Classical liberals naturally resist paternalism. Milton Friedman wrote that “the paternalistic ground for government activity is in many ways the most troublesome to a liberal; for it involves the acceptance of a principle—that some shall decide for others—which he finds objectionable in most applications.” Besides cases involving “children” and “madmen,” people identifying as liberal have historically considered any paternalistic policy to be unacceptable.

That’s not the case today. In their new book, Government Paternalism, Julian Le Grand of the London School of Economics and Bill New, an independent policy analyst, offer a powerful argument in support of government paternalism. In particular, they endeavor “to see if they could successfully meet what might be thought of as the John Stuart Mill challenge: are there circumstances in which the individual’s own good is sufficient warrant to justify a paternalistic intervention?” They define government paternalism, describe conditions under which they believe it is justified, and present three potential scenarios for paternalism.

When is paternalism warranted? / After reviewing the academic literature on paternalism, the authors “conclude that a government intervention is paternalistic with respect to an individual if it is intended to address a failure of judgment by that individual [and] further the individual’s own good.”

By “failure of judgment” or “reasoning failure,” they mean bad decisionmaking. They attribute such failures to four separate causes:

  • limited technical ability
  • limited imagination/​experience
  • limited willpower
  • limited objectivity

To possess “limited technical ability” is to struggle with—or simply ignore—math. For example, the buyer of a lottery ticket probably doesn’t bother to compute the probability of winning. Having “limited imagination/​experience” is to not appreciate how one’s perspective would change in alternative scenarios. For example, a healthy, financially secure young adult cannot fathom being a poor senior citizen in need of medical care. “Limited willpower” is straightforward enough. Young adults who can imagine being old and who know how to calculate how much income they’ll need in retirement might still have “limited willpower” to save regularly. As for “limited objectivity,” there are many types of this failure. Smokers, for instance, may lowball their increased risk of cancer, thinking simply that it would never afflict them. Le Grand and New, consistent with behavioral economists, expect us to think that reasoning failure afflicts almost everyone.

Consistent with conventional economists, the authors assume that individuals face tradeoffs—in the case of paternalism, between well-being and autonomy. The authors “acknowledge that autonomy can be ‘placed on the scales’ and weighed against an individual’s well-being.” It is possible to imagine special cases of individuals giving up autonomy in order to obtain greater well-being. People stranded on a desert island might be willing to give up some autonomy in order to acquire adequate food, clothing, and shelter. Le Grand and New try to convince us that not-so-desperate individuals will also give up some autonomy in order to achieve greater well-being.

Policy tools / The tools of government paternalism are “legal restrictions,” “taxation,” “subsidy,” and “nudging or framing.” By legal restrictions, the authors mean prohibition (of alcohol, for instance) or a mandate (the use of seatbelts, for instance). Paternalistic taxes, of course, aim to reduce behavior such as smoking or drinking. Paternalistic subsidies aim to promote activities such as “the cessation of smoking or the eating of healthful foods.” Nudging and framing reflect the idea of “libertarian paternalism” as described by Cass Sunstein and Richard Thaler. According to Le Grand and New, taxes, subsidies, and nudging produce a more favorable tradeoff between well-being and autonomy than legal restrictions. Three scenarios illustrate their analysis.

If a government paternalist can demonstrate that some individuals participating in a given activity suffer from some reasoning failure, he can then use a policy tool to counteract that failure. The authors assert that the following outcomes increase the likelihood that such intervention is justified:

  • People who suffer from reasoning failure “experience a large increase in well-being as a result” of the intervention.
  • People who do not suffer from reasoning failure but who nonetheless change their behavior because of government intervention “do not suffer greatly as a result” of the intervention.
  • Concerning all people who change their behavior because of intervention, some of whom suffer reasoning failure and some of whom don’t, the people with reasoning failure outnumber the people without.
  • Among all those who do not change their behavior despite the intervention, the people without reasoning failure outnumber the people with reasoning failure.

If those conditions seem abstruse, the authors’ scenarios should help illuminate.

First, consider smoking. According to the authors, many smokers have “limited willpower,” fail to anticipate becoming addicted, or lowball their chances of contracting cancer. Yet Le Grand and New reject outright prohibition of tobacco because of historical experience “of the prohibition of alcohol in the United States and of the ‘war’ against illegal drugs.” They reason that “it is likely that a significant portion of both those with reasoning failure and those without it will continue smoking” despite any legal restrictions. Among other reasons, they claim “there is a substantial impact on the autonomy of smokers of any kind, whether they suffer from reasoning failure or not, or whether they stop smoking or not.” Although the authors’ analysis appears long on speculation and short on data, they conclude that prohibition is worse than the alternatives.

An alternative policy is taxing the sale of tobacco. The main difference between a tax and a ban is the effect on autonomy: a government ban eliminates the choice to smoke legally, but taxation allows choice although it exacts a price for lighting up. Thus the authors favor a tax over a ban.

The authors also endorse the “libertarian paternalist idea” of a “smoking permit.” Under such an intervention, the smoker’s autonomy remains intact except for the nuisance of obtaining the permit, which makes the permit a better policy than a ban.

