Essays in Political Economy: Turnout Decline, Political Selection, and Accountability in Democracies under Strain

Advanced democratic systems are operating under mounting strain. Pressures arising from macro-level shocks, together with long-term trends in political participation, polarization, and demographic change put democratic resilience to the test. This dissertation examines how core components of electoral processes in democracies – who turns out, who governs, and who exerts control – operate under such conditions of stress. Across three chapters, it combines causal inference and computational methods in the analysis of large-scale administrative data, social media and legislative text corpora, and survey data to provide new evidence on the electoral consequences of turnout decline, the role of political selection in crisis, and the accountability costs of ideological defection. The first chapter studies the electoral consequences of a sudden contraction in voting access. I exploit a reform that cut the number of polling stations by 30 percent in Italy and use a fuzzy difference-in-differences design at the municipal level together with an original Wikipedia-based measure of party ideology and geo-referenced ITANES survey information. Three findings emerge. First, restricting access to the polls reduces turnout, with effects that persist over multiple elections. Second, this decline reflects selective demobilization: marginal voters with low political engagement, weak partisan anchoring, and specific economic attitudes display higher elasticity to voting costs and are the first to exit the electorate. Third, this access-induced turnout decline reshapes electoral outcomes, generating a lasting rightward shift in the ideological distribution of the electorate, redistributing vote shares from extreme to moderate party groups, lowering observed polarization, and boosting support for populist parties. The second chapter asks whether the characteristics of locally elected officials shaped excess mortality during the COVID-19 pandemic. Focusing on Italian municipalities, it studies whether having a college-educated mayor affected weekly excess deaths. I estimate excess mortality using official death statistics and address endogeneity in political selection with a close-election regression discontinuity design. The results show that college-educated mayors significantly reduced the probability and the level of excess mortality in the first wave, when local discretion mattered most, while effects fade in the second wave as national and regional policies converge. Drawing on fine-grained data on municipal decrees analyzed with text-as-data methods, I show that education proxies for a broader bundle of leadership traits – such as the ability to process complex information and to act early under uncertainty – that translated into better demographic outcomes. The third chapter turns to ideological polarization and electoral accountability. It studies whether voters sanction or reward incumbent mayors who change ideological camp between elections. Using three decades of Italian local elections, an original classification of parties into ideological poles, and detailed fiscal indicators, I compare the re-election prospects of ideological switchers with those of incumbents who remain loyal to their original camp. The analysis reveals an electoral penalty for mayors who cross ideological poles, on the order of 4 to 12 percentage points in re-election probability, even after conditioning on performance in office. Ideological switching emerges as an independent electoral liability, highlighting that voters value ideological congruence as a symbolic dimension of accountability, not merely policy outcomes.