About this Issue

Throughout the 20th century, the economics profession largely eschewed questions of value as part of the attempt to gain “scientific” status. Accordingly, when formal economic models imagine the nation-state as an agent that may either impede or improve the efficiency of an economy, the distinctively coercive character of state action is neither emphasized nor taken into the accounting of cost and benefits of various government policies. But shouldn’t there be a presumption in favor of voluntary action? If a policy violates that presumption, shouldn’t the loss of liberty be taken into account — even within a fully scientific economics? Is there an intelligible and morally compelling voluntary/coercive distinction that exists prior to and independent of considerations of efficiency? Can there even be coercion outside of a government’s framework for the assignment of legal rights?

In this month’s thought-provoking lead essay George Mason economist Daniel Klein, editor of the academically muckraking Econ Journal Watch, argues that not only do considerations of coercion and liberty figure into economists’ scientific judgment as a matter of fact, but that they ought to figure in with added clarity so that we can more directly and honestly consider the extent to which coercion is really worth it. Klein will face a wide-ranging array of eminent interlocutors, including NYU philosopher Liam Murphy, co-author (with Thomas Nagel) of The Myth of Ownership: Taxes and Justice; Harvard economist Edward Glaeser, editor of the Quarterly Journal of Economics; and University of Chicago law and economics powerhouse Richard Epstein, author of Skepticism and Freedom: A Modern Case for Classical Liberalism.


Lead Essay

Economics and the Distinction between Voluntary and Coercive Action

In 2006 there appeared a “raise the minimum wage” statement signed by 659 economists. I wanted to know why they favored the minimum wage, so I wrote up a questionnaire and sent it to them. But I also used the occasion to get their views on a very important matter: Did they view the minimum wage law as coercive?

Ninety-five graciously completed the survey. Very few of them simply accepted that the minimum wage law is coercive. More than half said the law is not coercive in any significant sense.[1]

But the minimum wage law (and concomitant enforcement) threatens the initiation of physical aggression against employers who pay less than the minimum wage. It threatens physical aggression against people for engaging in certain kinds of voluntary exchange. To me, that is coercion. Just imagine if your neighbor decided that he would impose a minimum wage law on us. Wouldn’t we all agree that he was coercing us? If it is coercion when he does it, why isn’t it coercion when the government does it?

Now, you might be muttering, “Yeah, whatever, but I’m interested in economics. I don’t care to ponder semantic issues about moral and political terminology. Let’s leave that to the philosophers.”

Hold on. We need the distinction between voluntary and coercive action to give meaning to “the free market.” We need it to identify an “intervention.” We need it to measure “economic freedom.”[2] We use it to categorize classes of action, to identify and define industries, to formulate theoretical parallels between one industry and another, and between one polity and another. We use it to formulate reform proposals. Our theories about human interaction make key distinctions based on whether the interaction is voluntary. We generally assume that the individual is bettering his situation in voluntary interaction, but we don’t make the same assumptions in coerced interaction. The distinction between voluntary and coercive is built into many of the key analytic distinctions we use in economics.[3] So it is important that we know what we mean by it.

And it is important that we be aware of the extent to which others reject the distinction. The minimum wage issue is very telling. Economists have been surveyed on the minimum wage. On average, economists are neutral, neither for nor against. But what is so striking is that opinion is not a hill in the middle of the spectrum of possible opinion. It is not even uniform across the spectrum. Rather, the pattern of opinion is U-shaped. A lot of economists oppose the minimum wage, a lot support it, and fewer are in the middle.[4] Clearly, there are major cleavages in visions and formulations. I submit that those cleavages relate intimately to the semantic issue raised here. The distinction between voluntary and coercive action lies at the heart of many economic controversies.

Again, most economists who support the minimum wage do not see the policy as an incursion on liberty. No doubt, most opponents of the minimum wage would say the reverse. Clearly, the economics and the semantics are linked.

I think good economics opposes the minimum wage. Good economics places the distinction at the center of the scientific enterprise. It develops understandings of the comparative consequences of freer versus less free arrangements. And the main way we develop our analysis about arrangements in an industry or sector is by analogizing to other industries or sectors, sometimes in other times and places, forming a web of understanding about how arrangements varying in freedom perform.

But not everyone sees it this way. Maybe they have not learned that the minimum wage is bad because they don’t share our semantics.

The Substance of the Distinction

You are coerced when someone brings physical aggression or threat thereof to your property. Your property is your stuff, including your person, and ownership implies a claim to your property good against the world. A claim, a benchmark, not an absolute or inviolable right.

Voluntary interaction is our consenting, in the absence of coercion, to alter our property arrangements and to form agreements such as contracts. As for who owns what, there are rather universal norms, beginning with the soul’s ownership of its person, extending to property acquired within the family and in trade, production, and gift relations.[5] Liberty is freedom from others messing with your stuff. Restrictions on voluntary interaction are diminutions of liberty.

Sure, there are holes and gray areas, and forms vary with social norms. But the basic ideas of tangible property, ownership, and consent are cogent and apply so widely that we may think of the exceptions as exceptions.

Within Liberal Civilization, the Distinction is Natural

The voluntary/coercive distinction is natural in the sense that, within liberal civilization, it is intuitive, emergent, and widely observed. Indeed, within liberal civilization, institutionalized coercion by private (nongovernmental) parties is almost never tolerated. One exception is “range country” rules in Montana and elsewhere, which entitle your neighbors to let their cows tromp and graze on your unfenced land. It’s up to you to fence out cows from coming onto your land. Another exception, in my view, is loud Harley-Davidsons. The idea of extending the principle to government action is natural enough.

The distinction sits plainly on the tabletop in our cultural living-room, and it has for centuries. It is available for analytic deployment even if the particular arrangements actually around you are highly coercive.

Natural Maxim versus Natural Axiom

Again, when it comes to private-to-private matters, the liberty principle comes close to being an absolute. It is just about 100 percent. But not so in governmental matters.

Government is a unique player in society, and rules and norms have emerged that recognize that uniqueness. We tolerate governmental coercion that we would not tolerate from private parties, and not only because the government is more resolute and better armed.[6] The liberty principle does not work as an axiom. Instead, it works as a maxim: In a choice between two policies (or reforms), choose the one that entails greater liberty. But it is only a rule of thumb, a presumption, something that we expect to be right ninety-something percent of the time.

The voluntary/coercive distinction spells the liberty principle, and, the liberty principle has often been represented as a moral axiom. In consequence, one of the great handicaps of making the voluntary/coercive distinction explicit and prominent in one’s economics is that it arouses suspicions and accusations of maintaining liberty as an axiom. It puts free-market economists into the position of having to explain that the distinction does not carry a necessary rejection of coercion. One can favor the distinction and some coercion.

Walter Block eloquently exclaims: “Coase, get your cattle off my land!”[7] Block is right about “my land.” But he is not necessarily right about “get off.” Maybe range country rules are good policy and good norms.

If we can diminish the close association of the distinction with the necessary recommendation of liberty, we will be in a better position to use the distinction as an analytic engine of inquiry into that central question: When should we endorse the liberty maxim and when not?

Many Dislike the Distinction

So, if an economist deploys the distinction he faces the problem of people mistaking maxim for axiom. But there is an even bigger problem. Even when clearly understood as a maxim, the distinction makes for a reading of the status quo that most economists find offensive. The distinction says that we live in a polity of pervasive coercion. The minimum wage, occupational licensing, FDA restrictions, gun control, drug prohibition, all forms of taxation, and myriad other government regulations are coercive. The distinction says that the New Deal was a watershed in institutionalized coercion. It exposes this fact to sunlight. The distinction-deploying economist might reassure his listeners, “Now, realize, just because I say the policy is coercive doesn’t necessarily mean I think it’s bad,” but nonetheless people will be offended. In everyone’s lexicon, “coercion” has a strong negative connotation.

Those who dislike the distinction try to get around it by redefining the key vocabulary: property, consent, freedom, rights, justice, equality, and equity. The central idea of their worldview is that the polity is one large voluntary organization, and its rules are entered into by consent. No one is forcing you to stay. Thus, when the government imposes a minimum wage law, it is not treading on your property or freedom, it is merely rearranging the rights that define your property. Your property, in this view, is the bundles of rights that the government says you have. The presumption behind this philosophy is that whatever stuff you have really belongs to the government, the organization, the state, and it is ‘yours’ only in the sense that they delegate to you certain powers over it. The state is the encompassing overlord, the real owner of all property in the polity, and we are just tenants.

The polity-as-voluntary-organization viewpoint allows many economists to dispose of the distinction. If an economist openly invokes the distinction as a fundamental analytic category, and hence implies that we live in a society of wholesale coercion, he runs the risk of being shut-out by the other types of economists. Sometimes they use the term “ideologue” and shut him out of their journals and institutions.

