From Populism to Neoliberalism:
Labor Unions and Market Reforms in Latin America

M. Victoria Murillo *

Tables

Appendix

A Tale of Three Carloses

Carlos Salinas, Carlos Andrés Pérez, and Carlos Menem were inaugurated between December 1988 and July 1989 as presidents of Mexico, Venezuela, and Argentina, respectively. They not only shared the timing of their administrations ¹ but also the task of turning populist labor parties toward neoliberalism. All three had been the candidates of populist labor-based parties that had advanced protectionism and state intervention in the postwar period. Once in office and facing tremendous fiscal crises and capital flight, all three presidents reduced state intervention and opened the economies of their countries. This shift in development strategy not only was the most important policy turnaround of the postwar era in all three Latin American countries, but it also moved their labor-based parties away from the policies upon [End Page 135] which their historic relationships with long-term union allies had been built.

Despite the parallel convergence of labor-based populism turning into neoliberalism and the common challenge faced by unions in Argentina, Mexico, and Venezuela, union responses to neoliberal reforms were diverse. Union-government interactions varied across these countries and across sectors within the same country. For instance, the Mexican Workers' Confederation subordinated to Salinas's policies, endorsing them in corporatist pacts and even campaigning actively for the North American Free Trade Agreement (NAFTA). To the contrary, the Venezuelan Workers' Confederation opposed Perez's reforms by organizing the first economic general strikes in the history of Venezuela. Argentine teachers unsuccessfully resisted the decentralization of schools; their militancy accounted for more than a third of the total strikes in the two years before the government, with no union input, finally implemented the reform. In Mexico, however, teachers successfully opposed key pieces of the decentralization process and limited its scope. Meanwhile, Mexican telephone workers negotiated with their government to support the privatization of their company in return for job stability, handsome social benefits, and easy financing for the purchase of almost 5 percent of the shares in the company.

Such variation in union-government interactions in response to market-oriented reforms has considerable political importance. First, unions can organize support for or opposition to the reforms, thus changing the costs of reform for policymakers. Their actions can affect the feasibility, design, and implementation of reforms, as shown by the delays in the reform of labor market regulations in all three countries. Unions should therefore be included in the analysis of the politics of economic reforms.

Second, unions can have a direct impact on the governance of administrations led by labor-based political parties. Unions not only have organized the core constituency of labor-based parties and provided them with political machines for electoral campaigns, but they also have shared a long-term partisan identity with governing politicians. As a result, the interaction between unions and labor-based parties implementing market-oriented reforms can influence the future of the electoral coalition that brought these parties to power. Argentina, Mexico, and Venezuela are key cases in that regard because the alliance between unions and labor-based parties had shaped the national party system fifty years earlier. ² [End Page 136] Thus, a theory that accounts for the cross-sectoral and cross-national variations in union-government interactions in these countries can help solve the puzzle of the transition from closed to open economies in Latin America. ³ Moreover, an explanation of such variation has broad comparative implications for our understanding of similar stories in other regions of the world where the pressure of globalization turned labor-parties toward policies that challenged their long-standing agreements with labor allies.

Theories of union politics suggest that union behavior varies according to either differences in national-level institutions or interests of economic sectors with regard to market policies and international integration. Yet, in Argentina, Mexico, and Venezuela, the interactions between unions and labor-based governments over market-oriented reforms cut across national borders and sectoral cleavages. Explanations based on either economic interest or national institutions are thus insufficient to understand such patterns of interaction. I argue instead that the incentives created by partisan loyalties, partisan competition, and union competition explain these interactions. Partisan loyalty results from the long-term affiliation of a union with a political party. Partisan competition is the struggle for control of the unions among union leaders affiliated with different political parties. Union competition is the rivalry among labor organizations for the representation of the same workers, which can take place in diverse national and sectoral contexts. Loyalty derived from a long-term affiliation with the incumbent party facilitates collaboration between labor unions and the government. Yet, if loyal union leaders are afraid of being replaced by activists affiliated with the opposition parties because of partisan competition, their incentives for militancy increase to show their responsiveness to the rank and file hurt by market reforms. Union competition for the representation of the same workers makes coordination more difficult thereby weakening unions and making them less likely to obtain concessions from the government despite their partisan loyalty. I use the empirical evidence from eighteen cases, including national confederations and individual unions in five economic sectors in Argentina, Mexico, and Venezuela to test this theory.

