The PES is the Bureau of Census Post-Enumeration Survey. This survey is used to statistically estimate the under- and overcounts for each state in the 2020 census. Their conclusion: six states had undercounts, while eight had overcounts. (An undercount means the official 2020 population for a state from the census is statistically significantly less than the PES estimate. You can figure out what an overcount means.) Here’s a link to the much-more-readable press release. This article examines (in excruciating detail) the relationship between Census PES and the House of Representatives.
Executive Summary
Many people have used the PES to claim that red states should have more seats and blue should lose seats. Using the actual apportionment methodology deployed by Census, here are the changes in House seats for each state.
Figure 1 Summary (click for larger image)
Blue states lose two seats, one each from Minnesota and Rhode Island. Red states gain two, Texas and Florida.
The critics are correct, but the changes are not nearly as extreme as they assume. Click here to download the summary Excel workbook.
Background
Release of the PES has led to some wild – and incorrect – speculation about what the House of Representatives would look like if the PES estimate had been used instead of the official Census numbers. Most fundamentally, using PES would be unconstitutional:
The actual Enumeration shall be made within three Years after the first Meeting of the Congress of the United States, and within every subsequent Term of ten Years, in such Manner as they shall by Law direct.
That’s from Article 1, Section 2. It describes how the population count should be done to determine the number of representatives for each state in the House. The phrase “actual Enumeration” has been interpreted as requiring an exact count. That excludes statistical adjustments. It’s also why the decennial census tries to be much more precise than the years between.
But there’s another issue. The PES estimates are being used incorrectly to calculate the changes in the number of representatives in each state. From the PES press release:
According to the PES, which states had undercounts?
- Arkansas (-5.04),
- Florida (-3.48),
- Illinois (-1.97),
- Mississippi (-4.11),
- Tennessee (-4.78), and
- Texas (-1.92).
And overcounts?
- Delaware (+5.45),
- Hawaii (+6.79),
- Massachusetts (+2.24),
- Minnesota (+3.84),
- New York (+3.44),
- Ohio (+1.49),
- Rhode Island (+5.05), and
- Utah (+2.59)
What many have noticed is that the states with overcounts are predominately blue (Democrat leaning) states. Those with undercounts are predominately red. Therefore, many conclude, there should be six fewer Democrats and six more Republicans in the House.
Methodology
This is a fundamental misunderstanding of how apportionment works. Fortunately, your humble economist has some expertise in that field. In fact, a significant chunk of my time over the last five years has been devoted to working with apportionment in the House.
I used my Excel apportionment calculator to figure out how many seats each state would get using PES population numbers. For better or worse, PES numbers are not the same as those Census uses to calculate apportionment. Let’s use Alabama as an example. The population used for apportionment was 5,024,279. The PES estimate was 4,896,000. The PES estimate of net coverage error was -1.05%, but that was not statistically significant. Nevertheless, the implied before PES population is 4,947,954. That’s 76,325 less than the apportionment population. I needed to find a different approach.
Fourteen states had statistically significant percentage changes. (The list is above.) Using the percentage changes for the fourteen states in question, I recalculated the apportionment populations for those states. For example, the Census estimate for Arkansas apportionment population was 3,011,524. PES says that figure is 5.04% too low. That implies a PES population using the apportionment population is 3,011,524*1.0504 = 3,163,305. I repeated this procedure for each of the remaining 13 states.
I then ran the software. The detailed results are in my usual Excel workbook and are summarized below. WordPress no longer allows uploads of macro-enabled Excel files so I can’t give you a link to download the file that crunches the numbers.. If you want to do the calculations yourself, sign the EULA and download the file TEMPLATE VB for apportionment.xlsm. Paste the population figures from the Working sheet in this workbook into the State Populations tab in the TEMPLATE workbook. Then run the ParseCopyPasteSortAssign macro.
Figure 1 Summary (click for larger image)
Conclusion
The moral of this tale is straightforward. Don’t rely on seat-of-the-pants calculations when you haven’t at least looked for possible complications. In this case, it makes a big difference.
