Slate.com's Timothy Noah has a good analysis of the white vote in presidential elections. Exit polls suggest that Barack Obama lost this group by 12 points, and that no Democrat since Lyndon Johnson has won in this category:
... when the Democrats go nearly a half-century without winning a majority of white votes in any presidential election, it's necessary to ask why, even after we've passed the remarkable milestone of electing our first black president.
As Noah points out, there all kinds of factors that influence a person's vote, so it's simplistic to say that the Democratic Party has a "white problem." (Just as it's simplistic to say that Obama "won back" the Catholic vote. Yes, exit polls on CNN.com show that Obama won 54 percent of the Catholic vote after Kerry narrowly lost it in 2004, but they also show that McCain won 52 percent of the "white Catholic vote." So the big shift to Obama may have come from Hispanic voters, not necessarily from Irish- and Italian-American parishioners.)
Noah says that it's also not quite right to say that "white resistance to voting Democratic [is] an entirely Southern phenomenon" that skews the national numbers. He points to a chart showing that Obama (narrowly) lost the white vote in several Northern states that he carried, including New Jersey, Ohio, and Pennsylvania.
But one thing that strikes me about the chart is that the most urban jurisdiction (the District of Columbia) is also where Obama got the highest share of the white vote (86 percent!). So is "white resistance" to the Obama and the Democratic Party a rural or suburban phenomenon, if not a Southern phenomenon? White residents in major cities tend not to own guns or hold conservative religious views, so they may be immune to the GOP's appeal on wedge issues such as gay marriage. And they probably don't respond well to rants against spending on "midnight basketball" and on that planetarium on Chicago's lakefront that McCain seemed so upset about.
Conversely, white residents in outlying suburbs may be deeply suspicious of spending programs that seem to benefit major cities (especially if opponents characterized them as "redistributionist"). Racism may well be a reason for this feeling, since most major cities have large black and Hispanic populations, so it's difficult to tell whether there's a hostility to the Democrats as the "big city party" as distinct from the "black party."
Unfortunately, "white urban" and "white suburban" voters are not among the dozens of categories listed among CNN's exit poll results. But the list below, of the 20 most densely populated counties in the US, suggests that Obama did well among non-Hispanic whites in major cities. Outside of Staten Island, he got at least 70 percent of the total vote regardless of the racial makeup of the county. In Manhattan, San Francisco, Boston, Philadelphia, and other cities, the collective vote for McCain and third-party candidates was less than half of the non-Hispanic white population.
MOST DENSELY POPULATED COUNTIES: ETHNIC BREAKDOWN AND 2008 VOTE
1. New York, NY (Manhattan): 48% non-Hispanic white; 14% non-Obama
2. Kings, NY (Brooklyn): 36% non-Hispanic white; 21% non-Obama
3. Bronx, NY: 13% non-Hispanic white; 12% non-Obama
4. Queens, NY: 31% non-Hispanic white; 25% non-Obama
5. San Francisco, CA: 45% non-Hispanic white; 16% non-Obama
6. Hudson, NJ: 34% non-Hispanic white; 27% non-Obama
7. Suffolk, MA: 52% non-Hispanic white; 23% non-Obama
8. Philadelphia, PA: 39% non-Hispanic white; 17% non-Obama
9. Washington, DC: 32% non-Hispanic white; 7% non-Obama
10. Alexandria City, VA: 59% non-Hispanic white; 28% non-Obama
11. Baltimore City, MD: 30% non-Hispanic white; 13% non-Obama
12. Richmond, NY (Staten Island): 67% non-Hispanic white; 53% non-Obama
13. Arlington, VA: 65% non-Hispanic white; 29% non-Obama
14. Essex, NJ: 36% non-Hispanic white; 24% non-Obama
15. Cook, IL: 45% non-Hispanic white; 24% non-Obama