October, 2010:

After one full week of early voting, the proportion of votes coming from Republican State Rep districts was tilted about two percentage points more towards the GOP than it was in 2006. I’ve argued that this is nothing to be particularly excited about one way or another, as I believe it represents a shift in behavior (more people voting early) than a change in the electorate. That’s largely predicated on Democratic voters coming out in equivalent numbers, of course. How do things look now, after three more days of Early Voting?

2010 Week One 

Strong R = 47.7%
Medium R =  9.4% 

Medium D = 17.9%
Strong D = 22.9% 

Total R = 57.1%
Total D = 40.8%  


2010 Week Two 

Strong R = 46.3%
Medium R =  9.3% 

Medium D = 17.0%
Strong D = 25.3% 

Total R = 55.6%
Total D = 42.3%  


2010 Overall 

Strong R = 47.2%
Medium R =  9.4% 

Medium D = 17.6%
Strong D = 23.8% 

Total R = 56.6%
Total D = 41.4%  


2006 Overall 

Strong R = 43.7%
Medium R = 11.2% 

Medium D = 19.2%
Strong D = 23.2%

Total R = 54.9%
Total D = 42.4%

So far, Week Two is about three points more Democratic than Week One. I should note that in terms of Strong D percentage, Tuesday was more Democratic than Monday, and Wednesday was more Democratic than Tuesday. There’s still a ways to go to get the overall mix closer to 2006 – and note that Week Two is still slightly redder than 2006 was – but if this continues, that will happen. Better news for Democrats, but still not where they want to be.

And then there’s still the matter of guessing how many people will be left to come out on Election Day itself. As of yesterday, 288,568 in person votes had been cast, with 48,498 mail ballots having been returned. That’s 337,066 total votes. To put this in some perspective, total Harris County turnout in 2006 was a smidge over 601,000, and in 2002 it was a bit over 652,000. The early vote total so far is almost 57% of the total vote from 2006, and nearly 52% of the total vote from 2002. I think final turnout this year will easily exceed 2006 levels, and may be higher than 2002, though probably not by much. Election Day itself may be a fairly quiet affair. Let’s see where we stand after Friday, and we can place our guesses for the final number then.

UPDATE: Another 44,324 votes were cast Thursday, bringing the total early in person vote count to 332,892. Of this, 180,984 were cast in the first seven days, and 151,908 have been cast since Monday. Thursday was the most Democratic day yet:

Strong R = 43.8%
Medium R = 9.1%

Medium D = 17.0%
Strong D = 27.5%

That’s 52.9% R, 44.5% D for the day, which is two points more Democratic than 2006. We’ll see what today holds, and I’ll run the final numbers tomorrow or Sunday.

Let’s start with the breakdown of the early voting numbers after one full week:

2010 Overall 

Strong R = 47.7% 
Medium R =  9.4% 

Medium D = 17.9% 
Strong D = 22.9% 

Total R = 57.1% 
Total D = 40.8%  


2006 Overall 

Strong R = 43.7% 
Medium R = 11.2% 

Medium D = 19.2% 
Strong D = 23.2%

Total R = 54.9% 
Total D = 42.4% 

Paul Burka, as is his wont, has proclaimed the early vote totals as gloom and doom for the Democrats. I think that’s a fallacious conclusion based on this limited evidence, for three reasons.

1. A surge in early voting does not mean a commensurate surge in final turnout. Didn’t we learn this lesson in 2008? Look what people, including me, were saying about Harris County turnout in 2008 based on the gigantic, unprecedented early vote numbers. County Clerk Beverly Kaufman predicted 73% turnout, on the assumption that as many people would show up on Election Day as had done during early voting. To put it mildly, that didn’t happen. Sure, final turnout was up a bit from 2004, but for the most part the huge early vote numbers represented a shift in behavior, not a surge in new participants. On Election Day itself, turnout was much lighter than expected because we’d basically run out of voters. In the absence of evidence that there’s a large number of non-habitual voters casting ballots this year, the most likely explanation of what we’re seeing is the same thing, a shift in behavior from Election Day voting to early voting. As was the case in 2008, there has been a concerted effort to get people to vote early this year, thanks to the fire that destroyed the county’s voting equipment. Why are we surprised that people are doing what they’ve been repeatedly told to do?

What we are seeing here is more people voting early than they did in 2006. Most likely, more of these people are Republicans, though we can’t really quantify that from these numbers; I’ll have more on that in a bit. That brings me to point 2:

2. There’s nothing unusual about the partisan pattern of early voting this year. Am I the only one who remembers that Republicans traditionally dominated early voting, with Democrats doing better on Election Day? No, Dr. Murray remembers that, too:

Murray said the numbers suggested to him that “Republicans look to be in better shape in Harris County, which, until 2008, has been the pattern. I don’t know what pattern we’re going to see in 2010.”

