Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS (oxfordre.com/politics). (c) Oxford University Press USA, 2019. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy and Legal Notice).

date: 14 October 2019

Clientelism in Latin American Politics

Summary and Keywords

Clientelism is a type of nonprogrammatic linkage strategy that political parties deploy to win elections. Specifically, the concept refers to the personalized and discretionary exchange of goods or favors for political support. Scholars of comparative politics investigate variation in the prevalence of clientelism across countries, as well as the organizations that parties create to distribute personalized gifts and favors. A large body of work also studies the types of voters more commonly targeted by machines. The debates about the determinants of clientelism and specific targeting patterns are important because they inform broader discussions about democratic quality in Latin America and other developing regions, where nonprogrammatic linkages such as clientelism are common. In particular, the literature on clientelism has implications for discussions about the use and misuse of public and private funds to support electoral efforts. It also raises questions about the ability of citizens to vote their conscience and hold politicians accountable in the privacy of the voting booth.

Keywords: clientelism, Latin America, vote buying, political behavior, brokers, Latin American politics

Introduction

Scholars of Latin America are at the forefront of research on clientelism in comparative politics. This article reviews the relevant literature, focusing on long-standing debates as well as more recent theoretical and methodological innovations. The first section defines clientelism as the personalized and discretionary exchange of goods and favors for political support. The second section discusses the advantages and shortcomings of different strategies used to measure clientelism in Latin America. The section also reports survey evidence that reveals significant variation in the pervasiveness of the practice across the region. The third section discusses studies that look at targeting dynamics. In particular, it focuses on the debate between those who think party machines target swing voters for vote buying and those who argue that clientelism is primarily about mobilizing and maintaining the base. The fourth section analyzes commitment problems among parties and clients, and presents an overview of different theories put forward to explain why parties engage in clientelism despite their inability to directly enforce contracts. Finally, the last section explores the role of brokers in machine politics, commitment problems between politicians and brokers, and how types of brokers vary across contexts. The article concludes with a discussion of possible avenues for future research.

What Is Clientelism?

Clientelism is a type of linkage strategy that parties deploy to win elections (Kitschelt, 2000).1 We define it as the personalized and discretionary exchange of goods or favors for political support.2 The characterization of the clientelistic exchange as personalized or individualized (and often face-to-face) helps distinguish clientelism from other forms of distributive politics in which the exchange involves groups of voters. In pork-barrel politics, for example, everyone who lives in a certain area receives a benefit. Politicians target these benefits with the hope that the group as a whole will be more likely to vote for them. By contrast, in the case of clientelism, individuals who are not part of the exchange are excluded from the benefit. This implies that clientelism is individually targeted, and therefore patrons know who receives the benefit. Political brokers, or middlemen, nurture and sustain the personal connections that underpin these transactions. They do so over long periods of time, during and between electoral campaigns, and thus get to know their clients (Auyero, 2000, 2001; Calvo & Murillo, 2013; Diaz-Cayeros, Estevez, & Magaloni, 2016; Stokes, Dunning, Nazareno, & Brusco, 2013; Szwarcberg, 2015; Zarazaga, 2014).3 Crucially, these intermediaries make sure to emphasize the personal nature of the connection, publicly boasting of their “service to the people” and highlighting “their particular efforts to obtain the goods . . . thus creating the appearance that were they not there, the benefits would not be delivered” (Auyero, Lapegna, & Poma, 2009, p. 5). The fact that the goods distributed through clientelism are often private goods makes credit-claiming comparatively easier (Desposato, 2007). Credit-claiming, however, might be harder in certain cases, such as in the discretionary distribution of targeted government programs (Weitz-Shapiro, 2014). In those cases, patrons and brokers have to make an additional effort to convince clients of the personalized and discretionary character of the benefit.4

In clientelistic transactions, patrons provide goods or favors, and clients are expected to reciprocate with political support. The existing literature provides a wide variety of examples of the types of goods and favors that are distributed through clientelistic exchanges in Latin America: clothing, mattresses, medicine, milk, corrugated metal, construction materials, blankets, hangers, utility bill payments, money, eyeglasses, chickens, trees, magnets (Brusco, Nazareno, & Stokes, 2004), favors (Oliveros, 2016), and even alcohol and drugs (Szwarcberg, 2015) in Argentina; medicine, health exams, dentures, wheelchairs, orthopedic boots, and female sterilization in Brazil (Nichter, 2011); property titles, subsidized housing and food, work opportunities, licenses to sell merchandise in flea markets (Magaloni, Diaz-Cayeros, & Estévez, 2007), supermarket gift cards (Cantú, 2019), and access to water (Herrera, 2017) in Mexico; and furniture, animals, food, tools, and construction materials in Nicaragua (Gonzalez‐Ocantos, Kiewiet de Jonge, Meléndez, Osorio, & Nickerson, 2012). The type of good or favor therefore does not determine the existence of a clientelistic exchange.

Although all of the above examples are private goods, it is of course possible to distribute private goods through programmatic politics rather than through clientelism.5 In programmatic politics, a set of rules stipulates the conditions under which citizens should receive certain benefits, and all citizens who satisfy those requirements receive the benefit, without any bias or discretion. The group is defined in abstract terms—for instance, the poor, the unemployed—and all people thus defined receive the benefit. No member of the group can be excluded. More importantly, in programmatic politics there is no explicit expectation that recipients will reciprocate with political support. By contrast, patrons expect clients to engage in a number of activities, including turning out on Election Day, voting for a specific candidate, attending rallies, or helping out with the campaign. Moreover, the clientelistic exchange is discretionary because patrons (or brokers) enjoy considerable leeway in deciding who receives the benefit or can make clients believe they have such leeway. There are no formal rules governing the distribution of goods or favors. Brokers and politicians target benefits based on the perceived responsiveness of certain voters, the need to sway some voters in order to win, to maintain the electoral base, or for personal and career goals.

Most definitions of clientelism emphasize the quid pro quo nature of the exchange in a very specific way: the provision of the goods and services is contingent upon the actions of the client. For instance, for Stokes et al. (2013, p. 13), “the party offers material benefits only on the condition that the recipient returns the favor with a vote or other form of political support.” Similarly, Kitschelt and Wilkinson (2007) consider a clientelistic exchange to be made up of three components: “contingent direct exchange, predictability, and monitoring” (p. 9).6 From this perspective, two more conditions are necessary before an exchange can be classified as clientelistic. First, the patron must know, infer, or (at the very least) be able to make the client believe that it is possible to monitor political behavior. This may involve some mechanism that violates the secrecy of the ballot or that makes the client believe that this is a real possibility. In these cases, party machines with an army of politicians and brokers “deeply embedded in social networks” are required to select appropriate clients and monitor their behavior (Stokes, 2009, p. 14). Second, the client should believe that she could be punished if she reneges on her side of the agreement. In other words, for a clientelistic exchange to take place, the patron should be able to identify noncompliers and credibly commit to punish them.

Our understanding of clientelism is different. We do not assume that clientelistic exchanges require monitoring of specific voting decisions or political behavior, or that clients’ fear of punishment is their main reason for fulfilling their side of the agreement. Whether monitoring and punishment (or the threat of punishment) actually occur is variable, and is not a necessary feature of the phenomenon. Put simply, in a clientelistic exchange both sides expect to obtain something from the other, but whether and why the contract is respected, and patrons and clients actually obtain what they want, is highly contingent and ultimately an empirical question. There are several explanations for why clients fulfill their side of the clientelistic agreement in addition to beliefs about patrons’ capacity to monitor. Indeed, clients can make an active and conscious decision to support their patron because they think it is in their best interest to do so (Baldwin, 2013; Diaz-Cayeros et al., 2016; Oliveros, 2016, 2018; Zarazaga, 2014, 2015) or because they feel morally compelled to return favors (Finan & Schechter, 2012; Lawson & Greene, 2014). It should be noted, however, that the fact that clientelism is not always sustained by fear or coercion does not mean it is not usually colored by acute power and status asymmetries (Auyero, 2001). The exchange may well be mutually beneficial and clients may willingly enter into such transactions, but the relationship is deeply unbalanced because the stakes for those on the receiving end, often found among the very poor, are much higher.

