Predicting stock market movements in the United States : the role of presidential approval ratings

Abstract

In this paper we analyze whether presidential approval ratings can predict the S&P 500 returns over the monthly period of July 1941 to April 2018, using a dynamic conditional correlation multivariate generalized autoregressive conditional heteroscedasticity (DCC-MGARCH) model. Our results show that standard linear Granger causality test fail to detect any evidence of predictability. However, the linear model is found to be misspecified due to structural breaks and nonlinearity, and hence, the result of no causality from presidential approval ratings to stock returns cannot be considered reliable. When we use the DCC-MGARCH model, which is robust to such misspecifications, in 69% of the sample period, approval ratings in fact do strongly predict the S&P 500 stock return. Moreover, using the DCC-MGARCH model we find that presidential approval rating is also a strong predictor of the realized volatility of S&P 500. Overall, our results highlight that presidential approval ratings is helpful in predicting stock return and volatility, when one accounts for nonlinearity and regime changes through a robust time-varying model.