Essays on the Impact of Presidential Elections on the U.S. Stock Market

Abstract

This dissertation consists of two essays that are organized as chapters. In the first chapter, I examine the effect of presidential elections on the timing of turning points of stock market cycles in the United States. The empirical results from duration analysis show that compared to at other times, a market trough, the end of a bear market, is more likely in the period before an election when the incumbent is a Republican; meanwhile a market peak, the end of a bull market, occurs sooner following a Republican election victory. There is also evidence suggesting that bear markets are less likely to end after an election of a Republican president than in other periods. Results from further examination reveal that political control, the political alignment between the president and Congress, has a vital role in the timing of the turning points relative to the elections. In particular, political control found to reduce the probability of a market trough in the pre-election period and this reduction in the hazards for a bear market prior to an election is more significant for a Democratic president. Alternatively, political control boots the prospect of the completion of a bull market in the post-election period, especially when a Democrat was elected. Finally, political control in Congress can substantially shorten the duration of a bear market in the post-election period when the Republicans control both the White House and Capitol Hill. In the second chapter, I develop a dynamic factor model to examine the relations among presidential elections, investor sentiment, and stock market returns simultaneously. Results in the study uncover that there is a sizeable improvement of investor sentiment prior to an election, and this pre-election upsurge in sentiment can explain a substantial portion of the presidential election cycle effect in the stock returns. However, data in the study fail to provide significant evidence that Democratic presidents can install more optimism in the stock market. My result does confirm that there was a market-wide panic during the height of the recent financial crisis in iv 2007-2008. Furthermore, results from the asset pricing tests show that in addition to the conventional risk factors, the market return factor and the factor of change in investment opportunity set, the factor of investor sentiment is a critical component in asset pricing and prediction. By including the sentiment factor, the proposed Augmented Intertemporal Capital Asset Pricing Model (AICAPM) in the paper improves upon the explanatory and predictive power of other competing models such as the Intertemporal Capital Asset Pricing Model of Merton (1973) (ICAPM) and the Fama-French (1993) 3-factor model (FF3).