Civiqs: Cruz +2

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

I am afraid that people are getting too optimistic here about Beto's chances based on a new internet poll from a new pollster (Civiqs) which is run by a liberal blog (Daily Kos) with an untested methodology.

Yet again, Charlie Brown doesn't realize that the football is going to be taken away? Will Charlie ever learn? Apparently not.

There are ALWAYS a few early polls showing Texas deceptively close. These polls are ALWAYS wrong. Why are they wrong? Because Texas is a hard state to poll due to very low turnout and an electorate that is very different from the overall population, and these polls are not polling the actual electorate.

There is a very simple point that you need to understand in order to understand elections in Texas. TEXAS HISPANICS DO NOT VOTE. PERIOD. NEVER HAVE, NEVER WILL (at least not for a few decades until they get older, citizenship rates go up, they actually register to vote, and a culture of actually voting develops). Texas Democrats (particularly Hispanics) vote like horses that are led to water drink. You can lead them to the polling place, you can show TV ads to them, you can knock on their door, your can call them on the phone, you can send them mail, but you can't make them vote.



2002 Senate

You must be logged in to read this quote.

2008 President

You must be logged in to read this quote.

"Obama is even competitive in North Dakota and Alaska, where he has campaign offices. Although he is not competing in Texas, he probably will in 2012, and it is only a matter of time before demographics turn Texas blue."



2010 Governor

You must be logged in to read this quote.

"Rick Perry is hugely unpopular, the energy of the 2008 Democratic primary finally activated Texas Democrats, and although Obama did lose, he swung Texas bigly and muh Obama coalition emerging Democratic majority in Austin, Dallas, and Houston. Bill White is a great candidate with strong crossover appeal."



2012 President

You must be logged in to read this quote.

"2010 was a disappointment, but Bill White was running against a Republican midterm tsunami. 2012 will be different, it is not a midterm and the Obama coalition will turn out and vote again. Obama may not win Texas in 2012, but he can come close, build on the progress he made in 2008, and build up Texas Democrats for future competitive races."



2014 Governor

You must be logged in to read this quote.

"After so many years of Republican rule, surely Texas must be getting tired of it. They have to elect a Democrat eventually, surely they can't have one party rule forever. Even states like Wyoming elect a Democratic governor occasionally. Wendy Davis has gotten lots of national support, fundraising, and is a rising Democratic star. And besides, we finally have the national Democratic party investing in Texas, Battleground Texas and Hispanics will turn Texas blue!"



Finally, most recently... I'll even post more than one poll for this one...

2016 President

You must be logged in to read this quote.

"Trump is historically unpopular, a horrible fit for Texas, and suburban Republican women and Hispanics won't vote for a pussy grabber. Could this be the year that Texas goes blue?"

Do you get it yet?



With regards to the Civiqs poll, the crosstabs do look quite plausible. Contrary to what some people incorrectly believe, Texas doesn't have any more Black Republicans than other states, and doesn't really have materially more Hispanic Republicans. While Hispanics do vote slightly more Republican than in some other states, this is not a very large difference. Texas Dems routinely win about 65-75% of the Hispanic vote, and this has been the case for at least the past decade with the earlier exceptions of Bush's mega-landslides, when he did a bit better (but not actually as much better as is sometimes claimed). Exit polls tend to overestimate how Republican Hispanics vote because of poor methodology, which has led to a myth that Texas has lots of Republican Hispanics.

The Civiqs methodology sounds interesting and might (or might not) actually work reasonably well, because they seem to do a quasi-voter file match -

https://civiqs.com/methodology/

It needs to be proven, though.

But while the crosstabs look reasonable, the overall result is likely overstating minority turnout. This supposed 2 point Cruz advantage is not among any sort of "likely voters," as far as I can see, and that being the case, of course it will overstate how well Beto is doing. If the entire population of Texas voted, then Beto would indeed probably do very well. But the entire population of Texas will most definitely not vote, and the people who do vote are not all that much like the overall population.

If Hispanics didn't vote in 2016 against Trump in a Presidential year, why would any sane analyst think that they will vote in 2018 in a midterm year?

Who DOES vote in Texas?

Angry white men. And who do they vote for? Republicans.

So the problem Democrats have is really not so much support levels, it is that nobody (at least nobody that is not a white Republican except for a few elderly black women in Dallas and Houston and a few white liberals in Austin) votes in Texas. In addition, even if minority turnout does increase by realistically plausible amounts, it has a comparatively small effect on the overall result, because Whites vote so heavily Republican.

