A few years ago, there was a bit of a controversy after a handful of people were discovered to have changed their racial identity from white to black. The new census suggests that changes in identity are actually quite common:
Notice that the number of Hispanics that identify as two or more races rose from 3 million in 2010 to 20 million in 2020. That sort of increase is much too big to be explained solely by demographic shifts, and instead implies that millions of Hispanics who identified as white in 2010 identified as mixed race in 2020. But why?
Noah Smith has an interesting post on this issue:
When people say “race is a social construct”, this is one of the things they mean — the whole racial identify of tens of millions of people can change in just a few short years.
What’s going on? One obvious possibility is that the Trump Era made Hispanic people feel more like a racial minority. The Right’s targeting of immigrants, Trump’s well-publicized disparaging remarks about Mexicans, and violence like the 2019 El Paso shooting might have sent a message that Hispanics are not in the White club, and Census respondents might simply be acknowledging that.
Another possibility is that Hispanics don’t feel rejected by whiteness, but are choosing to reject it:
That’s probably a part of the story, but the change is so large I suspect additional factors as well. One possibility is that the census changed the way it surveyed people.
At first glance, the 2010 and 2020 census forms look pretty similar, offering Americans the choice between the same set of language and race categories. But there is one minor change that might be significant. This is from the 2010 form:
With the 2020 form, the American Indian category had some additional explanation:
I’m not sure how the average Hispanic would interpret these two questions, but I suspect that many people view the term “American Indian” as something closer to “United States Indian.” That’s how I’ve always thought of the term. Indeed ‘America’ and ‘United States’ are often used synonymously. In the 2020 form, however, the census includes examples such as “Mayan” and “Aztec”, which makes it pretty clear that the census intended “American Indian” to mean “Western hemisphere Indian”.
Most of the Hispanics who immigrate from Mexico and Central America to the US have both white and Native American ancestors. (In Mexico, they use the term “Mestizo.”) Thus it’s not surprising that when the question was clarified, many more Hispanics chose more than one race on question 9.
As is often the case, media reports of a declining white population were somewhat misleading. The US white population is probably not shrinking, at least in the sense that most people think of the term ‘shrinking’. There are not fewer white people, maybe not even fewer people who identify as white, merely fewer people who check only the white box on their census form.
I suspect these distinctions will become increasingly meaningless over time. If someone were to ask me whether my daughter was white I wouldn’t even know how to answer the question. What does it mean to say someone is white? The race of their parents? How they identify? How they are seen by the public?
The census also suggests that 26 million Hispanics identify as a race other than Asian, white, black, Native American, Hawaiian or Alaskan Native, and yet claim not to be a mixture of several races. I wonder what that “other race” actually is?
READER COMMENTS
Mark Z
Aug 18 2021 at 9:44pm
It seems like a test of this might be to look at the rate of identity shift by country of origin. Do immigrants (or descendants) from countries with fewer Mestizos still identify mostly as white, even though they could opt out? Also, was there an increase in identification as mixed race among part-black or part-Asian people (some of whom otherwise could identify as white?) I think these questions might elucidate whether your or Noah’s explanation is more likely (my thinking is, if it’s a phenomenon specific to hispanics or more pronounced among very Mestizo nations of origin, yours is probably more likely).
Noah Smith: “When people say “race is a social construct”, this is one of the things they mean — the whole racial identify of tens of millions of people can change in just a few short years.”
This is like arguing that because different people (or even the same person at different times) identify a color sometimes as a shade of blue and sometimes as a shade of green, that this implies color is a social construct. Showing that the boundaries between categories is rather arbitrary does not reflect that the variable is socially constructed, but rather that it’s actually a continuous rather than discrete variable. Sorry, I had to complain; I see this kind of non-sequitur all the time and it kind of annoys me.
Scott Sumner
Aug 18 2021 at 10:48pm
I wouldn’t say it’s either/or; I believe there are multiple factors at work.
