Close elections

I was giving a lecture over the weekend about why this presidential election is so close. Given that the fundamentals of the race (fundraising, presidential popularity, voter registration, etc.) seem to overwhelmingly favor the Democrats this year, shouldn't Obama be winning by 20 points by now?

My main answer was that blowouts rarely happen in presidential elections. Here's a graph of presidential election margins since 1948:
To be sure, there were some big margins in there. LBJ beat Goldwater by more than 20 points, and Nixon beat McGovern by a similar margin. However, if you omit elections in which an incumbent president was running for reelection, there just aren't that many blowouts. Kennedy/Nixon, Humphrey/Nixon, and Gore/Bush were all nailbiters. The last time a non-incumbent won by more than 10 points was when Ike beat Stevenson in '52. So judging from this history, it would seem more than likely that McCain will get at least 45% of the vote. And, of course, my official forecast model only saw Obama winning by 4-6 points, and that's not accounting for any sort of race effect.

That said, this may not be a normal election. Obama's recent surge in the polls seems largely due to the financial crisis, but neither the crisis, nor the surge, seem to be abating. (See the graphs at the top right of this page, and your 401k.) The question, I think, is whether voters believe this is 1979 or 1929.