Early voting in Texas ended on Friday and today the Secretary of State estimated that approximately 60% of all voters expected to cast ballots in the 2008 Texas primary -- 2 million of the 3.3 million total -- had voted during the early voting period. Not surprisingly, the vast majority of those early ballots were in the Democratic primary.

The estimate is based on the slightly more than 1.2 million who voted in the 15 most populous counties during a 10-day period that ended Feb. 29.

The Secretary of State's Office tally for the 15 most populous counties showed that 890,188...voted early in the Democratic primary -- more than the Texas Democratic Party has seen in the past two presidential primaries statewide for early and Election Day voting. [...]

On the Republican side, 303,338...of registered voters in the 15 biggest counties cast early ballots.

As for whom the early voters favor, the results appear to be mixed. The final Public Strategies Belo tracking poll shows Obama up 54-46 for the entire early voting period (although the final polling period had Clinton closing the gap to 4 points) and SurveyUSA has Clinton up 50-48. For all the difference in the early vote numbers, however, both polls result in a virtual tie, Clinton up 1 in Belo and Obama up 1 in Survey USA. And I have to say, despite the fact that Hillary Clinton appears to have the late momentum in Texas, my gut is that we're going to see essentially a tie tomorrow (by which I mean, at best, a 1 or 2 point win for Clinton.)

First of all, the Survey USA and the Public Strategies poll have very different Hispanic turnout models (Survey USA: 32%, Belo: 25%.) The reason both polls still result in a tie though is that SUSA has Clinton up by 31% among Hispanics while Belo has a 44% lead. In other words, both polls agree with what we know to be true: the higher the Hispanic turnout, especially relative to African-American turn-out where Barack Obama enjoys prohibitive leads, the better Hillary Clinton will do. As Belo puts it:

...if African American turnout equals or surpasses Hispanic turnout, Obama will win.

Looking once again at the early vote, this graf from the Belo poll stood out:

Totaling all of the nights of tracking together, Barack Obama leads in early voting 54%-46%, and African American turnout has been equal to Hispanic turnout.

Add to that this bit from the Star Tribune article on actual early vote turnout:

The biggest percentage turnout among Democrats was Hidalgo County, in the predominantly Hispanic Rio Grande Valley, where 18 percent of registered voters cast ballots.

If it's indeed the case that so far Hispanics and African-Americans have voted early in equal numbers AND that early votes have come disproportionately from a region where Clinton is strongest, that doesn't bode well at all for Clinton tomorrow, the late Clinton surge notwithstanding. I would, of course, be happy to be wrong.

Update [2008-3-4 1:31:50 by Todd Beeton]:Since composing this post, I've come across a couple of interesting pieces of information that lead me to believe that Clinton may actually have a better day tomorrow than I originally thought.

First of all, I saw Burnt Orange Report's breakdown of early voting by county. Note that while Clinton's stronghold of Hidalgo County is responsible for the highest percentage of registered voters voting early (of those counties counted) that was perfectly in keeping with their 2004 early vote performance. Travis County on the other hand, home of Obama stronghold Austin, showed a huge percentage increase between 2004 and 2008, indicating that extra early vote efforts were made there. Note also the top two counties in raw early vote totals are Harris and Dallas Counties, home to Obama strongholds Houston and Dallas respectively. One may draw from this that as a result of the excellent Obama early vote efforts, the pool of Obama voters who have yet to vote is simply smaller.

Which leads me to the new ARG poll, which finds Clinton up by slightly more than I've posited, at 50-47%. Interestingly, ARG concurs with what is apparent from looking at where the early votes are coming from: that Obama is up among early voters, but finds Clinton's late surge is helping her immensely among those who intend to vote tomorrow.

Obama leads Clinton 52% to 47% among early voters (34% of likely Democratic primary voters), while Clinton leads Obama 52% to 45% among in-person voters.

It is ARG after all, so take with requisite grain of salt, but it's not insignificant that this 50-47% result represents a net gain for Clinton of 10% since last week.