Demographics Are Changing Slower Than We Thought

The Nation Magically Got Whiter Overnight

Today, the Census Bureau released new data showing how many people of each age, sex, race, and ethnicity reside in the US. It’s nifty. This is the benchmark data people cite when talking about demographic change. And The Narrative is strong in this one. Here’s the Census Bureau’s talking point:

The nation is becoming more diverse. Now, that’s true in terms of time the time series. A Herfindahl-Hirschman index of nonhispanic race groups and Hispanics as a group does show a steady decline from 2010 to 2016 as the nonhispanic white population posts consistently less impressive growth than other ethnic groups. So, as far as it goes, this story is right.

But… it’s only true, from a certain point of view. That is, if you compare Vintage 2016 estimates fro 2016 to Vintage 2016 estimates of 2015. But if you compare Vintage 2016 to Vintage 2015, you get a totally different story.

The Census Bureau revised away 515,000 minorities in 2015.

Here’s the annual raw revisions in population, 2016 vs 2015, by group:

As you can see, these estimates made very big downward revisions to Asians and Hispanics. They made upward revisions to whites until 2015. That is, Census had previously been underestimating the white population, and overestimating Hispanics, blacks, and Asians.


Well, this goes back to Census changing the way they estimate emigration. It turns out all the old estimates of race/ethnicity incorporated dubious assumptions about emigration with understated how many people also left. So that’s a big story. But that doesn’t explain the larger white share!

The larger white share is mor ambiguous, but we can’t think that Census was understating white fertility or overstating white mortality, because headline revisions to those actually cut the other way. Rather, the most likely explanation is a classic problem the Census bureau faces: a large number of people born to Hispanic, and sometimes other race/ethnic groups too, identify as white. That is, Census Bureau uses the fertility of a group to estimate how many new Hispanics will be born. So if you’re born to two Hispanics, you’re Hispanic. But it turns out, a fairly large number of people who Census Bureau identified as Race/Ethnicity X when they were kids end up identifying as just white when they’re old enough to make their own survey response.

This creates a really interesting racial politic. What is the political story of a Nonhispanic White population outperforming expectations, making the nation whiter… if the Nonhispanic White population is increasingly propped up by the children of Hispanics? What’s the point of these classifications?

Let me take a moment to reiterate my view on how we should track race. We should stop tracking it. We should require people to specify ancestry from the hundreds of options available, or an ancestry they define. We should require them to specify at least 2 ancestries. Then we should show them a color palette of skin tones and ask them to match which one best describes them. Race/ethnicity questions as they exist are a ridiculous and non-rigorous hodge-podge of skintone phenotypes, ethnic background, national origin, and cultural identification. They increasingly do not reflect the kind of people Americans are. Furthermore, Census questions serve as definitional for Americans; they are a form of public communication by the State. If the State asks about race, it communicates to people that race is real. But if the State asks about ancestry and requires multiple ancestry, then wholly separately asks about skintone, then it communicates that ancestry is real and plural, which is a better message to communicate to people, and that skintone continues to be culturally significant. Plus it allows us to neatly sidestep the infuriatingly circular question of whether white South Africans are African-Americans.

That digression aside, let’s look at the data again. The chart below shows the change in each race/ethnic group’s share of the total population from vintage 2015 to vintage 2016:

Here you can see that these revisions are pretty stark: Hispanics and Asians got cut, nonhispanic whites got boosted, other groups were basically unchanged.

Now, the magnitude is small. We’re talking 0.1% difference. But the point is, forecasting the demographic breakout for next year is not that hard. It’s no mystery that the US is getting more diverse and will continue to get more diverse. That’s baked into the age pyramid. What is most interesting is what this data says about how we should adjust our prior beliefs. And it turns out, diversification of America is happening slower than we thought.

Check out my Podcast about the history of American migration.

If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!

Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!

I’m a native of Wilmore, Kentucky, a graduate of Transylvania University, and also the George Washington University’s Elliott School. My real job is as an economist at USDA’s Foreign Agricultural Service, where I analyze and forecast cotton market conditions. I’m married to a kickass Kentucky woman named Ruth.

My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research.