Vying Conceptions of Economics

The deep tensions between free-market economics and the broad political culture help us understand why even free-market economists have promulgated conceptions of economics that skirt the distinction. Lionel Robbins championed the idea that economics is about the pure logic of choice, the efficient way of achieving exogenously given ends. In the same vein, George Stigler and Gary Becker assert that economics is about individual utility maximization within a theoretical framework of equilibrium. In my view, these conceptions teeter between vacuity and inhuman artificiality, and have contributed to the miscarriage of economics. One reason they have nonetheless been maintained is that they help market-oriented economists navigate the culture. George Stigler not only downplayed the distinction, but directly attacked it, arguing that wealth, utility maximization and efficiency have superseded any concerns about liberty, rendering the latter a meaningless and unimportant idea.[8]

“Spontaneous” Means Voluntary

Appreciation of the distinction helps to clarify good economics. Hayek is famous for developing ideas of local knowledge and spontaneous order.[9] His lessons against central planning have been taken to heart. But while people concede that the government should not centrally plan, they go along with a thousand forms of government tinkering, like the minimum wage. Let people act spontaneously, but alter the range or types of actions they may choose. That way we can utilize local knowledge but still ameliorate externalities, asymmetric information, and the like.

The distinction helps us see, however, that “spontaneous” principally means free. Even though restrictions like the minimum wage cannot be called central planning, they are in fact incursions on spontaneity. Hayek’s insights also have power in criticizing interventions. What the interventionists neglect is that the very problem posited to justify intervention would generate concern, awareness, and opportunity for new practices and institutions. The aberrations create new opportunities for mutual gains, opportunities that summon our entrepreneurial propensities to resolve or avoid the initial aberration. Occupational licensing, for example, is rationalized as consumer protection from quacks and charlatans. Yet we witness myriad private institutions and practices to certify practitioners and assure the quality of their services. Economists who study occupational licensing agree that, rather than protect consumers, the requirements hurt consumers by restricting the range and competition of spontaneous developments.

Behind our scientific sensibilities lies an educated, defensible faith in the potentiality of coincidence of interest, and that principle is defined in part by the distinction we are discussing. Hayek indeed made the distinction central in his economics, but it should be noted that he did so very diplomatically, often leaving it “between the lines.” To smooth things over, Hayek often employed “competition,” “decentralized action,” “the market,” and “spontaneous order.” Moreover, in his treatise on political philosophy he avoided the sound definition of liberty (rooted in ownership/property) and instead characterized liberty by reference to some of its important and appealing correlates.[10] Sometimes, in some discourse situations, obfuscation is appropriate, but in other situations we must propound the sound definition of liberty and the central role it plays in good economics.

Scientific Judgment Is a Matter of Sensibilities

Relinquishing liberty as 100 percent means that you have to judge whether this or that intervention is an exception. On what basis do you decide when to contravene the liberty maxim?

Here we draw on our broad sensibilities about consequences, including moral and cultural consequences. We make reasonable efforts to characterize our sensibilities, but we do not attempt to set out any complete or definitive characterization, any algorithm of desirability. Sometimes others demand a “foundation” or complete standard. It is good to articulate and clarify our deeper values and criteria, as best we can. But the deeper we go, the vaguer and more platitudinous our “foundations” become. You can’t characterize your sensibilities on economic policy any more explicitly or definitively than you can characterize your sensibilities on aesthetics. No one demands “foundations” in judging movies or poems. You should get used to a similar looseness in judging economic policies.

These Thoughts Agree with Adam Smith

The character of economics offered here could be called Smithian economics, for every important point finds support in the writings of Adam Smith:

  • George Stigler criticized Smithian economics for not being sufficiently Stiglerian.[11] Indeed, as Ronald Coase explained, Smith would have frowned on the characterizations of economics as utility maximization, “rational choice,” and the like.[12] Rather, Smith tended to see political economy “as a branch of the science of a statesman or legislator”.[13]
  • Central to Smith’s analysis in The Wealth of Nations is “the obvious and simple system of natural liberty,” which he closely associated with justice. Smith maintained the classical, intuitive “exclusion” idea of property, and his idea of liberty is built on the idea of property and freedom of voluntary agreement.[14] Natural liberty has a conceptual status quite independent of what the government rules actually are.[15] The conceptual system of natural liberty “establishes itself of its own accord.” [16]
  • The Wealth of Nations surveys policy issues quite comprehensively. Issues are formulated in terms of their conformance to natural liberty. Smith’s general scheme is to explain when to follow the liberty principle and when not to. Natural liberty is brickwork in Smith’s economics.
  • For Smith, the liberty principle was a maxim, not an axiom. In The Wealth of Nations he pauses to make explicit that he endorses certain specific contraventions of natural liberty (incidentally, J.B. Say did likewise).[17] Effectively, Smith emphasizes that the distinction is compatible with some endorsements of coercion. Smith said that the rules of commutative justice are analogous to the rules of grammar, and he implied that sometimes situations call for improper grammar—and being called for doesn’t make them proper grammar.
  • Smith would have been appalled by the subversion of the liberal lexicon and its implicit premise of government as encompassing overlord. Smith imputed a kind of legal positivism to Thomas Hobbes, and explained the errors of “so odious a doctrine.”[18] He also condemned Colbert’s management of the French economy “upon the same model as the departments of a public office,” and contrasted it with “allowing every man to pursue his own interest in his own way, upon the liberal plan of equality, liberty, and justice.”[19]
  • Smith’s scientific judgment supported the case for a culture holding a strong presumption of liberty. The liberty maxim holds ninety-something percent of the time, and it makes sense to use it as an analytic distinction and engine of criticism, and to develop theoretical categories in light of the possible exceptions to the maxim. Even when an intervention is established policy, its supporter should bear the burden of proof. The presumption of liberty—as opposed to a presumption of the status quo—is a feature of Smith’s economics. Smith sometimes supports status quo interventions (particularly, I notice, of the Scotland of his day), but he feels obligated to give good reasons. (Whether he always succeeded is another matter.)
  • Smith also would have upbraided demands for a definitive characterization of our sensibilities. Such sensibilities are not simple rules like the rules of grammar, but rather are like the rules of sublime and elegant writing, which necessarily are “loose, vague, and indeterminate.”[20] Yes, Smith saw commutative justice as analogous to the rules of grammar, but the warrant for a general adherence to and strong presumption of that justice lies in the loose, vague, indeterminate—but not arbitrary or meaningless—realm of aesthetio-political sensibilities. Smith wrote and rewrote two large books to develop and express what his sensibilities are like.

Concluding Remark

Smith saw clearly that economics had purpose: to address the most important things in economic policy and to edify the practitioner. Judgment about the most important things is, naturally, part of the science. How to formulate those things is also a paramount task of science. Again, he saw liberty as a natural concept with a status quite independent of judgment on any particular policy issue, so there was nothing irregular in using the concept in the formulation of issues and development of analysis.

Judgment enters into one’s analysis of the minimum wage and other economic policies. But, further, there are different ways to structure and develop the entire science of economics. The choice to use the voluntary/coercive distinction — both as brickwork and as engine of inquiry — flows in part from a judgment about the comparative value of the overall science that results. That judgment is all-important, and hence, too, a part of the science.

In my view, economic understanding, by experts and the general public alike, would gain by economists doing more of the following: (1) using the voluntary/coercive distinction in their formulations, analysis, and discourse; (2) making that utilization explicit and unabashed; (3) thinking hard about the content of that distinction, particularly by clarifying the holes and gray areas; (4) making it clear that, while they may promote a presumption of liberty, they do not mean to suggest that the distinction carries a necessary condemnation of coercion.

If the Smith-Hayek economists will admit that sometimes coercion is our friend, weakening the absolute negativity of “coercion,” then they might be able get others too to embrace the distinction. It would be very fruitful to have economic discourse agree on the distinction — to agree, for example, that the minimum wage is coercive — and then debate just where, why, and how much coercion is our friend.


[1] Daniel B. Klein and Stewart Dompe, “Reasons for Supporting the Minimum Wage: Asking Signatories of the ‘Raise the Minimum Wage’ Statement,” Econ Journal Watch 4, no.1 (January 2007): 125-167.

[2] James Gwartney and Robert Lawson, Economic Freedom of the World: 2006 Annual Report, (Vancouver: Fraser Institute, 2007); Tim Kane, Kim R. Holmes, and Mary Anastasia O’Grady, 2007 Index of Economic Freedom, (Washington, DC: Heritage Foundation, 2007).

[3] Jack High, “Is Economics Independent of Ethics?” Reason Papers 10, no. 1 (1985): 3-16.

[4] Daniel B. Klein and Stewart Dompe, “Reasons for Supporting the Minimum Wage”: 132.

[5] David D. Friedman, “A Positive Account of Property Rights,” Social Philosophy and Policy 11, no. 2: 1-16.

[6] Ibid.

[7] Block’s quip is reported and discussed in Gary North, “Undermining Property Rights: Coase and Becker,” Journal of Libertarian Studies 16, no. 4: 75-100 and developed in Walter Block, “Coase and Demsetz on Private Property Rights,” Journal of Libertarian Studies 1, no 2 (1997): 111-115.

[8] George J. Stigler, “Wealth, and Possibly Liberty,” Journal of Legal Studies 7, no. 2 (1978): 213-217.

[9] Friedrich A. Hayek, Law, Legislation and Liberty, Vol. 1, Rules and Order, (Chicago: University of Chicago Press, 1973).

[10] Friedrich A. Hayek, The Constitution of Liberty, (Chicago: University of Chicago Press, 1960); Daniel B. Klein “Mere Libertarianism: Blending Hayek and Rothbard,” Reason Papers 27 (2004): 7-43.

