How Much Does the Republican Nominee Matter?

Especially in the aftermath of tonight’s debate, we are going to hear a lot about whether Romney was successful in presenting Perry as unelectable (see here, here, and here for example). We’ve heard the same about Bachman, Paul, and Palin. So with that in mind, I want to throw out a provocative question: how much does it actually matter whom the Republicans nominate?

Here’s the spirit of the question. In alternate reality where unemployment falls to 6% by the 2010 mid-terms, the Democrats hold the house and expand on their majority in the Senate, and Obama has passed a few more signature pieces of legislation and then kills Osama Bin Laden to boot, we’re not talking about the Republican nominee making much difference. Similarly, if the Euro collapses tomorrow, financial panic spreads throughout the world and unemployment hits 20% by this summer, Obama is probably not getting re-elected. So at some places along the spectrum, the identity of the Republican nominee is fairly unimporant.

That being said, it is of course folly to argue that the indentity of the nominee doesn’t matter at all. We can make up similar counter-factuals (e.g. if the Republicans nominate a convicted axe-murderer) to prove the point that there is a certain type of candidate that can never get elected. But let’s limit our universe of potential candidates to those who participated in the debate tonight. How much of a boost would the most “electable” candidate provide in a general election as opposed to the least “electable”?

I am truly asking this as an open-ended question, and very much look forward to the comments. To start the discussion off, I will throw out the number that was given to me when I asked this same question this summer of a distinguished scholar of American politics: 2%. Does that strike readers as too low? too high? About right? To put this in perspective, if the “best” Republican candidate could win 52.5% of the vote, then it means any of them could probably win a majority of the popular vote.

Anyone have any research that speaks directly to this topic?

I look forward to your comments!

13 Responses to How Much Does the Republican Nominee Matter?

  1. Agnes Nov September 8, 2011 at 2:58 am #

    It seems to be important to what extent the candidate is able to adjust (in a credible way) after the nomination. Some candidates would not be able to move sufficiently close to the center (again credibly), and they will loose like Goldwater. So 2% may be too low.

  2. elgatgordo September 8, 2011 at 6:02 am #

    This piece is written on a high school level. Do better.

    • John Sides September 8, 2011 at 11:52 am #

      This comment is written on a kindergarten level. Do better.

  3. Matt Glassman September 8, 2011 at 8:22 am #


    My sense is that it still matter quite a bit, at least in one sense. Perhaps a better way to frame your questions is, “At what level of popularity would Obama have to be at in order for any minimally competent Republican be a lock to beat him?” This points us to two issues:

    1) First, what is a “minimally competent Republican.” I tend to think about this as, “the most radical candidate that the median quintile of voters, while maybe not agreeing with them much on policy, would still consider a serious person and would not lose sleep at night over if the candidate were running the executive branch.” It seems to me that not everyone on the stage last night passes that test yet.

    2) I think if the economy does not improve from it’s current point, then any minimally competent Republican will win next November.

    Therefore, my sense is that so long as the GOP nominates someone of minimum competence, the candidate does not matter. Romney, Huntsman, Gingrich — doesn’t matter. Cain, Paul, or Bachmann — I think they probably do not pass the minimum competency test as of now, and probably could not get there during the general. Perry or Santorum — I think their main hurdle would be passing that test in the general election campaign, which I think they could both plausibly do.


  4. James H Nunn September 8, 2011 at 9:37 am #

    Romney seems like the best candidate for the general elction but he will have a hard time in the primaries. The more conservative and religious fundamentalist candidates will only have a chance if the economy really tanks and Obama fails in conving the electorate that the Republicans prevented him from introducing needed economic stimulus. Republican insistance on no tax raises for the wealthy and immediate expense cuts, except in Defense, will be self-defeating except in the worse of times.

  5. Andrew Gelman September 8, 2011 at 10:08 am #


    1. You write, “In alternate reality where unemployment falls to 6% by the 2010 mid-terms, the Democrats hold the house and expand on their majority in the Senate…” I don’t know if that’s correct. My impression from the research of Erikson, Wlezien, and Bafumi is that, for midterm elections, balancing is more important than the economy.

    2. For an answer to your last question, see Rosenstone’s book on Forecasting Presidential Elections. I think he basically got it right in 1983.

  6. Scott Monje September 8, 2011 at 10:35 am #


    In your No. 1, I think those are three conditions, not one condition and two results. If all those things had happened in 2010, then Obama would win in 2012.

