In 2008, Colorado was the state that got Obama over 270 electoral votes. He won it by 9 points back then; today, though, Obama’s locked in an extremely tight race there. And to many who anticipated that Obama would build his reelection campaign around winning “new coalition” states where Democrats fared relatively well in 2010, Colorado’s descent down the tipping point state list has come as something of a surprise.

But there's a fairly simple reason for Colorado's tight race: unlike in many battleground states, Democrats don’t have a partisan edge in Colorado. That means Obama needs to win a majority of independent voters in the state, but Obama is struggling with independent voters nationally. To be sure, Colorado’s independent voters are relatively well-educated and socially moderate, so Obama will probably do better among independent voters in Colorado than he does nationally. But that still makes Colorado a state where Obama's chances depend more on persuasion than turnout.  

The early voting numbers in Colorado confirm that the race hinges on independent voters. Unlike Nevada or Iowa, where Democrats have already banked a large lead, registered Republicans are performing as well as their registration numbers in Colorado. Approximately 25 percent of the ’08 electorate has already cast ballots, and registered Republicans outnumber registered Democrats by 3 points, 39 to 36. That's an improvement compared to 2008, when registered Democrats outnumbered Republicans by 2 points in Colorado’s early voting.   

Obama’s ability to prevail depends on a better showing with moderate, college-educated women in the Denver suburbs than the polls currently suggest. A massive gender gap allowed Michael Bennet to overcome the GOP’s national wave in 2010, and he won women 56-39. In 2012, there were reasons to think that Obama’s effort to attack Romney on social issues would help generate big support among the well-educated and socially moderate women of Jefferson and Arapahoe Counties. But the polls don’t show Obama generating that type of margin, with NBC/Marist, PPP, SurveyUSA, and Quinnipiac each showing Obama up by 6 to 10 points and between 49 and 54 percent of the vote. That’s about the same as Obama’s lead among women nationally, but that probably won’t do it in Colorado, a state where there aren’t as many non-white voters to compensate for Obama’s deficit among white men.

Support thought-provoking, quality journalism. Join The New Republic for $3.99/month.

Perhaps Obama will outperform the polls as Bennet did two years ago. After all, the pre-election surveys didn’t show Bennet generating the advantage among women that ultimately powered him victory. But counting on systemic polling errors without an underlying methodological explanation isn’t especially wise. The argument that the polls missed Latino voters in Nevada isn’t as applicable to Colorado, since Colorado’s Latino population is far smaller and it’s not as heavily Democratic. In fact, Bennet outperformed the polls by about the same margin as other Democrats locked in tight races, suggesting that his surprise win wasn't due to Colorado-specific polling issues.

Four years ago, Obama did better much better in Colorado than he did in either Virginia or Ohio. Today, Obama could conceivably fare worse in Colorado than both states. Losses among independent voters have hit the president hard in a state with a larger Republican base and without rapid demographic changes that help Obama make up ground. The fate of Colorado seems to hinge on the socially moderate women of the Denver suburbs—and, so far, they aren’t supporting Obama to the extent that they supported Bennet.