Smoothed trendlines for Illinois gubernatorial race
Since we've unveiled the Daily Kos Elections Poll Explorer, we've always led off the discussion by talking about the Senate races. That makes sense, since it's very unusual to have an election where it's so totally unclear at this late a date which party will have control of the Senate. On the one hand, it's amazing to have the odds of a Republican Senate takeover trading reliably, for months on end, in a band between 45 and 55 percent odds. On the other hand, at some point, you start running out of dramatic things to say about the minor fluctuations within that band.
And that doesn't even address the boring reality of a Republican takeover. Sure, it would suck mightily from a "what's wrong with these people?!?" perspective, and it would throw a monkeywrench in the Democrats' ability to get judicial or executive branch nominees confirmed. However, from a policy standpoint, the Senate would just keep on doing what it does best, which is doing nothing; it would mostly just mean Harry Reid doing the filibustering, instead of Mitch McConnell, and if a budgetary matter happens to clear the Senate via reconciliation, then it would just give Barack Obama a chance to get better acquainted with his barely-used veto pen. That stultifying reality may be why the Senate race is generating so little interest in the broader public this year, even as it's totally captivating to pundits, modelers, and their ilk.
Which brings me to the gubernatorial races, which draw surprisingly little attention from the Beltway press. While there are at least half a dozen Senate models from major media organizations, our Poll Explorer is the only one that models overall gubernatorial odds. That's baffling, because the state houses are where it's at, these days. That's partly because Congress has ground to an effective halt, while there's less gridlock in state houses these days; where split control often gummed up many states, now there are more and more states with either an all-blue or all-red trifecta. You can see that in the increasing divergence in policy between blue and red states, with many blue states adopting same-sex marriage and making bold moves on the environmental front, or red states refusing Medicaid expansion and enacting stricter abortion restrictions.
And another consideration is that many of the governors who are elected this year will still be in power through 2022. That means they'll be the ones signing off on the next round of redistricting. The more Democratic governors we elect this year, the more governors we'll have in 2022 signing off on Dem-friendly Congressional maps (or vetoing maps drawn by GOP legislatures, forcing neutral court-drawn maps). Driving a stake through the heart of Republican gerrymandering is the only way Democrats will have a shot at taking back the House next decade.
With that in mind, two Democratic-held gubernatorial seats that were in woeful shape last week have suddenly burst back into competitive shape, thanks to new polling. We'll look at that those (and the Senate map) over the fold:
Distribution of likely number of Democratic gubernatorial seats
One of those races—as you can infer from the title—is Illinois. Pat Quinn's never had an easy go of it; he inherited the office when Rod Blagojevich was convicted, and then he won both the primary and general elections in 2010 by the skin of his teeth. He wasn't even expected to be the nominee this year, but when Attorney General Lisa Madigan surprised everyone by declining to run, Quinn found himself with the nomination, but facing a generic rich guy in Bruce Rauner instead of an abrasive Downstate social conservative like he did in 2010. Since then, Quinn has
consistently trailed Rauner in the polls, often by double digits (though most of those double-digit polls all came from We Ask America, the polling arm of the Republican-friendly Illinois Manufacturers' Association).
Then, two polls arrived; one was a Democratic internal poll for the DGA on Thursday, giving Quinn a 3-point lead. This was the first poll of 2014 giving Quinn a lead; that was a little surprising, but there had been a few previous internals showing Quinn down a point or two, so not totally unexpected. But then on Saturday came a second poll that was a) not an internal at all, but rather from the Chicago Tribune, a paper with a usually Republican-friendly editorial page, and b) a poll with Quinn leading outside the margin of error, 48-37.
Now, it's quite possible this poll is an outlier. The term "outlier" is a word that gets thrown around casually to mean any poll that smells a little weird, but it's actually pretty extreme. If all the other polls in a race were coming up 50-50 but you did a poll with a 5 percent margin of error that was 54-46, that's not so much an outlier as just pushing the limits on the margin of error; it could still be technically valid even if the race was truly 50-50, because it's within the MoE.
Almost all polls are constructed using a 95 percent confidence interval, which means that you can expect that 95 times out of 100, the candidates' true numbers fall within spread set by the margin of error. One out of every 20 times, though, even if you do everything completely right, sampling error dictates that you're going to produce numbers that fall outside the margin of error (before even getting to the possibility that you might have completely mis-modeled the electorate). So we may have gotten one of those 1-in-20 occurrences, or more likely, it may simply be a bungled sample, considering that the bulk of all polling hasn't shown a Quinn lead, let alone a double-digit lead. We'll never know if it's an "outlier" or not, though, and the proper response is not to discard it but to just throw it on the pile with everything else, which is exactly what we've done with it. Nevertheless, that moved Quinn's odds of victory from only 8 percent last week to 63 percent today.
