The ONS reports that inequality, measured by Gini coefficients for both pre-tax and post-tax incomes, has fallen slightly in the last ten years. I fear this is an example of how statistics sometime don't tell us very much.

- Even if inequality is now falling, the damage done by the previous rise in it still lingers. To paraphrase Joseph Schumpeter, if a man has been hit by a lorry you don't restore him to health by reversing the lorry. And this damage is considerable. It consists not just of slower economic growth but of increased distrust and coarsened politics: the oft-heard allegations that "elites" are out of touch with the "people" are the product of that earlier rise in inequality.

- ONS data tell us nothing about the incomes of the super-rich. But we know that the salaries of CEOs of big companies have risen far faster than average incomes, as have incomes of richer bankers. Rising incomes of the super-rich are consistent with a flat or falling Gini if inequalities between the moderately well-off and moderately poor narrow - as has happened.

- The same Gini coefficient can describe very different economies. A bourgeois society in which many are doing OK whilst only a few are very poor or very rich can have the same Gini as a winner-take-all society in which some have massive incomes whilst there's income support for the poor but a hollowed-out middle class. But these will be different societies with (ultimately) different cultures. I suspect that the stable-ish Gini hides the fact that we've shifted from the former to the latter.

- Gini coefficients tell us nothing about how inequality arises. The same Gini might arise from people happily paying a man for his great talents (as in Nozick's Wilt Chamberlain example) or from that man ripping people off. What troubles some of us is that our current inequality is due a lot to the latter.

- It's not just inequality of income that matters. Inequalities of status and power also do - not least because they contribute to bad decision-making and degrading working conditions. These are very much still with us.

My point here is a trivial one. We should ask of all statistics: what exactly is it that these are telling us, and what aren't they? I fear that inequality statistics might not be telling us much. Of course, my concerns about the causes and effects of inequality might be mistaken. There's a debate to be had here, but it won't be settled by the ONS's data.