Why Obama's Play in Arizona is Smart

Nate Silver, statistician that he is, is skeptical of the wisdom of Obama's decision to start spending advertising money in Arizona. He writes:

I have to say that I'm not a big fan of this from standpoint of marginal electoral strategy. A slew of recent polls in Arizona show the state close, by margins ranging from 1 to 8 points. However, this is the time of year when "close" means something very different from "functionally tied". A 3-to-5 point lead in a state, which is where the Arizona polls average out, is fairly significant at this stage of the contest. That lead still belongs to John McCain.

And needless to say, it is hard to elucidate a scenario in which Arizona serves as some sort of tipping point state. Obama will not perform better in Arizona than in New Mexico, Nevada, or Colorado, neighboring states that have been polling anywhere from 5-20 points more strongly for him. Suppose somehow that Obama were to insult the Pittsburgh Steelers or something and lose Pennsylvania; could Arizona matter then? Not really. The Kerry states less Pennsylvania, but plus Iowa, Colorado, Nevada, New Mexico and Arizona total 267 electoral votes, three fewer than Obama needs for victory. Obama would also have to win something like Montana for it to matter (while losing Virginia, Florida, Ohio, Missouri, North Carolina etc.). Our model thinks that the odds of this happening are something like 800,000-to-1 against.

I think this analysis misses some important considerations. The key point is that Obama is close in Arizona despite having devoted precisely zero attention or resources to that state.

That means there's potentially room for improvement there that there isn't in any of the other close states. By making a token ad buy in that state, Obama gets two immediate (and cheaply purchased) benefits. First, he ensures that the primary topic of conversation in the local Arizona media for the next three days is the sudden emergence of Arizona as a battleground state. That alone is likely to drive up Democratic turnout and--even if it doesn't result in an Obama win--could really help down ticket candidates like Bob Lord. Second, it may well cause the McCain campaign to devote its precious time and resources to defending Arizona. Losing is one thing; losing your home state is another. For pride's sake, McCain won't want to let that happen.

Moreover, as a general matter, there's a diminishing marginal utility to spending all your money in places like Pennsylvania, Virginia, and Colorado, where you've been saturating the airwaves for months now and have made scores of campaign stops. Any money spent in a place like Arizona, which is starved for attention, is going to go a lot further. Any ads played there will be seen by many voters for the first time. Many voters who felt ignored by the national campaign may appreciate the sudden attention. In short, it's going to be a lot easier to move the polls a few points in your direction in a place like Arizona than it is in one of the battle-hardened battleground states.

And finally, on the off chance Obama pulls off an upset in Arizona, it would have significant symbolic value. All winning presidents like to point to evidence of an electoral mandate. What better piece of evidence would there be than a victory in your opponent's home state.

I agree with Nate that Obama shouldn't go overboard here. But dumping a few bucks into an ad buy in Arizona is well worth the money.