Waldron, Sunstein, and nudge paternalism

It is not my policy to comment on articles published in the New York Review of Books, but Jeremy Waldron had a piece a little while back – a discussion of Cass Sunstein’s book, Why Nudge? – that I feel impelled to respond to. That’s because the view Waldron puts forward, in criticizing Sunstein, is a precise articulation of exactly the view that I think we need to be getting away from. One of the major objectives of my own recent book, Enlightenment 2.0, was to explain why we need to stop thinking this way. (That’s actually the reason for the “2.0” in the book title. What Waldron is urging upon us is what I like to think of as the “Enlightenment 1.0” position.) And Sunstein’s response, in the most recent issue, is too tepid by far.

Sunstein, it may be recalled, is a proponent of “nudge” paternalism, based on the observation that the way choices are presented to people, although seemingly neutral from the standpoint of economic rationality, often actually favour one option over some other. There may be no tangible economic cost associated with choosing x over y, but if getting x requires filling out a form, while y is the default, then the architecture of the choice is one that favours y. Even though the effort it takes to fill out a form may seem trivial, the psychological cost at the point of decision may be quite large, and so individuals may put it off, or ignore it entirely, and so wind up with y even if they would much prefer x.

This is something that companies have known about for a very long time, and seldom miss an opportunity to exploit. (To take just the most transparent example I have encountered, the Playstation 4 that I recently bought came with one month of free Playstation Plus membership. But in order to activate that membership, I had to provide my credit card information, then opt into an automatic monthly renewal plan. I was, of course, free to disable the automatic renewal later on. Sony’s calculation, however, was that establishing renewal as the default option was of sufficient value that it was worth their while to irritate many customers upfront in order to secure it.)

So “nudge” paternalism, in its least controversial form, is simply the observation that when the choice architecture is inevitably going to favour one option over some other, then the government, within its areas of jurisdiction, should advance its commitment to the “public interest” by setting things up so that the favoured option is the one most beneficial to the person who is doing the choosing. (The more controversial form suggests that the state may legitimately force private actors to structure choices this way as well.)

Most choices we confront, it should be noted, are not organized in a way that favours our interests. Consider, for example, the layout of a traditional supermarket. This is all very carefully planned out, with the sole and specific intention of maximizing sales, with particular emphasis on maximizing the sale of high-margin items. For example, the dairy section is typically at the furthest point from the entrance, requiring the person who just pops in to buy a litre of milk to walk past every end-aisle display in the entire store to get the milk, pay for it, and exit. At least a half-dozen powerful cognitive biases are being exploited, along the way, in order to ensure that practically no one will go in to buy milk and emerge having bought just milk.

So what if the government were to run supermarkets? Would it be acceptable, or desirable, for these supermarkets to be organized in the same way? Right now, supermarkets are organized for the benefit of the owners of supermarkets, not the customers. In fact, several features of the way they are organized right now is overtly hostile to the interests of customers – particularly the strategic positioning of a gauntlet of unhealthy snack foods on the way to the cashier. What would supermarkets be like if they were organized to the benefit of the customer? They would be practically unrecognizable. Most obviously, the layout would be completely different, so that the foods that people come in to buy most often would be the easiest (as opposed to the hardest) to get to. But more importantly, the stock would be reorganized in such a way as to give greater prominence to foods that are healthy and foods that offer good value for the money.

This is of course merely an illustration, since governments are not about to get into the supermarket business. But governments are in the business of running a lot of institutions with cafeterias, including many schools and hospitals. (And even when the actual operations are outsourced, they have the ability to set a lot of conditions on how the food service is to be operated.) The typical privately-run cafeteria is also organized to maximize sales, and high-margin sales, often in ways that promote unhealthy eating (both in terms of food choice and the quantities chosen). Thus Sunstein and his co-author, Richard Thaler, make the suggestion that institutional cafeterias should be organized differently, giving greater prominence to salad than dessert, and so on.