Ds have made long term minor progress in the Texas suburbs, but at the same time have absolutely cratered in rural areas to the point that Republicans win about 100% of the white rural vote. The "good" news is that it is not possible for Republicans to do better than 100% of the white rural vote, so maybe the marginal D gains in the suburbs will no longer be offset by R rural gains, as they have been for the past decade or two. But the hole is sufficiently large that it will take a while to climb out of.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

The Black vote is most Democratic in the Deep South and Northern Cities; in the Outer South and the Midwest, it tends to be more 80/20 rather than 90/10 or 95/5.

What actual evidence is there that the black vote is 80/20 or so in the Outer South/Midwest?

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

I believe your missing something in all of those election polls you posted.
1. Besides 2008, all occurred in either R wave years, or neutral years. And TX 2008 is a different state than TX 2018, just look at the Rust Belt, AZ, GA.

That is where you are going wrong. You are trying to analyze Texas as though it is another state, which is capable of shifting its support or splitting its ticket between candidates of different parties. Texas is not like that, and in this sense is different from every other state of which I am aware. Voters always support the same party, whereas in other states on occasion they will support the opposite party in order to have some checks and balances, Texas voters just do not believe in that concept at all.

You must be logged in to read this quote.

That is an artifice/coincidence of bad polling. In reality (and this goes for elections in general, not just for Texas). Many voters who say they are "undecided" in a poll also will end up just not voting at all. And many others who say they are undecided are not really "undecided" at all - they basically vote the same way every time, and just tell pollsters they are undecided so that they can think of themselves as being openminded good citizens, and thereby feel good about themselves and how much of a moderate hero they are. So it is not as though all undecided voters actually break in a way so that you can compare pre-election polls with the actual results.

But more specifically in the case of Texas, it goes beyond "undecided" voters not voting and also not actually being "undecided." In Texas, it is also that pre-election polls are including too many people who won't actually vote, which overstates the Democratic support. This artificially makes the Democratic support # in polls closer to the final support # that Democrats will get on election day, and artificially lowers the Republican support # in polls. So then, even if undecideds (the ones that are actually undecided and actually vote) break 50-50, it will appear after the fact if you are looking at it how you want to look at it, that "undecideds broke GOP." This is why, if you look at undecideds in polls and compare that to results, it will never look like undecideds are breaking Democratic in Texas.

You must be logged in to read this quote.

It sounds to me like our difference here is only really in how we define "suburb." I was thinking of suburb on a sub-county level (e.g. Cypress is a major growing suburb of Houston), whereas it sounds to me like you are probably thinking just of "suburban" counties such as Collin County, Williamson County, or Fort Bend County.

Much of the Democratic gains over the past decade or so in Texas have been in these suburban areas of urban counties. And there has also not been much true "urban" population growth in Texas for a while, even in "urban" counties. Keep in mind that the counties in Texas are sufficiently geographically large so that there is lots of empty land still even in urban counties, and that previously un-occupied land on the fringes of the counties is where most of the growth has been (in new suburbs, i.e. not true "urban" growth).

In reality, in the major "urban" counties Democrats have been getting a higher % of the vote in the core urban counties (Dallas and Houston especially) primarily because of population shifts within the larger metro areas. What is actually going on here (in Dallas county and Harris county) is that there is either basically no white population growth in the county or possibly even slightly negative. Whites that formerly were in Dallas/Harris counties are moving into the suburban counties such as Collin/Denton/Ellis/Kaufman/Montgomery/Brazoria/etc.

This also can create the impression that urban/suburban whites in Texas are getting significantly more Democratic, because you see predominantly white precincts in e.g. North Dallas or West Houston that are voting more Dem. To some limited extent that is true, but a lot of it is also simply the fact that the whites who formerly were voting Republican there just moved to a new suburb further out and are now still voting Republican there. And then the only Whites left in the initial area are Whites who were always voting Dem in those areas, plus maybe a few new White transplants from other states (which are not actually all that Democratic), and some new minorities. It is not, at least not to the extent that you might think, that people who used to vote R are now voting D. And the Rs are still voting R, they just moved to Rockwall/Montgomery/wherever.

Compare, for example, Dallas and Collin Counties in 2008 to 2016.

