Texas is up first on the list of first looks at redistricting. This look at Texas will feature some issues that are going to be recurring throughout this discussion and one issue that will impact Texas the most but might come into play in some other states.
The first issue is that we do not yet have the actual precinct and block level counts from the 2020 census. That means that this first look is based on the 2019 estimates. And, of course, estimates are not necessarily exact (as the state level numbers for 2020 showed). While the far right is upset about the national numbers and want to raise sinister suggestions that something happened behind the scenes to fudge the real numbers, it is equally likely that the previous administration was fudging the numbers in the estimates. What is most likely is that certain steps by red state governments and anti-government rhetoric led to an undercount of certain groups in red states in 2020 when the Trump Administration was running the Census. So, we have to expect that there will be some unexpected deviations within states when we get the numbers in November.
The second issue is the Donald Trump had a big impact on voters. There are various ways to measure partisan lean in a state. Most involve taking a composite of recent state-wide elections. The software that I am using is currently based on the 2012 through 2016 elections. In Texas, in 2012, both Senator Ted Cruz and Senator Mitt Romney carried the state by around 16%. In 2014, Senator Cornyn carried the state by about 27% and other Republicans were winning by around a 22% margin. In 2016, Donald Trump won by around 9%. In 2018, however, Senator Cruz only won by 2%. While Governor Greg Abbott won by around 13% with the other state-wide Republicans ranging between 3% and 11%. Finally, in 2020, President Trump only won by 6% and Senator Cornyn won by around 10%. In other words, what my software is showing as a 60-40 state based on the 2012-16 results is actually something more like a 54-46 state. And a good chunk of that swing was in suburban districts which probably went from something like 60-40 to very close to 50-50. Overall, there were three congressional districts (Fourteenth, Fifteenth, and Sixteenth) in which the Republican House candidate beat the 2012-16 Republican composite numbers in their district, and in some districts the Republican underperformed by around 10%.
A great example of this swing is the Seventh District. Way back in the day, the Seventh District was represented by George Herbert Walker Bush. It was a Republican pocket in Democratic state. Basically, it was (and largely still is) an enclave of the wealthy and the upper middle class. By the 2012-16 numbers, it is a 60-40 Republican District. In 2018, Representative Elizabeth Fletcher took the district away from the Republican incumbent by about 5%, and she managed to hold onto the seat in 2020 by around 3%. (That and the Thirty-second district are the only two nominally Republican districts that Democrats won in 2020.)
The bottom line is that, if you use pre-Trump numbers, it is not that difficult to draw a map in which Republicans improve on the 23-13 current margin. But, if you assume that the 2020 numbers are the new numbers (while that means some opportunity for the Republicans with Tejano voters in South Texas), is is going to be hard to draw favorable maps in the suburbs.
The final issue for Texas is actually a national issue, but it will have the most impact in Texas. The Constitution requires using the whole population to apportion House seats to the states, but is silent about what numbers states should use to draw district lines. Traditionally, states have used the whole population numbers. In the 1960s, when the Supreme Court held that the Equal Protection clause required that districts have roughly equal population, the Supreme Court used the symbolic phrase “one person, one vote.” After the last Census, some conservative activists made a leap to conclude that if the Equal Protection clause requires “one person, one vote,” then district lines should be based on eligible voters. The Supreme Court rejected the argument that states required to use eligible voters as the measure for drawing lines, but did not address whether states could use eligible voters.
Of course, conservative activists are not pushing this position because they think that it is correct. (And it almost certainly is not.) They are pushing it because immigrants have always tended to cluster in urban areas (which lean Democratic) and rural areas (which lean Republican) are aging faster than urban and suburban areas. Of course, if Republicans were serious about this theory, they would also exclude inmates (who are also ineligible to vote) from the count, but most prisons are in rural area; so, excluding inmates from the count would reduce representation for rural areas. As such, they want to use the count for voting age citizens.
