Evaluating the costs and benefits of standard government policies is straightforward, when we assume that citizens are rational and well-informed. Choices are guided by consumer and firm optimisation of clear-cut objective functions. Policy evaluation is based on how the policy changes the value of those objective functions. Assessing ‘nudges’ is a very different thing.
Nudges are modifications of people’s choice architecture (CA) that impact their behaviour without changing their material incentive structure or in any way coercing them (Thaler and Sunstein 2008).1 There is abundant evidence that they ‘work’ (Costa and Kahn 2010), but are they welfare improving?2
In fact, evaluating the costs and benefits of nudges entails some unavoidable ethical considerations. This column raises a number of issues that are widely neglected by both proponents and critics of nudging.3
I suggest four steps when it comes to evaluating a nudge policy.
- First, let’s see whether the nudges in question increase people’s wellbeing.
Unfortunately, already here we get into deep water.
It’s unclear how we should think about wellbeing in the ‘behavioural world’ – which is, in fact, the real world, and the world in which nudges work. Remember, in the behavioural world, people not only have limited mental resources – meaning computational capacities, willpower and attention – but also context-dependent, inconsistent, and incomplete preferences. As a consequence, the standard neoclassical notion that defines wellbeing as the technical degree of satisfaction of given and consistent preferences cannot be applied.
A key point is that there is no received-wisdom alternative. Ideas have been floated – from measurable happiness to ‘only perfectly informed preferences should count’, all the way to Bob Sugden’s ‘opportunity’ criterion (probably the most elaborate alternative concept at the moment).4 The jury, however, is basically still out on how to think about wellbeing in a behavioural world.5
Let’s leave this core question aside for the moment and push on. As Chetty (2015) puts it, nudges may still be ‘pragmatically’ useful (in concert with more traditional regulation) in achieving specific policy goals that citizens have somehow agreed upon beforehand.
- Second, we have to ask how nudges affect people’s autonomy.
Most critics agree that nudges compromise this key value by interfering with and manipulating people’s preference formation, and by addressing people’s lower instincts instead of reason. Individuals are then argued to lose ‘control’ over their own preferences (Hausman and Welch 2010).
Upon closer inspection, this argument looks a bit strange. Does autonomy really depend on the kind of hyper-rationality presupposed here? Aren’t we all subject to myriads of influences on a daily basis, most of which we’re even unaware of?6 Do we really lose our autonomy – and potentially our moral accountability with it – when acting thoughtlessly or in a way that is contrary to our sincerely held moral values (Buss 2012)? Suffice to say that whoever takes issue with nudges along these lines faces difficult conceptual and ethical questions (what is ‘manipulation’ anyway?).
- The third step invites you to check whether it’s maybe not autonomy after all, but rather people’s integrity that’s at stake in nudging?
After all, nudges are supposed to work in a setting where people haven’t yet made up their mind (consider the notorious cafeteria case); that is, they lack complete preferences. It may be a good idea, then, to have a closer look at the problem of preference formation, which, for economists, is akin to the problem of identity or character formation.
The late James Buchanan had suggested that we should take seriously the notion that human beings face the task of creating their preferences and assume responsibility for them (Buchanan 1999). As Korsgaard (2009) shows, a necessary condition for succeeding in this ongoing process of ‘self-constitution’ is active choosing. Some kinds of nudges clearly support informed active choosing.7 Others rather seem to discourage people from engaging in active choice.
Put differently, some nudges may produce ‘excessive convenience’. Consider a world with widespread adoption of public nudging. There, I, the consumer, don’t need to worry about my retirement savings, nor about mustering the little self-control that I have to avoid the chocolate bars in the cafeteria, nor about being wary about the tricks of door-to-door salespersons. In all these cases, some choice architect, somewhere behind the scenes, subtly steers me into the ‘right’ direction – by changing defaults and frames, and by implementing cooling-off periods. In other words, I’m outsourcing my choices to some external body.
