People aren't as daft as some behavioural economists would have us believe. This is the message of a new paper by Olivier Bargain and colleagues. They studied the choices of British workers of how long to work, and found that:

People behave on average as if they were maximizing SWB...A majority of individuals actually make decisions that are in line with the maximization of income-leisure satisfaction.

Of course, there are exceptions to this. But many of these are because people are constrained by labour market institutions and policies: some people work too much because they can't find decent part-time work, and others too little because they can't find work at all.

Granted, there's a caveat here. Perhaps people's preferences have adapted to their circumstances and prior choices. Nevertheless, this is evidence that people aren't as stupid as paternalists claim. Perhaps workers are badly off not so much because they make bad choices, but more because they are the victims of inequality and bad policies.

Bargain's finding is consistent with some other evidence. For example, Marc Fleurbaey and Hannes Schwandt found that only a minority of Americans could think of easy changes to their lives that would increase their subjective well-being - which is what we'd expect if most people were maximizing their well-being. And people's spending decisions do seem rational and forward-looking on average, in the important sense that ratios of consumer spending to wealth predict subsequent stock market and economic fluctuations.

How can we reconcile all this with the claim that voters are ill-informed and irrational?

Easy.

Our spending and working decisions are made regularly, so we can learn from experience to make better ones. And we have a strong incentive to do so: if we're wrong we end up poor and unhappy.

In voting, however, neither of these conditions is met. Even general elections are one-off events (let alone referenda) so we don't get the relevant feedback that would help us make better decisions in future. And we've little incentive to do so; if I make a stupid decision in the voting booth, I'll not suffer.

This brings me a paradox about behavioural economics, at least as it is used in politics. It invites us to distrust people when they should be trusted - when they make regular decisions about their everyday lives. But it trusts people when they should be distrusted - for example by deferring too much to unfiltered and unreflective public opinion*.

Downing Street's "nudge unit" should have worried less about the irrationalities of consumers, and more about the reckless overconfidence of David Cameron which gave us austerity and Brexit.

* There's a stronger case for worker democracy than there is for referendums.