likely to hold. It is easy to envision rules that are formally equal that have a disparate impact on the regulated population. One illustration is a prohibition against all further development in a community when some landowners have already built while others have vacant land. At this point, the hard administrative question, of greater concern for lawyers than for philosophers, is whether the additional administrative costs (of figuring out whether the formal rule cloaks redistribution between neighboring landowners) are justified by the level of covert redistribution through regulation that they prevent.

In dealing with the philosophical issues, however, the exact form of the nondiscrimination rule is not the key point relevant to Nozick's voluntaristic claims. Rather, what matters the most is that the move from voluntary association to political organization takes place by means of forced exchanges and not by consensual transactions. The fact that the transactions follow this path should lead us to be cautious in defining the powers that are extended to the state precisely because its justificatory apparatus is of necessity suspect: our criterion for state power is weaker than actual consent, and the difficulties in measuring compensation are far greater. But once the method is understood, then oddly enough the need to posit intermediate protective associations disappears as well. Rather, it becomes possible to revert to the progression of social contracts that mark the evolution through Hobbes, Locke, and Hume. We start in the state of nature, a world in which anything goes. We realize that all individuals are better off if they mutually renounce the use of force and fraud, as Locke did with the rejection of license in favor of liberty. We then further realize that this renunciation will be written on the wind unless social institutions are established to secure the new prohibitions that are put in place. We next recognize that a system of taxation proportionate to property protected can be used to supply state services, including courts free from bias, to administer these rules. Nozick, like Hume, does not like the idea of “social contract.” But in effect that is where the basic argument leads, so long as both words receive due emphasis. The word “contract” refers to the proposition that at each stage we are confident (but not positive) that the benefits conferred exceed the costs imposed, which is what happens in real contracts of hire and sale. We use the word “social” to connote the fact that these contracts are hypothetical, and we keep the four constraints on forced exchanges—high transaction costs, just compensation, net benefit, and proration of gain—to ensure that these hypothetical contracts do not run riot. The Nozickian effort to use voluntary sequential contracts to reach the minimal state thus fails. The effort to use forced exchanges with mutual benefits in high-transaction-costs environments succeeds, however imperfectly, in filling that gap. The question thus remains: How may that principle be used once the state is in formation?

VI. Beyond the Minimal State

We are now in a position to see the error in the second half of Nozick's argument, namely, that no state larger than the minimal state is permissible. We immediately should be suspicious of that position, insofar as it contravenes the common assumption that states routinely provide public goods for their citizens through the means of taxation, let alone a wide range of redistributive services, such as social security, that prove, on examination, to be far more controversial. That said, Nozick's conclusion would be correct if he could demonstrate that the voluntary approaches alone were sufficient to create the minimal state. But his argument loses all of its zip once it is recognized that forced exchanges are built into the foundations of any organized society. They are first built into society in the definitions of property and liberty and their correlative duties, and next in the creation of a single sovereign with monopoly power over all individuals within its territory. If, however, the method of forced exchanges is needed to explain and justify the creation of a principled state, there is no reason why the same methodology cannot be used to explain commonplace institutions that are developed once that state is established. Quite simply, the logic of forced exchanges is to allow for Pareto improvements that voluntary transactions cannot achieve. It is difficult to imagine, let alone identify, any moral theory that would sanction the use of this principle to account for the legitimate foundations of state authority only to turn around and reject the use of this principle once the state is created. The real test, therefore, is whether we can identify common institutions that satisfy this criterion. And we can.

In order to undertake this process, it is necessary to look at a set of midlevel institutions that are vital to the day-to-day operation of the system. These midlevel institutions often fly under the radar of philosophers and political theorists, but these institutions constitute key elements of any well-functioning state. I will begin with those that are not characteristic elements of the modern welfare state, but are considered to be elements of a sensible state devoted in large measure to the protection of liberty and property. Begin first with the security of transactions. To philosophers the important element about contract turns on the moral force of promises. Questions of proof and implementations are second-order questions that are best handled by lawyers with their superior technical skills. Three of the relevant devices of great note and importance are the statute of frauds, the various recordation statutes for real property (and some forms of personal property), and statutes of limitation. The statute of frauds identifies particular classes of contracts, such as those for the conveyance of land or the guarantees of personal contracts, that must be evidenced in writing.³³

Recordation statutes require that various deeds and liens (indicating a security interest for a loan or other debt) be recorded in order to be valid against subsequent purchasers of interests in the subject property. Statutes of limitations bar, upon the passage of time alone, claims that are otherwise valid. One key variation on the statute of limitations is the doctrine of adverse possession, which allows the unlawful possessor of land to trump the interest of the original landowner after the passage of time. All of these rules represent enormous areas of working law subject to many variations and exceptions. Their one common thread is that they show how no working legal system stops with the minimal state.

