Which sort of inequalities should we tolerate and which should we remedy?

One respectable philosophical position here is that of luck egalitarianism. This distinguishes between brute luck and option luck: option luck describes the outcome of gambles we have chosen, whereas brute luck describes luck beyond our control. The luck egalitarian would equalize the outcome of brute luck, but not that of option luck. He would compensate someone who loses their job in a recession, say, but not someone who loses their money in a casino.

You might not agree with this view, but it's surely tenable.

Here, then, is a surprise. In a laboratory experiment, almost nobody is a luck egalitarian.

Johanna Mollerstrom, Bjorn-Atle Reme and Erik Sorenson gave people $24 and told them they faced three equiprobable scenarios. In A, they would keep the money. In B and C, they would lose it. However, they were given the choice to buy insurance against scenario B, but not C. Scenario B is therefore a case of bad option luck, whereas scenario C is bad brute luck.

152 subjects at Harvard Business School were then faced with pairs of subjects who had experienced these scenarios and given the choice to equalize incomes within the pair. And here's the strange thing. Whilst there was a decent minority of strict egalitarians, who equalized in all scenarios, and of "libertarians" who never equalized, only one of the 152 subjects was a consistent luck egalitarian.

Instead, many subjects were what they call choice compensators. They were more likely to compensate people who suffered scenario C if they had bought insurance against scenario B than if they hadn't - even though that choice made no difference to their loss.

There might be a justification for this. Buying insurance signals that you are risk-averse, so losses hurt more - and subjects are keener to compensate people for if they face a painful loss.

However, something nastier might be going on. Maybe subjects want to reward "good" behaviour - buying insurance - even if that behaviour is irrelevant for the outcome in question. Worse still, some might be looking to blame the victim, and see not buying insurance as a sign of recklessness.

It's in this context that a lot of political rhetoric makes sense. If you think the poor are "hard-working families", you'll be think they made the right choices and so deserve redistribution. If instead you think they are lazy scroungers, you won't. Those of us who point out that laziness and scrounging might be irrelevant for actual outcomes are missing the point; people are looking for a moral basis for redistribution, not for a causal link between behaviour and outcomes.

Now, in these lab experiments, subjects didn't have to use their own money to equalize. In the real world, where there's a self-serving bias, people will be even keener to find reasons not to redistribute.

We should, however, remember that there is a difference between what's right and what's popular. The fact that almost nobody is a luck egalitarian is not sufficient to show that luck egalitarianism is wrong.