Within the Margin

There’s no shortage of good writing out there about the politics and policy of the Paul Ryan vice presidential pick. But electoral history fascinates me as much if not more than current races do, so what I enjoy most about the pick is that it will give us a new and useful data point about the relationship between vice presidential geography and electoral outcomes. That’s because this will be the first VP nominee in two decades to come from a state that will be seriously targeted by both campaigns in the fall. Recent presidential elections have featured VP picks from states that were safe, or perceived to be safe, for one party the other:

• In 2008, Alaska was actually looking competitive before Sarah Palin was named to the Republican ticket. Summer Polls showed Obama within a few points of McCain. After Palin joined the fray, Obama did not seriously contest the Last Frontier, so we can’t really take too much away from the result. Plus, it’s not like Alaska was considered to be a competitive state in any other recent presidential elections. Nor is it likely to be competitive this year. So we can’t view Palin as a swing state pick in 2008. Joe Biden, of course, came from Delaware, which has been safely blue for a while now.
• 2004 saw a Republican running mate from Wyoming, where the result was never in doubt. John Edwards was chosen by Kerry ostensibly to make the South competitive, but actually showed little interested in pursuing any state in the region besides Florida. In the event, Edwards impact on North Carolina was negligible, but we’d have learned more if the Kerry/Edwards campaign had actually tried to win there.
• Both running mate choices in 2000 were made without any nod toward geography. Wyoming (Dick Cheney) and Connecticut (Joe Lieberman) were never in play.
• In 1996, the Dole campaign was never going to carry New York, so picking Jack Kemp tells us little. Bill Clinton’s holdover running mate from 1992, however, was Al Gore: chosen in 1992 to double-down on Clinton’s appeal as a youthful moderate Southerner while making up some of the ticket’s deficit in military expertise and experience. Gore indeed succeeded at carrying his native  Tennessee in both 1992 and 1996, ending a streak of three consecutive Democratic losses there in presidential races. How much of that was Clinton’s appeal from next-door Arkansas, versus Gore’s strength as a native son? That’s hard to pin down, especially since Gore went on to lose Tennessee as the presidential nominee himself in 2000. But the Clinton/Gore win in 1992 represents the most recent time that a state flipped from one election to the next while running a VP nominee from said state.

While Wisconsin was an easy Obama win with 56% of the vote in 2008, and he has generally led this year’s Wisconsin polling by  mid-single digits margins or above,  the state has never appeared fully out of reach for Romney: Nate Silver at FiveThirtyEight currently projects a 4.4% or 5.4% Obama win, depending on whether you use the Now-cast or the November projection, while RealClearPolitics‘ Wisconsin polling average shows Obama up 5.4%. Certainly not a bad place for Obama to be, but the conventional wisdom is that the Great Lakes states are much tougher terrain for the president than they were four years ago.

Silver has worked the numbers (of course) and finds – with plenty of caveats owing to the nature of the data at hand – that from 1920 to 2008, a running mate was worth about 2.2 points on average in his home state, in terms of margin of victory (or defeat). In recent decades, it’s been more like four points – with an even stronger caveat about small sample sizes with many other factors involved. That’s a fairly small impact at the state level…unless we’re talking about a truly marginal state. Wisconsin was a blowout 14-point win for Obama in 2008, but John Kerry only won it by 0.38% in 2004, and Gore squeaked by with a 0.22% win in 2000. Silver has already adjusted his model to incorporate the historically-average 2.2% boost in Wisconsin; the RCP average to which I referred above is close enough that even a small additional impact could push Wisconsin deeper into toss-up territory, and at that point the Romney campaign may be even more inclined to devote its considerable financial resource to winning the state.

I’m guessing that the Romney/Ryan ticket will not see its Pennsylvania numbers improve much in the coming weeks, given that the Keystone State’s population skews older than much of the nation and will be hearing a lot of about the effect of Ryan’s budget policies on Medicare and Social Security. By mid-September, it might make a lot of sense for the campaign to concede those 20 electoral votes, while re-doubling their efforts to win 10 in Wisconsin.