Texas in 2020 (Part Two): County Level and Path to Victory
This is the second part in a series where I consider the data regarding the possibility of a blue wave in Texas. The first part (which can be read here) looked at state-level data and built several models that can be used for calculating the effects of various variables on Democratic vote share, with an emphasis on the effect of voter turnout. The results indicated that higher voter turnout does increase Democratic vote share at the state level, on average. I also ran a logistic regression model to calculate the win probability for Democratic candidates in any given race, depending on various variables.
Ultimately, my analysis found that even with promising results from Beto O’Rourke’s 2018 Senate campaign, the Democratic Party still has the deck stacked against them in terms of winning the Lone Star State. These models found that it’s very unlikely that Joe Biden will win Texas in the 2020 presidential election (at least for now).
In this part, I will take a closer look at Texas, examining the state at the county level. While I would have liked to consider data at the congressional district level (as they are units of roughly the same population), I had a ton of difficulty finding consistent data at that level over multiple years. And not to mention that congressional district boundaries are redrawn every ten years, making it impossible to compare historical data even if it could be found. So I had to settle with counties, bearing in mind that they can vary drastically in population, as they at least retain their borders over time. Considering there are 254 counties in Texas, the data collection was a time-consuming endeavor, but doable.
With this data, I will also run regression models, similar to what I did in the first part to determine what, on average, determines Democratic vote share within Texas counties. From there, I will diverge and lay out specific counties that a potential Democratic candidate should target if they hope to turn the state blue. None of this is to say that Texas will turn blue in 2020 per se, but I am interested to see how a “blue wave” could likely pan out in future elections.
More specifically about the data, I have collected additional data from the first part (which I may update closer to the election) in order to improve upon these models. The observations include the counties in each of the presidential elections since 1996; however, I also included results in various midterm elections (US House, Senate, and Governor) to gain a more robust sense of each counties propensity to vote for certain parties. I also included more robust economic data: the unemployment rate (and its change from the year before the election), the GDP growth from the year before the election, the poverty rate as defined by the US Census (and its change from the year before the election), the real income growth from the year before the election, and the growth in the Dow Jones Industrial Average from the year before the election. Furthermore, I included multiple dummy variables accounting for incumbency, the party’s legacy in holding the White House consecutive terms (since 1992, no party has held the White House for more than two consecutive terms), and whether the county is urban or suburban. Considering that Democrats typically perform better in urban and suburban counties than in Texas’s rural counties, I thought it’d be necessary to include these variables.
From these variables, I hope that I can produce more robust models for analysis moving forward.
Figure 1: Mean Values for Texas Counties
Variable All Counties Urban Counties Suburban Counties
Democratic Vote Share 0.3177 0.4870 0.3150
Gen. Voter Turnout 0.5060 0.4745 0.5180
Primary Voter Turnout 0.1367 0.0907 0.0826
Income $38,270 $44,650 $50,890
Unemployment 0.0568 0.0573 0.0506
Education 0.1542 0.2614 0.2061
Non-White 0.1969 0.3221 0.1866
Young 0.3308 0.4217 0.3588
Next, let’s run a series of fixed effect regression models to establish a more formal relationship (the coefficient results can be seen in Figure 2). One key difference from the regressions in the last post is that in addition to collecting data on the non-white population, I also considered the Hispanic population of each county (the data sources consider Hispanic an “ethnicity” category as opposed to its own race category. Because of this, there’s considerable overlap between whites and Hispanics, which wouldn’t count towards the non-white population. For this article, I include Hispanics of all race categories, including non-white categories).
Model 1 is fixed effects with turnout and change in turnout from the previous presidential election. Interestingly, the results indicate that the level of turnout doesn’t matter much, with a low and statistically insignificant coefficient. Despite this, the change in turnout is quite significant, as an increase in turnout from the previous election increases the county’s Democratic vote share. One reason for this could be that large increases in turnout relative to previous elections are indicative of strong campaigning or increased national interest in the election, causing people that normally don’t vote to turn out. In recent history, it seems like these infrequent voters turning out disproportionately lean Democratic.
Other factors that increase Democratic vote share include the county’s proportion of Hispanic residents, Democratic incumbency in the White House (Bill Clinton running for re-election in 1996 and Barack Obama running for re-election in 2012), and national Democratic performance (a “blue wave”). These factors appear to be consistent with existing evidence; however, two variables stand out and require explanation.
