US MIDTERM ELECTIONS 2010-2018: Socioeconomic Factors in Turnout and Candidate Ideology

Abstract

This paper seeks to contribute to the understanding of the three midterm elections of the decade: 2010, 2014, 2018. To this end the paper models the effect of local-level socioeconomic factors on electoral turnout, and the ideological position of the winning candidate. Previous scholarship on factors affecting turnout is considered, and brief histories on the elections and relevant political developments are presented. These lens inform the interpretation of 3 panel regressions: (1) modeling the effect of the congressional district’s economic situation, age, and education on the midterm turnout; (2) the effect of these same local socioeconomic factors on the ideology of the winner, as measured by Poole Rosenthal DW-Nominate score, and; (3) the effect of these local socioeconomic factors on an ideology score constructed for this paper using the project VoteSmart NPAT data set. Models include fixed-effects estimators for each US state. Model (1), turnout, finds statistically significant relationships between turnout and the variables for unemployment, inequality, median household income, and education. Amongst other findings, the estimated effects for higher education was found to be higher in 2018 than in previous years, controlling for other factors, which is interpreted as evidence for anti-Trump enthusiasm in more educated districts. Model (2), nominate score, suggests that congressional district unemployment levels are associated with more liberal candidates. This magnitude was found to increase across the period. Model (2) further suggests education was associated with more liberal winners, only in 2018. Model (3), NPAT score, corroborates the unemployment finding, but does not suggest a relationship with education. Model (3) is found to have flaws, which are discussed. Through these models and modern history, this paper examines a wide range of topics in the developing and exciting subject of 21st century American politics, including: record highs and lows for turnout, record high campaign spending, Obama, Trump, the rise of social media, gun control, health care, voting behavior, and ideology.