Studying politics in Rio de Janeiro’s favelas (shantytowns) is an extremely challenging enterprise in light of the immense violence facing these communities. During my first research trip to Rio in 1996 few wanted to help me, an inexperienced and young researcher, gain access to any favela. After nearly a month of efforts I visited the favela of Rocinha under the auspices of an internship program run by the Pontifica Universidade Católica (PUC), an elite private college. On one of my early visits I attended an end-of-the-semester party at PUC’s center in the community where I met Roberto, a young politician campaigning for city council. When I lamented my difficulties gaining access to favelas he told me that “you just have to know the right people” and invited me along to visit several that evening. I accepted and in minutes we had left on a journey that would take us to three favelas.

We ended the night in Rio’s working class Zona Norte (North Side) at a favela called A Ladeira where a party was just beginning. Standing in the plaza I could smell grilling meats from the many booths selling beers, sodas, and food. A DJ warmed up the crowd playing charme, a melodic and usually romantic variant of Rio’s more staccato, violent, and sexual funk (a type of hip-hop). As a dancing contest began someone released a small hot air balloon into the clear starry sky.

I was unsure of exactly why we had come to this party except that Roberto had to meet with someone about gaining access to the community. Before going to the party we had visited another social gathering in a nearby housing project populated with residents the military had displaced from the Praia do Pinto favela in the 1970s (Perlman, 1976, pp. 258–260). There we met a middle aged woman who took us to A Ladiera.

After a long wait at the party Túlio, a young stylishly dressed man, came down to meet with us and he and Roberto went off to have a private conversation. I was left with the politician’s short bull-necked driver and a very large powerfully built man who had arrived with Túlio. While they talked, the three of us silently drank beer, listened to the music, and ate tough grilled meats.

Eventually Roberto and Túlio returned, everyone shook hands, and we said good-bye. As Roberto, his chauffer, and I drove away, Roberto, reveling in the illicitness of his contact, said that Túlio was the dono (literally owner, head trafficker) of the favela and that he had asked Roberto to help improve the main quadra (public square) in the community in exchange for exclusive access during the election period. Roberto said he refused the deal because he did not work with traffickers though I wondered why we had gone to the meeting at all if he had planned from the outset not to make a deal. Several months later Roberto lost his bid for the city council.

Drug trafficking and clientelism in Rio de Janeiro

Clientelism is alive and well in Rio de Janeiro. This story, however, highlights the changing nature of this practice in light of the growing power of drug traffickers in the city’s poor communities. Brokers, patrons, and clients still work together to distribute favors and organize votes but today drug traffickers have stepped into the picture creating a new level of exchange in this process. To get into favelas politicians must first make an arrangement with traffickers who then, usually through interlocutors, present the politicians to the community. While politicians always had to gain access to favelas through intermediaries, this new situation is clearly more intimidating and potentially compromising. The result, as I will show in this article, is that politicians tend to distribute favors to traffickers and traffickers then redistribute those favors to residents resulting in a two-tiered clientelism in which traffickers are free to negotiate patronage with a number of politicians but in which ordinary favela residents obtain political patronage through a single trafficker who dominates the favela they live in partially through force. This decreases the ability of favela residents, who during much of the late twentieth century could choose which local leaders would represent them to politicians, to effectively negotiate for patronage during elections. It also reduces residents’ already precarious identification with politicians and political parties while at the same time increasing, through political patronage, ties between residents and traffickers.

In the past two decades, organized criminal groups have had a growing impact on the political life of poor communities and, more broadly, on Rio de Janeiro as a whole (Leeds, 1996, pp. 70–73; Gay, 2005, pp. 54–58). These drug gangs can become involved in extended conflicts with one another that kill many and bring chaos to the communities where they operate. Clientelism provided the primary way that Rio’s impoverished linked themselves into political organizations, redressed basic grievances, and provided a minimal infrastructure to the favelas where many of the poor live. We know a great deal about the history of clientelism in Rio and the ways that drug trafficking has affected the political life of favelas by providing services to residents, resolving conflicts, and imposing controls on the operation of Associações de Moradores (Residents’ Association, AMs) (Goldstein, 2003, pp. 174–225; Gay, 2005, pp. 54–58). Nevertheless, we know relatively little about the specific effects that drug trafficking has had on the practice of clientelism in contemporary Rio.

In this article I will argue that the growing prevalence of crime in Rio’s favelas has caused a bifurcation and transformation of clientelist practices. What has emerged today is a combination of fixed-patron clientelism, reminiscent of what Scott Mainwaring has described as “traditional rural” clientelism, and neo-clientelist practices based on more flexible patron-broker-client ties. Both of these systems exist simultaneously and help deliver some limited services to favela residents. At the same time, however, they do little to provide a meaningful link between favela residents and the political system. In what I refer to in this article as two-tiered clientelism drug traffickers act as relatively fixed patrons to favela residents by distributing limited goods and services in exchange for favela residents’ silence and protection. At the same time, traffickers obtain some, though by no means all, of these goods through flexible patronage relations with politicians. Through brokers, traffickers augment their resources for patronage in favelas by negotiating among a number of outside political patrons during each election cycle. The result is the distribution of electoral clientelism to favela residents through traffickers who, as a result of their more flexible relationship with outside patrons, can skim off a portion of patronage resources before passing them back along to their captive clients in the favela. Examining clientelist transactions during the 1998 elections in the favelas of Tubarão, Santa Ana, and Vigário Geral, this paper shows how political and criminal clientelism operate simultaneously and cooperatively in Rio today to decrease direct links between the poor and the state and to improve the political position of drug traffickers both inside and outside favelas.