One is the Economic Index, which account for the average in the changes in various economic indicators in the year leading up to the election. This Index is standardized from 0 to 1, where 0 points to negative changes in these indicators (growing poverty, growing unemployment, shrinking GDP, and shrinking per capita income) while 1 points to positive changes in these indicators (shrinking poverty, shrinking unemployment, growing GDP, and growing per capita income). What’s interesting is that Democratic vote share is positively associated with the Index, suggesting that Democratic candidates fare better when the economy appears to be improving. This appears to go against assumptions that Democrats benefit when the economy enters downturns (such as the Great Depression bringing in FDR in 1932); however, there are two possible explanations for this. One is that the Index only considers changes to these parameters as opposed to their absolute level. For example, a 5 percent increase in the unemployment rate may be a greater shock to a more affluent county that typically has a low rate and votes Republican than to a perpetually depressed county that’s used to high unemployment and to supporting Democratic candidates. To that end, while changes to economic indicators may adjust voters’ views on the status quo, absolute levels may say more about how much change is required to shake such perceptions; in other words, economic views are relative. And another explanation is that when changes to the economy does sour opinions on the status quo, such anger is usually directed towards the incumbent party, even if it’s the Democratic Party. While Democrats benefitted immensely from economic anxieties and anti-Bush sentiment in 2008 during the Great Recession, such dynamics can work the other way as well. For example, Ronald Reagan was able to mobilize anger against the Carter administration’s handling of the energy crisis and economic slowdown in 1980 through a “change” platform that promised to move America away from the status quo. While one could argue the economic merits of Reagan’s platform, his decisive victory indicates that poor economic conditions turns voters more against the incumbent party rather than gravitate them towards the same party every time. To that end, this can explain how the Index can produce interesting results.
And the other variable of note is the Democratic Other Office Aggregate, composed of the average of the county’s Democratic vote share in the previous presidential election as well as that of the gubernatorial, senatorial, and House elections in the preceding midterm cycle (for 2016, this would include President Obama’s performance in 2012 as well as the Democratic candidates’ performance in the corresponding elections from the 2014 cycle). Interestingly, while this variable is intended to capture the county’s overall partisan leaning, it has virtually no effect on the Democratic vote share in the presidential election. These variables are highly correlated with each other, which would be expected to lead to a higher coefficient; however, that doesn’t appear to be the case. This low coefficient is partly an indication that other variables (demographics and national vote share) do a better job at explaining variation between different elections, the midterm versus the presidential election.
Model 2 is fixed effects with only turnout. While the relationships exhibited in Model 1 are similar for the other variables, the main difference is that in the absence of the change in turnout between elections, the absolute level of turnout is strongly and positively associated with Democratic vote share. The results of Model 1 do suggest that this absolute level isn’t as important when considering the change in turnout; however, both models indicate that higher turnout is expected to increase Democratic vote share, on average.
Finally, Model 3 is a logistic regression seeking to determine the probability that a Democratic candidate will win a given county. Among the significant variables that increase the probability of a Democratic win in a given county include the national Democratic vote share, whether the Democratic candidate in the previous election won the county, whether the county is urban, and the proportion of Hispanic residents in the county. One surprising finding is that the voter turnout is negatively associated with the probability of winning a given county. This finding does go against the findings in the previous two models; however, its magnitude is offset by that of other variables, particularly the large coefficient of the national Democratic vote share.
Overall, these three models provide strong evidence that certain variables can increase the Democratic vote share in a given county. These include the coattails of a strong Democratic performance nationally, a high population of Hispanic residents, whether or not the Democratic candidate in the previous election won the county, a high voter turnout, and a large increase in voter turnout from the previous elections. Together, these variables lay out key conditions under which a Democratic candidate can benefit in a given election cycle within Texas.
Figure 2: Regression Estimates
Variable Model 1 Model 2 Model 3
Voter Turnout -0.0162 0.1082*** -5.128*
(0.0480) (0.0416) (2.711)
Turnout Change (from previous election) 0.1679***
(0.0332)
Previous Dem. Win 4.558***
(0.4047)
Dem. Incumbent 0.0624*** 0.0587*** 0.6773
(0.0241) (0.0248) (0.4796)
Dem. Other Office Aggregate -0.00001* -0.00001 0.0031***
(0.00001) (0.00001) (0.0005)
Economic Index 0.0498* 0.0504* 3.298
(0.0283) (0.0286) (2.222)
Hispanic 0.1961*** 0.1995*** 4.869***
(0.0517) (0.0522) (0.7784)
Urban 0.0091 0.0084 1.771***
(0.0078) (0.0079) (0.5386)
Suburban 0.0032 0.0035 -0.6758
(0.0038) (0.0038) (0.5479)
National Dem. Vote Share 2.8234** 2.6993** 46.10***
(1.1903) (1.2017) (12.10)
Intercept -29.06
Now with this background, let’s look more specifically at which counties should be targeted. First, I identified 16 “solidly Democratic” counties in Texas. I define these as counties where the Democratic candidate in the previous two statewide federal elections (Hillary Clinton in 2016 and Beto O’Rourke in 2020) received at least 60 percent of the county’s vote. These 17 counties serve as the backbone to any potential victory in Texas moving forward and I list each of them in Figure 4.
Figure 4: Solidly Democratic Counties
Brooks (southern part of the state)
Cameron (includes Brownsville, on Mexican border)
Culberson (western part of state)
Dallas (Dallas city proper)
Dimmit (southern part of state)
Duval (southern part of state)
El Paso (El Paso city proper)
Hidalgo (on Mexican border)
Jim Hogg (southern part of state)
Maverick (on Mexican border)
Presidio (western part of state, on Mexican border)
Starr (southern part of state, on Mexican border)
Travis (Austin city proper and surrounding suburbs)
Webb (southern part of state, on Mexican border)
Willacy (southern part of state)
Zapata (southern part of state, on Mexican border)
Zavala (southern part of state)
What’s interesting about these solidly Democratic counties is that almost all of them (with the exception of Dallas, El Paso, and Travis) are neither urban nor suburban. Many of them are smaller counties clustered in the southern part of the state close to the Mexican border. Several characteristics define these counties. For one, theyhese counties are predominantly Hispanic; in 2016, Hispanics made up 80 percent of the population in these counties. They are also less well-off economically. While their average Economic Index is slightly above the state average in 2016 (0.242 in these counties versus 0.233 across all counties), these counties have a stubbornly high average unemployment rate of 8.3 percent (versus 4.7 percent statewide), low average per capita personal income of $34,410 (versus $45,614 statewide), a staggering average poverty rate of 27 percent, and an average GDP of only $23.8 million (versus an average of $63 million across all counties).
These demographic and economic factors create prime feeding ground for strong Democratic support; however, they also complicate efforts to drive up turnout. Collectively, these counties had a 56.3 percent turnout rate in 2016 (above the statewide rate of 51.4 percent); however, if we exclude Dallas and Travis, the two largest and most affluent counties in this subset, the turnout rate drops to 48.6 percent. Considering these challenges, a Democratic candidate hoping to win Texas should focus on driving up turnout in these predominantly Hispanic regions. Luckily, there is encouragement in that turnout in the 2018 midterms was up from the previous 2014 midterms. Collectively, these counties had a 51.3 percent turnout rate in the 2018 Senate race (41.1 percent when excluding Dallas and Travis), which is up from 30.3 percent in 2014 (21.4 percent when excluding Dallas and Travis). In this respect, this demonstrates that a strong grassroots campaign can mobilize higher turnout in these solid Democratic counties, including those with historically low turnout.
In general, while both Clinton and O’Rourke performed well in these counties, the lower turnout in 2018 compared to 2016 often resulted in O’Rourke winning fewer raw votes than Clinton even if his vote share was larger than Clinton’s. Even so, there are some exceptions. On top of having a higher vote share, Dallas saw more raw votes in 2018 for O’Rourke despite a lower turnout. If we take the 2016 turnout figures, Hillary Clinton would have netted an additional 40,873 votes in Dallas County had her vote share equaled that of O’Rourke’s. This exercise is useful in visualizing how high voter turnout and increased Democratic enthusiasm could sweeten the electoral fortunes of the Democratic Party statewide.
To win statewide, Democratic candidates must rely on high turnout in these counties. In 2018, O’Rourke demonstrated that a strong campaign can increase turnout, producing a close statewide race. Even so, campaigns shouldn’t focus the bulk of their time or resources on increasing turnout in these counties. While such a strategy may increase turnout somewhat in these counties, such gains would likely be offset by the diminishing rate of return from such turnout (some of these counties are very small and have fewer additional voters to mobilize), lack of support in more competitive counties due to lack of organizing in these battlegrounds, and perceptions that the candidate lacks broad appeal and a realistic chance of statewide victory. Rather, campaigns are better suited putting their resources into more competitive counties to persuade voters there while relying on heightened enthusiasm statewide in order to drive up turnout in all counties. Such enthusiasm will likely attract more volunteers and fundraising as hopes of statewide victory become more realistic, meaning more resources can be diverted towards a get-out-the-vote operation in these more solid counties. None of this is to say that the solid counties should be ignored, but rather this observation points to the difficulty of a Democratic candidate winning statewide. Such an effort requires a strong campaign that has enough resources to invest in both the solid counties and the more competitive counties.
