The Machinery of Freedom
by David Friedman

The most well written introduction to libertarianism except for Rothbard's. It is much better than I had been led to expect by the reviews I read of it. The reviewers emphasized what is wrong with it, but failed to note that it is well written, well organized, full of good ideas, and persuasive.

David Friedman sees problems with both the libertarian natural rights approach to law and the utilitarian approach. Each leads to unacceptable conclusions in some situations. For example, the libertarian natural rights approach, which holds private property as an inviolable right, leads to the unacceptable conclusion that you have no right to steal A's rifle--even if it is the only way to stop B from murdering C, D, E, ... Z. The utilitarian approach handles this particular problem more satisfactorily by justifying your theft of A's rifle on the grounds that the benefit of saving all those lives outweighs the harm done by the theft--using the criteria that you should maximize happiness and minimize harm (171-173). A more significant difference between the libertarian natural rights approach and the utilitarian approach shows up in the issue of using missiles armed with hydrogen bombs as a means of national defense. The libertarian natural rights approach does not permit it, but the utilitarian approach can justify the threat of mass murder.

The utilitarian approach, however, is unsatisfactory in other situations. For example, suppose, after a series of brutal murders, an angry mob is ready to lynch three innocent suspects (A, B, and C) and the only way to stop them is to frame D, who is also innocent. Would you frame D? The utilitarian answer is definitely yes. David Friedman's answer is no, which means he is not a pure utilitarian┘he believes the immorality of punishing an innocent person in this situation overrules the utilitarian principle (180-181).

Despite David Friedman's refreshing candor about the imperfections of the utilitarian approach and despite his personal belief in libertarian ethics, he has concentrated his talents on evaluating laws from a utilitarian perspective, using economic theory as his primary tool. Since he doesn't really believe in utilitarianism as a moral philosophy, he devotes very little attention to defending it philosophically. He adopts it because it lets him reduce moral philosophy, which makes him uncomfortable, to economics, which he enjoys.

He wants to develop arguments to persuade others to become libertarians, and he believes utilitarian arguments are more persuasive than arguments based on natural rights because: (1) More people believe that happiness is good than believe in the libertarian theory of natural rights. (2) Economics is more well developed, more scientific, and more likely to make progress than the field of moral philosophy. (3) He knows more about economics than he does about moral philosophy (181-182).

He could use the same arguments to defend a physics approach to law. After all, more people believe in physics than believe in the libertarian theory of rights, and physics is a more well developed and scientific field than moral philosophy, and David Friedman has a Ph.D. in physics, so he probably knows more about it than he does about moral philosophy. Why did he choose an economics approach to law rather than one based on physics? This is a silly question, but let's answer it anyway. Economics is more relevant to law than physics is, because economics and law both deal with purposeful human action, but physics does not.

But if it makes more sense to analyze law from an economist's point of view than from a physicist's point of view because economics is more relevant to law, then why doesn't it make even more sense to analyze law from a moral philosopher's point of view, since law is a subset of moral philosophy?

I think David Friedman is like the drunk who lost his wallet in the middle of the block but is searching for it under the streetlight at the corner because he can see better there. He explicitly denies this charge, but not convincingly. He argues that even though utilitarianism is not true it may still be useful to libertarians because: (1) There is a close correlation between libertarian rules and rules that make people happy. (2) By figuring out what legal rules would be economically efficient we may learn what rules would be generated in a free market. (3) Demonstrating that a particular legal rule tends to increase human happiness is a strong argument for it. (Even if you do not care about the sum total of happiness, you still care about your own happiness. So if a particular legal rule increases the average happiness, there is a good chance it will increase your own happiness.) (4) He has no way to settle disagreements about values, so he bases his arguments on happiness, which is a value that most people share (184-185).

I am in favor of economic analysis of law, as long as it is not seen as a substitute for moral philosophy. Economic analysis of law can show that libertarian principles are efficient, and, although that is not a sufficient reason for adopting libertarian principles, it is, at least, a reason for not rejecting them.