How Pervasive Is Clientelism in Latin America?

Observing clientelistic exchanges and measuring their prevalence across countries is crucial to describing the bond between patrons and clients; testing propositions about who is targeted, when, and why; and understanding the rise, reproduction, and collapse of machine politics. In this sense, one of most vexing questions in clientelism research is how to document the existence and pervasiveness of an electoral strategy that is often illegal and “generally associated with a negative social stigma due to its disjuncture with democratic norms and association with poverty” (Gonzalez‐Ocantos et al., 2012, p. 203).

Latin Americanists have produced important innovations on this front. A few seminal studies deploy ethnographic methods to zoom in on the social relations that emerge around clientelistic transactions (e.g., Auyero, 2001; Gay, 1994; Szwarcberg, 2015). This approach allows scholars to build trust with informants. Brokers are thus willing to reveal stories that involve the distribution of goods and services for political gain. By the same token, clients feel comfortable admitting their participation in these exchanges and their dependence on discretionary state largesse for survival. The result is usually a rich and diachronic account of clientelistic transactions as they unfold during and between electoral campaigns in specific locales. But ethnographies are not designed to generate precise estimates of the pervasiveness of clientelism and its variation across theoretically relevant portions of the electorate. Rich description of clientelistic transactions in a few communities comes at the expense of studying representative samples of clients and brokers.

To address this shortcoming, scholars of clientelism in Latin America have turned to surveys that ask a sample of respondents about their experiences with clientelism. Researchers are thus able to present a more systematic depiction of the reach and characteristics of the clientelistic exchange. In sharp contrast to ethnographers, however, enumerators are only able to establish limited rapport with their interviewees. This increases the risk of untruthful responses as a result of social desirability bias. Indeed, studies show that asking people about stigmatized or illegal practices produces this kind of bias due to individuals’ propensity to depict themselves in a positive light to avoid sanctions (Bradburn, Sudman, Blair, & Stocking, 1978; DeMaio, 1984; Nadeau & Niemi, 1995; Tourangeau & Yan, 2007). In the case of clientelism, levels of social stigma attached to the practice are high, to the point that respondents tend to think it is unlikely that others will willingly admit to partaking in these exchanges (Gonzalez Ocantos, Kiewiet de Jonge, & Nickerson, 2014; Kiewiet de Jonge, 2015).

Brusco et al. (2004) and Stokes (2005) were among the first to recognize the problem of social desirability bias in this type of research. In a survey conducted in three Argentinean provinces in 2001–2002, they asked respondents whether they had received a gift or favor from a candidate or a party during the campaign. Only 7% of the sample provided a positive response. But they also asked whether parties had distributed goods to individuals in their neighborhoods during the campaign. A question about the behavior of others is more likely to lead to honest answers. In their study, 44% reported that parties distributed handouts in their neighborhood and 35% could name the items that were distributed and the parties that gave them out. For the authors, the individual and neighborhood estimates reflect the possible range of the extent of clientelism. While the self-reported measure likely underestimates clientelism, the neighborhood measure indicates an upper bound.

Due to the possibility of social desirability bias, surveys of clientelism in Latin America and beyond now tend to include these two measures. Scholars often find similar differences between them and assume that the “true” level of clientelism lies somewhere in between. The Latin American Public Opinion Project’s (LAPOP) AmericasBarometer, for instance, includes both types of questions, allowing us to compare the extent of the phenomenon across countries.7 The 2014 survey asked:

  • Thinking of the last national elections, did any candidate or political party offer a favor, gift, or other benefit to a person you know in exchange for that person’s support or vote?

  • And thinking about the last presidential elections of [year varies by country], did someone offer you something, like a favor, gift, or any other benefit in return for your vote or support?

Figure 1 presents the responses to these questions in 19 Latin American countries, and shows that there is significant variation in the pervasiveness of clientelism.8

Clientelism in Latin American PoliticsClick to view larger

Figure 1. Percentage of respondents reporting clientelism across 19 Latin American Countries.

Notes: DN/NA were coded as missing. Horizontal bars represent 95% confidence intervals.

Source: Data from the 2014 AmericasBarometer by the Latin American Public Opinion Project (LAPOP).

According to the self-reported measure, the Dominican Republic is the country with the highest prevalence of clientelism (24.4%), followed by Honduras (18.1%) and Mexico (15.1%). According to the indirect measure, the Dominican Republic is also the most clientelistic country with 31.5% of positive responses, followed by Mexico (22.7%), Haiti (22.3%), Honduras (21.7%), Brazil (21.6%), and Paraguay (21.6%). Note that the difference between the direct and indirect measures also varies across countries. For instance, in Uruguay 3.4% of respondents reported being personally targeted, while more than twice as many (10%) reported seeing others participate in these exchanges. Similarly, in El Salvador, 4% of respondents reported personal clientelism, but approximately three times more (11.4%) reported offers to others. By contrast, in Ecuador (8.8% personal, 9.9% others) and Bolivia (5.4% personal, 5.7% others) there are no significant differences between the measures.

While this empirical strategy is useful and has led to interesting national and cross-national research, it is not without problems. First, the direct and indirect estimates could be measuring different things. For example, when asked about clientelism in their neighborhood, respondents could weaponize their answers to sully the reputation of certain parties or voters. More important, however, is the fact that the indirect measure still remains the key instrument to test propositions about the impact of individual-level characteristics (e.g., partisanship, income, voting behavior) on brokers’ targeting decisions and the workings of clientelism more broadly. But since this measure is likely tainted by social desirability bias, and bias is likely to be nonrandom across theoretically relevant subgroups of voters, “empirical results about the dynamics of vote buying derived from surveys are on a shaky foundation” (Gonzalez‐Ocantos et al., 2012, p. 202).9

To address this issue, Gonzalez-Ocantos et al. (2012) introduced list experiments to the study of clientelism in Latin America. List experiments provide an unobtrusive alternative to direct questions and produce more reliable estimates of the pervasiveness of the practice. The technique allows respondents to reveal that they took part in these exchanges without directly admitting to it. The survey sample is divided into control and treatment groups. All participants are asked to count how many of a series of campaign activities mentioned in a list they experienced during a recent election. Crucially, the list in the treatment condition has an additional item, usually phrased as follows: “A party offered you a gift or favor in exchange for your vote.” Since respondents in both experimental groups are on average the same, any difference in the mean number of activities reported by each group has to be the result of including this additional item as part of the treatment list, with no difference indicating the absence of clientelism. While researchers cannot know which individuals were targeted, they can estimate the pervasiveness of clientelism across theoretically relevant subgroups.10 Moreover, it is possible to use list experiments to investigate the strength of social desirability bias in different countries. In this sense, differences between the direct measure and the list experiment estimate reflect the presence of bias and show that previous research based on direct questions might be more or less reliable depending on the context. Figure 2 shows the result of a series of list experiments fielded in Latin America along with the estimates from direct questions.11

Clientelism in Latin American Politics

Figure 2. Comparison of percentage of respondents reporting clientelism with the direct estimate and the list experiment estimate across selected countries.

Notes: DN/NA were coded as missing. Horizontal bars represent 95% confidence intervals.

Sources: Argentina (Oliveros, 2019); Bolivia, Chile, Guatemala, Mexico, and Uruguay (Kiewiet de Jonge, 2015), El Salvador (Gonzalez-Ocantos, Kiewiet de Jonge, & Meseguer, 2018), Honduras (Gonzalez-Ocantos et al., 2015), Nicaragua (Gonzalez‐Ocantos et al., 2012).