The initial point is that none of these well-nigh universal statutes is justified by reference to any basic moral principles of right conduct, including Nozick's. Quite the opposite: the statute of frauds denies the enforcement of serious oral promises solely for want of the formality. The recordation statutes for their part upset a central principle of justice in transfer, because first in time need not be higher in right. Let O first convey property to A, who does not record, after which O conveys the same property to B, who does record, and B will take priority over A, notwithstanding that at common law O had divested himself of all interest in the property by the initial conveyance. The statute of limitations blocks claims because of the staleness of evidence and the conviction that people should act to assert their rights within a reasonable period of time or forever hold their peace. I would be hard pressed to imagine any serious libertarian who would have an instinctive revulsion to statutes whose central purpose is to supply additional levels of security to property and contract even if they clashed with Nozick's rule of rectification, which does not explain how the passage of time purifies a legal wrong. Libertarians might, of course, seek to justify these statutes by some appeal to the common good. Although such an appeal gives these statutes safe passage, it spells the end of any constraint on government, for it is hard to see what legal doctrine, whether by statute or common law, could not pass muster under this test. That justification becomes an excuse factory for excessive state intervention unless it is tightly moored to the claims of the individuals within society. Indeed, just this loose test is used today in modern American constitutional law (where it travels under the name of the “rational basis” test), and it has had the (predictable) effect of fostering such intervention.³⁴

Under this test, so long as one group benefits from the passage of a statute, no court will look closely to see whether that benefit pales into insignificance in comparison with the harms that the statute imposes on others. Few statutes pass without some champions, so that in economic matters, the test has generally been interpreted as giving Congress and the states a free hand in the regulation of economic affairs.³⁵ The basic constitutional guarantees of life, liberty, and property can easily be circumvented under a capacious police power test molded in these terms.

But what makes these rules (statutes of frauds, recordation, and limitation) interesting is that they also satisfy the more stringent tests set out above for deviations from the libertarian norms. Here, the best way to see the point is to press John Rawls's famous veil of ignorance test into service for a cause quite different from his own.³⁶

Behind the veil of ignorance, one does not know whether he would be a buyer or a seller, a claimant or a defendant. Any rule that increases the security of transactions reduces the costs of doing business, so that all individuals are left better off in all future transactions from the implementation of the rule. The compensation received in the ex ante position comes from your ability as a buyer who records to receive protection from the institutions that block your claim, for want of recording, on this occasion. It is, therefore, hard to find anyone who would prefer to operate in a world without these protections; nor is it possible typically to find anyone who seems to gain disproportionately from their implementation. These statutes come in the generic form as close to Pareto improvements as one could imagine, which is why they have always escaped controversy and condemnation.

There are, of course, cases in which sensible statutes or laws could be turned to more dubious ends. Certain individuals could be selectively exempted from these various rules so that it is no longer possible to say that the rules raise all people equally to a higher level of utility. Thus, serious issues would arise if a statute of limitations were lifted after it barred claims of a certain class, or if the allowable periods for adverse possession or recordation were so short that in practice no one could comply with them. But these ploys do nothing to undermine the strength of the general theory. Quite the opposite: they only confirm its use. Each and every one of these changes has to promise a fair chance of overall improvement in order to pass muster. The selective and opportunistic application of the rules counts as a form of bias that virtually ensures that these conditions are not satisfied. Retroactive changes in the law are yet another illustration of the same problem. As with all empirical applications of the general test, there are only strong presumptions that certain rules bring about illicit redistributions between persons. Nonetheless, there are clear warning signs when something is amiss, as with any rule that imposes selective burdens while generating general benefits, or the reverse. Thus, a rule that taxed the landowners in one part of a town to provide road improvements everywhere would be suspect, whereas one that imposed local taxes for local benefits would not be. In all cases, however, the ultimate question is whether the gains are pro rata with the exaction, so that the apparent mismatch of benefits and costs might be justified, for example, if all the residents in one part of town made heavy, well-nigh exclusive, use of roads in the portion to which their special taxes were directed—an unlikely occurrence.

