On Monday, the Commerce Department declared that it would ask people in the upcoming census whether they were American citizens. Long rumored, this decision—assuming it survives the legal challenges that are already arising—has the potential to reshape our politics to the advantage of Republicans. To discuss why, as well as the other implications of this decision, I spoke with Andy Beveridge, a professor at Queens College and the CEO of the demographics research company Social Explorer. An edited and condensed version of our conversation is below.
Isaac Chotiner: Why exactly was Monday night’s announcement such a big deal?
Andy Beveridge: It’s very important for a number of reasons. I think a large concern in the advocacy community, particularly the immigrant-rights community, is that this, along with the hostile environment created by the Trump administration towards immigrants, is going to deter immigrants from responding to the census. If that’s the case, either not responding to this question or not responding to the census at all will make the census less accurate.
Who do we expect will be undercounted and why?
The groups that get undercounted or that don’t respond to the census very well are recent immigrants, to some extent undocumented immigrants, and people who are living in kind of odd household situations. It would mean recent Hispanic immigrants, recent Asian immigrants might be less likely to answer the census.
“I always tell my students that the census is really only about two things: power and money.”
— Andy Beveridge
Let’s say that there’s massive undercounting of recent immigrants. How will that manifest itself politically?
Well, it would affect reapportionment. In other words, let’s say states such as Florida and Texas have a larger undercount. They might end up with fewer congressmen, along with like California, Arizona, probably New York, states like that.
OK, so there are two separate issues that are related. One is how electoral votes are divided on the map of 538, and the second is within states where people are located, which could warp state legislative districts, which could then in turn warp congressional redistricting, correct?
Yeah. Congressional districts have to be exact. It might mean that areas with high concentrations of immigrants will lose out to areas with low concentrations of immigrants. Basically, the urban areas would be likely the places [where this happens], but also the valley in California where there are a lot of agricultural workers that are immigrants, and Los Angeles, along the Texas border, cities like Houston and Dallas. All of those places could have a lower count than they should, and therefore they would lose a proportion of representation to the rest of the state, which would likely be more rural, more likely affluent, far-flung suburbs, that sort of thing. [Social Explorer has a tool to let you view how different areas might be affected.]
That’s really one of the major things that could happen, and that would affect any set of statewide legislative districts, including Congress. It also could affect some of these larger areas, like when you have a big county or something like you do in, say, San Diego County. Then it could affect the distribution of representatives in that county, the City Council of Los Angeles and so forth and so on.
Then the other thing, which I think is hovering here, is the Supreme Court case called Evenwel, which was decided a couple of years ago, but that case [could be reopened to allow] states to try to base their distribution of seats not on total population, which is the general way it’s done now, but on the basis of citizen voting-age population. And so that would disadvantage both the areas with high numbers of noncitizens, and it would also disadvantage the areas with lots of kids, so that would have a huge effect.
We did a tool to look at what impact that would have, and it would push the balance in several states much more towards the Republicans. In fact, one reporter for the Washington Post told me at one point when this first came up that he had talked to some Republican redistrictors and they felt they’d gone as far as they could with partisan gerrymandering, and so they needed something like Evenwel to push further.
I know the judges in that case ruled that districts did not have to be divided in terms of eligible voters, rather than total people, but they said that states could make that decision in the future. Are you saying the reason that this would open up the case again is that it would offer states the data to do this?
Well, better data. If you’re going to use citizen voting-age data to divide districts, as opposed to use it to enforce the Voting Rights Act, which is the way people use it now, it probably would be [easier] if you had data that was collected at the same identical time that the census collected on citizenship.
What other problems are there with undercounting?
The census data is the baseline, really, for every single survey that’s done by the federal government, by private industry. It’s something that’s very important really for commerce, because you want to know how many people there are. It also is used to divide up government appropriations. I always tell my students that the census is really only about two things: power and money. It’s a way to divide power up, and it’s also a way to divide money up, and so if it’s screwed up for 10 years, it’s a real problem.
The 1920 census was rejected by Congress, so it was never used, so there’s a history of having a situation where the census was rejected. There is also the horrible situation, and I think that some of the immigrants are concerned about this, when they used the block-level data after the 1940 census to figure out where to round up the Japanese and intern them. I think there’s some concern in the immigrant community about the whole idea, that since there’s such interest in deportation now, that these data could be used to help target where to go find immigrants to deport. I’m sure it’s gone through people’s heads.