Finally, six weeks after the last primary, Pennsylvania has voted. Polls taken just before the race showed Hillary Clinton ahead of Barack Obama by 49.5 percent to 43.4 percent. The actual vote was 55 percent to 45 percent with Clinton winning (official county results here).
Some interesting patterns here. As I noted in a previous blog post, Obama carries blacks, academics, and state capitals—and not much else. He carried seven of Pennsylvania's 67 counties: Philadelphia County (44 percent black in 2000, by 65 percent to 35 percent), Delaware County (suburban Philadelphia, 15 percent black in 2000 and probably higher today, by 55 percent to 45 percent), Chester County (historically Republican, affluent suburban and exurban, by 55 percent to 45 percent), Lancaster County (heavily Republican, Pennsylvania Dutch country and exurban, by 64 percent to 36 percent), Dauphin County (state capital, 18 percent black in 2000, second-highest percentage in state, by 58 percent to 42 percent), Union County (Bucknell University, by 52 percent to 48 percent), and Centre County (Penn State University, by 60 percent to 40 percent).
And, as I noted in the same column, Clinton carries Jacksonians—the descendants of those Scots-Irish, Scots Lowlanders, and northern Englishmen who settled the Appalachian chain starting in Pennsylvania in the 18th century and heading southwest, ultimately to Texas, in the 19th. Clinton won 70 percent or more of the vote in 15 counties in the mountains, including the old anthracite country in the east and the bituminous coal country in the west. She won 74 percent in Lackawanna County (Scranton), the home base of Sen. Bob Casey, who endorsed Obama, and she won 79 percent in Fayette County (Monongahela Valley south of Pittsburgh). The latter is on the border of West Virginia, and the results here, as well as in earlier primaries in Ohio, Maryland, and Virginia counties adjacent to West Virginia, suggest that Clinton will win more than 70 percent of the vote in the primary there May 13. Sean Oxendine's excellent map makes this point graphically.
Here's another graphic from blogger Soren Dayton showing the black percentage by census tract in Philadelphia and the voting results by ward there. There's obviously a huge correlation between race and voting, with heavily white wards casting big margins against Obama. Note also that the most heavily Hispanic wards voted heavily for Clinton: ward 7 (70 percent to 30 percent) and ward 19 (65 percent to 35 percent). Hispanic voters may have contributed to Clinton's 60 percent in Lehigh County (Allentown) and 58 percent in Berks County (Reading); those counties were 10 percent Hispanic in 2000, the highest percentages in the state. As I noted in a previous blog post, Latinos, like Jacksonians, have proved a heavily pro-Clinton (anti-Obama?) constituency.
Is Pennsylvania a game-changer, as some Clinton backers proclaim? On the one hand, you could argue that it is just a duplication of the result in Ohio. In my March 28 post projecting the results in a way optimistic to Clinton, I projected a 60 percent to 40 percent margin for her in Pennsylvania and only a 55 percent to 45 percent margin for Obama in North Carolina. Now we know that Clinton's margin in Pennsylvania was only 55 percent to 45 percent, while recent polls show Obama with 59 percent of the two-candidate vote in North Carolina. Moreover, polling in the other May 6 state, Indiana, shows Clinton with nothing like the 60 percent to 40 percent lead I projected there (though the firms showing Obama ahead are unknown to me, so I am not convinced of their results).
My March 28 projections showed Clinton ending up, after the June 3 contests, with fewer pledged delegates (elected in primaries and caucuses) than Obama but more popular votes—whether or not you include the Florida and Michigan primary results (which the Democratic National Committee has ruled don't count because they were too early) and whether or not you include the imputed totals for the Iowa, Nevada, Washington, and Maine caucuses (where the state Democratic parties provided no count of those who participated). In Pennsylvania, Clinton fell short of the trajectory that would take her there.
Even so, Clinton now leads in the popular vote, if you include the Florida and Michigan results, by 121,943 votes. And even if you include the imputed totals for the Iowa, Nevada, Washington, and Maine caucuses, she's ahead by 11,721 votes. It seems to me that this provides the Clinton campaign with an important talking point, though one they're probably reluctant to use over the next two weeks. Reluctant, because the likely Obama victory in North Carolina could erase this popular-vote lead, and) an offsetting Clinton margin in Indiana seems unlikely (or at least risky to project from current polling). But looking ahead from May 6, Clinton is likely to regain that popular-vote lead (including Florida and Michigan) and quite possibly could gain a popular-vote lead counting just Florida and not the more problematic (because Obama was not on the ballot there) Michigan. She'll get big margins in West Virginia on May 13 and Kentucky on May 20, and it's not clear Obama will get a big margin in Oregon on May 20; Obama won the nonbinding February 19 primary in Washington only narrowly. If Clinton wins big in Puerto Rico on June 1, as the one poll I've seen there suggests, that will far outshadow in popular votes any Obama margins in South Dakota and Montana on June 3.
