Tuesday, 24 August 2010

A review of Nudge by Richard Thaler & Cass Sunstein

I first came across the term 'libertarian paternalism' in 2008 when Julian Le Grand proposed making smokers buy smoking permits. Le Grand was, and is, a classic Blairite—Eton-educated, champagne socialist, illiberal, wonk—and like all Blairites, he was obsessed with branding. Specifically, he was obsessed with finding friendly buzz-words to help sell unpleasant ideas.

Libertarian paternalism sounds better than authoritarianism. Nudging sounds better than shoving. And so it was that one of the least liberal public health policies imaginable was allowed to be prefixed with the word 'libertarian'. As I wrote in Velvet Glove, Iron Fist, this was an oxymoronic piece of doublespeak if ever there was one.

The idea did not go down well. As Julian the Great later said:

My e-mail inbox exploded. Mostly with pictures of Hitler, I have to say. People were very hostile to that sort of idea. So, although the nudge agenda, I think, does have possibilities I think care has to be taken that people don't feel that it's the nanny state, indeed the nanny state squared.

It is interesting that Le Grand has no voice on his shoulder telling him when he's crossed the line, and instead has to rely on the quantity of hate-mail he receives. Perhaps it's his focus-group mindset. Maybe he just lacks a moral compass. Most likely, as with many public health protagonists, he is an opportunist, lurching towards prohibition but drawing back when the public's apathy turns to disgust.

Le Grand was inspired by the ideas presented in Nudge, a book on behavioural economics written by Richard Thaler and Cass Sunstein. Thaler and Sunstein invented the phrase 'libertarian paternalism'. They also invented the phrase 'choice architecture'. You can see why this kind of management-speak appealed to the political class.

And not just in the Labour party. Nudge is said to be one of the Conservative party's favourite texts at the moment. Nudging is back in fashion and, in a shameless attempt to embrace the zeitgeist, the UK Faculty of Health titled its latest manifesto for authoritarian paternalism 'Healthy Nudges'. This document—which, in reality, is a glorified opinion poll—asserted that:

Sometimes legislation and regulation of the so-called ‘choice architecture’ are important tools for nudging people into making healthier choices.

As far the UK Faculty of Health are concerned, the kind of 'nudges' required are a ban on 'junk food' advertising, raising the minimum age of alcohol purchase to 21, reducing the speed limit in towns to 20mph, a minimum price on alcohol, banning smoking in cars and banning buy-one-get-one-free offers on food.

Since this was all I knew about Nudge, I began reading it this weekend fearing the worst. Nudging, I assumed, would be the latest euphemism for banning and prohibiting. But I was wrong. Nothing in Nudge justifies the kind of prohibitions being mooted by public health campaigners. Authoritarians will find little to console them in this book.

The underlying assumption of Nudge is that human beings are fallible, easily influenced and not always best-placed to judge what is in their own interests. Libertarians will find this a worrying premise from which to start. It echoes the Marxist concept of false consciousness and could be a charter for those who think they know what's best for people. But Thaler and Sunstein never view the public as unwitting dupes of corporations and peer-pressure. Rather, they view people as fundamentally sensible, if not always well informed. As such, they set significant limits on the kind of nudging that can be considered tolerable in a liberal society.

Libertarian paternalism, as defined by Thaler and Sunstein, aims to help people make the best decisions without obstructing those who still wish to make the 'wrong' decisions. In any government intervention, the costs to liberty and the economy should be close to zero. They would like people to save more for their retirement, take out the right mortgage, take advantage of pension schemes, save electricity and not get ripped off by high-pressure salesmen. In practice, this means making information readily available and comprehensible, cooling off periods, warnings and easy opt-outs.

When they talk about making 'good' choices easier, they actually mean it. Unlike the UK Faculty of Health and the last British government, they do not mean banning 'bad' choices. A paternalist solution to stop foreign tourists getting run over in central London would be to ban jaywalking. The libertarian paternalist solution would be to paint 'look right' on the street. It is difficult to see this as an unnecessary intrusion on personal freedom.

Nor do they appeal to the majority to decide the fate of the minority. Nudge almost exclusively deals with helping people avoid situations they could not conceivably desire. If asked, people would generally welcome a safer, healthier or cheaper option. You want to use a sun bed? No problem, but how about we fit the sun bed with a dial to make sure it turns off after ten minutes in case you fall asleep under it? You want to gamble? No problem, but if you know you have a gambling problem how about we enable you to be able to go to a casino and ask them to ban you?

