Rescuing Rationality? Essays on Voters, Parties, and Accountability in a Polarized World

Abstract

Over the past 25 years, political polarization has noticeably increased in advanced democracies. The four substantive essays in this dissertation ask four different but related questions that arise in the context of rising polarization. First, what explains the emergence and success of bad candidates in electoral politics? Second, once those bad candidates are on the ballot, what explains voters' decisions to support them, causing the breakdown of accountability? Third, does a rise in political polarization lead to a decline in electoral accountability? Fourth and finally, how do elected representatives' choices while in office create political inequalities in representation?

The first substantive essay (Chapter 2) addresses what I call the puzzle of bad candidates. If parties wish to maximize their electoral performance against a backdrop of rational and informed voters who want competent and honest politicians, the emergence and success of bad candidates constitute a puzzle. In this chapter, I offer an analytical examination of why bad candidates often emerge in electoral politics. Parties may select bad (valence) candidates due to a restricted supply of political candidates, imperfect information, the necessary balance between the parties' electoral and non-electoral motivations, and failures of coordination within the party. Established political parties, facing limited competition due to barriers to entry into the political system, can select sub-optimal candidates without risking excessively large electoral downfalls.

The second essay (Chapter 3) explains why voters may decide to vote for bad candidates once they appear on the ballot. Recent research explains the breakdown of accountability and the success of bad candidates with voters' lack of sufficient information or cognitive biases such as partisan-motivated reasoning. In contrast, I show that accountability might still break down even if those shortcomings are absent. Many voters, who are not partisan-motivated reasoners, still decide to forgo accountability and cast a vote for the bad candidate. Original experimental evidence from three countries (the United Kingdom, the United States, and Spain, N = 4200) uncovers three important findings. First, most voters are responsive to differences in valence, partisan, and policy between candidates. However, when faced with a hard trade-off, many forgo accountability even when fully aware they are voting for a suboptimal candidate. Second, moderate voters are less likely to forgo accountability than ideologues. Third, even partisan voters attribute significant positive value to both policy congruence and performance considerations, i.e., they are not blind partisans, revealing a pattern of accountability in parallel.

The third essay (Chapter 4) examines whether a rise in candidate polarization causes a decline in electoral accountability and a lower prevalence of economic voting. Building on the model of voter behavior introduced in the previous chapter, this chapter argues that ideological polarization between candidates causes poorly performing incumbents to have higher vote shares, decreasing electoral accountability. These effects are driven only by elite (candidate) polarization and occur even if mass polarization is low. My argument is empirically supported in two ways. First, original factorial survey experiments in the same three countries vary levels of polarization to study its effects on voter decision-making and accountability. Second, cross-national analyses of post-electoral surveys in 28 democracies between 1999 and 2021 suggest that, indeed, as voters' perceptions of party-system polarization increase, the marginal effect of economic evaluations on vote choice decreases, and poorly performing incumbents are more likely to be reelected.

The fourth essay (Chapter 5), in co-authored work, turns to elite behavior and asks whether the behavior of elected representatives creates inequalities in political representation. In particular, it argues that the constraints of time and resources that politicians face, coupled with their electoral motivations, generate systematic patterns of unequal representation between geographic places. This chapter uses text data obtained from all legislative speeches made in the Portuguese parliament over twenty years (1999-2019), coupled with Google Places and Named Entity Recognition, we are able to quantify the total amount of place-based representation that each place received during this period. The results show that even in a proportional, party-based system, electoral incentives and time-resource constraints lead politicians to focus disproportionately on both small towns and large cities, underrepresenting middle-sized towns and the suburban periphery of large cities.

This dissertation's research carries important contributions to multiple debates in current political science research and within contemporary democracies. I finish with a discussion of the normative implications of my theoretical and empirical contributions to the theory and study of democracy.