At a time of spectacular failures of sub-prime mortgages, catastrophic levels of personal debt, an epidemic spread of obesity, it is unsurprising that so many are now doubting citizens’ capacities to make sensible decisions. For related reasons, behavioural economists now find themselves with an audience of policy makers who are eager for advice on how to address social ills by revising institutions and policies to fit ordinary human biases and irrationalities.
The present book makes a particularly engaging case for a whole range of policy implications of behavioural economics. Behavioural economist Richard Thaler and legal theorist Cass Sunstein (now Director of the Office of Information and Regulatory Affairs in the US White House) have done an excellent job of presenting their ideas for governmental reform that, in their terms, include much-needed ‘paternalistic’ interventions without violating widely shared ‘libertarian’ principles. The rhetoric is highly compelling, and their approach is already having a significant impact. However, while the wider audience for whom the book is written may not be interested in the justification of the underlying principles, it is precisely the cracks in the foundations that pose the greatest threat to the project.
The book's central claim is this: since people are so susceptible to various counterproductive behavioural and cognitive tendencies, it is appropriate for social planners, policy makers, and other ‘choice architects’ to modify decision-making situations in ways that nudge people toward better choices, where ‘nudging’ is distinguished from ‘forcing’ by the fact that people can still choose the de-emphasized option relatively easily. For example, many employees have the opportunity to use pre-tax income to cover purchases (parking passes, commuter rail tickets, etc.) but never get around to filing the necessary forms. In light of this predictable inertia and the opportunity costs to employees, the choice architecture could be altered so that these purchases are made pre-tax by default, although employees still have the option of opting out. The number of options stays the same, but by changing the default structure, employees are benefitted – not just according to some imposed standard but by their own lights. And this sort of choice-preserving improvement in welfare represents, according to Thaler and Sunstein, an instance of ‘libertarian paternalism’. The book is chock-full of illustrations of human irrationality and nudge solutions, which makes it a very good read and a rich vein of inspiration to policymakers. As I shall argue below, however, the range of cases and the looseness of some of the terminology leave a wake of questions about the details of the underlying principles.
The book is organized into five parts. Part One contains the core of the argument, in which the key concepts are developed, including ‘libertarian paternalism’, ‘nudge’, ‘choice architecture’. And it is worth getting clear on a few of these basic claims before continuing.
Regarding our ‘biased’ nature, Thaler and Sunstein's starting point is the contrast between ‘Econs’ and ‘Humans’, the now-familiar contrast drawn by behavioural economists between the agent that is purportedly presupposed in rational choice theory and neoclassical economic theory, and real human beings. The book is very good on summarizing and amusingly illustrating huge amounts of empirical evidence for our foibles; but it is weak on showing how economic theory (let alone rational-choice policy approaches) is actually committed to the strongly idealized and methodologically naïve view that Thaler and Sunstein ridicule it for. Clearly, the point of the book is to present a vision of how to approach public policy rather than to provide a thorough analysis of various competing positions.
A ‘nudge’ is presented as any de facto influence on choice that is ‘easy and cheap to avoid’ (p. 6). We are constantly being nudged by a variety of influences such as hyperbolic discounting of future payoffs, status quo biases, and effects resulting from our susceptibility to various perceptual illusions. Importantly, by defining nudges so broadly, without reference to anyone actually doing the nudging, Thaler and Sunstein take the choice-framing effects out of the context of manipulative strategies and shift the focus to the actual effects that are realized (and the pragmatic questions of whether they are welcome outcomes or not).
The nudges that interest Thaler and Sunstein particularly are those affecting agents’ behaviour in situations of choice or decision-making. This ‘choice architecture’ includes, for example, the order in which options are presented, or the relative prominence of certain simultaneously presented options (in a school cafeteria line-up, for example, fruit ahead of snack-cake – or vice versa). The idea is to harness the biases to good effect, by redirecting the nudges built into choice architecture so that people end up making the choices they themselves want to make.
