Thursday 25 December 2014

The Rights and Wrongs of Libertarian Paternalism

‘Libertarian paternalism' is Richard Thaler and Cass Sunstein's big idea for improving individual choice-making while respecting our autonomy. It has inspired fierce and sustained academic criticism from philosophers and economists from both the left and the right - as well as from less distinguished commentators like Glen Beck. Ultimately though most of these critiques seem to be complaining more about the depressing findings of behavioural economics research than Thaler and Sunstein's positive proposals to nudge us to choose better.

Thaler and Sunstein begin from the behavioural economics evidence that people do not choose as rationally as standard economic theory (or our folk psychology) suggest. In particular, people's choices are often strongly influenced by what should be irrelevant features of how their choices are framed. People tend to put more food on their plate if the plate they are handed is bigger; most of us go along with whatever the default option is for choices about pension contributions and organ donation; and so on. In many cases it seems that we don't have an existing specific preference over outcomes and are therefore open to having our subjective valuations shaped or induced by how the options are presented.

In light of this, Thaler and Sunstein propose a governmental ethic of libertarian paternalism. Since how we make many of our choices is so influenced by their framing, it is right and proper for governments to try to design how choices appear to people in a helpful way (paternalism). Yet merely redesigning the presentation of choices does not force people to choose what the designers had in mind. People will still be as free as ever to choose what they want, if they know what they want (libertarian).

As they explain in their popular treatment of their idea, Nudge:
A nudge, as we will use the term, is any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid. Nudges are not mandates. Putting the fruit at eye level counts as a nudge. Banning junk food does not. 

I. The Critique: The Assault on our Autonomy

The critics of libertarian paternalism have landed a couple of good blows.

Of course more organ donation is the right answer
First, how do we know what the 'rational' choice is that people should be making? In order to design choice architectures that nudge people to make the 'correct' choice - e.g. about signing up as an organ donor - policymakers must believe that they know the correct answers to these. In other words, government civil servants (no doubt with the assistance of behavioural economics consultants) are supposed to be able to work out what is in the best interests of 'ordinary' people more accurately and reliably than we can.

The challenge is whether this is plausible, or whether governments are likely to do something else instead. Are governments really able and interested in working out what their citizenry really want, i.e. what we would choose if our cognitive biases didn't pull us in other directions? Or is it more likely that governments will instead take the opportunity to decide on what the correct choice should be - of course people should eat less, become organ donors, use less electricity, etc.

The second challenge relates to this. Given that policymakers are apt to substitute their own value scheme for those of the people concerned, the way that libertarian paternalist policies work becomes very worrying. Recall that the problem Sunstein and Thaler identify is that people's choices are influenced 'behind their back' by how those choices are presented to them. Yet the solution they propose does not attempt to address this problem. Instead they propose to use the exact same 'behind your back' method to manipulate people into making better choices.

This, as several critics have noted, makes libertarian paternalism a rather insidious project. Perhaps more objectionable even than regular paternalism. Normally when the government wants you to do something (or stop doing something) they use very clear and obvious measures. They either ban things (like heroin or driving without insurance) or put whacking great taxes on them (like tobacco). These sanctions work by constraining or reshaping the options people have and their costs and benefits. Liberals since JS Mill have a principled objection to such treatment. It is obnoxious to have the state forcing us to behave in certain ways purely 'for our own benefit'. What right does the government have to control what I do if it doesn't harm anyone else?

Yet there is an advantage to a government being so obnoxiously direct about its paternalism: one can politicise and challenge the government's limitation of our choices, such as in terms of Mill's harm principle. This may not be possible if the state instead tries to change how we behave by exploiting our cognitive weaknesses about how we understand the choices we face. How can you criticise such manipulation if you can't see it? How can you demand your freedom back if you never actually lost it? Democracy itself may seem to take on a reflexive character in this model: we choose a government on the basis of our understanding of our interests and values, but then the government shapes our understanding in turn.

II. The Immorality of Government Inaction

Such criticisms do make a case for restrictions around the practise of libertarian paternalism, such as the need for explicit public transparency and accountability in government nudging. Nevertheless they do not succeed in challenging its legitimacy in a principled way because they mostly do not engage with its fundamental logic. They suppose that the choice governments face is about whether or not to frame the choices their citizenry face, rather than about how choices should best be framed. That is, they assume that government inaction would be morally neutral, not recognising that once we come to understand the significance of choice architecture, inaction itself becomes a moral choice that should be justified. Knowledge can change the moral expectations and standards around government action, just as the scientific understanding of germ theory changed the morality of government inaction on public sanitation and food regulation.