Figure 2 shows that the degree of clientelism and social desirability bias vary significantly across countries. Indeed, in some cases the self-reported measure can be extremely misleading. Social desirability bias is greatest in the cases of Argentina, Honduras, and Nicaragua. For instance, in Nicaragua the direct estimate of clientelism is 2.4% while the estimate from the list experiment is 24.3%. Similarly, in Honduras the direct estimate is 3.5% while the list experiment estimate is 20.1%. By contrast, in the cases of Chile, Bolivia, and Uruguay, the list experiment estimates and the direct estimates are fairly similar, suggesting lower levels of social desirability bias.12

Differences in measurement strategies aside, large survey projects that gauge the extent of clientelism across Latin America share two important shortcomings. First, the timing of surveys is far from systematic. The list experiments reported in Figure 2 were not conducted at the same stage of the electoral process (some were conducted before Election Day, others after the election, and yet others in between rounds). Similarly, the AmericasBarometer surveys are not scheduled to coincide with elections. In fact, they are often timed to avoid election cycles. Consequently, phenomena that are particularly salient around elections, such as clientelism, may be underestimated (Lupu, Oliveros, & Schiumerini, 2019). Academic surveys conducted around elections and applied regularly across cycles are still rare in Latin America.13

Timing matters. When dealing with issues of electoral integrity, such as clientelism, there are good reasons to suspect that measures taken before the election capture expectations, while measures taken after the election capture both the experience of the election as well as voters’ feelings of approval or disappointment with the outcome (Anderson, Blais, Bowler, Donovan, & Listhaug, 2005; Cantú & García-Ponce, 2015). In Mexico, for instance, Cantú and García-Ponce (2015) draw on data collected over the course of the 2012 presidential campaign and find that confidence in the integrity of the electoral process varies over time and across partisan groups. In particular, they find that supporters of the incumbent party lose confidence in the electoral process only after their candidate is defeated. Similarly, using panel data from the 2015 Argentinean election, Oliveros (2019) finds that supporters of the incumbent party were more likely to believe that voting was secret before the election but not after their candidate lost.14 These differences could affect voters’ willingness to report vote buying. In this sense, recent panel studies fielded in Brazil (Ames et al., 2014, 2010), Mexico (Lawson, 2001, 2007), and Argentina (Lupu, Gervasoni, Oliveros, & Schiumerini, 2015), offer great opportunities to explore how reporting of clientelism, as well as targeting dynamics, vary during the electoral cycle.

The second shortcoming has to do with coverage. Little attention has been paid to where our measures of clientelism are taken. A lot of what we have learned from survey research about clientelism in Latin America is based on surveys that are supposed to be representative at the national level, but often exclude slum dwellers (Murillo, Oliveros, & Zarazaga, 2019). Sampling procedures require information from census maps and census data, but slums are rarely mapped or surveyed by government agencies (Auerbach, LeBas, Post, & Weitz-Shapiro, 2018). Moreover, for security and logistical reasons, polling firms in developing countries often substitute some sampling units for others that are more accessible or less dangerous (Lupu & Michelitch, 2018). Of course, omitting slum dwellers in studies of clientelism is particularly problematic since we know from ethnographic studies that slums are often the main sites of clientelism. Indeed, Murillo et al. (2019) call attention to this issue and show that not only is there more clientelism among Argentine slums dwellers but also that their understanding of the mechanisms of the clientelistic exchange is different from (equally poor) nonslum dwellers.

Who Are the Clients?

Scholars of clientelism in Latin America and beyond generally agree that poor voters are disproportionally targeted with clientelistic offers. For some scholars, poor voters are targeted by parties because of the higher marginal value of relatively inexpensive gifts for poorer voters. For others, they are targeted because the poor are more likely to value immediate assistance rather than programmatic promises (e.g., Calvo & Murillo, 2004; Kitschelt, 2000; Kitschelt & Wilkinson, 2007; Luna, 2014; Mares & Young, 2016; Stokes et al., 2013; Weitz-Shapiro, 2014).15 Beyond this general agreement, much of the debate around the dynamics of clientelism focuses on the types of clients favored by party machines. Do political parties target core supporters or swing voters? Relying on a formal model and survey data from Argentina, Stokes (2005) jump-started this debate in Latin America arguing that rational parties disproportionally target weak supporters and swing voters. Resources are wasted if parties invest in core supporters, who presumably will vote for the party regardless of whether or not they are rewarded with material incentives.16 Using the same survey data, Nichter (2008) challenged this proposition. First, he is more skeptical of parties’ ability to monitor the vote, which is needed when targeting nonpartisans. Second, he introduces a second dimension in the targeting calculus: voter’s turnout record. He concludes that rational political parties will use clientelism to buy turnout among voters whom they can trust to vote as expected (i.e., core supporters), but whose propensity to vote is low and therefore need additional incentives to mobilize.

Subsequent research led to an emerging consensus around the view that clientelism is mainly used to mobilize and maintain the base. Most notably, Stokes et al. (2013) analyze individual-level data from Argentina, Venezuela, Mexico, and India, and find that swing voters are targeted at very low rates. To explain why, they develop a theory of distribution that takes into account the preferences of brokers, who mediate between party bosses and voters. Party bosses might prefer to target swing voters to increase the margin of victory, but brokers face incentives to design different distribution strategies that allow them to pocket some of the resources fed into the machine. The fact that party bosses can only imperfectly monitor broker behavior means that brokers can get away with not being perfect agents of their principals. As Szwarcberg’s (2015) study of Argentina shows, bosses are only able to evaluate brokers’ performance by, for example, observing how many people they bring to rallies or mobilize to the polls, but cannot know with certainty whether these voters vote as instructed. As a result, brokers have an incentive to target core supporters, who are easier and cheaper to mobilize (Zarazaga, 2014, 2016), thus signaling their effectiveness to the leader while saving resources for personal enrichment.

More generally, clientelistic exchanges are based on long-term interactions. Politicians therefore care about their reputation and keeping their supporters satisfied. Politicians who systematically break their promises to their loyal supporters will be unable to sustain their electoral base over time (Diaz-Cayeros et al., 2016; Zarazaga, 2014). Indeed, Diaz-Cayeros et al. (2016) find that Mexican parties target supporters in an effort to cement loyalties. Parties only use clientelism to attract other voters when their core base of support is not enough to win an election. Similarly, Gonzalez-Ocantos, Kiewiet de Jonge, and Meseguer (2018) argue that poor voters in El Salvador who receive remittances from family members living abroad experience an ideological shift to the right, which leads rightist parties to target them with vote buying to reinforce emerging loyalties. A study of Honduras’s 2009 presidential election shows that when parties are worried about turnout levels, they target vote buying disproportionately at core supporters to ensure high participation rates and thus legitimize the election (Gonzalez-Ocantos, Kiewiet de Jonge, & Nickerson, 2015). Finally, Calvo and Murillo (2013) show that citizens’ perception of the likelihood of being offered a handout increases among Argentineans (but not Chileans) who are more connected to partisan networks (those more likely to be core supporters). These findings point to a different logic than the one in Stokes et al.’s (2013) theory of broker-mediated distribution. Far from being a distortion that brokers introduce to the distribution plan, targeting core supporters is often what party leaders want.

Moving the debate in a different direction, two recent papers have called attention to other characteristics of clients that matter beyond their poverty and their core or swing status. According to Holland and Palmer-Rubin (2015, p. 1187), “in many low- and middle-income democracies, organizational membership, not poverty or partisan activity, is the strongest predictor of exposure to vote buying.” Using data from the Latin American Public Opinion Project’s (LAPOP) AmericasBarometer, they show that participation in civic associations (e.g., religious meetings, school board meetings, and community improvement committee meetings, among others) is a more important predictor of clientelism than income. In their view, which they support with qualitative evidence from Colombia and Mexico, this association is explained by the “brokerage function” of some organization leaders. In an argument that also calls attention to the social networks of voters, Schaffer and Baker (2015) argue that parties target influential individuals who have access to large numbers of voters. Using the LAPOP data as well as panel data from Mexico, they show that voters who engage in “frequent political persuasion” and who are located in “large political discussion networks” are disproportionally targeted by clientelistic offers. Parties thus take advantage of the multiplier effect of “persuading” one well-connected individual. Finally, Carlin and Moseley (2015) suggest that political machines in Argentina tend to avoid citizens with strong democratic values, preferring those who are more ambivalent to or opposed to democracy. Clientelism carries a heavy stigma, especially among those who are firmly committed to democracy. These voters are likely to see “vote buying as a direct affront to their political belief system” (Carlin & Moseley, 2015, p. 15).

