Dems 2008: Predictions and number-crunching, full of sound and fury [Karl]
Former Clinton pollster Dick Morris believes that Sen. Barack Obama will defeat Sen. Hillary Clinton and win the Democratic presidential nomination.Ã‚Â Tucked into his commentary was this:
March 4th will, at worst, be a wash for Obama with his probable wins in Ohio, Rhode Island and Vermont offsetting his probable defeat in Texas. (Although in Texas’ open primary, Republicans and Independents may flock to the Dem primary to beat Hillary).Ã‚Â And then come a list of states almost all of which should go for Obama, including likely victories in Pennsylvania, North Carolina, and Indiana. By the convention, he will have more than enough delegates to overcome the expected margins Hillary may rack up among super delegates.
Morris, however, provides no data or argument for these predictions.
Bill Kristol claims that Obama will take Ohio, apparently based on a regression analysis at the Daily Kos.Ã‚Â However, Josh Patashnik at TNRÃ‚Â notes that this model did not fare well in last weekend’s contests, suggesting “overfitting” in the model.
Patashnik, meanwhile, gets his own comeuppance in his suggestion that Clinton has a shot in Virginia tomorrow:
No polls have been taken in the state since October, when he trailedÃ‚Â Clinton 49 percent to 25. Needless to say, those numbers are no longer accurate, but they do suggest a strong base of support for Clinton in the state–especially given that in 2004, 56 percent of Democratic primary voters were women. And even if Obama wins the statewide vote, the delegate count might not break his way, since his strongest supporters, African Americans, are packed heavily into two of the state’s eleven congressional districts.
In reality, there have been several new polls in Virginia, each giving Obama at least a 15 point lead.Ã‚Â He might have a point about the delegates if the margin ends up being smaller than that.
Jonathan Chait argues (also at TNR)Ã‚Â that Ohio and Texas really aren’t all that important, though only by ignoring Ohio’s swing-state status, which will weigh on the minds of voters and super-delegates.
The spreadsheet “inadvertently” leaked from camp Obama has him winning Wisconsin 53%-46% (contrary to the latestÃ‚Â ARGÃ‚Â poll, fwiw); losing Ohio by the same margin (whereÃ‚Â the Columbus DispatchÃ‚Â had Clinton way ahead, but before Edwards dropped out and 20% undecided); losing Texas by only 4 points (as opposed to the 10 point CW); and losing Pennsylvania by only 5 points (the Daily News had Clinton up by 20, again before Edwards withdrew).
Patrick Ruffini has posted a spreadsheet with intriguing estimates:
At the end of the primaries, I project 1,607 pledged delegates for Obama and 1,548 for Clinton, leaving 418 and 477 superdelegates respectively left to secure the nomination.
However, if Michigan and Florida are allowed to revote in a primary, and (conservatively speaking) Clinton reprises her 15 point margins in both states, ObamaÃ¢â‚¬â„¢s pledged delegate lead shrinks to 12 Ã¢â‚¬â€ 1,740 to 1,728.
At this point, ObamaÃ¢â‚¬â„¢s proposal to have the superdelegates vote according to the majority in their state would probably flip the lead to Clinton, since Clinton will have probably won more states narrowly, amplifying her lead.
That would be ironic, but may explain why Obama strategist David Axelrod was singing a different tune about super-delegates as independent for Matt Lauer on the Today Show.
Not ironic, but amusing, is Ruffini’s non-trivial argument that Puerto Rico may swing the nomination for Clinton.Ã‚Â He links to no less an expert than Michael Barone, who argues that the commonwealth’s delegates may not be proportionately allocated, which could make its bloc of 63 delegates decisive.