The Dissident
Donate | Contact | Masthead | Discuss


The Myth of Mandates

The Thoughtless Orthodoxy

The Priority of the Visible: How Democracy Empowers Terrorists

Ditching the M-word

Straight Talk about Economic Illiteracy

The Poverty of Good Intentions

Buddhist Capitalism

Theory Gets a Reality Check: Power, Money, and a Little Bit about Love

The Politics of Innocents





books + ideas + provocations
Theory Gets a Reality Check: Power, Money and a Little Bit about Love

discuss this article

I. The Limits of Political Philosophy

The mainstream of contemporary political philosophy is embodied in A Theory of Justice (1971), by the late John Rawls. Most readers find in Rawls’s book an attempted philosophical rationale for what Peter Berkowitz recently called “the progressive, liberal welfare state.”

According to this interpretation of Rawls, his book justifies the “redistribution of wealth to achieve a substantially more egalitarian society” than is possible under capitalism. This is ecause Rawls judges the basic institutions of a society according to whether they will benefit its worst-off members. And interpreters of Rawls’s book—including Rawls himself—assume that, by this standard, laissez-faire capitalism is unjust; and that something at least as intrusive as the modern state is needed to correct its injustices.

But this interpretation of Rawls is a little too quick. Even if Rawls’s criterion of justice is correct, it produces only questions about which institutions are just—not answers. Rawls’s questions boil down to this one: Which economic system is best at helping the poor?

Answering Rawls’s question requires knowledge of the real-world effects that various economic systems have on poverty. And that isn’t knowledge that’s available to philosophers, qua philosophers—such as Rawls. Rawls the man concluded that his theory justified government redistribution of capitalists’ wealth, but this was just his personal opinion. His interpreters have generally leapt to the same conclusion, but this bespeaks their unfamiliarity with alternative ways of looking at institutions. The whole question of what institutions follow from Rawls’s theory has been treated superficially.

Since what is at issue are economic institutions, taking Rawls’s question seriously requires a realistic version of economics (Part 2, below). And, since departures from laissez-faire capitalism aimed at bettering the condition of the worst-off would have to be achieved politically, it also requires a realistic version of political science (Part 3, below).

The realism of the economics and politics I will discuss is premised on making human ignorance their centerpiece. The solution to ignorance that I’ll propose is to maximize people’s power to take actions that might improve their situation without thinking about, let alone knowing, what they are doing.

As my inspiration I will take the blind but often effective operations of love; as my aspiration, justice as defined by Rawls. My goal is to see whether unfettered capitalism, by better dealing with ignorance than does politics, achieves Rawlsian objectives more effectively than “the progressive, liberal welfare state.”


Rawls’s best-known philosophical opponent, Robert Nozick, was much more knowledgeable about both economics and political science than Rawls. Yet Nozick, too, was a philosopher (his Harvard office was right down the hall from Rawls’s), and his case for capitalism made no use of this knowledge. In the 1960s, Nozick had started out on the left. He was a member of Students for a Democratic Society (SDS), who assumed (as Rawls did) that being compassionate about the poor meant being on the left. Nozick became a libertarian only in grad school, when a friend told him about arguments against the feasibility of socialism that had been made in the 1920s by an Austrian economist, Ludwig von Mises.

Hearing about these arguments led Nozick to do something people rarely do: he took “the other side’s” ideas seriously enough to actually read them. And as a result of his reading, Nozick switched sides. He became a proponent of laissez-faire capitalism.

In such books as Mises’s Socialism; Peter L. Berger’s The Capitalist Revolution; Walter Williams’s The State against Blacks; Deepak Lal’s The Poverty of Development Economics; Thomas Sowell’s Markets and Minorities; and Peter T. Bauer’s Equality, the Third World, and Economic Delusion, Nozick found political and economic arguments against what he—along with so many well-meaning students and professors—had assumed was true: that capitalism is evil at worst, neutral at best; and that it must be supplemented, tamed, or overthrown in the interest of justice. The listed books’ very titles convey the “practical” nature of the considerations they adduced against those still-popular assumptions. (Some of the actual books Nozick read are out of print, so my list is updated while reflecting the gist of the midcentury free-marketeers’ case.)

These books contend that only a few hundred years ago, almost everyone was as poor as residents of the Third World still are; and that relatively laissez-faire capitalism—not labor unions, not the welfare state, not progressive social activists—made the First World rich. Moreover, government intervention in the economy, and political activism in support of that intervention, while aimed at achieving Rawls’s goal—helping the poor— unintentionally ended up hurting the poor.

Nozick might have elaborated on the arguments made in such books in his response to Rawls: Anarchy, State, and Utopia (1974). “The best way to help the most impoverished,” Nozick might have written, “would be to turn capitalism loose.” But to make such a claim stick, he would have had to address matters of economic and political fact—which, as a philosopher, he did not want to do. Instead, Nozick attempted to combat Rawls on conceptual grounds. As a result, Nozick’s book made no practical arguments for the justice of laissez faire— i.e., for thinking that untrammeled capitalism best helps the poor. From the halls of a philosophy department, Nozick was no better able to answer the question of which institutions best help the poor than Rawls himself had been.

Let’s backtrack, then, and put ourselves in the position of Nozick the open-minded left-wing grad student who is about to begin his iconoclastic reading. Let’s treat Rawls’s question as legitimate, in other words—but treat the answer to it as open. Instead of pre-empting an examination of the effect of capitalism on the poor, let’s engage in such an examination in the only way possible: by studying economics and politics to shed light on the real-world tendencies of different institutions.

2. Economics They Don’t Teach at Harvard

Martin Feldstein’s introductory course in economics is the only class listed in the voluminous Harvard catalogue that has its own organized opposition. Demonstrators hand students walking into Feldstein’s classroom leaflets protesting what they are about to be taught.

Part of the reason is that Feldstein is a conservative who headed the Council of Economic Advisors under Ronald Reagan and now advises President Bush. But are the protests just another instance of political correctness run amok—a case of students who are outraged to hear any challenge to the notion, endorsed by virtually all their other professors, that capitalism (along with imperialism, racism, and sexism) is evil; but who, unlike Nozick, aren’t willing to hear what “the other side” believes?

In large part, the ongoing anti-Feldstein movement is just that. But the protesters also have a legitimate complaint: that Feldstein, like the vast majority of economists, uses a model of capitalism that is patently unrealistic.


In Feldstein’s quite standard “neoclassical” model, everyone is motivated by self-interest. Consumers self-interestedly pay the lowest price possible for whatever they buy, after comparing the various goods for sale against their own hierarchically ranked desires. Similarly, in the pursuit of profit, producers provide consumers with exactly what they want, and will thus pay for. The desire for profit makes producers channel their self-interest toward serving selfish consumers’ desires. As Adam Smith put it in The Wealth of Nations, “It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.”

Economists like Feldstein take Smith’s generalization and run with it—too far. Although they are careful to note that their model is an abstraction from reality, in practice, they treat the abstraction and reality as interchangeable. Only by assuming that people are always selfishly motivated—and only by adding the further assumptions that producers have perfect knowledge of what consumers want, and that consumers have perfect knowledge of what producers are selling—can economists depict as neat blackboard diagrams the Smithian transformation of self-serving behavior into behavior that serves others. These diagrams are required if economics is to acquire the appearance of precision: at the intersection of blackboard supply and demand curves is he perfect price for a product.

In this rendition, the advantage of capitalism equals the extent to which it approximates a world in which blackboard supply/demand curves meet in equilibrium—the result of perfect competition in every market, based on self-interested producers’ perfect knowledge of the demand that will bring them profits, consumers’ perfect knowledge of the supply offered by producers, and an infinite number of competing producers driving prices down to the supply curve/demand curve intersection points.

All it takes for a student protester to have a gripe against Feldstein is to notice the extent to which the real world departs from that model. If all capitalism had going for it is that it allows the convergence of knowable supply and demand curves upon the price for each product sold in a perfectly competitive market, then capitalism would have little going for it indeed—at least for residents of the real world.

The protesters themselves go too far, though, in their repudiation of capitalism. The fact that most economists have taken a turn toward mathematicized mythology does not entail that all of their conclusions are unsound.

