Does Maternal Race Impact Adult Outcomes?

Last week, I attended an interesting seminar about a new paper in the works by Peter Arcidiacono and Seth Sanders from Duke University titled Maternal Race and Black Outcomes. Peter Arcidiacono presented the research and did a great job handling a multitude of questions and challenges to the paper. I considered applying to Duke's MA Program, so I was excited to have a Duke professor come to present his work. Anyway, here is the abstract:

Differences between blacks and whites in test scores and labor market outcomes are stark. While much catchup occurred post-Civil rights, convergence has slowed. We examine how differences across education and labor market outcomes vary by maternal race and own race with identification coming from mixed-race families. While black students with white mothers come from families with similar demographics to black students with black mothers, their education and labor market outcomes are very different. There are no significant differences in test scores, grades, college graduation, and wages between black and white males with white mothers, yet large differences exist between these groups and black males with black mothers. These results are insensitive to alternative measures of own-race, using skin tone instead of own race, and including school fixed effects.

The paper is somewhat controversial depending on how the results are interpreted. One caveat is that these results are only statistically significant for boys, although girls' outcomes follow a similar pattern. Several major points stuck with me from the seminar.

1) Although black children with white mothers are demographically more similar to black children with black mothers, black children with white mothers have adolescent and adult outcomes that are more similar to outcomes of white children.

2) Skin tone of the child was insignificant in terms of impact on adult outcomes when maternal race of is accounted for.

3) These results held for a variety of outcomes including wages, college completion, test scores, etc.

As you can see, this paper could likely take a lot of heat simply due to the issue it is trying to tackle. The fact that the mother's racial background is statistically more important than than her child's is a novel idea in this field. Is it culture, social networks and access, the inter-generational impact of discrimination on families, etc. or a combination of many factors? What mechanism is at work here? Here is the authors' conclusion:

That fact that the results seem to be different depending upon whether race is coded as race of the mother or race of the child is suggestive that race of the mother may have an affect on outcomes distinct from its effect through the race of the child. This pattern is supported by the findings in this paper which points towards differential investment patterns across mothers of different races.