When mathematician Stanislaw Ulam challenged Paul Samuelson to point to anything in economics that was both universally true and non-obvious, Samuelson pointed to David Ricardo’s principle of comparative advantage: Even the least productive countries still benefit from trade.

Richard Thaler’s Nobel Prize is richly deserved for several substantive contributions in behavioural economics. His work in the early 1980s on consumer choice and the economics of self-control laid some of the foundations for the behavioural economics that was to come in the 1990s and 2000s.

But my favourite remains his paper with Cass Sunstein that formed the basis for his later work on nudges. The paper’s title, Libertarian Paternalism is not an Oxymoron, is true and non-obvious. And it could be added to the list of answers that could be given to questions like Ulam’s.

The proposition is simple. If there are predictable behavioural biases in individual choice, then choice context unavoidably influences choice. For example, if voters who do not have strong preferences among options on one part of a ballot paper tend to choose whoever is at the top of the list, then the choice of whether to present options in alphabetical order or in random order matters.

If it is impossible to avoid influencing choices when setting options or designing policy, even if you want to be neutral, then even a libertarian cannot help but be a little bit paternalistic.

In Sunstein and Thaler’s libertarian paternalism, in contrast to some of the Nudgers who followed them, choice must be respected. The default option should be the one that is best for most people, as those people themselves judge things.

Minimise costs

And so a libertarian paternalist would wish to minimise the overall costs of assigning the wrong default setting. A good default would require relatively few people to switch, and would not impose substantial costs on those assigned the wrong default. And switching from the default must be as costless as possible in order that choice be respected.

While the underlying psychological literature on which the nudge theory is based comes under increasing attack in what’s been called the replication crisis, nudge still holds much promise – or at least in contrast to more heavily paternalistic forms of regulation.

If the psychological literature is right about predictably irrational choices, then a libertarian paternalism at least offers an opt-out for those who do not welcome the nudge. And if the psychological literature is wrong, then the libertarian version of paternalism does far less harm.

So if default options are sticky, then opt-out savings schemes like KiwiSaver can do some good – so long as the right defaults are set. If default options are not sticky, then opt-out schemes do no harm.And we can think of plenty of areas where existing harder forms of paternalism could be rolled back in favour of approaches that pay at least a little respect to individual choice. Rather than ban gambling, why not allow problem gamblers to sign onto a list of people forbidden from entering casinos – with a fortnight’s stand-down period to get off of the list? Drug prohibition could be replaced with an opt-in scheme for those wishing to use illegal substances. And rigid health and safety regimes could be replaced with nudge-based alternatives allowing more choice.

But it is hard to stop bad interpretations of good ideas. Poor interpretations of Lord Keynes’ work led TW Hutchisan to write a book contrasting Keynes and the Keynesians. About 25 years later, Ed Glaeser wrote contrasting Coase’s ideas with the Coaseans who followed. And we may yet need a book contrasting Thaler’s work with the nudge-inspired policies that followed.

Default options

It is not always easy to tell what default options count as nudges. Wales proposed a presumed-consent organ donation model, which it advertised as a prime example of "nudge" policies in action; Thaler disagreed, calling it a bad idea. He preferred prompted choice models in which people are asked explicitly to choose among options.

The UK proposed an opt-out internet pornography filter in which, by default, internet service providers must block access to pornographic materials – but with the option for opting out. Thaler rightly replied that the proposed filter is hardly a nudge, tweeting that under that scheme “Families are forced to make a choice they might prefer not to make.”

When New York’s mayor, Michael Bloomberg, proposed banning large-sized sodas at the movies as a nudge, Thaler tweeted “To state the obvious: a BAN is not a NUDGE. The opposite in fact.” But the University of Michigan’s Professor Justin Wolfers disagreed, arguing that the ability to order two small glasses of soda instead was an opt-out.

It is obviously not always easy to tell whether you are nudging or shoving somebody. And that remains the danger with Thaler’s idea in practice. Even when governments try to do right in applying it, as New Zealand did in establishing KiwiSaver, it is easy to get things wrong by setting the wrong default fund allocation.

But it is more than that. Rather than using nudge to explore ways of easing existing prohibitions, policymakers have taken the opportunity to expand paternalism’s scope – albeit sometimes with opt-out clauses.

We perhaps need a better behavioural psychology of existing policy-making processes before empowering the bureaus with new nudge-based tools. And we should worry as well about behavioural biases among voters at the ballot box. Thaler’s ideas of behavioural biases apply most strongly in areas where choice is infrequent and where it is difficult for the person making the choice to tell whether it was really the right one. Voters might then need a few nudges as well.

Thaler’s prize remains richly deserved. But be careful with the ideas in practice.