[11] George J. Stigler, “Smith’s Travels on the Ship of State,” History of Political Economy 3, (1971); reprinted in Stigler’s The Economist as Preacher and Other Essays, (Chicago: University of Chicago Press, 1982), pp. 136-145.

[12] Ronald H. Coase, “Adam Smith’s View of Man,” in Essays on Economics and Economists, (Chicago: University of Chicago Press, 1994), pp. 95-116.

[13] Adam Smith, The Wealth of Nations, (Indianapolis: Liberty Fund, 1981), p. 138.

[14] Thomas W. Merrill and Henry E. Smith, “What Happened to Property in Law and Economics?,” Yale Law Journal 111, no. 2 (November 2001): 357-398.

[15] Adam Smith, The Theory of Moral Sentiments, (Indianapolis: Liberty Fund, 1982), p. 80.

[16] Smith, The Wealth of Nations, p. 687.

[17] Smith, The Wealth of Nations, p. 324.

[18] Smith, The Theory of Moral Sentiments,, p. 318.

[19] Smith, The Wealth of Nations, p. 664.

[20] Smith, The Theory of Moral Sentiments, p. 327.

Daniel Klein is professor of economics at George Mason University and chief editor of Econ Journal Watch.

Response Essays

Coercion as a Political Concept

The Relevance of Coercion to the Defense of Markets

The arguments that can be offered for and against minimum wage regulation mirror the arguments that can be offered for and against markets generally. There are those, such as Robert Nozick, who find ethical foundations for markets in a particular view about individual rights.[1] Assuming that people have pre-institutional (moral or natural, not just legal) rights to liberty and property, market exchange is just what happens when people’s rights are respected. Since, on this view, all interference with voluntary exchange is prima facie wrong, the objection to minimum wage legislation is easy to see.

A less thoroughgoing moral defense of market exchange is also possible. Daniel Klein proposes a moderate libertarianism that tolerates infringements of property and liberty rights by the government so long as enough is at stake in terms of social welfare.

In stark contrast, however, are purely instrumental defenses of market institutions. A freer market is generally better than a more regulated one because it is more efficient, resulting in a bigger social pie. Such a position is accepted even by most egalitarians these days. Though pretax market outcomes will tend to be, from that point of view, unacceptably unequal, properly designed tax and transfer institutions can produce a better result than the direct regulation of economic interaction. Whether the regulation of wages is an exception to this general rule divides egalitarian economists. But their dispute is the purely empirical one of whether setting a floor to legal wages makes poor people better or worse off.

Instrumental arguments for markets may be put to the service of different views about the (morally) right social aims—ranging from the neo-utilitarianism of cost-benefit analysis to strongly egalitarian views influenced by John Rawls. But in all cases the defense of markets depends entirely on the factual question of how well they serve the proposed social aims, and not at all on the idea that people have pre-institutional rights to property and its voluntary exchange.

I venture that instrumentalists generally, both those for and against minimum wage laws, would be inclined to react with puzzlement to Klein’s question about whether such laws are coercive. In one natural sense of the word, after all, the entire set of institutions that support market interaction is coercive; criminal prohibitions on theft, for example, are coercive, as is taxation.

Implicit in this kind of reaction is one of two views. Some believe that there are no natural rights and so the only question is whether government, in general, does more harm than good. Others believe that there are natural rights that constrain government’s pursuit of social goals but the questions of which such rights there are, and what strength they have, are not ones economists are likely to have any special insight about. Economists’ expertise starts downstream of the specification of both the proper social ends and the moral limits to governmental means. (Torture by the police is out, even if it somehow could increase the size of the welfare pie, but economics for the most part takes such restrictions on means for granted.)

Klein rightly notes that this reaction is in one way too quick. The notion of coercion, or free exchange, is at the foundation of the instrumentalist case for free markets because the claim that exchange of property is mutually beneficial and thus increases the size of the welfare pie is plausible only if it is limited to cases of uncoerced exchange.

But the instrumentalist can reply that the notion of coercion she needs may overlap only accidentally with the notion relevant to setting (prima facie) limits on government means. All the instrumentalist is saying is that there are general classes of transactions that we can (somehow) identify in which the assumption of mutual benefit breaks down. Label those for which the problem seems to be informational cases of fraud, and those for which the problem seems to have something to do with the decisional capacity of one of the parties cases of coercion; in both cases, the contours of the notion are shaped by the question of when the assumption of mutual benefit no longer holds.

There need be no attempt, on this approach, to argue from some independent understanding of the concept of coercion to the assumption of mutual benefit in exchange. The instrumentalist economist does not need to claim that there is no morally significant independent concept of coercion, with important applications in other contexts. That’s just not the notion of coercion she has in mind when she writes that uncoerced exchange benefits both parties. Klein’s attempt to expose all economists as market moralists in denial can be resisted.

Unpacking Political Concepts

In suggesting that there is a single notion of coercion that can play a foundational role in many different areas of social policy, Klein opens up a rather unpromising arena of debate. Key political and moral concepts such as those of liberty, coercion, democracy, and the rule of law, are well-established battle-sites within political theory. Suppose someone argues that the American combination of constitutionalized individual rights and judicial review is undemocratic. Defenders of this practice would rather be able to say that the critic doesn’t understand what democracy is than to have to say that democracy is not the only thing that matters; so they offer an account of what democracy is that is compatible with judicial review.

Sometimes this kind of move is rather obviously an attempt to disguise what is really at stake. You take a concept that has positive connotations for everyone and then build a lot of controversial political theory into the content of the concept. The claim that unrestricted markets maximize liberty falls into this category, since this can be made plausible only if you define liberty in terms of respect for the very (alleged) rights that are the basis of the moral argument for the market.

Of course two can play this game, and if arguments about the minimum wage become arguments about the concept of coercion, it is not obvious where we will end up. It could be argued that any contract of labor that pays less than a certain socially recognized minimum is coercive because uncoerced exchange occurs only when people have a range of options consistent with some standard of minimally fair social interaction. Thus minimum wage legislation could be defended on the ground that it is necessary to prevent coercion by private parties. Of course the possibility of such an account of coercion is precisely why Klein believes that theorists of the market need to think harder about this issue. They need to get the concept of coercion right.

As it happens, philosophers have thought rather a lot about the concept of coercion, starting with Nozick’s classic 1969 essay, “Coercion.” Nozick’s final but still tentative analysis of coercion goes roughly as follows: A proposal (such as “I’ll kill you if you don’t give me your money” or “I’ll pay you $5 per hour for washing the dishes”) counts as coercive if the person to whom the proposal is made would prefer never to have received it in the first place.[2]

So on Nozick’s account, labor contracts for less than the minimum wage are not coercive. But it is also doubtful that their prohibition is coercive, since the relevant description of the government’s proposal seems to be this: “We will enforce agreements where the wage rate exceeds $X, but agreements for less than $X will not be enforced and in fact will be otherwise sanctioned.” This is better than nothing, as far as government enforcement of labor agreements is concerned, and so rational employers and employees would not prefer that the proposal had never been made.

Those who feel that something has gone wrong here might be tempted by another approach that Nozick discusses in his article, and which has been very influential. On the so-called “moral baseline” approach, we can count proposals as coercive if carrying them out would make their recipients worse off than they would be if everyone acted as they morally ought to act. A government’s proposal to enforce only certain labor agreements will count as a coercive, on this view, just in case governments ought to enforce all labor agreements.

Klein seems to have in mind some version of the moral-baseline approach to coercion. If I understand him correctly, he holds that coercion is interference with liberty or a proposal to do this, where liberty is in turn understood in the specific way necessary for the slogan that free markets maximize liberty.

There are two fundamental objections to be made to Klein’s project. The first is that the only method we have for analyzing concepts is the one Nozick used: we offer examples in order to prompt intuitive judgments about whether those cases involve coercion or not. Analysis of any concept is hostage in the first instance to people’s intuitions about correct usage. Klein tells us what coercion is, but he does not attempt to explain away the fact that many people have a very different view about the correct application of the concept. All concepts are indeterminate to some extent, but some are more so than others. The concept of coercion, I am convinced, is deeply indeterminate, with disagreement about correct usage tracking exactly the fault lines that have political significance; so there is simply no right answer to such questions as whether a labor contract for below a minimum wage, or its prohibition, is coercive.

But the more important point is that this doesn’t matter. What is really at stake in the argument over the moral foundations of the market, and therefore minimum wage laws, is what rights people have, how much weight they have, and what the proper aims of government are. It is at best misleading to disguise discussion about these issues as a debate about the proper understanding of the notion of coercion. In this respect, I side with Nozick. The argument for libertarianism in Anarchy, State, and Utopia does not trade in rhetoric about the maximization of liberty and the word “coercion” does not appear in the index.


[1] Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974).

[2] Robert Nozick, “Coercion,” in Socratic Questions (Cambridge, Mass. and London: Harvard University Press, 1997).

Liam Murphy is a professor of philosophy and law at New York University.

Coercive Regulation and the Balance of Freedom

Daniel Klein has written an elegant essay arguing that minimum wage laws are coercive. He is obviously right. These laws threaten employers with state-sponsored violence if they have a contractual relationship with wages that are too low. To me, the most striking fact in the essay was that more than fifty percent of a survey of economists said that these laws are not coercive in any significant sense. That’s just silly.

