¹ In Karl's terms, this is a middle-range definition of democracy, in that it falls between the narrow Schumpeterian range of contestation needed for strictly intraelite competition and approaches that depend on particular socioeconomic or participatory outcomes. See Karl, Terry Lynn, “Dilemmas of Democratization in Latin America,” Comparative Politics 23 (October 1990)CrossRefGoogle Scholar.

² Democratization is defined here as the process of movement toward these conditions, while the consolidation of a democratic regime requires fulfilling all of them. Regimes can therefore be in transition to democracy—further along than liberalization—but still fall short of a democratic threshold. For further discussion, see Mainwaring, Scott, O'Donnell, Guillermo, and Valenzuela, J. Samuel, eds., Issues in Democratic Consolidation (Notre Dame, Ind.: University of Notre Dame Press, 1992)Google Scholar.

³ On associational autonomy as a democratic right, see Dahl, Robert, Dilemmas of Pluralist Democracy: Autonomy vs. Control (New Haven: Yale University Press, 1982)Google Scholar. Note that autonomy is inherently relational and is therefore a matter of degree. On the difference between autonomy and capacity, see Fox, Jonathan, The Politics of Food in Mexico: State Power and Social Mobilization (Ithaca, N.Y.: Cornell University Press, 1992), chap. 2Google Scholar.

⁴ See, e.g., Cohen, Joshua and Rogers, Joel, “Secondary Associations in Democratic Governance,” Politics and Society 20 (December 1992), on the U.S.CrossRefGoogle Scholar; Putnam, Robert, Maying Democracy Work, Civic Traditions in Modern Italy (Princeton: Princeton University Press, 1992), on ItalyGoogle Scholar; Fox, Jonathan and Hernández, Luis, “Mexico's Difficult Democracy: Grassroots Movements, NGOS and Local Government,” Alternatives 17 (Spring 1992), on MexicoCrossRefGoogle Scholar; and Lawson, Kay and Merkl, Peter, eds., When Parties Vail: Emerging Alternative Organizations (Princeton: Princeton University Press, 1988)Google Scholar, on broader problems of party representation. This paper treats intermediate associations as broadly representative though not necessarily democratic. On the problem of internal democracy within such organizations, see Fox, Jonathan, “Democratic Rural Development: Leadership Accountability in Regional Peasant Organizations,” Development and Change 23, no. 2 (1992)CrossRefGoogle Scholar. Note also that intermediate associations in developing countries often do not appear as formal organizations; they may be kinship or religiously based community associations, for example, as in many African or Middle Eastern societies.

⁵ The guaranteed rights of political citizenship in a democracy include basic civil and political freedoms, majority rule with minority rights, and the equitable administration of justice, as well as respect for associational autonomy.

⁶ Associational autonomy is an especially vital right for the poorest members of society, for two main reasons. First, they are usually the most vulnerable to state-sanctioned coercion should they express discontent. Second, their survival needs make them especially vulnerable to dientelistic incentives. Together, these threats and inducements inhibit autonomous collective action. See Collier, Ruth and Collier, David, “Inducements versus Constraints: Disaggregating ‘Corporatism,’ ” American Political Science Review 73 (January 1979)CrossRefGoogle Scholar.

⁷ The appearance of subordination should not be confused with actual submission, however; see Scott, James C.Weapons of the Weak; Everyday Forms of Peasant Resistance (New Haven: Yale University Press, 1985)Google Scholar; and idem, Domination and the Arts of Resistance (New Haven: Yale University Press, 1990)Google Scholar. Others stress the subjective importance of “trust” in such dependent relationships; see Roniger, Luis, Hierarchy and Trust in Modern Mexico and Brazil (Westport, Conn.: Greenwood, 1990)Google Scholar. For a view that stresses the role of coercion, see Flynn, Peter, “Class, Clientelism, and Coercion: Some Mechanisms of Internal Dependency and Control,” Journal of Commonwealth and Comparative Politics 12, no. 2 (1974)CrossRefGoogle Scholar. For overviews of political clientelism, see Clapham, Christopher, ed., Private Patronage and Public Power: Political Clientelism in the Modern State (New York: St. Martin's Press, 1982)Google Scholar; Eisenstadt, S. N. and Roniger, Luis, “Patron-Client Relations as a Model of Structuring Social Exchange,” Comparative Studies in Society and History 22 (1980)CrossRefGoogle Scholar; Eisenstadt, S. N. and Lemarchand, René, eds., Political Clientelism, Patronage and Development (Beverly Hills, Calif.: Sage Publications, 1981)Google Scholar; Gellner, Ernest and Waterbury, John, eds., Patrons and Clients in Mediterranean Societies (London: Duckworth, 1977)Google Scholar; Kaufman, Robert R., “A Patron-Client Concept and Macropolitics: Prospects and Problems,” Comparative Studies in Society and History 16, no. 3 (1974)CrossRefGoogle Scholar; Rouquié, Alain, “Client Control and Authoritarian Contexts,” in Hermet, Guy, Rose, Richard, and Rouquié, Alain, eds., Elections without Choice (New York: John Wiley, 1978)Google Scholar; Schmidt, Steffen W. et al., eds., Friends, Followers and Factions: A Reader in Political Clientelism (Berkeley: University of California Press, 1977)Google Scholar; Strickon, Arnold and Greenfield, Sidney, eds., Structure and Process in Latin America: Patronage, Clientage and Power Systems (Albuquerque: University of New Mexico Press, 1972)Google Scholar. Note that the literature on political clientelism virtually stopped over a decade ago, long before the flowering of analysis of regime change. Few analysts explore how the persistence of machine politics affects the nature of transitions to civilian rule. For exceptions, see Hagopian, Frances, “‘Democracy by Undemocratic Means?’ Elites, Political Pacts and Regime Transition in Brazil,” Comparative Political Studies 23 (July 1990)CrossRefGoogle Scholar; idem, “The Compromised Consolidation: The Political Class in the Brazilian Transition,” in Mainwaring, O'Donnell, and Valenzuela (fn. 2); and O'Donnell, Guillermo, “Challenges to Democratization in Brazil,” World Policy Journal 5, no. 2 (1988)Google Scholar.