History and operation of clientelism

Scholars of Latin America have relied heavily on clientelism to explain numerous cross-class political interactions. While for a long time most scholars saw this practice in a negative light (Stokes, 1995, pp. 131–132; Sives, 2002, pp. 66–79; Hagopian, 1992, pp. 251, 271–279) in recent years a few scholars conducting intensely focused field research have shown its enduring importance as one of the few modes available to the poor to make demands of the state and effectively provide for themselves and their communities in highly unequal political systems (Gay, 1998, pp. 7–24; Auyero, 2000, pp. 41–44; for an example see Scott & Kerkvliet, 1977, p. 440).

Most political science approaches suggest that clientelism is a set of unequal, reciprocal, non-institutionalized, face-to-face exchange relations (Stokes, 1995, p. 55; Sives, 2002, pp. 80–82; also see Chalmers, 1977, pp. 412–415; Scott & Kerkvliet, 1977 also offers a nuanced analysis of this, pp. 442–443). This definition encompasses three narrower sets of practices. The first, patrimonialism, refers to the treatment of state power as private, almost familial, resources in which clients work in the personal interest of their patrons. The second, what Scott Mainwaring has called “traditional clientelism,” is a practice in which powerful fixed patrons, often operating with the force of arms, dominate a particular region and negotiate the votes of their captive clientele in exchange for outside support, a small portion of which is passed back to clients. In Brazil the classic example of this were rural coroneis, powerful landowners, who dominated clientele networks made up of poor workers in order to make demands of the state. In return for delivering their votes, and hence shoring up their political position with higher-level patrons, the coroneis would provide some personal assistance to those who worked for them. In these relationships clients were beholden to patrons in relatively static, spatially linked relationships.(Graham, 1990; Leal, 1977) Finally, neo-clientelism refers to the post-war urban phenomena in which independent brokers representing groups of clients move flexibly among politicians negotiating deals to provide services to residents in exchange for votes (Mainwaring, 1995, pp. 175–181). In these relations, which Robert Gay has so well documented in the Rio de Janeiro of the 1980s and 1990s, groups of clients are not indefinitely stuck with a broker, there is no necessary reproduction of pseudo-familial relations, and brokers are free to negotiate independent agreements with politicians in different electoral cycles (Gay, 1994, pp. 54–55).

A short history of clientelism in Rio de Janeiro

Over the last seventy years, as the urban population and the political franchise have generally expanded, clientelist practices in Rio have approximated those observed in other parts of Latin America. During Brazil’s Second Republic (1945–1964) clientelism played a major role in linking poor communities to the state (for a description of politics in this period see Mainwaring, 1995, p. 359). This was an era of growing concern about the role of favelas in cities and the emergence of numerous policies designed to incorporate favelas more effectively into the urban landscape thereby reducing some of the problems which many political and social leaders believed they caused (Leeds and Leeds, 1978, pp. 198–229). While the Catholic Church, in conjunction with the state, worked to provide services to favela residents and resituate them in other forms of housing to prevent the spread of communism and provide them a better life, many actual improvements in services came from local organizing or contacts with politicians. At the behest of Catholic organizations and the government, favela residents formed Associações de Moradores which helped mediate contacts with politicians and stop removal efforts (Burgos, 1998, pp. 28–31). Over time the leaders of these groups became the primary locus of clientelist contacts and their presidents the main brokers of relations between politicians and residents.

From the early 1960s until the mid-1970s the government of Rio and later the Brazilian military regime adopted an active policy of favela removal. In response, AMs organized, often through statewide interest groups, to prevent violent government eviction efforts. While the dictatorship crushed mobilizations during its most brutal period in the 1970s this changed in the late-1970s and early-1980s as the political system began a gradual opening and favela residents organized collectively to demand concrete improvements in services (Gay, 1994, pp. 19–21; Perlman, 1977, pp. 206–209, 258–262; Burgos, 1998, p. 38).

Nevertheless, even during the dictatorship’s worst years state-level and some federal-level elections occurred and clientelism played an important role in linking politicians to poor communities. Under Governor Antônio de Padua Chagas Freitas, the leader of a powerful political machine that dominated Rio politics from 1970 until 1982, many state resources were distributed through brokers to favela residents in exchange for electoral support. While Chagas Freitas was formally a member of the opposition Movimento Democrático Brasiliero (Brazlian Democratic Movement), ideology played a weak second to seat machine politics in his efforts to obtain votes (Gay, 1994, pp. 21–26; see also Diniz, 1982). During this period and later into the twentieth century favela brokers, usually elected AM leaders, sought out and negotiated among several politicians for the best arrangement for the communities they represented. Brokers worked hard to get good deals because the communities they represented could replace them through the normal electoral cycle.

In 1982 Leonel Brizola, an eminent exile during the dictatorship, was elected governor of Rio with extensive support from favelas. Brizola’s efforts to gain support from the poor were not explicitly clientelistic in that services were provided not by individual politicians directly to favelas but rather through state agencies under the direction of the governor. Politicians from Brizola’s party, however, ensured that the state agencies they controlled distributed these services as favors to communities which had provided them and their allies with votes (Gay, 1994, p. 26–34). The result was that a political machine centralized in the state apparatus managed exchanges of favors and votes. Brizola’s populism became a grander, more centralized form of clientelism (Gay, 1990, pp. 452–454). During this period the state government regularized electricity in most established favelas and made other significant improvements. These shifts had the effect of weakening AM power since, with many collective demands met and without the income from the provision of home electricity, AMs and their leadership became less relevant to the day-to-day lives of residents (Alvito, 2001, p. 161).

Around 1985, narcotics trafficking exploded in Rio as a result of the expansion of the Andean drug trade. As the home of the second largest port in Brazil and an important tourist hub, Rio was an easy choice to become a major drug transshipment center. Often built on very steep hillsides, Rio’s favelas provided ideal places to hide drugs and arms. Traffickers started to employ many residents in their operations, provided needed assistance to the poor, and fortified their leadership role within favelas (Dowdney, 2003, pp. 60–62). The consolidation of drug gangs has increased crime and violence in the city, provoking aggressive military and police responses which have resulted in the violation of favela residents’ rights (Leeds, 1996, pp. 63–66; Dowdney, 2003, p. 32).