If Thaler and Sunstein are to have any chance of meriting ‘libertarian’ credentials, it is essential that the re-engineering of choice architecture not be guided by any particular political or moral agenda. Their claim here is that what makes a nudge beneficial is not that it steers people in the direction of behaviour that is objectively ‘good’ but rather in the direction of behaviour that they won't regret. That said, there are a number of points where it is hard to shake the suspicion that Thaler and Sunstein help themselves rather conveniently to the idea that, if a person's choices are largely a function of various framing effects, say, then overriding those choices violates no autonomy, since there was no autonomy there to be violated. Their idea seems to be that if, in choosing between two similar items, people tend to pick whatever product is on the right, then they ought to be happy to have things so arranged that the item on the right is the one that is more beneficial. I'll return to this below, but for the most part the authors are simply relying on the idea that anything that reduces the extent to which people make choices that they regret less than they otherwise would is best understood as respecting their internal sense of what is important.
Parts Two and Three discuss a series of examples that illustrate how the approach would work in financial and health-related contexts. One of these is Thaler's widely discussed proposal (‘Save More Tomorrow’) to increase savings by redesigning the retirement-planning situation as one in which the money one is deciding to save is ‘extra’ money that one will be getting in the future (as part of anticipated pay raises). Other proposals are aimed at improving the way in which insurance plans or mutual funds are offered, or at helping people officially make the commitment to donating organs that they say they want to make. Although not especially detailed (and I won't here try to assess the likely effectiveness of the proposed solutions), these chapters thought-provokingly illustrate how the basic principles of libertarian paternalism work.
Part Four takes a different angle, arguing in short, quick chapters that paying adequate attention to choice-architecture could open up room for approaches that are actually more libertarian. These are situations in which we have ended up with policies that offer a dysfunctionally narrow range of choice because policy-makers have been justifiably afraid of deregulation under laissez-faire approaches to choice-architecture.
Part Five wraps things up with a slate of further ideas and a rather breezy discussion of objections. The replies to objections display the agile pragmatism of the authors. No solutions are perfect for all contexts, but there are lots of cases in which nudges work. This generally seems pretty sensible, as long as they are choosing the examples. But what is lacking is sufficient attention to the principles underlying the cases, and why it is that we should go along with some nudges and not others. It can be an effective rhetorical strategy, for some audiences, to imply that the only opposition to your ideas comes from overly abstract theorists who care more about consistent principles than about effective policy. But Thaler and Sunstein also claim to be introducing a new theoretical underpinning for policy-making, and so the objections cannot be so easilydismissed.
In the remainder of this review, I will raise four further concerns not adequately addressed in the book.
(1) First, Thaler and Sunstein are misleading in their use of the term ‘paternalism’, which is standardly defined as a matter of interfering with a competent person's choices (e.g. by forcing them to act in a certain way, or reversing their choices), out of a motivation to improve that person's welfare. As Thaler and Sunstein use the term, however, it becomes equivalent to beneficence: when the government acts to improve people's welfare by influencing choices in any way, it is engaging in paternalism (p. 5). This trivializes the objections to paternalism, making opposition to paternalism the exclusive domain of hard-right libertarians. Insofar as they do acknowledge that the charge of paternalism involves claims about coercion, they insist that their version of paternalism is ‘soft’ paternalism (p. 5). Unfortunately, this is not quite true. Soft paternalism standardly refers to the cases in which the person being coerced is, for example, ignorant of certain dangers requiring immediate action (Mill's example of someone about to step, unawares, onto an unsafe bridge) or temporarily not in a sound state of mind (Dworkin Reference Dworkin1988: 107). But the appeal to soft paternalism will work for Thaler and Sunstein only if all choice-improving nudges are to be understood along the lines of misinformation and temporary insanity. And if that is the case, then it is unclear why they should put so much focus on the fact that nudges are forms of influence that are easy to resist.