If Republican could use voting age citizens instead of total population (and they will probably try in Texas), it would change what Republicans need to do. For example, even though Texas is gaining two districts, there are four districts (out of thirty-six districts) which are more than 10,000 residents short of the target number for each district in a thirty-eight district map. Three of the four districts have a lop-sided Republican majority with the other district currently being Democratic by a somewhat narrow margin (and the Democratic incumbent is not running for re-election). But the citizen voting age population (CVAP) is slightly less than two-thirds of the total population. Of the four districts, the only one that misses the CVAP target is the Democratic district.
Additionally, using CVAP instead of total population would potentially change voting rights targets. Using total population, Non-Hispanic whites are only around 42% of the population with Hispanics representing 39% of the population, African Americans representing 13% of the population, and Asian-Americans 5% of the population. Those numbers would suggest five African-American districts (if possible), fifteen Hispanic districts, and two Asian-American districts (if possible). On the other hand, Non-Hispanic whites are 52% of the CVAP with Hispanics falling to 30%. Those numbers would suggest only eleven or twelve Hispanic districts. With careful line drawing, the Republicans can be competitive in three of four Tejano-dominated Hispanic districts in South Texas, but the other Hispanic districts will likely be Democratic districts so keeping that number down is crucial. Currently, Texas has ten districts that are Hispanic majority, three districts in which Hispanics and African-Americans combine for the majority, and three districts in which Hispanics, African-Americans, and Asian-Americans combine for the majority.
With all of that for background, the two initial questions for Texas Republicans when the Texas Legislature meets in November to draw the new lines will be: 1) do we use total population or CVAP to draw the lines; and 2) do we assume that 2020 is the new normal.? My guess is that we will see two sets of maps out of Texas. The initial map will use total population and will serve as a back-up map (in case the courts reject the CVAP maps). After adopting the total population maps, the Legislature will then pass a set of CVAP maps. I also think that Republicans will have to draw maps on the assumption that 2020 may be the new normal. In other words, pack as many people who were Democrats in 2012, 2016, and 2020 into Democratic districts as possible shooting for 80% plus Democratic districts. For Republican districts, particularly in suburban and urban areas, you will aim for a higher margin than previously. Typically a 60-40 split is a pretty safe Republican district, but this cycle we may see attempts to get to a 70-30 split (at least for the 2012-16 composite).
The big problem in Texas is that in 2010 Republicans did a very good job at drawing a jigsaw puzzle of a map with lines zigging and zagging as necessary to push as many Democrats as possible into the Democratic districts leaving them with multiple lean Republican districts. If you want a list of the Democratic target districts for the past four cycles, almost all of them border the Austin-San Antonio region (Tenth, Twenty-first, Twenty-third, Twenty-fifth, and Thirty-first). The Republican’s best performance in this list in 2020 was the Thirty-first district with approximately 56% and the rest were under 53%. As all of these and the three Democratic districts (Fifteenth, Twentieth, and Thirty-third) will need to be shedding voters, it will take some very careful line drawing to avoid losing one of the Republican districts.
Additionally, incumbents tend to like to keep a solid base when you redraw the lines. They want to run in a district in which they have already built up name recognition and a relationship with their voters. But when you have multiple adjoining districts with excess population, there is a ripple effect of shifting. Some of this will be solved by creating new districts, but creating those new districts will create deficiencies in neighboring districts which will then need to shift into other districts to make up the new districts.
Looking at the current map and excesses, the new districts will be in two of three locations — Dallas-Forth Worth area, Houston area, or Austin-Antonio area. Whichever one does not get a district will see a shift of the districts from the other areas in their direction. (The following districts have over 100,000 in excess population — the Third District (northeast Dallas area), the Eighth District (suburbs north of Houston), Tenth District (from suburbs of Houston to Austin area), Twenty-second District (southwest Houston area), Twenty-sixth District (north Dallas area), Thirty-first District (north Austin and points further north). Of these districts, the Third, Tenth, Twenty-Second, and Thirty-first were barely Republican in 2020; the Eighth and Twenty-sixth was solidly Republican.) My hunch is that the districts will be in the Houston and Dallas-Forth Worth area. That means that there will be some shift westward from Houston and southwestward from Dallas-Fort Worth. At the same time, there will be some shift northward from San Antonio toward Austin. The end result is that there will be a lot of new people added to the Tenth, Twenty-fifth and Thirty-First. And because all three are vulnerable, the Republicans will have to be careful with how the high population Democratic areas get shifted to avoid flipping a district.