In 2008, Dallas had 740k votes and went 57-42 for Obama (423k to 310k). 113k vote margin.
In 2016, Dallas had 765k votes and went 60-34 for Clinton (461k to 263k). 198k vote margin.

(Note that the overall turnout is barely any higher, despite a decade of substantial overall population growth. Why? Because a lot of the White people (i.e. the people who actually vote) left, and were replaced by Hispanics, who generally don't vote.)

On its face, that looks like a pretty decent improvement. And when you compare to other years like 2004 (when Bush actually won Dallas County), the fact that Hillary Clinton won Dallas County by almost 2 to 1 looks phenomenal, and if you don't realize the within-metro-area population shifts that are occurring underneath the county level and between counties, suggests that Democrats are making major urban gains which could start to flip TX blue. And the overall shift from a 113k vote margin to 198k vote margin is 85k. That doesn't seem so bad at all. But that is deceptive.

Now look at Collin County.

In 2008, Collin had 297k votes and went 62-37 for McCain (185k to 109k). 76k vote margin.
In 2016, Collin had 364k votes and went 55-39 for Trump (201k to 141k). 60k vote margin.

Despite seemingly large gains, not much shift in the margin there; a shift of 16k votes, even with Gary Johnson helping to crater Trump's numbers in a way that is not likely to be the case in other races.

Then start looking at *other* suburban/exurban counties in the DFW metro area besides Collin County, in particular the ones with predominantly White population growth (lots of white Republicans fleeing the urban counties). Look at Wise County (Fort Worth exurbs, and btw Tarrant County itself had basically no shift in vote margin whatsoever from 2008 to 2016)...

In 2008, Wise had 48k votes and went 77-22 for McCain (37k to 11k). 26k vote margin.
In 2016, Wise had 57k votes and went 82-15 for Trump (46k to 8k). 38k vote margin.

So just Wise county alone basically offsets the seemingly sizeable Dem shift in Collin county.

Also look at other DFW suburban/exurban counties like Johnson, Ellis, Rockwall, Hood, and you will see the same story. When you add all these counties up, you start offsetting the Dem gains in Dallas county, and Dems have made hardly any overall gains in the greater DFW area after a full decade of off-the-charts population growth and demographic change.

And this - Dallas and its suburbs, especially focusing on the northern suburbs and Collin County - is pretty much the #1 example you could point to of an urban/suburban area of Metro Texas seeming to shift Democratic.

So when you look at the actual vote margins, despite the North Dallas/Collin County area shifting pretty strongly Democratic on the precinct level or on the county level in terms of the % of the vote that Ds receive, the actual overall vote margins are not shifting by as much as you might think.



That's the basic story in Dallas and also Houston. In Houston, you have to look more on the sub-county level, because Harris county is geographically so huge (larger than Dallas County), and so many of the Houston suburbs (and even exurbs) are actually all located within Harris County. But underlying the different sizes of counties is the same basic trend as in Dallas, just with West Houston (TX-7 area in congressional districts) playing the same role as North Dallas (TX-32 area in congressional districts). As a result, even though Democrats appear to be doing better in urban Houston if you are looking at things like the % swing in precincts or in counties, if you look at the overall vote margins in the greater Houston Metro area, the change is much less impressive.



However, it should be said that Austin is different from Dallas/Houston - there you do have white population growth, which is also Democratic, in the "urban" county (Travis county) as well as in "suburban" counties (Williamson mainly), where it is not quite so Democratic.

You must be logged in to read this quote.

TX minorities (at least Hispanics) have not been voting in particularly larger numbers over the years. True, minority votes are increasing, but only very gradually. Given the size of the Republican advantage, it takes a LOOONG time for that gradual increase to make a difference and bring TX to competitiveness, much less make it blue. My point is not that it will never happen, it is that it is not happening imminently, and in particular not happening in time for Beto O'Rourke in 2018.

You must be logged in to read this quote.

I agree that African American turnout is not bad, and that was especially the case when Obama was on the ballot. But Hispanics in Texas most definitely cannot "be easily motivated to vote." LMAO, the idea that you can get a single Hispanic person in Texas to vote is just about the craziest thing I have ever heard. You must never have paid attention to an election in Texas before. It might be possible to motivate Hispanics to vote in other states such as Nevada, California, or wherever else, but I assure you it is not possible to motivate Hispanics to vote in Texas.

