Thursday, August 07, 2008
The conventional wisdom has it down pat: A bad economy works against the candidate from the party in power as voters take out their rage and fear on the president's party and back the challenger, just like they did in 1992. But this is not a normal economic slowdown (or recession) and Obama is not a normal challenger. I think the conventional wisdom may be dead wrong.Voters tend to be more retrospective than prospective, in study after study of voting patterns. Models that predict voter behavior based on past economic performance tend to do pretty well. So Dismal.com in June forecasted Obama to win 380 Electoral College votes. Ray Fair's model generates now a very small gap between McCain and Obama (about 3% of the two-party vote share.) If you input data for third quarter GDP greater than 3.2% (i.e., there's good news about the economy on that last weekend before the election), McCain pulls ever-so-slightly ahead. (I don't know the error bands here, but I'll bet both results are within two standard deviations.) House Republicans are seen doing relatively poorly.
It is not so much that unemployment is so high (5.7%) or that the economy is in the tank (1% growth this quarter) as that everything seems to be falling apart.
It is interesting, though, that the economy (ex energy) has not gained more traction in the campaign. Even the quite different positions of the candidates on taxes has not to this point grabbed the attention of the public. Perhaps they're just not paying attention, though new polls say otherwise. Or perhaps it's what they are asked to pay attention to; the campaign is one more of personality than policy. Still, you'd think a campaign that focused on uneasiness about voters' future economic fortunes would favor the Obama campaign. I'm puzzled he has not done more with that. Perhaps it's because the economy isn't as bad as many think.
UPDATE: Obama fatigue? Can campaigns have a third stage of production? Bill Luksetich and I wrote a paper on campaign financing and House elections (Economic Inquiry, 1991) in which we modeled campaign spending as having diminishing marginal returns; I don't think we looked to see if anyone got to negative marginal returns. I'm traveling today and will have to check to see what we wrote when I get back. But negative returns might make the prediction equation perform less well.