The EU referendum raises a longstanding question in welfare economics discussed, for example, in this paper (pdf) by Daniel Hausman: why should we pay attention to people's preferences?

The standard answer is that people know what is best for themselves -or at least that their mistakes cancel out - and so satisfying people's preferences maximizes aggregate welfare. As Jeremy Bentham wrote, "no man can be so good a judge as the man himself, what it is gives him pleasure or displeasure."

However, whilst the idea that people are good judges is valid in some contexts, it seems assuredly untrue of the EU debate. There are three problems here*:

- Voters are wrong about the basic facts. For example, they over-estimate the number of EU migrants in the UK by a factor of three.

- Some of their views are shaped (pdf) by cognitive biases. A form of halo effect has bred hostility to "elites": because the Establishment has been wrong about many things, voters don't trust them even when they correctly warn of the costs of Brexit. The post hoc ergo propter hoc fallacy means voters over-estimate the costs of immigration, and blame immigrants for problems caused by the financial crisis and austerity - a process perhaps exacerbated by the "bad begets bad" heuristic: voters unsettled by the uncertainty caused by immigration also believe immigration is bad for the economy. And people's tendency to take risks when they have lost can cause them to become reckless.

- Many voters are simply angry and spiteful. 61% say they'd accept a weaker economy as the cost of reducing immigration**. As Dan Davies tweeted, we are a nation of 60-year-olds in crap towns complaining about GP appointments.

The problem is that these defective preferences are systematic. It is those who have lost from economic and social change, or who feel discomfited by it, who are most hostile to the EU.

Now, in saying all this I am not necessarily opposing referendums generally. In principle, it should be possible to have institutions which promote deliberative democracy, in which the public have a voice whilst ill-informed or nasty preferences are filtered out.

However, our referendum contains no such filters. Quite the opposite. Much of the media seems to be actively selecting in favour of the worst arguments. The BBC, to its shame, is complicit in this. Not only are its main reports impartial between lies and truth, but it also gives the oxygen of publicity to charlatans and anti-democratic neofascists rather than more thoughtful Brexiters such as Daniel Hannan or Andrew Lilico.

Herein lies a paradox. David Cameron set up the "nudge unit" because he was aware that people are sometimes irrational. And yet in calling a referendum he has unleashed the worst irrationalities of many voters.

As I write, we don't know the precise circumstances that led to the murder of Jo Cox. But we do know that nasty rhetoric can have nasty consequences. I suspect that heightened anti-immigrant sentiment means that foreigners feel less safe walking our streets. Such insecurity is the direct result of Cameron's decision to use a referendum in a - perhaps ill-judged - attempt to solve what is primarily a problem of internal Tory party politics. For this, he deserves our contempt.

* There might be a fourth: that some people's preferences are more intense than others. I'm leaving this aside.

** Far fewer are willing to pay a personal price, but this might be another example of a cognitive bias: wishful thinking leads folk to think that the costs of a weaker economy won't be paid by them personally.