⁸ On the construction of political machines, see, among others, Luigi Graziano, “PatronClient Relationships in Southern Italy,” in Schmidt et al. (fn. 7); Scott, James C., “Corruption, Machine Politics, and Political Change,” American Political Science Review 63 (December 1969)CrossRefGoogle Scholar; idem, Comparative Political Corruption (Englewood Cliffs, N.J.: Prentice-Hall, 1972)Google Scholar; Shefter, Martin, “The Electoral Foundations of the Political Machine: New York City, 1884–1897,” in Silbey, Joel et al., eds., The History of American Electoral Behavior (Princeton: Princeton University Press, 1978)Google Scholar; Schmidt, Steffen W., “The Transformation of Clientelism in Rural Colombia,” in Schmidt, et al. (fn. 7); Tarrow, Sidney, Peasant Communism in Southern Italy (New Haven: Yale University Press, 1967)Google Scholar; and Weiner, Myron, Party Building in a New Nation (Chicago: University of Chicago Press, 1967)Google Scholar.

⁹ For discussions of how clientelism can evolve from patrimonial to repressive, increasing the role of coercion, see Anthony Hall, “Patron-Client Relations: Concepts and Terms,” in Schmidt et al. (fn. 7); and James C. Scott and Benedict J. Kerkvliet, “How Traditional Rural Patrons Lose Legitimacy: A Theory with Special Reference to Southeast Asia,” in Schmidt et al. (fn. 7).

¹⁰ See the following essays in Fox, Jonathan, ed., The Challenge of Rural Democratization: Perspectives from Latin America and the Philippines (London: Frank Cass, 1990)Google Scholar (also published as Journal of Development Studies 26 [July 1990]Google Scholar): Candido Grzybowski, “Rural Workers and Democratization in Brazil”; Francisco Lara, Jr., and Horacio Morales, Jr., “The Peasant Movement and the Challenge of Democratization in the Philippines”; and Leon Zamosc, “The Political Crisis and the Prospects for Rural Democracy in Colombia.”

¹¹ Revolutions are the most obvious examples of political processes that can sweep away clientelistic systems, but they are rarely followed by respect for associational autonomy. New webs of clientelism can emerge in their wake, especially in rural areas. See, e.g., Oi, Jean C., “Communism and Clientelism: Rural Politics in China,” World Politics 37 (January 1985)CrossRefGoogle Scholar. On the tensions over associational autonomy between peasant movements and left-wing political parties that claim to represent them in Latin America, see Fox, Jonathan, “New Terrain for Rural Politics” Report on the Americas 25, no. 5 (1992)CrossRefGoogle Scholar.

¹² See O'Donnell, Guillermo and Schmitter, Philippe C., Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies (Baltimore: Johns Hopkins University Press, 1986)Google Scholar; and Przeworski, Adam, Democracy and the Market (Cambridge: Cambridge University Press, 1991)CrossRefGoogle Scholar.

¹³ Note that the political terms of access to social entitlements are distinct from the levels of benefits, which are sometimes referred to as “social citizenship” rights. The argument here does not attempt to explain the determinants of the levels of material entitlements; they are logically and historically distinct from regime type in general and from the right to associational autonomy in particular. Democratic regimes, for example, may offer access to a narrow range of social rights without attaching political conditions (food stamps in the U.S.), whereas authoritarian regimes may offer a broad range of material entitlements in exchange for deference (as in communist and populist regimes). Analysts of the construction of rights have tended to focus either on electoral enfranchisement or on the extension of social welfare benefits, but not on the political terms of access to the latter. See, e.g., Barbalet, J. M., Citizenship (Minneapolis: University of Minnesota Press, 1988)Google Scholar.

¹⁴ Note that this argument does not attempt to account for the emergence of each of these three actors but rather shows how certain patterns of interaction among them can explain the construction of respect for associational autonomy. On cycles of social mobilization and reform, see Piven, Frances Fox and Cloward, Richard, Poor People's Movements: How They Succeed and Why They Fail (New York: Pantheon, 1977)Google Scholar; and Tarrow, Sidney, Struggle, Politics and Reform: Collective Action, Social Movements and Cycles of Protest, Occasional Paper no. 21, Western Societies Program (Ithaca, N.Y.: Center for International Studies, Cornell University, 1989)Google Scholar. To frame the issue of associational autonomy as broadly as possible, this discussion does not detail the diverse range of repertoires of action and forms of representation among poor people's movements.

¹⁵ For the purposes of this discussion, the trigger for division within authoritarian regimes is considered here to be contingent. Possible causes include economic crisis, international pressures, military defeat, and the “excessive” use of repression.

¹⁶ When reformists deliberately encourage social mobilization to offset hard-line authoritarian rulers, they can be said to be pursuing a “sandwich strategy” for political change. See Fox (fn. 3). Other examples include the U.S. federal government's antipoverty and civil rights efforts in the early 1960s, Colombian agrarian reform policy in the late 1960s, and Gorbachev's glasnost policies of the late 1980s.