The Obama campaign has had much success selling the press and some superdelegates on the notion that the candidate who wins the most pledged delegates has some moral entitlement to superdelegates' votes. The idea is that the unelected superdelegates should not overturn the verdict of the elected pledged delegates. But I think there are serious arguments against as well as for this proposition. The superdelegates themselves were elected, and to positions that everyone knew carried an entitlement to be a delegate to the Democratic National Convention. A minority of them are elected public officials; most are state party chairmen and members of the Democratic National Committee. You can argue that the party officials were chosen by only a relatively few party insiders, but that's arguably true of pledged delegates selected in caucuses as well. In both cases, everyone knew or could have known the rules, and those who chose to participate had influence over the outcome, while those who didn't, didn't.
As it happens, something like three quarters of Obama's current advantage in pledged delegates comes from delegates elected in caucuses. The Clinton people might argue that these aren't as legitimate as delegates selected in primaries because they were chosen by so few people (perhaps 1.5 million as against 30-some million in primaries). But the Obama people have a perfectly good reply when they say that the Clinton people knew the rules and that if they didn't play competently, the side that did shouldn't be penalized. The problem is, the same argument could be deployed in favor of superdelegates' supporting Clinton. The Obama people could have lobbied these superdelegates better or, back when they were selected, acted to choose different superdelegates more amenable to Obama.
It seems to me that Clinton's current popular-vote lead (with Florida and Michigan) and her likely post-June 3 popular-vote lead (with Florida and Michigan) and possible post-June 3 popular-vote lead (with Florida but not Michigan) give her a talking point with superdelegates. The talking point is that she is the choice of the people. The Obama side can respond, plausibly, by saying that caucus wins produce only small popular-vote margins (or imputed popular-vote margins, as in Iowa, Nevada, Washington, and Maine) and that if those states had primaries, they would have produced bigger Obama margins. To which the Clinton people can reply that Obama has consistently done better in caucuses than in primaries (as in Texas, where he lost the primary and won the caucus) and that his percentages in caucus states if they had held primaries would have been smaller and that in some such cases he would have lost caucus states if they had held primaries. The Obama side can also say that Florida and Michigan shouldn't be counted, because they were held too early and because Obama wasn't even on the ballot in Michigan. But real people did vote there, and in Michigan prominent Obama supporters urged people to vote for "uncommitted," which got a respectable 40 percent of the vote, to Clinton's 55 percent. And then you can get into arguments about the imputed vote in the four caucus states.
The important thing here is that all, or almost all, these arguments, on both sides, are plausible. Reasonable people can advance and believe them. And supporters of the two candidates will, reasonably, advance the reasonable arguments that serve their cause. There is no entirely value-neutral basis, no one moral yardstick, to determine which set of arguments is the more legitimate. Reasonable people will disagree, as they call on the superdelegates to make their decisions.
My sense is that the superdelegates don't want to make their own decisions; they want to ratify someone else's decision. This was underscored when I watched North Carolina Rep. Brad Miller being interviewed on the Fox News election program last night. Miller has a district with a large black percentage and with white-majority rural counties. His black constituents are probably numerous enough to carry the district for Obama; his white constituents will probably vote heavily for Clinton. And he, to paraphrase an old joke, is always with his constituents. He wants to ratify, not decide. The problem is that the results are sending us to a situation in which superdelegates have to decide which decisions they will ratify.
The Pennsylvania primary seems to be sending us to a situation in which Obama will clearly have the most pledged delegates and Clinton will be able to claim that she has the most popular votes. Play by the rules! Count every vote! We will hear the same cries we heard in the struggle for Florida's electoral votes in November and December 2000. The news media are trying to keep careful count of the superdelegates' preferences. Which leads me to ask this question: How would you like to be the superdelegate who casts, or is presented by the media as casting, the decisive vote? The vote that will determine whether Barack Obama or Hillary Clinton is the Democratic nominee. The vote that will determine whether you are overriding the delegates elected by the people or whether you are overriding the people who have cast the votes. The vote that will determine whether the party rejects the first black with a serious chance to be elected president or the party rejects the first woman with a serious chance to be elected president.
Even to be part of a large group of superdelegates that is seen to have cast the decisive votes is to be in a position of political peril and the focus of furious discord. To be the single superdelegate seen as casting the decisive vote is to be in the position of the senator who cast the single decisive vote against the conviction and removal from office of President Andrew Johnson. He was not heard from again until John Kennedy wrote (or had ghostwritten for him) Profiles in Courage 87 years later. Which superdelegate wants to volunteer for that position or find himself or herself in it after a game of political musical chairs?