The growing strength of drug traffickers did not simply stem from the ways they met the needs of residents or the weakness of AMs. It also resulted from state corruption, a social policy dominated by personalist politics, organizational strategies learned by common criminals while incarcerated with political prisoners during the dictatorship, and the particular economic and political geography of Rio in an era of globalizing markets. Increasingly drug traffickers realized that to remain safe in favelas they had to control elements of local political life. Consequently, in the late 1980s, with approximately six-hundred favelas in the city of Rio, police recorded the murder of two-hundred and forty community leaders (Barcellos, 2003, p. 234). As a result fewer and fewer favela residents had any interest in running for leadership positions and, increasingly, trafficker supported AM presidents stayed indefinitely in local leadership positions whether ratified by uncontested elections or not.

Since the mid-1980s drug trafficking organizations have played increasingly important roles in favelas providing limited assistance to residents and involving themselves in the political realm. In 1996 Elizabeth Leeds wrote, “ … in learning to play the electoral game, the drug groups run the risk of distorting the democratic process with the help of vote-hungry politicians and of helping re-create an even more pernicious form of patronage politics (Leeds, 1996, p. 76).” While we know that traffickers are involved in politics it is less clear from the existing evidence exactly how they are involved and how they have transformed local clientelist practices. To understand favela politics we need to delve more deeply into roles that traffickers play in the political life of favelas.

Understanding contemporary clientelism in criminal environments

The most explicit analysis of clientelism and drug trafficking comes from the Jamaican context. From the 1960s to the 1980s a violent form of political clientelism operated in the Kingston-Spanish Town area in which politicians distributed resources to poor communities via brokers who also acted as enforcers in keeping opposition supporters out of the neighborhood. To this end, politicians gave weapons to some partisans. This led to cyclical electoral violence that in 1980 took over eight hundred lives in a city of five-hundred thousand (Gunst, 1998, p. 111; Stolzoff, 2000, p. 99).

With this explosion of violence, politicians began backing away from armed local supporters who, in a period in which the global cocaine trade expanded dramatically became involved in drug trafficking (Small, 1995, pp. 140–142). This occurred at the same time as state resources available for clientelist spending declined under the weight of demands for structural adjustment (Sives, 2002, pp. 80–82). Increasingly these criminals, many of whom were former political gunmen, began to independently provide resources earned through smuggling to the communities that they came from and played increasingly large roles in providing local security. The rivalries of these new criminal gangs often paralleled pre-existing lines of political conflict.

Amanda Sives has argued that this has led to a type of double or parallel clientelism in which drug traffickers play a primary role delivering services to poor areas and maintaining order. She notes that while “there is still some pork left in the barrel” there has been a weakening of clientelist ties with politicians as residents of poor neighborhoods have increasingly looked to drug dealers for basic assistance and security (Sives, 2002, p. 84). During elections, however, residents of these communities continue to receive some support from political patrons. At times, Sives argues, this has resulted in outright conflicts between politicians and traffickers (Sives, 2002, pp. 80–84). Her claims about a parallel clientelism resonate with the experience of Rio, where drug traffickers and politicians also simultaneously provide resources to poor residents. This analysis provides the basis for building a framework for understanding the operation of Carioca clientelism today.

It is widely known that traffickers in Rio de Janeiro provide services to favela residents in exchange for their protection from police and other traffickers. Evidence from the three favelas I studied suggested that traffickers regularly threw parties for residents in each of the communities. They also helped to build and improve residents’ homes, worked to improve water service, resolved disputes, stopped petty theft, maintained basic order, provided street lighting, and helped residents gain access to health care and burial services. In general this assistance was limited, sporadic, and very much based on personal relationships. In the abstract, residents would describe traffickers’ services in a way that left the impression that they really thought they could get assistance from the traffickers when they needed it but, in practice, traffickers’ decisions to help residents seemed much more capricious and depended on personal relationships and circumstances. To make matters more complicated residents had only a limited amount of choice in these matters since it is very difficult to remove traffickers and, even then, they would only be replaced by other traffickers who would probably follow similar principles in distributing services. Despite these limitations, traffickers are one regular possible source of assistance for many favela residents in an environment in which state services to the poor are erratic and extremely limited. Traffickers in Rio, as in Kingston, are a continuous presence in poor communities and offer the regular, if limited and capricious, possibility of patronage. All of this improves the stature of traffickers who many favela residents, nonetheless, still view with a high degree of fear and circumspection. The ongoing presence of criminal patrons within the favela affects the activities of political patrons during electoral cycles.

In Cores de Acarí, Marcos Alvito builds on the experiences of one group of favelas to offer insights into contemporary clientelism in Rio. Paralleling some of Sives observations of Kingston, Alvito argues that Rio’s primary brokers, AM presidents, see traffickers and politicians as similar types of patrons. He notes, for example, that when a police occupation reduced drug sales in Acarí and drained traffickers of resources to provide certain services, an AM president quickly secured support from a sympathetic politician to maintain some of the programs the traffickers had paid for. The growing power of criminals and the formalization through them, Alivto argues, of more capitalist and individualized (as opposed to symbolic and communal) exchange relations within favelas has led to a transformation of political clientelist relations from collective efforts to gain resources transacted through ritualized events to more individualized and overtly instrumental exchanges of jobs for votes. AM leaders, he goes on, are transformed from being brokers who act in the interest of the community into entrepreneurial free agents who serve traffickers and look after their own interests by delivering services from wealthy patrons to residents (Alvito, 2001, pp. 162–164). Traffickers and their allies step into this gap and distribute resources to favela residents. This process reaches its logical endpoint when traffickers take direct control of the few universal goods (such as a soccer field) that politicians occasionally build in a favela. Here Alvito offers important insights into how clientelism has evolved by showing the overlapping roles of traffickers and politicians in favelas and the seeming primacy of traffickers.