¹⁷ Note that in this argument, mass mobilization alone cannot win citizenship rights. If authoritarian elites remain united, they can simply respond with coercion rather than concessions. Even when reformists are present within the state, they can lose; cycles of bargaining may well fail to build democratic rights. If hard-liners prevail, they will repress social movements and purge reformists from the state (as in the downward spiral that followed El Salvador's 1979 reformist coup).

¹⁸ In Mexico, for example, citizen's groups from the north have had much greater success at winning official respect than in the much poorer, largely indigenous southern states. Even within the south, results vary greatly across bargaining arenas; winning access to social programs is much easier than ending impunity for violent officials. Similarly, even in relatively democratic Brazil, official respect for human rights varies directly by the victims' race and region.

¹⁹ Though analytically distinct, these categories often overlap in practice.

²⁰ Seen from the receiving end, the differences between the withdrawal of benefits and the threat of coercion are quite significant. Both can discourage autonomous collective action, but only one is potentially permanent in its effects.

²¹ The term pluralism refers here to respect for associational autonomy rather than to the political system as a whole. Access to social programs can be considered pluralistic when it is not conditioned on political subordination.

²² Taiwan is a case where political attitudes and opportunities are changing quickly, showing how political action can undermine clientelism. Vote buying is still pervasive in Taiwan, and traditional norms of gratitude used to be sufficient to produce compliance. In the last several years, however, partly as the result of effective civic education campaigns, increasing numbers of voters accept the money and only comply symbolically. According to a recent study by the Center for Policy Studies at Sun Yat-Sen University, 44.8% of the population of Taiwan's second largest city were given money for their vote, but only 12.7% of them said they would actually vote for the candidate who bribed them; see Herr, Robin, “A Call for Independence in Taiwan,' Christian Science Monitor, March 5, 1993Google Scholar. Reportedly, one member of the family may return the favor with a vote for the ruling party while the rest feel free to vote their preference.

²³ In Colombia, for example, an apparently small change in ballot procedures after 1990 significantly weakened clientelistic bosses. Until then separate ballots were cast for each party, allowing bosses to check how people were going to vote while they waited in line. After 1990 the system was changed to a single ballot, which greatly increased ballot secrecy. Even where individual ballots might be secret, however, communities that are united in voting for the democratic opposition still reveal their dissent to authoritarian elites. In the key 1987 congressional elections in the Philippines, for example, government military units regularly assembled farmworkers to threaten them with reprisals should their villages vote for pro-land-reform candidates (author's field interviews).

²⁴ Classic individualistic ideas of citizenship may be inappropriate for non-Western social actors. For a critique of the imposition of foreign notions of citizenship on indigenous societies, see Silvia Rivera Cusicanqui, “Liberal Democracy and Ayllu Democracy: The Case of Northern Potosi, Bolivia,” in Fox (fn. 10). In part for this reason, this study analyzes associational autonomy in terms of the state's respect for ethnic and community-based groups rather than in terms of individual members of those communities.

²⁵ New historical research stresses that the coverage of Mexico's well-known corporatist organizations was partial rather than complete; see Rubin, Jeffrey W., “Popular Mobilization and the Myth of State Corporatism,” in Foweraker, Joe and Craig, Ann, eds., Popular Movements and Political Change in Mexico (Boulder, Colo.: Lynne Rienner, 1990)Google Scholar. On rural bosses, see Bartra, Roger, ed., Caciquismo y poder politico en el México rural (Mexico City: Siglo XXI, 1975)Google Scholar; and Fox (fn. 3). On elite political dientelism, see Camp, Roderic A., “Camarillas in Mexican Politics: The Case of the Salinas Cabinet,” Mexican Studies/Estudios Mexicanos 6 (Winter 1990)CrossRefGoogle Scholar; Centeno, Miguel Angel and Maxfield, Sylvia, “The Marriage of Finance and Order: Changes in the Mexican Political Elite,” Journal of Latin American Studies 24 (February 1992)CrossRefGoogle Scholar; Purcell, Susan Kaufman and Purcell, John F. H., “State and Society in Mexico: Must a Stable Polity Be Institutionalized?” World Politics 32 (January 1980)CrossRefGoogle Scholar; and Smith, Peter, Labyrinths of Power: Political Recruitment in Twentieth-Century Mexico (Princeton: Princeton University Press, 1979)Google Scholar.

²⁶ This dichotomy of “official versus independent” social movements was especially pronounced in the 1970s and 1980s, as collective resistance to the state grew. By the 1990s social movements increasingly stressed autonomy from political parties in general, since contestational “independence” had often involved subordination to opposition parties. See Fox, Jonathan and Gordillo, Gustavo, “Between State and Market: The Campesinos' Quest for Autonomy,” in Cornelius, Wayne A., Gentleman, Judith, and Smith, Peter H., eds., Mexico's Alternative Political Futures (La Jolla: UCSD, Center for U.S.-Mexican Studies, 1989)Google Scholar; Foweraker and Craig (fn. 25); and Hellman, Judith Adler “The Study of New Social Movements in Latin America and the Question of Autonomy,” in Escobar, Arturo and Alvarez, Sonia, eds., The Maying of Social Movements in Latin America (Boulder, Colo.: Westview Press, 1992)Google Scholar.

²⁷ Distributing patronage through semiclientelistic means (i.e., nonenforceable deals) can also make fraudulent electoral outcomes seem more politically plausible to the electorate, since even individuals who accepted the incentives but voted their conscience anyway cannot be sure how many others did the same. This in turn undermines the potential for collective action in defense of clean elections.

