===== PAGE 1 ===== The Limits of Government. By David Schmidtz. Boulder, Colorado: Westview Press, 1991. examination a key argument used to justify governmental intervention into the free market. Some economic goods, it is alleged, the market cannot produce efficiently. Roughly speaking, goods will be produced if the gains expected by their producers exceed the costs involved in their manufacture. Some goods, e.g., national defense, have features that skew the market from an approach to efficiency. The goods in question, termed public goods, are character- ized by jointness of supply and nonexcludability. If produced, they are available to everyone within a given area. If, e.g., a drive-in movie is visible from the street, anyone who passes by may watch it, not just those who bought tickets to it. If, for example, some of the passersby would have been willing to purchase tickets had the film been otherwise unavailable, then (so it is claimed) a problem exists. The movie is underproduced. The owner of the theater takes into account only the gains from those who pay admission; he receives nothing from passersby who free ride. Private cost and public benefit in public goods diverge. To expunge the gap, various remedies have been proposed, most of which involve resort to the state. Schmidtz begins his assessment of the public goods problem by discussing several fundamental issues of political theory. He first analyzes the nature of justification, finding two basic sorts of reason by which political institutions can be supported. The first is teleological justification: here the institution has a feature or promotes a goal which counts as a reason for establishing or main- taining it. An emergent justification, by contrast, claims that the manner in which an institution has been established justifies it. D avid Schmidtz’s illuminating new book subjects to careful The Review of Austrian Economics, Vol. 6, No. 1 (1992): 125-135 ISSN 0889-3047 125 ===== PAGE 2 ===== 126 The Review of Austrian Economics, Vol. 6, No. 1 Many political theorists claim that the state promotes civil peace, welfare, etc., better than alternative arrangements. Others argue that if a state arises by popular consent, it holds power rightfully. The first of these claims is a teleological justification; the second, an emergent. Schmidtz cogently argues that these two patterns consti- tute the principal sorts of justification for the state. As he notes, the two forms are not the only conceivable types of justificatory argu- ment: Besides his own example, a theory that required the state to be justified both teleologically and emergently would fit neither type alone. Though neither mutually exclusive nor exhaustive, the two types are of great importance. Schmidtz uses his classification scheme to make an intriguing claim about a style of argument popular in recent political thought. Some writers maintain that institutions are justified by what people under given conditions would rationally find it in their interests to accept. Schmidtz maintains that the apparent form of the argument masks its true character. It is not an emergent argument but rather a teleological one, since its “consent” is hypothetical. More radically, Schmidtz avers that to grasp the true force of a hypothetical consent argument, one must cast aside the imaginary story in which it is cloaked. The consent that would occur, e.g., in Rawls’s original position, comes about for certain reasons. It is these reasons that bear the brunt of the argument. Schmidtz states the essence of his case with characteristic lucid- ity: “The hypothetical story adds nothing whatsoever. It certainly does not add consent, for the story is only hypothetical” (p. 6). I am not sure that Schmidtz’s provocative argument succeeds. Why cannot a teleological justification claim that an institution has the following good-making feature: it would be agreed upon by persons under stated hypothetical conditions? Perhaps this sort of teleological argument always fails, and I certainly have no wish to defend it. An advocate would need to show why the feature in question added to the moral value of the institution having it. But in form, at least, I can see nothing wrong with the argument. The good-making feature is not the reason the hypotheti- cal actors consent: it is rather their hypothetical consent itself. Schmidtz is surely correct that reasons need to be advanced in defense of the claim that consent would occur; but I cannot see that this always suffices to show the hypothetical consent does no work. Schmidtz now turns to a subject of concern to everyone interested in libertarian theory: the Lockean Proviso. In one of its forms, this principle has seemed impossible to satisfy; it requires initial appro- priators of property to leave “as much and as good” as they take for others. ===== PAGE 3 ===== Book Reviews 127 Suppose that a parcel of land is divided into one hundred lots, identical in all qualities. Further, assume that there are one hundred settlers who use the land in common. If someone takes one of the lots, he has left ninety-nine lots for everyone else. Has he left “as much and as good” for others? On one reading of the Proviso, he has not; by removing one lot, everyone else has one fewer choice available. The problem of satisfying the Proviso becomes even more severe if one takes into account future generations. One must now leave as much and as good for those who will come into existence. Schmidtz finds it plausible to extend the Proviso in this way, although he does not specify how many generations need to be considered. Schmidtz seems on the surface to have painted himself into