Does the Sotomayor Choice Make Texas a Swing State?

Matthew Dowd thinks it might, depending on how Republicans in the Senate handle the Sonia Sotomayor nomination to the Supreme Court by President Obama.

“If they don’t get back to a place where they are getting roughly 40 percent net of the Hispanic vote, there is no way they can ever win,” [Dowd] said.

Now, Dowd was talking in national terms when mentioning that 40% barrier, but let's look at this within the context of Texas. Hispanics made up a shade more than a third (36%) of the Lone Star state's population in 2007 according to the Census estimates and comprised approximately one-fifth of the 2008 electorate there (based on exit polls). In raw data terms, that's 8.8 million Hispanics in Texas, 1.6 million of which voted. In November, Obama won 63% of the Texas Hispanic vote while losing by just shy of one million votes to John McCain overall in the state. That left Obama with a +450,000 vote "Hispanic margin." In other words, despite beating McCain by nearly half a million votes among Hispanic, the president still lost by one million votes in Texas.

But the question is: How many Hispanic voters would be energized by a controversial Supreme Court confirmation process involving the first potential Hispanic justice, and would that be enough to overcome that one million vote deficit? Possibly. On one hand, Texas is growing at a pretty good clip and a lot of that growth is Hispanic growth. But on the other hand, Texas, in a more competitive environment, would likely see increased turnout. The former is much more difficult to simulate than the latter, but let's look at turnout first and see if we can get at least half way to an answer to this question.

First, let's construct a model based on the 2008 election data we have. If we regress FHQ's final polling margin averages by state, number of electoral votes, and a state's party lean (a dichotomous variable where 1 = Democratic lean and 0 = Republican lean)* on the final turnout figures from 2008, we get a decent model for the purposes of prediction (The R-squared isn't great -- .3 -- but let's keep this simple.). In reality, Texas had a 54.7% (voting eligible population) turnout rate in 2008. Under this model, however, the Lone Star state is predicted to have had a turnout rate of 57.6%. In other words, we have some error present; most likely due to some level of omitted variable bias. Again, though, simplicity is the goal here, not elegance. That said, if we assume that Texas was a dead-heat in the polls leading up to the election (I dropped the +11.66 McCain advantage in the polls down to a +0.66 McCain lead), the turnout rate would have increased to 60.1%.

Now, if we assume the same exit poll distribution among racial categories prevailed in the Texas electorate -- 20% Hispanic, 63% of which voted for Obama -- the president would have inched approximately 50,000 votes close John McCain. That's a drop in the bucket when compared to a nearly one million vote deficit. But if we assume that the GOP caucus in the Senate balks at the Sotomayor nomination, damaging the party's standing with Hispanics even further, that mere drop in the bucket may turn into a tide against the GOP in state's with a dense Hispanic population. For example, if we assume, based on 2008 population and turnout statistics, that the Texas electorate was 25% Hispanic (instead of 20%), 75% of whom voted for Obama (up from 63%), the president would have increased his Hispanic margin from 450,000 votes to over one million votes relative to McCain. That half a million vote difference would have cut McCain's statewide advantage in half assuming all other racial categories behave as they actually did (in terms of percentages of the electorate) in the 2008 election.

The problem here is that this simulation is done in terms of the 2008 election; an election that is obviously in the history books. What's missing, then, is an accounting of the population growth to occur between now and 2012 (Oh, and the actual level of incitement a fight over an Hispanic Supreme Court nominee triggered. But that's a different story.). If you look at the Election Data Services estimates (Table B), it looks as if Texas has gained between 400,000 and 500,000 people every year since 2000. That would place the state's population at somewhere around 26 million people in 2012. If the same 60.8% of the population was voting eligible at that point -- and that doesn't include latent Hispanic voters activated by a court nomination fight -- there would be approximately 16 million voters in Texas in 2012. One thing to note is that we are assuming uniform growth across all categories of ethnicity. In other words, the expansion of the Hispanic margin between the Democratic and Republican candidates would be counteracted by a similar increase in the white margin. As noted above, though, the population growth will not necessarily be uniform.

If, then, we further assume the same 20% of the electorate is Hispanic, 63% of whom vote for the Democratic candidate, the Democratic Hispanic margin over the Republicans would grow to around one million votes. Assuming a 25%/75% split, as was done above, would increase that Hispanic margin to about 2 million votes. But is that bar too high or too low to account for active Hispanic voters moving over to the Democratic column or latent Hispanics being activated by a partisan battle over Sotomayor?

And, as always, will it even matter three years down the road when the next presidential votes are cast?


*The hypothesis here is that Democratic states -- especially solid Democratic states -- would see increased turnout regardless of competitiveness while solid Republican states would witness lower turnout rates.


Recent Posts:
Two Huckabee Slips in One Day?

Is Charlie Crist Running for Senate or Vice President?

Memorial Day Travel and the 2012 Bumper Sticker Battle