The “Problem” Is Different and So Is the “Solution” Open Access

There is a statement at the end of Mario Rizzo and Glen Whitman’s commendable book that captures rather well their argument against those they call the “behavioral paternalists.”

The evidence in question is the behavioral research on individual decision making. People do not reveal consistent preferences and we frequently ignore Reverend Bayes’s advice when updating our beliefs. Such is life, suggest Rizzo and Whitman (R&W). Not so, say the “behavioral paternalists.” The “behavioral paternalists” hold, instead, either that we ought to be, or indeed are, but for psychological deflections of one kind or another, individuals who have well-behaved preferences and who update beliefs Bayesian-style. Their “solution” to this “problem” is to “nudge” us back to what we either ought to have done or what our inner, but currently dozy, rational choice selves would have done. Thus, they are doing us a favor when they wield their behavioral insights and rather resent the title of paternalist. Indeed, they are at pains to explain why they are not paternalist at all because their interventions will not influence the rational choice agents that we either are at heart or ought to be.¹

R&W will have none of this. There is neither evidence that we are inner rational choice agents of this consistent and Bayes loving-kind, nor are there any good normative arguments for why we should be like that. Indeed, most of the behavioral research is, R&W suggest, a testament to a perfectly plausible inclusive kind of ecologically adapted rationality. We’re simply discovering preferences, adjusting to what we experience using rules and heuristics that are mostly well adapted to the complexity of the environment and the costs of calculation. In short, we’re fine as we are. There is no “problem”. Quite the contrary, we should just be left alone. Thus, when “nudgers” decide we ought to behave in some way different to this, that’s paternalism and it arises from a false diagnosis that our behavior is in some way a “problem.” It’s not.

I find the R&W argument against the “nudgers” with respect to their paternalism entirely persuasive on these points. R&W set their case out in detail and cover almost all the imaginable re-joinders. After reading this, it would be difficult to think that “nudgers” were not both paternalist and mistaken in their diagnosis. I am not persuaded, however, by the R&W “solution.” I have one criticism; and this leads me to propose a different “solution.”

My criticism is that this argument by-passes ordinary politics, where policy efficacy rather than paternalism is what matters. The real “problem” is that, for this reason, “nudging” is attractive because it appears a cheap way for politicians to achieve their objective. R&W acknowledge this possibility and their “solution” to this larger difficulty, I shall argue is not persuasive – at least in so far as it is forced to rely, as it appears to in R&W’s argument, on J. S. Mill. I develop this criticism in the next section. I sketch an alternative “solution” to this more general “problem” in Section 3. My “solution” is to argue that politicians need persuading that the task of politics is to select rules and not engineer specific outcomes for particular individuals. The latter is what “nudging” does. Indeed, it is an important claim of the proponents of “nudging” that “nudging” is not a change in the rules in this sense –otherwise rational choice agents would also be potentially influenced by “nudges” and the claim is that they are not. Thus, if politics is about rule selection, “nudging” has no place in politics.

In my experience, policy makers do not deeply care about what troubles R&W. That is, whether “nudges” are a form of paternalism that arises from the diagnosis of what is a phantom “problem”. Policy makers seem, instead, to be centrally interested in any tool that might enable them to achieve their objectives effectively and cheaply. Arguments around libertarian or asymmetric paternalism, in my experience, had I raised them in front of a policy maker would have produced, I suspect, at best a wry and indulgent smile, or at worst some eye-rolling directed at the ivory tower.

Policy makers do not seem to think much about whether what they are doing is paternalist. They typically have ideas about what needs to be done (perhaps for the good of the nation, the good of their supporters or just the good of themselves; or some measure of all three), that may or may not have been well-aired in an election, and they set about doing them. Of course, it is never simple because, as the UK Prime minister Harold Macmillan famously remarked, there is the “opposition of events”. Unforeseen challenges arise in office and the politicians adjust their objectives accordingly, but they still want tools to deliver on those evolving objectives. The attraction of “nudges”, which explains their popularity, is that they appear to be a very cheap way to alter people’s behavior and achieve a politician’s objectives. For most politicians, therefore, an argument that “nudges” are paternalistic misses the point. Politicians do not embrace “nudging” because they think it is not paternalistic. Nudges are popular because they are perceived as cheap and effective ways for politicians to achieve their objectives.

R&W acknowledge that their argument is in this sense tangential to the practice of politics.

