Only a quarter of the nation lives in states where the African-American population is in "the high single digits" -- that is, where "racialized voting kicks in" but where the African-American population is not large enough "to make up for and often overcome the increased white resistance to voting for Obama." But 96 out of the 153 swing electoral votes belong to five key states where voting is racially polarized, and where the black population is not big enough for an increased black turnout to offset the white vote. The states, in order of black population by percentage, are Michigan and Florida (14 percent); Missouri and Ohio (11 percent); and Pennsylvania (10 percent).
The argument that Wilentz should have made is that this spring's primary results show that white reluctance to vote for an African-American candidate could be a real and important factor this November -- and potentially key in these five states, all of them crucial to Democratic hopes. Superdelegates should therefore make a coldblooded calculation to cater to the prejudices of the American electorate in swing states by choosing Clinton over Obama.
Is this argument true? Is it supported by statistical fact? As best as I can tell, no.
As Nyhan pointed out, there is no visible tendency for Obama to fare worse than Clinton as the African-American portion of the population increases. Nyhan presents a graph showing that the higher the black share of the population, the better Barack Obama has done in the primaries. Any increasing racial polarization as the black share of the population rises is offset by greater African-American turnout.
But would this same logic apply to the general election? I believe that it would. First of all, there is no sign that states with demographic compositions like the key five -- Pennsylvania, Ohio, Florida, Michigan and Missouri -- are necessarily hard terrain for Democratic politicians. Consider this graph, which is constructed to show the correlation between percentage of black population and the Democratic vote in the 50 states and the District of Columbia.
With the exception of Washington, D.C. -- that blue dot way off in the upper-right corner, which has both a black majority and a population that votes 80 percent Democratic -- the linear relationship between African-American population share and Democratic vote share is flat. There is some evidence, albeit weak, that states with demographics like the key five are the most hospitable to Democrats. Democratic presidential candidates do fine in states in the middle of the black population range, like New York. It's mostly the states with the lowest and the highest African-American shares -- both Idaho, less than 1 percent black, and Mississippi, nearly 40 percent black, are GOP fiefdoms -- that are the least hospitable to Democrats.
Second, wherever Barack Obama has campaigned this spring, the results in terms of voter turnout have been astonishing and phenomenal. There are two ways to win a general election: mobilizing and achieving a high turnout from your issue and demographic base, and persuading independent swing voters to come to your side. Barack Obama has shown a remarkable power to get independents who do not usually turn out for the Democratic primary to show up and vote for him. And he has shown a remarkable power to turn out his base. Both of these would have to vanish mysteriously in the general election before Obama could be called less "electable."
Thus my judgment is that the argument that superdelegates should support Hillary Clinton because Barack Obama is not very "electable" falls to the ground of its own weight.
I think that Wilentz agrees with me. He talks about a winner-take-all system that doesn't exist, but spends little time engaging the real-world issue of electability. And that is: Which Democratic candidate, Obama or Clinton, has a better chance of carrying Michigan, Pennsylvania, Minnesota, Oregon, Wisconsin, New Hampshire, Iowa, New Mexico, Florida, Ohio, Nevada, Missouri and Colorado against John McCain in November?
And, alas, the arguments that Clinton would fare worse in those states, and that she is less electable generally than Obama, are numerous and distressingly powerful:
- She is a Clinton, and hence will energize the Republican base against her nationwide as nobody else can.
- The press corps has never given her a fair shake, and as Machiavelli once said, we can never forgive and be fair to those to whom we have done injury.
- Barack Obama is a charismatic, historic figure.
- The positions that Clinton has been taking vis-à-vis Obama in the past month appear to open up major vulnerabilities in the fall. McCain's national security experience in Vietnam trumps Clinton's national security experience in Tuzla, Bosnia.
Now, none of these are Hillary Rodham Clinton's fault -- well, except for that last one, which is her fault, or perhaps Mark Penn's. None of these are fair. But they do make me believe that flinty-eyed Democratic superdelegates making coldblooded calculations about the national interest are making a better bet on the future if they decide to support Barack Obama.
About the writer
Brad DeLong is a professor of economics at UC-Berkeley, a blogger and a research associate of the National Bureau of Economic Research, and was a deputy assistant secretary of the U.S. Treasury from 1993 to 1995.
Related Stories
Why Hillary Clinton should be winning
Under a winner-take-all primary system, Hillary Clinton would have a wide lead over Barack Obama -- and enough delegates to clinch the nomination by June.
Story finder (3 ways to search Salon)
Salon Directory (browse by topic)
