14 Year Rule? Not So Much

Jonathan Rauch has been asking for it this week over at Sullivan's place, so he's gonna get. But not on the subject of blogging: on the subject of presidential elections. Rauch is pushing his idea that "you can't be elected president in America if it takes you longer than 14 years to make it from governor or senator to president or vice president." Sorry, but it's bunk.

Here's the original piece Rauch wrote, way back in 2003. At that point, he could argue that every president but one elected beginning with Teddy Roosevelt fit, and 12 of 18 losing major-party nominees from 1904 through 2000. Since then, he's one-for-one with newly elected presidents, but 0-for-2 on losing nominees: John Kerry has been in the Senate since 1985, and John McCain since 1987, so neither passing his test -- meaning that only 12 of 20 losing nominees fit.

As Rauch admitted in 2003, one can in fact come up with all sorts of patterns, although he later claims that it's hard to find one that works for a century. I disagree! Let's see if I can do it...we're talking 19 people, and he's giving himself one mulligan. Here's a list by date of birth -- guess what? None of the 19 were born in the three months from February 7 through May 7. That's a quarter of the year, and no presidents! Or: only one president was born from the 16th through the 26th day of the month. That's 11 days, over a third of all possible days -- and yet only a single exception defied the 16-26 rule. In fact, McKinley qualifies under both of those rules, so make it 19 or 20 out of 20. I'm too lazy to look up which day of the week they were born, but odds are that there's a day missing...try some stuff about the alphabet, name length, shoe size, whatever -- it's very, very easy to find that kind of thing. And while I'd be pressed to come up with a theory explaining the 16-26 thing (c'mon, creative commenters, I know you can do it), I'm sure I could think of some plausible-sounding reason that modern presidents aren't born in late winter/early spring.

OK, I should get to the real point here, which is that we know a whole lot about who wins presidential elections, and we can almost certainly eliminate any major proposed factor that is candidate-based and large. There just isn't enough unexplained variation, once you account for party and familiar "fundamentals" such as the economy and war, for it to be possible that there's a large candidate (or campaign factor). Some, yes. Two points, three, maybe. That's important, and well worth learning more about -- and if you're running a campaign, well worth working hard to exploit, if it's possible to exploit it. But if you think that the candidates' height, or hair color, or handedness, or whatever, matters, it just really can't, very much.

In the general election.

For the nomination, we actually know very little about what factors produce nominees, and we certainly can't rule out anything that makes some sense without testing it first. If you told me that the best looking candidate has a large advantage in nomination contests, I'd be open to the possibility that it's true. If you tell me that about the general election, and you're saying that large is more than a couple of points, then I'm going to tell you that it just isn't plausible. From that it's easy to conclude that if any candidate or campaign based effect "works" for the general election but not for nominations, it's almost certainly bunk. Which is exactly where I'll file Rauch's 14 year rule.