A Misconception About Income Mobility

It is often touted as a fact that it’s far less likely for someone to move out of the lowest income quintile than it is for someone in the middle class, and that it’s far less likely for someone at the top to move out than someone in the middle.

This is a fact. But, it’s a totally useless fact.

Consider this table:

The bold numbers show this fact. The table is drawn from this post by Alan Reynolds writing at Cato at Liberty. David Henderson made a similar point on Econlog. Reynolds is generally regarded as a conservative who is too partisan. Henderson is merely a Libertarian. I found there explanations a tad weak.

For my part, what we need to be looking at is the numbers just above or below the diagonal. Like so:

69	22	5	2	1
19	49	24	7	2
7	21	45	23	4
3	7	22	50	18
2	1	4	18	78

I’ve just repeated the numbers here (and rounded for simplicity). If we’re concerned about downward mobility, we need to highlight the 4 cells where it’s possible for someone to go down. When I look at those 4, I don’t see much evidence that downward income mobility depends much on your current position.

Doing the same for upward mobility I get:

69	22	5	2	1
19	49	24	7	2
7	21	45	23	4
3	7	22	50	18
2	1	4	18	78

Again, there's a small difference in the numbers, but not much.