Also, knowledge of economics might prevent some bad laws from being passed. For example, if everyone understood that minimum-wage laws either have no effect (when they set the minimum wage below the lowest market wage), or they create unemployment (when they set the minimum wage above the lowest market wage), the general public would oppose such laws, and their elected representatives might repeal such laws if labor unions do not have more influence with them on this issue. This is one case where utilitarian principles plus knowledge of economics lead to the same conclusion as the libertarian philosophy that you have the right to make contracts concerning your own property.

Another reason for applying economic analysis to law is that if you believe in the utilitarian principle of maximizing happiness, it is incumbent upon you to understand economics so you can avoid applying the principle incorrectly. This may explain why so many economists are utilitarians-utilitarians have a moral obligation to become economists. However, since economics is a separate field from law, whereas moral philosophy includes law, we should expect to find that economics is of relatively less use than moral philosophy in analyzing law.

The Limitations of Economic Analysis

Economics is worthwhile, but it cannot answer questions about what the laws should be, because economics as a science is neutral with regard to all moral principles, laws, and values.

When economists advocate the free market because it is efficient, they are no longer acting as value-free scientists, and they are not providing an alternative to moral philosophy. They are advocating the libertarian principles of justice.

The concept of the free market has moral principles built into it. The free market is based on private property rights, freedom of contract, voluntary consent, fulfillment of personal commitments, and honest dealing as opposed to theft, fraud, extortion, and cheating. So, whether they are aware of it or not, when economists analyze the free market, even though they avoid using words that sound moralistic, they are nonetheless assuming a society based on libertarian moral principles.

Champions of the free market who think they can avoid making moral judgments by stressing how efficient the market is and by avoiding such words as right, wrong, justice, guilt, responsibility, and honesty, are mistaken. By advocating the free market for any reason, you are advocating a system of justice that cannot operate unless people make libertarian moral judgments about being honest, keeping your commitments, stealing, cheating, and so on.

Utilitarian economists who concentrate on economic analysis of laws are more explicit than other economists about the fact that they are analyzing moral rules. At least they admit they are wading into moral philosophy, even though they haven't learned how to immerse themselves in it and swim.

David Friedman was a bit confused about this when he wrote:

"As a moral philosopher I am a libertarian, insofar as I am anything. As an economist I am a utilitarian." (183)
Strictly speaking, the last sentence is nonsense. It's like saying, "As a physicist I am a Republican." Utilitarianism is a moral philosophy. Economics is a value-free science. As a moral philosopher David Friedman can be a utilitarian or a natural-rights libertarian, but as an economist, as a scientist, he cannot make value judgments about his findings or advocate any policy whatsoever.

It would make more sense, and I think it would be true, if he said: "As a moral philosopher, I actually agree with the libertarian theory of rights and the non-aggression principle, but, for the reasons previously stated, I pretend to be a utilitarian. And since I am pretending to be a utilitarian who judges the morality of all actions according to their long-range consequences, I have to be an economist so I can predict what the consequences will be."

The purpose of economic analysis of law is to make logical inferences about the short- and long-term economic effects of laws. The purpose of utilitarian analysis of law is to promote laws that result in economic improvement and to oppose laws that make things economically worse.

David Friedman is excited by the fact that by starting with the goal of economic efficiency, the economic analysis of law produces what look like ethical rules such as "thou shalt not steal." It is as if he thinks he has found a way to assess laws without having to use moral philosophy. But this is a mistake. Economics as a science is neutral with regard to whether economic efficiency is desirable. Once you add the goal of economic efficiency to your economic analysis, you are crossing over from value-free economic science into utilitarian moral philosophy. It should not be surprising that you can get ethical rules from a moral philosophy, and it is nothing to get excited about.

The Limitations of Utilitarian Analysis

Although most people believe happiness is a good thing, they do not accept the utilitarian philosophy that maximization of happiness is the greatest good and that it outweighs all other good things. Not everybody accepts the libertarian theory of rights, but almost everybody accepts some theory of rights, at least at the emotional level, and they believe that justice is not the same thing as economic efficiency.

