UC Berkeley

Machine parties have been an important focus in political science during the past decade. The Argentine Partido Justicialista (the Peronist Party, or PJ) is a well-known case of an electorally successful party machine, and scholars have repeatedly noted the salience of brokers (called punteros in Argentina) for the PJ political strategy and its political hegemony. However, key political dynamics that make brokers in efficient political agents and the PJ a successful party machine remained unexplained. This study shows how party machines and their brokers operate by studying the case of the Peronist Party.

Peronism has recently achieved a remarkable consolidation of power through control of electoral districts by networks of municipal mayors and brokers. This has been accomplished especially in the Conurbano Bonaerense (CB), i.e., the 33 municipalities surrounding the city of Buenos Aires. Since the 1983 re-democratization up to 2010, the PJ has won 168 out of 212 (a remarkable 80 percent) mayoral elections in the CB.

Since 1990, two factors have increased the salience of mayors and the pyramidal structure of brokers they command. On the demand side, the poor population in the CB continued to expand, and with it the demand for clientelistic rewards. On the supply side, consolidation of the PJ's hegemonic position was accompanied by a sharp increase of competition within the party, with complex competition among national political candidates, mayors, and brokers to gain the support of the poor.

This study focuses on the brokers and the chain of relationships and strategies involved in a political machine. I analyze, with the help of three formal models, two resources--information and reputation--that influence how brokers practice clientelism, and particularly vote-buying. Given that vote-buying inherently involves uncertainty about voters' political behavior, these two resources become especially important for brokers. The insights from the model are complemented by information gleaned from 170 in-depth interviews I personally conducted with brokers and politicians, including 3 former governors of Buenos Aires, 5 mayors from the CB, 12 municipal directors and secretaries, 22 city council members, and 120 brokers.

The first three chapters present the qualitative analysis. Chapter 1 examines the historical trajectory of this party. Chapter 2 examines differences across CB municipalities on three dimensions: (a) the level of the PJ's electoral success and (b) the degree of intra-Peronist competition, which in turn are crucial in influencing (c) in the level of clientelism. Chapter 3 explores the full portfolio of brokers' strategies--both clientelistic and non-clientelistic.

Chapter 4 to Chapter 6 present formal models of information and reputation. Chapter 4 offers a probabilistic vote-buying model that captures how clientelistic parties allocate resources, according to their informational advantage, when they are uncertain about how their clients will actually cast their votes. This formal model helps show why clientelistic parties win elections with higher probabilities than their counterparts and accounts for why clientelistic parties often target their own constituencies, rather than attempting to win over swing voters. Chapter 5 and 6 develop reputational models showing the importance for PJ brokers of building reputations for accessing and delivering resources, as clients remain loyal to brokers with reputations for delivering. The reputational models capture how history matters for clientelistic deals; it affects voters' expected payoff.

Overall, this study provides new insight for explaining the salience of the networks of brokers for electoral competition in Argentina--where the Peronist party has enjoyed a persistent advantage--as well providing new leverage in understanding other cases of established machine parties.