Overall, the only consensus in the targeting literature seems to be around the proposition that clientelism mainly affects the poor, and increasingly so, around the idea that parties focus on core rather than swing voters. But scholars have also identified a variety of additional factors that affect distribution dynamics. Some of these claims are based on single case studies, and therefore cannot be generalized. In fact, recent work suggests there may be limited room for generalizations. For example, Gans-Morse, Mazzuca, and Nichter (2014) draw attention to a series of institutional and political factors that likely lead to variation in targeting strategies across contexts.

Enforcing Clientelistic Deals

Regardless of targeting strategies, clientelistic deals are in principle hard to enforce. First, they are sequenced. Parties distribute benefits during the campaign, or in between elections, and voters cast their ballot at a later date. Second, because these exchanges are illegal and clandestine, the party cannot rely on the law to punish the voter if she fails to comply. More fundamentally, where the ballot is secret, parties cannot know if the voter complied with her end of the deal. Political scientists tend to assume that parties are rational actors that spend campaign resources strategically in an effort to maximize electoral gains. So a logical question follows: why do parties spend resources buying votes if these deals cannot be enforced? In other words, given the commitment problems inherent to clientelism, why embrace such levels of uncertainty? Indeed, anti–vote-buying campaigns often make voters aware of the unenforceable nature of the deal, calling on people to “receive with one hand and vote with the other” (Szwarcberg, 2001).

One possible answer put forward by scholars of Latin America is that clientelism works as a result of voters’ fear of getting cut off if they fail to comply (e.g., Brusco et al., 2004; Stokes, 2005; Stokes et al., 2013; Weitz-Shapiro, 2014). Avoiding defection in the polling booth therefore requires parties to either devise monitoring strategies or persuade voters that their private actions can be observed. To do so candidates recruit brokers who work hard to minimize commitment problems by getting to know their clients and issuing credible threats. This is why Stokes (2005) refers to vote buying as a “perverse” form of accountability: parties check on voters, not the other way around. Brokers are sometimes able to devise sophisticated techniques to monitor the vote. For instance, in Mexico, there is anecdotal evidence that some brokers force voters to use mobile phones to record their behavior inside the polling station. Alternatively, brokers can try to convince their clients that the ballot is not secret. For example, sustained interactions between voters and brokers in Argentine municipalities allow Peronist operatives to gather valuable information that helps enforce contracts. Intimacy and local knowledge makes it possible to punish noncompliers. As one broker told Stokes (2005), “you know if a neighbor voted against your party if he can’t look you in the eye” (p. 317). These might be empty threats, but voters, especially poor ones, may conclude that it is simply too risky to ignore them.17 Brokers may also use activities they can readily observe, such as turnout (Nichter, 2008), attendance at rallies (Szwarcberg, 2015, pp. 63–70), or participation in local institutions (Gonzalez-Ocantos et al., 2012), to test their clients’ reliability and loyalty. Finally, some studies argue that clientelism is not based on monitoring, punishment, or fear at the individual level, but on the collective monitoring of small groups (Gingerich & Medina, 2013; Rueda, 2015, 2017).

For a second group of scholars, commitment problems are not that hard to solve. For some, this is because clientelism generates a sense of gratitude that makes the promise of compliance binding from the voter’s perspective (Graziano, 1976). In other words, these transactions are sustained by norms of reciprocity. While voters could in principle defect in the privacy of the polling station, they are compelled to comply by the moral underpinnings of gift giving and receiving. The repeated nature of the transactions leads to the emergence of shared cultural representations of vote buying, which obscure the inequalities inherent in such exchanges and reinforce recipients’ determination to preserve the culture of returning gifts and favors with votes. In his famous ethnography of an Argentine slum, Auyero (2001) depicts electoral machines as “problem solving networks,” and shows that poor voters appreciate the work they do. According to one client, brokers do not even have to ask for reciprocation: “Because she gave me medicine, or some milk, or a packet of yerba or sugar, I know that I have to go to her rally in order to fulfill my obligation to her, to show my gratitude” (Auyero, 2001, p. 160). In a similar vein, Gonzalez-Ocantos et al.’s (2014) study of five Latin American countries shows that while at the aggregate level vote buying might distort electoral democracy, at the individual level it is a mutually beneficial act, such that poor voters and those with strong reciprocity values tend to see these exchanges in a more positive light. Parties seem to be aware of the power of these norms. Studies conducted in Paraguay (Finan & Schechter, 2012) and Mexico (Lawson & Greene, 2014) show that brokers target reciprocal individuals because they expect them to be more compliant.

For others, it is not reciprocity but self-interest that solves commitment problems. Clientelistic transactions give voters access to resources they need, so they vote as expected because they want to guarantee a continuous flow of resources and not because they are forced to do so (Diaz-Cayeros et al., 2016; Oliveros, 2016, 2018; Zarazaga, 2014, 2015). Importantly, due to their economic insecurity, resource-poor constituencies are responsive to relatively inexpensive material rewards (Calvo & Murillo, 2004; Kitschelt, 2000; Stokes, 2007; Stokes et al., 2013). Their preferences for a given party or machine can therefore become endogenous to the exchange itself. In this sense, Diaz-Cayeros et al. (2016) argue that party loyalty is conditional: “partisan attachments are constructed through reciprocal material and symbolic exchanges, past, present, and future” (p. 10). If the flow of resources stops, voters will explore other options. In many ways, this argument is diametrically opposed to explanations that emphasize monitoring: it highlights the agency of the voter rather than that of the machine.

Finally, recent work that focuses on short-term clientelism argues that under certain conditions parties do not care about commitment issues at all, and often distribute gifts knowing full well that their ability to enforce deals is nonexistent. For example, Muñoz (2014, 2018) shows that this is the case in Peru, where parties are weak and do not have permanent or deeply embedded armies of brokers. The purpose in this case is not to buy votes directly, but to boost attendance at rallies. Well-attended rallies allow parties to engage voters and stage their mobilizing potential.

Principal-Agent Problems Between Party Bosses and Brokers

Most of the literature on clientelism that deals with commitment issues focuses on the relationship between voters and patrons/brokers. Some scholars, however, now study the commitment problems that can arise in the relationship between patrons (or party leaders) and brokers. Until very recently, scholars depicted machines as unitary actors, assuming that the goal of brokers and patrons was one and the same—electoral success. Recent studies, however, tell a more complicated story.

First, the prospect of electoral success may not be enough for party bosses to extract effort from brokers. According to Camp (2017), this is because of a collective action problem: “individual marginal efforts do not significantly impact electoral outcomes, and they [brokers] are inclined to free-ride on the efforts of other brokers” (p. 522). Second, brokers tend to have more information about voters than their bosses do. This informational asymmetry, coupled with party leaders’ imperfect ability to monitor broker behavior, incentivizes brokers to “misuse” resources by distributing gifts to core voters, rather than to swing voters, or to keep part of those resources to themselves (Stokes et al., 2013; Szwarcberg, 2015). “Disloyal” brokers in weak party systems can also switch parties (Novaes, 2018). Stokes et al. (2013) argue that these principal-agent problems, and the costs associated with them, can encourage political parties to abandon clientelistic practices altogether.

To solve these commitment problems and create incentives for brokers to deliver, political parties monitor brokers’ performance and offer them different types of private rewards. For instance, relying on extensive fieldwork and a data set tracing the political careers of local brokers in Argentina, Szwarcberg (2012, 2015) shows how bosses combine information from voter turnout at rallies and elections to reward and punish brokers’ efforts. Similarly, taking advantage of variation in the number of polling stations in Mexican electoral precincts, Larreguy, Marshall, and Querubin (2016) show that bosses’ greater monitoring capacity in smaller precincts increases turnout and votes for the Partido Acción Nacional (PAN) and the Partido Revolucionario Institucional (PRI).

Commitment problems, however, only arise when the electoral success of the patron is not enough to incentivize the broker to work for the machine. When their electoral interests are not aligned with those of the party that hires them, brokers might have incentives to shirk. But under certain conditions, incentives can be aligned, minimizing agency loss. This alignment might be ideological. For instance, Larreguy, Montiel Olea, and Querubin (2017) show that Mexican teachers affiliated with the Mexican National Educational Workers Union (SNTE) are effective brokers only when mobilizing voters in support of their preferred party. Another possibility for aligned incentives is when brokers are public employees. Oliveros (2018) finds that Argentine brokers with patronage jobs believe that their jobs might be in jeopardy if the opposition wins, which provides a strong incentive to work hard to keep the incumbent in office.

