The new paternalism does not replace older wisdom

Abstract

The Berlin and Lewis epigraphs draw out the moral psychology of paternalism. The paternalist, enlightened as he is, sees past your empirical shell and into the true essence of your being. Setting aside the protestations of the shell that you mistake for yourself, he steers you towards the good. The steering might involve violence, but that is a small price to pay in the grand scheme of things. Nobody enjoys being flogged for heresy; but eliciting conversion is worth the pain. The convicted paternalist persists in his drive to serve the good with the approval of his own conscience.

The liberal tradition rejects paternalism and exalts the dignity and freedom of the individual person. We have been gifted by our Creator with the inalienable rights to life, liberty, and the pursuit of happiness, Thomas Jefferson asserted. Each should be free, Adam Smith said, to pursue his interest in his own way. The question of the good is too discursive, in most cases, to warrant violence against those with differing opinions. And in any event, we ought to presumptively respect human freedom and agency per se, regardless of whether someone follows a life path that we approve.

The liberal rejection of paternalism notwithstanding, a new species of paternalism has in recent decades reared its head in the West. This new paternalism has come forth on the wings of inferences from laboratory experiments and surveys, in cognitive psychology and behavioral economics, about the extent of human rationality. Evidence, it is claimed, has accumulated that we human beings are consistently poor reasoners: we are “predictably irrational” (Ariely, 2010). Instead of properly calculating probabilities when assessing risks and making decisions, it has been discovered that we rely on a range of mental short cuts or heuristics. The heuristics work sometimes, but they systematically lead us to make avoidable mistakes in, for example, financial decisions (Tversky & Kahneman, 1974). We “think fast” when apparently we should be “thinking slow” (Kahneman, 2011). Irrationalities are said to abound in other areas, too. The salience of the present, for example, leads us to overestimate the benefits of present pleasures and underestimate the pain of future costs (Laibson, 1997; Read & van Leeuwen, 1998). Our ‘present bias’ erodes our self-control and subverts our goals. We save too little, smoke too much, and repeatedly renege on our New Year's resolutions.

Each of us predictably fails to pursue the good as we ourselves understand it. And we do so across a wide set of domains.

The new paternalism proposes to use knowledge of our decision-making processes to steer us towards what we ourselves judge to be good. Part of the steering involves the regulation of ‘choice architecture’. This refers to the framing of choice options in a situation to render certain options salient to the chooser, and hence more likely to be chosen. A classic example of choice architecture is product-placement decisions – products that brick-and-mortar businesses more urgently wish to sell are placed in physically prominent locations, with the goal of drawing the attention of consumers. Choice architecture is ubiquitous in life, in the sense that every choice situation, whether it is product arrangement decisions at the grocery store or the design of retirement savings enrolment processes, necessarily has an architecture or arrangement.

Choice architecture matters, intuitively. Candy placed in the checkout aisle at grocery stores almost certainly increases the likelihood in some cases that customers will purchase candy. The long-run effects of certain choice architectures, however, are less apparent. It is clear that people habituate to their environments and develop strategies to deal with, say, temptations to buy candy on every shopping excursion. Humans are remarkable creative in their “intimate contest for self-command” (Schelling, 1980).

Since the early 2000s, especially in the wake of Richard Thaler and Cass Sunstein's popularisation of the new paternalism in their groundbreaking 2008 book Nudge: Improving Decisions About Health, Wealth and Happiness, new paternalists have proposed regulations for private choice architecture. That is, they have proposed regulations to alter existing, private choice architectures in areas in which people, according to the new paternalists, are prone to self-subversion, areas such as tobacco, sugar, and alcohol consumption, financial risk assessment, and retirement-account contribution decisions. The new paternalists have emphasised the non-coercive and even “libertarian” (read: choice-preserving) character of these regulatory proposals (Camerer et al., 2003; Sunstein & Thaler, 2003a). Many of the proposed regulations do not, strictly speaking, truncate the range of consumer options – although they certainly raise the transaction costs of options disfavoured by the self-styled choice architects. However, regulations requiring, for example, the posting of calorie counts or the presence of sugars or trans fats in food products, the changing of default labour contracts from “at will” to “for cause” termination (Sunstein & Thaler, 2003b, p. 1175), the physical placement of desserts above eye-level or in the back of a store, and more explicit and repeated announcements of bank account overdraft fees (Bar-Gill, 2012, p. 95) are coercive to firms. Such restrictions require compliance and impose penalties for non-compliance.