²⁸ Although similar cycles of social mobilization can also be found in Mexican urban politics, their emergence in remote rural areas as well shows that the erosion of clientelism can be encouraged by strategic political action and is not driven exclusively by secular socioeconomic trends such as urbanization. On urban politics and the poor, see Vivienne Bennett, “The Evolution of Urban Popular Movements in Mexico between 1968 and 1988,” in Escobar and Alvarez (fn. 26); Cornelius, Wayne, Politics and the Migrant Poor in Mexico City (Stanford, Calif.: Stanford University Press, 1975)Google Scholar; Eckstein, Susan, The Poverty of Revolution, 2d ed. (Princeton: Princeton University Press, 1988)Google Scholar; Fox and Hernandez (fn. 4); and Ward, Peter, Welfare Politics in Mexico (London: Allen and Unwin, 1986)Google Scholar.

²⁹ According to one PIDER official: “If participation is stimulated too much it gets out of PIDER'S control and brings political problems. It becomes a political problem for PIDER when it begins to break up or threaten commercial interests”; quoted in Grindle, Merilee, “Official Interpretations of Rural Underdevelopment: Mexico in the 1970's” Working Papers in U.S.-Mexican Studies 20 (1981), 43Google Scholar. See also Cernea, Michael, A Social Methodology for Community Participation in Local Investment: The Experience of Mexico's PIDER Program, World Bank Staff Working Paper no. 598 (Washington D.C.: World Bank, August 1983)Google Scholar; and Lindheim, Daniel Noah, “Regional Development and Deliberate Social Change: Integrated Rural Development in Mexico” (Ph.D. diss., University of California, Berkeley, 1986)Google Scholar.

³⁰ In one notable case, the PIDER Brigade was so effective at encouraging autonomous mobilization that the governor expelled it from the state; see Fox (fn. 4).

³¹ Notably, Carlos Salinas de Gortari's dissertation concluded that because of the ineffectiveness and corruption of the conventional state apparatus, the regime lost the political payoff associated with increased antipoverty spending: “The State [must] rely on a corps of leaders of local development programs who will be attentive to the problems encountered in the delivery of development projects to targeted communities. … They must lead, not in the hierarchic sense of demanding obedience, but in the sense of coordinating and orienting a decision-making process in which the members of affected communities participate.” See de Gortari, Carlos Salinas, Political Participation, Public Investment, and Support for the System: A Comparative Study of Rural Communities in Mexico, Research Report Series 35 (La Jolla: UCSD, Center for U.S.-Mexican Studies, 1982), 41–42Google Scholar.

³² This section draws from the more detailed discussion in Fox (fn. 3).

³³ For the geographic distribution of autonomous Food Councils, see Fox (fn. 3).

³⁴ Concertación in the Mexican context has been translated in a variety of ways, ranging from “social dialogue” to “corporative agreements.”

³⁵ This shift was especially clear in the area of urban consumer food subsidies, which supported a wide range of staples until the mid-1980s. Then the government began limiting food subsidies to tortillas and milk for means-tested low-income city dwellers.

³⁶ Solidarity proclaimed that its “new dynamic … breaks with bureaucratic atavism and administrative rigidity. Public servants increasingly share a vocation for dialogue, agreement, concertacidn and direct, coresponsible work with the citizenry, which also assumes an increasingly active and leading role in the actions intended to improve their standard of living.” See Rojas, Carlos et al., Solidaridad a Debate (Mexico City: El Nacional, 1991), 23Google Scholar.

³⁷ For comprehensive overviews of PRONASOL politics, see Dresser, Denise, Neopopulist Solutions to Neoliberal Problems: Mexico's National Solidarity Program, Current Issues Brief no. 3 (La Jolla: UCSD, Center for U.S.-Mexican Studies, 1991)Google Scholar; idem, “Pronasol: los dilemas de la gobernabilidad,” El Cotidiano 49 (July-August 1992)Google Scholar. See also the Los Angeles Times poll, October 22, 1991. Solidarity spending rose sharply just before the 1991 midterm elections and was widely credited with helping to revive the official party's electoral fortunes, although its impact is difficult to disentangle from reduced inflation and the beginnings of economic growth. For journalistic accounts of direct electoral use of Solidarity funding, see Rio, Pascal Beltrán del, “Solidaridad, oxígeno para el PRI, en el rescate de votos,” Proceso 718 (August 6, 1990)Google Scholar; idem, “El memorandum de Pichardo, prueba de que el Pronasol es para servir al PRI,” Proceso 730 (October 29, 1990)Google Scholar; Correa, Guillermo, “EL PRONASOL, que nacío como esperanza, ha generado corrupción y protestas,” Proceso 727 (October 8, 1990)Google Scholar; and Leyva, Ciro Gómez, “Solidaridad gratuita en todas las pantallas,” Este País 7 (October 1991)Google Scholar.

³⁸ Golden, Tim, “Point of Attack for Mexico's Retooled Party Machine: The Leftist Stronghold,” New York Times, July 12, 1992Google Scholar. See also Cantú, Jesús, “Solidaridad, además de electorero, se manejó en Michoacán coercitivamente,” Proceso 819 (July 13, 1992)Google Scholar. In addition, the governor's election campaign expenses reportedly topped U.S. $30 million, almost $80 per vote officially cast for the PRI; see Chávez, Elías, “Michoacán: cada voto costó 239,188 pesos; cada voto del PRD costó 6,916 pesos,” Proceso 821 (July 27, 1992)Google Scholar. For a state-level statistical analysis of electoral targeting, see Juan Molinar and Jeffrey Weldon, “Electoral Determinants and Consequences of National Solidarity,” in Wayne Cornelius, Ann Craig, and Jonathan Fox, eds., Transforming State-Society Relations in Mexico: The National Solidarity Strategy (La Jolla: UCSD, Center for U.S.-Mexican Studies, forthcoming).

