Abstract:
Excess returns in the stock market are significantly higher during Democratic presidential administrations. Previous research concludes that partisan return differentials are anomalous since they are not due to differences in required returns. We find that partisan return differentials are, instead, likely due to differences in cash flows - capital income growth - during the first years of presidential administrations as predicted by the rational partisan model of the business cycle. The first major finding of this paper is that there is a statistically and economically significant partisan difference in capital income growth in the first year of presidential terms. The second finding of the paper is that significant partisan differences in excess returns are also found only in the first year of presidential terms. Further, it is differences in unexpected returns during that first year that is the source of partisan return differentials. We find no statistically significant partisan differences in unexpected returns during the rest of the term. This result holds across market capitalization deciles and book-to-market value deciles. The third finding is that there is a positive and statistically significant relationship between unexpected returns and capital income growth and real GDP growth one and two quarters ahead. Lastly, we find that the unexpected returns are related to the degree of electoral surprise as predicted by the rational partisan model. We conclude that that there is strong evidence in favor of the rational partisan model as an explanation for partisan return differences in the stock and bond markets.