Rio today lives with a two-tiered form of clientelism in which traffickers and politicians perform similar but different roles. They are comparable in that both provide patronage in exchange for support but diverge in terms of the type of support they seek. Five factors characterize this system. The first is that, as in Jamaica, traffickers are a constant presence in favelas who work to provide some degree of regular security and assistance to residents while politicians appear in the favela only around elections. Second, as Alvito notes, civic leaders mediate contacts between politicians, residents, and criminals. These leaders often have long-term associations with criminals and, as a result, bargain more in the interests of the criminals than the community as a whole. Third, the relationship between politicians and favela residents tends to be much less personal than the relationship between residents and traffickers. Indeed, major political investment today, such as with the well known Favela-Bairro program, seems to go into making improvements which collectively benefit whole communities rather than provisions which go to specific residents. While politicians hand out some jobs to garner support during election campaigns, those resources go through AM leaders who almost always have close ties to traffickers and, hence, hand out those jobs to residents on the basis of their relationship to traffickers rather than their commitment to the politician. Fourth, traffickers and their AM allies who mediate connections between the community and politicians, control which politicians gain access to the community, determine what gets built with political resources, and decide how to use those resources. This leads to criminals gaining disproportionately from political investment and actually controlling which residents benefit from the services that politicians bring to the community. Thus, political spending accrues more power and authority to the traffickers than it does to the politicians. Finally, traffickers, differently from brokers of an earlier area, provide an independent channel of assistance to residents that is much more personal and face-to-face than contacts residents have with politicians.

At the heart of these arrangements are elements of neo-clientelism, based on more flexible patron client ties, and a fixed-patron clientelism in which, because of the preponderance of force available to patrons, clients have little or no choice about who they will appeal to for resources. Traffickers and their brokers engage in neo-clientelist practices as they skillfully negotiate with different politicians during campaigns to get the best possible deal for themselves. Traffickers, however, then turn around and distribute those resources to residents on the basis of specific personal relationships within the favelas. Traffickers, often referred to in favelas as donos, serve as fixed patrons within favelas and can effectively charge a monopoly rent in terms of their patron-client exchanges. Residents must work with those traffickers whether they like them or not to gain access to resources. This fixed relationship decreases the bargaining ability of residents and forces them to maintain strong relationships with traffickers and their allies even if traffickers fail to act generously.

This is different from how clientelism operated in Rio in the 1970s and 1980s when, despite the efforts of some civic groups in Rio to eliminate clientelism, freely elected AM leaders, who were moderately accountable to residents, negotiated among politicians for the best pay-offs during elections. If the AMs leaders did not perform, residents could replace them before the next election in an attempt to get more out of political patrons. AM leaders, however, were not always removed after a poor showing in an election since they would receive some credit for their negotiating efforts even if benefits were not immediately obtained. Now, with the force of arms behind them, traffickers have much more flexibility in terms of what types of resources they distribute to residents.

The result of this particular practice of clientelism is that politicians gain votes but little else from poor communities. Traffickers gain access to resources which they then distribute to residents who are moderately grateful for the services but who see traffickers, more than politicians, as the source of those services. Traffickers and their allies skillfully negotiate bringing the resources in and then decide who gains access to those resources. In this sense, traffickers themselves become the clients in the relationship with the politicians. They then redistribute goods to the residents who, in turn, are their clients. AM leaders still act as brokers in each of these relationships but get little out of it. Since traffickers are criminals they have only a limited and fragile degree of respect within communities and they have difficulty directly interacting with state officials. What contacts politicians have with traffickers actually may hurt politicians’ image with residents who see them as even more corrupt and dishonest than before even though they will vote for them as a result of the traffickers’ support. When residents give politicians their support they believe they are getting something because of the skillful exploitation of political corruption by the traffickers and AM leaders rather than as a result of a legitimate political process. While this is similar to how residents viewed politician-broker relations prior to the rise of drug trafficker power in Rio (Gay, 1998), the particular involvement of criminals in the process and the specific way that criminals relate to residents only further increases favela residents alienation from the political system and further limits the few contacts favela residents would have had with politicians. To make matters worse, since AM leaders are really brokering a relationship between politicians and traffickers rather than between politicians and residents they may have more difficulty making appeals of politicians outside of electoral cycles than they did previously. Further, in an effort to avoid overt and potentially dangerous contact with criminals and their civic allies, politicians spend very little, if any, time in the communities which they provide assistance to and local leaders do very little to turn residents out to speeches or other events where they might build ties with other supporters of the politician. This contrasts markedly with the practices of clientelism that Javier Auyero has documented in Argentina (Auyero, 2000, pp. 5–11). The result is a political exchange that delivers votes but which limits the prestige of the patron politicians and prevents the deepening of political ties within these communities. At the same time traffickers increase their stature by delivering resources to favela residents.

These practices do not offer a rosy picture of democracy in Rio. Some time ago Guillermo O’Donnell wrote that Latin America could be seen as a map of blue, green, and brown areas. Blue reflected places where there was a regular rule of law and a real protection of basic rights. Brown areas reflected places where authoritarian practices persisted in the midst of a wider democratic polity. Green areas fell somewhere in between (O’Donnell, 1993, pp. 1361–1364). While O’Donnell offered this cartographic metaphor he has since lamented the limited understanding we have of these areas (O’Donnell, 1998, pp. 314–315).