Most contemporary studies of clientelism recognize the importance of public sector jobs as a way of financing the work of political brokers (Mares & Young, 2016). For instance, in what is probably the biggest survey of brokers conducted to date (800 interviews in four Argentinean provinces), 30% of respondents (excluding those who held public office) reported being public sector employees. In fact, public employment was the largest single occupation in the sample (Stokes et al., 2013). Similarly, in a 2003 survey of 112 Peronist grassroots organizations in Greater Buenos Aires and the City of Buenos Aires in Argentina, Levitsky (2003, pp. 195–197) found that two-thirds were run by brokers who worked in the public administration. In one-third of them, at least two other activists also held public sector jobs.18 Similarly, public school teachers broker votes in Mexico (Larreguy et al., 2017; Chambers-Ju, 2017) and Colombia (Eaton & Chambers-Ju, 2014).

Brokers with public sector jobs tend to be party brokers. They are committed to one party and have no ties to other interest groups. These are the types of brokers that have received most scholarly attention in Latin America. Recent research, however, shows that there are other types of brokers. Holland and Palmer-Rubin (2015) propose a typology of brokers based on their relationships to parties and interest groups. Party brokers are those who mobilize voters for a single party and are not embedded in any organization, such as most Peronist Party operatives in Argentina (e.g., Auyero, 2001; Levitsky, 2003; Stokes et al., 2013; Szwarcberg, 2015; Zarazaga, 2014). Organizational brokers, in turn, are those embedded in an organization who have weak (or no) party ties. Leaders of street vending associations in Colombia, who negotiate blocs of votes with multiple parties in each election, are a good example (Holland & Palmer-Rubin, 2015, pp. 1205–1208). The third category, hybrid brokers, consists of brokers who have ties both to a party and an organization. An example can be found in Oliveros’s (2016) study of Argentina. In addition to working for the local administration, a young broker described her activities as the head of a neighborhood association that organizes social and cultural events, runs a soup kitchen, and manages food assistance programs. The broker explained that she is very interested in politics because: “all the contacts [you need] to bring things to the neighborhood are political contacts.” (p. 385) Mexican teachers affiliated with the SNTE, a longtime ally of the PRI, are another case of hybrid brokers (Larreguy et al., 2017). Finally, the fourth category, independent brokers, includes brokers who are not embedded in interest organizations and have weak party commitments. Such brokers tend to “sell” their mobilization capacities to different parties in each electoral cycle. They are common in Peru, a country characterized by weak parties and party fluidity (Muñoz, 2014, 2018).19

Conclusion

By way of conclusion we would like to point to a few areas where more research is needed. First, while most of the literature studies contexts where clientelism is an ongoing relationship, we still know little about how the practice varies over time. Specifically, there are reasons to suspect that the types of goods that brokers distribute, the prevalence of the practice, and the profile of the targets, vary across the duration of the electoral cycle. For example, the mix between core and swing targets might shift as Election Day approaches. Indeed, Greene (2016) uses panel data from Mexico to show that effective campaigns can affect voters preferences, converting pre-campaign swing voters into post-campaign opposition or loyal voters. At the same time, there are good reasons to believe that the reporting of clientelism might also vary throughout the campaign (Cantú & García-Ponce, 2015; Oliveros, 2019). Ethnographies and panel surveys can contribute a great deal toward our understanding of these temporal dynamics. In addition, longitudinal studies can shed light on the extent to which partisan preferences or turnout propensities, which are key determinants of brokers’ targeting choices in most models of gift distribution, can indeed be treated as exogenous or are actually the product of sustained clientelistic interactions.

Second, most of the literature discussed here sees the prevalence and dynamics of clientelism as a function of the availability and types of brokers and, most crucially, of client characteristics, including level of material need, partisanship, and turnout propensities. But there are other factors that likely affect clientelism. In particular, clientelism is often just one element in the nonprogrammatic toolkit and parties tend to diversify campaign portfolios (Calvo & Murillo, 2019; Diaz-Cayeros et al., 2016). How do the prevalence and dynamics of clientelism change when we take into account the use of other strategies? We know, for example, that electoral intimidation is an effective, and much cheaper, form of demobilization, which on Election Day may prove the mathematical equivalent of buying votes (Collier & Vicente, 2014; Gonzalez-Ocantos et al., forthcoming). In this sense, the availability of violence infrastructures may reduce brokers’ incentives to engage in clientelism because prioritizing violence can free resources for personal enrichment. But party bosses presumably want voters to like them and not to fear them, and prefer brokers to use carrots rather than sticks. An interesting question is therefore how bosses varying monitoring capacities affect the mix of clientelistic and intimidatory tactics that brokers choose in such environments. Exploring how clientelism works amid violence is particularly relevant given the worsening security situation in many Latin American countries, as well as processes of democratic erosion across the region that lead to a rise in repression.

Third, recent developments in the field take seriously the agency of clients vis-à-vis the machine. This is an especially promising area for future inquiry. Most of the literature on clientelism has focused on elite-level strategies and “usually portrays clients as passive, myopic, nonstrategic, or mainly driven by short-term reactions to the actions of politicians” (Oliveros, 2018, p. 254). However, scholars recently have called attention to the fact that clients are often the ones who initiate the clientelistic exchange (e.g., Nichter, 2018; Nichter & Peress, 2017; Oliveros, 2016) and that they change loyalties if the party fails to deliver as promised (e.g., Diaz-Cayeros et al., 2016; Zarazaga, 2014). We still know very little about the incentive structure that clients face when making sophisticated, strategic choices. How do clients choose between brokers when there is more than one option? Under what conditions do clients decide to renege on their side of the deal? Do clients “take with one hand and vote with the other”? We also know little about how clients perceive clientelistic transactions, which is important to understand what they think they are getting from the machine. Few studies systematically explore how clients and the broader public evaluate clientelism from a normative point of view (Gonzalez-Ocantos et al., 2014; Weitz-Shapiro, 2014). Such research would provide understanding of whether anti–vote-buying campaigns resonate with the public and what can be done to curb clientelism. We know that mass tolerance for corruption helps lock in corrupt equilibria (Pavao, 2018). We must also determine if this is also true for clientelism or if clientelism can persist amid high levels of negative stigma.

References

Ames, Barry, Huberts, Alyssa, Machado, Fabiana, Rennó, Lúcio R., Samuels, David, Smith, Amy Erica, & Zucco, César. (2014). Brazilian electoral panel study: 2014 results (Working Paper). Washington, DC: IADB.Find this resource:

Ames, Barry, Machado, Fabiana, Rennó, Lúcio R., Samuels, David, Smith, Amy Erica, & Zucco, César. (2010). Brazilian electoral panel studies (BEPS): Brazilian public opinion in the 2010 presidential elections (Working Paper). Washington, DC: IADB.Find this resource:

Anderson, Christopher, Blais, André, Bowler, Shaun, Donovan, Todd, & Listhaug, Ola. (2005). Losers’ consent: Elections and democratic legitimacy. Oxford, U.K.: Oxford University Press.Find this resource:

Auerbach, Adam M., LeBas, Adrienne, Post, Alison E., & Weitz-Shapiro, Rebecca. (2018). State, society, and informality in cities of the Global South. Studies in Comparative International Development, 53, 261–88.Find this resource:

Auyero, Javier. (2000). The logic of clientelism in Argentina: An ethnographic account. Latin American Research Review, 35(3), 55–81.Find this resource:

Auyero, Javier. (2001). Poor people’s politics: Peronist survival networks and the legacy of Evita. Durham, NC: Duke University Press.Find this resource:

Auyero, Javier, Lapegna, Pablo, & Page Poma, Fernanda. (2009). Patronage politics and contentious collective action: A recursive relationship. Latin American Politics and Society, 51(3), 1–31.Find this resource:

Bahamonde, Héctor. (2018). Aiming right at you: Group versus individual clientelistic targeting in Brazil. Journal of Politics in Latin America, 10(2), 41–76.Find this resource:

Baldwin, Kate. (2013). Why vote with the chief? Political connections and public goods provision in Zambia. American Journal of Political Science, 57(4), 794–809.Find this resource:

Blair, Graeme, & Imai, Kosuke. (2012). Statistical analysis of list experiments. Political Analysis, 20(1), 47–77.Find this resource:

Bradburn, Norman M., Sudman, Seymour, Blair, Ed, & Stocking, Carol. (1978). Question threat and response bias. Public Opinion Quarterly, 42(2), 221–234.Find this resource:

Brusco, Valeria, Nazareno, Marcelo, & Stokes, Susan Carol. (2004). Vote buying in Argentina. Latin American Research Review, 39(2), 66–88.Find this resource:

Calvo, Ernesto, & Murillo, María Victoria. (2004). Who delivers? Partisan clients in the Argentine electoral market. American Journal of Political Science, 48(4), 742–757.Find this resource:

Calvo, Ernesto, & Murillo, María Victoria. (2013). When parties meet voters assessing political linkages through partisan networks and distributive expectations in Argentina and Chile. Comparative Political Studies, 46(7), 851–882.Find this resource:

Calvo, Ernesto, & Murillo, María Victoria. (2019). Non‐policy politics: Richer voter, poorer voter, and the diversification of parties’ electoral offers. New York, NY: Cambridge University Press.Find this resource:

Camp, Edwin. (2017). Cultivating effective brokers: A party leader’s dilemma. British Journal of Political Science, 47(3), 521–543.Find this resource:

Cantú, Francisco. (2019). Groceries for votes: The electoral returns of vote-buying. Journal of Politics, 81(3), 790–804.Find this resource:

Cantú, Francisco, & García-Ponce, Omar. (2015). Partisan losers’ effects: Perceptions of electoral integrity in Mexico. Electoral Studies, 39, 1–14.Find this resource:

Carlin, Ryan E., & Moseley, Mason. (2015). Good Democrats, bad targets: Democratic values and clientelistic vote buying. Journal of Politics, 77(1), 14–26.Find this resource:

Chambers-Ju, Christopher. (2017). Protest or politics? Varieties of teacher representation in Latin America. Berkeley, CA: University of California Berkeley.Find this resource:

Chandra, Kanchan. (2007). Counting heads: A theory of voter and elite behavior in patronage democracies. In Herbert Kitschelt, & Steve Wilkinson (Eds.) Patrons, clients, and policies: Patterns of democratic accountability and political competition, New York, NY: Cambridge University Press (pp. 84–109).Find this resource:

Collier, Paul, & Vicente, Pedro C. (2014). Votes and violence: Evidence from a field experiment in Nigeria. Economic Journal, 124(574), F327–F355.Find this resource:

Corstange, Daniel. (2009). Sensitive questions, truthful answers? Modeling the list experiment with LISTIT. Political Analysis, 17(1), 45–63.Find this resource:

De La O, Ana L. (2013). Do conditional cash transfers affect electoral behavior? Evidence from a randomized experiment in Mexico. American Journal of Political Science, 57(1), 1–14.Find this resource:

De La O, Ana L. (2015). Crafting policies to end poverty in Latin America. New York, NY: Cambridge University Press.Find this resource:

DeMaio, Theresa J. (1984). Social desirability and survey. In Charles F. Turner & Elizabeth Martin (Eds.), Surveying subjective phenomena (pp. 257–282). Thousand Oaks, CA: SAGE.Find this resource:

Desposato, Scott W. (2007). How does vote buying shape the legislative arena? In Frederic Charles Schaffer (Ed.), Elections for sale: The causes and consequences of vote buying (pp. 101–122). Boulder, CO: Lynne Rienner.Find this resource:

Diaz-Cayeros, Alberto, Estevez, Federico, & Magaloni, Beatriz. (2016). The political logic of poverty relief: Electoral strategies and social policy in Mexico. Cambridge, U.K.: Cambridge University Press.Find this resource:

Eaton, Kent, & Chambers-Ju, Christopher. (2014). Teachers, mayors, and the transformation of clientelism in Colombia. In Diego Abente Brun & Larry Diamond (Eds.), Clientelism, social policy, and the quality of democracy (pp. 88–113). Baltimore, MD: John Hopkins University Press.Find this resource:

Finan, Frederico, & Schechter, Laura. (2012). Vote‐buying and reciprocity. Econometrica, 80(2), 863–881.Find this resource:

Gans-Morse, Jordan, Mazzuca, Sebastián, & Nichter, Simeon. (2014). Varieties of clientelism: Machine politics during elections. American Journal of Political Science, 58(2), 415–432.Find this resource:

Gay, Robert. (1994). Popular organization and democracy in Rio de Janeiro: A tale of two favelas. Philadelphia, PA: Temple University Press.Find this resource:

Gingerich, Daniel W., & Medina, Luis Fernando. (2013). The endurance and eclipse of the controlled vote: A formal model of vote brokerage under the secret ballot. Economics & Politics, 25(3), 453–480.Find this resource:

Glynn, Adam N. (2013). What can we learn with statistical truth serum? Design and analysis of the list experiment. Public Opinion Quarterly, 77(S1), 159–172.Find this resource:

Gonzalez‐Ocantos, Ezequiel, Kiewiet de Jonge, Chad, Meléndez, Carlos, Osorio, Javier, & Nickerson, David W. (2012). Vote buying and social desirability bias: Experimental evidence from Nicaragua. American Journal of Political Science, 56(1), 202–217.Find this resource:

Gonzalez‐Ocantos, Ezequiel, Kiewiet de Jonge, Chad, Meléndez, Carlos, Osorio, Javier, & Nickerson, David W. (forthcoming). Carrots and sticks: Experimental evidence of vote buying and voter intimidation in Guatemala. Journal of Peace Research.Find this resource:

Gonzalez-Ocantos, Ezequiel, Kiewiet de Jonge, Chad P., & Meseguer, Covadonga. (2018). Vote buying and remittances. Latin American Research Review, 53(4), 689–707Find this resource:

Gonzalez Ocantos, Ezequiel, Kiewiet de Jonge, Chad P., & Nickerson, David W. (2014). The conditionality of vote‐buying norms: Experimental evidence from Latin America. American Journal of Political Science, 58(1), 197–211.Find this resource:

Gonzalez-Ocantos, Ezequiel, Kiewiet de Jonge, Chad P., & Nickerson, David W. (2015). Legitimacy buying: The dynamics of clientelism in the face of legitimacy challenges. Comparative Political Studies, 48(9), 1127–1158.Find this resource:

Graziano, Luigi. (1976). A conceptual framework for the study of clientelistic behavior. European Journal of Political Research, 4(2), 149–174.Find this resource:

Greene, Kenneth. (2016). Why vote buying fails: Campaign effects and the elusive swing voter. Unpublished Manuscript. University of Texas at Austin.Find this resource:

Grzymala-Busse, Anna. (2007). Rebuilding Leviathan: Party competition and state exploitation in post-communist democracies. Cambridge, U.K.: Cambridge University Press.Find this resource:

Grzymala-Busse, Anna. (2008). Beyond clientelism incumbent state capture and state formation. Comparative Political Studies, 41(4–5), 638–673.Find this resource:

Herrera, Veronica. (2017). Water and politics: Clientelism and reform in urban Mexico. Ann Arbor: University of Michigan Press.Find this resource:

Hicken, Allen. (2011). Clientelism. Annual Review of Political Science, 14, 289–310.Find this resource:

Holland, Alisha C., & Palmer-Rubin, Brian. (2015). Beyond the machine: Clientelist brokers and interest organizations in Latin America. Comparative Political Studies, 48(9), 1186–1223.Find this resource:

Imai, Kosuke. (2011). Multivariate regression analysis for the item count technique. Journal of the American Statistical Association, 106(494), 407–416.Find this resource:

Kiewiet de Jonge, Chad P. (2015). Who lies about electoral gifts?Public Opinion Quarterly, 79(3), 710–739.Find this resource:

Kitschelt, Herbert. (2000). Linkages between citizens and politicians in democratic polities. Comparative Political Studies, 33(6–7), 845–879.Find this resource:

Kitschelt, Herbert, & Altamirano, Melina. (2015). Clientelism in Latin America: In The Latin American voter: Pursuing representation and accountability in challenging contexts (pp. 246–274). Ann Arbor: University of Michigan Press.Find this resource:

Kitschelt, Herbert, & Wilkinson, Steven. (2007). Citizen-politician linkages: An introduction. In Herbert Kitschelt & Steven Wilkinson (Eds.), Patrons, clients, and policies: Patterns of democratic accountability and political competition (pp. 1–49). Cambridge, U.K.: Cambridge University Press.Find this resource:

Larreguy, Horacio, Marshall, John, & Querubin, Pablo. (2016). Parties, brokers, and voter mobilization: How turnout buying depends upon the party’s capacity to monitor brokers. American Political Science Review, 110(1), 160–179.Find this resource:

Larreguy, Horacio, Montiel Olea, Cesar E., & Querubin, Pablo. (2017). Political brokers: Partisans or agents? Evidence from the Mexican Teachers’ Union. American Journal of Political Science, 61(4), 877–891.Find this resource:

Lawson, Chappell. (2001). Mexico 2000 Panel Study.Find this resource:

Lawson, Chappell. (2007). Mexico 2006 Panel Study.Find this resource:

Lawson, Chappell, & Greene, Kenneth F. (2014). Making clientelism work: How norms of reciprocity increase voter compliance. Comparative Politics, 47(1), 61–85.Find this resource:

Levitsky, Steven. (2003). Transforming labor-based parties in Latin America: Argentine Peronism in comparative perspective. Cambridge, U.K.: Cambridge University Press.Find this resource:

Luna, Juan Pablo. (2014). Segmented representation: Political party strategies in unequal democracies. Oxford, U.K.: Oxford University Press.Find this resource:

Lupu, Noam, Gervasoni, Carlos, Oliveros, Virginia, & Schiumerini, Luis. (2015). Argentine panel election study.Find this resource:

Lupu, Noam, & Michelitch, Kristin. (2018). Advances in survey methods for the developing world. Annual Review of Political Science, 21, 195–214.Find this resource:

Lupu, Noam, Oliveros, Virginia, & Schiumerini, Luis. (2019). Toward a theory of campaigns and voters in developing democracies. In Noam Lupu, Virginia Oliveros, & Luis Schiumerini (Eds.), Campaigns and voters in developing democracies: Argentina in comparative perspective (pp. 1–27). Ann Arbor: Michigan University Press.Find this resource:

Magaloni, Beatriz, Diaz-Cayeros, Alberto, & Estévez, Federico. (2007). Clientelism and portfolio diversification: A model of electoral investment with applications to Mexico. In Herbert Kitschelt & Steven Wilkinson (Eds.), Patrons, clients, and policies: Patterns of democratic accountability and political competition (pp. 182–205). Cambridge, U.K.: Cambridge University Press.Find this resource:

Mares, Isabela, & Young, Lauren. (2016). Buying, expropriating, and stealing votes. Annual Review of Political Science, 19, 267–88.Find this resource:

Medina, Luis Fernando, & Stokes, Susan. (2007). Monopoly and monitoring: An approach to political clientelism. In Herbert Kitschelt & Steven I. Wilkinson (Eds.), Patrons, clients, and policies: Patterns of democratic accountability and political competition (pp. 68–83). Cambridge, U.K.: Cambridge University Press.Find this resource:

Muñoz, Paula. (2014). An informational theory of campaign clientelism: The case of Peru. Comparative Politics, 47(1), 79–98.Find this resource:

Muñoz, Paula. (2018). Buying audiences: Clientelism and electoral campaigns when parties are weak. New York, NY: Cambridge University Press.Find this resource:

Murillo, María Victoria, Oliveros, Virginia, & Zarazaga, Rodrigo. (2019). The most vulnerable poor: Clientelism among slum dwellers. Unpublished Manuscript.Find this resource:

Nadeau, Richard, & Niemi, Richard G. (1995). Educated guesses: The process of answering factual knowledge questions in surveys. Public Opinion Quarterly, 59(3), 323–346.Find this resource:

Nichter, Simeon. (2008). Vote buying or turnout buying? Machine politics and the secret ballot. American Political Science Review, 102(1), 19–31.Find this resource:

Nichter, Simeon. (2011). Electoral clientelism or relational clientelism? Healthcare and sterilization in Brazil. Paper Presented at the American Political Science Conference Meeting, Seattle, Washington.Find this resource:

Nichter, Simeon. (2014). Conceptualizing vote buying. Electoral Studies, 35, 315–327.Find this resource:

Nichter, Simeon. (2018). Votes for survival: Relational clientelism in Latin America. New York, NY: Cambridge University Press.Find this resource:

Nichter, Simeon, & Peress, Michael. (2017). Request fulfilling: When citizens demand clientelist benefits. Comparative Political Studies, 50(8), 1086–1117.Find this resource:

Novaes, Lucas M. (2018). Disloyal brokers and weak parties. American Journal of Political Science, 62(1), 84–98.Find this resource:

Oliveros, Virginia. (2016). Making it personal: Clientelism, favors, and the personalization of public administration in Argentina. Comparative Politics, 48(3), 373–391.Find this resource:

Oliveros, Virginia. (2018). Patronage at work: Public jobs and political services in Argentina. Unpublished Manuscript. Tulane University.Find this resource:

Oliveros, Virginia. (2019). Perceptions of ballot integrity and clientelism. In Noam Lupu, Virginia Oliveros, & Luis Schiumerini (Eds.), Campaigns and voters in developing democracies: Argentina in comparative perspective (pp. 213–238). Ann Arbor: Michigan University Press.Find this resource:

Pavao, Nara. (2018). Corruption as the only option: The limits to electoral accountability. The Journal of Politics, 80(3), 996–1010.Find this resource:

Rueda, Miguel R. (2015). Buying votes with imperfect local knowledge and a secret ballot. Journal of Theoretical Politics, 27(3), 428–456.Find this resource:

Rueda, Miguel R. (2017). Small aggregates, big manipulation: Vote buying enforcement and collective monitoring. American Journal of Political Science, 61(1), 163–177.Find this resource:

Schaffer, Frederic Charles, & Schedler, Andres. (2007). What is vote buying? In Frederic Charles Schaffer (Ed.), Elections for sale: The causes and consequences of vote buying. Boulder, CO: Lynne Rienner.Find this resource:

Schaffer, Joby, & Baker, Andy. (2015). Clientelism as persuasion-buying: Evidence from Latin America. Comparative Political Studies, 48(9), 1093–1126.Find this resource:

Scherlis, Gerardo. (2013). The contours of party patronage in Argentina. Latin American Research Review, 48(3), 63–84.Find this resource:

Stokes, Susan C. (2005). Perverse accountability: A formal model of machine politics with evidence from Argentina. American Political Science Review, 99(3), 315–325.Find this resource:

Stokes, Susan C. (2007). Political clientelism. In Susan Stokes, & Carles Boix (Eds.), The Oxford Handbook of Comparative Politics, Vol. 4 (pp. 604–627). New York, NY: Oxford University Press.Find this resource:

Stokes, Susan C. (2009). Pork, by any other name . . . Building a conceptual scheme of distributive politics. Paper presented at the American Political Science Association Conference, Toronto, Canada.Find this resource:

Stokes, Susan C., Dunning, Thad, Nazareno, Marcelo, & Brusco, Valeria. (2013). Brokers, voters, and clientelism: The puzzle of distributive politics. New York, NY: Cambridge University Press.Find this resource:

Szwarcberg, Mariela. (2001). Feeding loyalties: An analysis of clientelism‑The case of the Manzaneras. Buenos Aires: Universidad Di Tella.Find this resource:

Szwarcberg, Mariela. (2012). Uncertainty, political clientelism, and voter turnout in Latin America: Why parties conduct rallies in Argentina. Comparative Politics, 45(1), 88–106.Find this resource:

Szwarcberg, Mariela. (2015). Mobilizing poor voters: Machine politics, clientelism, and social networks in Argentina. New York, NY: Cambridge University Press.Find this resource:

Tourangeau, Roger, & Yan, Ting. (2007). Sensitive questions in surveys. Psychological Bulletin, 133(5), 859–883.Find this resource:

Weitz-Shapiro, Rebecca. (2012). What wins votes: Why some politicians opt out of clientelism. American Journal of Political Science, 56(3), 568–583.Find this resource:

Weitz-Shapiro, Rebecca. (2014). Curbing clientelism in Argentina: Politics, poverty, and social policy. New York, NY: Cambridge University Press.Find this resource:

Zarazaga, Rodrigo. (2014). Brokers beyond clientelism: A new perspective through the Argentine case. Latin American Politics and Society, 56(3), 23–45.Find this resource:

Zarazaga, Rodrigo. (2015). Plugged in brokers: A model of vote-buying and access to resources. Journal of Applied Economics, 18(2), 369–390.Find this resource:

Zarazaga, Rodrigo. (2016). Party machines and voter-customized rewards strategies. Journal of Theoretical Politics, 28(4), 678–701.Find this resource:

Zucco, Cesar. (2013). When payouts pay off: Conditional Cash transfers and voting behavior in Brazil 2002–10. American Journal of Political Science, 57(4), 810–822.Find this resource:

Appendix

Table A1. Percentage of Respondents Reporting Clientelism Across 19 Latin American Countries

Person you know

You

N

Argentina

5.91%

2.35%

1488/1488

(0.61)

(0.39)

Bolivia

5.68%

5.45%

3010/3011

(0.42)

(0.41)

Brazil

21.68%

10.69%

1485/1487

(1.07)

(0.80)

Chile

2.84%

1.6%

1548/1562

(0.42)

(0.32)

Colombia

14.48%

7.81%

1478/1472

(0.92)

(0.70)

Costa Rica

2.57%

2.16%

1520/1528

(0.41)

(0.37)

Dominican Republic

31.53%

24.42%

1497/1507

(1.20)

(1.11)

Ecuador

9.91%

8.83%

1473/1473

(0.78)

(0.74)

El Salvador

11.39%

4.03%

1501/1512

(0.82)

(0.51)

Guatemala

14.38%

12.45%

1474/1486

(0.91)

(0.86)

Haiti

22.36%

9.73%

1288/1418

(1.16)

(0.79)

Honduras

21.74%

18.06%

1559/1556

(1.05)

(0.98)

Mexico

22.72%

15.08%

1510/1519

(1.08)

(0.92)

Nicaragua

9.11%

5.51%

1537/1543

(0.73)

(0.58)

Panama

8.59%

6.13%

1478/1484

(0.73)

(0.62)

Paraguay

21.6%

12.35%

1472/1482

(1.07)

(0.85)

Peru

9.1%

5.77%

1473/1473

(0.75)

(0.61)

Uruguay

9.99%

3.39%

1501/1505

(0.77)

(0.47)

Venezuela

4.05%

2.7%

1482/1483

(0.51)

(0.42)

Notes. DN/NA were coded as missing; Standard errors in parentheses.

Source: The 2014 AmericasBarometer by the Latin American Public Opinion Project (LAPOP).

Table A2. Comparison of Percentage of Respondents Reporting Clientelism With the Direct Estimate and the List Experiment Estimate Across Selected Countries

Direct Estimate

List Experiment Estimate

N

Source

Argentina 2015

1.75%

14.88%

1381/1325

Oliveros (2019)

(0.35)

(5.27)

Bolivia 2009

5.24%

7.61%

2099/2063

Kiewiet de Jonge (2015)

(0.49)

(4.55)

Chile 2009

5.62%

2.88%

1941/1882

Kiewiet de Jonge (2015)

(0.52)

(5.61)

El Salvador 2014

2.80%

9.40%

994/992

Gonzalez-Ocantos et al. (2018)

(0.70)

(6.80)

Guatemala 2011

3.78%

14%

503/500

Kiewiet de Jonge (2015)

(0.85)

(8.06)

Honduras 2009

3.50%

20.7%

1005/993

Gonzalez-Ocantos et al. (2015)

(0.60)

(6.80)

Mexico 2009

8.11%

17.11%

1246/1183

Kiewiet de Jonge (2015)

(0.77)

(5.01)

Nicaragua 2008

2.39%

24.34%

1003/995

Gonzalez-Ocantos et al. (2012)

(0.58)

(5.53)

Uruguay 2009

0.68%

-1.77%

880/857

Kiewiet de Jonge (2015)

(0.28)

(5.41)

Notes. DN/NA were coded as missing. Standard errors in parentheses. List experiments from Kiewiet de Jonge (2015) were calculated with the available data using two-sample t-test with unequal variance.

Notes:

(1.) This section largely draws on Oliveros (2018, chapter 1).

(2.) This definition follows Kitschelt and Wilkinson (2007), and Stokes (2007). For a distinction between clientelism and other forms of electoral strategies (including programmatic ones) see Stokes et al. (2013, pp. 6–18).

(3.) Nichter (2014) distinguishes “electoral clientelism,” in which the clientelistic exchange takes place around elections, from “relational clientelism,” in which the exchange is not restricted to the electoral period.

(4.) For a description of the strategies that politicians can use to create or increase voters’ perception of discretion in the implementation of policy, see Weitz-Shapiro (2014, chapter 2) and Oliveros (2018, chapter 5).

(5.) In fact, in advanced democracies most welfare-state policies are targeted yet programmatic (Stokes, 2009). See Weitz-Shapiro (2012, 2014) for an example of how the same program (a targeted food distribution program in Argentina) can be distributed in a clientelistic or non-clientelistic way. Recent studies of conditional cash transfer programs in Mexico (De La O, 2013, 2015; Diaz-Cayeros, et al., 2016) and Brazil (Zucco, 2013), which show that distribution is in fact programmatic, provide further evidence that the type of good in itself does not necessarily imply clientelism.

(6.) For a similar perspective, see Grzymala-Busse (2007, 2008), Hicken (2011), Kitschelt and Altamirano (2015), Medina and Stokes (2007), Magaloni et al. (2007), Nichter (2008, 2014), Stokes (2007, 2009), and Weitz- Shapiro (2012, 2014).

(7.) We thank the LAPOP and its major supporters (the United States Agency for International Development, the Inter-American Development Bank, and Vanderbilt University) for making the data available.

(8.) Table A1 in the Appendix presents the numeric values.

(9.) For example, the indirect estimates lead to the conclusion that Uruguay has more vote buying than Argentina. Qualitative studies tell us this is highly unlikely.

(10.) Although there have been some developments on the use of list experiments with multivariate regression analysis (see, for instance, Blair & Imai, 2012; Corstange, 2009; Glynn, 2013; Imai, 2011), difference-in-means estimators have the advantage that no functional form assumption is needed.

(11.) Table A2 in the Appendix presents the numeric values. While it might be tempting to compare these figures with those reported by LAPOP, it is important to bear in mind significant differences in question wording and timing of the surveys. For example, LAPOP conducts its surveys during non-electoral periods, whereas most of these list experiments were fielded during campaigns.

(12.) For a detailed analysis of who lies about clientelism, see Kiewiet de Jonge (2015).

(13.) In Latin America, regular election studies are available only for Brazil, Chile, Mexico, and Peru (Lupu, Oliveros, & Schiumerini, 2019).

(14.) Note, however, that she does not find evidence of a winner-loser gap in the case of clientelism. Supporters of the incumbent party were not more likely to report clientelism after their candidate’s defeat (Oliveros, 2019).

(15.) For a discussion of the link between poverty and clientelism, see Stokes et al. (2013, chapter 6). Bahamonde (2018) adds an interesting twist to this debate, arguing that the salience of income as a determinant of clientelism is highly context dependent.

(16.) For a review of the literature on swing voter targeting, see Stokes et al. (2013, pp. 32–36).

(17.) See Chandra (2007, pp. 89–90), Kitschelt and Wilkinson (2007, pp. 15–18), and Schaffer and Schendler (2007, pp. 22–24), for examples of different techniques to monitor electoral preferences (or making voters believe that is possible).

(18.) On the use of public employment to finance the work of brokers and activists in Argentina, see also Auyero (2000, 2001), Oliveros (2016, 2018), Scherlis (2013), and Zarazaga (2014).

(19.) For a description of different types of brokers see also Mares and Young (2016).