Harvard's Greg Mankiw blogged today about Republicans, Democrats, and the Stock Market. His main point is that the efficient market hypothesis implies that the incoming president moves the market prior to taking office; therefore, it is meaningless to measure market performance from their respective start and end dates in office and conclude that Democrats are better for the market.
It seems to me an analysis that omitted the first and last years of each president's term would be a much better barometer. Investors with longer time horizons for their investments might skew this measure slightly, but not too much, I would think.
There's still the correlation-does-not-prove-causality issue...
1 hour ago