This has generated the predictable backlash, with people complaining about meddling bureaucrats trying to shove healthy food down their throats. (The fact that meddling private corporations are investing several orders of magnitude more time and energy trying to get people to eat unhealthy food appears to be something that we have all gotten used to, and so no longer provokes any outrage.) And so we have the spectacle of Sarah Palin showing up for a speech at a CPAC convention slurping on a Super Big Gulp (although, looking at this spread in Runner’s World, one gets the impression that Palin doesn’t normally drink over 500 calories worth of soda in a single serving.) More recently, it led to this perfectly knee-jerk column by Margaret Wente (which has all the signs of being based on reading just Waldron’s review, not Sunstein’s book).

These responses all miss the subtlety of Sunstein’s position. In particular, they invite the obvious rejoinder, that if you don’t want the government organizing things in ways that are good for you, what do you want, for them to organize things in ways that are bad for you? Because the government is perfectly capable of doing that as well, and indeed, there is certainly no shortage of business lobbyists encouraging them to do so.

Waldron’s objection, however, seems to reflect his sense that nudges are a type of cheat, since they don’t actually improve our choices, or our choice-making abilities:

Nudging doesn’t teach me not to use inappropriate heuristics or to abandon irrational intuitions or outdated rules of thumb. It does not try to educate my choosing, for maybe I am unteachable. Instead it builds on my foibles. It manipulates my sense of the situation so that some heuristic – for example, a lazy feeling that I don’t need to think about saving for retirement – which is in principle inappropriate for the choice that I face, will still, thanks to a nudge, yield the answer that rational reflection would yield.

Now Waldron does not exactly say what his view is about all this, but he is obviously sympathetic to the idea that this kind of external manipulation involves some violation of autonomy or dignity:

Nudging takes advantage of my deficiencies in the way one indulges a child. The people doing this (up in Government House) are not exactly using me as a mere means in violation of some Kantian imperative. They are supposed to be doing it for my own good. Still, my choosing is being made a mere means to my ends by somebody else – and I think this is what the concern about dignity is all about.

The basic framework here is one that draws a distinction between what goes on “inside the head” of the agent and what depends upon the environment. It then claims (or assumes) that individual autonomy is based upon individuals using “inside the head” resources to govern their decisions. If your choice rule is flawed, it is not sufficient just to reorganize the environment in such a way as to ensure that you get the outcome you want. Waldron suggests that this something of a cheat. He thinks that you should actually fix the choice rule. But the question is, why? Most of us, for example, have rather unreliable memory systems. Suppose that there are some papers that I need to bring to work tomorrow. Although right now I know that I need them, I suspect that tomorrow morning, in my hurry to get to work, I may forget them. And so I lean them up against the front door, so that I will trip over them tomorrow morning on my way out. Is this some kind of a cheat? After all, I’m not really fixing my deficient memory system. I’m just reorganizing the environment, so that my deficient memory system will nevertheless yield the outcome that I desire.

My basic objection to Waldron’s discussion, therefore, is that the way he privileges “onboard” resources over environmental ones in thinking about personal autonomy is arbitrary. If you look carefully at the habits of highly effective people, you will see that they are full of environmental manipulations (or environmental “scaffolding,” to use the lingo). This is the major lesson that I take from the “extended mind” thesis, advanced most influentially by Andy Clark and David Chalmers. Individuals are constantly “offloading” elements of cognition onto the environment. That anyone could regard recognition of this fact as an affront to our “dignity” suggests a serious misunderstanding of how our minds function.

Joel Anderson and I have adapted this idea to get our concept of “the extended will,” by showing some of the multifarious ways in which individuals achieve rational self-governance through environmental manipulation. For example, imagine two people, one of whom keeps piles of junk food in the house, but only ever eats it in moderation. The other has difficulty eating in moderation, and so never keeps junk food in the house – making it so that, if he wants to snack, he has to get up and go to the corner store. Both individuals have achieved some sort of “rational self-governance,” with respect to the desire to avoid an unhealthy diet. The first person, however, is doing so based almost entirely upon “onboard” resources of will-power. The second person is achieving it primarily through an environmental manipulation (keeping the house clear of junk food), combined with the fact that he is too lazy to get up and go to the corner store. Is there any basis – other than Kantian purism – for saying that the former is autonomous, while the latter is not? I am inclined rather to treat these as just two different styles of rational autonomy.