And I agree with Smith on race being at least partly socially constructed, at least as the concept of race is used in the US. Skin color is not socially constructed, but race is about much more than just skin color.
mbka
Aug 18 2021 at 10:38pm
Scott, I find it amazing that the form doesn’t blush as it is calling Chinese, Japanese, Vietnamese different races. Why is the word race even used anymore? There is a much better word for the intended purpose of labelling a person by belonging to a group of origin or identity. That word is ethnicity. And if we’re really honest, these identity groups are a mix of cultural and biological meanings. Black is not a race. Australian Aborigenes, South Indians, Africans of all sorts, are all very dark skinned, yet completely different ethnic groups. The genetic differences within Africa itself are higher than the average difference between “Africans” and non-Africans. Ethiopians, Zulus, Igbo, Bororo, Tuareg, Haussa, I mean, where do I start? African Americans did not come from “Africa”. Descendants of slaves and recent African immigrants to the US are often vastly different in ethnicity. Just because they’re “black” doesn’t mean that they cluster to the same “African” “ethnicity”.
All these “race” labels are absurd and really just a proxy for culture. I’ve heard of “black” people from the US being labelled “white” in Africa. Which is probably more correct than the nonsensical questions on that form. My son would also have no tick box on it, and pick something at random, if he had to.
And similarly to Mark Z, I find the absurdity here isn’t so much that “race is a social construct” but that “the alternatives offered on that form make neither biological nor cultural sense, the word “race” is wrong in this context, therefore the answers are arbitrarily wrong, but wrong differently depending which year you are asking them”.
Scott Sumner
Aug 19 2021 at 12:43pm
Yes, the way Americans approach “race” is borderline insane. It might make more sense to ask about 6 regional categories of ancestry:
European
East Asian (Burma to Japan or perhaps Hawaii)
South Asian (India, Pakistan, Bangladesh, etc.)
Middle Eastern (Afghanistan to Morocco)
Sub-Saharan African
Latin American
Within each group there’d be multiple “races”, but at least it would treat various regions in a more symmetrical fashion.
It’s weird that the census asks people if they are “Hispanic”, which is a silly category that includes Spaniards, Cubans and Mayans but not Brazilians.
robc
Aug 19 2021 at 5:04pm
I was earlier going to reply with a joke that I was mixed-race, both Anglo **AND** Saxon.
I am of the “race is a social construct” belief. I know how to test for sex. Genetic tests say Male, Female, and, for rare cases, Other. There is no genetic test for race.
And even your breakdown has some fuzzy lines. Are Burmese really closer in race to Japanese than they are to Bangladeshi? And what about Nepal or Bhutan? The lines are completely artificial.
TGGP
Aug 20 2021 at 10:21am
You may not know how to test for race, for forensic pathologists have long been able to do so with skeletons. Cluster analyses of DNA show matches of over 99% with self-reported race. A recent study freaked out its authors by showing that machine learning AI could have similarly high accuracy of categorizing CT scans even when they’d been deliberately blurred so much as to just look like gray squares to human eyes.
All that being said, race is FAR fuzzier than sex. The human sex system is often said to come down to a single gene on the Y chromosome, and you either have that or you don’t. Race is a very large number of correlated genes, and any of them individually don’t really determine a category by themselves.
mbka
Aug 19 2021 at 10:55pm
Scott, yes “Hispanics” cracks me up, to begin with it I am not even sure whether in US government parlance it includes Spaniards from Spain. Does it? When I lived in California, people told me that Spaniards are not Hispanics. Logic aside.
The whole thing reveals that the question on race is actually intended as a question on culture – how will people behave. In which case we should just ask, are you
Anglo Saxon (includes AU, NZ, UK, CA) and Skandinavian
Western European (includes Northern Italy) – weakly Catholic, Protestant
Eastern European
Southern European – strongly Catholic (includes Spain and Portugal, Southern Italy and Sicily)
Arab Middle Eastern and North African
Non Arab Middle Eastern, incl. Turkey, Lebanon, Israel, Iran
South Asian – weakly Buddhist
South Asian – Islamic (incl Pakistan and Bangladesh but not Myanmar)
Islamic Central Asian (incl the “stans”, but not say, Nepal, Tibet)
Central and East Asian – Confucian – Buddhist – Taoist (incl China, Korea, Japan)
SE Asian (incl Buddhist Thailand, Islamic Indonesia and Catholic Philippines)
Pacific Islander including Hawai’i
Subsaharan West African
Subsaharan East African excluding Ethiopia, Eritrea, Somalia
Central African incl. Angola, Mozambique
Ethiopian, Eritrean, Somalian
South African incl Namibia
The key when I drew this up, for me, was “inside each group, do people “get” each other?” Not great either, but less wrong.
mbka
Aug 20 2021 at 6:59am
Just realized that was all except the Americas! So to add
Latin American
Native North American
Native Central American
Native South American
That should do. But – should it be done at all?