Minimum wage laws, like most restraints on trade and like tax laws, are enforced with the power of the state. Coercion lies at the core of almost all government policies. Rarely is voluntary participation a reliable tool for enforcing rules. If we could count on voluntary participation, we probably wouldn’t need the government involved in the first place.

But, as Klein notes, just because something is coercive, doesn’t mean that it is wrong. The coercive power of the state is useful when it protects our lives and property from outside harm. If we think that state-sponsored redistribution is desirable, then we are willing to accept more coercion to help the less fortunate. We also rely on state-sponsored coercion regularly when writing private contracts. The ability of creditors to collect depends on the power of the state to coerce borrowers.

The great difficulty is that coercion is both necessary and terrifying. For millenia, governments have abused their control over the tools of violence. The historical track record insists that we treat any governmental intervention warily. What principles help us decide on the appropriate limits to government-sponsored coercion? Are minimum wage laws acceptable coercion or do they fall outside of the pale?

I start with the view that individual freedom is the ultimate goal for any government. The ultimate job of the state is to increase the range of options available to its citizens. To me, this is not a maxim, but an axiom that is justified by both philosophy and history. On a basic level, I believe that human beings are the best judges of what is best for themselves. I also believe that the right to make our own decisions is an intrinsically good thing. I also believe that people become better decision-makers through the course of regularly making their own decisions. Moreover, the historical track record looks a lot better for governments that put freedom first. The liberal democracies, defined by their affection for liberty, have been far better for their citizens, than alternatives, whether Communist or Fascist, that enforced state-sponsored visions of how people should live their lives.

A belief in the value of liberty flows strongly through mainstream neoclassical economics. Economists frequently speak about an aim of maximizing utility levels, and this is often mistranslated as maximizing happiness. Maximizing freedom would be a better translation. The only way that economists know that utility has increased is if a person has more options to choose from, and that sounds like freedom to me. It is this attachment to liberty that makes neoclassical economists fond of political liberty and making people richer, because more wealth means more choices.

There is a recent wave of scholarship suggesting that the government can help individuals be happy by reducing their choices. While happiness may be a very nice thing, it is neither the obvious central desiderata for private or public decision-making. On a private level, I make decisions all that time that I expect to lower my level of happiness, because I have other objectives. On a public level, I can’t imagine why we would want to privilege this emotion over all other goals. A much better objective for the state is to aim at giving people the biggest range of choices possible, and then let people decide what is best for them.

But putting freedom first doesn’t mean abandoning the state. At the very least, we rely on the government to protect our private property against incursions by others. Even most libertarians think that it is reasonable for the state to enforce contracts. This enforcement increases the range of contractual options and this is, in a way, expands liberty.

While these forms of state action are readily defensible, many of the thorniest questions involve tradeoffs between the liberty of one person and the liberty of another. Taking wealth from Peter and giving it to Paul increases the choices available to Peter and decreases the choices available to Paul. Governmental coercion to redistribute income cannot be opposed purely on the grounds that it restricts liberty. Certainly, redistribution reduces the freedom of the taxpayer but it increases the options of the recipient of governmental largesse.

With this lengthy preamble, let me switch to the minimum wage and related restrictions on the ability to contract. The minimum wage reduces the options available to the employer, who must pay his workers more. It also reduces the freedom of both employer and employee, both of whom lose the ability to contract at a lower wage. Opposing this loss of freedom is an increase in the options available to workers who remain employed and now earn a higher wage. While I am no fan of higher minimum wages, I can imagine settings in which the increase in the freedom of the still-employed workers could be more important than the offsetting losses to individual liberty. We cannot get to a clear answer on the minimum wage on the basis of an axiomatic desire to increase the range of choices available to individuals, because we are trading one person’s choices against the choices of another.

Perhaps one might come to a clear view on the minimum wage by hewing to an uncompromising belief in freedom to contract. While I certainly have sympathies for that belief, there are cases where freedom to contract is and should be imperfect. For example, many contracts rely on expensive government enforcement, and it is reasonable to set limits on the scope of government action in this, as in every, setting. An employment contract with a lifetime non-compete clause, for example, relies on governmental agents enforcing a prohibition against a worker for decades. I’m okay with the idea that the government doesn’t need to spend its resources, which are, after all, our resources, enforcing every extreme contract. I am also in favor of limits on contracts that encourage highly anti-social activities, like murder. Support for the freedom to contract is a good maxim, but it cannot be an axiom.

The case against the minimum wage or other related restrictions on contracting does not, in my view, come from clear anti-coercion axioms or even maxims, but from other more technical reasons that have been emphasized for decades. If we want the state to redistribute income, we have sensible means for doing that like Friedman’s negative income tax or the Earned Income Tax Credit. These tax-based approaches are also coercive, but they can increase the choice set of the poor with less of a reduction in the freedom of others. Obviously, these tax-based solutions don’t restrict the set of available contracts and that is a great plus. The fact that American minimum wages are too low to create large-scale unemployment shouldn’t blind us to the fact that, across the Atlantic, far more aggressive minimum wages are accompanied by vast numbers of unemployed youths. The minimum wage is also bad redistribution policy because it imposes the costs of redistribution on the employers of the poor, and on their customers who will have to pay higher prices to make up for higher wages. If we want to redistribute income to the poor, then it is appropriate that everyone with resources pay, not just employers in sectors that employ the less fortunate.

A final reason to reject further increases in the minimum wage is that it gets the government into the business of setting prices, and this requires competence that seems far beyond the limits to government. The case for laissez-faire comes ultimately not from unbounded faith in the power of the market, but rather in a realistic appraisal of the limitations of government. Price-setting is a difficult task that is prone to enormous abuse. The historical track record of price and rent controls is pretty terrible. It seems like this track record should make us further recoil from further governmental incursions into setting prices.

By reminding us about the coercive nature of the minimum wage and other similar regulations, Daniel Klein also reminds us that the centuries of sagacious concerns about the abuse of government power also apply in this case. Perhaps we should use redistributive taxes that reduce the freedom of the wealthy to increase the freedom of the poor, but it is hard to think that the minimum wage is a good tool for redistribution.

Edward Glaeser is the Fred and Eleanor Glimp Professor of Economics at Harvard University

Voluntary and Coercive Action: A Key Distinction in the Overall System of Liberty

Justifying (Some) Coercion

I have just had the pleasure of reading Daniel Klein’s provocative essay that seeks to unpack economics from the tradition of classical legal philosophy. He starts with the fundamental distinction between coercion and voluntary transactions, and investigates the extent to which it can explain the attitude that economists take toward the minimum wage in particular, and the overall question of regulation more generally. In my view, his instincts are sound, but I think that his overall exposition could be more coherent and compelling if he took into account the traditional legal analysis of coercion that has fallen somewhat beside the way.

On the first question, Klein is right to be puzzled with the common response of economists that the minimum wage is not coercive in any significant sense. It is not, I might add, an answer to say, with Liam Murphy, that the question is odd because “the entire set of institutions that support market interaction is coercive; criminal prohibitions on theft, for example, are coercive, as is taxation.” The reason Murphy misfires is that the initial inquiry only concerns what kinds of actions are coercive by individuals in nature. We can easily concede that some sort of state coercion should be used in civil society, which Edward Glaeser rightly stresses in his response to Klein. Ultimately, the question at hand concerns the ends to which coercion may be put, and that inquiry can only come after we have a coherent theory of the uses and limits of state power.

Common contemporary responses to that question indicate a profound shift in how economists, and often philosophers, think about the world. The classical liberal tradition started from the naïve libertarian view that the sole function of government should be to control the operation of force and fraud, and to enforce voluntary agreements. Given this mindset, a concern with the definition of coercion proves central to, for example, the question of whether we can make sense of the Millian proposition that the sole end of the use of public force is the prevention of harm to others. The difficulty with this view is that an analysis of the meaning of the term “coercion” will not settle the question of what forms of legal arrangements are appropriate, for there are, as Klein notes and Murphy and Glaeser echo, a wide range of circumstances in which the use of coercion, either public or private, could be allowed.

The fuller analysis, which requires some retreat from the hard-line libertarian position, identifies the cases of what might be termed “justified coercion.” The unmistakable import of that position is that the use of coercion is something that needs to be justified, usually by some fairly convincing claim. The presumption in favor of liberty of action replaces the sacredness of liberty as the starting point of analysis. The loss of some sense of absolutism carries with it an increased complexity which always proves uneasy to someone like me, who believes in “Simple Rules for a Complex World.” But this sort of complexity has to be accepted to make sense out of a social reality that has never (even at the height of laissez-faire) treated the line between voluntary and coercive action as the one true litmus test for all occasions.

From Liberty to Utility

Modern economists do not start typically from the presumption of liberty. Rather, they are usually self-conscious welfarists who seek to maximize social utility by the use of a variety of policy instruments that include various forms of taxes, regulations, and subsidies. It is the rare economist today who would give some special pride of place to autonomy or liberty. It is not that economists as a group do not care about people. It is that they model people as arguments in a (collective) utility function that should be maximized subject to some constraint. On this view, a preference for markets or freedom might fall out of the equations, but it is not at the center of their intellectual universe. Hence, the methodology of economics has, in my view, led to a gradual separation of the field from its traditional pro-market roots. The technology has taken over the discourse.

