Harry B. wants to know what I think of Ray Fair's model for presidential elections. It's a surprisingly accurate equation that takes, as inputs, economic factors, party control, etc, and spits out the expected results of the election. It's rarely wrong, and has missed the actual total by 4 percent exactly once. This year, it's predicting that McCain wins a mere 46.7 percent of the popular vote. The honest answer is I don't really know what I think of it. It's true that the economy is bad this year, and that should, in theory, offer a significant advantage to the Democrats. it's also true that Barack Obama is black. And that John McCain is over 70. And that Obama isn't very experienced. And that McCain has never been attacked from the left. And that we don't know everything Jeremiah Wright has ever said in Obama's presence. And so on, and so forth. The point of the model is to disprove exactly what I just did -- that it's misleading to get too hung up on the individual characteristics of any particular election. Intuitively, that seems wrong to me. But the model has been retty predictively successful. In any case, I think Harry B's original comment is way more interesting than my musings, so I'm attaching it below the fold. Folks with more knowledge of these models should weigh in at comments.