In a sense, favelas are brown areas par excellence (Goldstein, 2003). The evidence presented here suggests, however, that these brown areas are far more integral to the Brazilian political system than O’Donnell’s cartographic metaphor would imply. What the shift in the structure of clientelism suggests is that criminal violence has itself become embedded in the political process. This means that, as things stand now, more intense policing, police reform, or NGO efforts are unlikely to materially decrease violence in the city because traffickers have strong connections into the state and civil society. At a deeper level the growing tolerance of violence in poor areas among the upper classes and the ongoing relationships many politicians maintain with drug traffickers means that crime has become embedded within the political system and is unlikely to be controlled without some sort of major political change. The result, then, is the existence of a political system in Rio which needs a degree of violence to function. Politicians depend on relationships with criminals to provide them with secure access to poor communities and, as a result, do little to remove them once in office. All of this is exacerbated by extensive police corruption and impunity. This violence materially curtails the rights of a large percentage of the city’s citizenry thus raising the question of how deeply consolidated “democracy” is in Brazil or if “democracy” is even the appropriate concept to use to understand a political system that systematically tolerates the abuse of residents by state and non-state actors. The system seems set up to provide increasing benefits to criminals and their political allies. Residents of favelas only benefit tangentially. A comparative analysis of three favelas will help flesh out this model of state criminal relations.

The practice of clientelism in three Rio de Janeiro favelas

To understand political exchanges in a highly criminal environment, this section will use ethnographic analysis to examine clientelist practices in three favelas during the 1998 national elections. Tubarão, a community of about 10,000, is located on a steep hill overlooking Rio’s picturesque and wealthy Zona Sul (South Zone) and has long been a site of heavy political investment. Santa Ana, a favela of around 4,000, is located near downtown Rio. Both of these communities were the sites of very powerful, well connected, and violent criminal gangs during this electoral period. The final favela, Vigário Geral, a community of around 10,000 located on the far northern edge of the city, was the site of a massacre of twenty-one residents by police in 1993. After that a large local social mobilization took place that succeeded in reducing violence. During the 1998 elections traffickers were very weak and had little control over the electoral process in the community.

I obtained the data for this paper between 1997 and 1999 as part of a broader project on human rights and violence in favelas. During this period, I spent between three and six months in each community conducting ethnographic research involving participant observation and more formal interviewing. During the 1998 elections I spent most of my time in Tubarão but traveled regularly to the other two communities where I had earlier conducted ethnographic research.

My initial contact in all three favelas was through the local AMs. While community leaders received me warmly it usually took two months of participant observation before residents began to open up. During this time I generally sat in the AM, got to know the local leaders, and conducted historical interviews with older residents. Over time residents grew accustomed to my presence and began to speak more openly with me. Some of my best information, however, came after a period of intense field research in a particular community had ended and I would return to visit with contacts. During these trips I was often regaled with details about much of what had occurred during my absence.

With the rapid rise in crime and violence throughout the region ethnographic methods are becoming an essential tool for understanding politics in Latin America today. People experiencing regular violence, subject to police impunity, and criminal gangs are usually not tremendously forthcoming about their feelings and experiences. They are exposed to substantial risks which middle-class educated North Americans are not subject to in their day-to-day lives or even on their brief visits to developing countries. As a result it often takes a long period of close research to gain even a rudimentary understanding of the politics of many populations in the region. Ethnography is an essential tool in the political analysis of the conditions that exist in Latin America today.

The high levels of danger involved in this work and resident apprehension led me to choose not to tape record interviews. During visits to communities I would do my best to compose an outline of notes which I would write up in detail later that day. As a result the quotes which appear in this piece are not exact quotes from informants but, rather, detailed reconstructions based on sparse notes usually taken during or immediately after a conversation. Finally, with the exception of Vigário Geral all the names used in this article are pseudonyms. This was done to protect the privacy and safety of those who provided me with information. I used the actual name of Vigário as a result of the extensive literature that already existed on the community when I started work there.

Tubarão

The primary political broker in Tubarão was the thirty-six year old AM president Bernardo. A tall, heavy-set, Afro-Brazilian newly converted to Evangelical Christianity, Bernardo had come to power in the mid-1980s after a notoriously unstable trafficker forced many residents out of the community. Bernardo, then AM vice-president, was among the exiles but he reacted by making a deal with the Comando Vermelho, at the time an up-and-coming prison-based gang which grew into one of Rio’s major drug factions, in which he agreed to pave the way for them capturing the community if he could return and take up the AM presidency.Footnote 1 His gambit succeeded and from that time forward, until his death in an automobile accident in 2002, he ran the AM. This long incumbency resulted from Bernardo’s significant political skills and his unparalleled abilities to manage relations both inside the favela and between the favela and the outside. Despite numerous violent changes in trafficker leadership he remained in the AM presidency. Jorge, a former president of Ceuzinho, a nearby favela, said, “[Bernardo] is tied to one group of traffickers and he knows that their life is short so he is planning to leave the community when this one leaves. He used to run the community for Fernando but then Alberto [the current head trafficker] had him killed and [then] he worked with Alberto … He is his compadre [the godfather of the trafficker’s child].”Footnote 2 Jorge went on, “Bernardo has a lot of jogo [short for jogo de cintura or political skill] and changes between one [trafficker] and the other.”Footnote 3 In 1992 a prominent journalist even publicly accused Bernando of running for political office to do the traffickers’ bidding in the state government. Like one of the brokers examined in Gay’s work, other AM leaders seek him out for his connections to the political world and advice on how to negotiate agreements with politicians (Gay, 1994, p. 59). I attended one meeting in which he offered detailed advice to leaders from a neighboring favela about how to force politicians and bureaucrats to deal with persistent water problems there and on another occasion the president of an AM in the distant Zona Oeste (West Zone) traveled hours by bus just to meet with him about how he might negotiate with politicians in the upcoming elections.Footnote 4 His success in bringing resources into the favela led to some talk among residents of a neighboring favela that he should also take over their AM. One resident said, “Bernardo is going to be the president of [Ceuzinho] because no one here has confidence. The associação does nothing here. Alexandre [the current president] wants to leave. Bernardo is the only one who can take over who has everyone’s confidence.Footnote 5

During the 1998 elections Bernardo managed to make agreements with two candidates for state assembly. The first, a woman from one of Rio’s most powerful political families, agreed to hire around fifty residents to distribute flyers for her on the weekends on the wealthy and often crowded beaches near the community. While she stored numerous signs in the AM headquarters for residents to use when they worked for her, the agreement she made with Bernardo did not give her access to the favela. Bernardo negotiated a separate deal with a second well-connected politician notorious for working with traffickers. In exchange for monopoly access to the community during the election Bernardo demanded money to improve a set of decaying stairs and to rebuild an outdoor quadra (dance floor, square), resources to hire a small number of residents to distribute pamphlets and put up signs for the politician in the favela, and three “clean” (not directly traceable) cellular phones for use by him, Alberto (who was in prison at the time), and Alberto’s girlfriend.Footnote 6 While I am not sure if the politician delivered the telephones the campaign manager readily accepted the demands. Bernardo then hired a group of five women to work for the politician in the community. This group included the expectant mother of the jailed head trafficker who once discussed possible baby names with me, the wife of another jailed trafficker, as well as three other women associated with the AM leadership.