(2) This appeal to the resistibility of nudges is a centrepiece of the account and particularly of its claim to be ‘libertarian’. As noted, Thaler and Sunstein argue that nudges still leave people the option of taking a different route, by opting out of automatic enrolments, selecting non-default options, looking on the bottom shelf in the supermarket, and so on. But this move is also a source of a second set of non-clarities. There are at least three lines of objection here. To begin with, many libertarians will object to any choice-steering measures being labelled ‘liberty preserving’. Anticipating this, Thaler and Sunstein point out that it's not the case that switching the default option introduces influences into a situation previously devoid of influences; it simply substitutes one set of influences for another. Alterations of the choice architecture thus need not reduce liberty, since the status quo contains patterns of influence that cannot be assumed to be innocent in this regard simply because they emerged in free market interactions. This is an important argument and goes a long way toward undermining a facile conservative strain within libertarianism.
Two further aspects of the purported resistibility of nudges are more problematic. Sometimes, it is not plausible to view the modifications to choice architecture that Thaler and Sunstein propose viewing as a matter of swapping one nudge for another. Moving from a situation in which junk food is positioned at eye-level at the check-out line in a supermarket to one in which it is less prominently displayed might count as counter-nudging or nudge-swapping; the same cannot be said for moves from a situation in which people's own inertia leads them to delay signing an organ donor card to one in which they are required to make a yes/no choice when renewing a passport or driver's licence. When a policy change increases the level of steering, typically by introducing new nudges, those concerned with the preservation of liberty will demand assurances that those being nudged are, in fact, able to resist cheaply and easily. Thaler and Sunstein provide little evidence that this is the case. Indeed, given what they themselves write about how effective nudges are and how poor we are at even noticing the influences to which we are subject, it is entirely unclear why we should expect nudges to be easily resisted. Their explicit pragmatist claim is that they advocate only those influences that really can be easily and cheaply resisted and thus qualify as nudges. But the book contains meagre grounds, empirical or otherwise, for thinking that people will be able to resist their proposed policy nudges ‘easily and cheaply’.
This leads to a further issue, one that also affects cases that can be understood as cases of nudge-swapping. Suppose we start from the plausible assumption that part of what makes an influence coercive – and thus not a ‘nudge’ but a ‘shove’ – is individuals’ inability to swim against the current channelled by choice architecture. Since individuals vary in their capacities to resist, one and the same policy may be resistible for some and not for others. This already raises concerns (nowhere addressed in Nudge) about the equality of effects that nudges have. But the fundamental methodological question is what principles Thaler and Sunstein propose to use in sorting out not only what degree of resistibility nudges must have (and for whom) but also what advantages accrue to various degrees of resistibility. After all, given what they say about how disastrous laissez-faire approaches are, why should we think that those who feel strongly about resisting nudges will be members of a choice-making elite of quasi-‘Econs’ rather that narrow-minded and impulsive individuals with an allergy to whatever the default is? Thaler and Sunstein might simply have a principled commitment to letting people who feel strongly about bucking the tide make their own mistakes, but they provide no reason to think that this won't lead to the same problems that nudges are supposed to solve.
(3) A third line of objection concerns the marginal status of respect for the autonomy of those targeted for nudges. One can distinguish two worries here. The first has to do with the value of autonomy as self-governance, according to which citizens who are subjected to various governmental influences are entitled to the opportunity to consent. This is also the most straightforward way to avoid much of what makes paternalism so genuinely offensive. Thaler and Sunstein insist that they are strongly committed to publicity and transparency, which they present as follows: ‘The government should respect the people whom it governs, and if it adopts policies that it could not defend in public, it fails to manifest that respect’ (p. 245). Together with a few lines to the effect that government officials ought to be ‘happy to reveal both their methods and their motives’ (p. 245), this mention of respect for individuals stands out in a book geared almost exclusively towards optimizing outcomes. Even here, however, the consent is merely hypothetical, and makes clear how marginal the procedures of democratic self-governance are to their approach. After all, if the nudges really had the consent of those being nudged, it would no longer be clear why the approach would need to be called ‘paternalistic’ at all. Put somewhat differently, Thaler and Sunstein generally align themselves with the perspective of planners who half-pose the presumptuously rhetorical question, ‘Who would really mind?’ (e.g. p. 11).
