The persistent presidential dummy

Is a Republican or a Democratic party president better for the stock market? Although many observers discuss this issue in only casual terms, one recently published academic study claims to have documented that the market does significantly better under Democratic party administrations. Despite widespread publicity about this in the financial press, the study's results and conclusions are biased by faulty statistical tests. Once the methodology is corrected, the differences in stock market returns under different political parties turn out not to be meaningful. The lessons here extend well beyond the presidential political party effect. Proper research design is essential if we are to reach correct conclusions.