Cass Sunstein and “Libertarian” Paternalism
by David Gordon
Most often the state compels you to do things, not because these things are supposed to be good for you, but because they fulfill the state’s purposes. The state doesn’t take your money to help you. Sometimes, though, the state does pass laws that claim to restrict people for their own good, e.g., laws that forbid use of certain drugs that are supposed to be bad for your health. Laws of this kind are called paternalistic.
Libertarians of course oppose paternalism, but it is not only libertarians who reject it. It is at odds with the entire heritage of classical liberalism. John Stuart Mill famously opposed paternalism in On Liberty; and it is Cass Sunstein’s principal aim in Why Nudge?: The Politics of Libertarian Paternalism to cast doubt on Mill’s canonical statement of anti-paternalism, the Harm Principle. This principle is the following: “[T]he only purpose for which power may be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or mental, is not a sufficient warrant.” (Sunstein here is quoting Mill’s On Liberty.)
Sunstein challenges what he considers the two main arguments that support the Harm Principle. The first of these, which he considers the more important of the two, is the Epistemic Argument: “Because individuals know their tastes and situations better than officials do, they are in the best position to identify their own ends and the best means of obtaining them. … In my view, it [the Epistemic Argument] provides the strongest support that the Harm Principle can find.”
One way to challenge the Epistemic Argument would be to claim that certain things are good or bad for people, regardless of their own ends and desires. Smoking is bad for you and you shouldn’t do it, proponents of this position would say, even if after careful consideration you want to smoke. But Sunstein does not follow this path. He could hardly claim to be a “libertarian” paternalist if he did. Rather, he points to cognitive mistakes that people make. He is not trying to impose his view of what people should do on others: he is saying to them that paternalistic interventions can add to people’s well-being by helping them to act more rationally.
Even if this is what Sunstein aims to do, though, isn’t it still odd for him to claim to be a libertarian paternalist? Sunstein attempts to reduce the tension between adjective and noun by appealing to “nudges.” “In light of the pervasive risk of government error and the inescapable fact of human diversity, it is usually best to use the mildest and most choice-preserving forms of intervention. These forms include ‘nudges,’ understood as initiatives that maintain freedom of choice while also steering people’s decisions in the right direction (as judged by people themselves).” The nudges, as here explained, invade people’s freedom less than other paternalistic measures; but this hardly suffices to make them libertarian. By similar reasoning, one could call a robber who refrained from murdering his victims a “libertarian” aggressor.
Let us return to cognitive mistakes. Sunstein is a leading figure in behavioral economics, and he writes about these mistakes with especial authority. Following the psychologist (and Nobel Prize-winner) Daniel Kahneman, he distinguishes between two “cognitive systems” in the mind. “System 1 works fast. It is often on automatic pilot. Driven by habit, it can be emotional and intuitive.” By contrast, System 2 is “deliberative and reflective.” When we operate, as we often do, with System 1, we are subject to various sets of mistakes, which count as “behavioral market failures.” With the details of these mistakes, we are not here concerned, but the errors include “present bias and time inconsistencies,” “ignoring shrouded (but important) attributes,” “unrealistic optimism,” and “problems with probability.” What for our purposes is important is the conclusion Sunstein draws: “With respect to paternalism, the unified theme is that insofar as people are making the relevant errors, their choices will fail to promote their own ends. It follows that a successful effort to correct these errors would generally substitute an official judgment for that of choosers only with respect to means, not ends.”
Suppose, for the moment, that we accept Sunstein’s claim that these cognitive mistakes impede people from getting what they want. Does this give one reason to reject the Epistemic Argument? I do not think so. According to the Epistemic Argument, each person is in a better position than government officials to choose the appropriate means to satisfy his ends. This is entirely consistent with people’s making cognitive mistakes. The point of the Epistemic Argument is that people can better judge their situation than officials can, not that their judgment is without error.
Mises fully realized this point, and Sunstein would have profited from a reading of Mises’s comment in his essay “Laissez-Faire or Dictatorship” on J.E. Cairnes’s objection to laissez-faire: “Let us for the sake of argument accept the way in which Cairnes presents the problem and in which he argues. Human beings are fallible and therefore sometimes fail to learn what their true interests would require them to do. … It is very unfortunate that reality is such. But, we must ask, is there any means available to prevent mankind from being hurt by people’s bad judgment and malice? Is it not a non sequitur to assume that one could avoid the disastrous consequences of these human weaknesses by substituting the government’s discretion for that of the individual citizens?”
The objection I here have in mind differs from one that Sunstein does consider. Sunstein knows full well that government officials are also subject to cognitive mistakes and have their own agendas. Incredibly, his response is that technocrats are more likely than the public to be influenced by rational, System 2 thinking. “A large virtue of technocrats in government — specialists in science, economics, and law — is that they can help overcome some of the errors that might otherwise influence public or private judgments.” He does generously allow, though, that biases of the government officials require further study.
My objection, though, is not that the officials are biased and self-interested, though they are indeed that. It is rather that “nudging” people to act in ways they would not otherwise have chosen disregards the fact, to which the Epistemic Argument calls attention, that they are the best judges of how to deal with their individual situations. Only if their cognitive defects were so severe that they outweighed the force of the Epistemic Argument would Sunstein have a good argument for paternalism. To reiterate, the problem I have in mind is not whether the government officials are more likely than the public to suffer from cognitive defects. It is that the existence of cognitive mistakes does not by itself refute the Epistemic Argument.
There is a further problem with Sunstein’s use of cognitive mistakes to justify paternalistic interventions. He offers no evidence that people who act in ways he wants to modify have fallen victim to cognitive mistakes. Do people who smoke, or consume sodas in large quantities, or fail to buy fuel-efficient cars, suffer from cognitive mistakes? Perhaps they do, but the fact that people are susceptible to these mistakes does not show, for any particular choice, that it stems from a mistake.
Sunstein criticizes another argument for the Harm Principle. This argument appeals to autonomy: “We might insist that people have a right to choose and that government cannot legitimately intrude on that right even when it does in fact know best. … On this view, people should not be regarded as children; they should be treated with respect. They should be seen as ends, not means.”
This is of course the second formulation of Kant’s Categorical Imperative and is today widely accepted as a principle of morality, even by philosophers not in orientation Kantian. Unfortunately, Sunstein is tone deaf to its force. He thinks preference for freedom of choice is at best a component of welfare. If it is taken to be more than this, it stems from System 1 thinking: it is a “rapid, intuitive judgment about welfare.” Besides, many people do not want what they consider an overabundance of choices. (But aren’t such people free to seek situations where they would confront fewer alternatives?) I fear that Sunstein, like all-too-many economists, is so committed to welfare as the objective of morality that he is unable to understand respect for persons. This phenomenon is itself a cognitive defect, albeit one that has yet to attract the attention of behavioral economists. I do not recommend government intervention, even the mildest nudge, to correct it.