But what are the effects of a permit on well-being? Le Grand and New conjecture that smokers with reasoning failure who forgo the permit and quit smoking will “likely” experience a substantial increase in their well-being. Borrowing from another “libertarian paternalist” idea, they reason that if the number of spendthrifts with reasoning failure who remain automatically enrolled in a pension plan is a “reasonable guide” to the number of smokers with reasoning failure who would forgo the permit and quit smoking, there may be many of them. The authors offer no evidence, however, that the decision to forgo a smoking permit reflects the decision to remain enrolled in a pension plan.

It is easier to buy their argument that the number of smokers without reasoning failure who forgo a permit will be small. Given that those smokers believe the benefit of smoking is greater than the health risk, there’s nothing stopping them from obtaining a permit other than the nuisance.

Because Le Grand and New expect a large number of smokers with reasoning failure to decline the permit and quit smoking, one may infer that they expect a small number of smokers with reasoning failure to obtain the permit and continue smoking.

All those groups of people—those with and without reasoning failure, and those who forgo the permit and quit smoking and those who don’t—are at least as well off as they would be when faced with a ban or a tax on smoking. The authors conclude that “the permit idea at least seems to be worthy of serious consideration.”

The second scenario Le Grand and New consider is saving. “In general,” they claim, “people do not save enough for their pension.” Therefore the government paternalist aims to encourage individuals to save more.

The primary source of reasoning failure in this case is “limited imagination”: young people cannot foresee themselves as old. Le Grand and New dub those with reasoning failure, who save little of their incomes, “myopics.” In contrast, “farseers” save little of their incomes but they do not suffer from reasoning failure—apparently they have some good reason for near-term consumption.

Three policies aim to promote saving: “legal compulsion,” “tax relief,” and a “libertarian paternalistic policy whereby people are automatically enrolled in a pension plan unless they opt out.” If the government requires everyone to save, according to the authors, the well-being of myopics will increase and the well-being of farseers will decrease. They do not mention the well-being of myopics or farseers who ignore the mandate. They do mention that “there is obviously a significant loss in autonomy for all groups, both perceived and actual.”

Tax relief increases the well-being of myopics who take the tax break and save more. Their increase in well-being “is likely to be greater,” Le Grand and New add, than under a mandate because tax relief reduces the cost to them of saving more. They point out that tax relief will induce fewer myopics to save more than a mandate. Thus, it is unclear whether the increase in well-being per myopic who saves more, multiplied by the number of myopics who save, is greater under tax relief or a mandate.

Tax relief increases the well-being of farseers who take the tax break and save more. The authors do not mention the well-being of myopics who ignore the tax break; presumably it stays the same. The authors do mention that the well-being of farseers who decline the tax break and continue to save little remains the same. Tax relief causes little, if any, decrease in anyone’s autonomy. Le Grand and New point out one drawback: “Tax relief can be highly regressive.”

The “libertarian paternalistic policy” designed to promote saving switches the employee’s default decision from not participating to participating in a saving plan. In other words, employees are “automatically enrolled … unless they opt out.” The well-being of myopics who remain enrolled and save more increases. In theory there would be no farseers who remain enrolled, because farseers lack reasoning failure and would fill out the paperwork to avoid saving more than they prefer. On the other hand, there would be few myopics who fill out the paperwork in order to avoid saving because they do possess reasoning failure. As with tax relief, “there may be little loss in autonomy.” Le Grand and New implicitly endorse this policy of “changing the default position” over the alternatives of a mandate and tax incentives to encourage more saving.

Knights or knaves? / Le Grand and New are not overzealous; they demand evidence of reasoning failure before intervention and they require a gain in well-being that outweighs a loss in autonomy. They see the possibility of “too much paternalism.” Consider their analysis of a ban on assisted suicide. The authors reckon that people with reasoning failure who cannot commit assisted suicide will experience an increase in well-being, and people without reasoning failure will experience a decrease in well-being. They argue that the number of people without reasoning failure “is likely to be much larger than” the number with reasoning failure. Therefore the ban reduces overall well-being. Couple that with the loss in autonomy and Le Grand and New decide that “assisted suicide should not be prohibited.”

Credit the authors for addressing objections to government paternalism. They ask, “Might not [government officials] be subject to the very kinds of reasoning failure that we have ascribed to people engaging in self-damaging behavior?” According to them, reasoning failure applies to individuals making “their own decisions,” but not individuals making “decisions on behalf of others.” Therefore the authors are confident that citizens will not bear the burden of bad decisions made by disinterested government officials.

Le Grand and New also believe that government paternalism withstands the insights of public choice economics. They recognize that, for example, even if politicians are nominally responsible for promoting the well-being of citizens, they might do whatever garners the most votes. They nevertheless have faith that “elections,” “referendums,” and “sunset clauses” will make government officials act more like “knights” than “knaves.”

Le Grand and New recognize that government paternalism may weaken an individual’s ability to make good decisions. Their response, if my understanding is correct, is that a reduced capacity to make decisions is a part of the autonomy loss that must be offset by an increase in well-being. That, perhaps, is insufficient assurance that government paternalism will not cause decisionmaking ability to atrophy. And one wonders why we don’t leave paternalism to family and friends.

Le Grand and New conclude that government paternalism is a “helpful friend,” not a “nanny state,” because government paternalism “usually operates in an impersonal manner.” Though the authors are not overzealous, they might be overconfident that justifying even more politicization of decisionmaking carries as little risk as they seem to think it does.