Their riposte to this toolbox view is to draw attention to other challenges that have been discussed in the book with respect to “nudging”: the “deficiencies of the underlying research, lack of relevant knowledge, unintended consequences and political risks’ (RW, 429).

For the most part, these are cautions that apply to any kind of government intervention (e.g., unintended consequences) and so, if they have not checked government intervention before now, they seem unlikely suddenly to halt “nudging” in its tracks. There is one, though, that specifically applies to nudging: the “deficiencies of the underlying research.” But in so far as there are deficiencies in the underlying behavioral research, then a politician is unlikely to take these on trust (for exactly the reasons of inclusive rationality that R&W suggest guide decision making). Instead, politicians are inclined to “suck it and see.” Any mistake on this score will be self-correcting. If it turns out that the nudging bottle has been diluted with snake oil, then politicians will soon realize this is the case. But I doubt, as I write, that such reflection will cause some Damascene moment with respect to the uses of nudges. This is because I judge that politicians more often see “nudging” successes than failures. For example, there are notable successes with interventions like changes to the default in decision problems (e.g., in relation to pensions); and in so far as there are failures, they did not cost much if anything. Thus I cannot see much in this argument to prevent politicians embracing “nudging.” Something more is needed.

As a prelude to developing what that something more might be, it is helpful to set out the “political risks” that R&W refer to above to buttress their position. The political risks arise for them because one intervention of this kind tends to trigger another and so on until the slippery slope has delivered a collapse of the space in which people are free to do as they please. Their “solution,” in other words, to this bigger problem does not simply turn on the observation that the behavioral evidence suggests that “we’re fine.” It is based on a slippery slope injunction against state intervention per se. In this way, it is not unfair to see R&W’s book as belonging to a restrain-Leviathan tradition in political theory. It just happens to focus, and quite naturally but not exclusively, on the latest case of an advancing Leviathan: “nudging.”

To illustrate this anti-Leviathan anchoring of the R&W “solution,” consider what they say about fiscal externalities. This is revealing because, as they correctly say, the fiscal externalities argument for intervention has nothing to do with paternalism. It is an intervention that is undertaken to correct for third party spillovers from individual action (i.e., externalities). For this reason, this particular line argument is more like a baring of the authors’ souls because it does not impinge directly on their dispute with “nudging” and behavioral paternalism.

Their objection to this kind of intervention is that “it allows the state to create circumstances that then justify further interventions – specifically, interventions that tend to infringe on personal choice (RW, 431).” Their illustrations are various: cigarette smoking is constrained or helmet wearing is required because there are spillover health costs for third parties from the dangerous activities of smoking and riding without a helmet that the smoker and riders will not take into account. Their argument is that these spillovers can be traced to an earlier state intervention: socialized healthcare. In this way, the logic of a slippery slope is illustrated. One intervention (socialized healthcare) creates the ground for another (bans and taxes on smoking) and so on (wearing helmets when riding).

If this argument was robust, then the R&W anti-Leviathan anchored “solution” would make perfect sense. But is it?

R&W say that although they have much admiration for Mill’s defense of liberty, their case against paternalism plays into their comparative advantage as economists: it is, as a result, conceptual and consequentialist (e.g., see p. 23). This misses a trick, I suspect, because Mill supplies a powerful argument in his commendation of “experiments in living” against nudging. People have to be able to make mistakes, otherwise they cannot learn who or what they want to be as individuals. However, this is not the key point about their suggestion. Its claim to be consequentialist does not sound right. A consequential argument would surely allow intervention in genuine cases of externalities. One can argue about whether a particular instance is a genuine case, but their chosen example of socialized health care creating the illustrative slippery slope looks, nevertheless, like a classic instance of externalities. Public health is, as it were, something close to a public good. It’s difficult to escape, for instance, the effect of someone else’s TB, cholera, or COVID-19. We all benefit from each other’s health. So, saying no to the advancing Leviathan in such genuine cases cannot be on consequentialist grounds – at least if the consequences that matter in such cases are people’s health. If there are more such genuine externalities that flow from this one, then the logic of consequentialism is that the Leviathan should advance. Their case against the advancing Leviathan has, therefore, to be non-consequentialist. The only source for this that they seem to offer is indeed Mill: Mill is, they suggest, one example who makes “a straightforward appeal to freedom and autonomy” (RW, 23).

It is of course true that Mill is taken to value freedom intrinsically. Individual liberty is what allows individuals to be become individuals in his argument in On Liberty; and in this sense it is an intrinsic justification of liberty rather than an instrumental one. The difficulty with using Mill, however, to defend freedom in such circumstances is that his famous no-harm principle exactly allows constraints on liberty when there are negative spillovers from the untrammeled exercise of individual freedom. You are free to do as you please but only in so far as your actions do not cause harm to others. In practice, in liberal democratic societies, the devil has always been in the detail of deciding what counts as a harm for this purpose. However, I don’t think anyone doubts that if I give you COVID-19, I am harming you and therefore the “no-harm” principle, prima face, sanctions restrictions on my liberty in this regard.

In short, I cannot see the argument in R&W that is persuasive over why we should halt intervention because it spawns a slippery slope. Provided, of course, the externalities are genuine and significant at each stage, there is no slippery slope as such. There are grounds for intervention that are either met or not and these define the limits of intervention. There is nothing in Mill’s On Liberty, in other words, to counsel against intervention per se, on grounds of a slippery slope: quite the reverse. This is why I am not persuaded that R&W’s “solution,” with its anchor in an anti-Leviathan tradition, is the right one.

To summarize, we have a “problem.” Politicians do not materially care whether an intervention is paternalistic, so R&W’s demonstration that nudging is paternalistic passes most by. Furthermore, R&W’s other injunction against “nudging” and intervening more generally, which is “just don’t do it in the name of freedom”, is not persuasive – at least if JS Mill is its support. So, what is going to stop the typical politician “nudging”’ here, there and everywhere? In short, what is the shape of an alternative “solution” for diverting politicians away from “nudging” if it is not a wholesale shackling of the State that R&W seem to draw upon?

My argument, that I develop in the next section, is that you cannot deny there are reasons for state intervention because there are: spillovers are a powerful case in point. Indeed, if anything such spillovers seem likely to grow with the complexity of social and economic life. My “solution,” therefore, is not to deny intervention. Rather it should be embraced. What we need is a different kind of state intervention. We need intervention that is directed at the rules and not at achieving particular outcomes for specific individuals. We need to re-anchor politics in a constitutional, rule changing domain.

In short, politicians need to be persuaded to think differently about their objectives: the business of politics ought to be the choice of rules and not the achievement of specific outcomes for particular individuals.

There are two starting points for my alternative “solution.” The first is an interpretation of the behavioral research that follows that of R&W’s but also, importantly, adds something more. I follow R&W by taking much of the evidence of decision making under uncertainty, that is where individual’s make decisions in an environment defined by nature, to illustrate that people rarely have well defined preferences. They are mostly bumbling along, collecting rudimentary evidence and then correcting actions, beliefs and preferences every now and then when they plainly make a mistake and have time for reflection. However, this is not the only insight from behavioral research. The other comes from interactive decision making, where the decision environment is partially made up by other people’s decisions. Here, it is plain that our selfish instincts are cut by a concern to do what also seems right in that decision problem – and notably R&W appear less interested in this aspect of behavioral research in their book. Of course, there is great variety in what people reveal they believe to be “right.” But, I focus below and make use of what we know from the behavioral research about what is perceived as “just” in dictator and distributional games.

My second starting point, into which I shall feed the first, is a broad taxonomy of what politics does in liberal democratic societies. This is the hook for engaging politicians over how they should think about their interventions. For good reasons, as I have suggested, politicians will want to intervene, and my strategy is to argue they need to conceive of their interventions in such a way that it precludes “nudging.”

For this purpose, I identify two broad kinds of activities in politics. One relates to deciding on collective actions where this yields improvements for everyone and where these gains seem worth trading-off for the constraints on individual liberty that are entailed by taking collective action. This is where politicians sniff-out Hobbesian type improvements: people accept the constraint of collective rules because they all see merit in so doing. It is tempting to say, using Berlin’s (1958) terminology that they give up some liberty in the negative sense (they accept constraints on individual action) in return for a boost to liberty in the positive sense (the capacity to do things).