Beyond regulations imposed on firms, the new paternalist movement has shown itself to be quite unlibertarian in recent years towards consumers. It has moved closer and even openly towards classical paternalism. A forensic examination of the philosophy and methods of the new paternalists reveals that it has never, in fact, succeeded in being anything else – the new paternalism has always been classical paternalism informed by behavioural science. But before turning to the philosophical issues, it is important to briefly record the trend of the movement.

The new paternalists have described themselves as paternalists of means rather than ends. As Richard Thaler and Cass Sunstein wrote in the final edition of Nudge, “the paternalistic policies that we favor aim to influence choices in a way that will make choosers better off, as judged by the choosers themselves. This is a paternalism of means, not of ends; those policies help people reach their own preferred destination” (Thaler & Sunstein, 2021, p. 7; emphasis in original). But in some areas, some new paternalists have discovered that we are so prone to erring that outright coercion is warranted. We are too myopic, inconsistent, and powerless to do those things that we really want. In the extreme, some new paternalists have simply moved from these assumptions to embrace open coercion. Sarah Conly has argued that we are so unable to control ourselves, due to interactions between our psychological make-up and the environment of the modern commercial world, that “we should, for instance, ban cigarettes; ban trans-fats; require restaurants to reduce portion sizes to less elephantine dimensions; increase required savings; and control how much debt individuals can run up” (2013, p. 1; cf. Le Grand & New, 2015, pp. 167–76). Sunstein also seems to have moved somewhat in this direction recently, admitting that choice and welfare do not always dovetail, and when they part ways, it is welfare that ought to be prioritised. In other words, coercive regulations can sometimes be warranted on the grounds that they are good for the coerced parties (Sunstein, 2020).

One way to think about this confession of new paternalists is in terms of the unavoidable long-term convergence of ‘means paternalism’ and ‘ends paternalism’. The new paternalists attempted to help us do what we want to do by trying to reframe our choices. Reframing food consumption by mandating the provision of calorie information or indicating the presence of trans fats doesn't eliminate our ability to consume high-calorie or trans-fat-laden products, but it does, it is claimed, induce us to reconsider the best means for attaining our life goals: if we want to live a long, pleasurable life, is it really in our interest to consume such unhealthy products? Reframing the decision to smoke by requiring tobacco companies to put gruesome images on cigarette packages aims similarly to encourage reflection on the means of attaining our goals. Notwithstanding regulated reframing in food, the new paternalist still observes people consuming unhealthy foods. Notwithstanding the fact that everyone knows that smoking is associated with increased risk of disease – in part because they see pictures on every cigarette pack – people still smoke. A natural conclusion is that people have considered the downsides and judge the benefits and pleasures to be worth the associated risk. Another conclusion is that people are so biased in their decision-making in these areas, so unable to determine the best means given their final goals, that to help them help themselves we must further reframe the context of choice – and perhaps even ban products that lead them to undercut their aims entirely. Hence the current proposed generational tobacco ban in the United Kingdom. Thus, what starts as paternalism of means – a paternalism that takes individual desires seriously – ends up taking certain ends off the table, on the assumption that these ends are but inappropriately selected means of attaining higher, more significant ends.

The problem, however, is that we can refashion any paternalism of ends into a paternalism of means until we arrive at deeply disputed questions about the fundamental meaning of life itself. Flogging a man until he converts to a selected religion could be described as an attempt to ‘encourage’ him to select the proper means – the true religion – to his desired end: everlasting, abundant life. But it seems clear in the case of forced conversion that what the concrete, empirical person before us wants has been disregarded. The agency of that empirical person has been despised. Something similar can be said, I think, for the new paternalists, albeit at what seems to be a much more mundane scale. What the consumer claims he wants before the new paternalist arrives is disregarded in favour of a posited deeper set of desires that more truly represent his interests (Infante et al., 2016): you can't really want to smoke given that you would like to live a long life. The old paternalist argues that the person is deceived by occult forces and must be coerced into the favoured rituals and incantations; the new paternalist argues the person is deceived by psychological shocks and must be protected through regulations. There is a parallel, albeit an imperfect one.

