Inequality has fallen in the UK - which might be worrying.

This sounds like an odd thing to say. But it's the natural implication of what Ben Chu rightly says. He notes that whilst the Gini coefficient hasn't changed much since the early 90s, the share of income going to the top 1% has risen.

However, we can think of the Gini coefficient as a single measure of all inequalities: the gap between the top 1% and the second percentile, plus the gap between the second and third percentiles, and so on. For this reason, the same Gini coefficient can describe very different societies.

The Gini coefficient can therefore be stable if one inequality increases and another diminishes. As Ben says, inequality has risen in the sense that the top 1% has gotten relatively better off. But it follows, therefore, that some other inequality has fallen.

My chart, taken from the IFS, shows one of these inequalities - the ratio of the 80th to 20th percentiles, but neighbouring ratios such as the 70/30 show a similar pattern.

Now, these figures refer to incomes after tax and tax credits but before housing costs and they are adjusted for household size. A childless couple with a disposable income of £288 per week in 2012-13 (the latest date available) was at the 20th percentile, whilst £689 per week got the couple onto the 80th percentile; the latter is equivalent to a pre-tax income of just under £50,000.

What's going on here, I suspect, are two separate developments.The 20th percentile are often lower-wage workers, who have benefited from tax credits: since 1996-97 their incomes have risen by 23.7% - more than any group except the 1% who got a 40% rise. However, the 80th percentile has seen growth of less than 13% - as have the 65-85th percentiles. This, I suspect reflects a combination of the relative decline of the lower middle-class and job polarization hurting white-collar workers: you can picture the 80th percentile couple as a man on £30,000 and woman on £20,000 - or more for a couple with kids.

In fact, I suspect their relative decline might be worse than this, because working conditions for such people might also have deteriorated.

Worse still, it's possible that robotization (pdf) will worsen their relative position still further.

And here's the thing. It's these sort of people who have, in politico-speak, worked hard and played by the rules. And yet they've seen their position decline relative both to lower-paid workers and the top 1%. Their response to this, in some (many) cases has been resentment against the political class generally: when I picture a Clarkson-loving resentful white man, it's someone on the sort of decentish wage that gets him and his family into to the 65th-85th percentiles.

It is often said that a strong middle-class is necessary for a healthy, stable free society. Insofar as this is the case, then the sort of increased equality we've seen might be a problem.