1. The literature on the role of collective organizations in the process of political change in Brazil is vast. For a sample of the more optimistic theses, see M. G. M. Gohn, Reivindicações Populares Urbanas: Um Estudo sobre as Sociedades de Amigos de Bairros de São Paulo (São Paulo: Edições Cortes, 1981); J. A. Moisés, V. Martínez-Alier, F. de Oliveira, and S. de Souza Lima, Contradições Urbanos e Movimentos Sociais (Rio de Janeiro: Paz e Terra, 1977); and P. Singer “Movimentos de Bairro,” in São Paulo: O Povo em Movimento, edited by P. Singer and V. Caldeira Brant (Rio de Janeiro: Editora Vozes, 1981), 83–107. On the birth of the Partido dos Trabalhadores, see M. Keck, “Democratization and Dissension: The Formation of the Worker's Party,” Politics and Society 15, no. 1 (1986):67-95. On the connection between collective organizations and political parties, see M. H. Moreira Alves, “Grassroots Organizations, Trade-Unions, and the Church: A Challenge to Controlled Abertura in Brazil,” Latin American Perspectives 2, no. 11 (1984).

2. Clientelism is defined here as the distribution (or promise) of resources—public or private—by power seekers or holders in return for votes. Populism represents an appeal by political elites to popular discontent with the distribution of power in society. In a sense, populism is simply a more sophisticated form of clientelism in that both represent strategies for the political incorporation of the masses. But although populism is also essentially an exchange of “votes for patronage” in that the distribution of benefits is calculated on the basis of what is necessary for consolidating political power, populism is less transactional and more ideological than clientelism.

3. As an example, see M. C. Campello de Souza, “A Democracia Populista (1945/1964): Bases e Limites,” in Como Renascem as Democracias, edited by A. Rouquié, B. Lamounier, and J. Schvarzer (São Paulo: Brasiliense, 1985), 73–103; and A. Stepan, “State Power and the Strength of Civil Society in the Southern Cone of Latin America,” in Bringing the State Back In, edited by P. Evans, T. Skocpol, and D. Rueschemeyer (New York: Cambridge University Press, 1985), 317–43.

4. Among the best appraisals are R. Boschi, A Arte da Associação Política de Base e Democratização no Brasil (São Paulo: Edições Vertice, 1987); R. Boschi and L. Valladares, “Problemas Teóricos na Análise de Movimentos Sociais: Ação Coletiva e o Papel do Estado,” Espaço e Debates 8 (1983):64-77; R. Cardoso “Movimentos Sociais Urbanos: Balanço Crítico,” in Sociedade e Política no Brasil Pós-64, edited by V. S. Cruz (São Paulo: Brasiliense, 1983), 215–39; E. Diniz, “Favelas: Associativismo e Participação Social,” in Movimentos Colectivos no Brasil Urbano, edited by R. Boschi (Rio de Janeiro: Zahar, 1983), 27–74; and L. A. Machado da Silva and A. Ziccardi, “Notas para uma Discussão sobre Movimentos Sociais Urbanos,” Cadernos do Centro de Estudos Rurais e Urbanos 13 (1980).

5. For recent discussions of changes in the nature of political discourse in Brazil, see G. Banck, “Poverty, Politics, and the Shaping of Urban Space: A Brazilian Example,” International Journal of Urban and Regional Research 10, no. 4 (1986):522-40; and C. N. F. Santos, “Metrópoles e Outras Cidades Brasileiras: Bem antes de 60, Muito depois de 80,” Espaço e Debates 13 (1984):103-16.

6. This assumption was especially characteristic of those who borrowed from Marxist urban theory. So pervasive was the influence of this body of theory that few discussions of the “new urban social movements” in Brazil failed to cite the following works: J. Borja, Movimentos Sociales Urbanos (Buenos Aires: Ediciones SIAP, 1975); M. Castells, City, Class, and Power (London: Macmillan, 1978); M. Castells, The Urban Question (London: Edward Arnold, 1977); and J. Lojkine, O Estado Capitalista e a Questião Urbana (São Paulo: Martins Fontes, 1981).

7. See T. P. R. Caldeira, “Electoral Struggles in a Neighborhood on the Periphery of São Paulo,” Politics and Society 15, no. 1 (1986):43-66. Caldeira's work is particularly interesting in illustrating the problems that face the PT in low-income neighborhoods.

8. See N. V. T. Lima, “As Eleições de 1982 em Favelas de Rio de Janeiro,” paper presented at the meeting of the Associação Nacional de Pós-Graduação e Pesquisa em Ciências Sociais (ANPOCS), Rio de Janeiro, Oct. 1983.

9. See S. Oliveira, “O Movimento Associativo e o Debate sobre a Representação de Interesse: Principais Interpretações,” paper presented at the meeting of ANPOCS, Rio de Janeiro, 1984.

