This article belongs to the debate » Choice Architecture in Democracies
05 Januar 2015

Why not Nudge?

Now as ever, I agree with Cass Sunstein’s views on many matters. I above all agree that nudging is compatible with any defensible liberal idea of autonomy, and especially with the undeniable claim that nudges can often enhance autonomy in the empire of caveat emptor. Indeed, my concern is that libertarian paternalism is too libertarian, not too paternalistic. If one values current regulations on seatbelts, bicycle helmets, smoking, gambling, compulsory pension contributions, employment law, consumer protection, usury, as well as alcohol, drugs and medicines, then one ought to be very suspicious of how the idea is deployed in public policy. The crucial issue for social and economic policy is the role of mandatory, non-waivable regulation (mandates, penalties and taxes, also called ‘hard paternalism’) vs. waivable, optional default rules (nudges, also called ‘soft paternalism’). Nudges are better than nothing, but contrary to the leitmotif of Nudge (2009) and Why Nudge?: The Politics of Libertarian Paternalism (2014), mandates are almost always better in any social policy of large importance, and very few of them offend any defensible idea of liberal autonomy. But libertarian paternalism actually furnishes arguments against mandates. Let me explain these views by elaborating upon a few short propositions.

1. The argument for nudging is limited to regulation of harm to self, which properly understood is a relatively small policy space.

In Why Nudge? Sunstein seeks to qualify JS Mill’s harm principle, which holds that society may only use coercion to regulate a person’s harm to others, rather than harm to him or herself. Mill’s key argument was that each knows his or her own interests better than the community does. Drawing on empirical data on irrational decision-making, Sunstein shows why this is wrong-headed. But his arguments for nudging instead of mandates are all directed at (or are logically concerned with) self-regarding conduct. This is because nudging builds in respect for the idea of autonomy, which is the libertarian aspect of the idea. However, autonomy provides no reason for limiting coercive regulation of harm to others.  The problem here is that the line between harm to self and harm to others is notoriously difficult to maintain in any real social policy. With seat belt laws, for instance, the amount of public money spent on policing, medical and related services for accidents complicated by failure to wear a seat belt is quite considerable. In the domain of financial regulation, the US Securities Act of 1933 imposed strict disclosure rules on issuers and underwriters of debt and equity securities. A key justification was the protection of investors, many of whom were shamelessly defrauded in the speculation leading to the 1929 financial crisis. But we all know now what President Roosevelt saw then: that what rocks the securities market rocks the economy (or even the world). Now, if we look at policy areas that are traditionally considered paternalistic, such as seatbelts and helmets, risky sports, prescription medicines, smoking, obesity, and arguably pensions and consumer protection, it becomes clear that regulation of all of them is in significant part motivated by concerns over harm to others. If society has a non-waivable duty to rescue even those who recklessly hurt themselves (as almost everyone believes it does), then the reckless risk-takers would in fact be free riders on the costly public duty.

2. There is a tendency to advocate nudging instead of mandates for regulating harm to others.

In Why Nudge?, Sunstein repeatedly refers to ‘texting while driving’ and ‘fuel standards’ as areas where nudging is appropriate. These are obviously concerned with harm to others, so the argument for nudging is inapt. But if Sunstein – whose politics seem impeccable by social democratic standards – is liable to keep so long a leash on nudging, what will the (neo)liberals do with it? My fear is that they will avoid effective forms of regulation and favour nudging instead, because it offers the image of action without coercively regulating business or taxing citizens. For different but related reasons, liberal egalitarians and social democrats may opt for it as “politically possible” instead of fighting the tougher fight for effective penalties or taxes. These fears were stoked when the idea took root in public policy in Britain under the austerity government of Prime Minister David Cameron’s Conservative Party. The PM put Thaler and Sunstein’s book on his ‘summer reading list’ for party members. Now the idea is now also embraced by the Liberal Party Premier of New South Wales, Mike Baird, a former investment banker and ‘fiscally conservative Christian premier.’ conservative. Britain’s Nudge Unit (aka Behavioural Insights Team) has issued a string of policy innovations, most of which seem like improvements. But all of them, without exception from what I can tell, concern regulatory areas that involve harm or duties to others rather than harm to self. The head of the Nudge Unit, David Halpern, elsewhere described as ‘unashamedly pro-business’, told the Guardian newspaper ‘we try to avoid legislation and ordering…the real question is what is a more or less effective way of doing it?’ That is very much is the real question.