Of course, the examples that I have given so far involve individuals reorganizing the physical environment so that they wind up with outcomes they want, despite employing flawed decision heuristics. The point is to show that our conception of rational autonomy should not require us to overcome our “inappropriate heuristics.” We can also work around them – this is the essence of the kluge. The kind of nudges that Sunstein is talking about are different in that they involve changes in the social environment. One might think that the threat to autonomy is more apparent here, because it involves surrendering decision-making authority to others. And yet it is important to recognize that individuals often do this as well, on purpose. In the same way that they offload tasks onto the physical environment, they also offload them onto other people. This is particularly true with respect to will-power and self-control, where our onboard resources are especially deficient. (Anyone who has been in a long-term relationship should be able to rattle off a half-dozen examples quite easily, of cases where they offload an exercise of will-power onto their partner.)

As a result, interactions that may look like person x imposing his will upon person y, or substituting his judgment for that of person y, may in fact be quite different. If person y has offloaded a particular exercise of will-power or practical reasoning onto person x – and thus, licensed the imposition – then person y’s actions are part of the exercise of person x’s rational autonomy. So to focus on that fact that one person is constraining or influencing someone else’s choices is to focus on the wrong thing – therein lies no violation of that person’s autonomy. The normatively significant question is whether that person would endorse those constraints or influences.

One of the problems in our society is that, in the name of liberty, we have deprived ourselves of a lot of the scaffolding that people have traditionally used to achieve self-governance – while at the same time transforming the environment so that the number of temptations has increased substantially. In other words, the world is evolving in a direction where the demand for rational self-governance is constantly increasing, while the institutional resources that individuals have historically relied upon to achieve rational self-governance are being withdrawn. The two most obvious examples are in the area of diet and addictive substances, but there are many others.

Finally, there is lots of evidence to suggest that “fixing” our broken heuristics, or retiring “outdated rules of thumb,” in the way that Waldron suggests, is not possible. Fixing them was a central feature of the first Enlightenment project, and it was in several respects an abject failure (at least as far as the mass of the population is concerned). In my book, I draw on recent “dual process” psychology to explain why so many heuristics are so resistant to modification, even despite our awareness that they are flawed. (And for anyone wanting a primary source, Brian Wansink’s work on diet is particularly good.)

Waldron writes:

I’m not sure I want to live in nudge-world, though—as a notoriously poor chooser—I appreciate the good-hearted and intelligent efforts of choice architects such as Sunstein to make my autonomous life a little bit better. I wish, though, that I could be made a better chooser rather than having someone on high take advantage (even for my own benefit) of my current thoughtlessness and my shabby intuitions.

My temptation, in responding to this, is simply to observe that one does not become the Chichele Professor of Social and Political Theory at All Souls College, Oxford by being a “poor chooser.” Genuinely poor choosers wind up as drug addicts, high school dropouts, and murder victims. Waldron, like Sunstein, myself, and everyone else involved in this discussion, is a member of what I call the “self-control aristocracy.” Indeed, anyone who completes a PhD, gets a job in academia, and publishes not just one book but several, is obviously at the upper end of the top 1%, which it comes to self-control, deferred gratification, and rational life planning. The fact that it may not feel that way, when the alarm clock goes off at 6:30am, or when one contemplates having just one more finger of scotch, should provide some insight into what life is like most of the time for members of the 99%.

