So far the discussion of TX early voting data has been lacking partisan data on turnout by party, because Texas does not have party registration (unlike states like Florida).
However, TX does have primary voting history and some other demographic data, and that can be seen in the voter file when you look at which individual voters have voted so far.
Here is a pretty good report on that data for TX early voters so far. It also happens to be from a Republican consultant/strategist, so it definitely ought not to be skewed/selectively interpreted in favor of the Dems:
https://mcusercontent.com/d3064a2fadaf6089dc58a8393/files/862e2ef3-a87a-4640-8919-0137e43dc5b6/Statewide_Report_Day_8.04.pdfSo far, 31.2% of voters have Republican primary voting history but not any Dem primary voting history, and 27.5% have Dem primary voting history but no GOP primary voting history. 3.0% have mixed voting history but their last primary was a Dem primary, and 0.9% have mixed primary history but their last primary was an R primary.
That compares to 28.7% Republican (0.9% mixed with last R) and 25.1% Dem (1.5% mixed with last D) in 2018 early voting.
In 2016, it was 30.1% R (1% mixed last R) to 20.3% D (0.9% last D).
So it is clearly a very substantial improvement for Dems so far relative to 2016. It also appears to be an improvement over 2018 for Dems, if you add up the R + mixed R and compare to D + mixed D, it is 32.1% R to 30.5% D in 2020 so far (a margin of 1.6% for the Rs), compared to 29.6% R to 26.6% D in 2018. If people generally voted similarly to in the 2018 Senate race, with a shift in the partisan composition of the electorate of about that much, that would take TX squarely into the pure tossup zone.
Voters with Democratic primary voting history are consistently turning out to vote at higher rates than voters with Republican primary voting history. For example, 65.2% of people who voted in the last 4 R primaries have voted, compared to 76.35 of people who voted in the last 4 Dem primaries. 48.7% of people who voted in 1 of the last 4 D primaries have voted, compared to 42.2% of people who voted in 1 of the last 4 R primaries. Dem's also hold similar turnout advantages among 2 of 4 and 3 of 4 primary voters.
One interesting thing, however, is that voters so far are older than in 2018. Only 10.5% of early voters so far are age 18-29, compared to 12.5% in 2018 and 13.4% in 2016. Whereas 21.5% so far are age 70+, compared to 17% in 2018 and 15.1% in 2016.
There are different ways you could interpret this. One is that young voters are not voting and that seniors are turning out. In TX, if that is the case, that would definitely be good for Trump (even if Trump is doing a bit worse with Seniors nationally, if he wins TX it will absolutely be with the support of older voters bringing it home for him, because younger voters are much more diverse and more Dem). Another way to interpret that, however, is that we are only part way through the early voting period in 2020 so far, and the votes that we have so far are still disproportionately skewed towards mail ballots. In TX, only Seniors are automatically eligible to vote by mail, so it is unsurprising that they should be overrepresented when the data does not include all in-person votes yet, since there has been more time for mail ballots to be returned so far than for people to vote in person. My educated guess is that the latter is mostly the correct way to interpret this. I would be extremely shocked if the electorate doesn't get younger as more and more votes come in.
One implication of that is that the later early vote and the election day vote in TX may end up being less tilted towards the GOP than in some other states where more younger people are eligible for vote by mail. And another way to interpret this is also that if the partisan composition of voters is as favorable as it is to Dems (more favorable so far than 2018) even with the electorate being skewed towards older voters, then once the Senior vote share starts to go down towards what it was in 2018/2016, things will be even more favorable for Biden/Dems than they already are.
Another thing to look at is the gender of early voters. That also looks good for Biden; 52.1% so far are women, and only 43.1% men (4.8% unknown). This is more skewed towards women than overall voter registration in TX, which is 50.9% women, 45.3% men, and 3.8% unknown. A more female electorate is definitely good for Dems.
Voters with Democratic primary voting history also seem to be out-voting people with Republican primary voting history not just in the largest urban/suburban/Beto-trending counties, but also in basically all the significantly sized counties where data is included in the report.
In Harris County (Houston, 60.5% of D primary voters have voted, compared to 52.9% of R primary voters. Dems have similar sorts of advantages across the big urban counties such as Dallas, Tarrant, Bexar, and Travis.