a corner. He has presented a version of the Proviso which seems on its face to rule out the initial acquisition of property; yet he wishes to defend property acquisition. With a bold stroke, Schmidtz extricates himself from the difficulty; the Proviso actually mandates the acqui- sition of property. “Leaving goods in the commons fails to satisfy the Proviso. In fact, leaving goods in the commons practically insures their ruin” (p. 21, emphasis in original). The argument Schmidtz deploys is straightforward. If land is worked in common, then individuals have little incentive to conserve resources. Persons do not bear the full cost of their actions; if, e.g., someone grazes cattle on common land in a wasteful manner, most of the ensuing costs will be borne by others. The upshot will be what Garrett Hardin has famously called a “tragedy of the commons.” Since property rights have not been established, the available resources will quickly sink to nothing. If so, Schmidtz ingeniously contends, the Proviso mandates es- tablishment of a system of property rights. The erstwhile irremovable bar to property rights has become their chief support. If it is objected that individuals who arrive on the scene later than the original appropriators can no longer themselves acquire unowned property, Schmidtz has a ready response. The Proviso requires that future generations have resources to use, not to appropriate. They are not harmed by having to acquire property from others. Schmidtz’s brilliant discussion will be a must for all future work on the Lockean Proviso. A few clarifications and objections, however, require mention. First, a point I think Schmidtz recognizes, he has not argued that the Proviso is morally required. It is not at all obvious why original appropriators must leave as much and as good for others: if one holds a strong labor-mixture theory of original acquisition, e.g., the limitation imposed by the Proviso will seem implausible. ===== PAGE 4 ===== 128 The Review of Austrian Economics, Vol. 6, No. 1 It might be objected that Schmidtz has given an argument for the Proviso. Unless it is satisfied disaster will ensue through a tragedy of the commons. Is this not a very strong reason for accepting the Proviso as Schmidtz construes it? But what Schmidtz’s point shows is that a society ought to have a system of property rights rather than allow unrestricted common use of resources. One need not fulfill the Proviso to meet this requirement. A system that requires individuals to bear the cost of their own waste of resources will prevent a tragedy of the commons, even if it leaves some people worse off than they would have been before the system was established. Suppose, e.g., that a few people in a society are totally denied any access to resources. The rest maintain a standard free-market system. There is no tragedy of the commons, but the Proviso is not satisfied. Further, it is not apparent why Schmidtz deems it obvious that the Proviso must be extended to future generations. Why is it irratio- nal to hold that one’s obligations extend only to actual persons, not to persons who do not now exist? I do not argue that Schmidtz is wrong: it just is unclear why he thinks the issue requires no discus- sion. To turn now to the substance of his argument; does the Proviso in fact require that property be removed from the commons? Schmidtz is of course correct that land is not being used efficiently in the situation he depicts. But it does not follow at once that a tragedy will ensue; this depends on circumstances. How much of a deviation will there be from efficient use? If the deviation is not substantial, then how strong is the requirement that enough and as good be left for others? It seems implausible that we are morally obligated to install a system providing for the highest possible degree of efficiency. If the use of land does not prevent successors from doing about as well as current users, why is there cause for complaint? (Perhaps if one took into account many future generations, however, a sorties problem might reinstate the Proviso.) It may well be, though, that Schmidtz is correct about most real-world situations, in which case his argument has outstanding practical importance. One more difficulty needs to be treated, but this fortunately can easily be remedied. Suppose that common use does in fact result in rapid impoverishment of everyone. Has the Proviso been violated? No, since as stated it restricts only those who appropriate property; by hypothesis, there are none in the present case. But the modifica- tion of the Proviso which Schmidtz requires to meet this minor point is obvious and will be left “as an exercise for the reader.” Schmidtz continues with more provocative and insightful analy- sis in his third chapter; there the subject is the right of punishment. ===== PAGE 5 ===== Book Reviews 129 His surprising conclusion is that “a state can acquire an exclusive right to punish even if individuals also have a right to punish and even if they do not relinquish that right” (p. 34). Although Schmidtz developed his view of punishment as a re- sponse to a problem in Robert Nozick’s treatment of risk, he ap- proaches matters in a very different way from Nozick. The risk that concerns Schmidtz is danger to innocent bystanders—presumably when someone tries to enforce a sentence upon a recalcitrant of- fender. Although Schmidtz avoids postulating a right not to be put at risk, he does assume that one has no right to impose unnecessary risks on others. If the state can reduce risks to innocent bystanders below the level prevalent in self-help enforcement, then people are no longer at liberty