It is hard to see even hardcore libertarians objecting to most of the ideas presented, although there are exceptions. Paying teenage girls a dollar a day to not get pregnant might save the taxpayer money in the long-term but it smacks of the nanny state and encourages public to see themselves as (to quote Alexis de Tocqueville) "a flock of timid and industrious animals, of which the government is the shepherd." Presumed consent for organ donations also sits uneasily with many people—me included—who are uncomfortable with the notion of the state owning its citizens bodies. 

Presumed consent is a classic Nudge problem. There is little doubt that many people would be happy to be organ donors but have not yet registered and will never do so. There is also little doubt that presumed consent would increase the number of organs available and save many lives. If opting out was as easy as making a one-off phone call or visiting a website (as the authors propose), it is difficult to muster up too much antipathy towards it. Presumed consent is not something I am fervently against, but I would like to see a more proactive attempt to get people to register voluntarily before we contemplate coercion. Years ago, I remember donor cards being available at shop counters. All you had to do was pick one up, fill it in and keep it with you. I haven't seem them for years and I doubt this is because budgets have been slashed. There are, in other words, better ways to help make choices easier without the presumption of consent.

The strongest argument against nudging is the slippery slope. Although most of Thaler and Sunstein's ideas would be of no practical detriment to either the target group or the wider public, they do concede that the impact cannot always be reduced to zero. Having accepted that small impositions on liberty and choice are not always avoidable, where does it end?

To be fair, Thaler and Sunstein address the slippery slope argument towards the end of the book, using an apt example:

Governments that start with education might end with stiff fines and even prison terms. The case of cigarettes offers a possible example. Some nations have gone from modest warning labels to much more aggressive information campaigns to high cigarette taxes to bans on smoking in public places, and a smoker would not have to be paranoid to think that the day might eventually come when one or another nation heavily regulates or even bans cigarettes altogether... Faced with the risk of overreaching, critics might think it is better to avoid starting to slide at all. (p. 235-6)

Thaler and Sunstein's counter-argument is that (a) individual proposals should be judged on their own merits; (b) their philosophy automatically puts a limit on further legislation by insisting on low-cost opt-outs; (c) some kind of action is inevitable, so it is better for it to be a nudge than a shove.

I am not fully convinced by these assurances. It seems to me to they overlook how the slippery slope works in practice. In reality, policies are not always judged on their own merits and the appeal to precedent is powerful. How often do we find campaigners demanding that 'loopholes' be closed, despite the fact that these 'loopholes' are actually deliberate exemptions created to make the legislation reasonable in the first place? And how easy has it been for health campaigners to apply the policies first used against the "unique case" of cigarettes to a whole range of products that carry an element of risk?

It may be fallacious to claim that because we are forced to wear seat-belts, we should be forced not to smoke or eat fatty foods, but campaigners who call it 'the next logical step' are not wholly wrong. Those who opposed seat-belt laws in the 1970s did indeed warn of a slippery slope and, what's more, they turned out to be right. One needs to look no further than 'Healthy Nudges' to see the truth of this. It refers to seat-belts no fewer than six times in eight pages. The message is clear—it was for your own good then and it's for your own good now.

When the idea is raised of regulating against allowing smoking in cars when children are on board, politicians are often concerned about legislation in the “private space”. Of course, there are clear parallels with seat-belt legislation...

It is often far cheaper and longer-lasting to introduce regulations or change the law. The classic example of this is the use of seat-belts...

Did politicians of the 1970s intend compulsory seat-belts to be used as a precedent for smoking bans and tax rises? Surely not. Would they have passed the legislation had they known what it would lead to? Probably not. Saying 'we did that therefore we must do this' might be fallacious, but people are susceptible to fallacious arguments. Since Nudge revolves around the idea that human beings are fallible, this point could have been more thoroughly explored.

But perhaps this is to blame the leader for the sins of his followers (see here for an amusing example of how nudging can be abused for malign purposes). If politicians stuck to both the spirit and the letter of Thaler and Sunstein's philosophy, the nudge agenda would be largely benign and almost certainly beneficial. Far from supporting the kind of policies being pursued by the UK Faculty of Health, any British government that was genuinely committed to the Nudge agenda would have no choice but to repeal whole swaths of legislation that already cross the line between libertarianism and paternalism. But as Le Grand and the UK Faculty for Health amply demonstrated, once an idea reaches the mainstream, you no longer get to make the rules.


