Race as a bundle and its implications

As I mentioned in my last post, I have been given the topic of race increased thought recently.

One of the recent developments in political science has been thinking of race not as a dichotomous variable, but as a bundle of related but distinct characteristics. Race is not simply phenotype, but a mixture of such things as one’s dialect, diet, and socioeconomic status among other things.

RaceBundle

The idea to me seems obvious, which makes me inclined to believe it. The thing is, if we take this broader approach to what race is, what are the implications for prior work not only in regards to race but the effect of demographic characteristics generally.

Race is already difficult to conduct research in because it is assigned at birth which makes it difficult to manipulate and which influences other characteristics we would ordinarily ‘control’ for in statistical analysis. To my knowledge there isn’t a ‘race ray’ that we can use to randomly assign being ‘black’ in an experiment. Tracing causality is possible, but difficult enough even in ideal situations.

Take for example the gender wage gap argument. When you control for education, presence of children, and other characteristics the gap in wages between males and females vanishes. However many of these characteristics are impacted by one’s gender. While females are not discriminated against ceteris paribus, being female does increase one’s likelihood of having to be the primary care taker for children and has historically decreased educational outcomes. In this broader sense there is a gender wage gap.

What can be done about it though? Men can try to share more of the house duties with their wives, but my general observation in life has been that children prefer being cared for by their mothers over their fathers. Should we try to do something about it? Are there advantages to one member of the household specializing in housework?

Or, if you prefer to think of the question purely in regards to race let us consider crime rates by race. I am not convinced that blacks have any higher propensity to crime than whites. However blacks are more likely to grow up in poverty and have lower educational outcomes than other races, which in turn leads to higher crime rates statistically speaking. Where should the arrow of causality be pointed towards: race, education, socioeconomic status?

Race is a difficult concept to think about. However it is precisely the difficulty with discussing it which begs that it be thought about more. I believe we liberals have a particular duty to think about race more because if we don’t then our ideological rivals will continue to dominate the conversation.

See here for an un-gated draft of the relevant paper: Sen, Maya, and Omar Wasow. “Race as a Bundle of Sticks: Designs that Estimate Effects of Seemingly Immutable Characteristics.” Annual Review of Political Science 19 (2016): 499-522.