Consider what would happen to economics if its unrealistic assumptions were discarded. First let’s drop the assumption that entrepreneurs have perfect knowledge of consumers’ “demand curves.” As a consumer, isn’t it true that you often don’t know what you want until you stumble across it? And aren’t your “preferences,” far from being hierarchically fixed (let alone both fixed and known to entrepreneurs), often plastic enough to be shaped by all sorts of environmental influences, such as advertising and fashion?

I edit a scholarly journal, Critical Review. Even though it is a nonprofit, I try to maximize the revenues that subscriptions bring in, so that I need to rely on philanthropists for as small a portion of the budget as possible. (Note to philanthropists: your help is needed anyway!) In setting Critical Review’s subscription price, however, I can only guess what would happen to total revenue if the price were higher (or lower) by 1 or 2 or 5 or 10 dollars. Would the higher price be offset by fewer subscribers? The only way to tell is to try it and see what happens.

Nor could I ever do more than that, no matter how sophisticated my market research. Until potential subscribers are faced—in a real, historical time and place—with the possibility of actually subscribing to Critical Review, they themselves can’t possibly know (as opposed to guessing) what price they would be willing to pay for it—if any. Even if I were to give possible subscribers a list of specific circumstances that might affect their decision about whether to subscribe, the factors that influence one’s decisions are too many and too opaque for me to be able to list them all, or for consumers themselves to predict how those factors would work themselves out.

It is true that, in guessing the optimal price to charge for the product I’m selling, I am, in effect, guessing at the shape of the demand curve for it. But the notion that this curve is more than speculative goes too far. Demand curves are about as realistic as the romantic“types” we construct from past experience: one might imagine that she is looking for someone tall, dark, handsome, etc.—until along comes someone who, defying her expectations, captures her heart. Our busily constructedtypologies may serve a retrodictive function and, if we’re lucky, a predictive one—but only if we’re lucky. The same goes for capitalism—which raises the possibility that capitalism works in the same blindly fortuitous way that love does.

Next, let us discard the reverse assumption: that consumers have perfect knowledge of what businesses want to sell them. Do even the most conscientious shoppers—those, for example, who study Consumer Reports—really know what they’re buying before they buy it, let alone what better deal they might have gotten elsewhere? In an uncertain world—the real world—the answer is surely no. As with potential mates, you have to experience a product in order to know what you’re really getting, and whether and how much you’ll like it; and as with unmet potential mates, all the products you haven’t encountered remain mysteries.

Advertising can be helpful in diminishing consumer ignorance, because it can alert consumers to the goods being sold by the advertisers. (Draw your own comparisons to love.) But even the barrage of commercials to which we are subjected is not sufficient—even if they were accurate!–to make more than a dent in our ignorance. Our lives are too short, our brains too small, and our experiences too restricted for us to be truly expert in more than a few subjects, if any. Considering the nearly infinite array of things we might want to buy, it’s impossible for us to be expert consumers. Such expertise would require us to be almost literally omniscient not only about what we “prefer,” but about the price at which we might buy it from among all the sellers in the world.

Along with the assumption of perfect consumer knowledge of supply, and the assumption of perfect producer knowledge of demand, we can jettison the assumption of universal self-interest.

Among consumers, self-interest is no more universal than omniscience. A parent supporting his children; an art-lover bidding for a painting; a scholar buying a book—are these people acting out of self-interest? Admittedly, actions taken in pursuit of love, beauty, or truth can be put under the rubric of “self-interest”—but only if we define that term so broadly that every possible human action counts. And this renders the term meaningless.

Likewise for the producers of what consumers buy. Some production, rather than being driven by the pursuit of profit, is accidental. The founder of Apple Computer didn’t tinker with electronics in order to become a billionaire; he was simply having fun. Likewise, many people do their jobs not primarily out of financial self-interest, but to serve a political cause, or for love of the activity itself, love for someone to whom they donate their earnings, love of beauty, or love of truth. Self-interest does not begin to capture what producers do—unless, of course, self-interest is not a desire for financial gain, but an empty tautology.


Now we can take a preliminary stab at the question of whether laissez-faire capitalism is the system that, as the Rawlsian criterion of justice would require, best helps the poorest. The answer offered by a realistic form of economics may initially seem to be negative.

The inequalities among, and limitations of, real-world human beings are reflected in the limitations of realworld markets. Some people are born unfairly handicapped, or impoverished, or to unloving or incompetent parents. As a result, they may have little or nothing to sell that others are willing to buy.

What is for sale is not perfectly known to consumers, and what consumers will buy is not perfectly known to producers. Because of imperfect knowledge and unpredictable motives, market processes are never perfect; nor is competition; nor are outcomes. People make mistakes in capitalist economies, and those mistakes can be disastrous. Producers and investors who guess incorrectly about customers’ “demand curves” may go broke. Likewise, consumer desires may go unrecognized; a finite supply of producers cannot meet those desires as quickly and cheaply as economists’ idealized models assume; and path dependency may lock in relatively inefficient ways of meeting consumer demand.

Given these imperfections, can anything be said in favor of capitalism?

Something indeed can.

Once we’ve subtracted the assumptions that make the standard model of economics so mathematically elegant, so susceptible to chalkboard diagrams, and so vulnerable to legitimate student complaint, we are left only with what Albert O. Hirschman famously labeled “exit.” Exit is the ability of individuals to leave (and conversely to enter into) any “deal,” for any reason—or for no reason at all.

Exit is exemplified by a consumer’s ability to purchase something that’s for sale. Apurchase may be a mistake. But exit is the power not to repeat a given purchase—and that ability is essential.

Sometimes, a producer will be fortunate enough to hit on a product that consumers like enough to buy again and again. The producer will stay in business, and the consumers will get what they want. Neither perfect producer knowledge of consumer demand nor perfect competition ensures this result. It is ensured by the ability of consumers to exit from whatever a producer is selling, and by the ability of producers to exit from producing an unpopular product.

This market process creates a flood of desirable material goods, and it is exit that ensures that these goods benefit not only producers, but the consumers who continue to buy from them and the employees who continue work for them. Consequently, the vast wealth that we in the West have come to take for granted has assumed the form not only of profits for the few, but of rising standards of living for the many. The former cannot occur without the latter.

While it’s possible that a successful producer knows or cares what consumers want, it is also possible that he doesn’t. Arestaurateur may go into business for the sheer joy of it or out of family tradition, not in order to please his customers or make a profit. If he does make a profit, he may think it’s because of demand for the type of food he serves, when perhaps his customers really just like the ambience. The customers themselves may not know why they like his restaurant. No particular set of motives, and no particular knowledge, is required of anyone.

All that matters is that if enough customers don’t like the restaurant—whatever the actual reason may be, and whatever the reason the restaurant was opened—its owner will not make a profit, because his customers will head for the exits. In this way, the operation of the exit mechanism tends to select for behavior that does satisfy real people’s actual wants, and to select against behavior that doesn’t. The overall trend of laissez faire would therefore be to weed out enterprises that don’t satisfy people’s wants, and to weed in enterprises that do.

The option of exiting from any exchange one finds undesirable is advantageous because it is as undemanding of people’s knowledge, and of their reasoning abilities, as it is of their good will. The neoclassical economists, with their heroic assumptions about the levels of knowledge and competition that would have to characterize a beneficent capitalist economy, and their oddly anti-heroic assumptions about the motives of those who’d populate that economy, have in both cases put the emphasis backwards. The advantage of capitalism can be seen in the way it doesn’t really require any assumptions, however heroic, about either people’s motives or their knowledge. We don’t need to speculate about whether there is a universal tendency to egoism or to altruism, or about which motive is likely to prevail in all economies or in all polities. All that capitalism actually requires is institutions grounded on people’s ability to exit.

Here I am taking my cue from the argument against socialism that Nozick read about while in grad school. This argument had nothing to do with socialists’ own heroic assumptions about people’s selflessness; nor did it require analogous assumptions about the perfect knowledge possessed by capitalist producers or consumers. Ludwig von Mises, the Austrian economist, had contended that—even with the best intentions in the world—socialist central planners couldn’t know what they would need to know about supply and demand in order to keep an advanced industrial economy functioning. The reasons are precisely the same reasons that a capitalist producer can only guess what consumers want, or how much they will pay for it, before he actually puts a product on the market. Supply and demand curves do not exist apart from actual, historical purchases. And even if they did, people’s many and opaque desires, and the innumerable influences upon them, are data beyond the ability of even the most massive bureaucracy to deduce, collect, quantify, and understand.

