Rumors of Mitt’s Death Are Greatly Exaggerated

I know. I know. “Where have you been?!”   I appreciate the email inquiries.  The short answer is I’ve been trapped in my office, fending off an onslaught of students.  Such is the life of a departmental chair at a nationally-ranked liberal arts college.   There’s been other distractions as well associated with the start of the semester (teaching a new election class!), and with giving election-year talks.   So that’s my excuse for the scarcity of posts.  I’ll try to do better now as academic-related activities begin to slow.

Meanwhile, in my blogging absence, the presidential campaign has, apparently, all but ended, with Romney suffering an ignominious defeat.  Or so the pundits tell me.  Evidently, Mitt’s political “death” was precipitated by several causes.  First there was the disastrous Republican convention, lowlighted by Clint Eastwood talking to a chair.   That was followed by the brilliantly orchestrated Democratic Convention, highlighted by the Big Dog’s mesmerizing recitation of his…er….Obama’s accomplishments.  Then Romney dug himself a deeper hole by seeming to politicize Libyan Ambassador Chris Steven’s death through some ill-timed remarks.   Romney dumped the final shovelful on his own political grave by accusing 47% of voters – many of them presumably his own supporters – of suffering from  an “entitlement ethos” that makes them overly dependent on government programs.

That last “devastating” gaffe was enough to convince several pundits  (see here and here) that  Romney had “lost the election.”   Forgive me if I’m not persuaded, and why I think you should not be either.   Not surprisingly, of course, the pundits who are certifying that the campaign is over are all Obama supporters.  More importantly, however, is that the polling data, while indicating that Obama may have gained a couple of percentage points over Romney compared to the pre-convention polls, still show this as a tight race.  The latest aggregate poll has Obama up by 3%, at 48.1%-45.1.%.  On the day before the Republican convention, Obama led by 1.4% in their composite poll.  At RealClearPolitics, Obama leads by 3.9%, 48.6%-44.7%.  He was up by 1.4% there before the conventions.  So there is evidence that the cumulative polling impact of the “devastating” period (for Romney) has cost him about 1.6%-2.5% in the polls.  That’s not insignificant, particularly in a tight race, but I don’t see this as proof of Romney’s demise either.

In assessing the claims that we have just witnessed a turning point in the campaign, I suggest keeping several factors in mind.  First, neither candidate got a huge convention boost as measured  by historical standards although my read of the polls is that the net polling advantage in convention bumps went to Obama.  Already, however, we see signs that some of that initial Democratic convention bump has dissipated.

Second , as John Sides argues, the shelf-life of presidential candidates’ rhetorical gaffes is surprisingly short.   Here’s John’s chart showing just how little previous rhetorical gaffes, such as Obama’s “You didn’t build that” remark, have actually moved the polling needle.  Romney may have gained a percentage point or two due to Obama’s statements, but it’s hard to say that permanently changed the race, particularly since Romney’s support dipped down again shortly after.

As I’ve discussed previously, these remarks tend not to have much impact largely because they are filtered through voters’ preexisting ideological beliefs. For this reason, I doubt Mitt’s 47% comment is the game changer that partisan pundits predict/hope it will be.  Remember, campaigns tend not to change votes so much as they activate latent predispositions among voters.  Yes, it’s possible this time will be different, and that Mitt’s remarks really are a turning point. But in the absence of evidence indicating why this time should be different, forgive me if I don’t take the partisan pundits’ words for it.

Already, the talking heads are debating just how bad a candidate Mitt is.  But, while he may not be the most well-liked guy, it is not clear to me that he is underperforming the economic fundamentals by all that much, if at all, based on current polls.  Much depends, of course, on which forecast model you believe.   As I’ve discussed in several previous posts, more than one forecast model has Obama winning this race by a very close margin.  Taken as a whole, as I’ll discuss in a future post, the forecast models see this race as a toss-up.  And that’s not far from where the aggregate polling has it right now. Remember, whenever a candidate appears to be losing, media pundits invariably point to failures in candidate strategy and/or in the candidates’ perceived personal shortcomings.   But that doesn’t mean that assessment is right.  And I don’t think it is right this time either.