More than 200 years ago Adam Smith explained that utility is not the source of moral approval even for those who come to realize that virtue and utility can coincide:

When those authors describe the innumerable advantages of a cultivated and social, above a savage and solitary life; when they expatiate upon the necessity of virtue and good order for the maintenance of the one, and demonstrate how infallibly the prevalence of vice and disobedience to the laws tend to bring back the other, the reader is charmed with the novelty and grandeur of those views which they open to him: he sees plainly a new beauty in virtue, and a new deformity in vice, which he had never taken notice of before, and is commonly so delighted with the discovery, that he seldom takes time to reflect, that this political view having never occurred to him before, cannot possibly be the ground of that approbation and disapprobation with which he has always been accustomed to consider those different qualities. (Adam Smith, in British Moralists Volume 1 pp. 321-322.)
Most people make moral decisions based on what they believe is deserved or not deserved, not on the basis of efficiency. When they do make decisions based on efficiency, they distinguish those decisions from moral decisions. When they think others are making decisions based on economic efficiency, they become suspicious about the morality of those decisions.

Most people regard economics as a dismal science. It is beyond their comprehension. They cannot follow the long chains of reasoning that utilitarians use to reach conclusions. Therefore, they cannot base their moral decisions on utilitarianism. They need a simpler, more direct, philosophy. See my article Normal People Believe in Natural Rights.

These facts limit the effectiveness of utilitarian arguments.

Another limitation on utilitarian analysis is that we don't know how to measure happiness. In attempting to measure happiness, economists tend to count things that are countable, such as money, and they tend to omit things that cannot be counted or measured, such as emotions. When you realize that happiness is more of an emotional state than a financial state, our inability to measure and compare emotions is a particularly significant defect in a philosophy whose goal is to maximize happiness.

Utilitarians need a unit of measurement for happiness that they can use in mathematical equations so they can determine whether a policy increases or decreases the sum of happiness in society. They use money for this, not because money is happiness, but because money is well suited for mathematical calculations and happiness isn't.

In some of their calculations, utilitarians assume that a dollar represents the same amount of happiness to everybody. This means that a man who has $200 is exactly $50 happier than a man who has $150. Most people do not believe this is necessarily true, but utilitarians ignore the likelihood that it is false, because they have no other way to measure happiness. Is this scientific?

David Friedman refers to Alfred Marshall's argument to justify substituting dollars for happiness in approximations of total utility. The argument is that the approximation in dollars should be a good one as long as we are considering large groups of people and have no reason to believe that, on average, one large group values dollars differently than another. That makes sense. But it restricts the applicability of utilitarianism to large groups, which means it isn't an appropriate philosophy for judging individual cases.

Economists infer information about a person's values from the way he acts in the marketplace. When a person freely exchanges A for B, he demonstrates that at the moment of the trade he would rather have B than A, which means B was more valuable to him. Utilitarians draw the conclusion that every voluntary trade increases the happiness of both parties to the trade. This assumes either (1) everyone's subjective values are objectively good, or (2) our goal is to maximize subjective happiness rather than objective happiness. The former assumption is almost definitely wrong. The latter assumption is less compelling, and it further assumes that each trader's expectations about his own satisfaction with the trade are infallibly correct. That is, it overlooks the fact that traders sometimes make mistakes and are actually disappointed and unhappy about trades. We can't find out about bad trades very easily, so utilitarians ignore them. Is this really a scientific approach? I don't think so, and I think conclusions based on unrealistic assumptions are suspect.

Utilitarian Punishment versus Retribution

David Friedman uses cost/benefit analysis to address the difficult problem of assigning punishments for crimes. I believe he is more successful using this approach than philosophers have been with the natural rights approach. I have argued in other articles that the natural rights philosophy has totally failed to demonstrate why one punishment is more appropriate than a thousand other possible punishments for any particular crime. See The State as Penalizer, The Anticrime Industry in a Free Nation, and Dialog: The Market for Punishment.

The natural rights approach to punishment is based on retribution, which is the idea that it is right to punish a person, as an end in itself, if he deserves it. There are many theories of retribution, but they all have a few ideas in common: (1) Punishment must be deserved. (2) When punishment is deserved it should be imposed. (3) When punishment is not deserved it should not be imposed.

Utilitarianism, on the other hand, does not make any of these assumptions. It may turn out that these are good rules in many cases, but if it is sometimes more efficient to punish the innocent or let the guilty go unpunished, then that is what utilitarianism requires.