For the purpose of (relative) brevity, I’ll take it for granted that Mises was right about full-blown socialism; the actual documents are reprinted in F. A. Hayek’s Collectivist Economic Planning (1935) and analyzed in Don Lavoie’s Rivalry and Central Planning (1985). But having, in the interests of space, bracketed the validity of Mises’s cognitive case against socialism, I want to take a similarly cognitive approach to the two main remaining alternatives to laissez faire: attempts to regulate capitalism, and attempts to redistribute the wealth that it generates. Can either regulatory or redistributive capitalism (or both) do what Rawls’s theory of justice requires—move people out of poverty—better than can laissez faire?


What economic regulation does is block exit. It forbids producers from offering certain products to consumers, or stops employers from offering certain jobs to employees. The converse is that economic regulations keep consumers from buying certain things, and prevent workers from taking certain kinds of jobs. On that basis, regulation should be rejected.

Consider the effect of economic regulations on people unlucky enough to be born into severe poverty—exactly the people Rawls would have us consider when deciding which institutions are just.

A not-untypical case, reported on the front page of the New York Times, is that of a woman named Mrs. Tratiwoon, who was born in the slums of Jakarta and who barely supports her son by rummaging through garbage dumps. Several years ago, Nike opened a sweatshop nearby.

If (contrary to the economist’s hypothesis of universal self-interest) our hearts go out to Mrs. Tratiwoon and her son, we might be inclined to impose a regulation that would ban sweatshops, or at least would upgrade their working conditions, in order to prevent people like Mrs. Tratiwoon and her son from being exploited. Or we might boycott Nike until it eliminated or regulated its own sweatshops.

But the Times reporters discovered (to their surprise) that Mrs. Tratiwoon’s highest aspiration was to get her son a job in the unregulated Nike sweatshop. That job would, she thinks, be a dramatic improvement over the other alternative open to him: following his mother into the garbage dumps of Jakarta. If Mrs. Tratiwoon is wrong, and working in a sweatshop turns out to be a bad alternative, capitalism allows her son to exit the factory and return to the garbage dump. But if she is right and he, like she, comes to see the unregulated sweatshop as an alternative worth taking, then its presence in Jakarta would allow him to exit from a terrible circumstance, exchanging it for one that is better (even if it’s still bad).

That exit option would be ruled out if sweatshops were banned. Even a regulation or boycott designed to leave the Nike sweatshop in place but improve its conditions might, by virtue of the cost of the mandated improvements, price the shoes the sweatshop manufactures out of the market—closing it down and sending Mrs. Tratiwoon’s son back to the garbage dump. Obviously that wouldn’t help him—unless he, like his mother, is mistaken about the relative desirability of the sweatshop.

By Rawlsian criteria, we should intervene in capitalism only if by doing so, we can enhance rather than block the ability of the poor to exit from deals they don’t like. That is a relatively sure way to guarantee that the position of the worst off improves. On these grounds, we have a presumptive case against the regulation of capitalism.


But what if, instead of blocking exit opportunities, we could offer Mrs. Tratiwoon’s son a better alternative than either the dump or the sweatshop?

That, in theory, is the difference between regulating capitalism and redistributing the wealth it creates. Rather than banning sweatshops or imposing regulations (or boycotts) that make them less economically viable, why not expand the opportunities open to the poor by simply taxing the rich and sending the proceeds to people like Mrs. Tratiwoon’s son? That way, he might exit from both the garbage dump and the sweatshop.

Such remedies for poverty have long been proposed by Milton Friedman (no relation). Instead of regulating capitalism through sweatshop bans, worker-safety regulations, minimum-wage laws, or compulsory unionization—all of which restrict people’s exit options—Friedman proposes a negative income tax. By this he means that people in poverty would receive income from the government rather than paying taxes to it. This redistribution would expand the menu of choices available to the worst off by giving them greater means to achieve their ends. The same exit-enhancing logic drives Friedman’s proposal for school vouchers, through which wealthy taxpayers would subsidize the education of the poor—who could spend their vouchers on any school, public or private, that they chose, and exit from any school with which they were dissatisfied.

Fully examining the desirability of redistributive measures like these requires looking at the real world of politics, for it is only through politics that such measures would (or wouldn’t) be designed and enacted. Taking this realistic look at politics is the task of Part 3 below.

For now, simply notice that the appearance in Jakarta of a sweatshop that might help Mrs. Tratiwoon’s son isn’t a conjuring trick made possible by some “magic of the market.” There is no legerdemain involved; there aren’t even “economic laws.” But there are tendencies, produced by the exit option. Nike appears to have guessed that a certain type of factory would, however deplorable to us, bring enough of an improvement in the lives of Jakartans that they would want to work there producing shoes at prices that (Nike guesses) people would want to buy. If this guess is incorrect, then the consumers can exit from buying the shoes, and N ike can exit by closing down the sweatshop. The employees guess that their lives would be improved by working in the sweatshop; if this guess is incorrect, they can exit by returning to their previous situation. The sweatshop cannot exist for long unless it tends to provide both its customers and its employees an improvement in their condition.

Although one could easily go through college (unless one is an econ major) without ever thinking about these tendencies—being taught, instead, that the immense material progress of the West is due to the efforts of unions and anti-business political movements—these tendencies are what made the West rich; and they are, under the heading of “globalization,” what is now making the Third World rich, too. The prosperity of the First World originated in an Industrial Revolution that generated the original sweatshops, in the nineteenth century. The opportunity to work in them prompted millions of people to uproot themselves from the countryside and move to the industrial cities of England and, soon afterwards, America. For that reason, it was to our greatgrandparents’ great advantage, and our own, that in the West (unlike in the Soviet Union and the rest of the Second World), political indignation against sweatshops never fully caught up with the spiraling process of capitalist growth.

Relatively unregulated capitalism—in the relevant sense, laissez-faire capitalism (wages and working conditions unregulated)—produced the real-world sweatshops in which my own great-grandparents worked when they came from Rumania and Russia to New York 100 years ago. Prosperity and religious toleration drew them here; both prosperity and toleration amounted to the ability to choose how to live, and thereby to improve how one lived. What distinguished early twentieth-century America economically was its abundant opportunity to do work that would be, for me, unimaginably tedious, but that was, for my great-grandparents—in comparison with the conditions of their birthplaces—a huge improvement. That’s why they exited from Europe.

But even if you grant that blocking exit opportunities through economic regulation is a bad idea, why shouldn’t governments follow Milton Friedman’s advice and redistribute wealth within a capitalist framework, thereby multiplying rather than restricting the exit opportunities of the poor? An answer to that question requires a “model” of politics that is as free of oversimplifications as is the exit model of economics.

3. Politics: Blinder than Love

The most consistent finding in political science, and perhaps in all of social science, is this: modern electorates, in every country studied, are almost completely uninformed about politics.

In 1964, for instance, at the height of the Cold War—after the world was nearly incinerated during the Cuban Missile Crisis, when Russia tried to place nuclear missiles off the coast of Florida—62 percent of the U.S. public failed to realize that Russia was not a member of the anti-Soviet military alliance, NATO.

In 1979, 76 percent of the public could not explain the First Amendment (in even the vaguest way).

In 1989, 43 percent of the public did not know what a recession is.

Also in 1989, 71 percent could not identify their U.S. Representative.

In 1994, after the Republicans took control of Congress for the first time in decades and elected Newt Gingrich Speaker of the House, 57 percent of the public had never heard of him.

Last year, 58 percent admitted that they knew “very little” about the USA Patriot Act.

Those are just a few of my favorite findings from the vast and always-growing public-ignorance literature. An appreciation of such findings is crucial in attaining a balanced answer to Rawls’s question: Which institutions best serve the worst-off? Thus far, however, outside the small circle of public-opinion specialists, such appreciation has been a one-sided, partisan affair.

With the re-election of President Bush, it has suddenly become fashionable in Democratic circles to berate the public for being uninformed—or stupid (they are not the same thing, but are often treated as such). Op-eds with such titles as “The Unteachable Ignorance of the Red States” abound, as do declarations like this one, from New York Times columnist Bob Herbert: “Ignorance played at least as big a role in the election’s outcome as values.”