In parts of the critical literature, this specific variant of moral hazard makes an appearance as the ‘infantilisation effect’ of nudging (Bovens 2009, White 2013). Note what’s at stake here –when preferences depend on people’s context, and policies can (partly) change that context, we face the problem that policies can impact preferences, which would present us with the awkward task to ponder over which kinds of preferences we want to promote (Hargreaves Heap 2013). The Buchanan-Korsgaard focus on identity formation may help us bypass this question (which is impossible to answer), without losing the ability to identify normative costs associated with nudging.8
- Fourth and finally, we should think about what all this means in terms of practical policy implications.
Ideally, citizens should be informed about the normative costs involved in public nudging before voting on its implementation (Schubert 2014).
Consider integrity. People seem to face a trade-off between ‘excessive convenience’ on the one hand (which discourages active choosing, to the detriment of character formation), and ‘too little’ convenience on the other hand (leaving them overwhelmed with complex choices). This trade-off looks different for different kinds of goods. With primary goods that satisfy basic needs, we may conjecture that most people will favour delegating choices, at least partly, to trusted external bodies. Consider basic retirement savings. The demand to form idiosyncratic preferences on issues related to basic retirement savings seems rather limited. In contrast, preferences on morally charged issues such as whether to donate organs post mortem – a popular example of effective nudging (Smith et al 2013) – don’t easily generalise. In that latter case, integrity arguments speak against the use of nudges as a regulatory policy tool.9
In most cases, nudges are likely to be implemented as complements to more traditional incentive-based tools. Research on the interplay between modifications of different parts of people’s choice architecture is still in its early stages. What’s striking, though, is that economists interested in deriving behavioural policy implications apparently require much more ethical input than their neoclassical predecessors were accustomed to. The behavioural economist may turn out to be the moral philosopher’s best friend.
Akerlof, G A and R J Shiller (2015) Phishing for phools: The economics of manipulation and deception, Princeton, Princeton University Press.
Berg, N (2014) “The consistency and ecological rationality approaches to normative bounded rationality”, Journal of Economic Methodology, 21: 375-395.
Bovens, L (2009) “The ethics of nudge”, in Preference change: Approaches from philosophy, economics and psychology, T Grüne-Yanoff and S O Hansson (eds), 207-220, Berlin, Springer.
Buchanan, J M (1999) “Natural and artifactual man”, in The logical foundations of constitutional liberty, Vol. I, J M Buchanan, 246-259, Indianapolis, Liberty Fund.
Buss, S (2012) “Autonomous action: Self-determination in the passive mode”, Ethics, 122. 647-691.
Costa, D L and M E Kahn (2010) “Energy conservation ‘nudges’ and environmentalist ideology: Evidence from a randomized residential electricity field experiment”, VoxEU.org, 19 May.
Hansen, P G (2015) “The definition of nudge and libertarian paternalism – does the hand fit the glove?”, European Journal of Risk Regulation, forthcoming.
Hargreaves Heap, S (2013) “What is the meaning of behavioural economics?”, Cambridge Journal of Economics,37: 985-1000.
Hausman, D M and B Welsh (2010) “Debate: To nudge or not to nudge?”, Journal of Political Philosophy, 18: 123-136.
Korsgaard, C M (2009) Self-constitution – Agency, identity, and integrity, Oxford, Oxford University Press.
Loewenstein, G, C Bryce, D Hagmann and S Rajpal (2014) “You are about to be nudged”, Working Paper, http.//ssrn.com/abstract=2417383.
Reiss, J (2013) Philosophy of economics: A contemporary introduction, London, Routledge.
Rothenberg, J (1962) “Consumers’ sovereignty revisited and the hospitability of freedom of choice”, American Economic Review, Papers & Proceedings, 52: 269-283.
Schubert, C (2014) “Evolutionary economics and the case for a constitutional libertarian paternalism”, Journal of Evolutionary Economics, 24: 1107-1113.