With that said, we should ask how the minimum wage question should be approached if we wish to blend the best of the classical liberal approach, which sets the presumption of liberty against state intervention, and the modern economic approach, which starts with social maximization models. Here is one way this could be done.

First, we have to think of what coercion means in the easy case. A first approximation might be that the use of force against another individual always counts as coercion. There is much truth in this proposition, but it is a bit too broad, as there are all sorts of cases where force accidentally causes harm to others. Such conduct might be subject to legal sanctions, typically not in the criminal law, but the actions so sanctioned are not coercive. What marks off coercion is that it involves the threat to use force in order to achieve some conscious end. The most powerful forms of coercion are often those that require no actual use of force because the threat is so powerful that compliance is the norm.

Using this definition of coercion is meant to remove from the picture other kinds of threats that do not involve the use of force, of which the most common is the refusal to deal with individuals who do not do what you want. Many people have insisted that coercion does go this far, so that virtually all market behavior counts as coercive. However, when stretched this thin, the term loses its meaning. The refusal to deal with others in thick markets is what makes competition possible, and this broad definition of coercion does not see any difference between the various devices of persuasion at hand. This position has got to be a nonstarter: for if true, the class of voluntary transactions would be empty. If rejecting an offer is coercive, then negotiation itself becomes a dubious activity.

The intuition behind this notion of coercion does have some bite in other areas, however. The threat of a monopolist not to deal with others is more potent than it would be under conditions of competition – under conditions where alternatives abound. These complications have led to an enormous body of law dealing with common carriers and other “natural” monopolies. But this is not relevant to the minimum wage question as such; most labor markets are intensely competitive, with lots of players on both sides of the employer/employee divide.

The Coercive Nature of the Minimum Wage

The task is now to show why the minimum wage law is coercive, even if we reject the broader definitions of coercion. In the case of the minimum wage, the law governs a relationship between two parties, an employer and an employee, but the sanctions it imposes are directed only to one side of that relationship: the employer’s. It is easy to see why conduct is coercive when the threat of force is directed to only one person to get him to do a particular act. But in the case of the minimum wage, the broad regulation covers all employers, and is ostensibly (don’t believe it!) intended to protect people on the employee side of the relationship. Defenders of the minimum wage could offer three obvious grounds on which to distinguish this regulatory command from the core case of coercion — the threat of force made by one private person to another.

The first of these grounds is that the law is the result of democratic processes, so it does not represent the unbridled preferences of one person.

True enough. But this point is not decisive. If it was, no law supported by the political process could ever be coercive. Democratic support might justify legal coercion, but it hardly establishes that there is none.

The second ground is that the law is general in its impact while ordinary private coercion is directed against one individual or a small number of individuals.

Again, the point is hardly decisive. The broader sweep of the law suggests that it involves coercion against many, not that it involves no coercion. This still leaves open the insistent question of justification, at least within the classical liberal framework.

The third ground is that the use of state force is directed only against one side of the relationship, but since this benefits persons on the other side, it cannot be regarded as coercive.

Again, this cannot be correct. If I threaten A against doing business with B, then I have limited B’s options as well, even if I purport to protect him. Indeed, an old precedent says that if a gun-wielding competitor scares off the pupils of an “antient school,” the schoolmaster can sue for the loss of his customers, even though the gun was not pointed at him. The same rule should surely apply here. One limits the choices of employees by directing force against the employer. The more effective the sanctions against employers, the less willing they will be to extend offers to workers. The law necessarily limits choices on both sides of the relationship. Its actual benefits most likely go to outsiders who face reduced competition from firms hobbled by the minimum wage law.

The Presumption of Liberty Applied

Within the classical liberal framework, therefore, there is no way to escape the conclusion that the minimum wage law is coercive. From that conclusion follows the presumption that it should be repealed. The prohibition of certain contracts limits voluntary actions and hence the gains from trade to both employers and employees. The winners are not the parties to the transaction, but those who face less competition from the regulated firm.

It is quite possible that some firms will see their employment levels rise in consequence of the statute, namely those firms that depend less on minimum wage workers and more on other inputs unaffected by the regulation. But that shift in balance is not to the overall good. The standard economic models show that competition produces net gains to the participants of exchange, and these in turn generate further gains to third persons. Coercion destroys all those gains, such that a party to a potential transaction withdraws because the restriction has raised his costs above his expected gains. In general, coercion yields negative sum games. That troublesome proposition should be the working premise of any social welfarist, even if, sadly, today it is not.

So what is stage two of the argument? Here it can only be said that cases in which we can find external gains that justify the losses are special ones, and that the minimum wage isn’t one of them. We know that the floor on legal wages leads (if there is no strategic behavior) to supplies that consistently outstrip demand, with no way to figure out how to handle the excess. That disequilibrium either eliminates gains from trade or it induces individuals to use various inefficient devices to escape or minimize the law. Administrative costs go up. Selective enforcement is likely in practice. Legitimacy is undermined.

The situation with the minimum wage is not like one in which the limitation on the ability of competitors to cartelize an industry is (or may be) justified by the higher overall levels of output that come with the return to competition. Nor is it justified on the grounds that small people cannot defend themselves in the marketplace, given the wide range of choices that are left open by free entry of other potential employers who will bid up wages if the current set of employers are garnering supracompetitive returns.

When Then Coercion?

Thus far I have applied the classical liberal model to only one case. But clearly in other situations there have to be some justifications for the use of coercion, and the question is how to frame them. Here, look back at the list of necessarily coercive state functions that were identified by Murphy and see how they relate to each other and the minimum wage. The prohibition against theft is the use of state coercion to prevent the private threat of or use of force. To the extent that it reinforces libertarian norms, it is just the kind of rule that every state should put in place. And, notwithstanding the vagaries of modern welfare economics, that is exactly what happens. Murphy next mentions the use of coercion to support market institutions, and notes that this is also coercive. True enough, but coercion here is justified because it expands the envelope for gains from trade through voluntary exchange, especially in sequential exchanges that take place over time, for which legal enforcement is critical. And where exchange does not serve that end — as with contracts to achieve murder or theft — then we outlaw these contracts as a form of conspiracy. We fear conspiracy precisely because it allows two individuals together to wreak more misery than they could alone. But with most contracts for the sale of goods and services, the social externality is positive, which is why they are enforced.

Last, there is taxation, which is a form of coercion: hand over some sum of money or we shall seize your property. And the justification of taxation is surely more nuanced because there is little doubt that it can be used as an instrument of theft, as libertarians are fond of saying. But they typically overstate their categorical case against taxation, because they refuse to look at the benefit side of the equation. At one extreme we should condemn those forms of taxation that allow a despot to enrich his private coffers at the expense of the public at large. At the other extreme, we understand that the prevention of theft and the enforcement of contracts requires social institutions strong enough to stop private coercion, but not so strong as to put illicit state coercion in its place. Therefore, if a tax is designed to provide some public good, and we can say all persons who pay some tax receive a benefit that they value in excess of the amount of the tax, then the arrangement is a form of justified coercion, as long as coordination problems make it impossible to achieve that same favorable distribution of cost and benefits through voluntary means.

This approach can be applied to other arrangements. Klein quotes Walter Block who protests, “Coase, get your cattle off my land!” The clear implication of Block’s position is that the Coasean framework will undermine the system of property rights by introducing a level of relativism into causation so that no one knows who is responsible for what. The system of property has as one of its key components the right to exclude, which the Coasean analysis of reciprocal causation tends to undermine, if not used very carefully. But Klein is right to note that systems of open range have been used, whereby the owner is under a duty to fence out cattle. But he does not ask this follow-up question: does the open range system produce gains to all landowners exceeding the value of the right to exclude and compensating for its loss? Oddly enough, if land is cheap, extensive and arid, and suitable only for a single use, then a system that allows everyone to let cattle roam could well make each landowner better off than before. But once multiple uses become feasible for land of increased value, this system will die, because no one will make investments if he must first get the consent of all. So what we need is an equilibrium analysis which shows when such open-access systems are likely to emerge and survive.

None of this is new. No less a figure than Adam Smith noted that once all the world was a commons for nomads and cattle, until the rise of agriculture made the right to exclude highly valuable to cultivation and long-term use. In a sense, Smith answered a question — the mirror image of Block’s — normally associated with Pierre-Joseph Proudhon, who famously said “property is theft.” Proudhon would ask this question of Block: How dare you make this property, which has always been common, private? And the answer, in all cases, is: because it is to the long-term advantage of all. Understanding property rights, in fine, does not lead us away from understanding utility. It allows us a principled approach to understanding how they operate, why they are needed, and why they incorporate into the analysis a distinction between coercion and voluntariness that the skeptics on the minimum wage law ignore.

Richard A. Epstein is the James Parker Hall Distinguished Service Professor of Law at the University of Chicago and the Peter and Kirsten Bedford Senior Fellow at the Hoover Institution, Stanford University.

The Conversation

Liberty and Semantics: Response to Murphy, Glaeser, and Epstein

I am grateful to Will Wilkinson, Brink Lindsey, and the Cato Institute for organizing this exchange on the semantics of liberty, or the distinction between voluntary and coercive action.

I think that, by and large, if our society were freer, we would have better housing, food, and healthcare.

