Andrew Gelman notes that the political scientists Robert S. Erikson and Christopher Wlezien have released a paper projecting an Obama win with 53% of the two-party presidential vote based on leading economics indicators and current trial heat polling. The economist Ray Fair, who has a well-known model that I’ve criticized in the past (see this Larry Bartels paper [PDF] for more), updated his projection on July 31 and now predicts that Obama will get 51.5% of the two-party vote. The more respected Bread and Peace model of Douglas Hibbs projects an Obama win with 52% of the two-party vote.

While these may seem like relatively narrow projected victories, Gelman notes in a separate post that incumbent parties rarely lose the popular vote by more than a few points. This may be encouraging to the Democrats who think Obama “should” have a bigger lead in the polls right now given the favorable political environment. On the other hand, it highlights the incredible tightrope he has to walk as the first African American presidential nominee of a major party. If Obama’s race costs him more than a point or two in the polls (something the above models cannot forecast and do not incorporate), it’s likely to be decisive.