Schubert, C (2015a) “Opportunity and preference learning”, Economics and Philosophy, 31: 275-295.
Schubert, C (2015b) “On the ethics of public nudging: Autonomy and agency”, Working paper, http.//papers.ssrn.com/sol3/papers.cfm?abstract_id=2672970.
Schubert, C (2016) “Green nudges: Do they work? Are they ethical?” Working Paper.
Smith, N C, D G Goldstein and E J Johnson (2013) Journal of Public Policy & Marketing 32: 159-172.
Sugden, R (2004) “The opportunity criterion, consumer sovereignty without the assumption of coherent preferences”, American Economic Review, 94: 1014-1033.
Sugden, R (2008) “Why incoherent preferences do not justify paternalism”, Constitutional Political Economy, 19: 226-248.
Sunstein, C R (2014a) Why nudge? The politics of libertarian paternalism, New Haven, Yale University Press.
Sunstein, C R (2014b) “Choosing not to choose”, Duke Law Journal, 64: 1-52.
Sunstein, C R (2015) “Nudging and choice architecture: Ethical considerations”, Working Paper (version 17 Jan 2015), http.//papers.ssrn.com/sol3/papers.cfm?abstract_id=2551264m, forthcoming, Yale Journal of Regulation.
Sunstein, C R and L A Reisch (2013) “Green by default”, Kyklos, 66: 398-402.
Thaler, R H and C R Sunstein (2003) “Liberterian paternalism”, American Economic Review, Papers & Proceedings, 93: 175-179.
Thaler, R H and C R Sunstein (2008) Nudge: Improving decisions about health, wealth and happiness, New Haven, Yale University Press.
White, M D (2013) The manipulation of choice: Ethics and libertarian paternalism, New York, Palgrave.
1 Nudges have been widely associated with the overarching normative programme of ‘Libertarian Paternalism’ (Thaler and Sunstein 2003), which restricts the set of legitimate CA modifications. They can however be applied in order to pursue non-paternalistic goals as well, such as protecting the environment (e.g. Sunstein and Reisch 2013, Schubert 2016). Nudges are supposed to be transparent, perhaps in the sense that an alert agent should be able to identify them and the channels through which they operate. That condition excludes, for instance, ‘subliminal advertising’ (Bovens 2009). Importantly, there is evidence that even perfectly transparent nudges can be highly effective (Loewenstein et al. 2014). As Sunstein (2014a: 13) puts it, the general idea is to develop “sensible, low-cost policies with close reference to how human beings actually think and behave”. Nudges have become very popular among practical policymakers, particularly in the US and the UK.
2 They manage to do so by either harnessing people’s cognitive biases or by responding to them (Hansen 2015).
3 See Schubert (2015b) for an elaboration.
4 See Sugden (2004, 2008) and Schubert (2015a) for a critical discussion.
5 This problem is closely related to the issue of conflicting understandings of rationality. While Thaler and Sunstein (2003, 2008) stick to the neoclassical variant (even elevating homo economicus to a normative role model!), others suggest the alternative notion of ‘ecological rationality’ (e.g., Berg 2014).
6 See.Reiss (2013: 299). It’s an open question whether competitive markets foster deceptive private commercial nudging; e.g. Akerlof and Shiller (2015).
7 Reminders and simplifications are obvious examples. To be sure, the whole nudge agenda raises awareness of the behavioural power of the choice architecture, which may promote informed choice overall. Note also that mandatory choice is a (non-nudge) element in the behavioural policymakers’ toolset (Sunstein 2014b).
8 Here’s a heretic thought I dare only express in a footnote – maybe it’s not people’s actual preferences that should be centre stage in normative economics, but rather their ongoing ability to cultivate them? Rothenberg (1962: 282-83) floated that idea long ago.
9 Welfare arguments may be weighted against integrity concerns here, but as we have seen, no one really knows what ‘welfare’ stands for in our behavioural world, at least as far as the level of the individual is concerned.