The thinking behind such statements will itself provide a good case study in the subtleties and sources of political ignorance.


What is it, exactly, that Bush supporters are supposed to be so ignorant of?

The answer, of course, is the war—primarily the absence of weapons of mass destruction in Iraq, and therefore, it is thought, the Bush administration’s manipulation and dissimulation about WMD. According to this line of analysis, if the public re-elected the president, it must have been ignorant of the fact that Bush’s lies and distortions about WMD had been exposed during the year before the election by a series of books, commissions, and news reports.

But in truth, none of the commissions, reports, or books gave any reason to think the Bush administration lied about WMD. The “Bush lied” assertion is an inference, drawn from the fact that Bush turned out to be wrong in asserting that Iraq had WMD-development programs and stockpiles. Bush’s assertion, however, was itself an inference drawn by almost every well-informed observer, prior to the war.

Neither France nor Germany nor Russia argued that Saddam Hussein had no WMD. Like virtually everyone else who had studied the matter, their intelligence agencies thought he had them. So did most American critics of the war. One of the main antiwar arguments the critics made, in fact, wa s that the ca sualties were likely to be horrendous on the American side, because an attempt to unseat Saddam would prompt him to use his WMD against the invaders. A secondary antiwar argument was that Iraq was unlikely to give its WMD to Osama bin Laden, because Saddam was a secularist, not an Islamist. Both of these arguments presupposed—just as President Bush did—that Saddam had WMD.

Because they didn’t contest that there were Iraqi WMD, most opponents of the war favored continuing or strengthening UN sanctions against Iraq. They argued that since the threat posed by Iraqi WMD against invading U.S. troops would be grave, and since there was no“imminent” threat of their use against the United States, it was better to let UN sanctions restrain Iraqi WMD than to let U.S. soldiers be killed by them.

The reasons for this consensus about Iraqi WMD were excellent—even though they turned out to be incorrect. In trying to demonstrate this, I’m deliberately picking a topic that is not only paradoxical at many levels, but one about which you’re likely to disagree with me (at least initially). I’m doing this for a reason, and the reason isn’t to try to make you into an ex-post-facto supporter of the war.

There are many possible grounds for opposing or supporting the war. My aim is to address just one of them—the “Bush lied” hypothesis—because it’s one about which you’ve likely thought a great deal. If I can show you that the thinking about this subject, perhaps even your thinking about it, was faulty, it will say something, I think, about the difficulty of being well informed and logical in political deliberation. And that will have serious implications for the desirability of using politics to redistribute wealth or, in general, to solve social problems.


Why was there such a wide consensus before the war that Iraq probably had WMD, or at least WMD-development programs?

At the end of the first Gulf War in 1991, the retreating Iraqi army left behind so much evidence of WMD stockpiles and programs that in response, the United Nations—not the United States—imposed drastic economic sanctions and an intrusive inspections regime on Iraq. (It was to allow the Iraqi population to survive the sanctions that the UN’s notorious Oil for Food program was established.) In the years that followed, UN inspectors uncovered Iraqi laboratories devoted to refining uranium into weapons-grade plutonium; shells and missiles modified to carry weapons of mass destruction; and vast quantities of weaponized anthrax and mustard gas, as well as the components of VX gas and botulinum toxin.

In 1998, following a long string of such discoveries, Iraq effectively expelled the UN inspectors, who were readmitted in 2002 only after the United States and Britain had begun assembling a huge army on Iraq’s borders. In the interim, high-ranking Iraqi scientists and officials—including Saddam Hussein’s son-in-law, Lt. Gen. Hussein Kamel (who later returned to Iraq and was executed, along with his family)—escaped from Iraq and claimed, in the pages of such publications as the New York Times, that even while the inspectors had been in Iraq, the Iraqi government had become adept at concealing its WMD research by using both hidden and mobile facilities.

The direct evidence for the continued use of such facilities as of 2002 was weak, but the inferential evidence was overwhelming. If nothing else (and there was plenty), there was this: in August 2002, the UN Security Council unanimously passed Resolution 1441, threatening“serious consequences” if Iraq failed to allow readmitted UN inspectors unimpeded access to any place and anyone in their attempt to discover what had happened to the WMD previously discovered in Iraq. Iraq failed to comply. For reasons that still aren’t clear, Saddam Hussein thereby tempted invasion by misleading everyone (even Iraqi Army commanders apparently thought that other units of the army had WMD) into thinking Iraq’s WMD programs and stockpiles were alive and well.

As suggested by the iconography of his regime, and by his recent professed turn toward Islam (overlooked by critics of possible cooperation between Iraq and Al-Qaeda), Hussein may have been angling for the unofficial position of leader of the Arab world by standing up to the United States. He may also not have believed that, in the face of protest demonstrations, the United States would really invade. Whatever the reason, Hussein failed to do what he could have done to stop the invasion in its tracks: comply with Resolution 1441.

The inference from this failure, and from Saddam’s past behavior in creating (and using) vast WMD programs and delivery systems, seemed clear at the time. Why would he have expelled the inspectors in 1998, and why would he fail to cooperate once they were readmitted, if he had discontinued Iraq’s WMD programs? This is the logic that drove nearly everyone who knew about his past behavior, and who was well informed about the UN’s current demands, to conclude that he had something to hide.

After the war, however, when no WMD were found, opponents of the war forgot all of this. Ironically, in arguing that “Bush lied” about WMD, and that Bush voters were ignorant of these lies, Bush’s opponents themselves arguably displayed ignorance of the facts, or at least forgetfulness of them.

I say “arguably” because “the facts” don’t interpret themselves. The facts I have detailed culminate only in an inference about Iraqi WMD that turned out to be wrong! I don’t claim that those who drew the inference were right; I claim only that they were relatively well informed and logical. A second irony of the “Bush lied” thesis is that its proponents implicitly claim not only that their enemy, Bush, was well informed and logical, but that he was (on this matter) literally omniscient.

Nobody has ever found evidence that Bush “lied.” For him to have done so, he would have had to know in advance that there were no Iraqi WMD. What is supposed to justify the “Bush lied” inference is that, as it turns out, Hussein didn’t have WMD. From that premise, however, the conclusion that Bush lied follows only if one first assumes that the world is so transparent that whatever we know now has always been known; or rather, in this case, that whatever we know now, the president of the United States has always known.

In short, the proponents of the “Bush lied” thesis are unwittingly crediting Bush with the ability to have predicted the future with certitude. Thus, if his publicly stated prediction turns out to have been incorrect, it must have been a deliberate falsehood—not just a well-intentioned error. Bush couldn’t merely have made an incorrect inference about the future from the past; he must have known that the inference was incorrect.

That is hardly a logical conclusion, or a charitable one (despite its attribution to Bush of godlike prescience). But it seems to make sense—if one does not take human ignorance seriously.

One would think that everyone takes ignorance seriously—especially those who now accuse Bush voters of being ignoramuses (and who formerly accused Bush himself of being one). But using “ignorance” as a cudgel against one’s political opponents is not the same thing as taking ignorance seriously. Taking ignorance seriously requires being charitable enough to allow that one’s proponents may simply be mistaken. We find in the charge that Bush voters are ignorant of “Bush’s lies,” then, two kinds of ignorance: ignorance of facts, and ignorance of human fallibility—ignorance, that is to say, of the possibility of ignorance.

To complete our taxonomy of ignorance, let us assume for the sake of argument that Bush was indeed lying about Iraqi WMD. So what?

Suppose Bush wasn’t motivated in the least by the possibility that Iraq would acquire, or give to terrorists, WMD. Suppose Bush was merely greedy for oil (although, in the event, nobody seems to have gotten any). Does the legitimacy of an argument depend on the sincerity of the person voicing it? If the war had foreclosed the possibility that WMD might kill millions of people, then of what relevance would be the motives of those who led the war? Can good consequences never follow from bad intentions?

By fixating on the (allegedly malevolent) intentions of the warmakers instead of the results the war might have achieved, proponents of the “Bush lied” hypothesis are arguing ad hominem. That is, they are committing a plain logical fallacy. The motives of someone (in this case, Bush) who urges a certain conclusion—whether a conclusion about a fact (such as an inferred presence of WMD) or about an action (such as the need to strike preemptively against a potential threat)—are irrelevant in evaluating whether that conclusion is sound. To care about whether Bush lied, then, suggests a third type of ignorance: ignorance of logic.


