Nudging for Changing Selves

When we go into a car showroom and we’re greeted by a series of sales assistants, one after the other, we tend to end up buying from the first person we met when we came in the door. And when we don’t have a strong preference between a range of political candidates on a ballot, we tend to pick the one at the top of the list. The American psychologists Dana Carney and Mahzarin Banaji call this general phenomenon ‘first is best’. Some people think that empirical results like this, which tell us how our choices are sensitive to the ways in which they are presented, can be used by governments to improve people’s decision-making in important contexts. Many of the classic results in this area have failed to replicate when later researchers have carried out new versions of the original experiments, so we should exercise caution when we appeal to any of them while the field is still in a state of flux, but let’s assume for the sake of argument that this one is secure. My question in this post is this: how might a government use this result? Here’s a standard, simple example. As the civil servant charged with designing the menu for the cafeteria in some institution, you might have found that people do, upon reflection, want to eat healthily, but also know that, when faced with a choice between a chocolate bar and an apple, they’ll be swayed by temptation to choose against their considered preferences and take the chocolate. In this case, you might appeal to Carney and Banaji’s research and design your menu so that the healthier options are listed at the top, in this way hoping to sway customers back from temptation to make the choice they would endorse upon reflection.

Were you to do this, you’d be engaged in what behavioural economists call nudging. Deliberately ordering the menu items in this way, rather than picking the order at random or deliberately putting the unhealthy items at the top, is said to have a number of advantages. For one thing, it’s supposed to be compatible with libertarianism, a political viewpoint that says that the government should never restrict its citizens’ freedom of choice. By listing the healthy options at the top, you might make someone more likely to choose them, but you don’t remove any options, nor make any options significantly more difficult to choose—it’s hardly a drag to read further down the menu for the chocolate bar selection. For another thing, it’s supposed to improve the lives of the individuals who are successfully nudged; and, most importantly, it improves them by their very own lights, or as the nudge enthusiasts often say, as judged by themselves—I do, upon reflection, want to eat more healthily. So, while this is assuredly a paternalist move in which the government tries to improve the lives of its citizens directly rather than merely providing them with the means to do so themselves, it is what economists call means paternalist, not ends paternalist. It tries to affect the means that you take when pursuing your ends, but it leaves your ends untouched—it doesn’t try to make you want to eat healthily if you don’t already; it just tries to make you choose in ways that better serve the end of healthy eating if you already have that goal.

In the previous blogpost, I talked about what Edna Ullmann-Margalit calls big decisions. These are decisions in which, if you choose one of the options, it will lead you to have what L. A. Paul calls a personally transformative experience, that is, one that will change some of your core values. For instance, when I am deciding whether or not to adopt a child, I might know that, if I do, my core values will be changed and I’ll come to value bringing up this particular child a great deal more than I currently do—at the moment, say, I’m pretty keen to stay child-free. But now an interesting question arises: is it ever OK for the government to nudge you, when faced with a big decision such as whether to adopt a child, into choosing the option that is likely to lead to a personally transformative experience? The question arises because of the test offered by two of the most enthusiastic proponents of nudging for whether a government should administer a nudge. Richard Thaler and Cass Sunstein, the authors of the book Nudge that popularised the practice and extolls its virtues, say that a government may nudge someone if that person would assent to it after the nudge has taken place were they to have plenty of time to think about and reflect on it, were they not burdened with any cognitive limitations such as biases, fuzzy thinking, poor numeracy, etc., and were they appraised of all available information pertaining to their choice. But of course, if the government were to nudge me into adopting a child, I would indeed assent to it after the event, even in this idealised conditions Thaler and Sunstein describe. After all, by that time, my values would have changed, and I would assign great value to bringing up that child. So such a nudge is OK by Thaler and Sunstein’s lights. But that seems wrong. Surely any government that did this would be overstepping their remit. Or, at the very least, their intervention would no longer count as means paternalist in the way that nudging is advertised to be. Instead of leaving my ends untouched and nudging me to pick a means that would better achieve those ends, I’ve been nudged into changing my ends. And yet, because the test for legitimacy is administered after the nudge, it would pass. What’s more, even very heavy-handed nudges might pass. If the government had subjected me to subliminal messaging or emotional blackmail for some lengthy period before I made my choice, I might now feel that the outcome thereby achieved is so valuable, I’m happy they did so.

All of this seems too much. This is not in the spirit of nudging at all. But how to prevent it? A natural response is just to say that we should administer Thaler and Sunstein’s test both before and after the nudge. So we’d say that a government is allowed to nudge me towards a decision if, were I to consider the decision both before and after the nudge with plenty of time for reflection, none of the cognitive limitations I actually have, and all the relevant information, I’d be happy for it to go ahead. Since I was inclined towards staying child-free before the choice, I wouldn’t assent to any nudge that led me to adopt, and certainly not a heavy-handed one. Of course, we couldn’t actually administer the test beforehand, since many nudges only work if you aren’t aware you’re being nudged. But the test is an idealisation anyway, so that doesn’t seem to be a problem.

A bigger problem is that there are cases where this goes wrong as well. (Krister Bykvist and I both offer versions of this case; he in this paper, me in my book.) Suppose that, at the moment, I give slightly higher value to being a parent than to remaining child-free. I value raising a child a great deal, and I value remaining child-free just slightly less. What’s more, I know I’ll retain these values if I were to remain child-free—I’ll always slightly wish I’d become a parent, but I’ll value my child-free life a great deal all the same. I also know that, if I were to become a parent, I’d come to value it dramatically less than I currently do; but I’d also come to value the child-free alternative even less than that. So, in this situation, both before becoming a parent and afterwards, I value that life more than the child-free life. So I’ll pass Thaler and Sunstein’s test before and after I’m nudged—at both times, I’ll be grateful for the nudge. And yet that nudge will send me down a path to which I’ll assign very low value once I’m on it. And that seems wrong.

So the search for a workable test for when it’s OK to nudge someone into a personally transformative experience continues…