    • Andrew Gelman September 8, 2011 at 10:55 am #

      Ahhh, the ambiguities of written language! But in that case the relevant condition is the economy in 2012, not 2010. A recovery in 2010 followed by a crash in 2012 would have been bad for Obama’s prospects for re-election, or at least so goes the general belief. In fact, one of my theories for why Obama was not more aggressive about the economy in 2009 is that I suspect he or his economic advisors wanted to avoid a premature recovery followed by a slump in years 3 and 4.

      • Joshua Tucker September 8, 2011 at 12:23 pm #

        Scott was correct – I was just throwing that out there as three characteristics of an alternate “pro-Obama re-election world” – could have had others, there was nothing special about those. Point was simply that we can create scenarios where an incumbent landslide or an opposition landslide is very likely, and in those scenarios it would be easy to say that the quality of the opposition candidate is unimportant. My question is how much is it worth in what we can call an “uncertain” election environment, whereby we could imagine the president being re-elected or defeated.

        To maybe try this a different way: we believe that the state of the economy and presidential popularity play a huge role in determining whether the incumbent is re-elected in a US presidential election, and both of these factors are (mostly) exogenous to the identity of the opposition candidate. Now we’re about to spend a ton of time focusing on who gets to be the opposition candidate (primaries), and we are going to hear a lot about the “electability” of different candidates. From a personal standpoint, I know this is something I think a lot about when I vote in primaries, and we can make a very rational argument for doing so (e.g., the difference between any of the different players on my team is trivial between the difference between my team and the other team.). My question was, how much is this really worth in US presidential elections??? Is “electability” (within reason, e.g., the axe-murderer) actually worth the time and effort we are going to spend discussing it as a predictor of the ultimate outcome of the election. Probably it is, but I was interested in (a) attempts to quantify how much this is worth and (b) if anyone had ever/could ever come up with a good way to measure it. I was taken aback this summer when the person I querried on this said 2% – seemed instinctively low to me, but what do I know? I study Russian politics, where the blessing of Vladimir Putin gets you 50% off the bat. So that’s the question I was trying to ask.

        • Andrew Gelman September 8, 2011 at 12:52 pm #


          I think 2% is reasonable, maybe on the high end. I’ve written more about this somewhere, but the basic idea is that a U.S. presidential general election is somewhat of an extreme case in which the fundamentals matter: 2 candidates (so no strategic voting), strong parties (unlike Russia), parties that with clearly distinct economic policies (unlike in much of Europe), a long campaign time, rough balance in financial resources between the parties, probably a few other factors that I’m forgetting.

          I think that a party is better off selecting the politically moderate candidate they think will govern best, rather than worrying about personalities.

  7. Michael September 8, 2011 at 12:11 pm #


    In fact I think it’s even less than you assume. If Perry gets the nod, it will fire up the GOP base to vote in the general election but it will also drive the left to fight to keep the lunatic out. If Romney somehow gets the nomination, the Republican base would be less enthusiastic but some on the left would be less concerned with fighting him since he is viewed (wrongly) as moderate.

    Mind you, I also assumed that there was no way Americans would elect a Republican majority in congress just two years after the Bush implosion so…

  8. Justin September 8, 2011 at 12:26 pm #

    Well I don’t know much about political theory, but just looking at the effect of Rick Perry entering the race had on the blogosphere. Sure, it might not be representative of the population, but a polarizing candidate will definitely be harder to predict and I think there is a higher chance for variation. The group probably affected the most is those who are not enthusiastic about politics and voting or are on the fence about it. In the case of Rick Perry, this group will probably be larger on the left than the right, because I assume the sort of people who would get “fired up” positively about Perry are more likely to be active, i.e. Tea Partiers, but Perry’s broad and radical antigovernment statements will hit harder with more of the left, including those who typically don’t show up at elections.

  9. Sanchit Kumar September 8, 2011 at 2:56 pm #

    I’m basing this answer a little on Jonathan Bernstein’s remarks on his post on Plain Blog, “Does your Vote ‘Count'”. I think that the identity of the Republican nominee loses importance if swing voters are increasingly divided or uncertain. If every voter was partisan, then the identity of the candidate wouldn’t matter one bit. As uncertainty increases – the bloc of swing voters becomes increasingly large, and their preferences muddled – then identity and comparisons will increase (within parties, ‘electability’ becomes more important, and at the general election, ‘appeal’).

    I don’t have polling on hand, but it seems like the swing vote is sizable, more so than it was in 2008 (which reminds me, turnout will probably also be important – if Dems don’t show up at the polls, identity won’t matter). In which case, the Republican candidate’s identity in this election at least will be important.