to punish those who have injured them. To reiterate, Schmidtz does not contend that one forfeits the right to punish in these circumstances. Rather, one cannot exercise one’s right, since an alternative exists less liable to place bystanders at risk. Once again, Schmidtz’s ingenuity compels admiration. He again and again comes up with original ideas that challenge and instruct. But should we accept his argument? I am inclined to think not. First, an analogous point to one raised in the discussion of the Proviso emerges here as well. Why should we be required to maximize safety to innocent bystanders? So long as private enforcement does not place innocent bystanders at substantially more risk than does state enforcement, why should one insist on the least possible risk (consistent with having sentencing at all)? As will later be apparent, Schmidtz himself uses the distinction between “good enough” and “best possible” in his later discussion of public goods; does it not also apply here? Further, why does Schmidtz think it likely that private enforce- ment will be riskier to bystanders than state enforcement? Perhaps he has in mind something like this: I attempt to take back my wallet from a thief. He sees no reason to return it to me, and in the ensuing struggle, people are injured. Had the police dealt with the thief, the bystanders could breathe a sigh of relief. Although on occasion crim- inals “shoot it out” with police, resistance is much less likely to occur. No doubt matters often take place in the way just described. But it does not follow from this that individual enforcers need abandon their activities. They need only choose their opportunities carefully. They can, e.g., take care to have sufficient force to deter the criminal from resisting. One way to do this is to hire a private protection agency to enforce judgment: If the agency were sufficiently powerful, resistance seems an unlikely prospect for the same reason that ===== PAGE 6 ===== 130 The Review of Austrian Economics, Vol. 6, No. 1 relatively few people resist arrest by the police. Another possibility is to pick one’s spots carefully so that no innocents stand in danger. Suppose, however, that Schmidtz’s picture perfectly matches re- ality: A single dominant protection agency is vastly less liable to expose bystanders to danger than any combination of independents. Have we obtained a state? I cannot see that we have. It might be the case that the single agency able to enforce sentences could do nothing else. The mandate for punishment would always come from an indepen- dent. I do not claim this possibility is more likely to arise than the one Schmidtz envisions, but only that his argument does not exclude it. A strong point of Schmidtz’s work is that he attempts to meet in advance objections to his argument. One such objection suggests, following Guido Calabresi and A. Douglas Melamed, that injury to innocent bystanders be handled differently from Schmidtz’s ap- proach. Instead of prohibiting independents from enforcement, why not compensate innocents who are injured? To this Schmidtz replies that the bystanders are being subject to pointless risks. Thus, the case should be governed bv a property rather than a liability rule, to use Calabresi’s terms: Prohibition rather than compensation is the indicated course. Schmidtz’s supposition of pointless risk does not cover all rele- vant cases. What if the state does not impose exactly the same punishment as the independent would have done? What if the state finds the criminal innocent, or fails to take up the case at all? Here one cannot say the risk is pointless—even if Schmidtz’s argument were otherwise acceptable, it would not apply to independents in these circumstances. But what of instances in which Schmidtz’s account exactly fits? I am not sure that even in this type of case, the risk is pointless: what if someone values protecting his own rights? But what if the risk is pointless? Schmidtz gives no argument that prohibition rather than compensation is the proper course to follow. But I have no argument to offer that he is wrong. Schmidtz introduces the principal topic of his book, the provision of public goods, with an excellent analysis of the Prisoner’s Dilemma. Many writers view the supply of public goods as raising the same problem as the classic Dilemma: supplying the good is collectively rational; but, for each person, refusal to cooperate in its supply dominates contributing. Since no person’s actions will substantially affect supply of the good, everyone individually seems better off refusing to contribute, regardless of what others do. Schmidtz very usefully divides the difficulty of securing coopera- tion into two parts. On the one hand, the familiar free-rider problem arises: Although everyone is better off if the public good is supplied, ===== PAGE 7 ===== Book Reviews 131 for each person refusal to contribute dominates contributing. On the other hand, the free-rider problem, as Schmidtz convincingly shows, does not exhaust the difficulty of the Dilemma. He draws attention to the assurance problem: People willing to contribute do not wish their contributions to go to waste by failure to secure enough donors to produce the good. Does the problem of supplying a public good, analyzed as a Prisoner’s Dilemma, require forcible intervention for its resolution? Schmidtz does not think so; he shows that voluntary agreement suffices. His solution accents the assurance problem. A universal contract will suffice to prevent contributions from being forfeited without result. The terms of such an assurance contract provide that everyone agrees to contribute to the