Why Texas, Florida won’t get more House seats from census miscount

The U.S. Census Bureau this week said the decennial tabulation of every American resident had miscounted populations in 14 states by enough that a handful of seats in the House of Representatives should have gone to other states.

But even with evidence in hand that at least a few seats should have shifted between states — Texas and Florida, for example, may have deserved additional seats — there is nothing those states, the Census Bureau, the Commerce Department or Congress can do about it.

On Thursday, the Bureau released the results of its post-enumeration survey, an in-depth look at the accuracy of the decennial count that took place beginning in 2020.

The latest survey showed the decennial count missed, or undercounted, a statistically significant portion of the populations of six states: Arkansas, Mississippi, Tennessee, Florida, Texas and Illinois.

At the same time, the census over-counted populations in eight states: Delaware, Hawaii, Massachusetts, Minnesota, New York, Ohio, Rhode Island and Utah.

The statistical errors can happen for all sorts of reasons. People are undercounted when they refuse to engage with census takers or when public data from other databases that the Census Bureau uses to fill in gaps is incomplete or incorrect. People can be overcounted when they are tallied both at their permanent homes and in second homes, or when a child of divorced parents is counted in both homes.

The accuracy of the broader count may have been undermined, too, when the Trump administration forced the bureau to rush to its conclusions even in the midst of a global pandemic that slowed the counting process. And many of the states that suffered undercounts did not put substantial efforts into promoting participation in the census, known in some states as complete count campaigns.

The resulting errors can mean the difference between a state winning an extra seat in Congress or just missing out on deserved representation.

When the Census Bureau uses its results to allocate House seats, it does so through a mathematical formula called the Huntington-Hill method that prioritizes states by population. Every state begins with one seat in the House, and then the remaining 385 seats are allocated in order of what is called a state’s priority value.

In 2020, that meant California, the nation’s largest state, won the 51st seat in Congress. Texas, the second-largest state, took the 52nd seat. California is so large that it also took the 53rd seat, while Florida got seat No. 54.

By the time the final seats are allocated, the difference in priority values are tiny. This decade, Minnesota won the 435th and final seat in Congress by a difference of just 3.4 points over New York.

Put in real terms, according to calculations by the demographer Kimball Brace of Election Data Services Inc., Minnesota would have lost its eighth seat in Congress if the Census Bureau had counted only 26 fewer people there. New York would have kept its 27th District if the census had tallied just 89 more people there.

Farther down the list, Texas would have needed an additional 189,000 residents to gain a third new House seat; Florida needed 171,000 additional residents to pick up a second new seat; and Tennessee would have needed another 321,000 residents to add a new seat to its nine-member delegation.

The post-enumeration survey released this week appears to show that Texas and Florida may have actually deserved those seats. The decennial census undercounted Texas’s population by 1.9 percentage points, or about half a million residents. Florida suffered an undercount of 3.5 percent, or about 700,000 residents. Tennessee, where the undercount was more than 4 percent, would have been right on the brink of adding a new seat as well.

But even with the new figures, Texas, Florida and Tennessee will not be getting a new seat in Congress, thanks to a two-decade old Supreme Court ruling authored by former Justice Sandra Day O’Connor.

That ruling, in a case called U.S. Department of Commerce v. U.S. House of Representatives, held that a clause in the Constitution relating to the census and the Census Act of 1976 prohibited the Census Bureau from using statistical sampling to calculate apportionment of congressional seats.

Said another way, the court held that only the decennial count of actual residents — rather than modeling aimed at producing a more accurate count — could be used to determine House seats. The decennial count is the actual enumeration that must be used; the post-enumeration survey is a statistical sample that cannot be used.

“The conventional understanding is that the Constitution requires use of the census for apportionment,” said Michael Li, a redistricting and reapportionment expert at New York University’s Brennan Center for Justice.

There is nothing stopping those states that missed out on a new seat from bringing a new suit. Texas Attorney General Ken Paxton’s (R) office did not immediately respond to a request for comment on the Census Bureau’s undercount.

Given that a new apportionment would likely reward Southern conservative states, the conservative majority on the Supreme Court may have a new interest in considering a challenge, said Michael McDonald, a political scientist and census expert at the University of Florida.

But any challenge would likely take years to wind its way through the court system. Instead, those states will have to wait another decade — and hope their populations hold — to see their congressional delegations grow once again.