³⁹ So far, much of the debate surrounding Solidarity's political character has been based more on ideological polemic than on empirical evidence, but the most plausible hypothesis is that, on balance, most of the electorally targeted spending was probably delivered through semiclientelist means. The basis for this general proposition, which will not be tested here, is that since most Solidarity programs were delivered from outside the community, they lacked the official party's once-powerful capacity to monitor and punish noncompliance at the individual level. As long as fraud remains an option for the regime, however, it can reduce the importance of individual compliance. Recently, most electoral manipulation appears to have occurred before election day. For example, over one hundred thousand likely opposition voters were allegedly “shaved” from the registration rolls in Michoacán, especially in urban PRD strongholds. See Bardacke, Ted, “The Lion Learns New Tricks,” El Financiero International, July 20, 1992Google Scholar. It must also be noted that the regime still uses sticks as well as carrots; selective political violence against the Left also continues with impunity. The PRD reported that 230 of its members had been killed for political reasons since 1988 (La Jornada, May 11, 1993). See also Watch, America's, Human Rights in Mexico: A Policy of Impunity (New York: America's Watch, 1990)Google Scholar; idem, “Unceasing Abuses: Human Rights in Mexico One Year after the Introduction of Reform” (New York: America's Watch, September 1991)Google Scholar; and PRD Human Rights Commission, The Political Violence in Mexico: A Human Rights Affair (Mexico City: Human Rights Commission Parliamentary Group, April 1992)Google Scholar.

⁴⁰ In addition to provision of public goods, this wide range of programs also included many with benefits that were much more divisible and therefore more vulnerable to local elite diversion (for example, soft loans for peasant producers). This distinction is central to the analysis of targeting. See Tendler, Judith, Rural Projects through Urban Eyes: An Interpretation of the World Band's New-Style Rural Development Projects Working Paper 532 (Washington, D.C.: World Bank, 1982)Google Scholar.

⁴¹ Solidarity spending is allocated at the president's discretion, as distinct from Mexico's official revenue-sharing, which is allocated according to technical formulas. See John Bailey, “Centralism and Political Change in Mexico: The Case of National Solidarity,” in Cornelius, Craig, and Fox (fn. 38).

⁴² See Cantú (fn. 38). On the PRD, see Jonathan Fox and Julio Moguel, “Pluralism and Anti-Poverty Policy in Mexico: The Experience of Left Opposition Municipal Governments, ” in Victoria Rodríguez and Peter Ward, eds., Opposition Government in Mexico: Past Experiences and Future Opportunities (Albuquerque: University of New Mexico Press, forthcoming); and on the National Action Party state government in Baja California Norte, see de Alba, Gerardo Albarrán, “Con Pronasol, la necesidad de la gente se usa electoralmente: Ruffo,” Proceso, no. 829, September 21, 1992Google Scholar. The pattern was not consistent, since some opposition municipalities of both Right and Left managed to bargain for control over PRONASOL resources.

⁴³ See Dresser (fn. 37, 1991 and 1992); and Fernández, Jorge, “El PRI ante su propia transitión,” Unomásuno, November 7, 1991Google Scholar.

⁴⁴ In this context, the president reportedly once told the following to a longtime friend, a historic radical leader of the urban popular movement: “You were my teacher: everywhere I go I leave a base of support.” At a meeting of five hundred representatives of five thousand urban Solidarity committees, for example, the president called for the creation of a national neighborhood network outside the ruling party. See Lomas, Emilio, “La democracia ya no es de las cúpulas, afirma Salinas,” La Jornada, September 13, 1991Google Scholar; idem, “Salinas: nueva relación Estado-sociedad civil,” La Jornada, September 15, 1991. Midway through the Salinas presidency, his advisers secretly debated three options—use the committees to build a new reform party, fold them openly into the official party, or keep them relatively nonpartisan. The latter position reportedly won.

⁴⁵ The CDP won the mayoralty in 1992 without allying with either the PAN or the PRD, leading some observers to suggest its access to Solidarity resources moderated its approach to national politics. See Haber, Paul, “Collective Dissent in Mexico: The Political Outcome of Contemporary Urban Popular Movements” (Ph.D. diss., Columbia University, 1992)Google Scholar; idem, “Political Change in Durango: The Role of National Solidarity,” in Cornelius, Craig, and Fox (fn. 38).

⁴⁶ According to INI'S annual report, its budget for the 1991 fiscal year was U.S. $140 million. President Salinas named Arturo Warman, PIDER veteran and one of Mexico's most distinguished anthropologists, as INI director. Solidarity's overall coordinator, Carlos Rojas, had worked for INI in Veracruz.

⁴⁷ Many Mexican anthropologists see indigenous voting patterns in terms of local “shortterm considerations that have nothing to do with political programs that propose alternative models for the future. The vote is seen more as a resource for here and now, [for] finishing a road, building a school or a drinking water system; [the] small benefits that help to resolve ancestral problems which shape their daily lives”; see Bonfil, Guillermo, Mexico Profundo: Una civilización negada (Mexico City: Grijalbo/CONACULT, 1990), iiiGoogle Scholar. Indeed, parties are not present in most indigenous regions (though this began to change after 1988). The analytical problem is to distinguish cause from effect. If opposition political parties fail to champion indigenous rights, then isolated villagers have few incentives to take the serious risks inherent in partisan collective action, especially when it so often appears unviable. As voters, they may not lack national political preferences as much as they lack meaningful national political choices.

