The Politics of Polarization: Governance And Party System Change in Latin America, 1990–2010
Samuel Handlin
Paper originally published as: Kellogg Institute for International Studies Working Paper 401, University of Notre Dame, November 2014
This paper uses case studies of Brazil and Venezuela to demonstrate how the quality of governance during the 1990s in each country impacted patterns of political polarization through a particular mechanism: by shaping factional conflict and strategic adaptation among and within left parties, such that the left consolidated in a moderate or radical direction. In conducting these case studies, the paper draws upon a large amount of primary source evidence. The paper is part of a larger book-length research project for which over 300 party sources (documents, memoirs, and contemporaneous writings of left leaders) and over 500 news articles were collected and analyzed. These primary sources are used to bolster both descriptive and causal inferences. The paper draws upon several sources to simply give evidence for important descriptive claims regarding dynamics within the left. The paper also draws upon multiple sources for causal inference, particularly to show how the contextual variable of governance quality impacted the ability of radical forces on the left to convincingly articulate alternative paths to electoral viability that did not involve left-right programmatic moderation, thus bolstering or undermining the relative fortunes of radicals and moderates in internal conflicts over the direction of left parties and movements.
Research Trajectory
The hypothesis that the quality of governance or state performance during 1990s had a decisive impact on whether the left in Venezuela and Brazil (as well as other countries in Latin America) moderated or radicalized was developed prior to the collection of these data: It was first inductively generated from observing general characteristics of the cases, as well as a correlation between explanatory variables and outcomes. This hypothesis was then elaborated in preliminary form, while not empirically supported to a substantial degree through primary-source evidence, as part of a doctoral dissertation. At this stage, the hypothesis was discussed in terms of the presence (or not) of a “legitimacy crisis” in each country that fostered anti-systemic and anti-status quo politics. Later, as the work developed further, I modified this hypothesis to capture the underlying political factors that might be driving such a crisis, now conceptualized as the quality of governance or state performance.
Data Collection
Primary-source data on factional conflicts and programmatic adaptation within the partisan left in Brazil and Venezuela was collected in order to test the governance hypothesis, as well as alternative theories. The goal of the data collection was to gather information relevant to factional conflict and adaption within major left parties, which was seen as a logically necessary step to examine for any explanation of whether the left consolidated in a radical or moderate direction. As such, by design the data collection was not specifically geared toward any specific hypothesis (such as the quality of governance) that might help explain patterns of conflict and adaptation among the left.
I decided that both news sources and party sources (documents, memoirs, and the contemporaneous writings of left leaders) would be valuable for the project. For news sources, I picked one major newspaper in each country that was known for relative political neutrality (El Universal in Venezuela, Folha de São Paulo in Brazil), defined the time period during which decisive changes and factional disputes occurred within the major left parties (1993-1998 in Venezuela, 1994-2002 in Brazil), and searched for articles that bore on questions of factional contestation and programmatic adaptation. I intentionally cast a wide net in this search, such that many articles only touched on the topics tangentially. The archives of El Universal were only available on microfilm (obtained from the Center for Research Libraries), such that the search was conducted manually. The archives of Folha de São Paulo were available online, such that the process could be conducted through keyword searches. With respect to party-related documents and primary sources, the goal was to gather as much information as possible that seemed relevant, from whatever sources could be found. For Brazil, a major archive of PT documents and internal party press materials was available on microfilm (93 reels). An exhaustive search through this archive was the basis for data collection. For Venezuela, sources had to be accumulated in a more ad hoc fashion, searching for relevant party documents in various locations and acquiring collections of essays and memoirs by left party leaders. As with the news sources, the strategy was to cast a relatively wide net and acquire anything that seemed potentially relevant, not to look only for specific documents likely to tell a particular story.
Data Analysis
Data analysis has involved sifting through this large collection of news and media sources, conducting process tracing in order to weigh the probative value of evidence that bears on the governance hypothesis (affirmatively or not) and evidence that might also support or undercut particular alternative hypotheses. In this respect, little weight has been placed on the particular number of documents that do or do not provide probative value, for several reasons. First, the data collection was designed to cast a wide net and therefore provide a diverse set of sources. Some of these documents provide support on important descriptive claims, others provide evidence regarding specific hypotheses, and many others turned out to have little value whatsoever. Second, as scholars of process tracing emphasize, it is often the case that one or two pieces of critical evidence – generated at critical moments and decisively bearing on the questions at hand – carry more probative value than a numerically greater amount of evidence that doesn’t directly bear on or adjudicate between hypotheses."
This paper uses case studies of Brazil and Venezuela to demonstrate how the quality of governance during the 1990s in each country impacted patterns of political polarization through a particular mechanism: by shaping factional conflict and strategic adaptation among and within left parties, such that the left consolidated in a moderate or radical direction. In conducting these case studies, the paper draws upon a large amount of primary source evidence. The paper is part of a larger book-length research project for which over 300 party sources (documents, memoirs, and contemporaneous writings of left leaders) and over 500 news articles were collected and analyzed. These primary sources are used to bolster both descriptive and causal inferences. The paper draws upon several sources to simply give evidence for important descriptive claims regarding dynamics within the left. The paper also draws upon multiple sources for causal inference, particularly to show how the contextual variable of governance quality impacted the ability of radical forces on the left to convincingly articulate alternative paths to electoral viability that did not involve left-right programmatic moderation, thus bolstering or undermining the relative fortunes of radicals and moderates in internal conflicts over the direction of left parties and movements.