In older democracies, political parties often traded individualized rewards--food, cash, clothing--for voters' support. Such strategies, while rare in these countries today, are still deployed in developing countries such as India, Argentina, and the Philippines. In Brazil, vote buying was once widespread but now is in decline. What explains this shift? Answering this question is essential to understanding why some democratic governments seek to respond to the needs of their citizens.
This dissertation explores the process by which vote buying is replaced by programmatic politics, where politicians seek to attract voters' support with policy proposals. In the NSF-funded component of the project, the researcher will conduct two large sample surveys in Brazil in order to test original hypotheses about why voters shift their support to programmatic politicians. A nationally representative survey will identify the type of people who sell their vote. Another survey will focus on the state of Bahia, which has undergone dramatic political changes over the past decade. This targeted survey will allow the researcher to test hypotheses about local characteristics that support a shift to programmatic politics. Data from this project will also be used to conduct cross-national comparisons with other countries where vote buying has been prevalent.
The literature on vote buying, which tends to focus on improving understanding of how party machines work, does not offer a good explanation for why this strategy gives way to programmatic appeals. One long-running debate, for example, pits scholars who believe that party machines target apolitical "swing voters" against others who think that politicians focus on cultivating relationships with "core voters." Why politicians might forsake vote buying altogether--or lose to those who do--is left unasked.
Some older research did explore why vote buying declines, but explanations were generally tied to broader social shifts and lacked rigorous empirical testing. Such research does not fully account for the regional disparities in how politics is practiced in Brazil, nor does it provide sufficient insight into why particular political machines break down.
This project takes prior research and current debates as a point of departure. While most contemporary research on vote buying assumes that current political systems are immalleable, this project breaks new ground by providing data that will allow for rigorous quantitative analysis of the factors that lead voters to sell their vote. By explaining regional variation in the prevalence of vote buying in Brazil, this research will provide insight into what causes vote buying to decline--both in some regions of contemporary Brazil and in other contexts. This study will also provide important but previously unavailable data to other researchers.
In addition to advancing the literature on vote buying and democratic consolidation and generating new data on topics that have received little scrutiny, this research will have important implications beyond academia. Understanding why some political systems adopt programmatic politics while others do not is an important step towards understanding--and overcoming--the impediments to constructing a democracy that truly represents its citizens.
How do clientelist regimes, where politicians mobilize support through personal favors, transition to programmatic ones, where politicians rely on electoral platforms and policies to sway voters? Clientelism constrains democratic representation and economic growth across the developing world, yet little is known about why clientelist regimes transition to programmatic ones. In this dissertation, I draw upon the case of Brazil to develop our understanding of why transitions occur. Brazil offers an excellent case for study as long dominant political machines recently have given way to programmatic policymaking. I present evidence that clientelism is waning in Brazil. As poverty relief traditionally took the form of handouts from politicians to voters, I pay special attention to new and purportedly technocratic antipoverty programs like Bolsa Família, which provides a stipend to impoverished families and reaches a quarter of Brazil’s population. My finding that many policy innovations in Brazil are programmatic helps explain why clientelist networks are deteriorating: with the utter poverty that once characterized Brazil eradicated, voters considering whether to vote against clientelist politicians have less to fear from retaliation. What explains the growth of programmatic policymaking? I hypothesize that the adoption of programmatic policies requires sufficient bureaucratic capacity. Bureaucratic capacity describes the ability of the civil service to implement programmatic policies according to technocratic (rather than political) criteria. This capacity makes politicians’ claims that they will enact such programs more credible. I use a multifaceted approach to test for this link. I begin by tracing the historical relationship between civil service professionalization and policymaking in Brazil. Early attempts at reform generally were overwhelmed by demands for clientelist resources. Brazil’s military, which had adopted a Weberian administrative ethos following World War II, was aghast at the venality, corruption, and clientelism that characterized Brazil’s political system and civil service. After the 1964 coup, Brazil’s military rulers expanded the role of technocratic ministries and agencies. These agencies were insulated from (clientelist) politics and handled economic management while social policies remained oriented towards patronage and clientelism. Future democratically-elected leaders could draw on this technocratic capacity should they choose to enact programmatic social policies. I provide micro-level evidence that bureaucratic capacity can encourage programmatic voting. I use a novel measure of programmatic voting and show that municipalities with greater capacity have higher levels of programmatic voting. I also developed a survey to investigate whether individuals’ perceptions of bureaucratic capacity affect the credibility of campaign promises and their willingness to support a programmatic candidate. This survey uses experimental vignettes to show that higher capacity increases voters’ willingness to vote for programmatic candidates. As clientelism subverts the democratic process, my dissertation reveals how polities can attain superior forms of democratic representation. In addition, since programmatic systems provide higher levels of public goods, they also foster improved economic outcomes and democratic consolidation because voters are more likely to reap the gains associated with democracy. Finally, by highlighting the importance of bureaucratic capacity, this dissertation has policy implications for other political systems that find themselves burdened by clientelist politics.
