Monday, July 22, 2013

What about from riches to rags?

This is quite an interesting piece on the relationship between parent income and (adult) child income, segmented by geographic region.  This is done on a quintile basis (and for the lowest quintile on a decile basis).  The lower endpoint of the top quintile is $70,000 in income by age 30 or $100,000 by age 45, where income is measured in 2012 dollars.  I don't think I made that age 30 number, though that would have been 1985 and the inflation adjustment is hard to figure (also I really don't remember my salary then).  I did make the age 45 number.

There must be some regression to the mean in this stuff.  In other words, if parent income didn't matter at all, then irrespective of parent income the child's expected place in the income distribution would have to be the 50th percentile. That extreme presumably represents "perfect income mobility."  The opposite extreme is where the child's place in the income distribution is exactly the same as the parent's.  That would be the no mobility case.  Reality is some (convex) combination of the two.

If as a parent you are in the top quintile (not the focus of this piece) and there is some mobility, then it is possible for your child to backslide in the income distribution and thereby end up in a lower quintile.  This backsliding doesn't get much attention at all in our discourse on income inequality.  Think of recent college grads who haven't found a job and have returned to living at home.  As I wrote to one of my economist friends earlier this morning, nobody really knows if now we are in the trough of an elongated business cycle, or whether this is the new normal for the labor market.  There is a lot of economic research that suggests graduates who don't find employment after graduation will have substantially lower lifetime earnings.  This is the group for whom the backslide possibility seems most likely.  The data that the research group has collected indicates about an 8% chance of backslide from the top quintile to the the bottom quintile and about a 13% chance of backslide from the top quintile to next to the bottom quintile.

There is not a perfect equivalence here but for simplicity the bottom quintile might be thought of as those living under the Poverty Line and the next quintile those who are the "Near Poor."  (I think there are some Near Poor in the middle quintile and there may be some poor in the second quintile.)  So the sort of backsliding I'm taking about here is not from upper middle class to middle class.  That transition is happening too, with some likelihood, but is not as much of a concern.  Above some minimum income threshold, doing meaningful work may be more important than generating higher income.  Below that income threshold, however, the ingredients for having a decent life are quite possibly absent, in which case earning more is all that matters.

The focus of the research, of course, is on bottom quintile and what the prospects are for their children.  It is an important question, clearly.  The prospect of mobility shows the system is fair and encourages people to play by the rules of the game.  Lack of mobility does the reverse. 

Here are two other caveats.  If the poor are mainly recent immigrants who have come to the U.S. to pursue the American dream, that may be more acceptable in the national psyche, as it is consistent with the myths about income distribution that we grew up with.  However, if the alternative is true, where the poor have been Americans for many generations but the parents were also poor as were the grandparents, then poverty starts to look like an "absorbing state."  It would be interesting to determine whether the geographic variations in mobility that the research finds can be at least partly explained by one or the other of these.

The other caveat is that by using relative ranking in the distribution only and then doing so with a high degree of aggregation (quintiles) the situation appears more symmetric than it really is.  In particular the income bands for each quintile are relatively narrow, except for the top quintile where the income band is huge.  One might redo the exercise but on a percentile basis or by having fixed income bands, say in $10,000 increments, and then coming up with the transition matrix for that.  Were that done I suspect the results would show less mobility than what is indicated here.  (Crossing the line to the 80th percentile might be reasonably likely, but to the 90th percentile much less likely, ditto from the 90th percentile to the 95th percentile, etc.)  The downward mobility risks are probably just as skewed. 

Let me close with a point I've made previously.  As a parent, you'd like to minimize the backsliding risk for your children.  This is why you try to make sure they've had a good education.  That gives self-protection against the back sliding risk.  But there are clearly limits to how much self-protection can be attained this way.  It is why parents in the top two quintiles will come to care for the welfare of those in the bottom three quintiles.  In turn, that may provide a realistic impetus for lessening income inequality in our society. 

No comments: