Wednesday, November 6, 2013

PostElection: Virginia Governor

I'm a bit late to this, and despite that I haven't come close to reading everything that people have written so I'm not sure whether people have covered this or not, but that disclaimer aside:

The interpretation of the Virginia governor's race, with Terry McAuliffe defeating Ken Cuccinelli by (based on results so far) 2.5 percentage points, appears to be dominated by questions about why it's so close -- rather than questions about why McAuliffe won.

That's because of two things. One is because polls had McAuliffe with a bigger lead; it was estimated at 7.2% by HuffPollster, which is pretty much the gold standard for these things. So there's certainly something to be explained about the difference between the polling and the actual results.

The other is the order in which the votes were counted. Republican precincts reported early -- which is the normal pattern in Virginia, as veteran election-watchers were saying all last night -- and so Cuccinelli actually was leading until very late in the counting. It's impossible to prove exactly how this played out, but I'm fairly confident that if the votes had been counted in the opposite way, with McAuliffe getting out to a big lead, settling down for most of the night into the 5-7% range, and then Cuccinelli catching up quite a bit after the networks had called the race and the candidates had given their speeches, that the election would have been perceived as a much more solid win.

Getting back to the first step here: the legitimate question here about the difference between the polls and the results is not the same question as the question of who did well and why. Which is why I agree with Greg Sargent and Ezra Klein that it's a bit bizarre to call the election a win for ACA foes.

As far as the polling shift, here's my two cents that I'd begin with. Compared with the HuffPollster estimate, the actual results (so far -- remember that they sometimes shift for several days after an election) are the following changes:

McAuliffe  + 2.7%
Cuccinelli  + 7.4%
Sarvis         - 2.5%

If I'm reading the exit polls correctly, most voters who supported the Libertarian Party candidate would have supported McAuliffe in a two-candidate race. I haven't looked into the details of the pre-election polls, but at least one strong possibility is that conservative libertarians (or perhaps conservatives who just didn't like the Republican candidate) came home, while liberals did not. That certainly could have something to do with the themes candidates campaigned on in the final days, but I'd suggest that Ron Paul's high-profile appearance for Cuccinelli is a more likely suspect. And then I'd also guess, and this is a bit more speculative, that Democrats were more likely to trust the polling (and therefore assume that it was safe to vote against a candidate they preferred to win) than Republicans (who, believing it was a close race, came home even if they weren't thrilled about it).

One more thing. While I was checking that one in the exit polls, I noticed something that I did not expect at all: outgoing Republican governor Bob McDonnell had a pretty solid 52% approval rating among exit polled voters. If our interest is in who won and why -- and not in the real but less interesting question about the polling -- then that suggests that Virginia was in fact probably ready to elect a Republican, and that the net effect of the candidates and campaigns probably favored McAuliffe, not Cuccinelli. Which is certainly long way from nailing down the effect of any specific issue, of course, but does suggest that a focus on "why was it so close" really is a backwards way of trying to understand this election.


  1. I wasn't fallowing the polling to heavily, but it looks like there may have been a small but significant number of voters who ultimately came home for Cuccinelli but kept telling pollsters they were undecided. That is they were Republican partisans of some degree or another but didn't want to admit to a pollster they were voting for Cuccinelli. If you add that group to the 2.5% group of libertarians who came home from Sarvis that could easily be most of the swing right there.

  2. Don't forget, though, that the exit polls are a biased sample of Virginians: namely, those that voted.

    Since Cuccinelli out-performed the polls, I would suspect that McDonnell did the same. The number of Cuccinelli voters that didn't approve of McDonnell would be pretty small.

    Looking at the exit polls...whoa, those are wierd.

    61% college grads? 85% college attended? 40% make over 100K? 28% tea party supporters? 20% abortion is the most important issue?

    This is a turnout story.

    1. 40% make over 100K? Couldn't anybody else take time off from work?

  3. I have not been following Bob McDonnell's polling very closely, either, but it seems as if his recent numbers have been quite a bit lower than 52%. If I am right about that (a big if), then much might be explained by problems with likely voter models. It's probably very hard to determine likely voters in an off-off year election in any case, and there are all sorts of things that might have led to an electorate that was significantly more pro-Republican (and thence pro-McDonnell) than expected. I guess anti-ACA agitation might account for some of it, but perhaps its also the case that some of the margin McAuliffe was getting in pre-election polls was pretty soft -- i.e. more of a vague anti-Cucinelli vote than anything else. As numerous commentators have pointed out, McAuliffe was not exactly the best candidate, in fact I think Chait said he is exactly the kind of Democrat many on the left would like to vote against and then the GOP went and nominated Cucinelli. Maybe a lot of those people saw the polls, shrugged, said "well, that's decided," and just stayed home.

