Why do people tolerate inequality? A new paper by Karl Ove Moene and colleagues provides a new explanation: it's because a small role for merit in income inequality causes a big drop in demand for redistribution.

They showed this in some experiments. Over 1000 people were asked to observe pairs of people working on a task for ten minutes, with earnings divided between the pairs in different ways. The spectators were then invited to redistribute some or all of those earnings.

When told that one worker got everything and another nothing simply because of a lottery, 68% of spectators chose to equalize earnings. However, when told that everything went to the worker who had done best, only 28% equalized earnings. This is consistent with most people being luck egalitarians; they want to eliminate inequalities that are due to luck, but accept those due to differences in merit or effort.

But here's the quirk. Other subjects were told that earnings were divided 90% by luck and 10% by merit. You'd expect these to redistribute almost as much as those subjects faced with inequalities due entirely to luck. But they didn't. Only 22% chose to equalize earnings. "A little bit of merit makes people significantly more inequality accepting" conclude the authors.

A tiny role for merit reduces demand for redistribution as much as a big role does.

Of course, sceptics will question the external validity of this result. I suspect, though, that in the real world it might be amplified by the hindsight bias. When we see successful people, we infer that they must have done something to earn that success and so we downgrade the importance of dumb luck.

This, of course, is not the only reason why people tolerate inequality. In fact, this phenomenon might be overdetermined. Here are some other theories:

- Ignorance. "People dramatically underestimate actual pay inequality" say Sorapap Kiatpongsan and Michael Norton. Ratios of CEO pay to that of unskilled workers are far higher than people think, and even higher than they believe desirable.

- Anchoring. Kris-Stella Trump has shown that our idea of what is an acceptable level of inequality is shaped by actual inequality. As inequality rises, therefore, so too does our view of tolerable inequality. In the same vein, Jimmy Charitie, Raymond Fisman and Ilyana Kuziemko have shown how preferences for redistribution are influenced (pdf) by reference points. If people expect incomes to be equal, they'll demand more redistribution than if they expect unequal incomes.

- System justification. John Jost and colleagues have shown (pdf) how people tell themselves stories to justify inequality, such as blaming the victim or believing injustice to be natural.

I say all this to make a point to both left and right. To the right, I suggest that a lack of demand for redistribution need not be evidence that inequalities are just. It might instead be that attitudes are distorted by cognitive biases. To the left, I suggest that we worry less about people being influenced by the right-wing media - the election result suggests such influence is less than previously thought - and more about how pro-inequality ideology can emerge endogenously from the interplay of capitalism and cognitive biases.