I've been having a hard time getting my head around the 'hollowing out of the middle class' theme that were seeing so much of. What, exactly, does that mean? And is it actually happening? In a Maclean's post a couple of months ago, I tried looking at it in terms of the proportion of the income distribution that was within a certain distance of the median, and this is what I got:
If the middle class were hollowing out, I would have expected to see a trend to smaller proportions of incomes 'close' to the median. But I don't see that trend in the data.
Soon after that post, Kevin Milligan brought this paper to my attention, which includes these remarkable graphs:
The lower right-hand graphs tell a pretty stark tale: for both sexes in the 25 years between 1981 and 2006, the higher you were in the income distribution, the faster your wages were rising. For men, the picture is particularly striking: the only income growth occurred for those above the median. The dynamics are pretty clear: these income distributions were spreading out.
So where does that first graph - which is also based on the same census Public Use Micro Datafiles - come from? The answer seems to lie in the relative positions of the bottom-right profiles: for a given percentile, the increase in women's wages was systematically higher than for men. Basically, the men who drifted below the median were replaced by women whose incomes were initially lower, but had grown faster.
As I typed that last sentence, it occurred to me that I should go back and check the gender composition of the definitions of the middle class in that first graph. Sure enough, the female share of the middle class (as defined in the graphs) has increased over time. What surprised me is that the female share of the middle class was always greater than 50%.
I'm not sure what to make of this. There is definitely something going very badly wrong with low-skilled men, but I don't think it's helpful to think of it as the hollowing-out of the middle class.