Next, let’s consider some potential pickup opportunities for the Democrats. I have identified 19 counties where the Democratic candidate won between 45 and 55 percent of the vote in the 2016 and/or 2018 elections. I’ll break them into two categories: the counties that were not won in either election and the counties that were won in one or both elections.
Figure 5: Narrowly Won Counties
Bexar (San Antonio city proper and surrounding suburbs)
Brewster (western part of the state, on Mexican border)
Fort Bend (suburbs of Houston metro area)
Frio (southern part of state)
Harris (Houston city proper and surrounding suburbs)
Hays (suburbs of Austin metro area)
Jefferson (eastern part of state, on Gulf of Mexico)
Jim Wells (southern part of state)
Kenedy (southern part of state)
Kleberg (southern part of state)
Lasalle (southern part of state)
Nueces (Corpus Christi city proper and surrounding suburbs)
Reeves (western part of state)
Tarrant (Fort Worth city proper and surrounding suburbs)
Val Verde (western part of state, on Mexican border)
Williamson (suburbs of Austin metro area)
Figure 6: Narrowly Lost Counties
Caldwell (suburbs of Austin metro area)
Collin (suburbs of Dallas/Fort Worth metro area)
Denton (suburbs of Dallas/Fort Worth metro area)
As indicated, many of these counties are located within the state’s major metropolitan areas, including Dallas/Fort Worth, Austin, Houston, San Antonio, and Corpus Christi. Considering that many commentators the suburbs of these metro areas to pave a critical path for any Democratic statewide win, it shouldn’t be too surprising that many of these closely contested counties are primarily suburban. With an average county Economic Index of 0.232 in 2016, these counties are in line with the state average in terms of change to indicators in the year leading up to the election, but the actual magnitude of those indicators reveal a less favorable economic picture on average (although there’s a decent degree of variance across individual counties). In 2016, the average unemployment rate in these counties was 5 percent, above the statewide rate of 4.7 percent (although nine of these counties boasted an unemployment rate below the statewide figure). The average county per capita personal income of $41,399 was also below the statewide figure of $45,654 (six counties were above the statewide figure and four had a per capita personal income above $50,000). The average county poverty rate of 16.5 percent was higher than the statewide rate of 15.6 percent (although eight counties were below the statewide rate and four had a poverty rate under 10 percent). And the combined GDP of these counties shrunk by 0.7 percent in the year leading up to the election, slightly larger than the 0.2 percent contraction statewide (although eight counties saw positive GDP growth, four saw GDP growth greater than 5 percent, and Reeves County saw its GDP grow by 21.7 percent). Although the regression models suggest that more favorable economic conditions increase Democratic vote share, these close competitive counties are less well-off as an aggregate than the state across multiple economic indicators. To this end, it’s possible for Democratic candidates to make successful appeals to certain counties depending on the broad economic conditions.
In terms of demographics, the average proportion of nonwhite residents across these counties in 2016 was about 16 percent and the average proportion of Hispanic residents is 51.7 percent, both of which are above the statewide figures. Overall, this marks an increase in these counties’ racial and ethnic diversity; in 1996, the average county proportion of Hispanic residents was only 45 percent. And for some of these counties, diversity has expanded more rapidly. Among the most dramatic transformations, Jefferson went from only 6.4 percent Hispanic in 1996 to 20.1 percent in 2016 and Tarrant grew its Hispanic population from 14.2 percent in 1996 to 28.4 percent by 2016. Considering that the regression models found a statisitcally significant positive relationship between the proportion of Hispanic residents and Democratic vote share, this growth in diversity indicates that these counties are increasingly becoming favorable territory for Democratic candidates.