In a recent work, New Paternalism Meets Older Wisdom (Matson, 2024), I take up philosophical and practical challenges facing paternalistic regulations, challenges that draw on the thought of Adam Smith and David Hume. The goal of the work is to complement recent critical analyses (e.g. Rizzo & Whitman, 2020; Sugden, 2018) of the new forms of paternalism that have been inspired by developments in the behavioural sciences. In my view, Smith and Hume continue to provide us with insights important to the case for liberty in the contemporary world; their ideas provide us with resources to reflect on contemporary tendencies towards paternalism.

We can continue to learn from the wisdom of the past and it can and should inform our present discussions (for some of my reflections on this, see Matson, 2023). Not everything of use in the history of political economy has found its way into the contemporary practice of the discipline. In some cases, it is worth retracing our steps and considering paths not taken. In the context of behavioural economics, especially in the realm of normative discussions, tracing the footpaths back to the Scottish Enlightenment is especially worthwhile, for Smith and Hume were plainly aware of many aspects of human behaviour emphasised in today's behavioural economics. Richard Thaler said as much in his 2016 presidential address to the American Economic Association: “George Stigler was fond of saying that there was nothing new in economics, as it had all been said by Adam Smith. It turns out that was true for behavioral economics as well” (Thaler, 2016, p. 1578; see also Ashraf et al., 2005; Khalil, 2010; Palacious-Huerta, 2003). But although Smith and Hume understood that we often exhibit time-inconsistent preferences, fail to properly assess risks, inappropriately privilege present over future consumption, and are affected by the preferences of those around us, they did not leap to the conclusion that we are hopelessly irrational and in need of saving by judicious government intervention. Their ideas, I argue, marshal against such intervention, even though they clearly appreciated the same kinds of behavioural patterns cited by paternalism in our own day.

It is true that drawing Smith and Hume into present policy debates can be anachronistic. But their philosophical formulations – especially their reflections on the nature of welfare and its political implications – are highly relevant, and it is here that we can use their ideas to argue that many of the new paternalistic interventions are wrongheaded.

The most fundamental question that welfare economics, whether neoclassical or behavioural, must answer is: what do people want? Economists have traditionally answered, ‘pleasure’, and they have conceived of pleasure as subjective. Nineteenth-century political economists, including Jeremy Bentham, Francis Edgeworth and Stanley Jevons, sought to measure subjective pleasures or utilities directly – this involved contact with hedonic psychology (e.g. Colander, 2007). In the twentieth century, economists looked with scepticism on such efforts and attempted to de-psychologise their discipline. The American economist Irving Fisher wrote that the “foisting of Psychology on Economics seems to me inappropriate and vicious” (Fisher, 1892, p. 5). By the middle of the twentieth century, economists had, for the most part, agreed to conceive of welfare simply in terms of the satisfaction of the preferences revealed in action.

A challenge to the asserted connection between the satisfaction of revealed preferences and welfare emerged when it was realised that individuals' revealed preferences sometimes conflict. This violates the strictures of the axioms of economists' conception of rationality, associated with the theoretical work of John von Neumann and Oskar Morgenstern (2004 [1944]). This conception of rationality, developed for modelling purposes (Rizzo & Whitman, 2020, p. 53), essentially maintains that a rational individual must have complete and consistent preferences.

Upon brief reflection, it is clear that no individual is rational in the von Neumann–Morgenstern sense – in many situations, our preferences are inchoate, we very often reverse or at least change our minds and desired courses of action, and we do not always calibrate our choices properly given our risk preferences and long-term goals. It was these sorts of insights that motivated the first waves of behavioural economics, and then the paternalistic policy paradigm that followed. The goal of the new paternalists, as expressed memorably in the first edition of Thaler and Sunstein's Nudge, is to prevent individuals from making “decisions they would not [make] if they … paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control” (Thaler & Sunstein, 2009, p. 5). The programme of behavioural welfare economics, which then became closely allied with the new paternalism, was to nudge individuals in line with the supposed rational internal economic agent dwelling within (Infante et al., 2016).