Importantly, since people were wondering about whether high turnout in places like Denton County and Collin County is good for Ds or for Rs, the Dem turnout advantage is also present in those counties. i.e.:
Collin: 66.3% D - 59.3% R
Denton: 65.8% D - 60.6% R
Williamson: 66.7% D - 57.0% R
Fort Bend: 53.5% D - 42.5% R
Galveston: 64.7% D - 59.7% R
Brazoria: 67.1% D - 60.3% R
It is even the case in GOP juggernaut Montgomery County: 54.0% D - 47.5% R
This is all very consistent with what we would expect to see if these suburban counties continue to trend Dem as they did in 2016 and 2018. And regardless of whether Biden wins, it is very bad for GOP chances to hold the State House and a bunch of those competitive suburban congressional seats... Honestly right now I would probably rate the TX State House at tilt/lean Dem rather than tossup, unless something unexpected changes.
Importantly, there is also a similar sort of Dem advantage in most of the smaller city counties outside of the major megacity metros (note however these are all in central or east TX). For example:
Bell (Killeen/Fort Hood Area): 56.0% D - 37.7% R
Jefferson (Beaumont area): 59.1% D - 54.8% R
McLennan (Waco area): 60.5% D - 47.8% R
Smith (Tyler area): 63.0% D - 47.1% R
Brazos (College Station area): 46.7% D - 39.2% R
Comal (New Braunfels/San Antonio exurbs): 61.9% D - 53.7% R
In Fort Worth Exurb/rural counties there is also a Dem advantage, though a small one. For example:
Johnson (South Fort Worth exurbs/semi-rural): 48.2% D - 43.0% R
Parker (West Fort Worth exurbs/Weatherford/semi-rural bleeding towards West TX): 47.7% D - 42.8% R
These areas are the deeply R of the deeply R. The fact that there is anything remotely positive at all to say about these areas for Dems is a very positive thing for Dems and very bad for Trump/Rs.
The turnout in these sorts of counties, along with the previous category of smaller city counties in central/east TX, is also suggestive of how things may be going in smaller rural TX counties. They suggest that insofar as there are Democrats in rural TX, they are voting, and the white rural non-college vote (which was already pretty much maxed out at very close to 95-100% Republican) seems like it is quite unlikely to offset the very clear urban/suburban trend we are seeing in the data. In other words, I would expect to see a similar sort of thing as in 2018 in these sorts of counties. Yes, Biden should get demolished similarly to Hillary/Beto, but I would not expect any sort of sharp Republican swing to offset the Dem swing elsewhere, and some smaller city & rural counties may even swing very slightly Dem.
There seem to be only 2 sets of counties where anything is possibly the slightest bit amiss for Dems. The first of these are West TX counties:
Lubbock (Lubbock area): 59.8% D - 59.9% R
Randall (Amarillo area): 57.0% D - 56.4% R
The second set that looks less than completely 100% ideal for Dems are South TX Hispanic counties.
Hidalgo (Edinburg/McAllen) - 48.9% D - 51.7% R
Cameron (Brownsville) - 51.6% D - 54.6% R
El Paso (not South TX and Beto country, looking more similar to the non-Hispanic urban/suburban counties): 55.6% D - 47.9% R
Webb (Larado): 33.4% D - 34.7% R
These counties are clear outliers (with El Paso being a clear outlier amongst the outliers). One thing to keep in mind, though, is that there are very few Republican primary voters in these counties in the first place. If you are a Republican primary voter in these counties, there is something that is making you be a committed voter and go against the grain. For example, even with Rs having a turnout edge in Cameron County, they are only 12.8% of voters there (compared to 43.2% D primary voters). And in all the other counties, it is lower than that 12.8% R, down to 4.8% R in Webb County.
It has often been the case in many of these counties that the Dem primary has been effectively the general election, and it has not been at all unusual for Dem primary turnout to actually be higher than general election turnout.
Nevertheless, the bottom line with regards to the South TX counties is there are a lot of Hispanic voters in South TX who have voted in previous Dem primary voters but who have not yet voted in the 2020 General election. This is pretty much the one and only thing that could be very favorable for Dems that could theoretically happen that is not clearly visible as already happening.
So overall, the early vote demographics look pretty favorable for Biden here and seem quite consistent with the polls that have been showing TX as a true tossup. I would not say they are favorable enough to guarantee a win or for TX to be lean or even tilt D, but given the comparison to 2018/2016 and how we know people voted in 2018 both of those years, it does appear that TX is truly competitive and Biden has a very real shot to win it.
The one area where there is clear room for improvement, though, is Democratic turnout in heavily Hispanic counties in South TX. Anything that can be done to increase that ought to be done ASAP, that could make all the difference in the world and could easily be the difference between Trump winning TX by a fraction of a point and Biden winning it by a fraction of a point.
