Black immigrants are upwardly mobile

The other day, I noticed something disturbing in a graph from a Brookings report on immigrant mobility:

We see that Hispanics are strongly upwardly mobile from the first to the second generation. Asians are slightly upwardly mobile, but from a pretty high base. Those are both good news. But black immigrants, on average, appear to show downward mobility.

Why would black immigrants be downwardly mobile? I posed the question on Twitter. A smart person called Abraham Bloodshack immediately tweeted this to me:

generational effects? i.e., recent increase in African migration could mean second gen are all quite early in careers

That was smart. We'll follow up on that later. But first, let's review some possible explanations for the mobility disparity:

1. Household size decrease. 1st gen. African immigrant families are probably really big, since Africa is a super-high-fertility place in general, while 2nd gen. families probably have drastically lower fertility.

2. Cohort effect. Recent changes in immigration composition might account for the effect. The average age difference that Abraham Bloodshack mentioned is also a kind of cohort effect. Age differences would affect the black immigrant average much more than the Hispanic or Asian immigrant average if African immigrant families, like the typical African family, are extremely large.

3. Downward assimilation. With many immigrant groups we see the 2nd generation picking up a lot of "bad" behaviors - or at least "bad" for earning power - from their decadent rich-world peers. These include things like not getting married, sponging off parents, and getting involved in the underground economy. 2nd-generation black Americans might be especially susceptible to this sort of thing. (And yes, I know "downward" might be a loaded word; if you want to sponge off your parents and play League of Legends, more power to you.*)

4. Racism. Negative attitudes toward African-Americans might not apply to people with African or Carribean accents, but might be applied toward their more American-sounding kids. (Update: See this excellent comment for more.)

I did a bit of digging on the ol' internet, and turned up this Tyler Cowen post on the subject, from two years ago. Cowen links to two papers (paper 1, paper 2) by Alison Rauh, a Chicago econ PhD, now a research associate at Cornerstone. 

The papers look at personal income, so we don't have to worry about the household size issue. They find broadly the same average income decrease as the Brookings graph, though to a lesser extent. Rauh's first paper attributes the difference to "idleness", her word for "being out of the labor force". Conditional on having a job, 2nd-generation black Americans earn a lot more than 1st-generation - for men, 29 percent more. That's roughly comparable to the average Hispanic increase from the Brookings graph, but from a much higher base. But so many 2nd-generation black Americans are out of the labor force that the overall average income goes down!

That would seem to point to the "downward assimilation" story, perhaps with some racism mixed in. Tyler goes for downward assimilation:

I take this to be a “peer effects are really really important” paper, namely that many of the virtues of immigrant culture are swallowed as the second generation assimilates.  

But this isn't the whole story. In her second paper, Rauh looks at what happens when the generations are adjusted for age, as Abraham Bloodshack might have suggested doing. Here is what she finds:

Note, however, that the average second generation black is more than 8 years younger than the first generation immigrant. Since earnings increase steeply until the mid forties, column 5 uses inverse probability weighting to equalize the age distribution of the first and second generation. Now sons of black immigrants earn $3000 or 8% more than the average first generation black immigrant. The fraction of second generation blacks with a college degree is 35% and therefore 4 percentage point higher than those of the first generation...For women both of these trends seem to be even more pronounced (Panel B). The second generation, once adjusted to have the same age distribution as the first generation, has an earnings premium of $8,600 over native blacks, $6,700 over first generation, and $3,600 over whites. (emphasis mine)

Lesson: Always read through the papers, don't just skim the Abstract/Intro/Conclusion! 

So if we adjust for age, we see that black immigrants are upwardly mobile, in terms of both income and education. And that upward mobility is from a decently high base for income and a very high base for education. 

In other words, the Brookings graph tells us the wrong story! A lot of those 2nd generation black Americans are either in college, or just starting at the bottom of the career ladder. They will eventually make a lot more money. This is great news.

But although black immigrants are upwardly mobile, they are not as upwardly mobile as they should be. Rauh observes that high incarceration rates play a large part in the fraction of 2nd-generation black Americans not in the labor force. Marriage rates are also much lower. 2nd-generation men are also much less upwardly mobile than 2nd-generation women.

These facts all point to a cultural effect - societal racism and/or downward assimilation, in some combination. Black immigrants are thriving in our society overall, but it's in spite of some headwinds.

Anyway, let this be a lesson: Before you go looking for theories to explain a fact, make sure you've gotten the fact right in the first place!


* Haha, I'm kidding of course. League of Legends sucks.