There are several things to note about such Hobbes-like interventions. First and most important, they relate to the rules that constrain action and not the determination of one specific action rather than another from within a given set of possible actions. This is obviously the case for Hobbesian, law and order-type collective actions but it is also the case for a tax on a commodity that has negative externalities: the tax changes the available set of actions that every individual in society faces, just like a change in the law. In this sense, these interventions are constitutional. In contrast, and this is the lynchpin of my argument, “nudging” is not constitutional in this sense. “Nudging” does not change the constraints on action that people face (or such is the argument of the “behavioral paternalists” because, otherwise, rational choice agents would also be affected by the “nudges”). Instead, “nudging” is designed to change what some individuals select from a given set of possible actions.

Second, to engage in such constitutional interventions, there has to be some way of identifying whether such a rule change will yield improvements for all. This is a difficult task, but it need not depend on the rational choice model and the assumption that people have well-behaved preferences. I will take up this last point first because it connects with R&W’s interpretation of the behavioral evidence which I endorse: that is, we are not well described by the rational choice model of preference satisfaction. The point here is that while economists typically use cost-benefit analysis because it establishes potential Pareto improvements and so depends implicitly on the rational choice model because the Pareto criterion is framed in terms of preference satisfaction, there is no need to hold the rational choice model to be able to use cost-benefit analysis. Without any assumption about people having preferences, cost-benefit analysis more weakly, and so also more generally, identifies interventions where the aggregate budget set of a society/community improves. This is because, at least in theory, the relative prices used in costbenefit analysis identify the marginal rate of transformation in production between goods. It is only the further connection of prices to the marginal rate of substitution in consumption that relies on the rational choice model to deliver the interpretation of a potential pareto improvement. The shift to the budget set interpretation of cost-benefit analysis provides a more general criterion for improvement because it does not depend on the further assumption of the rational choice model that people have well defined preferences. It is also more appropriate when we have no reason to suppose, from behavioral research as R&W suggest, that people typically behave “as if” they were preference satisfiers. The difficulty, that I now return to, is that rule changes which satisfy the cost-benefit net benefit test to produce an improvement in the aggregate budget set, will often in practice create individual winners and losers.

It may be possible to identify who are the winners and loser in the short run but even this may be difficult and tracking winners and losers in the longer run is near impossible for reasons that Hayek and the Austrian school have made clear. This means that it is impractical to set up precise compensation schemes to secure the condition that everyone actually benefits from such rule changes. The condition can only come close to being satisfied in practice by the introduction of background institutions that limit the downside for all individuals. Thus, for example, individuals in liberal democratic societies have collectively agreed to be constrained by the rules of a market society with private property rights (or strictly, they find these rules legitimate) because these rules, when accompanied by the institutions of the modern welfare state, plausibly secure better living standards for all than would otherwise be the case. This is the argument Sugden (2018) has recently made. The market order is based on mutual benefit and so this guarantees that the parties to a market exchange will always be better-off through any action they take. However, the fact that the parties to the exchange benefit does not guarantee a general improvement in the presence of third-party negative spillover effects. In practice, one can hope that any such negative third-party spillover effects tend to cancel out, but whether they do is ultimately an empirical question and, when they don’t, the market order can only be defended in these terms because it has developed the institutions of a welfare state to ensure, so to speak that when the tide rises, it does indeed raise all the boats.

To summarize, the first broad activity of liberal democratic politics is to engage in rule changes that plausibly secure improvements for all. The second broad function of politics is to resolve distributional indeterminacies. The latter is inevitably drawn into the former and it is on the latter that I want to draw the behavioral research on distributional decisions into my argument.

It is well known from dictator games that while many people decide selfishly when acting as the dictator, most people give something to the person they have been paired with. There is a tradition in economics that has interpreted such pro-social decisions as being motivated by an altruistic concern for the welfare of the other person. It is a natural extension or application of the utilitarian or social welfare function approach in welfare economics: it is “as if” the individual had internalized some version of utilitarian ethics and taken a weighted average of their own and the other person’s welfare when make the dictator’s decision. If this model of pro-sociality explained behavior well in dictator games, then it would lend general support to politicians when acting in this distributional domain of politics to be guided by how any intervention affects people’s welfare. However, the evidence on dictator giving is difficult to reconcile with this altruistic interpretation of why we give.