You must be logged in to read this quote.

Here again I think we are in basic agreement on this point, it is just that you were thinking of "suburbs" in terms of suburban COUNTIES, whereas I was thinking of them in terms of sub-county areas (including suburbs within "urban" counties). Republicans carry TX by such large margins thanks in no small part to counties such as Montgomery County/Brazoria (Houston), Rockwall/Ellis/Johnson/Kaufman/Wise (DFW).

Although it is also true that the rural counties do add up. The ones in West Texas are very small, but East and Central TX rural counties (and counties with smaller cities such as Tyler, College Station, etc) add up to a significant amount of population. And rural/smaller town central and east TX is all now voting at least 70%, 80%, or even getting up towards 90% Republican in some places, so when you add up all those counties, they do create margins that take a lot to overcome.

You must be logged in to read this quote.

Now here, again, is where you go wrong. Go back to my example of Dallas County, which is the prime example of an "urban" county, or as you put it "cities are rapidly bluing." Yes, it looks impressive that Bush won Dallas County in 2004 and now Hillary Clinton beat Trump almost 2 to 1 in that county in 2016. But most of what is happening there is just that the White Rs are moving to Rockwall County or wherever and voting R there instead of voting R in Dallas County, and so the overall vote margin shifts across the metro area shifts much less.

You must be logged in to read this quote.

Yes, the R numbers in Presidential elections have been fairly stagnant, but even with the fluctuations in D votes, in no case have Democrats been anywhere close to the R numbers.

This is also only Presidential elections. Midterms are quite another matter, and D turnout (especially minority turnout) is persistently lower in midterm years. 2018 is a midterm year... Sure, that will be helped somewhat by it being a midterm with an R President, but it is still a midterm year...

You must be logged in to read this quote.

He has a chance, just not a realistic/very likely chance. For example, he could win if the national congressional generic ballot is significantly above 10%, maybe if it is close to 15%. The chances of that are maybe 1%-5% at the very most, I would guess.

Yes, TX demographics do cause it to trend D, but only very slowly. By the time demographics have come into play sufficiently to make TX an actual Democratic state (not just competitive), the political coalitions of the D and R parties could well be entirely different than they currently are. Although this is less likely now than it was pre-Trump, it is possible that in the post-Trump era Republicans could recover enough and shift their coalition sufficiently so as to gain enough minority support to keep Texas Republican in, say, 2040 or 2050.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

There is reason to have hope that did not exist before, a bit anecdotal perhaps but city council seats in Tarrant County (where I live and basically the premier swing county of Texas in that the margin is never more than a couple of points off in Tarrant from where it is statewide) have been flipping to the D's in places where the D's never win, such as in Euless and Hurst.

Hmm, I wasn't aware of that.

If Dems are actually breaking into the north-Tarrant suburbs, that is pretty significant and definitely new. Hurst more so than Euless - Euless has been going the way of Grand Praire/Irving since 2008 or so and has had precincts that vote Democratic near the DFW airport because they are majority-minority, but not so much Hurst and more the heavily white North Tarrant suburbs (also North Richland Hills, Bedford, Wataugua, etc). Those sorts of areas have previously been entirely impervious and completely inelastic.

Dems breaking into the North Dallas suburbs is old news, and likewise breaking more into Arlington and southern Tarrant county. In those places, the Dem gains are largely from minorities moving in and White Rs moving to other suburbs/exurbs.

But Dems popping up in northern Tarrant would be a new thing. Is there suddenly much more minority population growth there (as in e.g. Mansfield/SE-Arlington)? If so, that would be the same pattern as has previously existed in DFW and Houston, going back to 2008 or so.

Or are White suburbanites actually flipping? If there are legitimate #s of White suburbanites actually flipping to Dem, that is what would be needed for TX to start becoming more competitive.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

The Black vote is most Democratic in the Deep South and Northern Cities; in the Outer South and the Midwest, it tends to be more 80/20 rather than 90/10 or 95/5.

What actual evidence is there that the black vote is 80/20 or so in the Outer South/Midwest?

