We are often told that inequality is a big problem. Perhaps it doesn't itself slow economic growth, cause social and political instability, or set off damaging "path dependence". Perhaps CEO pay is actually reasonable! Whether or not that's true, the argument runs, people just dislike it, and since preferences are what matter, we should cut it. But it turns out people don't really dislike inequality, they just dislike unfairness.

Or so say three recent papers—two economic experiments and a review in Nature: Human Behaviour. The first (pdf) finds that people in lab experiments have a tendency, when given the option, to burn others' income when inequality rises. As an egalitarian might predict, it is low income subjects who are the burners, and high income subjects who are the burnees. But burning is substantially higher when the inequality is down to unfair rule tweaking, and lower when it's down to extra effort.

So far, so arcane, an egalitarian might argue. Fair enough. But a second study, this time conducted in India, and accepted for publication in the QJE, finds precisely the same thing. This time the authors conduct a month-long experiment with manufacturing workers. People don't like inequality when it doesn't seem to come from unequal productivity: they don't turn up for work and they work less hard. But when it clearly comes from higher output the workers don't seem to mind.

Finally, the Nature paper:

Drawing upon laboratory studies, cross-cultural research, and experiments with babies and young children, we argue that humans naturally favour fair distributions, not equal ones, and that when fairness and equality clash, people prefer fair inequality over unfair equality.

The recent turn against inequality may not be evidence against this trend. Recent rises in wealth inequality seem to be driven by gaps in housing wealth, exaggerated by strict rules on housing supply. If this is true, then inequality is coming from unfair rules that rig things against one group and in favour of another. If that in turn is true then the "solution" to inequality really is just to fix the problem rules that cause some of it, but leave the bits that come from fair rules and uneven inputs alone. And, who knows, this might have some pretty nice side benefits...