  1. Matt,

    I agree with all of your points. Two questions:

    1. What about Obama’s lead in almost all of the swing states, save NC and Mizzou? Isn’t that really problematic for team Rommney and portend an electoral college landslide even if the popular vote remains close?

    2. What about the other fundamentals, namely the fact that Obama is now better financed and they have a far superior ground game?

  2. Matt, we forgive you, from far away, for leaving us without your commentary for oh-so-many days. It occurred to me today that you are the Dan Patrick of political bloggers, going against the conventional wisdom, and whatever is reported widely. Look forward to the next 46 days–

  3. Orion,

    I’ve posted on the swing state question before, but my short answer is not to think about the swing states as separate election contests. Instead, they are all likely to be affected to varying degrees by any changes in the national tide. Put another way, history says the winner of the national popular vote is likely to win the electoral college too. This isn’t always true, of course. So, if undecideds break by 2-3% to Romney, they are likely to do so – again with some variation – in the swing states as well. Of course, if Obama wins the popular vote (and I think he’s leading now), that could translate into a comfortable electoral college victory. Either way, I think the national popular vote is a pretty good barometer of who will win the Electoral College vote.

    As for the other fundamentals: it’s not clear to me that once you add in the Superpacs, 501(c)’s, etc., that Obama is going to be better funded. He has certainly outspent Mitt to this point, but in the end I don’t think money will be the deciding factor here. The ground game, on the other hand, could be decisive in a close election. Beyond some statistics regarding field offices in various swing states, I don’t have much independent evidence to determine who has the better GOTV organization, but it wouldn’t surprise me if the Obama campaign has the edge here due to their experience in 2008. But that is a reminder that turnout is particularly crucial in this election, and there is evidence that the “enthusiasm” gap which had favored the Republicans may have tightened a bit since the Democratic National Convention.

  4. Jeff,

    Much of the reason for why I am beating off the students with a stick is because my colleagues have fled to warmer climes, where they persist in telling me how wonderful their life is right now while I am left to deal with the hungry hordes.

  5. Dale,

    Thanks for this link – I hadn’t see this article. Again, it is difficult to evaluate the relative efficacy of the two campaigns’ ground games, but I am willing to believe that Romney’s is probably better than McCain’s was in 2008. Is it better than Obama’s? Just as good? Slightly worse? I don’t know if it is possible to make a blanket summary. I’m guessing there are variations in the relative organizational effectiveness within each state.

  6. While I agree that the election is far from over, there is a real cost to Romney in the poll decline over the past few weeks. He is losing and as any sports fan knows, when you are losing the clock is your enemy (well except in baseball of course). Every day that he is not making up ground has a cost (as Nate Silver ably points out on a daily basis). Now it may be that the polls will regress to where they were pre-convention but even if that is the case it still has Romney losing by 1-2 points. And I’m not sure that is close enough for ads or ground game to change (even if he has the advantage in those — a dubious proposition). I think he increasingly needs a boost from the debates or an October surprise.

  7. What kind of election numbers in November would make you question the “fundamentals”-based models? If Obama won 55-45? If Romney won 55-45? Even though you can find some individual models that come close to predicting both of those results, it seems like calling both those results “close elections” is pushing it.

  8. Stuart – I don’t disagree with anything you say here. Obama is, in all likelihood, ahead in this race, and it is better to be ahead than behind. The main thrust of my post was to suggest that, contrary to what some of the pundits were claiming, the race is far from over, and the 47% comment isn’t likely to have a huge impact.

  9. Gerald,

    You raise a great point, which I should probably address in a separate post. Generally, when I talk about the forecast models, I talk about them in the aggregate – that is, I tend to average the results of the ones that have been historically most accurate. But we shouldn’t forget that there is some degree of variation in the individual projections. So, how we judge the accuracy of the forecasts models depends, of course, on whether you are looking at the average, as I tend to do, or you are wedded to any single model. In either case, however, a model (or average of models) that is off by 5% is not close. Nor would I call an election in which either candidate wins by 10% close! But a forecast model can be close if the final tally comes within some specified margin of error in its projection, even if the election is not… .

  10. “You didn’t build that” isn’t a gaffe, get your facts straight. You can’t just take a part of a sentence out of a paragraph of text from a speech and call it a gaffe. He was clearly referring to the roads, bridges, and other infrastructure that individual business people did not build – that they were government services. “We built that” shows just how dumb people really are – they can’t even watch a 20 second clip of what was actually said.