In 2008, the Democratic Party relentlessly flogged the early vote message, and Democratic voters responded. Prior to that, Republicans won the early vote. They also did pretty well in Harris County in those days, and I’m sure they’re feeling good about that possibility. But again, the question is about how much of the vote has shifted to the EV period, and how much they’ll have in reserve on Election Day. As giddy as Democrats were about the early vote numbers in 2008, that feeling turned to nervousness as it was clear that the Republicans were catching up on Election Day. With a normal level of Democratic turnout, things will largely even out.

The Republican districts aren’t as Republican as the Democratic districts are Democratic. Putting this another way, here’s Dr. Murray again, analyzing some 2008 numbers:

One way to look at this is to take precincts that the Republican and Democratic presidential candidates won decisively in 2004 and compare registration gains this year in these precincts since the March 4th party primaries. I selected the top 20 precincts that the Republican nominee George W. Bush won with a total vote of more than 2,340. I then pulled out the top 20 precincts for Democrat John Kerry where he received more than 1,380 votes four years ago. Since March 4, the voter rolls in the top Republican precincts from 2004 have added 7.52% new total registrants, and the top Democratic precincts have added 7.85%. So by this measure, the registrations over the last six months look like a push for the two parties.

But, there is a very big difference in these two sets of precincts. The top 20 Democratic precincts were, in 2004, and remain, heavily minority boxes with very few Republican voters. For example, in the March 4, 2008 party primaries these precincts cast 25,676 votes in Democratic primary and just 1,097 in the Republican primary. This means the registration gains this year will almost certainly add to the total vote for Harris County Democratic candidates. The top twenty 2004 Republican precincts were, of course, carried by George W. Bush, but there was a sizeable Democratic minority (16,990 of 75,583 voters) in these predominately GOP boxes four years ago. That Democratic minority has grown since November 2004, if the March primary is any indication. For example, in precinct 764, which has had the largest registration gain in the county since November 2004 (+4288 as of yesterday), the vote in the March 4 Democratic primary was 2,185 compared to just 999 Republican primary voters (Precinct 764 in 2004 was split for the 2008 election and now also includes precinct 388). Overall, the top 20 Republican precincts in 2004 had almost as many Democratic voters in the March primary (18,869) as Republican voters (19,551).

And here’s my way of looking at that, which is to compare the percentages that Bill Moody and JR Molina got in the strong and medium partisan districts from 2002 to 2006:

Dist 02 Moody  06 Moody  02 Molina  06 Molina
=============================================
126      27.8      35.2       26.3       32.9
127      27.0      33.4       25.0       30.6
128      36.2      40.7       34.1       37.9
129      30.2      39.2       28.1       36.6
130      20.8      28.8       19.0       26.3
132      26.2      36.1       24.4       33.3
135      32.1      40.2       30.5       37.5
136      22.2      32.5       20.4       29.1
150      25.4      32.9       23.3       30.7

Dist 02 Moody  06 Moody  02 Molina  06 Molina
=============================================
131      78.3      81.2       78.0       79.8
139      81.3      83.1       81.1       81.5
140      68.8      68.9       69.3       68.1
141      77.0      76.9       76.4       75.5
142      82.8      80.9       82.6       79.3
143      71.8      69.4       71.8       68.6
145      68.3      68.9       69.9       68.5
146      71.6      74.8       70.8       72.4
147      81.6      82.6       81.6       80.6
148      63.5      67.7       63.0       65.3

Dist 02 Moody  06 Moody  02 Molina  06 Molina
=============================================
133      36.7      43.8       34.7       41.1
134      38.2      51.7       35.5       47.2
137      51.1      55.8       49.8       53.8
138      37.5      45.1       35.2       41.8
144      39.5      44.9       37.8       42.8
149      42.4      48.7       40.7       46.6

That’s a lot of growth in Democratic strength in four years, at a time when there was less organization in the county, and no funded GOTV effort. It’s demographic change, and it’s something I’ve referenced before: The Republican base in Harris County is stagnant, while the Democratic base is growing. The result was that in 2006, a Democrat in a lower profile race might get 75% of the vote in the Democratic districts, and 35% in the Republican ones. That means that for every 1000 votes, the Dem would have a net of 400 in the D districts and -300 in the R ones. At those rates, you’d need 1333 votes in a Republican district to get the same +400 net for a Republican candidate. And as it happens, 1333 is 57.1% of 2333, which is the combined vote total. And that’s before we take into account four more years of this kind of change.

Now again, these are all very rough and approximate calculations, which rely on a number of assumptions. You’re much better off getting hold of the roster of people who’ve actually cast votes and figuring out their partisan history rather than relying on this kind of hocus pocus. My point is not to put a smiley face on what has happened so far – I do agree that Dems are lagging, and based on the conversations I’ve had with people over the weekend, I’m the optimist – it’s simply to knock down the assumption that because more votes have been cast in Republican districts that all hope is lost. Hell, in 2008, more early votes were cast in the Republican districts, and we know how that turned out. The Democrats’ job is no different now than it was before early voting started, and that’s to get their voters to the polls. Whether that happens this week or next Tuesday doesn’t really matter, as long as it happens. If it does, we’ll be okay. If not…I don’t want to go there. Just vote, and make sure everyone you know does so as well.