The story of the 1998 elections in Tubarão, however, does not stop there. In the days immediately before the election the traffickers made an agreement with a second set of politicians and they ordered Bernardo to remove the signs for the politician who had paid for the quadra and put up a new set of signs for the other politicians. Bernardo angrily ordered the five residents who worked for the first candidate to take down the old signs and put up the new ones. As he was frantically trying to get the signs changed he said that he wanted to get himself out of “these problems” and that he takes “a subsidy from them [the traffickers] so when they decide to support a candidate I have to also. This was our [the favelas] chance to be respected again. We had a candidate from the same party as the mayor and had he been elected people would have paid attention … Now we were forced to support a candidate who they [the traffickers] had closed [the agreement with] without consulting me.” Footnote 7 It was never clear what the last group of politicians had agreed to give the traffickers.

This last minute change created difficulties. Residents were now confused about who to vote for and, capitalizing on the division between the AM and the traffickers, one of the local Neo-Pentecostal churches backed a third candidate. As a result, the favela’s votes split among at least three separate contenders none of whom won election. Despite the shift in candidates both the quadra and the stairway were finished by mid-October, just after the first round of voting.Footnote 8 The pressure to finish the quadra came from the traffickers who used the space frequently to hold parties and dances for residents and who wanted to host a large party for the International Day of the Child. The celebration, which drew a huge attendance among the community’s children, was administered by the AM and presided over by traffickers from a “luxury” box over the quadra.Footnote 9

The outcome of this event reflects the skillfulness of the traffickers in manipulating clientelist politics. They used Bernardo’s credibility and skills to bring in resources to improve the quadra where they hosted parties for residents. Thus, Bernardo brought resources into the community which they employed to directly deliver services to residents. By cutting the deal with the second set of politicians, however, the traffickers gained some other direct reward from politicians while at the same time undermining the possibility that the community and the AM would develop a powerful political patron who could compete with them.

All of this also reflects highly on Bernardo’s skills as a broker. By asking for the cell phones Bernardo communicated to the politician’s manager that he had connections with the traffickers without ever directly saying that he represented the traffickers. He was, thus, selling the traffickers’ protection to the politician’s campaign without ever formally engaging the traffickers. At the same time, Bernardo brought improvements to the community, found short-term employment for many residents, and helped the traffickers in their efforts to deliver services.

Santa Ana

During the 1998 elections the primary political broker in Santa Ana was Josias, the father of Doca, the manager of the drug trade in the favela and a relatively powerful trafficker in Rio. Josias was a small, wiry, and tough man with always perfectly brillanteened hair and a ready, usually very crude, joke. He had lost a leg when, as a sanitation worker, it was mangled in the back of a garbage truck. As a result he hobbled around the community on crutches as he tried to manage the sometimes tense relationships in this very violent favela that for much of the time I worked there experienced nightly gunfights.

Unlike Bernardo, Josias was not a natural political broker. In addition to his injuries which limited his mobility, his relationship with his son and the inadvertent murder of his wife by the police the year before the election made him very apprehensive about interacting with outsiders. While he was quite gregarious with other residents, he was relatively quiet and very formal when dealing with outsiders he did not know well. As with other favelas, however, politicians sought out links with Santa Ana to gain access to votes. Joselino, a voluble evangelical hardware store owner, said “Politicians can only get access to a community if they make deals with traffickers. The work between the traffickers and the politicians is limited to the election.”Footnote 10

The primary contacts which Josias made with the outside came through Doca’s attorney who otherwise spent most of his time trying to bribe police to release Doca’s employees. The lawyer put Doca and Josias in touch with two political candidates who, after negotiations, agreed to provide the materials to build a large covered quadra with the understanding that the traffickers would provide the manpower to build it and would guarantee them monopoly access to the community. Joselino noted “ … they [politicians] are all liars. Nelson [a politician] agreed to pay for the materials for the quadra if the traffickers paid for the mão de obra [labor]. Rodrigo [another politician] came in with him and knew what was going on.”Footnote 11 After the candidates and the AM leadership concluded the agreement a large banner went up over the middle of the small favela’s main street saying that Josias and the trafficker’s lawyer supported the two politicians, building supplies arrived, and men in Doca’s employ began carrying supplies up to the construction site. Within a year after the elections, local laborers finished building a huge dance floor capable of holding hundreds of people. Josias and Doca found enough resources build a small bar on the dance floor from which Josias could sell refreshments during parties. With its high roof traffickers and other community groups could use the quadra for festivities and sports such as volleyball and indoor soccer. These efforts, according to one resident, brought about 2,400 votes to each candidate, one of whom, with substantial support from other parts of the city, won election to the federal legislature.Footnote 12

Despite its apparently public nature, the traffickers controlled access to the quadra and used it primarily to host baile funks (hip-hop balls). One young Evangelical Christian noted that “ … they [the traffickers] have bailes all the time now and they are very loud. You can hear the beats up here. The houses around the quadra have all been devalued … [no one] who lives there could sell one of those houses now. Traffickers patrol around the dance with large weapons. You can’t get the dances changed because they run them. Eventually the dance will end because the police don’t like them.”Footnote 13 Doca’s girlfriend held the keys and controlled access to the space. One resident used the quadra to offer soccer classes but stopped as a result of his discomfort with constantly having to ask her and, hence, the traffickers for permission to teach the class. He claimed that he did not want to provide that type of example to children. As the 2002 elections approached the federal deputy who had originally provided money to build the quadra came back to campaign again and this time provided money to offer sports classes in the quadra to community children. The return of the candidate, who was reelected to the federal legislature with one of the highest vote totals in the state, confirmed that candidate’s perceptions about the success of his relationship with Santa Ana.

