As we have covered many of the state and local races throughout the past campaign cycle — you can check out some of the information we’ve gathered on our War Room — I have found a useful tool in the Cook Partisan Voting Index (PVI).
The Cook PVI, first introduced in 1997, gives an indication of the partisan leanings of each congressional district in the United States using a system of a letter to indicate the party leaning (i.e. “R” or “D”) followed by a plus sign and the amount of points above the national average (e.g. “+12”).
While the Cook PVI is useful when looking at the leanings of congressional districts, its use is limited when analyzing more local races.
Frustrated by this, I set out to perform my own analysis of the counties in Texas and assign them similar PVI values.
Fortunately, in this day and age, the work of calculating votes and percentages is minimal — spreadsheets do all of the math for me. But before the numbers could be calculated, I needed to figure out what numbers to calculate.
The Cook PVI formula is fairly simple: the average results of the past two presidential elections in a given district are compared against the national average in the same time period. The difference between the two numbers results in the PVI.
Only votes cast for Democrats and Republicans are considered.
While I could have followed this formula at the state level, I saw an opportunity to build upon its limitations.
Namely, while the Cook PVI can only analyze data from one nation-wide election (the presidency) every four years, there are several state-wide elections every two years that can be used to compare counties.
If the purpose of a PVI is to determine how consistently voters in a given region vote for a particular party, considering the average results across multiple elections will help account for inclinations or aversions toward specific candidates.
Another benefit of looking at data from a two-year period instead of a four-year period is that it presents a more accurate picture of what the political makeup of a county currently looks like, while still tempering any skewed results from an atypical voter turnout in a given cycle.
All in all, my county PVI formula is slightly more complicated than Cook’s, but operates on the same concept: the average results of all statewide elections in the past two election cycles in each given county are compared to the complete statewide average of the same elections.
Thus, the elections considered for this analysis are as follows:
- 2016: President and Railroad Commissioner
- 2018: Senate, Governor, Lt. Gov., Attorney General, Comptroller, Land Commissioner, Agriculture Commissioner, and Railroad Commissioner
While the results produced under this formula will be far from perfect and doubtlessly have some drawbacks, they should still provide a good indication of the leanings of specific counties.
Keep in mind that with this formula, the number following the party indicator is the amount above the statewide average.
Given that the statewide average in Texas already favors Republicans by five points, the numbers may be confusing in the most competitive districts. For example, a county with a PVI of D+1 is still likely to vote slightly more Republican, while a county with a PVI of D+5 will be the most competitive in the state.
Looking at the extreme ends of the results doesn’t reveal anything too groundbreaking or unexpected.
The map is covered with strongly Republican counties. 163 in total have a PVI of R+20 or greater, blanketing all of rural East and West Texas.
King and Roberts counties get the crown for the most Republican, both at R+40. Both are in the 13th Congressional District, the most conservative in the nation based on the Cook PVI.
At the other extreme, the most Democratic counties in the state are concentrated in South Texas along the border. This includes Zavala and Starr counties, which top out the list at D+35 and D+34, respectively.
The major urban areas of the state — Austin, San Antonio, Houston, and Dallas — all draw a large number of Democratic voters, though the counties themselves are not quite as solidly Democratic as the border counties.
Of those more urban areas, El Paso and Travis counties lead the pack, with D+26 and D+25, respectively.
Dallas County is at D+17, Bexar County is at D+11, and Harris County is at D+10.
At D+5, the most competitive county in the state is Jefferson County, sitting just east of Houston and home of Beaumont and Port Arthur.
Other pockets of competitive areas include the coastal counties near Corpus Christi (Nueces, Kleberg, and Kenedy counties), some rural counties near the border in West Texas (Brewster, Val Verde, and Reeves counties), and counties surrounding the major population centers.
Denton, Collin, and Tarrant counties are at R+5, R+4, and D+1, respectively.
Hays County is at D+6, making it one of the most competitive in the state, but its counterpart to the north of Austin, Williamson County, serves as the bellwether for the state average at R+0.
See the full results below:
Daniel Friend is a reporter for The Texan. While recently finishing his degree in Political Science from Azusa Pacific University, he also interned in the U.S. Senate and co-authored a book on C. S. Lewis’s science fiction trilogy. In his spare time, he might be reading up on Dostoevsky or attempting to write a novel.