Before Romney's victory in the first presidential debate, Obama led by 4 points with 49 percent of the popular vote. On Sunday, Romney possessed about 47.5 percent of the popular vote, his highest share of the cycle, while Obama also lost three points, falling near 46 percent. This is unabashedly good news for the Romney campaign, which hadn't led in the national polls for the entire general election campaign. A few weeks ago, I wrote that a Romney comeback would be unprecedented, even though the race was within five points. I hope election watchers properly appreciate Romney's achievement as a first in the history of presidential politics. His bounce is far more impressive than Kerry's four years ago, which basically just brought Democratic-leaners back into Kerry's fold.
But although Romney's bounce is very impressive, especially in historical context, it’s important to consider the limits of his improved position. Even after soundly defeating the president, Romney is still only at 47.5 percent of the vote. To be sure, candidates usually finish above their share of the vote in the polls, so one would expect Romney to exceed 47.5 percent on Election Day. But Obama’s best moments have yielded a larger share of the vote than Romney’s, at least to date. Indeed, Obama held between 49 and 49.5 percent of the popular vote for nearly one month following the DNC.
Whether Obama remains a favorite depends, at least in part, on whether Obama’s former supporters are likely to return to his column, assuming that Obama’s losses mainly reflect a genuine shift in voter preference and not a decline in Democratic response rates or enthusiasm. Put differently: would a candidate rather lead by a one point or have previously demonstrated the ability to assemble 49 percent of the vote? There isn't an abstract answer to this question and, in the absence of more data, there isn't necessarily a specific answer for this election. After the DNC, a wave of national polls revealed that Romney’s favorability ratings remained low and the president’s approval rating spiked along with a surge in economic confidence. As these numbers held steady for a few weeks, it seemed likely that the post-DNC shift was durable; it was resilient and underpinned by indicators of the so-called fundamentals.
Unfortunately, there haven’t been many national surveys since the first presidential debate. The polls tentatively suggest that Romney’s favorability rating has improved and perhaps significantly, but there aren’t enough surveys to gauge just how much Romney improved or whether it was primarily due to Republicans (some polls hint at this), while the president’s approval ratings and perceptions of the direction of the country appear unchanged. Without more data to judge changes in the so-called “fundamentals,” interpreting Romney’s bounce and Obama’s decline is more likely to turn on hunches and inference than hard evidence.
One example is Sean Trende’s piece “Can Obama Resist the Forces of Gravity,” which scrutinizes the movement of the RCP average and concludes that favorable events (like the 47 percent video) allow Obama to overcome fundamentals (gravity) that would otherwise yield an extremely tight race or a slight Romney lead. From this perspective, Obama’s post-DNC bump was an ephemeral moment of bandwagoning and Obama’s post-debate fall represents a reversion to the mean. To seize the lead again, Obama needs a well-timed event to defy gravity on Election Day.
It is easy to envision how Obama’s pre-debate lead defied gravity, as Trende suggests. The two percent of voters who joined Obama’s side after the DNC were not necessarily strong Obama supporters. After all, they were undecided for most of the summer, suggesting that they harbored reservations about the president’s performance. And if an unsustainably poor initial impression of Romney informed the decision of these voters to move toward the president, then Romney’s strong debate performance might genuinely cause these voters to reconsider. If this explains movement in the polls, as it very well might, then Obama could use another “47 percent” video between now and Election Day.
But peaks and bounces aren't necessarily anomalies in the gravitational field of electoral politics. For the most part, candidate's finish above their share of the vote, which suggests that "bounces" can reflect occasional moments of clarity when voters with weak preferences fleetingly realize the candidate that they'll probably support in November. The DNC was a three day informercial when the Obama campaign was able to make its best case to voters and, unsurprisingly, they attained the support of voters receptive to that message. Since these voters have proven receptive to Obama's message, their willingness to support Obama last month might presage an instinct to support him in November, as well. Is it possible that these voters only reluctantly moved to Obama after becoming prematurely convinced that Romney was a prohibitively bad candidate? Absolutely. But it might also be the case that these voters like Obama and decided that he was doing well enough to merit reelection (after the DNC, they approved of Obama's performance). While they are calling themselves undecided after the debate, they might remain likely to return to Obama's side. After all, they have not yet indicated a willingness to support the Republican nominee.