And this raises a related set of deep and neglected issues. For the question is not just whether people would object, if they were consulted. People also care about whether they are consulted and whether they understand the influences to which they are subject. Many of us would be annoyed to discover that the presentation of food at the office cafeteria was altered, unannounced, in order to nudge us toward healthier choices; and we'd be annoyed even if we were in favour of the measures. Thaler and Sunstein advise choice architects against being secretive about nudges, but this espousal of principle, in a late chapter responding to objections, is not borne out by the discussion in the rest of the book, which focuses almost exclusively and rather effusively on the possibilities for channelling individuals’ choice-behaviour toward better outcomes. They do, of course, regularly mention how much they themselves need these nudges, and this serves to create the impression that they are allied with ordinary folk. But the espousal of transparency and publicity constraints comes across as an artificial and ad hoc declaration of values that belies a lack of real interest in the importance of ensuring that those subjected to these subtle forms of state power understand the underlying rationale.
(4) One response that Thaler and Sunstein can make is that effective, affordable and politically realistic solutions to urgent social problems require the sort of pragmatic, managerial, outcomes-oriented approach that libertarian paternalism (or, more accurately, I think, ‘nudge paternalism’) advocates. But this is only true if there aren't good alternatives. And most of the alternative approaches discussed in Nudge tend to be naïve and unrealistic. There are, in fact, alternative approaches to improving decision making that are perfectly compatible with the behavioural economics and social psychology discussed by Thaler and Sunstein. For example, there is a growing body of research on how to construct contexts of interpersonal deliberation to generate stable solutions to collective action problems, by improving people's collective decision-making capacities, rather than subjecting them to nudges (see, with special reference to Thaler and Sunstein: John et al. Reference John, Smith and Stoker2009). More generally, one comes up with very different recommendations if one thinks of the problems to which nudges are supposed to be a solution not as a matter of human failings per se, but rather as a socially and historically contingent misfit between the decision-making capacities that particular policies and institutions require of individuals and the capacities that people actually have. Conceived in the latter way – in terms of what I have elsewhere dubbed ‘autonomy gaps’ (Anderson Reference Anderson and Forst2009) – it becomes an open question whether what is called for is a choice architecture that makes it easier to avoid regrettable decisions or, rather, various measures to improve individuals’ decision-making capacities, say, through education, ‘buddy’ arrangements, decision-making heuristics, etc. Given the concerns raised above regarding Thaler and Sunstein's approach, they owe readers more of an explanation of why nudges are necessary when such alternatives exist.
In closing, it is perhaps worth noting the political context of the book. Thaler and Sunstein hope to provide a way to transcend the highly politicized divide between Left and Right (particularly in the USA) in a way that will break up the logjam blocking much-needed policy changes. Their hope is that by avoiding both coercive mandates and laissez-faire negligence, their science-based pragmatism can create political common ground for policy reforms that will improve outcomes in ways that will be welcomed across the political spectrum. This is a laudable goal, and the book is on its way to having a significant and probably quite beneficial impact. But the authors’ enthusiasm for policy implications and their pragmatist avoidance of issues of principle frequently conspire to muddy the conceptual waters in troubling ways. When virtually any sensible policy initiative starts to look like an instance of a beneficent redesign of choice architecture, it becomes difficult to spot unduly paternalistic implications.
It is difficult to deny that there is an important role for nudges in a sensible overall approach to policy-making. Thaler and Sunstein are to be commended not merely for an engaging and innovative book, but also for adding nudges to the toolkit of policy makers. But much work still needs to be done in identifying the principles for implementing nudges appropriately and in a way that cuts out the lingering paternalism.