The credo of The Dissident, drawn from New Republic founder Walter Lippmann, is applicable to the “Bush lied” theory. As Lippmann put it, “the opponent”—in this case, Bush—“has always to be explained, and the last explanation that we ever look for is that he sees a different set of facts.”

Lippmann’s aphorism explains the intolerance manifest in contemporary politics—and exemplified in the view that only fools could vote for Bush. If one is ignorant of the possibility of human ignorance (not just the possibility of one’s opponents’ ignorance, but of one’s own), one won’t take seriously the possibilities (a) that there are facts of which one is, oneself, ignorant; and (b) that one’s reasoning about the facts may be deficient. Without that awareness, your own view of “the facts” will seem so self-evident that anyone who disagrees with you must be lying, or worse. In Lippmann’s words, “out of the opposition we make villains and conspiracies.”

Conversely, to the extent that you see the facts of the matter in dispute as being less than “obvious,” you will be tolerant of your opponent’s disagreement with you, because you will take seriously the possibility that your opponent’s interpretation of the facts is honestly mistaken. But as Lippmann points out, the possibility that your opponent is mistaken—i.e., fallible—entails that you might be mistaken, too. And that’s not something most of us are prepared to accept, except as an abstract possibility.

For this reason, Lippmann writes, “he who denies either my moral judgments or my version of the facts, is to me perverse, alien, dangerous.” Simple disagreement over the facts of the world “saps the very foundation of our own assurance that we have seen life steadily and seen it whole.” Thus, while people “are willing to admit that there are two sides to a ‘question,’ they do not believe that there are two sides to what they regard as a ‘fact.’” One’s political opponents, then, cannot just be wrong in their assessment of the facts about how to achieve a good end. They must be deliberately pursuing
an evil end.

It gets worse. In determining how best to achieve a good end, we need to be able to predict the consequences that various means to that end would produce. Such predictions require that we know not only which“set of facts” is true, but also how facts interact with each other, so we can determine which actions will produce beneficial effects. In short, we need both knowledge of the facts and rigorously reasoned theories about the facts—no illogical, ad hominem arguments allowed. Compounding this problem, from different theories about the interaction of facts will stem different understandings of which facts are relevant; and, often, different inferences about which “set of facts” is true. Our perceptions of the facts themselves will be colored by theories that may be uninformed or illogical.


In arriving at knowledge of which facts are true, and knowledge of which theories about the interaction of facts are sound, we face enormous barriers: limited experience, limited time, and limited powers of concentration. These are the cognitive (as opposed to emotional) sources of human fallibility. It is in the face of these barriers that the political world begins to look more complex than if decisions with which we disagree were made only by liars, villains, or conspirators.

Lippmann gives this example of how difficult it is for us to grapple with political complexity: “There are few big issues in public life where cause and effect are obvious at once.… It is not surprising,” therefore, that“the commonest form of reasoning is the intuitive post hoc ergo propter hoc. The more untrained a mind, the more readily it works out a theory that two things that catch its attention at the same time are causally related.” Like ad hominem arguments, post hoc ergo propter hoc reasoning is fallacious. And yet, Lippmann claims, it is intuitive.

In confirmation of Lippmann’s conjecture, public opinion research has found that most people vote in accordance with their judgment of “the nature of the times.” For instance, if the voter’s country is prosperous (a disputable fact), then the incumbent party must be responsible for the good times (an illogical theory about that fact). It is plain, on reflection, that the incumbent party’s policies may have nothing to do with the prosperity. But given the complexity of a modern economy, how can anyone—at least anyone without an Economics Ph.D. (or, given the argument of Part 2, maybe even with an Economics Ph.D.)—be anything but ignorant about what is really responsible for prosperity?

Post hoc ergo propter hoc reasoning is a heuristic—a substitute for, in this case, the type of elaborate economic theory one would need in order to justify voting for or against the incumbent party. The heuristic is as illogical as the argumentum ad hominem. More specifically, attributing good times to the incumbent party, post hoc ergo propter hoc, is magical thinking: for it identifies no logical connection between cause and effect.

Now, think back to the initial plausibility that antisweatshop regulations may have had when you began considering them in Part 2. Against the logic of exit, which purports to show that such regulations will tend to harm the very people they are designed to help, what is the basis of opposition to sweatshops, and to globalization more generally—if not magical thinking? The thinking goes something like this: “It is bad that people must work in sweatshops [a compassionate and valid sentiment]; therefore, let us ban the sweatshops [an unsound conclusion].” Exactly how does banning the visible manifestation of the problem (the sweatshop) solve the problem itself (poverty)? That the question is so rarely asked impels the conclusion that the thinking involved is tacitly magical. (In the modern world, few resort explicitly to magical thinking; the magic is buried in theories that tacitly justify various political prescriptions while keeping questions of their causal efficacy from even being asked.)

The implications go far beyond sweatshop regulation. How much of politics—how much not only of your opponents’ politics, but of yours—consists in favoring policies based on what are taken to be their goals (or, in the case of the war, opposing a policy based on what are taken to be its architects’ nefarious “real” goals), without expending any serious effort on inferring precisely how the means to be deployed are supposed to achieve those goals? Such thinking—ad hominem thinking writ large— is as magical as the post hoc ergo propter hoc.

Both post-hoc-ergo-propter-hoc and ad-hominem arguments are non sequiturs. It doesn’t follow from the temporal sequence of two factors that one caused the other; correlation is not causation. And it doesn’t follow from somebody’s (putatively) malevolent motives for doing something that that thing should not be done; or from their good motives, that it should be done. Gaps between cause and effect are the common denominator in these two logical errors, and magic is what closes such gaps.

More worrisome even than the ubiquity of such magical closures in political reasoning, however, is that they seem to be our default option. No matter how logically fallacious they really are, they seem, as Lippmann put it, “intuitive.”

A mismatch between our intuitions and the modern world is not surprising. Our ancestors evolved to make decisions that would contribute to their reproductive fitness in the relatively simple and personal situations hunters and gatherers faced for hundreds of thousands of years. But once the spread of agriculture created a surplus of food that made vast, impersonal civilizations possible, our genetic evolution slowed and possibly stopped: the agricultural surplus made possible not only civilization, but the sustenance of those who otherwise would have died off before reproducing. In the mere 6000 years since the emergence of “civilization,” would it not be remarkable if our minds had somehow adapted to reason clearly about the facts of far-more-complicated forms of society than those in which human beings evolved?

No hunter or gatherer would obtain a reproductive advantage in being able to theorize rigorously about the workings of impersonal economies that did not yet exist. So we are unlikely to find that such theorizing comes naturally to descendants of hunters and gatherers—such as us. By contrast, in hunter-gatherer societies, each individual’s power over the others was visible as the immediate cause of such effects as life or death, misery or ecstasy. When cause and effect are this directly related, post hoc may imply propter hoc, as a rule of thumb. Moreover, in such intimate circumstances, the discernment of motives assumes paramount importance—ad hominem—as a reliable indicator of the type of post-hoc treatment one will likely receive.

Mental habits that made good sense for our ancestors, though, are merely superstitious when applied to vastly different societies like ours. In impersonal civilizations, millions or even billions of people affect each other (e.g., through the economy) in ways that are largely invisible. Cause and effect are too indirect to untangle with ease; causes may have nothing at all to do with motives; and motives are, in any event, well-nigh impossible to discern, given the anonymous nature of our connections with the overwhelming majority of the people whose actions affect us.


Lippmann’s reference to “the untrained mind” suggests that intensive logical tutelage, and perhaps tutelage in the complexities of modern civilization—for instance, in economics—can combat the tendency to engage in spurious political reasoning. And to some extent, they can.

But our propensity to reason poorly and without adequate information is not a prescription for replacing democracy with rule by well-trained social-science experts (although that was the prescription toward which Lippmann himself was inclined).

