I’ve written two posts trying to debunk this idea but the fact that pols as savvy as Clinton and Haley Barbour are still pushing it is alarming. One of the lessons of the election, I thought, was that we should trust the math nerds when they read the chicken entrails of polling. Okay, well, the math nerds say that the hurricane wasn’t decisive. Mitt may be the only person in the GOP right now who benefits from believing that it was. He’s done running for office so he doesn’t need to worry about figuring out what went wrong with the larger operation or message.
In fairness to him, though, even he doesn’t sound completely convinced of Clinton’s theory, noting that BC was perhaps being overly “effusive” in his praise during their chat. He doesn’t offer a reason why but I’ll give you two possibilities (beyond the obvious “Clinton was just trying to be nice” explanation). One: Obama said at yesterday’s presser that he’ll talk to Romney before end of the year and praised some of his career accomplishments, like running the Olympics. I’m skeptical that he’d overcome his alleged disdain for Mitt enough to offer him a position of some kind — although people have been spitballing about that online — but maybe he thinks Romney can be useful to him somehow. At the very least, he’d get some points for comity and bipartisanship from the general public by bringing him onboard. Maybe Clinton is anticipating that and this was an early step to try to heal the breach. Two: Don’t forget that Romney spoke at the annual conference held by Clinton’s Global Initiative in the heat of the campaign, just a short time after Clinton destroyed him in his speech at the Democratic convention. Could be that Clinton felt obliged to reach out now since Romney was gracious enough to do him that favor. (He’s become good friends with the Bush family despite beating Bush 41.) If nothing else, it’s smart to stay in the good graces of a potential donor as generous and well connected as Romney.
But I digress. Here’s a new problem with the hurricane theory that I haven’t mentioned before. How likely is it that the race changed in the last few days when, by and large, it hadn’t changed much for the previous three months? To put it differently, what if the campaigns themselves just didn’t really matter?
We all know that most voters decide who to vote for well before the campaigns begin. In political science research, this is called the “minimal effects” thesis. Basically the vast majority of the voters vote how we would expect them to long before the election. The first study to investigate this phenomenon focused on voters during the 1940 election. Researchers found that only 8 percent of voters changed their preference over the course of the campaign. In 70 years, not much has changed…
Well, according to the best statistical models out there—no. This isn’t to say that Obama should have sat the race out as Romney dragged himself across the country. Indeed, if the campaigns were not equally run by the top professionals in the field with endless cash on hand, one campaign would matter. But when the campaigns have equal access to financial and intellectual capital, both could call it a day after the convention. Their activities simply cancel each other out. We could have all ignored the news since August and the election results would have come out the same.
There is one caveat though: as all of this evidence shows, elections are not won by convincing people to vote for a candidate for whom they are not previously inclined. The importance of the campaigns is to get more of their supporters to the polls than the other team. To paraphrase one Obama campaign aide, presidential appearances have nothing to do with convincing voters, they are about rallying the troops and keeping supporters passionate so that they actually turn out on Election Day (and bring their friends).
Reminds me of Drew Linzer’s boast that his Votomatic statistical model was predicting the eventual electoral college totals — Obama 332, Romney 206, give or take — as early as June. Another statistical model I saw claimed that the only thing that really moved the polls much either way over the last month or two was Romney’s performance at the first debate, but of course that wasn’t a campaign event. I think it’s theoretically true that two smart, well funded campaigns will more or less neutralize each other; problem is, both sides didn’t have “equal access to financial and intellectual capital” this time, or they didn’t utilize it equally if they did. Obama had more data, was smarter about how he targeted voters, was evidently more efficient in turning them out, and was shrewder strategically in defining Romney as a kleptocratic tax-cheating layoff artist early in the campaign. (Romney was handicapped in all those things by the fact that he had to worry about a contested primary.) That was the real hurricane. It blew away the traditional “likely voter” turnout model and replaced it with something that looked more like a “registered voter” model. If you’re a professional campaign “meteorologist” for the GOP, how on earth did you not see it on the radar?