  4. "If I'm reading the exit polls correctly, most voters who supported the Libertarian Party candidate would have supported McAuliffe in a two-candidate race."

    Yea, I saw that too. More than twice as many liberals as conservatives voted for the libertarian.

    1. I expect many liberals and others saw this election choice as the lesser of two jerks. If liberals read the polls and saw that Mc. was safely ahead, they may have failed to hold their noses and add to his numbers. Why they then voted for the Lib Party is beyond me or this analysis.

    2. Hmm, why vote for Sarvis?

      Looking at his website I notice this on his 'issues' scale:

      -Supports SSM.
      -Wants to reduce taxes but without talk of a 'flat tax' or some other "stick it to the poor" concept.
      -Supports getting 'fad teaching systems' and regulations off of teachers
      -Supports Marijuana legalization. Bah, he has the "if it doesn't hurt others, it's ok" line.

      He's against the ACA, but there's a lot of liberals who are against it. I'm guessing that they assume his desire to turn Medi into a block grant won't go anywhere

      -supports abortion (probably more so than most of the left)

      -Spotted a line saying he supports 'internet freedom'. That's a winner with a small group right there

      - His environmental stance is actually rather leftish sounding. It's basically "If you pollute, you pay for it." I would've expected a libertarian to say to just let the market handle it all.
      -Nothing about standing against the civil right's act or other such items that tends to turn the left away from libertarians

      Bah, he found a way to get to my views on the Death Penalty from the opposite end (He supports it but realizes that the 'government being able to kill people at will' may not be a good idea. I'm against it but understand that the public feels its necessary. The result for both is "Allow it but if anything make it harder to make it happen")

      There's not too much to hate about him from the left, especially if you either believe he won't have a chance to touch Medi or don't consider it a dealbreaker. I get a little iffy on voting third party, but I'd vote a democrat who had Sarvis' Issues page (losing Medi is cringe-worthy, but so was Obama loving the Patriot act and thinking he could do a better NCLB),

    3. A quick note though that I may not be the best one to judge on this. I actually have this out of left field idea of the libertarians joining the Democratic party especially since they seem to have fully separated from the Tea Party now. They'd end up being a more moderate wing that dislikes the hyperPC and finds a way to both support civil rights and take power from the "NSA powered Patriot Act emboldened" Fed.

      Though if they also find a way to handle the poor in non Social Darwinian terms and without using the word 'bootstraps' then I fear I might end up identifying as them. oh well

    4. Dakarian, thanks for the response. I used to be in the Libertarian Party. Sarvis is pretty typical of the party moderates (the other main division being radical-anarchists) There may be some areas in the US with more Constitutionalist-type libertarians, but a gay friend of mine recently told me that she attended a Libertarian Party event in Alabama that was extremely gay-friendly. In any case, hardcore conservatives usually end up in the Constitution Party.

      A lot of voters don't fall into the tidy ideological boxes that party leaders prepare for us. For example, I'd imagine that some Sarvis "liberals" might come down like this: support legal marijuana/gay marriage, think Snowdon has the right idea, want strong environmental protections, believe in gun rights but ok with some new regulation, support Clintonian economics and a Ron Paul-type foreign policy. In other words, an economic moderate who leans libertarian on most other issues. This person is definitely closer to Democrats than Republicans, but might also be comfortable with a Gary Jonson-type libertarian.

      "I actually have this out of left field idea of the libertarians joining the Democratic party especially since they seem to have fully separated from the Tea Party now."

      Interesting idea, but the movement libertarians seem to be pretty much following Rand Paul right now. If true liberals decided to take back the Democratic Party on civil liberty and foreign policy issues, I'd be very tempted to join them. I think there's a real appetite among voters for that kind of candidate, but it seems like that message may have a better chance of taking hold in the GOP right now. What I'd really like to see is an NRA-like organization for civil liberty issues (the ACLU is mostly about litigation, not politics). Targeting these issues specifically would be really effective, I think.

  5. Isn't there a "hack pundit" rule that Virginia always goes against the incumbent presidential party in governor's races? The pattern has held since 1977. And off years are always bad for the incumbent party, especially when approval ratings are as low as Obama's are now. So this should have been the GOP's race to lose.

  6. This comment has been removed by the author.


Note: Only a member of this blog may post a comment.