A given expert’s theoretical paradigm can typically be accounted for not by his positivistic study of “the facts,” but by the ideas about which facts are relevant, and how to interpret them, that he has absorbed from his culture, including his undergraduate and graduate education. His research will tend to be directed toward topics that spin out the logic—however illogical it may be—of that culture. So far, the situation is the same in social science as in natural science. The difference between the two is that, without the ability to conduct controlled experiments, social scientists are much less capable than natural scientists of transcending their cultural predispositions. As a result, social-scientific paradigm shifts, which seem so momentous to those undertaking them and to those who subsequently learn about them, are overwhelmingly instances of what Freud dubbed “the narcissism of small differences.”

Marx, for example, took from classical economics the assumption that individuals pursue self-interest (narrowly defined), and married it to a simplistic view of how easy it is to tell which social institutions do, and which don’t, serve one’s interests. This “revolutionary advance” accounts for Marx’s claim that members of the proletariat would intuit from the exploitative conditions of their labor that communism would be the means to the end of their self-interest; and for his equally naïve assumption that, after having overthrown capitalism, the proletariat would be able somehow to infer the allocation of labor and natural resources that would meet everyone’s needs. Both class consciousness and consciousness of how to administer a communist society, being functional to the interests of the proletariat, would essentially be self-evident to proletarians—not by virtue of reading books by Marx, but simply by the direct perception of their economic position.

Durkheim went a step further, positing that social needs and the institutions that meet them are self-evident to “society” as a whole. In this way he accounted for a noteworthy fact: the growth of regulatory states in nineteenth-century capitalist societies. Since (he assumed, without argument) “any fact of a vital nature… cannot survive if it does not serve a purpose or correspond to some need,” the regulatory state must be in some way functional. Durkheim therefore explained the regulatory state by means of an assumption with which we are, 110 years later, still thoroughly imbued: the assumption that modern capitalism must be in need of correction by the modern state.

Foucault, who is widely thought to be another great iconoclast, simply provides the reductio ad absurdum of functionalism. Like Marx, Foucault assumes that whatever institutions exist—“carceral” institutions like prisons, for example—must be functional to the oppressive interests of the ruling class. But Foucault extends Durkheim’s tacit admission that functionalism means history as the product of vague “social forces,” not of the interaction of specific people with first and last names who may have dysfunctional (mistaken) ideas about what institutions are needed, and may also have different criteria of what is needed than their selfinterest. Thus, Foucault narrates the transition from one status-quo-justifying “discourse” to another without providing evidence of conscious human intervention, each successive discourse somehow managing to prop up a new politico-economic order without any actual human beings conceptualizing the self-interested needs of the successive ruling classes. Since he provides no evidence that institutions and their attendant discourses serve class interests, Foucault’s casual references to the functionality of carceral society for “the bourgeoisie” must be credited solely to his unexamined Marxist assumptions. Although he was a profound observer of the subtleties of interpersonal coercion, Foucault was not at his best when it came to explaining human action in particular times and places—which is to say, all human action— even though, as an historian, human action in particular times and places was Foucault’s field of “expertise.”

Like his predecessors, Foucault failed to take seriously the ideas—the theories about reality—that may move human beings to act. When Foucault does occasionally eschew the passive voice and discuss the reasoning of real people who act in the political realm, such as Jeremy Bentham and other prison reformers, he quotes them selectively enough that he can discount their humanitarian and indeed radical aims (but read Foucault’s footnotes, where the occasional quotations from actual human beings give the game away!), since those aims contradict his premise that self-interest-serving discourses, not theorizing people, move the world. Durkheim, similarly, did not prove that the regulatory state was necessary; he simply assumed that it was the result of automatic “societal” adjustment to self-evident social needs, rather than of actions taken by people who believed, even if only at the level of assumption, the possibly mistaken theory that such a state was necessary—people like Durkheim. Likewise, in Marx’s case it turned out that the proletariat never did recognize its supposedly self-evident class interest in overthrowing capitalism. The only so-called communist revolutions to occur were coups d’etat carried out in agricultural rather than industrial societies by revolutionary vanguards moved to action by theories derived from reading Marx, not by some direct apprehension of proletarian interests. And having attained power, it turned out that it was no more self-evident to these vanguards how to plan the allocation of resources so as to meet everyone’s needs than Ludwig von Mises—the economist whose work prompted Robert Nozick’s switch from Left to Right—had predicted it would be.

The logical error of functionalism is precisely to reason post hoc, ergo propter hoc. If, á la Foucault, in modern societies workers are well disciplined, then (Foucault assumes) it is not due to a contingent chain of cultural causation such as the one described in Max Weber’s Protestant Ethic and the Spirit of Capitalism. Rather, it must be due to the emergence of a disciplinary discourse that necessarily coincides with the economic interests of the bourgeoisie. The correlation of two “facts” is mistaken for the causation of one by the other—as if by magic.

While some social scientists have noted the logical defects of functionalism, and others have pointed to the resulting errors in the interpretation of facts that have been produced by the likes of Marx, Durkheim, and Foucault, most have not. Functionalism of this sort is, as any undergraduate knows, alive and well (which is not to suggest that there are not other, equally erroneous social-scientific doctrines that are alive and well). Little social science that is both nontrivial and rigorously demonstrated has, in fact, emerged in the two centuries in which modern societies have been studied; and what profound discoveries have been made—for example, the discovery of public ignorance about most things political—have lain dormant. Considering that the work of such giants as Weber, Pareto, and Mises occurred early in the twentieth century and has been largely forgotten, even as half a dozen fads have poured sour old wine like functionalism into fashionable new bottles, “regress” might be better than “progress” to describe the direction of social science.

Given the dismal track record of the social sciences, how can the real social-scientific experts—the ones who are right—be distinguished from the frauds?

It is in the nature of our ignorance that even when we are aware of it in the abstract, we cannot pinpoint exactly what it is (not just the field, but the information from that field) that we should know but don’t. If we knew what we didn’t know, we would already know it. By the same token, we cannot tell which experts are right, or we wouldn’t need their help. We can only infer their rightness by means of such heuristics as doctoral degrees, universities attended, pedantry, eloquence, and wit. The inadequacy of these proxies for true expertise should be evident from considering the great intelligence and flair, the extremely pedantic absorption in the doctrines of their mentors, and the unsurpassed educational pedigrees of the “experts” who have developed and embraced such fallacious doctrines as functionalism.

It may be objected that evolutionary psychology itself gives a functionalist account of people’s cognitive capacities. But all functionalisms are not created equal. Unlike the types of functionalism I have criticized, evolutionary psychology posits a mechanism—natural selection through inadvertent “competition” between individuals with slightly varying genes—that explains the survival of traits that would have been functional in a hunter-gatherer context (but not so functional in a civilizational context). While Marx’s and Durkheim’s views were influenced by the Hegelian and Darwinian versions of historical evolution, Marx and Durkheim provided no more of a substitute than Foucault did for the Hegelian Spirit or for Darwinian natural selection as a mechanism that would screen out dysfunctional perceptions of where the interests of a class or of a society lay. Behind the immense sophistication of social-scientific experts, then, one finds with disheartening frequency the same type of magical thinking employed by members of the general public.

The fact that experts disagree with each other is enough to remind us that they, too, are human; at least some of them must be wrong. Just by virtue of the odds, then, we cannot rely on experts. And even when there is an expert consensus in a given time and place, consideration of the expert consensuses that were taken for granted in the past (for example, that homosexuality is a disease) should disillusion us of the notion that intensive study of the world produces reliable knowledge of it. Indeed, one of the least appreciated findings of the public-opinion literature is that through the selective perception and retention of information, people who are better informed also become proportionately more rigid in their beliefs. They know more facts, but they primarily know facts that tend to justify their dogmatically held theories.

The Hobson’s choice we face in politics, then, is governance by a relatively ignorant but open-minded general public or by a relatively well-informed but doctrinaire elite.


In Part 2, I argued that the complexities of the real world render unrealistic the economists’ assumptions of predictable supply and demand, predictably motivated suppliers and demanders, and perfect competition. The problem to which I am attributing political ignorance—and political dogmatism—is very similar to the problem facing real-world economic decision makers. Ignorance and dogmatism are human responses to the complexity of the modern world—at least when viewed through hunter-gatherer eyes such as ours.

