It is likely that the apparent size of white-black educational and economic gaps will shrink over the next few decades, simply because the percentage of black Americans who are immigrants and children of immigrants has increased in recent decades, and these groups perform better than the children of native born black Americans on a number of economic and educational measures. Whether this represents a real amelioration of racial inequality or simply an obscuring of it is, I think, not an easy question to answer. My belief is that the changes in the composition of the black population, away from the descendants of slaves kept in America and towards West Indians, recent African immigrants and multi-racial individuals, are going ultimately to render the system of affirmative action preferences in selective colleges totally untenable, but that the colleges themselves will delay that reconciliation as long as possible, for self-serving reasons.
To start with: at least some groups of recent African and West Indian immigrants have above median (and above median-white) household incomes:
Among black adults with wages in the top quarter, the percent of foreign born individuals has more than doubled in the last two decades:
Among black children in households with an earner in the top quartile, the percent with foreign-born parents has also roughly doubled:
Among the young adults in the National Longitudinal Survey: 1997 who attended high school in the United States, the percent of foreign born blacks with a BA or higher is more than twice the percent of native born blacks:
For this same large group of older Millennials (born between 1982 and 1986), this difference in educational proficiency was already apparent by the time they took a standardized math and reading test in 1998; higher-scorers were disproportionately not native born:
At a county level, the average grade 3-8 performance of black students in school districts with more foreign born black students is significantly higher:
Among current college-aged college students who were in 9th grade in 2009, the top 10% in GPA and math scores are almost 3 times as likely to be an immigrant or child of immigrants for boys and almost 4 times as likely for girls, relative to the mean of all black students:
Even conditioning on high school GPA and test scores, girls who are immigrants or the daughters of immigrants are more likely to be currently (as of 2016) enrolled in a selective college:
The causes of these educational and economic gaps between immigrant and non-immigrant black families are multiple and hard to disentangle. Marriage is probably one of them; foreign-born black mothers of teenagers are somewhat over twice as likely to be married (CPS 2017). The selectivity of emigration (and the relative difficulty African and West Indian immigrants have in coming here, relative to Latin American immigrants) probably partially also explains it. Lastly, it is likely that differential fertility patterns, in which poorer and lower scoring black native born women had more children than richer and higher scoring black women, starting at least among the cohort born in the 1960s, made racial gaps among native born adults larger.
These patterns of differential fertility appear to be recapitulated among more recent cohorts (as well as among recent cohorts of native born whites), which might not make racial gaps themselves still larger but would tend to depress the performance of the children of native born Americans overall relative to newer immigrants.
In 2004, Henry Louis Gates and Lani Guinier pointed out that only about a third of black Harvard undergraduates had four grandparents who were the descendants of American slaves. My guess is that, given these overall trends towards foreign born black students (along with trends towards greater intermarriage) that percentage has only diminished since then. Whether or not this will serve the ostensible purposes of affirmative action is an open question, but not one I would trust Harvard to answer.
Update: the graphs above show the probability of being an immigrant/child of immigrants conditional on a certain level of high school performance, for black students who were in 9th grade in 2009. Here are average weighted GPA and test scores for this same group, conditional on being an immigrant/child of immigrants (vs US born teens with US born parents):