⁴⁸ E.g., see Flores, Juan, “Proyectos de Etnodesarrollo = los ricos más ricos y los pobres mas pobres,” Etnias 2 (January 1991)Google Scholar; and González, Alvaro, Valdivia, Teresa, and Rees, Martha, “Evaluación de los Programas Agrícolas del INI: Chiapas, Puebla y Oaxaca” (Paper presented at the Society for Applied Anthropology, Oaxaca, April 1987)Google Scholar.

⁴⁹ INI'S other new initiatives during the Salinas administration included a human rights program that released over six thousand indigenous prisoners, as well as promotion of a constitutional amendment that for the first time officially recognized Mexico as a multicultural society. As of mid-1993 human rights groups campaigned for the freedom of several thousand indigenous prisoners who remained in jail without due process.

⁵⁰ One INI official also stressed that his staff was different from that of most Solidarity programs because they were “usually not in any political party. It's very unusual that INI personnel are in the PRI—but they aren't in the [opposition] PRD either. They aren't people who are going to induce [i.e., manipulate] or condition.” He claimed that because they work in such remote regions, “they will work with existing organizations—they can't invent others.”

⁵¹ INI, 1990, 41–42; emphasis added. Participation is limited to policy implementation here. INI continued to reject the long-standing demands of indigenous groups for greater involvement in the policy process itself. Some of INI'S most reformist policymakers tried this in 1983, but they were quickly purged.

⁵² INI'S other main economic program was its support for coffee producers after the abrupt withdrawal of the Mexican Coffee Institute from the market. Two-thirds of coffee producers are indigenous, accounting for 30 percent of national production and one-third of coffee lands; see INI, Programa National de Desarrollo de los Pueblos Indigenas, 1991–1994 (Mexico City: INI, 1990), 17Google Scholar. INI'S coffee program involved both pluralistic relations with autonomous producer organizations and semiclientelist relations with iNi-sponsored local Solidarity committees. See Luis Hernández and Fernando Celis, “Solidarity and the New Campesino Movement: The Case of Coffee Production,” in Cornelius, Craig, and Fox (fn. 38).

⁵³ INI described the FRS'S in explicitly political terms: “The Funds are an innovative process [to] increase the participatory role of civil society in decision making and in the definition of policy, which reflects a change in state-society relations. The relationship of coresponsibility established between the government and the indigenous population implies a turnaround in the role of [government] institutions to avoid reproducing paternalistic and vertical attitudes that interfere with indigenous peoples' development”; see INI, “Manual de Operación de los Fondos Regionales de Solidaridad para el Desarrollo de los Pueblos Indígenas” (Unpublished manuscript, Mexico City, November 1991), 2Google Scholar. This statement was dropped from the version eventually published for mass distribution in 1993.

⁵⁴ INI also encouraged the FRS'S to go beyond economic support for production projects and become advocates for indigenous communities in the broader public investment allocation process; the effort was largely unsuccessful.

⁵⁵ The more consolidated Regional Funds were reportedly in Veracruz, Chiapas, and Oaxaca, while those in the Huasteca, Chihuahua, and the Yucatan peninsula did poorly. Tabasco was especially disastrous: the governor tried to impose a corrupt crony as local INI director, which provoked a mass protest movement, and then rejected any development aid that could possibly reach potential opposition sympathizers. Not coincidentally, the state PRD leader, Manuel López Obrador, had won a broad indigenous following during his tenure as local INI director in the early 1980s.

⁵⁶ Oaxaca is one of Mexico's poorest states and at least 44% of the state's population speaks one of the state's seventeen indigenous languages; Rivera, Rafael Blanco, “Oaxaca, 1980,” Cuadernos de Demografía lndígena (Mexico City: INI, Dirección de Investigación y Promoción Cultural, 1991)Google Scholar.

⁵⁷ These leadership councils were based in Jamiltepec, Miahuatlán, Huautla, Tlacolula, Guelatao, and Cuicatlan. It must be stressed that “consolidation” does not imply that all or even most member groups of an LC were representative grassroots groups. Five of Oaxaca's twenty LCS were not “test” cases because of the lack of autonomous indigenous producer organizations in those regions as of mid-1992. For details, see Jonathan Fox, “Targeting the Poorest: The Role of the National Indigenous Institute in Mexico's Solidarity Program,” in Cornelius, Craig, and Fox (fn. 38).

⁵⁸ Most CEPCO member groups are nonpartisan or operate within the PRI, although a few sympathize with the PRD. CEPCO'S main activity is buying, processing, and selling coffee, setting a floor price following the withdrawal of the state from the market in 1989, and representing about one-third of Oaxaca's small coffee producers. Much of the state government and the corporatist apparatus felt threatened by CEPCO'S success at providing an alternative. See Moguel, Julio and Aranda, Josefina, “La Coordinadora Estatal de Productores de Café de Oaxaca,” in Moguel, Julio, Botey, Carlota, and Hernández, Luis, eds., Autonomía y nuevos sujetos sociales en el desarrollo rural (Mexico City: Siglo XXI/CEHAM, 1992)Google Scholar.

⁵⁹ A robust notion of pluralism would go beyond this inclusion/exclusion dichotomy and involve some degree of proportional representation. Funds include groups ranging in size from tiny kinship groups to producer associations representing thousands of families, yet in most leadership councils each has the same vote. Some INI directors used their clientele as counterweights to keep more broadly representative groups in the minority. The Mazateca highlands leadership council led the first experiment in institutionalizing proportional representation in the leadership councils, weighting the number of assembly delegates according to the membership of each participating organization. The INI convened this process in an apparent effort to undermine the outspoken CEPco-affiliated leadership of the Mazateca region's LC and to strengthen the official corporatist group, but the independent coffee producers swept the elections.

