Texas elections are under tension between rural counties with a strongly Republican lean, and urban concentrations that are Democratic. In previous elections, the net margin of Democratic votes in urban areas has been erased by Republican margins elsewhere.
Texas has made nearly real-time early voting response available on a per County granular basis. The news story of the weekend is the exceptionally strong early voting turnout, which exceeds 80% of the total 2016 vote statewide. I cross tabulated the reported county by county early voting tally expressed as a percentage of the total votes cast in 2016, in 2018 and the base “Clinton” vote and O’Rourke margins.
I display the O’Rourke vote fraction on the y axis (assumed to represent base Democratic vote) and the tally of votes *already* cast on the x axis, expressed as a percentage of the total 2016 presidential vote. Shown are the top 40 populated counties (of 244 counties), this represents as of 10/24, 5.9 million of the early voters 7.2 million votes tallied.
Two trend clusters are apparent in this presentation. Ex-urban counties surrounding the major Texas cities have early voting participation as of Oct. 24 of 80-105% of the total votes cast, and the participation falls closely along a recognizable positive trend-line conditioned by the 2018 O’Rourke (and alternatively, the Clinton %) vote percentage. O’Rourke outperformed Clinton on a percentage basis in most counties.
A second cluster represents other counties and cities, again with a positive trendline with early voting tally falling between 70 to about 78 percent of the recorded total 2016 votes cast. These are a collection of the 1) dense metropolis and 2) some smaller cities, and the traditionally Democratic lower Rio Grande. Again, as the assumed “Democratic” base vote goes up, the early voting percentage rises.
The positive trendline which closely correlates early voting participation with Democratic vote fraction in earlier elections indicates that early votes are likely trending Democratic.
This is presumptive evidence that Texas early voters are “eager” Democratic votes, and the net margin in early votes will exceed the general polling fractions. Early voting turnout can represent voters whom are long term committed voters, in which case, the residual in-person voting will strongly bias Republican as the “pool” of available Democratic voters exhausts. Alternatively, early voting turnout may represent a new cohort of expanded voters, and the available “pool” has been expanded, in which case the residual in-person vote fractions should mimic long term party affiliation.
The upper steeper trendline cluster represents “ex-urban” outer-ring suburban counties. Dallas suburban counties: Collin, Denton, Ellis. Houston counties: Fort Bend, Montgomery, Brazoria. Austin: Hays, Williamson. San Antonio: Comal, Guadalupe.
The early voting bias of the suburban counties may represent wealth and education factors that are quasi-independent of political affiliation, but the trend line derived from the historic Democratic fraction indicates the early voting participation is biased to Democrats, as is the similar, but less aggressive regression line of the collection of other large urban counties.
Two reliable polls have published recent cross tabs for Biden votes and early voting. Dallas Morning News (UT Tyler) and Quinnipiac show the intention to use Early and absentee voting at 81-87% of the polling sample, and the Democratic early voting intent at 87-90%. The Dallas-Tyler poll indicates Biden will win the early vote at 53-47%. The Quinnipiac poll puts the early voting Biden fraction at 53.4- 46.6%.
If Biden is winning early votes at the same fraction as O’Rourke won in 2018 for the final tally in the sampled counties, the banked margin in the early votes is 52.6%. O’Rourke turnout was 94% of the 2016 presidential vote for the sampled counties. The anticipated net margin in early turnout votes, perhaps 375,000 votes, brings Texas into “toss up” status.