Thus far, I have discussed the use of forced exchanges only in connection with those legal rules whose purpose is to strengthen the system of property and contract rights that libertarians themselves would endorse. But the program here goes much further, for it often allows for sensible redefinitions of property rights when those redefinitions would work for the long-term advantage of all individuals in circumstances in which voluntary renegotiation is practically impossible. In an earlier essay, I discussed at length the various modifications of the boundary conditions associated with the law of trespass that work in this fashion.³⁷

The rules of live-and-let-live with respect to low-level invasions of noise and dust offer yet another example, as too do the rules that require individuals to provide lateral support for the land of their neighbors.³⁸ Indeed, regime changes in property rights also cover important restrictions on the ability of individuals to withdraw (in the manner that Locke had contemplated) common pool assets, such as oil and gas, or wildlife. Here the broad statement of “enough and as good” yields to very precise delineation of who can take how much and under what circumstances. The gains to all participants in the process can be quite startling. Overdrilling of oil and gas fields results in two forms of losses: the unnecessary reduction in output and the concomitant increase in the cost of removal. A sensible regime of well-spacing or unitization can reduce the number of wells and increase the total output of the field. If some surface owners are shut out from drilling altogether, then they could be given a fractional interest in the produce of other wells. So long as the net gains in production exceed the costs of running the cooperative scheme, it is possible to create a system that meets the basic criteria set out above. As with all such schemes, it is easy to envision variations that appear to be neutral on their face, but that in fact work prejudice to some participants in the system, at which point the proper judicial response is to block implementation of the system until the losers are compensated for their losses.³⁹

The same approach accounts for much of intellectual property law, and helps explain why it is so difficult to fasten on the optimal rules for copyrights and patents, both of which are statutory creations. Here the ordinary theories of possession allow a person only to take hold of the thing that he has captured or occupied. These words do not capture the situation with respect to those forms of property for which these physical modes of acquisition are inappropriate. Nonetheless, it is a commonplace argument that individuals will not invest in future writings or inventions unless they are secured a return for their labor, which cannot be achieved by exclusive control of a prototype or manuscript that anyone can build or reproduce at will. In essence, therefore, the great systems of intellectual property require all individuals to surrender their rights of imitation in exchange for the greater security that they receive for their own writings and inventions.⁴⁰

Even those who do not write or invent are beneficiaries of this system because they will be able to obtain access to things of value that would never be produced in the absence of these statutory schemes for intellectual property rights. The exact contours of these systems are by no means easy to define, either as to scope or as to duration, and the dangers of special pleading are evident here as everywhere else. The Copyright Term Extension Act of 1998, which tacked on twenty additional years to existing copyrights so as to delay the entrance of currently protected works into the public domain, is but one example of these dangers. The extension of the monopoly reduces the level of consumption—a social loss—without giving a new incentive to innovate, which is the functional justification for copyright in the first instance.⁴¹

For my present purposes, the ornate details of these property rights systems again are beside the point. What matters is how they tie in with Nozick's theory. In chapters 6 and 7 of Anarchy, State, and Utopia, Nozick makes some brief mention of these intellectual property rights,⁴²

but does not show in that brief compass how they fit in with his general views on justice in acquisition or transfer. Nor does he offer any account for these differences between different types of intellectual property. Rather, Nozick seems (his language is not clear) to make some elementary mistakes. Independent creation is a defense to an action for copyright infringement, but independence is not a defense, as he appears to suggest, to a suit for patent infringement. One reason is that copying is far easier to prove in the former case than in the latter. The compromises that are required in this area create important differences between the various subheads of intellectual property that can only be integrated into a general theory of property rights. Forced exchange from the common law baseline of property rights is the key tool that allows for understanding this area.