Smith and Hume are relevant in this context in large part because they provide different answers to the question ‘what do people want?’, and those answers challenge paternalistic interventions, especially through the political process. Humans desire not just to satisfy some fixed set of preferences but to discover and satisfy the right kinds of preferences. Thus, we should expect people to exhibit seemingly inconsistent revealed preferences as they experiment and work towards the good as they understand it. Welfare is, we might say, a discovery process (Rizzo & Whitman, 2018). Even if we hold to an objective theory of the good, we need to discover how that objective theory applies in different circumstances, and what kinds of trade-offs we are willing to accept. Even if we were to agree that welfare is only about longevity (which it clearly is not), there are trade-offs and uncertainties – should I spend more money on a safer car and less money on wholesome food? If we respect the fact that individuals might pursue differing notions of the good, we must accept that our ability to identify and correct errors through political regulation is extremely limited.

These points can be developed by way of Adam Smith's formulation of self-judgment. Smith describes, in his Theory of Moral Sentiments (1982 [1790]), self-judgment as involving an artificial separation of the person into an acting and a judging self, and he presumptively cedes moral authority to the judging self – the ‘man within’. It might initially seem as if this formulation is amenable to the new paternalism. The goal, it might be claimed, is simply to help each individual's man within command the acting self towards the individual's overall conception of the good. The problem, however, is that the man within is a judge, not a rule. He is a spirit, a personality. He deliberates in different situations about what the proper course of action is. We can well imagine the man within approving of what appears to be inconsistent actions for contextual reasons. Thus, the inference from an observation of inconsistency – a professed desire not to smoke followed by a choice to smoke, say – to irrationality or impotence may not always be warranted.

A second problem from a Smithian point of view emerges when we realise that the man within is not simply deliberating about how his values cash out in specific circumstances, but about what his values actually are or should be. As the acting self converses with the man within about the proper course of action, so too does the man within converse with the Impartial Spectator (a god-like figure or, for the religious person, God himself) about what kind of life he should be steering the acting self towards. The point is to draw out the fact that life involves deliberations not just about how to achieve a set of goals but about what goals are actually worth setting and pursuing. The implication is that, if we claim to be liberals with respect to our notions of the good, we must allow for experiments in living, to borrow a phrase from John Stuart Mill. And even if we believe in an objective moral good, it still stands that the way individuals realise that good – how it manifests in their life circumstances – may differ. This too poses challenges for regulating people to better meet their own goals.

Hume's ideas complement these discussions in that he emphasises the satisfaction that comes from active life and the pursuit of new and better preferences. Here his thought resembles aspects of ‘Old Chicago’ welfare economics, especially the reflections of Frank Knight and James Buchanan (Dold, 2018; Dold & Rizzo, 2021). An implication of Humean thought is that we ought to respect and favour opportunities for individual dynamism and growth, given the inherent satisfaction that flows from action per se and the wide uncertainty of the kind of person we might yet become (cf. Sugden, 2018).

The behavioural scientist has valuable things to contribute to conversations about the good life, but she must not claim to hold privileged knowledge about what is in each individual's interest. Philosophers “confine too much their principles, and make no account of that vast variety, which nature has so much affected in all her operations,” Hume complained (1994, p. 159). The same complaint could be lodged against behavioural scientists and paternalistic policymakers today. The answer to the question ‘what does the good life look life?’ for Hume is chatter.

As we reflect on Smith and Hume's formulations, we come to a conclusion that has been reached by many critics of the new paternalism. That conclusion is that new paternalist regulations are not – and cannot possibly be – a simple matter of helping individuals achieve their subjective goals. Discerning what individuals want is no easy matter, even for the individuals themselves. New paternalist policy regulations are simply efforts to nudge, budge, or shove citizens to live in accordance with a folk theory of the good (Hausman, 2018) that the regulators themselves hold (Rizzo & Whitman, 2020, p. 400). Regulators have rarely applied their behavioural findings and claims of irrationality to themselves (Berggren, 2012; Rizzo & Whitman, 2020, pp. 329–47).

There certainly exists an informal, folk consensus that we should eat less sugar, smoke less, save more, and so forth. But should we steer people towards those ends, even if they might not want to be steered? Perhaps there are some cases where such paternalism is warranted, but we ought to tread very carefully here. We would do well in these matters to cleave to the presumption of liberty, the presumption of voluntary association, and the presumption against coercion afforded us by the writings of David Hume and Adam Smith.