One notable trend with these counties is that, historically, they have been competitive between Democrats and Republicans in presidential elections. In 2016, the Democratic vote share in these counties was 48.3 percent (42.1 percent when only counting the suburban counties in this category). While the two Bush elections were slightly less favorable for Democrats (the average county Democratic vote share for those elections was only 43.8 percent), these counties have remained competitive since 2008 with an average county Democratic vote share of 48.6 percent since 2008. Considering this, it would not seem that Donald Trump being on the ballot in 2016 shifted these voters away from the Republican side, or at least not immediately. Outside presidential elections, these counties have not been as reliably competitive for Democrats. In the 2014 midterm elections, a very unfavorable cycle for Democrats, Governor candidate Wendy Davis only won 43.2 percent of the vote in these counties, Senate candidate David Alameel fared even worse with only 38.1 percent, and House candidates only received a combined 37.3 percent (voter turnout in these counties was only 34 percent). While these numbers may reflect the anti-Democratic sentiment nationwide in 2014, they also suggest that these counties generally distinguish Democratic candidates at the presidential level from those running for others. 2018, however, marked a notable shift in the Democratic direction as Beto O’Rourke won 53.8 percent of the vote in these counties. While the nationwide climate in 2018 was more favorable for Democrats, this doesn’t take away from the significant improvement O’Rourke made over previous candidates in driving up voter turnout and improving his vote share. To this end, these counties serve as a critical roadmap for future candidates.
Of these counties, six were narrowly lost in 2016 but were picked up in 2018 (Brewster, Hays, Jefferson, Nueces, Tarrant, and Williamson). And of the counties won in 2016, all of them (except Kenedy) were also won in 2018. What’s impressive about the performance in these counties is that while 2018 saw fewer votes were cast in these counties than in 2016, the Democratic candidates received more raw votes in 2018, indicating that the vote share increase is not merely attributed to fewer votes being cast overall. Such gains are especially notable in Hays (which netted 12,360 Democratic votes between 2016 and 2018), Tarrant (netted 25,105 votes), and Williamson (netted 21,382 votes). Using the 2016 turnout figures, Hillary Clinton would have won an additional 77,321 votes in just these six pickup counties had her vote share within these counties equaled that of Beto O’Rourke’s. And if we include all 19 “close” counties in this exercise, Clinton would have added 226,326 votes to her total (statewide, Clinton’s margin of defeat in Texas was 807,179).
Considering this, a statewide campaign should focus on these competitive counties given the immense number of votes at stake and the favorable conditions for a post-Trump Democratic candidate to play into (large metropolitan areas and considerable pockets of potential Hispanic voters). O’Rourke received immense credit for performing well in these counties in 2018, allowing him to come within striking distance of winning statewide.
Finally, an interesting question to pose is whether Clinton could have won the state of Texas had her vote share across all 254 counties equaled that of O’Rourke’s. Obviously, it should be noted that Clinton and O’Rourke are two very different candidates with different levels of appeal across the electorate. But considering that I ran a similar exercise for certain counties, it begs the question of how many potential votes Clinton could have won statewide. And in doing this exercise, it can shine some light on what higher voter turnout might mean for a Democratic candidate in presidential contests (versus a midterm election, where one party usually enjoys an advantage).
After running through the calculations using 2016 turnout figures, it’s estimated that Clinton would have added 419,530 votes to her statewide total had her vote share in every county matched that of O’Rourke’s, increasing her vote count from 3,877,868 (43.24 percent) to 4,297,398 (47.91 percent). Would this had been enough to put Clinton over the top? That depends on which candidates she takes votes away from in this scenario, but it would probably be very close. While Donald Trump was the clear frontrunner statewide with 4,685,047 votes (52.23 percent), we also have to consider the vote shares of third-party candidates statewide. There’s Libertarian candidate Gary Johnson with 283,492 votes (3.16 percent), Green Party candidate Jill Stein with 71,558 votes (0.80 percent), Independent Evan McMullin with 42,366 votes (0.47 percent), and 8,895 votes for write-ins and other third-party candidates (0.10 percent). While these vote shares appear trivial, they actually determine the outcome of the race in this scenario depending on how the blows are distributed. Figure 6 below outlines several different distributions and the resultant outcomes.