For the most part, David Friedman ignores retribution in his analysis. (Although in his paper "Should the Character of Victims and Criminals Count?" in Boston College Law Review XXXIV No. 4, he shows wide-ranging familiarity with the legal, economic, and philosophical scholarship on punishment.) Instead, he regards deterrence of crime as the only justification for punishment, with reform of criminals and restitution to victims as possible by-products of deterrence. (The reason for wanting to deter crime is that, according to utilitarian assumptions, crime generally results is less overall happiness.) This means the problem is to find the optimal schedule of punishments--the schedule of punishments that will deter the largest amount of nonproductive crime at the least cost. Some of the factors that go into the calculation are the costs of the crime and the punishment, the probability of being punished, the elasticity of the supply of particular kinds of crimes, and the costs of administering the crime-reduction system.

He shows that he is aware of many complexities that have to be considered in calculating the correct punishment. For example, he discusses the possibility that a dollar might be less valuable to a rich man than a poor man and that a day in jail might cost a rich man more than a poor man. He draws out some of the implications of these possibilities in calculating punishments. He also considers the possibility that murder victims vary in their value to society and that this should be factored into the equations. But he doesn't say how we can measure these things other than by trial and error.

In David Friedman's scheme, the punishment needs to be more severe when the probability of being caught and punished is low, and less punishment is needed when the chance of being punished is high. This can lead to a schedule of punishments that most people would regard as unfair. For example, in one of his papers David Friedman and his co-author reached the following conclusion:

Where a criminal is choosing between two alternative crimes but where some criminals may prefer (punishment aside) the less serious of the two, the optimal schedule of punishments might punish the less serious crime more severely... (David Friedman and William Sjostromi, "Hanged for a Sheep--The Economics of Marginal Deterrence" in Journal of Legal Studies vol XII, June 1993.)
Some other implications of the utilitarian theory of punishment are unacceptable to many people. For example: Although the mathematical formulas for calculating efficient punishments include unrealistic assumptions, and we usually do not know the actual numbers to plug into the equations, we can still approach the optimal schedule of punishments by trial and error. We could simply try a schedule of punishments, see how well it deters crimes, adjust it accordingly, see the results, and adjust it again and again, until we can't improve it any more.

If we do not believe people have rights, this approach to punishment is rational. It sets a clear and measurable goal for punishment (to deter nonproductive crimes), and by freeing us from the restraint of respecting people's rights, it allows us to experiment with different schedules of punishments until we arrive at the optimal one.

Unfortunately for this theory, most people reject the underlying assumption that they have no rights, and most people reject some of the results that follow from it.

So even though utilitarianism can provide a more rational basis for adopting a schedule of punishments than theories of natural rights can, it cannot find this schedule of punishments in a morally acceptable way, and this schedule of punishments itself would contain anomalies that violate our sense of fairness.

Anarcho-Capitalism versus Libertarianism

One theme that runs through The Machinery of Freedom is that anarcho-capitalism is independent of libertarianism. David Friedman envisions anarcho-capitalism as a free market that includes private protection agencies competing for customers and making arrangements with each other about which private courts to use to settle disputes. His vision of anarcho-capitalism includes private courts that use competing legal principles. In other words, anarcho-capitalism implies competing systems of law.

Since there are competing systems of law, some of the systems of law, if not all of them, will differ from libertarian principles, and it is legitimate to ask, "How libertarian would an anarcho-capitalist economy be?" To answer this question, he considers what would be likely to happen if puritanical people used private courts that treat sins, such as pornography, prostitution, and use of narcotics, as crimes.

Laws against such victimless crimes are anti-libertarian, but under anarcho-capitalism as David Friedman sees it, people would be free to hire protection agencies and courts that attempt to enforce laws against these activities. But, on the other hand, those who enjoy sin would be free to hire protection agencies that use courts that recognize the rights of people to engage in victimless sins. Which set of conflicting laws would prevail?

