Nate Silver does an in depth analysis of economic factors wrt presidential elections. The conclusion? Economic indicators are not very good predictors of election results.
The broader point — and one thing this evidence is fairly definitive upon — is that the rate of change is what counts. Americans will give a fair amount of credit to a president in an economy that is still below its full productive capacity provided that it seems to be getting better.
Read the whole post (long and detailed) at The New York Times
Which Economic Indicators Best Predict Presidential Elections?