Monthly Returns During Presidential Election Years
Do the hopes and fears of presidential elections in the U.S. affect the “normal” seasonal variation in monthly stock market returns? To check, we compare average returns and volatilities (standard deviations of returns) by calendar month for the Dow Jones Industrial Average (DJIA) during years with and without quadrennial U.S. presidential elections. As a robustness check, we also check years with and without biennial U.S. congressional elections. Using monthly closes for the DJIA over the period October 1928 through October 2012 (about 84 years and 20 presidential elections), we find that:
The following chart compares average DJIA returns by calendar month for four subsamples:
- Years during which there is a presidential election (about 20 observations).
- Years during which there is a congressional election (about 41 observations, including presidential election years).
- Years during which there is no presidential election (63 observations).
- Years during which there is no congressional election (42 observations).
The average return for all months over this period is 0.54%. Results suggest that presidential elections may have special effects on January, April, May and August. However, the presidential election subsample is small.
What about an effect on stock market volatility by calendar month?
The final chart shows the standard deviations of DJIA returns by calendar month for the same four subsamples. Results suggest elevated stock market volatility during May, July and August, and perhaps, subdued volatility in September. However, as noted, the presidential election subsample is small.
In summary, evidence from simple tests supports some belief that presidential elections introduce some disruptions into typical U.S. stock market seasonality.
Cautions regarding findings include:
- As noted, subsample size is small, limiting reliability of any findings.
- Changes in U.S. political processes and the stock market environment could affect seasonality.