VII. Prohibited Forced Exchanges

As I have noted, the conditions under which forced exchanges may be used to deviate from libertarian baselines are quite stringent. Even though these are satisfied in many of the cases noted above, it by no means follows that every invitation to invoke the theory of forced exchanges to alter the basic structure of individual rights to liberty and property will meet with similar success. It is important, therefore, to indicate those instances where the older, hard-edge common law rules continue to hold firm because the proposed forced exchanges do not represent any overall improvement to the status quo ante. One broad and important class of arrangements in which the common law rules hold firm are ordinary private contracts for matters such as sale, lease, and hire. In each of these situations, the creation of a competitive market tends to exhaust all the gains from trade. It is difficult to conceive of any move that trumps the competitive equilibrium that emerges from the interplay of voluntary forces. Hence, the classical principle of freedom of contract exerts a tenacious influence in this area. Thus, the usual array of wage and hour laws, the sort that were so contentious at the beginning of the last century, should all be rejected on the grounds that they cannot do better for workers than the workers could do for themselves. The same is true of efforts to impose wage, rent, and price controls on particular industries or on the economy as a whole.⁴³

Likewise under this model, it is not possible to make out a credible case for antidiscrimination laws with respect to employment, lending, and housing so long as ordinary individuals and firms have the full range of choices available in competitive markets. Ironically, in this area, the source of so much uneasiness rests in monopoly structures that government creates: when the state makes a given union the exclusive bargaining agent for all workers by majority vote, then some nondiscrimination principle may well be needed, especially in cases of discrimination by race, to counter potential abuses of that power.⁴⁴ Nor is there any case for either an antidiscrimination norm or usury laws in any segment of the lending market, for their effect is to freeze out high-risk customers who cannot compensate the lender in advance for the risk of default. To be sure, it is sometimes urged that these limits are needed to serve traditional libertarian ends, namely to block fraud or duress, but that remedy is wildly overinclusive when a simple disclosure requirement (e.g., the annual percentage rate on a loan) can negate any shortfall in information.

Stated generally, it is just these cases of contractual interference that give rise to the strongest passions of libertarians, and they are vindicated twice over. Not only are the outcomes of these interferences consistent with the libertarians' limited theory of state power, but the case against these interventions remains every bit as strong even after we allow forced exchanges that generate social improvements. Nor does this outcome depend on some pre-political norm of pure competition. If someone looking forward from the state of nature thought that the baseline norm required wage, rent, or price controls, we could still insist on their removal on the grounds that such policies work for the long-run average disadvantage of all, especially when people have no knowledge of the roles that they will occupy once the veil of ignorance is removed.

The second area where the libertarian norms remain tenacious is on the question of state coercion for the redistribution of wealth from one person to another. Here the case is not so simple as a denunciation of theft. We do have state mechanisms that can determine conditions of eligibility for funds and trigger those according to need. But the hard question to ask is whether we can develop, in the style of Rawls, an argument that redistribution is justified as a system of social insurance that makes all better off in the state of nature than they would otherwise be. Here we find a complex trade-off between the inferior incentives that are created relative to the ability to even out the ups and downs in ways that provide gains across the board. The pervasive presence of health, life, and disability insurance in ordinary markets shows that there are demands for the smoothing of wealth across different states of the world, so the Rawlsian proposal here cannot be dismissed out of hand. But when one considers the voluntary mechanisms that have been devised to deal with questions of need, and the political risks that are associated with introducing social insurance programs of this sort, a credible case can be made that these programs do not work any kind of overall social improvement, and should be generally rejected on these grounds, notwithstanding their widespread adoption in modern times. The case for their severe limitation is, moreover, strengthened by the adoption of the rest of the program urged here, which promises a level of output and opportunity that should remove virtually all individuals from the conditions of poverty and want that could justify mandated assistance. Some modest program that deals with certain categorical disabilities (blindness, deafness, manifest incompetence, orphan status, etc.) and the like could still pass muster because such programs are relatively easy to monitor on the one hand, and give rise to little if any moral hazard from potential recipients on the other. Most generally then, the case for redistribution is far harder to make than the case for forced exchanges that produce ex post gains for all persons within the system. As a matter of political prudence, therefore, it is best to secure the programs of property rights and voluntary exchange before moving onto the treacherous shoals of forced redistribution of income or wealth. The guiding maxim is, redistribution last.