According to David Friedman, the answer depends on the relative economic strengths of the pro-sin and the anti-sin customers. That is, it depends on how much each group is willing to spend to get the kind of laws they want. If the anti-sin group outbids the pro-sin group, the laws against sin will prevail, and vice versa. Since the laws against sin harm sinners more than they benefit puritans, sinners would be willing to spend more per capita to make sin legal than puritans would be willing to spend to make it illegal. So he predicts that the laws under anarcho-capitalism will tend to be libertarian. (I find it interesting that in this utilitarian framework it is necessary to assume the libertarian natural rights definition of crime in order to understand what he is talking about. Other "crimes" have no victims, do not reduce overall happiness, and should not be deterred.)

As an example for analysis, he singles out heroin use. He predicts that under anarcho-capitalism the laws about heroin would be tailored to fit the demographics of local markets, because that would maximize returns to protection companies. In areas where there are very few heroin users and very many puritans, the cost of enforcing laws against heroin use would be low, because there would be nobody on the other side bidding to have heroin use legal. In metropolitan areas where heroin users tend to congregate, the cost of outlawing heroin would be bid up so high that the local puritans probably would not be willing to pay the costs of prohibition. So heroin use would be legal in places like New York City. (127-128)

As a libertarian, David Friedman would prefer to use private courts that base their judgments on libertarian legal principles. But he believes that under anarcho-capitalism, which he regards as the best possible economic system, there would be no guarantee that all courts, or even the majority of courts, will use only libertarian legal principles. He believes the only way that libertarian legal principles would be adopted by all the courts in an anarcho-capitalist economy is if virtually all people in the society are philosophically libertarian, and they spend their money for law and protection services accordingly.

I think he is correct in saying that libertarian legal principles will not be adopted in all the courts in a society unless the people in that society are libertarians. But I think he is mistaken in his belief that a purely anarcho-capitalist economy could exist in a non-libertarian society. In so far as an anarcho-capitalist economy includes courts that make decisions based on non-libertarian principles, the economy is not anarcho-capitalistic. It could still be anarchic, but it is not capitalistic in the relevant sense of the word, because if courts and the protection agencies that employ them use legal principles that justify prohibition of victimless crimes, those courts and protection agencies are part of a conspiracy against free trade with regard to those victimless crimes, and to the extent they are successful in prohibiting victimless crimes, the market is not free and the economy is not capitalistic. In other words, an economy that is purely anarcho-capitalistic can only exist in a society whose laws are purely based on libertarian principles.

Joseph Salerno made the same point years ago in a review of the first edition of The Machinery of Freedom:

In essaying to banish ethics from the purview of his analysis, Friedman has effected a monstrous bifurcation between anarcho-capitalism and libertarianism. He posits an anarcho-capitalist society in a political and ethical vacuum, and then goes on to analyze law "production" in economic terms, blithely unaware of his transgression against the most elemental dictates of common sense. For it is absurd to assume the existence of the economic institutions of anarcho-capitalism outside the politico-ethical framework of libertarianism. An objective, libertarian legal code, predicated on the Spencer-Rothbard axiom of nonaggression, and its acceptance by a large proportion of the populace, is the sine qua non of the establishment of anarcho-capitalism. Viewed in this light, Friedman's attempt in chapter 31 to adduce proof that anarcho-capitalism would be libertarian is at best supererogatory. (The Libertarian Forum, December 1973. p. 5.)
It should come as no surprise that an economic analysis of the laws developed by the free market would reveal that these laws are compatible with libertarianism. We don't need to understand economics to predict this outcome. Libertarian laws have to be established and enforced in order to have a free market in the first place. If the market develops any other kind of laws, it won't be a free market anymore.


Economic analysis of law can give us some interesting insights. For instance, it can show that a free-market economy is efficient and that crime can be deterred optimally by developing a schedule of punishments through a trial-and-error process that ignores individual rights. But it cannot provide any reason for a person to favor efficiency or optimal deterrence of crime versus other things. People form their opinions based on their personal values. They cannot get their values from the value-free science of economics, but they can clarify or even change their values by studying moral philosophy. That is why moral philosophy is more important than economics in deciding what the laws should be.

Year Read: 1974, 1990

Back to Libertarian Essays by Roy Halliday
Back to Nonfiction Book Notes
Back to Fiction Book Notes
Back to Book Notes by Author

This page was last updated on October 3, 2011.
This site is maintained by Roy Halliday. If you have any comments or suggestions, please send them to