The one-drop ethnicity
| Gabriel |
One of the findings from the 2010 Census that has been the subject of much discussion is the large increase in the Latino population since 2000, most of which is from natural increase rather than immigration. This is obviously a real trend and is to be expected as Latinos tend to be younger (read: many are in their prime fertility years) and to have higher total fertility than do Anglos. However I was left wondering that some of this increase might be somewhat exaggerated by how we collect the data.
As is described in Cristina Mora’s fantastic dissertation, during the 1970s there were all sorts of twists and contingencies as “Hispanic” was institutionalized as a category. One of the oddities is that “Hispanic” is not a race, but an “ethnicity” that cross-cuts race. A few decades later, we switched to a “check all that apply” racial taxonomy, but since Hispanic is not a race a Census respondent can check both “black” and “Asian” but not both “Hispanic” and “not Hispanic.” This effectively creates a one-drop rule for Hispanic ethnicity that creates systematic biases in our understanding of Hispanics and has the ironic effect of blinding us to some important ways that this always somewhat artificial distinction is blurring.
My own household consists of an adult Anglo (me), an adult Hispanic white (my wife), and our daughter who the Census Bureau considers a Hispanic white, just the same as if both her parents were Latino. In contrast if my wife were not Mexican but African American, the Census Bureau would make salient this admixture by coding our daughter as both “white” and “black.” Nor is my sort of marriage especially uncommon. In the 2000 Census, 28% of current marriages involving at least one Latino spouse also involved a non-Latino spouse. Thus as a quick and dirty ballpark estimate* we can say that something like a quarter to a third of the natural increase in Latinos involves children who are in some sense “half Latino,” but whom the Census records as simply “Latino.”
Thus implicitly our data have a “one-drop” rule for Latino ethnicity, even as we have moved away from comparable assumptions for racial distinctions. The theoretical question is whether this is an assumption that reflects social reality. In a sense it may be performative — if social institutions treat the offspring of Latino/non-Latino unions as simply “Latino” this may actually push such people towards identifying as Latino. (See a series of fascinating articles by Saperstein and Penner on social institutions and longitudinal racial identity). However I am rather skeptical that a one-drop rule for Latino necessarily reflects reality. Latino has always been a hodge-podge category defined almost entirely by a shared history of the Spanish language and we know that Latinos rapidly lose Spanish language ability when they do not grow up in households with an immigrant member. This Spanish-language loss is especially pronounced for people with a non-Latino parent and absent this defining characteristic we might expect many people who are half Latino and half Anglo to just sort of blend in, with an ethnic identity that is as unmarked and low salience as someone who is Irish/German or Levantine/Scots-Irish. Another way to say this is that we need to be careful about reifying our current arbitrary and historically contingent understanding of ethnicity and understand that a lot of what we see as growth among the Latino minority is really part of an emerging 21st century “beige” majority, no matter what they check on close-form surveys rooted in the politics, economics, and social sciences of the 1970s.
[Update: Also see similar thoughts from Matt Yglesias].
* Note that extrapolating marriages to births assumes that (a) the patterns are roughly comparable for non-marital fertility and (b) endogamous Latinos have similar fertility to exogamous Latinos.