What are the odds that President Obama wins reelection? Generally incumbents running for their party's second presidential term fare pretty well, as Alan Abramowitz notes. If the economy continues to recover, Obama would be a prohibitive favorite.
But Sean Trende at the right-leaning Real Clear Politics paints a darker picture of Obama's reelection chances. Trende's basic thesis, culled from a very few data points, is that Obama is a kind of post-partisan figure whose popularity runs ahead of his electoral success. It could be true, I suppose. The trouble is that Trende's analysis, as often seems to be the case at RCP, tilts the board clearly, if not implausibly, in the GOP's favor. The underlying assumption of his analysis is that the 2008 election was an outlier, and the 2010 midterm a return to normal -- when of course a first midterm election during a deep economic crisis following two straight wave elections is a recipe for massive out-party gains.
So, for instance, Trende argues that some of Obama's 2008 states are practically out of reach:
But the president has endured many, many bumps since winning his 365 electoral votes. This has weakened him significantly in many states. For example, he won North Carolina and Indiana by a combined 40,000 votes, when the Republican Party was in the midst of a perfect storm. It seems highly unlikely that he would win these states today, especially after the 2010 election results in these states.
We don't have any polling in Indiana. We do, however, have some North Carolina polling. A poll last November showed Obama running even with Republicans in the state (ahead of Sarah Palin, edging Newt Gingrich, tied with Mitt Romney and trailing Mike Huckabee.) A poll last week had Obama running ahead of all four, by margins ranging from 3 to 9 points. There's no evidence to say that Obama is "highly unlikely" to carry the state.
Meanwhile, here's Trende arguing that the 2010 election is a reasonable proxy for 2012:
if you take the 2008 turnout levels, but model the white vote to resemble 2010, President Obama would lose, albeit narrowly. Set the Democrats' share of the white vote to halfway between 2008 and 2010, and President Obama would win narrowly.
The 2010 electorate was far whiter than the 2008 electorate, and that accounts for much of the GOP gain. Trende properly adjusts for that change. But the 2010 electorate was different in another way -- it was much older. As I wrote on the night of the 2010 elections:
The exit polls from today's election show how this happened. The non-white share of the electorate fell from 24% in 2008 to 19% in 2010. But the age gap is the real tidal shift. In 2008, Republicans won voters over 65 years old by 8 points, but were crushed among voters under 30 by more than 30 points. The under 30 vote outnumbered the over 65 vote.
In 2010, Democrats still crushed Republicans among the under 30 vote, albeit by just 20 points. But the over 65 vote went Republican by a massive 20 point margin. What's more, in today's election, senior citizens constituted more than twice as high a share of the electorate compared to voters under 30. In 2008, the young were 18% of the electorate, and the old were 16% of the electorate. In 2010, the young were 10% of the electorate, and the old were 24% of the electorate.
Midterm elections tend to be low-turnout affairs with disproportionately old voters. This fact had little partisan significance until very recently, when age suddenly became a highly predictive factor. The whites who showed up in 2010 were far older than the whites who voted in 2008. Now, Trende suggests that it will be hard for Obama to match the enthusiasm of his pathbreaking 2008 run, and he's probably right. But the 2012 electorate is still going to look a lot more like 2008 than 2010.
A lot of conservatives have come to believe that the 2010 election was some kind of return to normal, an electorate awakening from its 2008 fever. The reality is that 2010 was overwhelmingly the story of the old and white turning out while the young and nonwhite stayed home. Now, old people's votes count, too. (Some of my best friends...) But to count on a recurrance in 2012 is to set yourself up for a rude awakening.