Throughout his career on the national stage, Obama has struggled among white men without a college education. But in these latest surveys, he has fallen to a level of support among them lower than any Democratic nominee has attracted in any election since 1980, according to an upcoming National Journal analysis of exit polls from presidential elections.
Though pollsters at each organization caution that the margins of error are substantial when looking at subgroups such as this, each poll shows erosion within that margin of error for Obama with these working-class white men. The new Quinnipiac poll shows Obama attracting just 29 percent of non-college white men, down from 32 percent in their most recent national survey in April, according to figures provided by Douglas Schwartz, April Radocchio and Ralph Hansen of Quinnipiac. The ABC/Washington Post survey found Obama drawing just 28 percent of non-college white men, down from 34 percent in their May survey, according to figures provided by ABC Pollster Gary Langer. Romney drew 56 percent of the non-college white men in Quinnipiac and 65 percent in the ABC/Washington Post survey.
No one expects Obama to win these blue-collar men, who are now among the most reliably Republican segments of the electorate. But even so, these numbers, if sustained through Election Day, would represent a modern nadir for Democrats. Since 1980, the worst performance for any Democratic nominee among these working-class white men was the 31 percent Walter Mondale managed against Ronald Reagan in 1984; the meager 39 percent Obama drew in 2008 was actually the party's best showing over that period. These new surveys show Obama that these non-college white men represent Obama's largest source of decline in the white electorate since 2008.
Still, Obama is also facing weak numbers among working-class white women. The Quinnipiac Poll shows him drawing just 37 percent of white women without a college education, and the ABC/Post poll puts him at 40 percent with those women. In each poll that's up five percentage points from his showing in the most recent national survey, a change within the margin of error. But even so, Obama's performance in the new polls shows the continued Democratic struggles with those "waitress moms" that Bill Clinton and then Al Gore targeted successfully (Clinton won 48 percent of their vote in 1996 and Gore 45 percent in 2000). Obama appears on track to do no better, and possibly slightly worse, than the modest 41 percent he won with those women in 2008, which was itself essentially unchanged from John Kerry's weak 40 percent showing in 2004.
Such a poor performance among working-class whites would enormously complicate Obama's hopes in older Rust Belt states where they predominate, including Ohio, Iowa, Michigan and Wisconsin (and to a slightly lesser extent Pennsylvania). But in several of those states, recent polls (like Quinnipiac's most recent Ohio and Pennsylvania surveys) show Obama running slightly better with non-college whites than he's run nationally, which has allowed him to maintain a lead.
By now, the reasons for Obama's struggles with working-class whites are familiar. Many are culturally-conservative; more are deeply skeptical of government (including his health care plan); and most are struggling in the sustained economic downturn. Polls also consistently show many working-class whites are deeply uneasy about the propulsive racial and ethnic changes that Obama uniquely embodies.
Michael Podhorzer, political director of the AFL-CIO, worries that Democrats are too resigned to further losses among working-class whites. "It's unfortunate," he says. "Karl Rove looks at the Latino numbers and says 'How do I get another couple percent' and focuses on that like a laser beam. And people tend to just be fatalistic about this." To reach these alienated voters, Podhorzer argues, the key for Obama is move the race "back to a choice from a referendum....For a lot of these white men, another part of the message is turning Romney from the 'not-Obama' candidate into a more sharply defined candidate who wants to take the country in the exact opposite direction. Obama may not have gotten far enough, but Romney wants to go back to even worse."
With college-educated whites, Obama's situation appears somewhat more stable. The two new surveys show Obama's support among college educated white men dipping slightly from its level in 2008, when he won 42 percent of them. Quinnipiac puts his showing with these men at 40 percent (essentially unchanged from 39 percent in its April poll) while the ABC/Post survey found him attracting only 37 percent, down from 44 percent in its May survey.
A source of better news for Obama in both the Quinnipiac Poll (which gave him a narrow 46 percent to 43 percent lead), and the ABC/Post survey (which showed him and Romney tied at 47 percent) is the persistence of his support among college-educated white women, consistently the most Democratic-leaning segment of the white electorate. Quinnipiac shows Obama drawing 52 percent of those women, unchanged from April; in the ABC/Post poll he attracts 49 percent, down a tick from 51 percent in May. Obama won 52 percent of those women in 2008, and Democrats have carried them in each of the past five elections except for 2004, when the split almost evenly between Kerry and George W. Bush. Obama's success at holding them, though, marks a major improvement from 2010, when exit polls showed these women joining all other whites in moving sharply toward the GOP in Congressional races.
Both surveys also show Obama maintaining strong support among minority voters (though Quinnipiac records a decline that may be a statistical blip.) In 2008, Obama won a combined four-fifths of non-white voters; in the ABC/Post survey, he leads Romney among all non-whites by 76 percent to 17 percent, essentially unchanged from May. Quinnipiac shows Obama leading among minorities by 69 percent to 21 percent, down from 77 percent in its April poll.
Demographic change is the silver lining for Obama in these shifting patterns of political support. As recently as 1992, non-college whites cast an absolute 53 majority of all votes, and minorities just 12 percent, with college whites contributing the remaining 35 percent. By 2008, minorities had increased to 26 percent and non-college whites fallen sharply to 39 percent of the votes, with college whites holding steady at 35 percent, according to the exit polls. (Democrats also have benefited from the shifting internal dynamics of the white population; in 1980, non-college white men outnumbered college-educated white women in the electorate by about three-to-one. By 2008, the two groups cast almost exactly the same share of ballots and in 2012, it's not inconceivable that the upscale women could equal or exceed the downscale men as a proportion of voters.)
Demographer William Frey and political analyst Ruy Teixeira, both affiliated with the Brookings Institution, saw the same trends continuing when they recently analyzed data from the November 2008 and May 2012 Census Current Population Surveys, to project the eligible voting population in November. Their conclusion is that since 2008 minorities have increased their share of the eligible voter population by about three percentage points, while the working-class white share has declined by a comparable amount and the college-plus whites remain essentially unchanged.
But it's far from clear that the change in the actual electorate will match the change in the eligible voter pool. Internally, the Obama campaign has estimated that the minority share of the vote, which has grown steadily since the late 1980s, will increase to 28 percent this year, which would obviously smooth his path to reelection if it occurs. But privately, some key Democratic strategists worry that the combination of a turnout surge among conservative whites, lagging enthusiasm among the (heavily minority) Millennial Generation, Republican-driven new voter identification laws, and faltering mobilization efforts among Hispanics could cause the minority vote share in 2012 to remain stuck at its 2008 level-or, incredibly, even decline. If that happens, almost needless to say, Obama's chances of a second term would surely decline as well.