provision of the public good, if and only if everyone else does. Further, with a unanimous agreement, the remaining obstacle to a cooperative solution also dissolves. There are no free riders, and individual rationality now coincides with what best secures the welfare of the group (p. 57). One might at first object that Schmidtz has by his assurance contract merely pushed back the difficulty. How is compliance with the contract to be secured? Will this not in turn generate another Prisoner’s Dilemma, this time to be resolved only by force? To this difficulty Schmidtz has a response: Contract enforcement is not a public good since it is not characterized by jointness of supply. Those who do not purchase this good will not secure the enforcement of their contracts. If it is now objected that contract enforcement depends upon force, Schmidtz once more has an ingenious reply. Signers of a contract cannot rightfully complain if they are compelled to observe its terms, since they have agreed to do so. The agency that in fact enforces contractual arrangements cannot then be legitimately ques- tioned, since it might have arisen voluntarily. But what if it did not? Schmidtz now produces the ingenious counter just mentioned. Even if the enforcement agency has not arisen through voluntary agreement, its activities do not depend on its having arisen through the use of force. The reason for this is just that an enforcement agency might have arisen voluntarily: How the agency arose does not affect the legitimacy of its present activities. (How it arose may of course affect whether it, rather than some other group, is best entitled to perform these activities.) Thus Schmidtz has virtually completed his case. Public goods can be produced voluntarily, and nothing in the means required to do this need violate rights. This part of Schmidtz’s argument seems to me eminently successful. One point that might require further ===== PAGE 8 ===== 132 The Review of Austrian Economics, Vol. 6, No. 1 development has to do with the existence of individual property rights, necessary for a free market to function. To fully complete his case, Schmidtz needs to show that these rights do not depend essen- tially on the use of force for their establishment. I see no reason to think that this cannot be done. A unanimous assurance contract by no means puts paid to all problems about public goods. Not all public goods can be analyzed as a Prisoner’s Dilemma game, the players of which are a society’s entire population. Some people may not regard the public good as desirable at all: Opponents of United States foreign policy, e.g., will not wish to purchase the “public good” of national defense. Others may rate certain public goods as worth less to them than the contribution required through an assurance contract. Why not then confine the assurance contract to those who do think the public good worth securing? In Schmidtz’s view, this solu- tion generates a further problem. Some people may actually want the good; but, knowing that those who do not are not required to contrib- ute, they may disguise their true wishes. Falsely pretending to be genuine holdouts, they hope by their strategic behavior to gain the good for nothing. The free-rider problem reemerges. A further difficulty complicates the public goods problem even more. Unanimous consent is in practice intolerably cumbersome to achieve: It is likely that the market solution will consist of an assurance contract of fewer than the total number of those who want the good. The number required will vary with the minimum number needed to produce the good (unanimous consent will presumably still be required for a pure public good, in which no coalition suffices for its production). Once more free riders threaten. Schmidtz suggests several principles to aid in the solution of these problems. He sets forward conditions under which he thinks people stand under an obligation to contribute to the provision of public goods they want. But obligation does not generate by itself liability to be coerced by others, should one fail to fulfill one’s duty. Because of the tendency of government officials to advance their own interests, abundantly documented by public choice economists, it is highly unlikely that government intervention will bring about a better state of affairs than the free market. On the contrary, it is a safe bet that the government will worsen things. Schmidtz accord- ingly holds that the government should stay out of public goods provision, except in emergencies in which immediate action is re- quired to prevent the destruction of society. Schmidtz does not confine himself to a purely speculative account of how the market might produce public goods. He describes several ===== PAGE 9 ===== Book Reviews 133 experiments, some of which he helped to design himself, which have endeavored to determine how people might handle public goods provision. The experiments provide some evidence for the view that people will voluntarily agree to produce public goods, although their behavior does not conform to theoretical models of rational choice. To those who object that the experiments are artificial, not genuine tests of market behavior, Schmidtz has a clever answer. Theories of public goods production should, if correct, apply to experimental situations, since these fall within the scope of the theories’ postulates. The experiments Schmidtz describes thus are more than simulation of actual market behavior: they are themselves “real-life” tests. Austrians may be inclined to look on experimental economics with a jaundiced eye, but I think Schmidtz escapes danger from this front. He does not claim