These contacts, however, undermined residents support for politicians. No one seemed very excited about their activities in the community or particularly attached to either of the two candidates or their ideas. Indeed, residents seemed not to really trust the politicians at all as they became identified as outsiders whose support the traffickers had skillfully bought in exchange for their votes. Joselino, concerned about the complicity between politicians and drug traffickers, noted sarcastically “the guy who was elected will work with the Medillín cartel.”Footnote 14

Santa Ana provides interesting evidence about how clientelism has adapted to drug trafficking. The AM acts as a broker between the drug traffickers and the politicians rather than between the residents and politicians. The goods the politicians give to the community mirror the way politicians had previously provided communal assistance in improving water services or building health clinics but instead of turning those services over to an AM or some other elected body, politicians directly delivered those resources to the traffickers who in turn used them to put together events to bolster their position in the community and tighten their relationship to residents. As a result, residents really did not receive anything from the politicians. Rather, the traffickers received a quadra which they then used to distribute services to residents. Further, traffickers paid for all the work to build the quadra. Direct personal assistance to residents came solely from the traffickers. As a result, residents felt no interest in the politician or loyalty to him. They voted for a politician and, in turn, received benefits from traffickers who used that transaction to shore up their position in the favela.

Vigário Geral

During the 1998 elections Vigário Geral had no active political broker. The community had come through a turbulent period with the massacre of twenty-one residents by police in 1993, extensive police efforts to kill traffickers involved in the murder of several police officers, and a social movement that organized to bring peace to this very violent community. This movement eventually led to the creation of the Casa da Paz, an important NGO that operated in the community until 1998, and also involved the Grupo Cultural Afro-Reggaes and Doctors without Borders. By 1998 police under special orders from the governor had occupied Vigário and violence had dropped dramatically. The natural person for the job of broker was Seu Almeida, the president of the AM. Miguel, a community plumber formerly in the employ of the traffickers, noted that Almeida’s political connections had led to the traffickers strongly supporting his candidacy for AM president in the years before their hold on the favela had weakened. “Seu Almeida had always believed himself to have the skills to be a good president. He had connection in politics and politicians knew him. The trafficker liked that.”Footnote 15 In other words, traffickers chose Almeida to lead the AM in order to exploit his connections with politicians. The subsequent social movement and political intervention, however, weakened traffickers. While Almeida succeeded in holding on to his office, the complicated situation with the traffickers and Almeida’s extensive efforts to bring services in through contacts with NGOs and other state agencies caused him to sit out this electoral cycle.

Vigário’s unlikely broker was Rúbia, the owner of a small restaurant that catered mainly to the NGO and government workers who had descended on the community as a result of the massacre. Rúbia, a friendly outgoing woman, had moved to the community when she left an abusive husband who lived in the neighborhood that bordered the favela. At the time she was trying to run her nascent business and raise a young daughter. Her restaurant, however, became a popular hub among politically engaged residents who often would sit with her drinking beer late into the night. Many of these activists complained that Almeida had not done enough to help the community and, partially as a result of anger stemming from how he came to power, plotted to replace him with his Vice-president, a man who had joined the AM after trafficker power had waned. In this environment, Rúbia took the lead in making contact with a powerful state legislator to help improve a soccer quadra and makeshift movie theatre that traffickers had built years before and that had fallen into disrepair.Footnote 16

The patron, the longest serving state assemblyman in Rio de Janeiro, was from a conservative party that backed Cesar Maia, Rio’s former and future hard-line mayor, as its gubernatorial candidate. He provided Rúbia with money to hire five workers to put up signs and distribute flyers which bore his name, face, and, in smaller print, Cesar Maia’s name as his choice for governor. Shortly after the signs went up, the traffickers told Rúbia to take them down because they supported the election of Maia’s main rival Anthony Garotinho of the left-leaning PDT. When Rúbia refused a more powerful trafficker arrived from another favela to deal with her intransigence. The traffickers then kidnapped her boyfriend Edivaldo and held him at gun point as they insisted Rúbia take down the signs. Still she refused. Eventually the traffickers and Rúbia compromised. They agreed to let her leave the signs up if she would remove Maia’s name. I noticed that signs for conservative candidates in other favelas I visited had been similarly altered.Footnote 17 All of this left Rúbia and Edivaldo very distraught and soon afterwards they broke up. Rúbia moved out of the favela less then a year later and reconciled with her husband.

Vigário contrasts with the other two communities examined in this paper. Here Rúbia established a clientelist network where she brokered a deal with an outside patron who would provide both personal and communal assistance directly through her to residents. The traffickers, however, could not abide this arrangement and threatened the use of force against Edivaldo creating a very tense situation for Rúbia. Their efforts contributed to breaking up a nascent political network in the community thereby maintaining conditions in which traffickers could retake power several years later. This story also makes clear that the drug traffickers in this part of Rio had a definite preference for the election of Garotinho who they may have believed would not be as tough on crime as Maia. It is interesting to note that when Garotinho visited Tubarão during this campaign a low-level trafficker asked Bernardo to lobby the candidate to eliminate the police presence in that community clearly believing that this request might not fall on deaf ears. Bernardo disagreed and rebuked the adolescent.

