Ray Fair is a Yale economist who has been using an econometric model for a number of years to predict the outcomes of presidential elections. It has a fairly good predictive value. His prediction for this November's election is that Bush will win in a landslide, with almost 59% of the total vote.
Fair's model is based on several economic factors such as anticipated or experienced GDP growth, GDP growth greater than 3.2% (he calls this the "good news" factor), inflation rate, duration in office of the incumbent (if the incumbent is running), etc.
The discussion of his model can be found at http://fairmodel.econ.yale.edu/RAYFAIR/PDF/2002DHTM.HTM
I am not endorsing his prediction. I find it Gawd-awful. If it comes true I may retire a year earlier than planned and move to Canada. A landslide would result in a veto-proof Congress.
Now take this and chew on it.