Before drawing the new districts, the Texas Legislature will first have to deal with the parts of the state that actually have to few people. This process will not help the Republicans much. For the most part, the short districts (the First (East Texas), the Thirteenth (Panhandle and North Texas), an the Nineteenth (Panhandle and North Texas) are in safe Republican parts of the state and are bordered by other safe Republican districts. This process will shift 20,000-40,000 mostly Republicans from neighboring districts into these “short” districts. The biggest problem for the Republicans is that some of the voters in the new First will come from the Fifth which will in turn have to get around 5,000-10,000 from the Sixth which is a little more at risk than the other districts involved in the process. And while the Twelfth District (Fort Worth area) — which will probably have to shed voters to the Thirteenth District — is currently a safe Republican seat, the maneuvering around with the other districts may put it at risk.
The room for Republicans to make a little mischief will be with the Sixteenth and the Thirty-Fourth. The Sixteenth is just a little bit short, but the only district that the Sixteenth (El Paso) can gain population from is the Twenty-Third District (Upper Rio Grande Valley). The Sixteenth — Beto O’Rourke’s old district — is already a pretty safe Democratic district, and the adjoining precincts in the Twenty-third are solidly Democratic precincts. Thus, bringing the Sixteenth up to the required numbers will make the Twenty-third a bit safer for the Republicans.
The Thirty-Fourth is where there is a lot of potential for creative redistricting. The Fifteenth District (Central Rio Grande Valley-San Antonio suburbs) is a lean Democratic district with a significant excess. The Thirty-Fourth District (lower Rio Grand Valley) is a lean Democratic district with a significant deficiency. The Twenty-seventh District (Corpus Christi) is a safe Republican district with a deficiency. In 2020, the Republicans actually got more votes than the Democrats if you combined the votes from these three districts, but two of the three seats are held by the Democrats. With the Democratic incumbent in the Thirty-fourth opting against reelections, the Republicans will probably take a look at what they can do to swing one of the two Democratic seats. To achieve this goal, the Twenty-seventh District can take some Democratic precincts from the Fifteenth and Thirty-Fourth and give some Republican precincts to the Fifteenth and Thirty-Fourth. with the Fifteenth making up the difference in the Thirty-Fourth by shedding some tossup precincts. The Republicans probably can’t make enough changes to get three likely Republican districts, but they probably can make the Fifteenth and Thirty-fourth more competitive without putting the Twenty-seventh at risk.
Then comes the difficult task, having already done the best that they could do in 2010 to move as many Democrats into the Democratic districts leaving the neighboring districts as lean Republicans, the Republicans know face excess population in the Democratic districts which have to be move to neighboring districts while still trying to get Republican voters from the safe Republican districts into the new districts. My first draft, while slightly improving the Republican chances of getting the Seventh back, actually made the Thirty-second District slightly bluer, while putting the Tenth, Seventeenth, and Twenty-Fifth at risk. Both of the two new districts were nominally Republican, but with the swing in recent years, the Thirty-seventh District that I drew in the Dallas area might be competitive. After some modifications, I ended up with a map in which the Republicans should have one safe and one lean Republican district from the new district, but the Tenth became a true toss-up district as did the Thirty-first.
The bottom line from my map-playing is that my expectation is that there are ways that Republicans can draw the two new districts in Republican leaning areas but only by pushing current Republican leaning districts further into blue precincts. So 23-13 probably becomes 24-14. But in that 24-14, there will be three or four Democratic-held swing districts and two or three Republican-held swing districts, meaning anything from a 28-10 Republican advantage to a 21-17 Republican advantage.
I am certain that Republicans will do their best to pack as many Democrats into the Democratic districts as possible, but the geography in Texas makes this harder than it is in other states. Weird pockets of gentrification or areas that have stayed upper class adjoin areas in which poverty is significant. Even the most Democratic districts only a few had expected Democratic votes above 75%, and there were as many rural Republican districts with similar percentages. If Texas truly did go blue at the presidential level, it would be hard for Republicans to maintain their advantage in Congress and the legislature even with creative lines.