[Elsewhere, The Guardian (which originally gave it a glowing review) has turned against Nudge for being too libertarian. For an alternative view, Spiked criticised Nudge for being too paternalistic. Who you agree with will depend on your politics.]


13 comments:

Simon Cooke said...

Brilliant post. Might I just add that behavioural economics (at least the theory) is not very robust. Lack of information isn't a justification for arguing that we do not respond to incentives (i.e. behave rationally)which is what behavioural economics suggests.

Anonymous said...

I think people repeat mistakes because they are protected from the full consequences - i.e. they can get blotto on a Saturday night knowing that the local NHS will patch them up free.
There is also a presumption that the authorities do have all the information and do know what is best. Whereas they do make misinformed decisions all the time and we have to deal with the unintended consequences.

Eric Crampton said...

Do check the work by Whitman and Rizzo on Sunstein's libertarian paternalism. It's all up on Whitman's blog - <a href="http://agoraphilia.blogspot.com/search/label/paternalism>here</a>.

Snowdon said...

Cheers Eric! This part sums it up:

The pattern repeats through the rest of the op-ed. If gallons-per-mile laws don’t induce people to choose different vehicles, then we need higher gas taxes. If telling people how much electricity their neighbors use doesn’t cause them to turn out the lights, then we need a carbon tax...

In our first paper on paternalist slopes, Mario Rizzo and I warned about precisely this kind of process. When a policy is enacted to achieve a specific goal and then fails to achieve it, further policies are justified on grounds of achieving the goal that “we” have already agreed upon.

Anonymous said...

Cool web site, I had not noticed velvetgloveironfist.blogspot.com previously during my searches!
Continue the fantastic work!

Anonymous said...

To the last Anonymous discoverer of Chris's blog...welcome aboard. Rarely a post goes by here without it dripping clarity, rationality, and erudition. And no, I am not 'the author'! Just an observer, and admirer, of good writing (though it aways helps when you fall squarely into Chris's philosophy!)

Anon1 said...

OT
There is a leaked “template” for reporting on tobacco issues. It explains why we’ve seen standard antismoking slogans in media reporting for the last few decades.

http://smokles.wordpress.com/2010/08/26/scoop-leaked-upa-guidelines-for-tobacco-reporting/

Anonymous said...

So who tells the United Press association to do this ?
Who controls them ?
Next logical atep.

Anon1 said...

The smokles reference is possibly a “send-up”. However, looking at a few decades of newspaper antismoking articles, it –unfortunately - very well describes the “formula” for putting together such articles. Regardless of a story, there must be a spattering of the requisite slogans - “the leading cause of preventable death”, the “4000 carcinogenic/toxic chemicals”, the “great financial cost to society”, the “chemicals that are also found in lead batteries and rat poison, etc”.

Anon1 said...

The smokles “scoop” appears to be a “send up”. However, if you peruse Tobacco Control media-advocacy manuals (e.g., as in the Godber Blueprint) they are an exercise in the manufacture of inflammatory propaganda. These manuals advise in the use of the most inflammatory terms possible that go far beyond the implications of fact. A scrutiny of [antismoking] articles appearing in newspapers over the last many years certainly highlights a “formula”. The formula is to include a number of successful, albeit fraudulent, slogans – even if disjointed to the story at hand - such as “the leading cause of preventable death”, the “4000 carcinogenic/toxic chemicals”, the “great financial cost to society”, the “chemicals that are also found in lead batteries and rat poison, etc”. Also included should be comments from a few nonsmokers that believe that antismoking policies are wonderful and comments from smokers who believe that their denormalization is also wonderful. The impression must always be that smoking should “justifiably” be eradicated from public and that everyone, smokers and nonsmokers alike, think it’s a grand, noble quest. The “formula”, however, is most probably not with the journalists, but with the TC advocates. Journalists simply print what they are fed by TC advocates.
www.rampant-antismoking.com

Anonymous said...

Personally, I find this "Nudge" approach to matters of public policy to be extremely troubling. This is mostly due to the reasons that CS has outlined above, like the "slippery slope".

I read CS' post the day before yesterday, and I read it again yesterday, and skimmed it again today. Something's bothering me about it. I have the suspicion that CS was trying to suss this out himself while writing this blog entry, because it's a puzzling matter.