3. Mandates are more effective than nudges.

Nudging is proposed as a soft alternative to things like mandatory pension plans and taxation of tobacco, or, to put it more formally in Sunstein’s language, as an alternative to any ‘material cost’ imposed on a regulated person. (Why Nudge?, p.55-61). Now, if we remove taxation and civil fines from our regulatory toolbox, then our ability to regulate smoking, alcohol, drugs, helmets and seat belts would be radically constricted, as would non-waivable rules in consumer protection law. Yet in most of these cases, there is ample evidence showing not only the superiority of mandates, but the vast superiority. Smoking provides a dramatic example. The International Tobacco Control Policy Evaluation Project reports on the comparative efficacy of health warning labels and of taxation. The reports make clear that taxes are clearly ‘the most effective tobacco control method,’ over and above their revenue generation that can offset the smoking related harms to the public purse. Warnings led to an average of between 10-15% of smokers in wealthy countries giving up a cigarette ‘at least once,’ whereas the American Lung Association reports that a ten percent increase in the price of cigarettes reduces consumption among adults by four percent and among youth by seven percent. Sunstein in fact reviews the tools available to regulate smoking, and while he avoids taking a clear position, he does use the span of regulatory tools to show that the less effective means of regulation (i.e. the soft ones) are those that are compatible with his general approach: Why Nudge? pp. 82-86. In any event, he does introduce an important caveat about ‘hard paternalism’ that would in fact allow mandates and taxes in such situations: ‘If the benefits of the mandate plainly outweigh the costs, it would seem justified, at least if that assessment can be trusted, and if there is no real problem from the standpoint of autonomy.’ (p.142) Well, the rest of his book gives us good reasons to think that the latter two concerns are generally not a significant problem in democratic societies, so what remains seems like a straightforward balancing exercise. In my view, mandates will win that exercise most of the time, and that is plainly inconsistent with his general argument, which is that ‘nudges are usually the best response.’ (p.17).

Let me take one further, and important example of where the logic of nudging may inadvertently lead us. A much publicized and used example by Thaler and Sunstein is the promotion of employee self-saving for retirement by defaulting them into a voluntary 401(k) scheme. On its face, it is an excellent idea and not remotely paternalistic – it prods people with an educated guess about what they would want if they knew the facts, and if they do in fact know the facts and don’t want the option, they untick the box. Brilliant. However, the real world options are not nudged 401(k) saving or bust. The real options at the level of social policy are whether to either (1) increase compulsory social security contributions or taxation rates order to fund a dignified retirement and eliminate elder poverty, or (2) use a nudging device to encourage people to opt in to more (tax-assisted) private saving. Which is better? At what? Let’s start with the observation that the system of social insurance is basically compulsory, and can probably only be defended on hard paternalistic grounds (a debated issue in the philosophical literature). At any rate, if our goal is reducing elder poverty, including by saving the improvident from themselves, then higher social insurance contribution rates, supplemented by robust, tax-funded social assistance and health care services (those that do not make retirement income contingent on life-long earnings income), are far more effective. But neither higher mandatory contributions nor tax increases are nudge-friendly. I suspect that both are chiefly to account for the fact that elder poverty is much higher in the US (and Britain) than in most northern European countries.

4. Autonomy rarely precludes mandates.

Sunstein argues that ‘social welfare is the master concept’ (p.164, chs. 3 and 4), and hence his arguments in Why Nudge? appear to be an autonomy-respecting form of consequentialism. Some readers, perhaps especially in Germany, will be concerned that everything said above sounds consequentialist and thus misses the point about the inviolability of our rights, including the right to autonomy. Admittedly, the best defences of hard paternalism do come from consequentialist philosophers – Robert Goodin (‘In Defence of the Nanny State’ (1995)), and recently, Sarah Conly (Against Autonomy (2012)). But liberal egalitarian rights theorists also tend to endorse the forms of hard paternalism that I contrasted with nudging above. In The Morality of Freedom (1986), Joseph Raz sets out a liberal perfectionist version of the harm principle that would permit forms of hard paternalism that do not criminalise the self harm at issue. Ronald Dworkin, who offers a prominent non–perfectionist account of liberal equality, not only accepts but supports the kinds of ‘openly paternalistic’ mandates under consideration here, such as seatbelt laws and compulsory insurance: see his ‘Sovereign Virtue Revisited’ (2002) 113 Ethics 106, 113-115. Simply put, these philosophers do not accept that autonomy requires that a person’s immediate desires should be sovereign, come what may for their own well-being. They explore how such choices might be regulated (coercion vs. taxation or discouragement), whether they are compatible with the person’s own deeper goals, whether such goals change over time, whether the choices actually promote the person’s own autonomy, and so on. And of course, they explain why autonomy does protect people against coercive paternalistic interventions on issues of critical personal or political importance, as well as moralistic interventions where self-harm is not even at issue (e.g. gender identity and sexual relationships). 


Nudging offers a compromise in the perennial tradeoff between promoting welfare and respecting autonomy over self-regarding choices. No doubt there are some, perhaps many, instances where nudging strikes the right balance.   But we must keep in mind that this is a purely contingent matter. Though uncompromising, hard paternalism — mandates, taxation, and the like — will strike the right balance in a great many cases and both should and does play a more central role than libertarian paternalism gives it credit for.

The Author thanks Dr. Bryan Thomas for very helpful feedback and contributions to drafting.


SUGGESTED CITATION  King, Jeff: Why not Nudge?, VerfBlog, 2015/1/05,, DOI: 10.17176/20181004-133144-0.

Leave A Comment


1. We welcome your comments but you do so as our guest. Please note that we will exercise our property rights to make sure that Verfassungsblog remains a safe and attractive place for everyone. Your comment will not appear immediately but will be moderated by us. Just as with posts, we make a choice. That means not all submitted comments will be published.

2. We expect comments to be matter-of-fact, on-topic and free of sarcasm, innuendo and ad personam arguments.

3. Racist, sexist and otherwise discriminatory comments will not be published.

4. Comments under pseudonym are allowed but a valid email address is obligatory. The use of more than one pseudonym is not allowed.

No Comments Join the discussion