In making personal economic decisions, however, human beings who are insufficiently informed about the complexities they face, and who are prone to errors in reasoning about them, have one advantage that they usually lack in politics, where they have to evaluate theories about the economy (or the civilization) as a whole. In private decision making, people don’t need to be well informed about any facts except their own direct experience of the product or service or job from which they have the power to exit. They don’t even need to reason about these facts rigorously. All they need do is react to negative stimuli by exiting—or to positive stimuli by staying—with no more of an accurate grasp of what they are doing, or of why, than is possessed by rats in a maze.

By contrast, in making political decisions, the only way to address a negative situation is to exercise the feeble power that Albert Hirschman contrasted against exit: “voice.” Voice requires us to react to negative stimuli not by leaving them behind, but by thinking about their causes, and inferring which policies would cure them, so we can “advocate” our diagnoses and prescribe remedies to our fellow citizens. Only if we have an accurate account of the cause of the problem and make sound inferences about the solution is politics likely to be productive. If, however, we are wrong about the source of the negative stimulus or its cure, we are stuck with it—there is “no exit” from politics, since it sets the boundaries of the private sphere—and the problem may even worsen as a result of our efforts. Think, for example, of the War on Drugs. Politics requires us all, in effect, to act like (omniscient) social scientists, with the (infinite) time necessary to investigate the world in all its fullness, and the (infallible) minds necessary to draw proper inferences from that investigation.

In Part 2, I left hanging the comparison between the exit mechanism in the economic and the romantic spheres, because only in the context of comparing economics to politics does the analogy come into its own. Consider that—tempting as it is to stay in bad relationships, trying to infer the reason for the problem—our ultimate recourse is to exit from them (at least when there are no third parties, such as children, who would thereby be affected). After we exit, we may spend the rest of our lives theorizing about what went wrong. The beauty of the exit mechanism, however, is that it allows our lives to continue, and for the lucky among us to do better next time.

Exit lets us leave failure behind, without having to figure out what failed. This may sound a bit less callous if you consider that it happens at an unconscious or semiconscious level hundreds of times in the course of a week, as you pass up the chance to flirt with every single person you encounter. Without thinking about it, you bypass most people because something about them just doesn’t interest you. How impossible your task would be, by contrast, if you had to justify, by articulating a theory, the rationale for choosing to bypass this stranger, then that one. That is just what political “voice” demands, and what economic “exit” does not—as suggested by the need to voice an intricate theoretical defense, against would-be regulators, of Mrs. Tratiwoon’s son’s potential choice of exiting the garbage dump in favor of a sweatshop.

Just as the flip side of economic exit is the ability of an employee or a consumer to stay with a job or a product that improves his life, the romantic flip side of exit is the ability to stay in a good relationship. And just as the restaurateur or his customers may not know why they don’t head for the exits, a lover need not know why he is in love. As Weber wrote, love is “subjective in the highest imaginable sense; and it must be absolutely incommunicable.” In love, as in economics, positive stimuli can lead to good outcomes without being understood; negative stimuli can be avoided through exit, also without need of understanding. But short of physically emigrating from the political jurisdiction in which one lives—as my Rumanian and Russian great-grandparents did, a century ago—political problems must be solved by understanding a world that human beings are ill equipped to grasp.

In a passage of his Capitalism, Socialism, and Democracy that is politically incorrect at too many levels to analyze, Joseph Schumpeter put the contrast between exit and voice succinctly. Comparing the blandishments of advertisers to those of politicians, he wrote that “the picture of the prettiest girl that ever lived will prove powerless in the long run to maintain the sales of a bad cigarette.” What Schumpeter meant is that no advertisement, however alluring, can overcome the negative stimulus of a cigarette that tastes bad. The bad taste will tend to lead disillusioned smokers to exit, by switching brands—as long as there are other brands to which they can switch, and that they can afford to buy.

But in contrast, Schumpeter continued,

there is no equally effective safeguard in the case of political decisions. Many decisions of fateful importance are of a nature that makes it impossible for the public to experiment with them at its leisure and at moderate cost. Even if that is possible, however, judgment is as a rule not so easy to arrive at as it is in the case of the cigarette, because effects are less easy to interpret.

Schumpeter is making two points. First, in politics, information about the negative (or positive) consequences of our actions usually requires second-hand reporting—mediation, usually by the mass media—before decision makers, such as voters, can perceive it. This allows the theoretical proclivities and leaps of logic of media personnel (whether the employees of advertising agencies or, I would add, newspapers, magazines, web sites, and TV networks) to help select the information they filter to us about the world. In making nonpolitical decisions, however, we often have unmediated access to the information we need: how the situation from which we might exit “tastes” to us. Second, in making political decisions, not only the information we perceive, but our view of what information is needed and of how best to interpret it, must be re-processed in our own minds by theories that disentangle the various possible chains of causation, but without the benefit of controlled experimentation.

If a Marlboro cigarette tastes bad, the smoker doesn’t need what NBC News ads call “the complete picture” to know it, and he doesn’t need NBC News to give him the partial picture he requires. Relatively speaking, the bad sensation is immediate: it is direct feedback from the decision to smoke a Marlboro—something that is rarely available from political decisions. What’s more, our inability to “experiment at leisure” with alternative public policies means that neither the journalist nor the social scientist nor the member of the voting public can perform tests that would establish which policies would have good effects, and which bad.

Even in the relatively simple world of foreign policy, where at least the motives of one’s adversaries matter (not in themselves, but as inferential causes of military actions), and where feedback is relatively clear (war or peace? victory or defeat?), it is impossible to determine the cause of this feedback with any precision. One can be misled by hard evidence and valid inference—misled by the likes of Saddam Hussein—because one can’t run the same historic scenario twice in order to isolate one variable (say, by invading Iraq just to see if it is really hiding WMD or not) before deciding what to do. Likewise, one can’t perform an experiment to determine whether victory in a particular war was due to lavish military spending, high morale, good strategy, superior technology, a combination of these factors, or something else. Our fallible, often fallacious, but intuitively plausible inferences—constructed in the face of our ignorance of a complex world—are all we have.

The same is true, only more so, of economic policy. Is a recession the result of a president’s tax cuts, or of Federal Reserve policy? Maybe more tax cuts would have stimulated enough prosperity to overcome the deficits they may have created. No counterfactual experiments can test which inference is right.

By contrast, in exiting from the bad cigarette, the smoker needn’t concern himself with any counterfactual except the one that is susceptible to experiment: will a Camel taste better than a Marlboro? As a result, once the smoker senses his distaste, he need not produce any theory about why the cigarette tastes bad (to him). All he has to do is try a different brand. Indeed, even an “expert” cigarette manufacturer may not be able to produce a sound theory about what makes for a good-tasting cigarette (which is why Marlboro may, in that case, go broke).

Exit tends to select for good decisions because, unlike the standard economic model, its star ting point is fallibility, not omniscience. When either consumers or producers, as fallible human beings, act on fallacious reasoning or misinformation, or both, they can exit from their mistakes without even being aware of what they are doing. Exit is an error-correction mechanism that spares us the need to infer the cause of our error.

The ignorance in which economic and personal choices are made when exit is available is not a degraded, animalistic condition imposed on people by capitalism (or by love). It coexists with being able to improve one’s situation piecemeal, one step at a time, without even giving it much thought. Ignorance plus exit is relatively blissful not in an absolute sense—omniscience would surely be better—but only given that ignorance is the fate of everyone but gods.

Consider by way of contrast what happens when we have to be well informed and think clearly about cause and effect. If the government produced the bad-tasting cigarette, politics would have to solve the problem through the voice mechanism. Voters would have to decide which of the theories advanced by various politicians (and media-interviewed “experts”) was the right one. Different political parties might call for a change in the tobacco, or the rolling paper, or the conditions in the factory; each of these intuitively plausible solutions, some demagogue might promise, would surely put a stop to the bad taste. Lacking the ability to experiment with the various proposals and to rigorously interpret the results of the experiments, no progress would be likely. With the exit mechanism, though, cigarette consumers need not think about such things, and cigarette producers may be mistaken about them, too. The consumer’s ability to leave behind a bad taste, none the wiser as to its cause, will tend, over time, to select for better-tasting products, no matter what anyone blames for the problem or credits for the improvement.

