2019 Census estimate changes congressional reapportionment projections

New population estimates indicate that Texas may gain three seats and Florida two seats in the US House of Representatives after next year's US Census, while New York and California may lose a seat each. A handful of other states, mainly in the West and Sun Belt are projected to gain a seat, and a handful, mainly in the Rust Belt, are likely to lose a seat. Oklahoma is projected to keep its five congressional seats.

Shortly before New Year's Day, the US Census Bureau released their estimate of each state's population as of July 1, 2019. This is not the actual enumeration required by the Constitution every ten years for reapportionment of the House of Representatives -- that census will be taken on April 1 this year. This is an estimate based on birth and death statistics, and federal records (e.g., tax returns or Medicare enrollments showing a Social Security number in a different state than previously) which show internal migration. (A 14-page paper about the population estimate methodology may be found on the Census Bureau website.) The recent data release is only for population by state. Over the course of the first half of this year, 2019 population estimates will be released for counties, metropolitan areas, and cities and towns.

Sean Trende, senior elections analyst for RealClearPolitics, has used a weighted average of the rate of growth across the 2017, 2018, and 2019 estimates to project a 2020 population for each state and then applied the Method of Equal Proportions, the method that the US has used since the 1930 census, to calculate congressional apportionment based on his 2020 population projection:

• Gain 3: Texas
• Gain 2: Florida
• Gain 1: North Carolina, Colorado, Arizona, Montana, Oregon
• Lose 1: Alabama, California, Illinois, Minnesota, Michigan, New York, Ohio, Pennsylvania, Rhode Island, West Virginia

The allocation of seats under this method is something like dealing out an entire deck of 435 cards. Each state gets dealt a seat, then the remaining 385 seats are dealt in priority order. A state's priority is calculated as its population divided by the geometric mean of the number of seats it's been dealt and the number of seats if it's dealt one more. In C syntax: priority = population / sqrt ( current_seats * ( current_seats + 1 ) ). As soon as a state is dealt a seat, its priority is recalculated for the next seat. This method produces an ordered list of seat assignments, and so you'll hear analysts talk about which state will get the 435th seat, which states are "on the bubble" to gain or lose seats, depending on small changes in the final, official count. States that are shown by estimates to be in that 430th to 440th seat range have a strong incentive to push their people to respond to the Census.

The ordered nature of passing out seats also means you can estimate the effect of increasing the size of the House, which has been fixed at 435 for over a century (with a brief exception after Alaska and Hawaii joined the union in the late 1950s). Prior to the 2000 census, when Oklahoma was expected to lose a seat after barely keeping its 6th seat in 1990, Congressman Ernest Istook proposed increasing the House by about 30 seats, enough so that no state would lose a seat, but equal proportions would be maintained.

With the size of the US House fixed by law at 435, a state will keep its congressional seats if its population growth keeps pace with the rest of the country. Growing more slowly puts you at risk of losing a seat. Losing population, as Illinois, Vermont, and West Virginia have done, pretty much guarantees the loss of a seat, unless, like Vermont, you're already down to the constitutional minimum of one.

MORE:

Back in December 2010, after the official state counts were released, I wrote a piece about the 2010 reapportionment process, with numerous links to background material.

The website Thirty-Thousand.org advocates for a House of Representatives closer to the Constitutional number of one representative for every 30,000 people. Had the first article of the Bill of Rights been ratified (the other 11 have been), we'd have about 6,000 representatives. The site provides a deep dive on the history of apportionment methods.

Wikipedia has the numbers for apportionment of the US House of Representatives since 1789.