Presidential Cycles and Time-Varying Bond-Stock Market Correlations: Evidence from More than Two Centuries of Data
11 Pages Posted: 15 Mar 2018 Last revised: 23 Sep 2018
Date Written: March 12, 2018
This paper examines the effect of presidential cycles on financial market correlations using monthly data for the U.S. stock and government bond returns over the historical period of 1791:09-2017:12. Utilizing a dynamic conditional correlation generalized autoregressive conditional heteroskedasticity (DCC-GARCH) model to capture the time-varying correlations, we show that Democratic administrations are generally associated with lower degree of co-movement between the stock and government bond returns. The observed negative presidential cycle effect is robust over various sub-samples identified by structural break tests. The findings are in line with the documented presidential cycle effect on stock market returns and corroborate recent evidence that, when risk aversion is high, agents tend to elect the Democratic Party.
Keywords: Conditional correlation, GARCH, Bond and Stock Returns Comovement, US Presidential Cycles
JEL Classification: C22, C32, D72, G10, G12
Suggested Citation: Suggested Citation