But I feel sure that paramount in life are interesting and meaningful lives, loves, losses, and friendships. Better stuff matters mainly as inputs to better culture, including household micro-culture. Culture is the meaning we attach to signs, symbols, and actions. It is paramount to social welfare.

The paramount things vary according to semantic content. Social improvement may come by better semantics, even if public policy isn’t any better. Given the policies we endure, better culture means we can better communicate and commiserate. We can better overcome cultural separation.

So, the next time someone says, “Oh, but that is just a semantic issue,” kindly smack him upside the head for me (an instance of justifiable coercion).

But the quality of semantics does affect the quality of public policy. Confucius spoke of downward movements: “When words lose their meaning, people will lose their liberty.”[1] As for the upward movements, better semantics will mean less coercion.

So I am delighted that three top-flight minds and mine are exploring the semantics of voluntary/coercive on the worldwide web. I am very grateful to Liam Murphy, Ed Glaeser, and Richard Epstein for the confabulation.

I want to confess. I am one of those who think that our political culture has deep-seated cleavages, and that it is far too statist. My view basically follows the Hayekian narrative [2] : First, evolution made us collectivist (and in that sense modern collectivism is natural [3]), then there emerged a normative and institutional revolution called liberalism, and then came the reaction of collectivist tendencies in the form of nationalist and social-democratic cultures, creeds, and institutions. Yes, I see much of modern academic culture as deeply wrongheaded. There are very few Milton Friedmans and Richard Epsteins. In academe, standard formulations and official modes of discourse effectively work to deny the idea of liberty and the integral voluntary/coercive distinction (the syndrome in economics is noted by Richard). Much of political culture represents, or at least accommodates, the cultural reaction to liberalism. Particularly after 1890, there occurred an epochal alteration of the liberal lexicon. Much of it was essentially subversion (see Confucius).

Liberty has become taboo in academia, politics, and polite society in general. Embarrassed by invoking it soundly, politicians invoke the term primarily when they push for something that is actually its contravention.

Yet the distinction remains as vital as Adam Smith, William Lloyd Garrison, and myriad others, and lives proudly and loudly among institutions like the Cato Institute.

In their cable TV series, Penn & Teller recently looked at illegal immigration. The program was predicated on the distinction, and the moral of the story appealed to a partiality towards liberty. The makers banked on viewers getting it. Their apparent success, like John Stossel’s, seems to indicate that they guessed right.

I offer other evidence. In teaching introductory economics, I quiz students on locating examples within the following classification:

Voluntary Coercive
Legal (A) (B)
Illegal (C) (D)

I clarify the person/property basis—“this is my hand”—the focus on coercion as the initiation of force or fraud (or threat thereof), and that an activity is to be dubbed “coercive” if any of the parties involved are so coerced, and otherwise “voluntary.” I locate several examples for them. Then I quiz them. I ask them to locate examples— “smoking pot,” “paying workers less than the minimum wage,” “playing chess with a friend,” “mugging someone,” etc.

College students catch on fine. They get it. They also get clearly that in cell (C) the government initiates coercion. The distinction is intuitive and coherent.

We see the distinction in use all around us. Without it, words like “the free market” and “intervention” mean nothing. But we rarely see it right on the surface of the examining table.

Before the subversion of the liberal lexicon, one often saw socialists and social democrats speaking quite candidly about the compulsory and coercive nature of their proposals. The culture was in transition.

The plea of classical liberals/libertarians is to own up to it. We are not saying it is absolute. We are not saying that it alone is focal. But we ask that laws against consensual activity be recognized as initiating coercion.

I believe that participants of this exchange all believe that semantics are worth arguing over. And all reject deontological grammars of the desirable (deon in Greek mean duty or obligation). We all recognize and count moral and cultural consequences, and we all embrace the “loose, vague, and indeterminate” nature of making big calls—like those made by Admiral Adama in Battlestar Galactica.

But of course that does not diminish the wisdom of consecrating and sticking close to simple rules, natural rules, especially in a complex world. Richard Epstein is a great figure, and I am honored by his participation and reassured in our concordance.

Where Richard addresses “the third argument,” a clarification may be useful. Does the minimum wage coerce any employees? I think Richard would agree that the answer is “no.” It is important to keep “coercion” clearly rooted in the threat or use of force against the coercee’s property/person. In characterizing the law in relation to employees, I like to call it “step-coercion.” But, just as your step-mother is not your mother, step-coercion is not coercion to the step-coercee.

I second much of what Ed Glaeser writes, and like the separation of happiness and utility. But some of his word usage gives me pause. Although he never takes issue with me, he uses liberty and freedom in ways at odds with the semantics I follow. I hope that Ed wants to be like Milton Friedman and Richard Epstein, but, if so, he still needs to make some important steps in reconnecting to the classical liberal tradition.

Liam Murphy’s essay is thoughtful and civil, and more fundamentally at odds with classical liberalism. Good and brave of him to join a Cato discussion in which he is outnumbered.

At one point, Liam seems to suggest that the definition of “voluntary” is contoured by mutual benefit. But it should have been clear that I based it on ownership and consent. He writes as though I would insist that cash for dope cannot be disadvantageous for either party. But that is something on which, like Smith, I certainly would not insist.

It seems to me, that, basically, Liam casts doubt on the coherence and focalness of the distinction I am following. His supposedly “relevant description” of the minimum wage (“We will enforce agreements where …”) is fraught with problems and does way too much. Basically, Liam is saying that liberty and coercion don’t really mean much of anything. Again, I think the denial of the distinction falls back on the idea of the state as encompassing overlord—we are tenants of its property. Is Liam ready to bite that bullet?

And yet, quietly, Liam employs the distinction. He writes of “unrestricted markets” and the “case for free markets.” If the meaning is not per the distinction, then what is it? Despite his argumentation, I bet that as a college student Liam would have aced the voluntary/coercive quiz.

On Robert Nozick: In Anarchy, State, and Utopia, he used “aggression” where others use “coercion” (a choice that I have agonized over). Nozick is a funny figure. In many ways, he supplied a brittle caricature of classical liberalism, and then promptly withdrew from political philosophy, deeds that brought him great renown.

In closing, I want to emphasize the general agreement: We all reject axiomatic libertarianism. If we could agree on the relevant semantics, we would all agree that sometimes coercion—contravention of the liberty principle—is what each of us would favor.

That sensibility, further, is now typical of libertarians. It becomes more and more unfair and irresponsible to characterize libertarians as people with an axiom to grind.


[1] Quoted by F.A. Hayek, The Fatal Conceit: The Errors of Socialism, (Chicago: University of Chicago Press), p. 106.

[2] The narrative suggested here broadly follows Hayek’s works, e.g., “The Atavism of Social Justice,” in his New Studies in Philosophy, Politics, Economics, and the History of Ideas, (Chicago: University of Chicago Press, 1978); The Mirage of Social Justice. Vol. 2 of Law, Legislation and Liberty, Chicago: University of Chicago Press); and The Fatal Conceit.

3 On the idea that many of our evolutionarily selected instincts are anti-libertarian, see Hayek, “Atavism;” Paul H. Rubin, Darwinian Politics: The Evolutionary Origin of Freedom, (Piscataway, NJ: Rutgers University Press, 2002); Bryan Caplan, The Myth of the Rational Voter: Why Democracies Choose Bad Policies, (New York: Princeton University Press, 2006); and Daniel B. Klein, “The People’s Romance: Why People Love Government (As Much as They Do),” The Independent Review 10(1), Spring 2005: 5-37.

Strategies of Cooperation

I am pleased that Daniel Klein has added yet further wrinkles to our discussion, on which I shall comment briefly on two.

First, there is much truth in Paul Rubin’s provocative suggestion that ordinary individuals evolved in group settings that were not entirely congenial to markets. The point of this observation is that families are bound together by strong genetic ties such that each person within the group takes into account the utility of others as though those utilities were their own. That is a consequence of the common genetic origins. What this proposition entails is that we can expect to see some higher level of cooperation among these individuals than we would expect to see among strangers, so that there is less need for, and less likelihood of, a price system emerging to mediate differences among members of family groups. That proposition is as true of the modern family as it was of our prehistoric ancestors. And it suggests that there is little in these forms of cooperative behaviors that help people reach the kinds of voluntary markets that are needed among strangers.

But by the same token, this is an incomplete account, for even if most interactions take place within the context of a family, there are always some that have to take place among strangers, and for these there are only two ways to proceed, by coercion or consent. The evolutionary impulses on this choice are not all that clear. If there is dominance and you can get what you want, then you will take it, or so the theory goes. But that strategy has serious risks because if the aggression fails, then defeat could happen in the individual case, or, even if the attackers “win,” they could easily be weakened and thus exposed to real losses at the hands of third persons in some subsequent encounters. The cooperative strategy may not give the same instantaneous uptick as aggressive behaviors when they work, but it does not have the same larger downside. So there are pressures toward cooperation in stable arrangements among strangers as a way to avoid the ultimate confrontation. And those instincts also have some genetic base, I suppose, which means that the willingness to engage in market-like behavior is not wholly foreign to the species.

On the normative side, of course, the choice is generally easy: cooperation produces joint gains, and aggression one-sided ones. We therefore should move strongly to endorse the former over the later, and shift that presumption only in rare cases. It is this basic insight that explains why the classical liberal position is so attractive. It is the only political philosophy that, if consistently maintained, promises a steady, relatively risk-free course of human progress.

