by David Epstein
As the primary season winds down, it's useful to take a look at the whole picture on the Democrat's side. In just about all respects, the race was a tie, as illustrated in the figure below. After the last set of primaries, Obama leads in the popular vote by 50.1% to 49.9%.* Due to the arcane delegate allocation formulas, this lead is magnified to 52.5% to 47.5% in the pledged delegate count.
As the figure shows, these leads are miniscule compared with the range of possible outcomes. Still, if that were all we had to go on, Obama could make a claim that, no matter how close, he's the winner, and that's that.
But this isn't taking place in a vacuum. The point of a primary is to pick the candidate with the greatest chance of winning the general election. Period. And to this end, Clinton has the strongest cards on her side of the table. First, as the figure shows, she leads in the electoral college votes of the states she's won by a whopping 308 to 221, or 58.2% to 41.8%. Moreover, as I explained in my last post on this topic, Obama has a "J-curve" problem; he's winning the states that were either strongly democratic or strongly republican in the past two elections, while Clinton is winning the swing states. If you're a democratic strategist, this makes you worry that he won't run as well in Clinton in the states that count for reversing the results of the last two elections.
Again, this argument has nothing to do with the race of voters, the race of candidates, their gender, and so on. It's just the obvious set of numbers to look at when you're trying to prove which candidate will do better in the general election. Obama should make a substantive argument as to why he thinks he's better positioned to win in November, or so the right thing by the party and just exit the race in favor of Clinton.
* Of course, these days, you can't give primary results without explaining how you're dealing with all the "tricky" state results. I'm counting Iowa, Nevada, Maine and Washington by estimates of the popular vote, since no official results were released. I'm counting Florida straight up, as this is, statistically, about where the result would have been if you extrapolate from similar states. And for Michigan I'm using the "10% rule," which says Hillary would have won by about 10%, as she did in the other big Midwestern industrial states: Pennsylvania, Ohio, New Jersey. Note that this is less favorable to Hillary than counting only her votes in Michigan and none for Obama, or counting her votes and giving him the "Uncommitted" vote. Either of these methods put her in the lead for the popular vote, but I don't think they're realistic, and as a superdelegate I'd want a fair estimate of what the result would have been if those two states had voted regularly, like all the others, and this is probably the best guess.