Since Andreoni (1989) there has been a large experimental literature that tests for this kind of pure altruism. A common finding is that people are instead impurely altruistic as one donation is only imperfectly crowded-out by another – whereas pure altruism/utilitarianism predicts full crowding-out because what matters to a pure altruist/utilitarian is the other person’s welfare/pay-offs and not their origin (e.g., see Engel, 2011). It seems people in labs give for other reasons, like a “warm glow” effect. One possible way of understanding how such a “warm glow” effect might vary across dictator-like decisions and so influence the Dictator’s gift is to posit a social preference for equity. In particular an aversive preference towards inequality has been formalized in a similarly consequential manner to that of altruism: with the difference that inequality aversion focuses on the gap in pay-offs between the dictator and the recipient and not some addition of the two (see Fehr and Schmidt, 1999). This fits the data in such distribution decisions better than simple forms of altruism/utilitarianism (e.g., see Charness and Rabin, 2002). But, it begs a question as to why people value equity in this sense. Furthermore, whatever the “warm glow” from acting in support of equity, while this particular formulation of inequality aversion may do better than pure altruism, it does not explain other aspects of the data on giving (e.g., Engelmann and Strobel, 2004; Chowdhury and Jeon, 2014).

One possible explanation of why a consequential theory like inequality aversion still explains giving in the lab imperfectly is that what motivates the dictator to give is a sense of fairness or justice and this will typically take account of how the initial endowments of the dictator and the recipient arose in a way that the particular theory of inequality aversion does not. In other words, to answer the earlier question about why people might value equity, it is because they have a preference for fairness and the theory of inequality aversion does not capture the procedural aspect of fairness because it focuses only on the outcomes and not how they arose. There is considerable evidence that the decisions in dictator and other distribution games are affected by the procedure in this sense that is responsible for generating the differences in endowment in these experiments. Thus, for example, Eckel et al. (2005) find that the dictators decisions vary depending on whether they understand that endowments arose from an early tax and redistribution decision. Likewise, many people are known to subscribe to a version of fairness where merit matters and Hoffman et al. (1999) find that when a dictator earns the endowment over which they then decide what to give their co-player, they give less. We also know from several experiments that people are more redistributive when outcomes arise from luck than individual decisions in the past (see Cappelen et al., 2013). This is important because it suggests that the relative success of inequality aversion in explaining dictator decisions with reference to the outcomes alone may be because the endowments in these experiments are largely the result of luck: that is, subjects typically arrive in the lab and are randomly assigned to the role of dictator who makes decisions with a variety of possible endowment configurations and one of these is randomly selected for implementation.

I draw the conclusion from this brief run through the experimental evidence on distribution decisions that people are plausibly motivated by a sense of fairness and this makes these decisions sensitive to the procedures that have generated outcomes:that is, the background rules of the game under which distributions arose and not simply the actual end-state outcomes themselves. Or to put this in a slightly different way, it is predominantly the character of the rules under which outcomes arise and not the outcomes per se that people care about in dictator games. If this is right, then the object of political choice in matters of distribution should again be the rules and not the specific outcomes.

Of course, there is also a Hayekian reason again why distribution decisions ought to be made with respect to rules and not outcomes. This is a replay of his argument that I made earlier. A particular intervention may achieve some desired distribution of income across particular individuals, but the actions of individuals will quickly undo that desirable distribution in ways that cannot be readily anticipated and taken into account through a more complicated form of intervention that triggers further compensating changes as these conditions evolve. Given the uncertainties about how an economy will evolve, any attempt to achieve a particular distribution of income for particular groups of people is a fool’s errand. This is, of course, also the argument made by Nozick (1974) about the mistake in being guided consequentially to achieve what he calls patterned outcomes. The best that can be sensibly discussed and done is to change the character of the rules.

Of course, it would be foolish to claim that the behavioral evidence on pro-sociality points only in the direction of people being concerned about the character of the rules generating outcomes and not the outcomes themselves. There is too much evidence, for example from the in-group bias in pro-sociality, that we have stronger fellow feeling for individuals that are closer to us than others who are more socially distant (e.g., see Hargreaves Heap and Zizzo, 2009). We are motivated, in other words, self-evidently in these cases by how an outcome affects a particular individual and not just the rules under which the outcome arose. My argument is different. It is that such motives should not be in play in politics. Personal considerations in the form of specific feelings of altruism or fellow feeling for particular people belong outside politics. They belong in the weft of social life.