Indeed, in this concluding paragraph, Waldron seems to have forgotten entirely that he is reviewing a book on public policy. I mean, what are the chances that the government is in a position to elevate the mass of the population to even the (apparently inadequate) level of rational self-mastery enjoyed by the Chichele Professor of Social and Political Theory? The sort of false modesty that is common in the professorial class (university professors sit around saying “I’m so absent-minded,” in much the same way that fashion models all say “I was so ugly in high school”) is in this case seriously interfering with Waldron’s ability to grapple with the important issue. If he finds his own decision-making “thoughtless” and “shabby,” then what sort of standard does he imagine the rest of the population is able to achieve? Pausing for a moment to contemplate just how bad everyone else must be, and what the chances are of improving their decision-making (raising them even to the level of the current Chichele Professor), should, I think, make one much more sympathetic to the nudge approach.

Finally, I should note that Waldron does make one incontrovertibly good point. One can find in discussions of behavioural economics and public policy an incredibly common fallacy, where it is assumed that demonstrating a bias in individual decision-making is itself sufficient to make the case for government interference. This is analogous, in many ways, to the fallacy exhibited by those who think that merely diagnosing an inefficiency in private markets is sufficient to make the case for state involvement. In both cases, what gets overlooked is the need to show that the state will somehow be able to do better. Now in the case of private markets, there is some reason to think that the state will sometimes be able to do better, because the state has powers of compulsion that private actors lack. But in the case of failures of rationality, what reason is there to think that the state will do better? After all, the state is made up of people. To make the case for state intervention, it seems, one would have to show that people who work for the state are somehow “more rational” than people who do not.

My formulation here is slightly narrower than Waldron’s (who brings up issues of incentives and trust as well), because I want to make the very basic point that, from a public policy perspective, the discovery that human decision-making is flawed in some way is neither here nor there. As such, it certainly does not speak in favour of state action – until some reason is given to think that the state will be able to correct it (and that the individuals working for the state, in charge of formulating a solution, will not be subject to the same biases, or worse). Fortunately, there is such an argument, and it has to do (again) with environmental scaffolding. The idea, roughly, is that state bureaucrats have better scaffolding for making such judgments, and so (under certain conditions) are more likely to be rational than private individuals. I summarized the argument briefly in Enlightenment 2.0, but could scarcely do it justice in a popular book. The idea comes from Andy Clark, who sets it out in one of my favourite articles of all time: “Economic Reason: The Interplay of Individual Learning and External Structure.” I would invite those with a professional interest in the question to consult Clark’s article.


Waldron, Sunstein, and nudge paternalism — 4 Comments

  1. Dear Prof. Heath,

    You make a compelling argument against Waldron’s suspicion of nudging, but shy away from explicitly meeting his central complaint about dignity.

    Waldron’s basic point is that it is worrying, from the perspective of dignity, when the government tries to work around our irrational laziness rather than confronting it head on. Pointing out that confronting it head on rarely works, while important, won’t calm anyone worried about Government House condescension. People may, for good reason, be willing to sacrifice efficiency for dignity.

    Your main argument, that Waldron arbitrarily privileges ‘onboard’ resources over environmental ones might also be seen as suspicions by someone concerned about dignity. It is true, as you point out, that people can use both kinds of resources to control their behaviour and lead better lives without sacrificing their autonomy.

    Something important happens, however, when you switch from talking about individuals deliberately outsourcing their self-control (say by enlisting a loved one) to talking about the government outsourcing people’s self-control for them (because it is in their best interest). It is possible to agree with your point that the former preserves autonomy while worrying that the latter may compromise it. When you write, “the normatively significant question is whether that person would endorse those constraints or influences” alarm bells go off for people who worry that what a person would endorse is not the same as what they do endorse.

    (There is an interestingly similar leap in Phillip Pettit’s new book ‘On the People’s Terms’ where he uses the example of handing over the key to your liquor cabinet to a trusted friend to set up an argument that properly democratic governments don’t limit the freedom of their citizens.)

    A better argument, for dealing with worries about dignity, is the one you make implicitly in the first part of your reply. Waldron assumes that in every choice the government offers us there is some default way of organizing the choice that is non-nudging (and dignity preserving) and that trying to deliberately improve this organization is dignity eroding. But it is not clear that there is such a default way of organizing a choice.

