The Trump administration spent an extraordinary amount of time and effort messing with the census over four years. Its apparent aim: It wanted to exclude undocumented immigrants from population counts, thereby allowing Republicans to draw more favourable maps in redistricting and gain more ground in the House.
The effort ultimately failed. And it now seems possible that these efforts might have played a role in costing red states seats.
A new report from the US Census Bureau found that 14 states were significantly miscounted in the 2020 Census, including six by 4 percentage points or more. At the high end, Hawaii’s population was overcounted by an estimated 6.8 percentage points, while Arkansas’ was undercounted by 5 points.
A trend you might notice if you peruse the data is that most of the states with significant overcounts were blue states like Hawaii (e.g., Delaware, Rhode Island, New York and Massachusetts), while most with undercounts were red, Southern states like Arkansas (e.g., Tennessee, Florida, Mississippi and Texas).
What that means: When it comes to the post-census awarding of seats - a process known as apportionment - those red states might have been at a deficit, because they weren’t given credit for their full populations. Blue states, on the other hand, might have been given more seats because the Census Bureau thought they contained more people than they actually did.
Exactly how the new data might have changed things isn’t totally clear. It’s difficult to gauge that with any certainty, given the margins of error involved in the new survey, the interplay between states’ populations in the apportionment calculations, and the fact that the new data are limited in some key ways, as The Post’s Tara Bahrampour writes.
But it seems pretty evident that incorrect numbers probably allowed two blue-leaning states with overcounts - Minnesota and Rhode Island - to keep seats they shouldn’t have, given that they just barely cleared the bar for keeping those seats. Minnesota kept its seat by a scant 26 people, and both states had been expected to lose seats before the bureau announced otherwise.
On the flip side, the undercounts in Florida and Texas might well have cost those two red-leaning states seats that they were on the cusp of adding. Texas did gain two other seats, but its 1.9% undercount was enough to deprive it of half a million people in apportionment. In pre-census population projections, both states had been on track to gain an additional seat.
(Some estimates Thursday suggested the miscounts might also have moved another seat from a red state to a blue state. We’ve checked in with some authoritative experts on the subject and will update with their assessments.)
So how does this trace to the Trump administration? You might recall hearing, in the heat of the 2020 election, that it decided to cut the census count short by a month. Again, the goal was pretty evident: It wanted to get the data early enough to use for apportionment before the administration might change - and ultimately did change - hands. The data were to be delivered to Trump by Dec. 31, three weeks before he would ultimately leave office.
(Earlier, the administration had tried and failed to get a citizenship question on the census, which probably would have dissuaded at least some undocumented immigrants from responding. It could even have allowed for districts to be drawn without regard for undocumented immigrants. Trump later signed a presidential memorandum in favour of barring undocumented immigrants from apportionment, specifically. It soon became clear the administration was going to attempt to overlay its own estimates of undocumented immigrants on top of census data.)
Census officials internally bemoaned the rushed timeline. They complained that it seemed to reflect a political agenda and - crucially - that it would sacrifice accuracy. Political officials ultimately failed to get the apportionment data or even some more cursory estimates they had been demanding to try to push the scheme through in time. But they pushed right up until the closing days of Trump’s presidency.
It’s impossible to say whether that’s the reason for the miscounts rather than, say, the pandemic alone or the fact that some of the undercounted states, like Texas, didn’t spend as much to encourage people to respond to the census.
Miscounts do happen: In the 2000 census, for example, more than 20 states were overcounted and the District of Columbia was undercounted. But those miscounts were smaller than the ones we saw in most of the 14 states Thursday, and the 2010 Census included no statistically significant miscounts.
It should also be noted that adding a seat to a blue state doesn’t necessarily translate to a blue seat. That depends on how the maps are drawn in a particular state, given the distribution of population.
But given the fine margins involved, the chaotic handling of the census and having data collection trimmed by multiple weeks, so late in the game, can’t have helped.
The Washington Post
