UC Berkeley

Elections in the wake of transitions to democracy are often structured by formal and informal institutions that benefit anti-democratic elites and that reduce the potential of expanded suffrage to affect policy. While most of the writing on counter-majoritarian institutions focuses on formal rules, the political consequences of informal institutions that can distort elections' capacity to accurately represent the electorate is less well understood. Drawing upon historical and recent evidence from Brazil, this study analyzes the specific aspects of the mechanics of Brazilian elections that interacted with informal practices to over-represent rural conservative interests, increased the ability of conservative political machines to win elections, and de-facto dis-enfranchised large swaths of the Brazilian electorate. These informal practices had substantial consequences for the quality of representation both during Brazil's first experiment with mass democracy between 1945 and 1964, as well as its most recent experience with widespread suffrage. Furthermore, this analysis considers the conditions under which interventions--such as the provision of information--can improve the extent to which elections can induce accountability and representation.

he first chapter examines a long standing anti-democratic practice in Brazil: the de-facto disenfranchisement of millions of Brazilian voters and widespread voting fraud caused by the interaction of a difficult paper ballot and permissive electoral rules. To provide such evidence, this analysis exploits the phased adoption of electronic voting in Brazil, a reform that increased the effective franchise in legislative elections by about 33 percent and eliminated fraud in the vote counting process. The research design relies on the fact that the reform was initially implemented in municipalities with an electorate over an arbitrary threshold and consequently allows for a regression discontinuity design. The two distinct effects of electronic voting--the enfranchisement of illiterates and other low information voters and the dramatic reduction of fraud--had consequences for the composition of the national legislature. Against the predictions of recent economic models of democratization, the data show that the enfranchisement of illiterates and other low information voters caused a small increase in the vote shares of right-wing candidates. More importantly, newly enfranchised voters were dramatically more likely to cast a "party list" or partisan ballot as opposed to a personal or candidate ballot, which benefitted Brazil's more programmatic and ideologically coherent parties. In states with hegemonic conservative parties, the introduction of electronic voting induced a roughly 20 percentage point swing against "political machine" candidates, which is attributable to a substantial reduction in fraud. Overall, the most important consequences of the reform was the strengthening of Brazil's major parties and a weakening of dominant subnational conservative political machines.

The third chapter explores how interventions that increase the amount of information available to the electorate can affect political accountability. An underlying assumption of much of the literature on political corruption is that if voters are provided with information about the performance of politicians by actors in civil society such as the media and non-governmental organizations, then the election of corrupt politicians is less likely. Yet, heterogeneous views about the importance of corruption can determine whether increased information changes electoral outcomes. If partisan cleavages correlate with the importance voters place on corruption, then the consequences of information may vary by candidate, even when voters identify multiple candidates as corrupt. This chapter provides evidence of this mechanism from a field experiment in a mayoral election in São Paulo where a reputable interest group declared both candidates corrupt. Informing voters about the challenger's record reduced turnout by 1.9 percentage points and increased the opponent's vote by 2.6 percentage points. Informing voters about the incumbent's record had no effect on behavior. This divergent finding is attributable to differences in how each candidate's supporters view corruption. Survey data and a survey experiment show that the challengers' supporters are more willing to punish their candidate for corruption, while the incumbent's supporters lack this inclination.