Take a look at the exit polls from 2004 & 2016 (Obama got a lot of Black Republicans, plus he turned out a lot of infrequent Black voters so it's a bit hard to use 2008 and 2012 data):

http://www.cnn.com/ELECTION/2004/pages/results/states/OH/P/00/epolls.0.html
http://www.cnn.com/ELECTION/2004/pages/results/states/PA/P/00/epolls.0.html
http://www.cnn.com/ELECTION/2004/pages/results/states/TX/P/00/epolls.0.html
https://www.cnn.com/election/2016/results/exit-polls/texas/president
https://www.cnn.com/election/2016/results/exit-polls/indiana/president

... okay I was wrong. Maybe it's more accurate to say 85/15 or 85/10/5.

Exit polls are not accurate for small subgroups. In states where African Americans are not much more than 10% of the population or so, national exit polls are unlikely to include any precincts that are heavily African American. In those precincts, if you look at precinct results for states like OH/PA/TX/IN, you will see that African Americans are most definitely voting more than 80-85% Dem. It is true that African Americans who live in less homogeneously African American precincts probably vote a bit less heavily D (more like the 80% number you have), but the problem is that exit polls are likely to ONLY include those voters (because they use stratified samples of precincts, rather than random samples of voters). So the African American voters who are most likely to be included in exit polls in states with small African American populations are not representative of African Americans in those states overall, and that biases the exit poll estimate for African Americans in those states towards Republicans.

This is the same story as with the Hispanic vote and exit poll methodological problems (google Latino Decisions' articles about methodology for more).

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Here are the presidential numbers of TX from 2000 to 2016(rounded to the thousand)
2000
D-2,434,000
R-3,800,000
2004
D-2,833,000(+399,000)
R-4,527,000(+727,000)
2008(D WAVE)
D-3,529,000(+696,000)
R-4,479,000(-48,000)
2012
D-3,308,000(-221,000)
R-4,570,000(+91,000)
2016
D-3,878,000(+570,000)
R-4,685,000(+115,000)
As you can see, the R numbers have actually been stagnant for a while, and its been the Ds who have seen voter fluctuation and gains. Its also important to note that 2016 was a poor year for D turnout, and so these numbers could be have been larger with a better candidate.

I am not saying Beto is going to win by 10%, what I am saying is that he has a chance, and that TX will not always be an R state.

So let's go back and take another look at these numbers. If you put them all in a spreadsheet and then draw a linear best fit for both the Democratic votes and the Republican votes, it looks like this:



i.e. if you extrapolate linear trends of vote gains, Texas is safe R/lean R until 2040, and is only a tossup (not even lean D, much less safe D) by 2040. That hardly looks like a blue state.

If you throw out the 2000 data and start with 2004, it crosses in 2032.

If you throw out both 2000 and 2004 and start with 2008, it crosses in about 2070...


Believe me, I would like to be wrong about this, but there is just no real support in previous election results to think that Texas is quickly shifting blue.

It is possible, of course, that it could shift more quickly, but if so, the shift will not be based upon just extrapolating previous Dem gains. It will be a different magnitude/degree/trend of Dem gains than has been seen before over the past decade or two. The best case for this happening is the possibility that Republican gains in rural and exurban areas have finally been maxed out, and that White millennials in TX replacing older voters in the electorate will start voting significantly less heavily R than their parents/grandparents. But even if that is the case, it is hard to see that coming into effect sufficiently to make TX a tossup (or lean D) before 2028/2032 or so. Which is not soon enough to make a difference for 2018, or 2020...


Other sunbelt states, such as Georgia (or Virginia) are different. In those states, demographic change turns more quickly into Democratic gains, because the demographic trends are based more on African American population growth and White Liberal population growth. Because African Americans and White Liberals vote at much higher rates than Hispanics do (especially Texas Hispanics), those states have much more of a substantive and faster Democratic trend. For example, GA is on track to pretty much be a tossup by 2020, and has been on that track since at least 2008, for a candidate who can get Obama-level turnout from African Americans and/or Clinton level gains among suburban whites. But for Texas, the timeline is more like 2040 or so.

Colorado is also different, because Democratic gains there are largely based on white liberals/moderates in Denver/Boulder flipping Dem. Minority/Hispanic growth makes some difference, but is comparatively small and slow (like in Texas). If you cut out East Texas, West Texas, and South Texas and had a state centered on Austin, the trend there would be a lot more similar to Colorado.

