There’s been a decent amount of head-scratching since the results rolled in Tuesday night about why Obama didn’t do better in the Philly suburbs, where he was expected to rack up healthy margins. The thinking was that these are relatively affluent, educated, politically-moderate areas—the kinds of places Obama has carried in pretty much every state so far. If, suddenly, he’s started slipping among these voters, it raises real concerns about his viability. NBC’s “First Read” called this development “the first evidence in weeks that Clinton has finally cut into Obama's coalition.” Looking at the same results, my colleague John Judis wondered if “the electoral premise of Obama's campaign--that he can attract middle class Republicans and Independents--is being undermined.”
So, is it? What exactly happened in suburban Philly? And what does it say about Obama’s coalition?
I can’t really answer any of these questions definitively; really getting to the bottom of them would require the sort of granular Pennsylvania political knowledge I don’t come close to having. (Pennsylvania residents should feel free to weigh in.) But, after looking over a decent amount of publicly-available data--organized neatly in the (graphically crude) chart below--I’d offer the following speculations:
Judis writes that “Obama won my own Montgomery County, Maryland by 55 to 43 percent but he lost suburban Philadelphia's very similar Montgomery County by 51 to 49 percent to Clinton.” Actually, the two Montgomery counties aren’t really so similar. The Judis ancestral home sits in a county that’s about 17 percent black, and where, as of 2000, 55 percent of people 25 and over had bachelor’s degrees. The Montgomery County Obama lost Tuesday night is only 8.4 percent black, and only 39 percent of people 25 and up had bachelor’s degrees there as of 2000. That’s a big difference, and probably accounts for much of the 14-point swing against Obama. (In fairness, Judis’s Montgomery County does have a lot more Hispanics, a group Obama has struggled with in other states. But he only lost them 55-45 in Maryland, so they probably weren’t much of a drag there.)
The same goes for most of the other counties in suburban Philly. Of the eight suburban counties that mayor-turned-governor Ed Rendell won in his 2002 statewide primary against current Senator Bob Casey, there’s only one (Chester) where more than 40 percent of 25-and-overs had a bachelor’s degree, and Obama won it by a respectable 55-45 margin. Likewise, in only one of these eight counties (Delaware) do African Americans make up more than 15 percent of the population, and Obama won it by the same 10-point margin.
Conversely, college grads account for less than 30 percent of the 25-and-up population in five of the counties (basically half the proportion of college grads as Judis’s Montgomery County), and African Americans accounted for 6.5 percent or less of the population in six of the eight counties. Lehigh and Northampton, two counties where Hillary throttled Obama, are heavily white working-class areas--only about 5 percent black and under 25 percent college-educated. Simply put, this wasn’t really Obama country.
2.) The second key factor here was probably Rendell himself. One way to see this is to look at the difference between two of the suburban counties, Lancaster and Berks, both of which Rendell carried comfortably in the 2002 primary against Casey, but only one of which (Lancaster) Obama carried this week. Moreover, Obama didn’t just lose Berks, he got pounded there.
What’s interesting about the two counties is that their raw demographic data are pretty similar: Only around 20 percent of adults are college-educated in each; African-Americans account for less than five percent of the population in both places. My guess is that Democrats in these places look pretty similar (though they may be slightly more affluent and liberal in Lancaster).
So what explains the difference? I’d guess it’s that Lancaster is a much, much more Republican county than Berks. In 2004, George Bush killed John Kerry in the former (66-34), but only narrowly beat him in the latter (53-47). Likewise, Lancaster was the only suburban-Philadelphia county Rendell lost to Republican Mike Fisher in the 2002 gubernatorial election. (And he lost it by a 2-1 margin). To me, that suggests Berks has a decent Democratic Party infrastructure, while Lancaster has little to speak of.
Being the popular mayor of Philadelphia, Rendell was going to win both of these counties in his 2002 primary against Casey—he wasn’t dependent on a local machine. By contrast, the presence or absence of a local party apparatus had huge implications in this race, since it that apparatus was thoroughly behind Hillary Clinton. In Berks, the place where there’s presumably a strong Democratic Party which presumably defers to Rendell, Hillary won 58-42. In Lancaster, where the party doesn’t have much of a presence, Obama won 55-45.
3.) Putting this all together, you notice a rough U-shaped pattern across the Philly metro area. Obama tends to win the counties that are either strongly Republican (like Lancaster) or strongly Democratic (like Delaware, or Philadelphia itself), while Hillary tends to do better in counties that are either narrowly Republican or narrowly Democratic—and, within that band, the more Democratic the better. Which makes sense. The narrowly Democratic counties have strong Democratic parties and are therefore places where Rendell’s help would have really mattered. But, by definition, they’re not *so* demographically favorable to Democrats—meaning they don’t have the huge numbers of college grads and African Americans that strongly favor Obama—which is why Republicans still have a chance there.
As with Pennsylvania generally, winning these places might have been a sign of the growth Obama's supporters would like to see. But that doesn't mean losing them was a step back.