    It seems peculiar that when bureaucrats organize our choices without concern for how the organization itself might affect our welfare they respect our dignity but if they organize our choices with an eye to how that organization might help us they compromise our dignity. It might be concerning that the government so frequently gets to organize our choices. Whether they do a thoughtful or slapdash job of it does’t seem to make it worse.

    A useful parallel here is one that engineers and designers are already familiar with. Kim Vicente, for example, in his popular book ‘The Human Factor’ points out how laughably badly designed systems like the layout of controls and indicators in machinery and control rooms can be. In designing machines that require human input, he argues, engineers need to respect basic facts about human nature, such as the length of our limbs, our need for sleep, and the types of mistakes we tend to make. Yet no one objects that by making airplane cockpits more intuitive engineers are compromising the dignity of pilots. This is because there is nothing sacred or natural about the way airplane cockpits were laid out to begin with. It’s just that the original engineers didn’t put enough thought into it. Waldron worries that nudgers are “taking advantage of our deficiencies.” So are engineers like Vicente. To do otherwise would be bad design.

    Perhaps the issue, then, is with the way Sustein frames his argument. Instead of arguing that we need to start nudging, he could instead claim that we need to start nudging more intelligently. That way when Waldron writes, “I’m not sure I want to live in nudge-world” Sustein can reply that he already does.

  2. Unlike (perhaps) Wente, I have read neither Waldron nor Sunstein, but this summary leads me to think:

    Fair enough that Waldron’s argument sure sounds excessively abstract and ivory-toweresque. But when I got to the end of the critique of Waldron, I was left thinking: but isn’t Waldron still playing off of a pretty common and important implicit principal here? And that principal is: if freedom is to be meaningful it must include the freedom to be stupid.

    One can try to draw various lines, of course, but all of these are question-begging – the question being, when are you willing to let people be stupid.

    (This issue is implied obviously enough by the examples given in favour of nudging: yes, of course, any right-minded liberal would welcome and even insist on preventing people from making “bad” dietary choices. But attempting to do so with respect to sexual choices would be seen as the slavery, Holocaust, and apartheid all rolled into one.)

    One way of making the necessary distinctions might actually be implied by the dietary examples: don’t most people *say* they want to eat healthier? (Actually, no, which is still a problem – both for the nudgers and the unwillingly nudged, in my opinion! – but certainly many people say so.) So in that case we could help people do what they say they want to do. OK (if tricky to execute on a mass scale). But even then: aren’t all the choices we make in pursuit of a larger aim still an essential part of our freedom?

    (Caveat: I’m sure that Waldron and Sunstein have both addressed these kinds of objections, I’m just pointing out why I presume that “freedom must include the freedom to be stupid” must be a pretty important part of Waldron’s argument, although not reflected in the summary of it.)

  3. Arguably an interesting twist on the “nudge” phenomenon is also Ezra Klein’s argument favouring California’s new rape laws:

    “Critics worry that colleges will fill with cases in which campus boards convict young men (and, occasionally, young women) of sexual assault for genuinely ambiguous situations. Sadly, that’s necessary for the law’s success. It’s those cases — particularly the ones that feel genuinely unclear and maybe even unfair, the ones that become lore in frats and cautionary tales that fathers e-mail to their sons — that will convince men that they better Be Pretty Damn Sure… Colleges have settled into an equilibrium where too little counts as sexual assault, where the ambiguity of consent gives rapists loopholes in which to hide, and forces women to spend their lives afraid. The Yes Means Yes laws creates an equilibrium where too much counts as sexual assault. Bad as it is, that’s a necessary change. A culture where one-in-five women is assaulted isn’t going to be dislodged with a gentle nudge.”

  4. The supermarket example gets confusing because although it is the consumer who we are trying to protect or aid by having the default option (milk at the front) it is the producer whose behaviour we’re regulating. The nudge thing seems like a no-brainer when it is default consumer choice with consumer ability to pick the alternative (only the consumer suffers the consequence of the decision), but as you say breaks down when it would pick a winner and loser as between two parties…then we seem to be back into plain old regulatory debates over whose interests prevail in which circumstances…