Arizona is also different because Democratic gains there... have not yet actually materialized... If Democratic gains do materialize there, it seems to me that this is again mostly a result of a shift in White voters, flipping from R to D (similar to Colorado). Yes, Hispanic population growth helps, but only on the margins. Presuming Sinema wins, she is not going to win just because Hispanics suddenly start voting in huge numbers in AZ. She is going to win because suburban whites are turned off by Trump and flip Dem, and because (in comparison to Texas), Arizona has many, many fewer exurban White Rs and rural White Rs.

While Arizona has also been fairly inelastic (similar to Texas), in comparison to Texas it has more of a recent history of splitting tickets and occasionally supporting Democrats. White voters (in particular white suburban voters) in Arizona have previously been much more elastic than white voters in Texas, and willing to elect people such as former Governor Janet Napolitano. AZ was also considered sufficiently elastic to be briefly somewhat contested in the 2004 Presidential election, and probably voted a bit more R than it otherwise would have in 2008 (McCain homestate) and 2012 (Romney mormon vote), so on a baseline level it is a tad less Republican than the last decade's Presidential results would indicate.

From 2004 - https://www.thenation.com/article/arizona-turning-blue-kerry/

Whereas white suburbanites in AZ voted sufficiently for Napolitano for her to win in a less rural/exurban state than Texas, white suburban voters in Texas would previously never have considered voting for a Democrat for anything, going back to roughly 1970 or so.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

First of all, it is wonderful that Beto raised $10 million etc. I am not in any way trying to say that is bad or a wasted effort, just that demographics work slowly and changing a state like TX is a long term process.

It is also great that he is energetically campaigning and visiting rural counties etc to try to reduce margins. However, previous candidates such as Bill White did that as well (and also raised a pretty significant amount of $). Despite his energetic campaigning, I would be extremely (pleasantly) surprised if Beto does anything other than lose the rural areas in a similarly sized landslide to what has happened with previous Dem candidates.

Im sorry, I just cant resist.
First of all the graph, I would just like to know
1. Where did you make that, I would love the software
2. Do you think voter growth is linear? Cause in the sunbelt, its not. Its, as SCNCmod put it

1) It is a simple graph made from an excel spreadsheet.
2) Of course, strictly speaking, basically any trend is not going to be strictly linear. But you are not being clear enough (as far as I can understand what you are saying) on what exactly you think is nonlinear and in what particular ways that are important for me to be able to really comment on it. There are multiple different nonlinearities of different importances that push in different directions, and I am not really sure what specifically you are referring to. But in terms of the effect of demographics on what % of the vote states give to Ds or Rs, for the purposes of a general discussion on an internet forum, I would say effect of demographics is not really too far from being linear. In principle, to do a full proper analysis, you would want to make a complete demographic model breaking TX down into population cohorts of different ages/races, and then you could calculate the effects of different hyptothetical future turnout rates and rates of supporting different parties among each cohort. The result of this sort of analysis would be to show TX getting a small amount more Democratic each election cycle - about a percentage point or two - (other things being equal).

BTW if you change the graph I posted to use exponential trends rather than linear, it crosses earlier, but only a few years earlier (it is still a matter of decades).

You must be logged in to read this quote.

The metro areas in Texas have grown quickly, not the cities per se. Most growth is in suburban areas. Whether Texas' growth has slowed down, sped up, or remained constant depends on whether you are thinking in terms of percentage growth rates or raw numbers - but regardless of how you want to look at it, my argument is that while it does have a real effect, the effect of demographic change is gradual, and not the sort of thing that can create massive political shifts by itself in the course of just one or two election cycles.

Yes, turnout (Democratic/Hispanic turnout) did increase from 2012-2016, but you should remember that is also a result of how abysmally low Democratic turnout was in 2012. The other thing is in quoting specific #s like 38% to 41% from the ACS voting supplement data, you should keep in mind the limitations of the data quality.

You must be logged in to read this quote.

I am sorry, but you will need to explain more clearly in what particular way you think the #s are nonlinear/parabolic/whatever if you want me to be able to respond intelligently. When you are arguing that the change in the D #s is

#1 ---


#2 ---


I am not sure if you are trying to argue that the change in Dem votes is like #1, but to me, if anything it looks more like #2.

