As the middle class shrinks, where are formerly middle-class people going?

by Ryan Streeter on May 12, 2016. Follow Ryan on Twitter.

In this Wonkblog overview of the new Pew report on the diminishing middle class, the authors write:

As the middle class has shrunk, Pew points out, the lower and upper classes in America have grown in size and significance. In some metros, the middle class is dwindling primarily because families are falling out of it and into the lower class…But in other places, the shrinking middle class is actually a sign of economic gains, as more people who were once middle class have joined the ranks at the top.

The latter point is one we don’t hear as much about in our public debates, namely that one reason the middle class is getting smaller is that a significant number of formerly middle class families and/or their children are getting wealthier. It’s at least as significant as the other problem whereby people fall out of the middle class into the lower ranks:

In total, 172 of [the] 229 metros [in the study] saw a growing share of households in the upper-income tier. About as many — 160 — saw a growing share at the bottom. And 108 experienced both: The middle class shrank as the ranks of both the poor and the rich grew.

In some cases the two trends appear to be directly related. For instance, look at the green (upper income) and blue (middle income) lines in this infographic:

Screen Shot 2016-05-12 at 12.12.28 PM

Understanding the reasons for the growth in the upper tier seems just as important as understanding the forces shaping the lower tier. The trend is more widespread than the infographic might suggest. You see the same blue-green trends above in cities such as Minneapolis, Midland, TX, and New Orleans, not just the super-charged areas in and around New York, DC, and San Francisco. Presumably, household formation together with growing percentages of metro economies favoring specialized and higher forms of education are a big part of this. But getting into the gritty details region by region would help us understand more about that thesis, and how correct it is in varying places.