The Meanings of “Liberty”

I do agree with Daniel Klein that “semantics” is worth arguing about, but only in the sense that it’s good to know what people mean and it’s good to be on the lookout for conceptual sleights of hand.

There are at least three important notions of liberty (and it is liberty, I think, not coercion, that is at the heart of this discussion). The classical liberal notion is simply the negative liberty of being able to do what you want to do without interference. This is by no means the notion of liberty required by libertarians in their attempt to construct a political theory around the idea of liberty. What is needed for that is a moralized conception of liberty according to which my liberty is not infringed when I am prevented from violating your rights. This account of liberty, however, moves us directly to the controversial issue of which rights we have. Nozick saw this clearly and, as I have already indicated, I agree with him about the true location of the debate.

The third notion of liberty is one usually reviled by libertarians as the root of all evil. So it is pleasing, in this place, to see Edward Glaeser invoking it whole-heartedly. Glaeser’s discussion is built around the idea that social policy should be guided not by the utilitarian aim of promoting welfare, but by the aim of promoting positive liberty. On this notion of liberty, people are freer the more options they have available to them. The reason this is anathema to libertarians is that it implies that transfers in wealth increase the liberty of the recipient. When we add in the diminishing marginal contribution money make to positive freedom, the implications of Glaeser’s view are clearly egalitarian.

These are three fundamentally different ideas associated with the same word. So, again, it is obviously important to be clear about meaning. But where I disagree with Klein is that he appears to believe that we can establish that one of these senses of the word is the semantically correct one. My own view is that “liberty” is simply an ambiguous word. But even if this is wrong, it is obvious that nothing of significance for political theory turns on the correct meanings of words. Once we know what we mean, we can turn to the substantive issues. In this sense, the complaint, “that’s just a semantic issue” is entirely reasonable.

The obvious substantive issue in dispute between libertarians and the rest of us (including Glaeser?) is the little matter of whether there are pre-institutional property rights and whether, if there were, actual legal holdings would bear any relation to them. This is obviously not the right time to enter into this debate. But let me just respond to Klein’s suggestion that those of us who believe that property rights are entirely conventional must therefore think that the government owns everything and doles out little bits of stuff to citizens as a kind of charity. This absurdity misunderstands the rejection of pre-institutional property rights. On the view first laid out by that great conservative philosopher, David Hume, there are no natural property rights, so the government can’t have them any more than individuals can. To find out who owns what, we have to look at what the law is.

What We Mean by “Liberty” — and “Wealth”

Let me just make a few comments to the various objections that Liam Murphy has raised to the definitions of “liberty” in use in classical liberal theory. I do not think that the term “liberty” has the deep philosophical ambiguity that he attributes to it. A person does not become more free because he has more wealth; he becomes wealthier, which confers on him more opportunities to use the liberty that he has. Any system of wealth transfers from one person to another, moreover, cannot increase aggregate liberty, because if the transferees are made, by Murphy’s lights, better off, then the transferors are made worse off, so that we do not have the kind of social improvement that is derived, say, from voluntary contracts that typically produce gains to the parties and positive external effects as well.

The question therefore is whether we can justify the limitation on the classical liberal definition of liberty on the grounds that wealth transfers produce some kind of social improvement. That improvement cannot be found on a strong Paretian standard because of the losses to the parties from whom the transfers are made — an objection that cannot be made against voluntary gifts. The harder question is whether the case for transfers can be made on a Kaldor-Hicks standard whereby we think that the gains to the winners exceed the losses to the losers, such that in principle transfer payments could bridge the gaps. This is possible with changes that produce overall wealth increases, but is much harder to achieve with wealth transfers; even with the diminishing utility of wealth, we have to hand everything back to the loser to put that person back in the prior state of affairs.

The only way out of this problem is to assume that from some ex ante position there is a global insurance contract whereby everyone thinks that some protection against adverse states is better than no protection at all. Yet here too the obstacles are severe because we know that, in the real world, transfer payments are costly to administer; that they provoke partisan squabbles that produce political strife; and that they dull the incentives for production at both ends of the income scale. Another way to put the point is this: even if we thought that the diminishing marginal utility of wealth justified coercive transfers, the question we would have to ask is not whether $100 in the hands of the poor is worth more socially than $100 in the hands of the rich, assuming interpersonal comparisons of utility are permissible. Rather the question is whether $100 – X in the hands of the poor is worth $100 in the hands of the rich, and until we know X that is hard to answer. My own sense is that there is no categorical answer to this question, but that politically the transfers will move either in the wrong direction, or will be larger than is optimal. There is no categorical case against redistribution as a deviation from libertarian conceptions of liberty, but it is a hard row to hoe to find transfer systems that do work. The minimum wage is not one of them.

Extra-Legal Institutions and Classical Liberal “Liberty”

Liam writes:

The obvious substantive issue in dispute between libertarians and the rest of us … is the little matter of whether there are pre-institutional property rights …

Is Liam trying to boil it all down to one dichotomy, namely, either: (1) the governmental legal rules that actually exist, or (2) nonsense-upon-stilts (Jeremy Bentham’s epithet for natural rights)?

If so, that is ill-considered.

I too reject the idea of a grammar of absolute deontological rights, prior and invariant to human experience and institutions. Such a view resembles a theological theory of rights. I have faulted the libertarian school of Murray Rothbard for tending to see it that way (something for which Richard and I knocked Walter Block). I favor Rothbard’s definition of liberty (and would credit him with defining it more thoroughly than anyone else, primarily in The Ethics of Liberty), but I do not favor the claims he made for liberty.

But if Liam is saying that there are no other important distinctions to be made, if, more specifically, he is saying that the voluntary/coercive distinction doesn’t really cohere and pertain, then I disagree. I am concerned that Liam wants to caricature all libertarians as Rothbardians (or perhaps ASU Nozickians) and then limit his critique to that caricature.

The main problem with the view that there are only two things, actual governmental legal rights and nonsense, is that it denies all other distinctions and categories that work to refine and advance our understanding and evaluation of rules. Rules and principles emerge from sources other than government. They often exist and function apart from and even in opposition to the government rules (e.g., black markets, as Walter Block has highlighted, are often heroic, and Robert Ellickson’s classic study of ranching in Shasta County, California). Moreover, rules and principles can exist in our minds even when contravened in institutional practice. The most pervasive and important constable is the impartial spectator. In fact, she too is an institution of sorts.

Regarding “positive liberty,” like Richard, I urge people not to call human capabilities “liberty.” Otherwise, you have to say that catching the flu is a diminution of liberty. It cheapens the term. It dilutes and confuses the important classical liberal meaning of liberty, a tradition to which we owe so, so much. I see Liam’s expansive usage of “liberty” as part of that semantic subversion mentioned previously. I disagree with his statement: “it is obvious that nothing of significance for political theory turns on the correct meaning of words.” Semantic confusion can destroy ideas and distinctions that have been crucial in the cultural traditions within which our theorizing is embedded.

Liam writes: “The classical liberal notion is simply the negative liberty of being able to do what you want to do without interference.” If he means to say that the classical liberal notion of liberty holds that sanctions against murder are a diminution of liberty, I say he’s wrong. You find things to that effect in Hobbes, Bentham, and Austin, but not in the people who define classical liberalism—e.g., Locke, Sidney, the Scots, the American founders, von Humboldt, the French liberals (many generations), the Abolitionists, J.S. Mill, Cobden, Bright, Spencer, Morley, and so on. What you find there is quite contrary to what Liam claims. The classical liberal notion of liberty is what Liam says the libertarian notion is (though we would have to be careful about “rights”). If there is one man who forms the bridge between classical liberalism and modern libertarianism, it is Ludwig von Mises, and, on the definition of liberty, he represents only continuity.

A quick follow-up to Richard’s latest:

Hear, hear. State redistribution coerces the taxed party and it does not enlarge anyone’s liberty (it makes them wealthier), and hence clearly contravenes the classical liberal/libertarian conception of liberty. I mostly oppose the welfare state, but the case against it is nothing like a silver bullet. A dollar means more to a poor person than to a rich person.

The Innocuous Indeterminacy of “Liberty”

My main point, that the concepts of coercion and liberty are indeterminate along politically significant fault-lines, seems innocuous enough; but it is not being received that way in this discussion. One wonders why this is, since this simple and hardly deep point is not in itself an objection to anyone’s political theory.

The business end of the libertarian position can be very clearly and unambiguously stated. People have natural property rights and the right to enter into agreements as they see fit. Violation of these rights, as with any rights, requires special justification. And, yes, we can rephrase all this, without fault, in terms of a particular idea of liberty. The idea of liberty in question requires a theory of rights for its statement, but that’s all right, since libertarians already have one of those.

What is the threat in saying that this particular sense of liberty is distinct from others in wide currency and that those others are no less linguistically or conceptually valid or correct? I can only assume that it comes from the difficulty of carrying the audience once the various different senses of “liberty” are untangled and the dependence of the libertarian sense of the word on a controversial theory of rights is brought out into the open. Far better to try to convince someone that libertarianism just follows from taking seriously the liberty he has always loved—the liberty, what’s more, that was also loved by such profound and profoundly different figures as Locke, Hume, Mill, and Rousseau. (OK, maybe not Rousseau.)