The puzzling matter is that the "slippery slope" that "nudging" invites is quick and steep, and it isn't so much a matter of public policy as it is human psychology.

So, trying to put my finger on what precisely so offends me about "nudging" beyond my libertarian political feelings, I have to make the odd and uncomfortable confession that I simply don't want to be subjected to such psychological reductionism by my government, regardless of what it is my government actually does in practice.

I know that corporations engage in all kinds of subtle psychology to sell products, but I have less of a problem with that, simply because I know that I have a choice. Even if I've been subtly manipulated, I have a chance of becoming aware of the manipulation.

Government doesn't work that way. Government is the entity that embodies the moral and foundational bedrock of our society at large, whether we like it or not. For example, only government could lead the vast majority to believe that secondhand smoke is a highly dangerous health menace that people should take an active effort to defend themselves against without any thought of the liberty of another human being.

Once "nudging" is accepted by the public to be a useful strategy for government policy, anyone can make any argument for government enforced "nudging" according to their desires.

While nudging may not infringe on our Rights to speak, worship, assemble and/or report, "nudging" amounts to what is really a confession that government can use modern techniques to manipulate how you think, thereby finding a back door way to bypass any of the active freedoms like speech, assembly, religion, etc.

These freedoms are supposed to ultimately guarantee freedom of thought. Even if, and perhaps especially, those thoughts are completely aberrant to any and all societal norms.

So, to me, there is something inherently illiberal about applying psychological findings and behavioral economics in the government/public sphere.

If government knows what I am 95% likely to think, the world can never know what I'm 5% likely to think.

-WS

Anonymous said...

Personally, I find this "Nudge" approach to matters of public policy to be extremely troubling. This is mostly due to the reasons that CS has outlined above, like the "slippery slope".

I read CS' post the day before yesterday, and I read it again yesterday, and skimmed it again today. Something's bothering me about it. I have the suspicion that CS was trying to suss this out himself while writing this blog entry, because it's a puzzling matter.

The puzzling matter is that the "slippery slope" that "nudging" invites is quick and steep, and it isn't so much a matter of public policy as it is human psychology.

So, trying to put my finger on what precisely so offends me about "nudging" beyond my libertarian political feelings, I have to make the odd and uncomfortable confession that I simply don't want to be subjected to such psychological reductionism by my government, regardless of what it is my government actually does in practice.

I know that corporations engage in all kinds of subtle psychology to sell products, but I have less of a problem with that, simply because I know that I have a choice. Even if I've been subtly manipulated, I have a chance of becoming aware of the manipulation.

Government doesn't work that way. Government is the entity that embodies the moral and foundational bedrock of our society at large, whether we like it or not. For example, only government could lead the vast majority to believe that secondhand smoke is a highly dangerous health menace that people should take an active effort to defend themselves against without any thought of the liberty of another human being.

Once "nudging" is accepted by the public to be a useful strategy for government policy, anyone can make any argument for government enforced "nudging" according to their desires.

While nudging may not infringe on our Rights to speak, worship, assemble and/or report, "nudging" amounts to what is really a confession that government can use modern techniques to manipulate how you think, thereby finding a back door way to bypass any of the active freedoms like speech, assembly, religion, etc.

These freedoms are supposed to ultimately guarantee freedom of thought. Even if, and perhaps especially, those thoughts are completely aberrant to any and all societal norms.

So, to me, there is something inherently illiberal about applying psychological findings and behavioral economics in the government/public sphere.

If government knows what I am 95% likely to think, the world can never know what I'm 5% likely to think.

-WS

blingmun said...

Clear, thoughtful writing as always. However, I do think this article misses a more fundamental objection.

Nudge is wrong from a libertarian perspective because it is no business of the State whether or not I have a healthy diet or decide to risk my life sky diving.

To put it in some context, the governor of Hong Kong prevented colleagues from gathering statistics because he knew they would be used to justify State intervention of one kind or another.

As long as we accept that the State's role goes beyond protecting us from foreign invasion and things like theft, murder, fraud etc. (specifically those things that curtail other people's freedoms) then we have already lost the whole argument. Even allowing the State to gather stats on a problem invites calls for "the government to do something about it".

We shouldn't be paying taxes to produce the statistics that would justify any nudges. We certainly shouldn't be paying for the subtle creativity and endless man hours that would produce the desired 'choice architecture'.