On the other hand, it seems possible that Quinn truly is closing the gap. The two polls listed here are the only polls that were in the field after the news broke on Sept. 2 that Rauner belonged to a wine club that costs $100,000 to join. While you wouldn't think that, in itself, would be a game changer, it got a huge amount of attention in the local press, and does seem give validation to Quinn's line of attack ads presenting Rauner, Mitt Romney-style, as an out-of-touch ultra-rich-guy out for himself. That was compounded by the revelation on Sept. 4 that Rauner had to fess up that he once supported reducing or eliminating the minimum wage. So it seems possible that that one-two punch could have suddenly crystalized things for voters, just as they started paying real attention to the election, post-Labor Day.
The other gubernatorial race that got shaken up was in Hawaii, where until now the small handful of polls polling a David Ige vs. Duke Aiona vs. Mufi Hannemann matchup gave the Republican, Aiona, a significant lead. This race is one, unlike Illinois (where Quinn looked very dead), where we were confident that Ige would soon pull into the lead, once he consolidated Democratic support from the primary, where he knocked off unpopular incumbent Neil Abercrombie. And that's exactly what happened in the last half a week, with polls from YouGov and Rasmussen Reports giving Ige small leads of 37-35 and 40-39, respectively. That's not a slam dunk (and it probably never will be, so long as ex-Dem Hannemann is in the race), but it's enough to boost Ige's chances from 16 percent last week to 54 percent now.
The Dems' overall odds of gaining gubernatorial seats, though, is unchanged from last week at 58 percent, despite the good news in Illinois and Hawaii; similarly, the median number of gubernatorial seats holds steady at 22 (for a one-seat gain). That's because that same round of YouGov polls gave Republicans the lead in three races where there's been a string of Dem-friendly polls: Kansas, Maine, and Wisconsin. Those results are counteracted by the bulk of all previous polling, so in each of those three cases, the Democrats' odds in each race fell from the 80 percent range to only the 60 percent range.
Interestingly, though, the declining odds in those races, combined with improving odds in some sketchy-looking retentions, are reshaping the histogram. Notice above how the gubernatorial histogram is now a very tidy-looking bell curve, compared with where we were
last month, with a curve with a very long tail on the right. That long tail was caused by a map with a few slam dunks and a lot of races with long odds, but things are looking more balanced now.
Finally, let's turn to the Senate map, where there's good news too. The Democrats' odds of holding the Senate ticked back over the 50 percent mark, from 47 percent last Thursday to 54 percent now. Similarly, the median number of Democratic-held seats in the Senate ticked up from 49 to 50, which makes all the difference in terms of control.
While the natural inclination among Democratic partisans would be to say "Woo hoo! We're back in the lead!" and start celebrating, I have to be my usual buzzkill self here and reiterate that the difference between a 47 percent and 54 percent probability is utterly unimportant; it's still a complete coin flip. As Josh Katz of the NYT's Upshot pointed out last week, you would need to flip that coin nearly 400 times before you could figure out the pattern that the coin is slightly weighted one way or the other.
For once, Kansas has nothing to do with the fluctuations in the Senate model. Instead, the big movers are North Carolina and Alaska, which are both critical holds for the Dems to maintain control. Kay Hagan's odds in North Carolina moved up from 55 percent to 74 percent, thanks to leads in new Rasmussen Reports and SurveyUSA polls (45-39 and 46-43, respectively). In the Last Frontier, Mark Begich's odds ticked back above water, from 40 percent to 51 percent. That's entirely thanks to a Democratic poll (on behalf of Senate Majority PAC, from Harstad Research) that gave him a 45-40 lead over Dan Sullivan, after having previously gotten drubbed by a YouGov poll last week that had him down 6.
With both Alaska and Louisiana presenting almost exactly 50-50 odds, it's looking clear that not only will we not know who controls the Senate on Election Night (while Alaska counts its votes after everyone else has gone to bed), but probably not for weeks beyond that (while we wait for the inevitable Louisiana runoff election). So stuck up now on alcohol and/or antacids for a long, bumpy ride.