In contrast, in politics, the two key activities are sniffing-out Hobbesian opportunities for collective action and resolving distributional indeterminacies and both can only and should only sensibly focus on the selection of the appropriate rules. This is a practical consideration as far as the first Hobbesian activity is concerned that arises because of Hayek’s knowledge problem. With distributional questions it is the same point but bolstered by the behavioral evidence on individual decision making as this points, at least significantly, to people thinking in terms of rules of fairness. The deep connection between distribution and rules of fairness is also found the tradition in political theory that associates justice with the choice of rules because impartiality is constitutive of justice (e.g, see Rawls, 1971). I have not focused on the latter here only because my argument has naturally been tethered to the behavioral evidence.

Thus, and in short to bring this argument to its conclusion, “nudging” is directed at changing the welfare of particular individuals and it specifically does not involve a change in the rules for rational choice individuals. In contrast, politics’s twin activities both essentially should entail choosing rules and not specific outcomes for particular individuals. Therefore, “nudging” has no place in politics.

R&W have mounted a splendid attack, using behavioral evidence, on the academic premise underpinning “nudging”: that we are or ought to be rational choice agents. My concern is that their argument will not persuade politicians to stop “nudging.” For the politician who is not deeply moved by whether or not their interventions are paternalist, something more is needed. R&W recognize this weakness and largely buttress their anti-“nudging” position by presenting a slippery-slope argument for being predisposed against any advance in Leviathan, including “nudging.” I have suggested this argument does not work – at least if J. S. Mill is its authority. I offer instead an argument that politics is about the selection of rules and not outcomes for specific individuals. “Nudging” does precisely the latter and not the former. It personalises politics. That “nudging” is not about a change in the rules can be readily appreciated because the proponents of “nudging” have made clear these interventions should not affect the decisions of rational choice agents. If “nudging” was a form of rule change then it must, in principle, affect rational choice agents. ERGO: “nudging” does not belong in politics.

In other words, if the practical question is how to persuade politicians to stop “nudging,” then my “solution” is to remind politicians what politics is about. It should be concerned with the choice of rules that both enable and constrain individual action. It should not be targeted personally and consequentially at changing welfare of particular individuals with tweaks to their choice architecture.

I conclude with an illustration of how this approach to politics might have implications beyond the “nudging” debate. I have argued earlier that we should expect the rules chosen by politicians in a liberal democratic society to involve a market order buttressed by the institutions of a welfare state that put a cap on the downside risks of living in a market society. The latter ensures, given the uncertainty over how third-party spillovers will evolve that, so to speak, as the tide rise, so do all the boats. This is why, in part, I suggest an anti-Leviathan general argument against any further intervention, “nudging” included, won’t work. The question I wish to focus on now, briefly, is the character of the welfare state.

The welfare state has always been torn between universal and targeted forms of benefit and the targeted side of this debate has until recently had the upper hand in this argument. The central reason behind the move towards targeting is that it delivers scarce tax-payers money to where it is most needed. This is an interesting argument because it is consequential and clearly relates to the needs of specific individuals. In other words, these targeted parts of the welfare state look less and less like a change in the rules and more and more like “nudging” in the sense that the aim is change the outcomes for specific individuals. Of course, any set of rules can become so nuanced in their application, depending on the conditions, that they become in effect a scheme for affecting the outcomes for particular individuals and so the distinction between rules/procedures and consequences/outcomes can become moot. This is the point. As the welfare state has shifted towards targeting it has become, in effect, more and more like a “nudge” type, person specific, intervention and less and less like a change in the rules that affects everyone’s decision making.

The alternative rule like version of a welfare state is to have universal forms of benefit. Universal benefits are rule-like in their application to everyone and they provide the background against which everyone makes decisions. The NHS in the UK is one such example and so were the original national insurance in the UK and social insurance schemes elsewhere that were designed to fund this and other universal benefits. While the NHS has survived, there has been a proliferation of other means tested benefits (e.g., housing benefits and income support) which are entirely funded from general taxation. The alternative, universal rule change that would substitute for this proliferation of targeting and avoid this drift to personalizing benefits is a Universal Basic Income (UBI).

I mention the connection with the current debate over UBI and targeting of benefits because it points to the way that the problematic personalization of politics I detect in the practice of “nudging” is not simply confined to “nudging.” This makes my approach to tackling “nudging” more difficult because it becomes part of what is a bigger problem of persuading politicians to see politics as primarily an activity in rule selection and not one of helping specific individuals. The upside, however, is that it means shifting this perspective on what politics is about is also more important than simply defeating “nudging.”