Figure 6: How Different Vote Distributions Affect the Texas Race Outcome
Clinton Net Gain: 419,530
Distribution of Votes Subtracted Resultant Trump Vote Share Resultant Outcome
Only Trump 4,265,517 votes (47.56 percent) Clinton Win (+31,881 votes)
Trump and McMullin 4,307,883 votes (48.03 percent) Clinton Loss (-10,485 votes)
Trump and Stein 4,337,075 votes (48.36 percent) Clinton Loss (-39,677 votes)
Trump and Johnson 4,549,009 votes (50.72 percent) Clinton Loss (-251,611 votes)
Trump, Johnson, and Stein 4,620,567 votes (51.52 percent) Clinton Loss (-323,169 votes)
Trump and All Non-Clinton Votes 4,671,828 votes (52.09 percent) Clinton Loss (-374,430 votes)
This figure indicates that while Clinton may have increased her own vote share in this scenario, the margin of her victory (or defeat) is still uncertain. The most realistic scenario for how the subtracted votes would be distributed is difficult to tell because there’s little polling on second choice preferences for Texas in 2016.
The best evidence available is the exit polls for Texas from 2016, which asked respondents who they would vote for in a two-person race. 5 percent said they would not have voted at all in a two-person race, 3 percent of those that said Clinton (40 percent of the entire sample) actually voted third-party, 2 percent of those that said Trump (53 percent of the entire sample) actually voted third-party. Overall, about 2.3 percent of the entire sample voted for a third-party candidate and still would have voted in a two-person race; however, 5 percent of the sample said they would not have voted at all in a two-person race. While the data doesn’t specify who this 5 percent of “no two-person race” respondents actually voted for, the nature of the question suggests that most of them voted for one of the third-party candidates. In other words, it’s unlikely from this polling data that Clinton would have siphoned off a significant number of third-party voters to her side, suggesting that most of Clinton’s increased vote share would come from Trump voters.
But on the other hand, this scenario would require Clinton to be a stronger candidate that could generate significant enthusiasm from voters. One reason why many voters said they may have stayed home in a two-person race is that they just weren’t excited by either of the candidates (this problem was especially true for Clinton, who struggled to win over Sanders supporters after the primaries). Looking at the Texas exit polls, only 38 percent of respondents said they strongly favored the candidate they voted for (within that cross-section, 49 percent voted for Clinton) while 34 percent of respondents said they were more motivated by disliking their opponent than motivated by liking the candidate they did vote for (within this cross-section, 39 percent voted for Clinton). And given the rise of party polarization, it’s unlikely that many Republican voters would flock to the other side even if they dislike Trump. In this alternative scenario, Clinton would be a stronger candidate that can mobilize the Democratic base and urge a decent chunk of those that otherwise wouldn’t vote (or would vote third-party) to instead back her. So in that respect, it can also be argued that Clinton would consolidate third-party support, even if doing so makes it more difficult to win statewide.
At the end of the day, what matters isn’t what happened in 2016 or 2018, but what happens in 2020 and beyond. While it’s fun to play with past data and imagine what could have been, it’s more constructive to use this data and lay out a road map for future campaigns to consider. Since 2016, there have been critical shifts in the demographics and political climate in Texas that raise the possibility of flipping the state “blue”. But we also shouldn’t underestimate how difficult it’ll be for the Democratic Party to win the state, let alone make it a fixture in its electoral coalition. As impressive as O’Rourke’s 2018 campaign was, he benefited from a relatively unpopular opponent in Ted Cruz (who, in October 2018, had a 47 percent favorable rating vs a 42 percent unfavorable rating despite being a Republican incumbent), a midterm cycle that benefited Democratic candidates across the board as sentiment turned against the incumbent Republican Party (a common trend for midterm elections), and unusually high voter turnout across the board that’s typically absent from midterm elections. And of course, it should be mentioned that despite these advantages, O’Rourke still couldn’t cross the finish line in Texas.
Given the current polling situation in Texas, there’s considerable optimism that anti-Trump sentiment could put Biden over the top and make him the first Democrat to win a statewide election in Texas since former Governor Ann Richards in 1990 (and the first Democrat to win the state in a presidential general election since Jimmy Carter in 1976). I will make a follow-up post closer to November that considers this polling information. This post seeks to examine how that victory will materialize when we look under the hood of Texas’s geography.
And then beyond 2020, there remains critical questions on how much of a swing state or stronghold Texas will be. Even if Biden wins the state in 2020 (and the general election), then how competitive will Texas be in the 2022 midterm elections? How much of these electoral gains in Texas is merely anti-Trump sentiment and at what point will the demographic shift (assuming the Republican Party doesn’t address this issue) lead to consistent electoral gains for the Democratic Party? These are valid questions to consider as well when interpreting the post-2016 election results in a state that has reliably supported Republicans for over a quarter century. Even so, the question of Texas will be an important one for the 2020 election and its outcome could be pivotal in deciding who will be taking the oath of office in January 2021.