Almeida’s role in all of this is very interesting. He chose to stay out of the political game that year. He may have thought that his participation would lead to similar difficulties between him and the weakened traffickers that Rúbia had encountered. Another reason is that he was actively engaged in a different sort of politics working through NGOs and government agencies to make improvements in the community and bring jobs and other services to residents. During the two years in which I conducted this research he brought in a major program to improve water service that employed tens of residents and also brought in a computer training class to help residents gain qualifications for white collar jobs. Almeida had an advantage over other favela leaders in these efforts as a result of the immense attention that had focused on Vigário in the wake of the massacre. The significant NGO presence connected him into a number of different funding sources (see Arias, 2004, pp. 15–16). Almeida’s choice not to enter the competitive clientelist fray of the elections was a savvy one which avoided conflict with the traffickers and freed up his time to bring in other types of resources.

Clientelist practices in Rio today

Much of the existing scholarship on clientelism focuses on how clientelism creates vertical identifications between the poor and political elites thereby strengthening state legitimacy, reducing horizontal bonds based on class, and limiting opportunities for radical political change. Many see clientelism as a fundamental component in building the hegemony of the dominant classes and reinforcing state power against possible insurrections by the poor (Stokes, 1995, pp. 131–132; Sives, 2002, pp. 66–79; Scott & Kerkvliet, 1977, p. 440; Hagopian, 1992, pp. 251, 271–279). Auyero has shown with considerable nuance how clientelism goes beyond just problem-solving into the realm of identifying voters positively with the patron (Auyero, 2000, p. 123). In these circumstances, clientelist exchanges through networks help to solve problems by facilitating the exchange of favors and, in the process, build cross-class political identities that contribute to state legitimacy.

The endemic presence of powerful violent criminals has displaced this process in Rio de Janeiro. The evidence from the cases examined here shows that clientelism in Rio today works on two separate levels. In the first instance it is a relationship between politicians and favelas in which favors are granted through a broker, usually an AM leader, to the community as a whole in the form of improvements to public goods. Often, however, control of that public good passes from the broker into the hands of criminals rather than residents. Traffickers then, again often through the AM president, redistribute these and other goods to residents. Politicians spend relatively little time in favelas and develop few, if any, personal relationships with residents. Rather, it is traffickers, through AM leaders, who build personal relations with residents by delivering services directly to them. Jobs are either given to residents through the traffickers or their allies in the AMs. All of this creates conditions in which residents feel little commitment directly to politicians or political parties. Instead they provide votes to politicians the traffickers and their allies support in exchange for assistance from traffickers who they then support in their efforts to avoid arrest or murder by police and other traffickers (this system has something in common with how Archer, 1995, pp. 189–190 has described broker clientelism in Colombia). These data also suggest that politicians and traffickers today still provide substantial communal goods to favelas. As a result some of Alvito’s observations about the end of the distribution of communal goods through clientelism may not be generalizable to all favelas in Rio.

At the heart of the complications that traffickers bring to the clientelist process is the complexity of their social status. While they do have some prestige among certain sectors of favela residents and may even be seen as having higher prestige among these sectors than most politicians, traffickers are criminals who have extremely low status in society as a whole and even among many residents of favelas who resent the fact that they do not hold regular jobs and that they bring so much violence to their communities. This is at odds with typical clientelist practice in which relations operate exclusively between higher and lower status actors (even if low-status clients sometimes have a dim view of their exploitative high-status patrons). In the case of favelas, traffickers, as criminals, act as an intermediate patron that has lower social status in society as a whole than many of the favela residents who are their clients. This official public status differential causes politicians to avoid openly appearing with criminals and to minimize their contact with the residents of the communities where traffickers operate. The result is that favelas are increasingly pushed to the edges of the political process and little effort is made to bring them into contact with others who voted for a particular politician.

This contemporary practice of clientelism has not done much to build support for the existing Brazilian political system. As Bête, a resident and NGO activist in Santa Ana said, “Democracy has done little for the residents of the favela other than provide them with the right to talk. Prior [to that, during the military regime,] individuals were not allowed to talk but the politics have not changed. The poor do not have what they need and politicians tend to provide things to the poor via clientelism. The poor have little more ability to interact with politicians than was available to them under the dictatorship (on changing conditions for the poor in Brazil see Perlman, 2004).”Footnote 18 The ongoing violence propagated by traffickers and police limits the ability of residents to organize within favelas, co-opts favelas’ civic leaders into criminal networks, and causes life in many favelas to be as bad or worse for residents that it was under the military dictatorship.

Conclusion

This paper has shown that the practice of clientelism is alive and well in Rio but that it has adapted to the emergence of powerful locally based narcotics traffickers. Drug dealers have become inserted into the process through which votes are exchanged for favors. This results in a two-tiered clientelist system in which drug traffickers make exchanges with politicians and then turn around and deliver services to residents of favelas themselves. While criminals can engage in the flexible negotiations with multiple politicians that characterize neo-clientelist arrangements, favela residents cannot choose which trafficker they will ask for assistance since each favela is dominated by one gang. As a result residents negotiating power during electoral periods is weakened by the fusion of a fixed-patron clientelism in which clients have virtually no choice about who their patron will be and of the more flexible neo-clientelism that characterized Rio politics after the Second World War. This results in vote exchanges and the election of politicians skilled at making arrangements with criminals. It does little to build state legitimacy among the poor and leads to a situation in which many elected officials tolerate criminal activity.

These insights provide a perspective on the micro-politics of voting in Brazil that shed considerable light on the institutional and relational mechanisms that lead to voting decisions. This information can only be obtained through a close examination of local-level politics. Given the growing interest in political behavior and, more broadly, the application of neo-institutional frameworks to the Brazilian case this paper points to the critical need for more micro-level ethnographic examinations of political practice in urban and rural Brazil. Such an effort would not only tell us about how traffickers participate in elections in Rio but would also provide insights into such issues as machine politics in the Brazilian Northeast, the continuing practice of coronelismo in parts of rural Brazil, and how the growing power of evangelical churches affects political practice in major cities.