I'm reluctant to give attention...

...to anything put out with Robert Prechter's name on it (he was a bit of a crank back in the 1980s; I can't believe he's still around), but if it's good enough for Andrew Ross Sorkin of the New York Times, then what the heck. 

Sorkin's column today refers to this study by the Social Science Research Network:

The best single predictor of presidential re-election results that we found was the percentage change in the stock market during the three years that preceded Election Day. Changes in stock prices had a positive, substantial and statistically significant association with incumbents' performances in re-elections. We found that they accounted for more than a quarter of the variation in incumbents' popular vote margins.