that he has an airtight model, deductively derived from self-evident axioms, specifying when it is rational for persons to contribute to public goods. He need not be taken as committing himself to the view that all economic theory requires empirical verification. On another point, though, Schmidtz’s analysis sharply contrasts with a Misesian position. On the “demonstrated preference” view, only persons’ actual choices are counted as genuine preferences. Counterfactual preferences are banned from praxeological analysis. As one can readily imagine, this has quite drastic consequences for the public goods problem; Hans Hoppe, most notably, has used dem- onstrated preference in an attempt to undermine standard accounts of public goods. I strongly suspect that Schmidtz thinks this view mistaken. I wish he had spent at least a short time discussing the issue, as the question is of concern to many philosophers and economists of free- market orientation. But given the wide range of topics Schmidtz does cover, I am no doubt unfair in my complaint. To return from Austrianism to Schmidtz, his skepticism about the role of government in the supply of public goods seems eminently justified. His careful distinction between obligation and liability to coercion is particularly excellent. On one point, however, Schmidtz might usefully have extended his skepticism. He assumes that free riding, and perhaps certain other kinds of strategic behavior, are morally questionable. Thus, he holds that if one benefits to a sufficient extent from a public good, one ought to contribute to its production. Further, though I may be mistaken, I think he finds morally problematic pretending for stra- tegic reasons to be a holdout. ===== PAGE 10 ===== 134 The Review of Austrian Economics, Vol. 6, No. 1 Though most philosophers interested in this area agree with Schmidtz, I do not think it at all obvious that something is wrong with free riding. Prima facie, the free rider (or strategic bargainer) is acting to secure his own interests: what exactly is his moral offense? It might be said that free riding prevents a group from maximiz- ing its collective welfare. But why is this a reason to hold the practice morally wrong? Is one under an obligation to avoid interfering with the use of resources to maximize collective welfare? On certain consequentialist views, the answer is yes; but these positions cannot simply be assumed as given. Schmidtz might reply that this objection underestimates the force of the considerations against the free rider. It is not only the group that will benefit if no one free rides: With an assurance contract, everyone, including potential free riders, will be better off than if each person is free to act as he pleases. But this at most shows that one would be irrational (or not rational in Schmidtz’s sense [p. xvi]) in refusing to sign an assurance contract. An additional premise is needed to generate a moral obligation. Schmidtz, as one might expect, knows this full well and has just the moral premise that will do the trick. His feedback theory of morality tests moral norms by the question: “Is this norm conducive to a cooperative society’s success, given how people will react both to it and to agents guided by it” (p. 153)? Granted that social cooperation is a conditio sine qua non of human happiness and survival, does it follow that one must take maximizing cooperation as the test of morality? Why is it not suffi- cient to secure some threshold of cooperation, after which this value might find itself in competition with other values? This point gener- alizes a criticism made earlier about Schmidtz’s discussion of the Proviso. From “x is good,” the further statement “maximize x” does not follow. Of course, I have not shown that one should not aim to maximize social cooperation: The point, rather, is that the matter is more controversial than Schmidtz allows. The case against the free rider suffers from a more fundamental flaw. The argument against him, very roughly, is that cooperation is desirable; free riding impedes cooperation; therefore, free riding is prima facie objectionable. A problem with this argument emerges if one considers once more Schmidtz’s convincing analysis of the Prisoner’s Dilemma. In that analysis, two problems produce the Dilemma: besides the free rider difficulty, there is also an assurance problem. If free riding is objectionable just because it impedes cooperation, why is there no similar objection to someone wishing assurance that his contributions ===== PAGE 11 ===== Book Reviews 135 will not be wasted? Schmidtz does not call for the curtailment or reduction of the desire for assurance; instead, he caters to this desire through the assurance contract. Why the difference? The objection to free riding cannot just be that it blocks cooperation. What then is it? I fear that I have gone on too long, but this is an endlessly fascinating book. Many of its themes I have had to omit: I think particularly of the illuminating discussion of what Schmidtz terms the concatenated Prisoner’s Dilemma (pp. 101-02)’ and the use of his feedback theory to derive a rule of reciprocity in cooperative ventures. The Limits of Government is an outstanding work, manifesting the remarkable philosophic talent of its author. David Gordon The Ludwig von Mises Institute 1Schmidtz suggests, following David Kreps, that someone playing an iterated or concatenated Dilemma with an opponent having an established reputation for reciproc- ity ought rationally to cooperate rather than use the backward induction paradox to justify noncooperation. But might the paradox interfere with developing such a repu- tation?