POLITICS: How Clientelism Actually Works

Berkeley Review of Latin American Studies, Fall 2015

People waiting for handouts from the Peña Nieto campaign during the 2012 election in San Luis Potosí, Mexico. (Photo by Germán Canseco/Procesofoto.)
People waiting for handouts from the Peña Nieto campaign during the 2012 election in San Luis Potosí, Mexico. (Photo by Germán Canseco/Procesofoto.)

Read this article on ISSUU
Download as printed (.pdf)
More about this event 

POLITICS: How Clientelism Actually Works

 In many parts of the developing world, political leaders provide goods and services to secure votes from their constituents. The conventional and theoretical logic of such clientelistic exchanges vary in scope and reach, often — but not exclusively — mediated and coordinated by on-the-ground brokers with extensive networks in their communities. In some settings, politicians or their brokers target the electorate at large, while in other cases these exchanges target particular voting groups, such as swing or absentee voters.

By engaging in a quid pro quo with voters, politicians or their brokers seek to secure electoral victory by persuading and mobilizing voters in their favor. However, it is unclear whether targeting the at-large electorate or specific swing groups is more feasible, and which strategy — if any — maximizes the politicians’ chances of electoral victory. Since brokers often mediate the exchanges, politicians may have little idea about what happens when voters are courted. For example, brokers motivating swing voters or reinforcing existing ties with loyalists? Despite the best efforts by social scientists to document this phenomenon, with and without brokers, there are conflicting views on how clientelism is carried out across the globe and particularly in developing countries.

Examining three countries with widespread clientelistic exchanges between the electorate and their politicians — Venezuela, India, and Argentina — Thad Dunning, the Robson Professor of Political Science at UC Berkeley, discussed the puzzle of distributive politics at the Center for Latin American Studies. His presentation focused on the logic of mobilization and persuasion in clientelism: how these exchanges are focused to reach different electoral constituencies.

Dunning identified four types of citizens from the perspectives of both the politicians and the brokers: potential voters, certain voters, loyalists, and swing voters. The first two types describe whether a voter is likely to show up to the polling booth or not, while the latter two categories characterize the voters’ political leanings. The four categories interact dynamically. Voters can either be potential or certain and either loyalist or swing. Conventional theories on clientelism tell us that politicians and brokers will target potential/ loyal voters to mobilize them and certain/swing voters to persuade them. For Dunning, however, this view fails to convey how distribution actually occurs through these practices. More importantly, in the cases where brokers act as the middlemen between politicians and their constituencies, conventional theories do not fully account for brokers’ agency in the exchanges.

In Venezuela, for example, Dunning describes how Chávez and his supporters used a spending redistribution program, Misiones, to target voters when faced with a competitive reelection campaign in 2004. Chavistas built a database, Maisanta, with information on whether individuals had voted in the past and if they were considered loyalist or not. Conventional clientelism theories would hypothesize that Chávez and his followers would have targeted individuals who had voted previously and were potential swing voters (certain/swing) and individuals who had not voted in the past and were loyalists (potential/loyal) to maximize Chávez’s reelection prospects. However, Dunning found that certain/loyalist voters overwhelmingly received the funds from the Misiones spending program. Instead of targeting indecisive voters, Chavez’s on-the-ground operators opted to reinforce existing ties with loyalists who had given them support in the past. For the most part, chavistas opted out of gaining a foothold in new constituencies. Their patronage strategy largely reinforced existing ties with potential/loyal voters. Although this seems limiting in trying to gain votes, the chavistas’ strategy worked: they gained control of the National Assembly and re-exerted their control on the Presidency. In an effort to unravel this contradiction on conventional expectations of clientelism, Dunning studied these practices in two additional settings.

Through a survey questionnaire in India, Dunning explored how voters reacted to clientelistic offers. Randomizing partisan and caste ties between survey respondents and hypothetical candidates for village council president, Dunning asked different groups of voters who they expected to give them some type of good or generate some sort of policy in exchange for their vote. When voters and politicians were copartisans, Dunning found that the former were slightly more likely to expect a quid pro quo. Not only do more targeted goods go to partisan voters, but there is also a greater demand from partisans for benefits in return for their continued support.

Dunning further tested his argument by surveying brokers in Argentina. Building a network of politicians and brokers across various provinces in Argentina, Dunning randomly sampled the broker population and conducted extensive surveys with them on how they mobilized voters. This strategy allowed him to generate a relatively representative sample of the broker population in Argentina. Presenting a hypothetical situation in which politicians gave them authority over how to allocate spending programs in order to secure votes, Dunning asked the brokers whether they would target loyalists or swing voters during the campaign. Once more, against conventional theories and expectations of clientelism, Dunning found that an overwhelming majority of brokers (more than 65 percent) would target loyalist over swing voters. Since politicians expect brokers to secure votes — and the former cannot fully distinguish between loyalist and swing voters — Dunning argued that brokers opt for the most secure venue in courting voters.

From his three empirical exercises, Dunning concluded that clientelism operates in situations with varying degrees of information among politicians, brokers, and voters. Targeting certain/loyalist voters might be a good strategy — against conventional expectations — because not only are those type of voters more likely to show up on the date of the elections, they are more likely to make brokers look as if they are actually doing their job. For Dunning, understanding information asymmetries and agency are crucial in unraveling the machine behind targeted distribution schemes.

Thad Dunning is the Robson Professor of Political Science at UC Berkeley, and directs the Center on the Politics of Development. He spoke for CLAS as part of the Bay Area Latin America Forum on October 21, 2015.

Carlos Schmidt Padilla is a graduate student in the Charles and Louise Travers Department of Political Science at UC Berkeley.