So I repeat that the real issue between libertarians and their opponents is what rights people have and what our social goals should be. I do not believe that when it comes to rights it’s either all conventional or all nonsense. I believe that there are, in fact, real natural, moral, preinstitutional rights, certain rights against aggression being among them. But I don’t believe that property rights are among them.

What about promissory or contractual rights? There is no plausible general right to freedom of contract; no one really believes that I have a right to enter into an agreement of any content I like. The general recognition of the value of freedom of contract is qualified by our ability to identify certain classes of agreements that, as we believe, typically retard our social goals or are in some way ethically objectionable. One might like to say that there is a right to freedom of contract, but it can be outweighed by considerations of social policy. But if a net social gain is sufficient to justify overriding this “right,” it is not worthy of the name.

In the end, a purely instrumentalist approach to the minimum wage issue seems inescapable. (Here I echo some of what Edward Glaeser wrote.) Not only is this not an issue where the concepts of coercion or liberty can play a useful role, it is not even an issue that requires us to think about the relative significance of rights within social policy. There are no rights involved. It all comes down to the question of whether a minimum wage makes poor workers better off, and whether we care about that.

Classical Liberal “Liberty” and the Dangers of Indeterminacy

Liam once again raises the proposition that libertarians and classical liberals do not have any special claim to the use of the term “liberty,” but only can claim use of one peculiar sense of the term, which then has to do battle with others. I disagree with that position. The proposition that each person is entitled to have the maximum liberty consistent with the like liberty of others is a position that only makes sense within the framework of the classical liberal system. The moment that increases in wealth through state transfers count as increases in liberty, then there is no way to preserve the like liberty between persons. The newer definition of liberty is one which lets one side be better off and the other worse off, which is, after all, what transfer systems are supposed to do.

And the difference does make a difference. The principle of freedom of contract is not idle because we believe that it can be abridged with the use of rules that improve both sides to the transaction in proportionate degree, as is often the case with formality. Nor does it become otiose if we assume that certain contracts, such as horizontal restraints on trade, should not be enforced universally because their negative effects outweigh their positive ones. The initial baseline that joint gains from contracts count as a social good is morally defensible and socially wise. No other baseline can have that result.

So much I think follows from Murphy’s last sentence, where he says, “It all comes down to the question of whether a minimum wage makes poor workers better off, and whether we care about that.” That is not the correct formulation of the problem, for it ignores the possibility that the improvement of (some) poor workers is a justification for a program that produces widespread social losses, both to other workers, some of whom are also poor, and to firms and their shareholders, and to customers and suppliers of the firms in question. The danger of Murphy’s excessive nominalism is that it now directs us to the wrong social inquiry by allowing gains to some preferred clientele to offset far greater losses to another group of individuals. Anyone who starts with a presumption in favor of freedom of contract will not make that mistake, which should be reason enough to avoid the dangerous claims of “indeterminacy” which all too often cloud social theory.

If We Must Have “Rights,” Don’t Forget the Subscripts

If one says that the minimum wage is coercive, is one saying that the law violates the only true rights (e.g., freedom of contract)?


But Liam continues to write as though someone in this exchange says “yes.”

If the law says employers may not pay less than $X, then, obviously, in one significant sense, employers do not have a right to pay less than $X. In another significant sense, they do have a right to pay less than $X.

Furthermore, the law has conferred a right, of sorts, on workers, to not have competitors work legally for less than $X. In that sense, repealing the minimum wage would strip workers of certain rights.

So “rights” come in different sorts, and emerge from different sources. If we want to do this conversation in terms of “rights,” then we need to put a subscript i on “rights,” making it rightsi, where i would correspond to the many different senses or sorts. And we would need to get used to finding them in conflict, and without any one sort being supreme or exclusively “real.”

In that sense, “rights” are highly plural, conflicting, indeterminate.

But the distinction between “voluntary” and “coercive” neither depends on nor necessarily implies a singular well-ordered scheme of rights. Indeed, we can do the distinction without the word “rights” at all (I think I managed pretty well in my essay, for example).

Pretending that the distinction does depend on or imply a singular well-ordered scheme of rights is one way to attack liberty as indeterminate. But it is only a pretense.

Where’s the Nominalism?

I doubt that Richard Epstein misunderstands the point I keep making, but he writes as if he does:

The proposition that each person is entitled to have the maximum liberty consistent with the like liberty of others is a position that only makes sense within the framework of the classical liberal system. The moment that increases in wealth through state transfers count as increases in liberty, then there is no way to preserve the like liberty between persons. The newer definition of liberty is one which lets one side be better off and the other worse off, which is, after all, what transfer systems are supposed to do.

Of course, if one is talking about positive liberty, as Edward Glaeser did in his contribution, one will not offer the slogan, “maximum liberty consistent with the like liberty of others.” Glaeser’s essentially consequentialist argument was that (positive) liberty should be optimized overall. But if Epstein does not like this view, he can hardly blame Glaeser’s definition of liberty. The argument can be rephrased in neutral terms: Maximize individual options for choice. Epstein does not like this because he is committed to a libertarian system of rights which implies that a society’s resources are not a common pool available for just (re)distribution.

The danger of Murphy’s excessive nominalism is that it now directs us to the wrong social inquiry by allowing gains to some preferred clientele to offset far greater losses to another group of individuals. Anyone who starts with a presumption in favor of freedom of contract will not make that mistake, which should be reason enough to avoid the dangerous claims of “indeterminacy” which all too often cloud social theory.

This is classic Epstein: few can aspire to this kind of rhetorical density.

First, I have argued that there is ineradicable conceptual indeterminacy in connection with coercion and liberty. This is not nominalism. I have not said that any of the senses in use refer to nothing real. Second, an acknowledgement of conceptual indeterminacy cannot direct us to any social inquiry; for that we need a normative claim. I personally endorse redistribution away from laissez-faire returns; but I could certainly do that even if I agreed with Epstein and Klein on the meanings of “coercion” and “liberty.” Last, the only indeterminacy I am claiming is conceptual—I am not claiming any kind of moral or social scientific indeterminacy.

As for the indeterminacy I am claiming, neither Epstein nor Klein has written a word about how it might be argued against. Exactly what method of conceptual analysis do they propose? I fear the method is the following: define terms to suit the rhetorical needs of your argument, and then insist that this and only this is what those terms mean.

A final point. I do start with a presumption in favor of freedom of contract. What I reject is that there is a right to freedom of contract. Despite what he writes, Epstein rejects such a right too. Epstein long ago defended infringements of rights when this can be said to benefit all concerned. Perhaps such rights are worthy of the name, since it isn’t just overall social gain that justifies the infringement, but benefit to each. But when it comes to prohibition of horizontal restraint of trade, benefit to each is not to be assumed, and yet Epstein, unsurprisingly, is in favor of infringement of the “right” to freedom of contract in such cases too. But now we see that what is called a right gives way when there is social gain to be had from its infringement. If having a right means anything, however, it means that the relevant interests are not to be sacrificed just when the positive social effects outweigh the negative. If there is nominalism in this discussion, it is to be found in Epstein’s account of rights.

The Centrality of Liberty

There are two good reasons that liberty should be central in any discussion of public policy. First, freedom is the best candidate available to be the central goal of social policy. Second, any sensible policy discussion recognizes that emphasizing liberty provides a needed safeguard against the excesses of government power. While a philosophical discussion of natural rights gives us little clear guidance on what to do about regulations, like the minimum wage, these two central reasons for valuing freedom point to a pretty clear policy conclusion against such regulations.

Why should freedom by the primary goal of government? There are really two alternatives for the proper objective of government: liberty and everything else. The advocates of liberty argue that individuals, not the state, should be the primary arbiter of how to live their lives. They argue that private individuals make better decisions than governments and that individual decision-making is inherently good. This view does not give any clear answers about tradeoffs between the freedoms of two people, but it does put liberty first. The alternative view is that the state should select objectives, like psychological well-being, longevity, or racial purity, and push the population towards those objectives. The case for liberty is as much a case against these alternatives. The case for liberty is best made by emphasizing the imperfections of governmental decision making, and a historical track record where liberal democracies have far out-performed the alternatives.

Accepting that the goal of government is to maximize the range of choices available to private individuals tells us little about cases where there is a tradeoff between two peoples’ choices. But regulations like the minimum wage are not problematic because they increase one individual’s choices at the expense of another, but because they redistribute choices badly. If the goal is to take money from some people and give it to others, then standard tax-based redistribution accomplishes this more effectively, without restricting the ability to freely contract and to employ less skilled workers. The minimum wage is bad not because it redistributes choices across people, but because it reduces freedom for everyone, at least relative to a better redistribution system.

The second reason to treasure liberty is the need to be vigilant against the abuse of power. Every expanded regulation provides yet another opportunity for the state to put special interests ahead of the public. Every increase in redistribution will be accompanied by inefficiency and inequity. Without a clear and compelling reason to increase the coercive

acts of government, our bias should be against more intervention. It is hard to see why the need for the minimum wage is so vital that it leaps over this threshold, especially when other forms of redistribution exist.

Whether or not one can spin philosophical justifications for this type of regulation, it is hard to see how this is a sensible way to redistribute income that puts the burden fairly on all taxpayers, instead of on the employers of less skilled labor and the consumers who buy their products.