⁶⁰ Recent indigenous movements have been most intense in Chiapas, Oaxaca, Hidalgo, Veracruz, and Guerrero. See Pineiros, Maria Consuelo Mejía and Sarmiento, Sergio, La lucha indígena: un reto a la ortodoxia (Mexico City: Siglo XXI, 1987)Google Scholar; Nagengast, Carole and Kearney, Michael, “Mixtec Ethnicity: Social Identity, Political Consciousness and Political Activism,” Latin American Research Review 25, no. 2 (1990)Google Scholar; Sarmiento, Sergio, “Movimiento indio y modernización,” Cuadernos Agrarios 2, new series (1991)Google Scholar; and the journals Etnias and Ojarasca (formerly México Indígena).

⁶¹ The rate at which presidents remove governors is an excellent indicator of the degree of intrastate conflict in Mexico. During the first three years of the Salinas administration, nine of the thirty-one governors had been forced to resign.

⁶² Rojas, Rosa, “Indigenas de Chiapas piden se libere a 3 funcionarios del INI,” La Jornada, March 21, 1992Google Scholar. Leaders of the Chiapas funds were also involved in the successful Xi 'Nich human rights protest march to Mexico City in early 1992. The national leader of Mexico's Independent Front of Indian Peoples (FIPI)—a frequent INI critic—confirmed that the Chiapas Regional Funds were remarkably open to independent groups. See Hernández, Margarito Xib Ruiz, “Todo indigenismo es lo mismo,” Ojarasca (February 1993)Google Scholar.

⁶³ See Pérez, Matilde, “El ejido es un sistema equitativo y eficaz: INI,” La Jornada, October 23, 1991Google Scholar. For a political analysis of the 1991 reform of Mexico's land-tenure system, see Jonathan Fox, “Political Change in Mexico's New Peasant Economy,” in Maria Lorena Cook, Kevin Middlebrook, and Juan Molinar Horcacitas, eds., The Politics of Economic Restructuring in Mexico (La Jolla: UCSD, Center for U.S.-Mexican Studies, forthcoming).

⁶⁴ As of mid-1993 the future of the Regional Solidarity Fund program was in doubt. The Social Development Ministry, which controls overall Solidarity funding, had frozen most of INI'S 1992 allocations for the Regional Solidarity Funds, blaming lagging repayment rates. Difficulties with repayment were not surprising, given the problems of profitability throughout the agricultural sector; but since the government was very flexible with much larger debts from other agricultural borrowers, such as owners of large coffee plantations or the buyers of privatized sugar mills, slow repayment rates alone were not an especially credible explanation for defunding the program, INI had been politically weakened by the transfer of its influential director to the newly created post of agrarian attorney general. This left INI'S Regional Funds vulnerable to powerful antipluralist elements within the Social Development Ministry itself, which wanted to take project decision-making power away from the leadership councils.

⁶⁵ Distributive reform thus became political reform, as Przeworski defines it: “a modification of the organization of conflicts that alters the prior probabilities of realizing group interests given their resources.” See Przeworski, Adam, “Some Problems in the Study of the Transition to Democracy,” in O'Donnell, Guillermo, Schmitter, Philippe C., and Whitehead, Laurence, eds., Transition from Authoritarian Rule: Comparative Perspectives (Baltimore: Johns Hopkins University Press, 1986), 58Google Scholar.

⁶⁶ See Cornelius, Wayne, “Political Liberalization in an Authoritarian Regime: Mexico, 1976–1985,” in Gentleman, Judith, ed., Mexican Politics in Transition (Boulder, Colo.: Westview Press, 1987)Google Scholar; and Kevin Middlebrook, “Political Liberalization in an Authoritarian Regime: The Case of Mexico,” in O'Donnell, Schmitter, and Whitehead (fn. 65). On electoral change in the 1980s, see Cornelius, Gentleman, and Smith (fn. 26). For a comparative discussion of Mexico's transition that explores the distinction between the formal regime and the actual political system, see Camou, Antonio, “Once tesis sobre la ‘transición’ mexicana: Gobernabilidad y democracia,” Nexos 55 (February 1992)Google Scholar.

⁶⁷ The regime was able to manage this uncertain process largely because the most contested races—for governors and mayors—were staggered so that the ruling party faced only one or two difficult states at a time.

⁶⁸ On the numerous irregularities, including widespread reports of attempts to condition Solidarity funding on PRI votes, see the election observer report by the Convergencia de Organismos Civiles por la Democracia, “Informe de observación electoral,” Perfil de la Jornada, August 16, 1992.

⁶⁹ The regime's willingness to cede legitimacy to some autonomous citizens' groups while continuing to manipulate elections also sharpened divisions within the left-leaning electoral opposition. When the wounds of the 1988 electoral conflict were still fresh, the PRD harshly condemned social organizations that bargained for Solidarity funds, asserting that they were implicitly recognizing the president's legitimacy. The PRD'S stance later softened, but its relationship with important social movements was damaged. See Dresser (fn. 37, 1991); and Haber (fn. 45).

⁷⁰ So far, two scenarios predicted by Cornelius and his colleagues are combining: “modernization of authoritarianism with selective pluralism” and “limited power sharing,” along the lines of the Indian Congress Party model. See Cornelius, Gentleman, and Smith (fn. 26); and Cornelius, Wayne and Craig, Ann, The Mexican Political System in Transition (La Jolla: UCSD, Center for U.S.-Mexican Studies, 1991), 118–19Google Scholar.