Indeed, it is the discomfiting research findings of behavioural economics about the limits of human cognitive powers that the critics seem to be really complaining about, rather than Thaler and Sunstein's specific proposal. As if complaining about inconvenient facts will make them go away.

Liberalism is built on the normative ideal that human persons are sovereign and that we should have the right to decide for ourselves on matters that primarily concern us. But it turns out that many liberals also make an empirical assumption that individuals are in fact the best judges of our own interests and are therefore best able to make decisions for ourselves. Behavioural economics research has shown that this convenient conjunction doesn't hold. This is upsetting to many people, including not only Kantians and Millians but also more generic liberals, such as mainstream economists.

The core assumption of neoclassical economics is that agents are rational and markets are neutral. Market outcomes are naturalised because they are the outcome of the laws of supply and demand operating over rational agents; they are efficient because they allocate on the basis of rational effective demand; they are just because that allocation reflects agents' own free and rational decisions in furthering their own interests (of which only they can judge). If people aren't rational after all, and 'markets' are manufactured contexts (like supermarkets) that are not neutral, then markets cannot be counted upon to produce either efficiency or justice. They may need supplementing by technocrats who know better than you what you should choose to secure your own interests

Some critics of framing suppose that forcing people to make an explicit choice would be neutral, and also better for people in the long run. For example, requiring people to choose whether they want to donate their organs after death without any default option. This misunderstands the new science of choice. There is no way to present a choice that doesn't influence choosing, which only reveals people's real preferences. And forcing people to make open explicit choices because you think that is what is best for them, for example in terms of the development and exercise of their personal autonomy, is still an instance of government paternalism. The way we choose is shaped and influenced by an external power that is not ourselves.

The inconvenient fact is that because most of us don't have pre-existing preference rankings for a great many options we come across, we make our decisions about what we want in the face of the choice we are presented with in a way that is always systematically biased by the framing of that choice. This has long been known to profit making companies, who exploit it to pull us into choosing expensive credit cards, the strangely expensive chocolate bars at cash registers, and so on. Indeed, this line of academic research into our cognitive biases has done much to help such companies organise their ad hoc methods more systematically.

The general point is that if we care about autonomy then we should care about people's real ability to choose in a way that secures and advances their interests. Pretending that our choices necessarily only reflect our own reasons may make us feel better, a more superior form of life, but does nothing to protect ourselves and our interests from such exploitation. Thaler and Sunstein's essential argument is that once one recognises the real world significance of framing, governments have an ethical obligation to step in to mitigate its potential to undermine people's ability to choose prudently - to safeguard and further our interests in health and wealth, say. Choosing not to try to frame choices better does not leave citizens free to make their own choices. It leaves us vulnerable to contingency and to exploitation by the corporations who employ framing effects to maximise their profits not our interests.

Many people associate paternalism with the odious idea that citizens should be treated like children, with the state taking over the responsibility for judging what is and isn't in their interests, like one of those Arab tyrannies in which the king or president for life portrays himself as the father of his people. Such paternalism is incompatible with political liberalism. But some versions of paternalism, like the softer 'libertarian paternalism' proposed by Sunstein and Thaler seem much less obnoxious. One can even see them as an important component in the individual autonomy so prized by liberalism.

This is a point particularly emphasised by the development economist, Esther Duflo. In her 2012 Tanner Lectures she notes that rich world critics of paternalistic policies generally don't realise how far they themselves benefit from their own government having made default decisions in a helpful way, such as the drinkable water that comes straight out of the tap, immunisations arranged and scheduled at each stage of a child's life, automatic enrolment in state pension systems, and so on. Poor people are often criticised for their lack of will power and judgement. But this relates to the hard fact that the world's poor face the stress of vast numbers of significant choices every day where the default option is bad. Clean water is something they have to go out and find, likewise immunisations and schooling for their children, healthcare that isn't quackery, and so on. Choosing well about relatively basic things is difficult and exhausting, a burden that often takes over their life to the extent that they have little resources left over for making more important long-term choices about who they want to be and how to get there.


The ability to easily choose prudently is something we take for granted. We'd like to think that it is something remarkable about us, but in general it turns out to a feature of the framing of the choices we face, like the rumble strips on motorway exit ramps designed to make you feel how fast you are still going and slow down. The idea that life in a liberal society should consist in figuring all this out for yourself is a strange and unsustainable one. This is what institutions were invented for. In any case, there is a reason Sunstein and Thaler's approach is called libertarian paternalism: if you already know what you want, or you are so remarkably smart that you can see through framing effects all by yourself, you remain just as free as ever to figure out what choice is best for you.

A previous version of this essay was published on 3 Quarks Daily.