Electoral Vote Prediction: Romney 338, Obama 200?

Sorry to throw another horserace/poll post at you, but this one is special.

ConArt critic has been making the case that blue states are in play; I personally have been questioning the conventional ill-wisdom that "it's going to be close." It's often not close at all, and yet every election prediction is "it's going to be close."

Well, if that's the eternal prediction, and yet never takes into account that sometimes it's not close at all, then that's not really a prediction. That's just something people say, because it sounds safe.

Two posts about this, the possibility of the race not being particularly close.

First, Henrik Temp wonders why every analysis begins with the 2008 election as the baseline. The 2010 election was more recent, and conditions have not significantly changed since 2010.

So what happens if you begin your analysis with the expectation that 2012 will be more like 2010 than 2008? That's the 338 - 200 electoral vote prediction.

Personally I've been splitting the difference myself, assuming that 2012 will be about halfway between 2010 and 2008.

But Michael Barone does some digging into his own almanac and discovers something interesting.

t seems to be a standard rule in assessing the prospects of Barack Obama and Mitt Romney in particular states to use the November 2008 numbers as a benchmark. However, as I have pointed out, in the last three presidential elections, the winning candidate has won a percentage of the popular vote identical to or within 1% of the percentage of the popular vote for the House of Representatives in the election held two years before. In this case, the November 2010 results are very different from 2008. In 2008 Obama won 53% of the popular vote. In 2010 House Democrats won 45% of the popular vote.

Thus, while most seem to discount the midterms as predictive -- and most seem to almost completely ignore 2010 as if it's an obvious aberration -- in fact, at least over the past three cycles, the midterms have been highly predictive. They've predicted the presidential vote to within 1%.

Given that circumstances from 2010 have not improved -- indeed, it seems likely Obama's position has deteriorated -- why would we expect 2012 to break this pattern?

Because Obama's on the ballot, officially, now? But he urged his supporters to treat voting for Democrats in 2010 as a proxy for voting for him. He nationalized the election, or tried to. (Then again, individual Democrats had a different idea, and tried to localize it.)

Barone charts out Obama's current support levels in the swing states versus the Democratic vote share in 2010.

The first thing to note is that Obama’s current percentage is closer to the 2010 Democratic percentage than to Obama’s 2008 percentage in every state but three. The exceptions are Nevada and Arizona, where the current Obama percentage is right in the middle of the two, and Florida, where the Democratic percentage in 2010 was very low because Democrats failed to contest three of the then 25 districts and because the Republican districting plan then in effect left few target seats for Democrats to seriously contest.

Second, it’s worth noting that in only four states is Obama at 50% or 51%. It should be added that he leads Romney by double digits in New Jersey, New Mexico and Minnesota; for the moment, at least, those look pretty safe for Obama.

Barone finds that if the indicator of 2010 does predict 2012 within 1%, the vote count will be similar to that predicted by Temp.

posted by Ace at 03:54 PM

| Access Comments