⁷¹ As Pye put it, “We need finer shades of typologies of political systems between the classical polar opposites of authoritarian and democratic. In the wake of the crisis of authoritarianism we can expect a wide variety of systems that will become part authoritarian and part free and that will fall far short of any reasonable definitions of democracy.” See Pye, Lucian W., “Political Science and the Crisis of Authoritarianism,” American Political Science Review 84 (March 1990), 13CrossRefGoogle Scholar.

⁷² See Terry Lynn Karl (fn. 1) on “electoralism”; Schmitter, Philippe and Karl, Terry Lynn, “What Democracy Is … and Is Not,” Journal of Democracy 2 (Summer 1991)CrossRefGoogle Scholar; Rosenberg, Tina, “Beyond Elections,” Foreign Policy 84 (Fall 1991)Google Scholar; and Hermet, Guy, Rose, Richard, and Rouquié, Alain, eds., Elections without Choice (New York: John Wiley, 1978)CrossRefGoogle Scholar. Stable electoral competition is sometimes confused with political democracy; see, e.g., Higley, John and Gunther, Richard, eds., Elites and Democratic Consolidation in Latin America and Southern Europe (Cambridge: Cambridge University Press, 1992)Google Scholar. Thus Mexico's ruling party holds the world's record for stability, having presided over electoral presidential successions since 1929. Some analysts fall into the opposite trap, assuming that unfair elections are politically meaningless exercises. Note, for example, the surprise military split and subsequent civic uprising following Philippine president Marcos's fraudulent “snap” elections in 1986. See also the debate over the relevance of El Salvador's sharply constrained wartime elections of the mid1980s in Herman, Edward and Brodhead, Frank, Demonstration Elections (Boston: South End Press, 1984)Google Scholar; Karl, Terry Lynn, “Imposing Consent? Electoralism vs. Democratization in El Salvador,” in Drake, Paul and Silva, Eduardo, eds., Elections and Democratization in Latin America, 1980–1985 (La Jolla: UCSD-CILAS, 1986)Google Scholar.

⁷³ On social investment funds, see Graham, Carol, “The APRA Government and the Urban Poor: The PIAT Programme in Lima's Pueblos Jóvenes,” Journal of Latin American Studies 23 (February 1991)CrossRefGoogle Scholar; idem, “The Politics of Protecting the Poor during Adjustment: Bolivia's Emergency Social Fund,” World Development 20 (September 1992)Google Scholar; idem, “Mexico's National Solidarity Program in Comparative Perspective: Demand-Based Poverty Alleviation Programs in Latin America, Africa, and Eastern Europe,” in Cornelius, Craig, and Fox (fn. 38). One could argue that opposition state governments in India carried out comparable programs earlier. See Kohli, Atul, The State and Poverty in India: The Politics of Reform (Cambridge: Cambridge University Press, 1987)CrossRefGoogle Scholar; and Echeverrí-Gent, John, “Public Participation and Poverty Alleviation: The Experience of Reform Communists in India's West Bengal,” World Development 20 (October 1992)CrossRefGoogle Scholar.

⁷⁴ See Herr (fn. 23); and Shenon, Philip, “It's Business as Usual in Thailand (Votes for Sale),” New York Times, March 18, 1992Google Scholar; as well as Scott (fn. 8, 1992).

⁷⁵ Thus, in 1988 Colombia appeared to take a major step toward greater pluralism by permitting citizens to elect their mayors for the first time. But once elected, many opposition mayors were assassinated by state-sanctioned death squads. See Leah Carroll, “Repression and the Limits to Rural Democratization: The Experience of Leftist County Executives in Colombia, 1988–1990” (Paper presented at the Latin American Studies Association, April 1991).

⁷⁶ By contrast, some analysts consider elections to be democratic if the bulk of the population participates; see, e.g., Huntington, Samuel P., The Third Wave: Democratization in the Late Twentieth Century (Norman: University of Oklahoma Press, 1991)Google Scholar.

⁷⁷ See Loveman, Brian, “Political Participation and Rural Labor in Chile,” in Seligson, Mitchell A. and Booth, John A., eds., Political Participation in Latin America, vol. 2, Politics and the Poor (New York: Holmes and Meier, 1979)Google Scholar.

⁷⁸ See Cotrim, Fernando Da Silveira, A geografia do voto no Brasil: Eleições de 1989 (Rio de Janeiro: IBASE, 1990)Google Scholar. On the persistent clout ot the traditional political class, see Hagopian (fn. 7, 1990 and 1992).

⁷⁹ See López, Arturo et al., Geografía de las elecciones presidenciales de México, 1988 (Mexico City: Fundación Arturo Rosenblueth, 1989)Google Scholar.

⁸⁰ Because these dimensions evolve along such different paths, Schmitter suggests that it may be useful to understand democracy as a “composite of ‘partial regimes,’ each of which [is] institutionalized around distinctive sites for the representation of social groups.” See Schmitter, Philippe, “The Consolidation of Democracy and Representation of Social Groups,” American Behavioral Scientist 35 (March-June 1992), 427CrossRefGoogle Scholar.

⁸¹ For general discussions of measurement issues, see Inkeles, Alex, ed., On Measuring Democracy (New Brunswick, N.J.: Transaction Publishers, 1991)Google Scholar. For new studies of degrees of civilian control over the military, see Pion-Berlin, David, “Military Autonomy and Emerging Democracies in South America,” Comparative Politics 25 (October 1992)CrossRefGoogle Scholar; and Zaverucha, Jorge, “The Degree of Military Political Autonomy during the Spanish, Argentine, and Brazilian Transitions,” Journal of Latin American Studies 25 (May 1993)CrossRefGoogle Scholar.