10. The MDB became the PMDB (Partido do Movimento Democrático Brasileiro) following the electoral reforms of 1979.

11. Estimates of the total population of Rio's more than four hundred favelas vary. While a recent municipal survey put the total at seven hundred and twenty-two thousand, popular estimates range between one to two million. For the most recent and one of the more reliable surveys of conditions in Rio's favelas, see P. Cavallieri, “Favelas Cariocas: Mudanças na Infra-Estrutura,” in 4 Estudos (Rio de Janeiro: IPLANRIO, 1986), 19–38.

12. For an analysis of the MDB in the city of Rio de Janeiro and its relationship with its clientele, see E. Diniz, Voto e Máquina Política: Clientelismo e Patronagem no Rio de Janeiro (Rio de Janeiro: Paz e Terra, 1982).

13. On the relationship between the state and FAFERJ from its formation until 1982, see ibid. For a discussion of more recent events see, R. Gay, “Political Clientelism and Urban Social Movements in Rio de Janeiro,” Ph.D. diss., Brown University, 1988.

14. Brizola engaged in and lost a dispute over use of the pre-coup party name of the PTB (Partido Trabalhista Brasileira).

15. The PDT's ideology consists of an awkward blend of democratic socialism and traditional laborism associated with the pre-coup PTB. The PDT is relatively progressive in the ideological spectrum of political parties in Brazil, but like most parties (excepting the PT), the PDT is organized from the top down. While not a clientelist party, it retains certain populist elements.

16. These projects included programs to install sewage, drainage, and water systems in the favelas, a pilot project for legalizing the tenure of favela communities, and a major project to improve educational opportunities for the working class by constructing special schools. For a description of the administration's programs, see Cavallieri, “Favelas Cariocas.”

17. The community studies were conducted during seven months of intensive field research in the two favelas between May and November of 1986.

18. Although the association president worked for Jorge Leite for five years, he expressed no affinity with the man or his party. In 1985 Jorge Leite won the PMDB candidacy for the mayoral elections in the city of Rio, and the president worked for him during the campaign. This time, however, the president received no payoff because the clientelist party was out of power. Jorge Leite suffered a crushing defeat at the hands of the PDT, and soon after, the president decided to switch his political allegiance. It would have been unwise for him to continue working for a politician who had recently suffered a major political reversal and was unlikely to regain power in the near future. It was also inconvenient for the president to continue to support a known politician from the losing party because it would be an obstacle in gaining access to the state bureaucracy, which was now under the control of the opposing party.

19. During this time, the president received calls from candidates for whom he had worked in the past, including Jorge Leite. Getting the favela's roads paved in 1982 was the president's greatest victory, and the only one directly linked to elections thus far, but he had achieved other “understandings” in the past.

20. It is important to point out that while the PDT administration at the time was involved in a certain kind of electioneering, it did not engage in naked clientelism of the kind employed by the PMDB machine in power until 1982.

21. This is not to say that the association was unwilling to accept “gifts” designed to generate political support. It was made clear, however, that acceptance of such donations would in no way oblige the community to vote for the donor. As a result, most of the offers that the association received during the election campaign did not materialize. The soccer club, however, which was run by the same individuals as the association, did receive money for a barbecue and a set of shirts.

22. Of the PDT candidates who had contact with Vidigal prior to the election in their capacities as the administrators of favela improvement programs, only one was favored by respondents in the survey. According to Vidigal leaders, this particular candidate was popular precisely because he never used his position to pursue personal political ambitions.

23. Age and sex of respondents were controlled for in the sample.

24. Because I am interested primarily in the relationship between the two major political parties or alliances in the election, table 1 lists neither preferences for other parties nor “undecided” responses. In the two favelas, only 8 percent of preferences for governor, 6 percent for senator, 7 percent for federal deputy, and 6 percent for state deputy named alternative parties (predominantly the PT). Those who remained undecided one week before the election totaled almost 16 percent of respondents for the election for governor, 42 percent for senator, 41 percent for federal deputy, and 43 percent for state deputy.

25. Log-linear analysis revealed that the variation in political preferences in the two favelas was statistically significant at each level.

26. This finding demonstrates the effectiveness of this type of transaction in that the survey probably underestimated the proportion of respondents who were actually going to vote for the two candidates, especially for Oswaldo. The survey was conducted two days after the candidate's “guided tour” of the favela and before the distribution of campaign publicity. Furthermore, many of those who were as yet undecided would have ended up opting for the most familiar names on the ballot sheet when they arrived at the polling booth. The influence of the neighborhood association was clear in that 52 percent of those who planned to vote for Oswaldo said they were doing so because the association president had instructed them to vote in this manner.

27. The election for governor differed in that both candidates enjoyed high visibility in the local press.