Broadly though, the #s that you posted for GA and AZ are the result of a variety of interacting factors (as for TX). I would caution about is trying to explain increases or decreases in those in particular years in terms of demographics. Demographics plays an important roll in explaining long term trends, but it has a more limited roll in explaining why e.g. the GOP # went up by more in 2004 than other years. There are multiple factors here besides demographics, such as e.g. Bush having the best ground game of any recent GOP candidate, 2004 being the best recent GOP Presidential year, turnout generally being up in 2004 (in large part because of Iraq) and also in elections since then, as compared to 2000 and the 1990s. I would certainly agree that e.g. the declining white share of the population also plays a roll in GOP votes going up by less in subsequent years, but it is not the only factor here and not the only part of the story.



Finally, here is something that can maybe help us look at the question. I made a quick demographic vote calculator spreadsheet for the 2016 Presidential vote in TX:



And here's a link to the spreadsheet so you can try changing numbers:

http://s000.tinyupload.com/?file_id=26275627319247150679

The numbers at the top are citizen voting age population. The light blue are the same turnout numbers you are using from ACS census data. From that, the # of people estimated to have voted in 2016 from each racial group is calculated. However, if you just use the raw ACS turnout percentage, this overestimates actual votes cast by about 600k, because the ACS data is not accurate and relies on self-reported voting (people lie or incorrectly claim to have voted).

So when you have "Y" in the yellow box, it uses the numbers (a bit lower than the ACS turnout numbers, so that the number of votes cast matches the actual number of votes cast in the Presidential race).

The green #s are my estimate for how each. This is slightly different from exit polls (mainly whites are more Republican and Hispanics and African Americans are more Democratic than exit polls indicate), but while I don't know exactly what the true numbers are, I am pretty confident that these are more accurate than exit polls. If you think the exit polls are closer to reality, you can use something closer to the exit polls by filling in numbers yourself (but you need to adjust something to get the overall Dem and R vote to match what Clinton and Trump actually got.

The orange #s calculate the estimated # of votes cast for Ds and Rs within each racial group, and the purple #s add them all up to the total vote. Red #s show the share of the electorate made up by each racial group, given the citizen VAP and the assumed turnout %s based on the ACS data.


Anyway, so from this starting point, we can change numbers to see how the vote would be different, hypothetically, if we change turnout, underlying population, or the crosstab %s each racial group gives to Ds and Rs.

In a previous post you said:

You must be logged in to read this quote.

This spreadsheet can allow us to see what the effect of increasing Hispanic turnout by 2% or 4% (e.g. increasing from 41% to 45%.

Since I am using the slightly lower turnout #s (to match the overall vote) rather than the raw ACS data, I increase the Hispanic turnout from 37.1% to 41.1%. The effect of this is to increase Hillary Clinton's vote % in TX from 43.12% to (something like) 43.73%. This is nothing remotely close to enough, by itself, to make Texas a lean D state (much less a tossup).

In fact, even if you raise Hispanic turnout (of Citizen VAP) all the way to the same level as White Non-Hispanic turnout (of Citizen VAP), something that is not remotely possible, Trump still would have won TX in 2016 by about 49.21% to 46.19%.

Note that if you use the exit polls estimate that Clinton won more like 60% of Hispanics rather than my estimate that she won more like 72%, then this would have even less of an effect.

And even if you raise Black turnout, Asian Turnout, and Other turnout ALSO to the same level as White Non-Hispanic turnout, Trump STILL narrowly wins TX in 2016 by about 48.31%-47.14% over Clinton.



So you can see very clearly from this spreadsheet that just increasing Hispanic turnout by any amount that is remotely plausible is just not remotely enough to make Texas competitive, much less to make it a blue state. In order for it to become a blue state, you need either:

1) To wait quite a while longer for long term demographic change for the White Non-Hispanic share of the Citizen Voting Age Population to go down by quite a bit.
2) To increase Democratic support, in particular among White Non-Hispanics. White Non-Hispanics are still a majority of Texas voters, and they vote very heavily Republican. In theory, Dems could also win a higher % of the Hispanic/Asian votes (or the Black vote, but surely that is maxed out), so in practice increasing Democratic support means increasing Democratic support among Whites.
3) Get all the people who voted 3rd party to vote Dem.

You can, in principle, accelerate this slightly by increasing turnout a bit, but that only makes a marginal difference of at most a few points.


Remember, this is all with Presidential year turnout. The math is more daunting still in a midterm year like 2018 when there will be lower minority turnout. This is why it is very hard to see how Beto can win - if he does, the primary thing that can make a difference is if he can do MUCH better with white voters than TX Dem candidates (including Clinton 2016) have done recently.