  11. Jeff,

    I think a lot of Obama supporters made that argument at the time that Obama made those comments. Alas, that’s not how it was played in much of the press. Similarly, a lot (but certainly not all) of Romney supporters are arguing that Romney should not only stand by the 47% comment, he should use it to differentiate his governing philosophy from the President’s. Nonetheless, I don’t think the media is playing it that way.

    My broader point, however, is that despite the media coverage, these “gaffes” generally don’t have a significant longterm impact on candidates’ standing in the polls.

  12. What surprises me here is that there is no discussion of the technical deficiencies of many (most) polling techniques. The weighting process looks backward, not forward. By way of short example, if the pollster only gets a limited amount of people from one party or the other, the results are re-weighted to conform to the last presidential election turnout numbers to get a result that will be (in my humble opinion) quite different than what the votes in November 2016 will show. Why?

    In 2008, there was an unprecedented turnout for Obama; does anyone really think that it will happen again?

    And, it is my contention that the turnout for McCain was lukewarm; Republicans did not realize how dangerous to their thinking Obama would turn out to be. Republicans consistently vote more reliably than Democrats, but in this election, I contend that they are super motivated to see Obama go back to Chicago and start planning his library.

    On the other hand, many, many of the 2008 Obama supporters are disillusioned and while them may not vote for Romney, it is more likely they just may not vote at all.

    One more point; someone said money isn’t going to be important; I disagree. Obama has been burning at a rate twice that of Romney in Ohio, Florida and Virginia, which may somewhat explain the small lead he has in those three key states. Romney has been husbanding his money, and so has Rove and other SuperPac gurus. I believe that when the Romney forces money is unleashed, things will change. That is probably timed to coincide with the debates which will occur October 3, 11 (VP), 16 and 22. Arguably, that’s when most peoples’ interest begins to peak, and the persuadable group is most attentive.

    Romney didn’t get wealthy and successful by mistake or luck; he analyzed problems and came up with winning solutions. Why would anyone think he dumbed down now?

    Get your bets down now while the odds are best. I think Romney/Ryan will will win decisively on November 6, ala Ronald Reagan v. Jimmy Carter.

    Oh, I left out the similarity of the Mideast on fire to the hostage crisis of 1980; the weaknesses of Carter and Obama are just now being drawn. Again, this does not bode well for the president and his foreign policy. But I will leave that discussion for another post by Professor Dickinson.

  13. “Taken as a whole, as I’ll discuss in a future post, the forecast models see this race as a toss-up”

    Good stuff.

    The final hold-out was Gallup which has Obama +6.

    At what point do you admit Obama is ahead?

  14. Adam – Keep in mind that the forecast models – at least the ones that I discuss in terms of modeling the “fundamentals” – do not change once the forecast is issued. As we get closer to the election, the poll-based forecasts will be increasingly accurate, so that by the eve of the election they will be far more reliable than the political science fundamentals-driven forecast models. But, of course, poll-based predictions are inherently uninteresting (at least to me) because they don’t say why one candidate won. That’s why political scientists spend time constructing these fundamental-based models in the first place – to see whether they truly do understand why people vote the way they do. If all you want is the bottom line, go visit Sam Wang’s site – he did pretty well in predicting the outcome using state-level polls in 2008. All a long way of saying that from an explanatory perspective, I’m not very interested in who the polls say is “ahead” right now – I’m only interested in seeing how close the final results are to what the forecast models predicted on Labor Day. Let the pundits and journalists debate who is “ahead” now…. If we have it right, in the end the various polls should converge to what the models predict. If they don’t – we have it wrong, and it will be interesting to see why!

  15. I am trying to make sense of these latest battleground polls, but simply can’t seem to find any rhyme or reason to them… why do these pollsters seem to think that we will see record Democrat turnouts over Republicans this November? Haven’t we seen, instead, that Republicans are generally more enthusiastic this time around? And more to the point, how is it that Romney is winning independents in these polls (though barely), and is still down by 8-12 pts in Florida, Ohio and the like? It seems to me that the sampling in these polls have been misleading, to say the least. Am I wrong?

Leave a Reply

Your email address will not be published. Required fields are marked *