In a democracy, the people being regulated—in their capacity as consumers, entrepreneurs, or workers—are the same ones who would, ultimately, be making the regulations in their capacity as voters. I can think of no plausible reason why the regulations they vote for would be likelier to aim at a truly Good end, whatever it is, than would the activities these same people would have undertaken if not for exit-blocking paternalism. And when it comes to regulating the actions people take as means to their putatively Good ends, we, the people, would, if we were to play paternalist, need a non-magical understanding of the causal interactions between facts about, say, the economy as a whole, which is unlikely enough. We would also need a knowledge of the direct effects that closing off exit actions would have on the lives of those we would presume to regulate that is superior to those people’s own knowledge of these effects—as expressed not in their political opinions (people benefited by the ability to work in a sweatshop, as evidenced by their decision to do so, might nonetheless seethe with resentment at the sweatshop’s greedy corporate owner), but in their decisions about whether or not to exit. That is even less likely.

You might be inspired by Schumpeter’s example to yearn for case-by-case regulatory paternalism—of the sort that would ban cigarettes, no matter how good they taste; but not, say, marijuana. Clearly, however, something about politics makes it not work out that well very often.

My candidate explanation is public ignorance. The public can easily be scared by demagogues about situations it has never experienced, and will demand legislative action against these situations if it has the power to regulate any private affairs at all. Legitimizing an exit-blocking regulation that may be thought desirable on paternalist grounds opens the door to bad regulations, on the same grounds. And if people’s knowledge and reasoning about politics tends to be deficient, it follows that there will tend to be more bad regulations than good ones.


It is just that type of consideration that allows us, at last, to return to Milton Friedman’s answer to Rawls’s question (how best to help the worst off?): not regulating capitalism by diminishing the exit options of the poor; but expanding their choices through wealth redistribution, so that nobody would face the choice between a garbage dump and a sweatshop to begin with.

Economists have come up with no reason to oppose every single penny of redistribution. (This is the task that philosophers such as Nozick set themselves—unsuccessfully, in my view.) Sometimes—given that the government imposes a high enough tax, or that people have a certain set of motives—redistribution may discourage enterprise, or encourage dependency. But if that is the only way to prevent starvation, it may be a price worth paying.

However, the workings of public ignorance do provide a reason to be at least very dubious about the prospect that redistribution would enhance the exit opportunities of the worst off. This reason has both a logical and an historical dimension.

Logically, if we open the political door to redistributive taxation, how do we close it to intuitively appealing but counterproductive regulations such as those deployed against sweatshops? In both cases, the role of the political system is to address social problems. If voters could see the perverse but indirect effects of public policies clearly enough to endorse only policies that increase exit opportunities rather than diminishing them, then they might draw the line at redistribution that gives the least advantaged more options, ruling out regulations—and forms of redistribution—that close options off. But an ignorant public—a human public—is unlikely to draw such lines. People who haven’t even heard of their senator cannot be expected to have heard of “exit” or voice,” or to see any reason why social problems might not have political solutions. People (which is to say, human beings) who assume that their version of “the facts” and their inferences about cause and effect are so obvious that those who disagree must be liars; people who reason magically; people who selectively perceive and retain only information friendly to their predispositions, are unlikely to be able to draw fine lines in the right places.

What I have been doing in this long article is considering the effect of a key element of the human condition, ignorance, in the realms of economics and politics. One small example of that condition is that people’s impulse—the intuitive human impulse—is to ban or otherwise regulate sweatshops. Focusing on the visible evil (at least once the media have brought it to their attention), and perhaps imagining the greedy motives of the sweatshop owner, or having been taught that the world’s ills can be attributed to “corporations” or a vaguely defined “capitalism,” the perverse indirect effect of regulating corporate behavior in order to stop the “exploitation” of Mrs. Tratiwoon’s son is invisible to them: namely, the effect of consigning him to the garbage dump. Given the tendency to think this way, if the state is thought to have the competence to redistribute wealth in order to solve social problems, it is unlikely that this power will be used in ways that actually help the worst off.

That’s why historically, there have been many more vigorous attempts to close down than to open up exit options. The intuitive political response to social problems seems to be to pass a magical law. Thus, with few exceptions around the world, regulation has widely been preferred to wealth redistribution. Moreover, wealth redistribution has tended to flow from the poor to the rich. In the United States, this perverse redistributive flow takes the form of mortgage subsidies; interstate highway construction that eases access to the suburbs; Social Security benefits skewed to wealthy retirees; a panoply of loans, credits, and direct subsidies to farms and other businesses; and the always-growing list of“middle-class entitlements,” including the recent trillion- dollar prescription-drug benefit. Even when “we,” and other polities of the First World, try to reach out to the Third World in an exit-enhancing manner through foreign aid, the money tends to be diverted from the poor to their corrupt rulers.

Those historical assertions, undemonstrable in this space, are the type of claim made in the books I listed in Part 1. But for now, if you doubt that exit-stifling forms of regulation and redistribution have proven to be overwhelmingly popular in country after country, ask yourself why Milton Friedman’s proposal for school vouchers has been adopted virtually nowhere. Instead, one nostrum after another is seized upon as “the” answer to an education crisis that drags on decade after decade, spreading from country to country and worsening all the time.

An economist of the Feldstein school might explain opposition to vouchers by inferring that rich and middle-class people don’t want to pay the taxes that would fund downward wealth redistribution, and there may be some truth to this view. But a more charitable explanation (which happens to be borne out by survey research) is cognitive, not motivational. If the humanly intuitive response to a problem is to have the government “do something” about it, then the most obvious thing to do is ban it, regulate it, or have the government administer it in some unnamed, magical way that will solve it. Given the propensity to think this way, voters who come to recognize that publicly administered education is, in fact, less than magical are put in the position of having to endorse one theory or another about what must be the problem (not enough tests; teaching to the tests; class size too large; fundamentals underemphasized; not enough money; passive principals; undereducated teachers; uninvolved parents… the list is endless).

That is the nature of politics: it requires us (in effect) to voice one theory or another about what has gone wrong. The possibilities that the reasons for bad education could vary from school to school, or that different children could flourish in different environments (and that exit might allow them to happen into an environment that worked well for them)—and the even more troubling possibility that we are ill-equipped to understand the causes of poor education—can’t be accommodated when politics is the only way to address failing schools: that is, when the choice of enhancing the opportunity to exit from a bad school, as vouchers would do, must compete with plausible-sounding fix-its that can be uniformly imposed through regulation. The reason to "privatize” education is that it would tend to solve educational problems by sparing voters the cognitive demands imposed on them by education politics.

I don’t have to be an automotive engineer to buy a car that runs well. If it doesn’t, I can trade it in—or enough people would have exited from the bad brand that it would already have ceased to exist. Likewise, with vouchers, a child can exit from a bad school without his parents needing to be education experts who understand why the school is bad. But the merits of exit seem too counterintuitive for the parents themselves, in their capacity as voters, to grasp. Which brings us back to Mrs. Tratiwoon’s son.

From his perspective—the perspective of the poorest of the poor—the noblest thing that compassionate students can do is not demonstrate against Nike, but buy its shoes so as to give him a better job than his current position in the garbage dump. But what could be more counterintuitive than that? Reflection isn’t what people do well, and appreciating the trial-and-error feedback process enabled by exit requires reflection, not intuition.

Democracy is, as Winston Churchill famously pointed out, the worst political system except all the others. It is the best known surety against horrible tyrants. But democracy isn’t good at producing effective public policy. To the extent that (within politically set legal limits) capitalism offers an alternative to all “political” systems, including democracy—and, therefore, an alternative to the excessive cognitive demands that politics makes on us—we should embrace that alternative, in the spirit of realism if not enthusiasm.

Once we recognize that real-world politics is an arena of ignorance and irrationality whose only means of error correction relies on ignorant and irrational people to diagnose and remedy their own errors, our conception of idealism itself ought to change. The true idealist, interested in genuinely helping others rather than making symbolic gestures of assistance, should (ironically) direct his political activism in an antipolitical direction to help unleash as much unregulated capitalism as possible—thereby helping the poor to exit from their misery as fast as they humanly can.

discuss this article

Jeffrey Friedman, who has taught history and political science at Berkeley, Yale, Harvard, and Barnard, is the editor of Critical Review and of The Rational-Choice Controversy: Economic Models of Politics Reconsidered (Yale University Press). He runs summer seminars on the political economy of modern democracy at Princeton (click on the seminar icon above).




the articles

on your campus