Nudging and Rationality

I’ve promised myself I’d write something about ‘nudging’ on here, so here goes.  It’s really only some initial thoughts, and it’s more about the principle than the detail, but hopefully that will come across and be interesting.

I’ve written here before about how nudging isn’t necessarily new (look at James Kneale’s work on the interest during the nineteenth century in governing the shape and design of pubs), and as part of this how it’s consistent with ideas of neoliberalism.  Sometime in the future I’d also quite like to write about how it doesn’t really mark anything of a ‘third way’ between intervention and laissez-faire, but that’s not too dissimilar an argument to the one about neoliberalism.

What I want to talk about here is how I see a misunderstanding of human rationality and desires at the heart of the ‘nudging’ approach.

Underlying the idea of nudging as outlined by Thaler and Sunstein and those analysing their work (like Marteau and colleagues) is the understanding that two contrasting systems drive human behaviour.  One is a reflective, goal-oriented system; the other an automatic, unconscious system based on emotions/affect.  I found this idea to be explained clearly and simply by Dan Gardner in his book Risk.  He uses the terms Head and Gut to describe the two systems.

Thaler and Sunstein relate this duality to classical versus behavioural economics.  Classical economics, they say, treated people as if they only used Head – as perfectly rational, calculating individuals with access to perfect information and plenty of time to think.  In reality, we humans are flawed and use Gut quite a lot – often to ill effect.  Behavioural economics is seen as an improvement on classical economics, as it accepts that people are irrational, but continues to seek to model them.  As Jones and colleagues put it neatly, it seeks to render them ‘predictably irrational’.*

Thaler and Sunstein then take these insights as to how people behave irrationally in order to change their actions so that they are ‘rational’.

However, in my view this is slightly to misunderstand the motivation of classical economics.  It is not concerned with thinking about people’s inner monologues; it is concerned with modelling how they actually behave.  This is a common theme in social research, perhaps most vociferously put forward by David Silverman.  (This is not to say that at other times we shouldn’t be concerned as much with what people say as what they do.  I would of course humbly suggest that my work on drinking and distinction, and its implications for alcohol policy, is a case in point…)

In classical economics we don’t always have to worry whether someone ‘really’ wants the chocolate bar in the long run; we’re more concerned with how much they’re prepared to pay for it in practice when they get to the checkout counter.  That is, whether it’s Head or Gut, the point is modelling the decision and what this tells us about people’s expressed preferences.  This decision can’t really be ‘irrational’ in any meaningful sense, from this perspective; it’s merely expressing the way this person has (unconsciously) balanced present pleasure with future health and wealth.  That is, there isn’t really a reason to see the person as ‘predictably irrational’ unless we’re concerned with understanding the disjuncture between what they say and what they do.

This is revealing of the ‘nudging’ perspective on rationality.  Rather than assuming, like conservative philosophers such as Michael Oakeshott and John Gray, that humans are incorrigibly irrational, Thaler and Sunstein want to re-mould them as more rational.

Oakeshott and Gray would consider this not just a vain, but a dangerous aspiration.  I don’t want to get into that argument now (though I have some sympathies with it), apart from to note that ‘rationality’ as an aspiration can be problematic for policymakers as they often see rationality (like Thaler and Sunstein) as decided by final outcomes rather than the process of thinking.  Smoking or drinking to drunkenness are not by definition irrational; they are only irrational if they don’t reflect how a person ‘really’ values their future health compared to their current pleasure.

This brings us round to this idea of ‘real’ preferences again.**  Thaler and Sunstein are concerned to demonstrate the falsity of the (straw man) claim that ‘almost all people, almost all of the time, make choices that are in their best interest or at the very least are better than the choices that would be made by someone else’.

For this statement to make any sense, it seems to me that we first need to understand what ‘best’ choices are.  Mill might make the suggestion that there are many different ways of living, and any individual with a ‘tolerable’ amount of common sense should be trusted to make their own choices.  Thaler and Sunstein don’t agree, as shown by their analogy of making life choices: a novice playing an experienced player at chess.

Not only are the aims of chess clear (to win), but the rules are also laid down formally.  We can imagine that different people might have different rules (or conventions or scruples) for going through life, but more importantly we certainly don’t agree on what the ‘aim’ of it all is.  The analogy is bizarre – who is a grand master at living?  Can we all agree who that might be?

One might generously assume that they mean something more like every individual is playing their own ‘game of life’, with their own aim and rules.  However, there remains this assumption that even for an individual there is a clear goal.  This is clear from the writings of the LSE philosopher Luc Bovens.

Bovens is very much in favour of nudges – he goes so far as to say that we should be considering ‘whatever works’ and these could well work so we should not be concerned about ‘ideological commitments to . . . human agency’, for example.  This raises important questions about how we might assess whether something ‘works’, which I’ll come back to in a later post, but for the moment I want to think about how Bovens judges whether or not nudges are legitimate.

Rather than some idea of preserving liberty, or an individual’s personal choices, he takes Thaler and Sunstein’s approach that we should be trying to help people make choices that are ‘better’ than the ones they make for themselves, and judges that an intervention is reasonable so long as it maintains a person’s ‘preference structure’.  This is defined as a consistent, coherent ‘conception of the good’.

Note that Bovens is not suggesting we do actually behave in accordance with such a structure; the whole reason we need to be ‘nudged’ is that we don’t.  As he puts it, ‘We choose on the background of a fragmented self’.  The aim of nudging is to put us together into a coherent self.

To me it seems completely impractical that we could work out what someone’s putative ‘preference structure’ is (what actions should be discounted as mistakes, and what as reflecting the underlying ‘reality’?), but the issue isn’t one of practice but theory.  That is, I don’t believe this is even theoretically possible – still less desirable.

As John Gray suggests, there are many competing desires, identities and communities within societies – and indeed within individuals themselves – and no single rational solution to balance them up.  Moreover, these desires aren’t consistent over time and they won’t all reflect a coherent view.

Given that the most outstanding philosophers across the ages have struggled to put forward a coherent, consistent worldview, it seems unlikely that the rest of us will have mastered one in our everyday lives.  What is the singly expressed ‘aim’ of your life, by which your ‘rationality’ can be judged?  How can intoxication be weighed against health; or reading against playing football; or cake against cheese?  Which should I be ‘nudged’ towards?

Although Thaler and Sunstein mock classical economists with their homo economicus, they are actually lamenting the nature of the human condition – they would like people to be more like this fiction.  Nudge is a manual for how to deal with the fact that humans make ‘bad’ decisions – defined as those they wouldn’t make ‘if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control.’  As they would be the first to point out, this calculating ideal doesn’t sound like a human – and I’m not sure I would want to live in a world where people behaved like this, even if it were only through ‘nudges’ from on high.  Is a world led solely by the Head an attractive vision?  (And whose Head?)  To err is human.

* I’m not sure this phrase is theirs originally, but it’s through their work that I’ve come across it.

** This is particularly interesting in the field of addiction.  Thaler and Sunstein, like many others, cite the example of smoking, and how the ‘overwhelming’ majority of people want to quit.  This is fascinating for me, as in discussions of addiction and treatment reference is often made to the idea that a person needs to have a ‘genuine’ desire and commitment to change, and this is quite different from simply stating that one would like to quit (particularly in a world where addiction is stigmatised).

Although I’ve not referenced them directly, my thinking on nudging has probably been coloured by reading Amir & Lobel, Burgess, Goodwin, Grüne-Yanoff and Selinger & Whyte.