Similarly, the DNC and the 47 percent remarks moved many voters from Romney to undecided, but their previous support for the Republican nominee indicated that they were likelier to return to his column than seriously consider the president. Almost all of them disapproved of the president's performance and Obama never reached 53 percent of the vote, indicating that they never supported the president's reelection. As a result, I frequently referred to these as "latent Romney voters" who were likely to return to his side by election day. With a flood of ex-Obama supporters now in the undecided column, there might be plenty of "latent Obama voters" who are less sure of their support in the aftermath of a crushing defeat for a candidate that they're still likely to eventually support.
The demographic composition of Obama's bounce and subsequent decline provides additional evidence to support this possibility. Obama did not make big gains among Republicans, conservatives, or other groups where Obama's gains might seem fleeting. Instead, many of the groups where Obama made his largest gains came from groups where Democrats have a strong and resonant message, like Hispanics and low-income voters. Of course, there are plenty of marginal swing voters in Democratic-leaning groups; after all, Bush did quite well with Hispanics eight years ago. But winning Latino swing voters will require Romney to overcome a Democratic message with strong appeal. There's a reason they moved to the president in the first place.
One related type of voter has been written about extensively: the disappointed Obama '08 voter that likes the president personally and thinks he was dealt a tough hand. The articles written about these voters contend that it is difficult to get them to break their emotional attachment to the president and suggest that they are not convinced that he is a failure. It is easy to envision how many of these voters might have flocked to Obama's side after the DNC as they gained confidence in his performance and the direction of the economy, but reverted to "undecided" following another disappointment when high expectations confronted the reality of a feckless performance. Are these voters an easy sell for Romney or are they likely to return to the president? Given their willingness to support Obama after the DNC, it is easy to see how they could return to his column on Election Day, especially if Obama can restore their confidence in later debates.
The 2004 election also suggests that a collapse in a president's support after a debate doesn't mean that the incumbent is unlikely to win those voters back. Bush held 49.5 percent of the vote heading into the first presidential debate and fell as low as 47.3 percent after his defeat. By election day, he returned to 48.9 percent of the vote, held 49 percent a few times over the final few weeks, and won 50.7 percent on Election Day. If Obama returns to something short of his pre-debate average, he'd hold 48.5 to 49 percent of the vote on Election Day. And although Kerry continued to pick up support after the first debate, rising from 43 before to 46 after the debate and even 47.4 by Election Day, Kerry was largely soaking up his own latent voters--48 percent supported Kerry after the debate and 47 percent of voters disapproved of Bush's performance. Put differently, there were plenty of latent or likely Kerry supporters for the Democrats to soak up over the final month of the campaign.
With both parties all but assured of 47 percent of the vote, every voter on the route to 50 percent gets more and more difficult. With Romney finally matching Obama's disapproval rating and holding his highest share of the vote to date, additional gains won't come easily. Obama's past flirtations with 49 percent of the vote suggest that he might have plenty of latent supporters on the route to 49 or 50 percent. If Obama's losses are the result of marginal voters doubting the president after a mediocre performance and accompanying media coverage, then Obama will probably win many of these voters back over the next few weeks. But the decision of these voters to embrace the president may have been informed by a more negative impression of the Republican nominee than voters will receive over the final month, between strong Romney debate performances and more balanced ad spending. If Obama's losses are more a product of Romney disrupting negative preconceptions than voters doubting the president after a weak performance, then Romney might be able to fight with Obama voter for voter through Election Day.