Scott Sumner
Aug 20 2021 at 12:52pm
Yes, the census form explicitly includes “Spaniard” as an example of Hispanic.
Frank
Aug 19 2021 at 12:03am
Identity is a question of incentives.
Scott Sumner
Aug 19 2021 at 12:32pm
Yes.
Brian Donohue
Aug 19 2021 at 11:27am
A couple generations ago, “passing as white” was all the rage. Now it’s the opposite. Clearly, white nationalists have the whip hand, and we need to pivot from Afghanistan to exterminate the enemy within.
Floccina
Aug 19 2021 at 2:58pm
Another reason Identity might change:
My wife is from Honduras, my children usually select white, but my older son was having trouble finding a job so he was thinking of switching to selecting Hispanic on job aplications.
Rebes
Aug 19 2021 at 8:25pm
If he wants to increase his chances further, also tell him to write LGBQT+ on his application.
fiona
Aug 19 2021 at 6:16pm
For the last 3 censuses I have selected Other: human as my race. Several people have suggested that since we as homo sapiens originated in Africa, we are all African Americans. Still others state that since they were born in America, they are native Americans. I suppose this is the source of the followup question on tribal identity. 23 and Me and other genetic testing facilities have a lot to answer for.
DeservingPorcupine
Aug 20 2021 at 12:17am
I object to the phrase “social construct” entirely. It’s meaningless and should never, ever be used.
Scott Sumner
Aug 20 2021 at 12:53pm
Good point.
Rajat
Aug 20 2021 at 10:16pm
You say you think “these distinctions will become increasingly meaningless over time”, but the fact that questions like this are being increasingly asked suggests not. My (government) workplace has soft targets for ‘CALD’ employees (culturally and linguistically diverse). In my lived experience, racism in Australia has reduced dramatically over the last 40 years, but I hear about how racist Australia is a lot more than I used to. The racism we have that remains is directed at the most recent cohort(s) of migrants and/or people with foreign (non-British or American) accents, which these sorts of distinctions are useless for tackling.
Linda Seebach
Aug 20 2021 at 11:27pm
After the 2000 census, we checked how the “hispanic” block subdivided by the answer to the race question, and roughly half of the Hispanics chose white and the other half chose Other (with a very small percentage who chose black or another specific category). From which we inferred that for Hispanics, Hispanic was their race, no matter what the census said about it.
Peter
Aug 21 2021 at 3:38am
At the opposite end of the spectrum, and this was discussed a lot here after the recent census data, in Hawaii you have a vast over reporting of people claiming to be White as well as Native Hawaiian and Hispanics are highly undercounted. Everyone here wants to not be East Asian nor Hispanic.
PS: I’m curious why you refused to capitalize white and black when using them as proper nouns but you did every other ethnicity.
Mark Brophy
Aug 21 2021 at 12:42pm
There are so many obnoxious people who capitalize “Black” and refuse to capitalize “white” that it’s best to use the existing convention to not capitalize black or white.
Mark Brophy
Aug 21 2021 at 12:40pm
“White” means not black. Mexicans, Asians and Indians are white. There are so many white people that some want further details.
E. Harding
Aug 21 2021 at 5:02pm
“The US white population is probably not shrinking, at least in the sense that most people think of the term ‘shrinking’.”
Of course it’s shrinking. White people have had below-replacement fertility for decades, and there’s hardly any White immigration into the United States.
In my opinion, the census should collect people’s autosomal DNA in order to make it clear to everyone that what is happening is not just a social construction.
Comments are closed.