The article is divided into six sections. The first section introduces the policy shift of labor-based parties from populism to neoliberalism [End Page 137] in Argentina, Mexico, and Venezuela. The second examines the relevant literature upon which I build my analysis. The third describes the patterns of union-government interaction in the context of market-oriented reforms. The fourth presents my hypothesis for explaining those patterns. The fifth section reports the test of my hypothesis on the case studies, and the last offers some conclusions.

From Populism to Neoliberalism

On October 17, 1945, a multitude of workers organized by their unions marched into downtown Buenos Aires to demand the release of imprisoned Colonel Juan Perón and to defend the social legislation he had implemented as secretary of labor. In the elections of February 1946, the Labor Party--organized by the leaders of those same unions--led the political coalition that elected Perón to the presidency. Most workers voted for Perón and taught their children to be Peronist, creating one of the strongest partisan loyalties in Latin America. After all, Perón and the Peronist unions had changed their lives by providing better wages and labor benefits, social security, and even paid vacations at union resorts. Since labor benefits compensated them for previous frustration in dealing with indifferent employers and hostile governments, the unions identified with Peronism and served as Peronist political machines. Peronism, which became the symbol of workers' rights, promoted import substitution industrialization and state intervention as development strategies. ⁴ It also turned unions into key players in the Argentine political system.

In 1989 a Peronist candidate, Carlos Menem, won the presidential elections. During his populist electoral campaign, Menem promised wage hikes and social justice, and he threatened not to pay the external debt. After his inauguration, however, he delivered austerity, followed by trade liberalization, privatization, and adjustment of the public sector. As state intervention had done forty years earlier, the retreat of the state and market-oriented reforms reshaped state-society relations. The national labor confederation, the Peronist-dominated General Confederation of Labor (CGT), although surprised by the policy turnaround, accepted market-oriented reforms and reduced the number of general [End Page 138] strikes from thirteen against the previous administration of the Civic Radical Union (UCR) (1983-89) to only one under the first Menem administration (1989-95). Peronist unions, though, were able to negotiate concessions on the reforms of social security, labor legislation, and privatization.

In Mexico during the same period, the story of the policy turnaround of labor parties in government was similar to that in Argentina. Unions entered the political arena during the Mexican Revolution. They first formed six "Red Battalions" that fought with the army of President Venustiano Carranza in 1915. In 1919 they organized a Labor Party, which supported the elections of Presidents Alvaro Obregón and Plutarco E. Calles. In return, unions obtained political appointments and favorable labor legislation to compensate for their weakness in collective bargaining with private employers. In 1936 President Lázaro Cárdenas promoted the organization of the Mexican Workers Confederation (CTM). In 1938 he founded the Party of the Mexican Revolution (PRM) and integrated labor into its functional structure. Unions became political machines for the PRM, which would become the Institutional Revolutionary Party (PRI), and were included in the party structure. Unions also gave Cárdenas a suitable excuse--in the form of a labor conflict--for the nationalization of oil in 1938. In return, workers received social benefits, and union leaders gained political influence they could use in the industrial arena. The PRI also implemented policies of import substitution industrialization while using the state as the main instrument for economic development.

Half a century later, however, PRI labor leaders witnessed a PRI president, Carlos Salinas, become the champion of economic liberalization, state shrinkage, and market-oriented reforms. The main national confederation, the PRI-dominated CTM, explicitly supported Salinas's stabilization plan and structural reforms both in the annual corporatist pacts signed with government and business representatives and in an agreement to increase labor productivity signed in 1992. Moreover, it campaigned for Salinas's proposal to integrate Mexico into NAFTA. In spite of the CTM's acquiescence, the administration ignored most of its demands even while paying lip service to the alliance between the PRI and the labor movement.

During the same period, a similar story of policy turnaround was being written in Venezuela, where Democratic Action (AD) had traditionally been a champion of democracy and nationalism as well as of union organization. During its first administration (1946-48), AD promoted an upsurge in union organization and social mobilization. [End Page 139] Democratic Action union leaders promoted the founding of the Venezuelan Workers Confederation (CTV), and its labor bureau integrated union leaders into the party structure. Unions provided AD with political machines while channeling workers' loyalty and supporting AD development policies based on state intervention and import substitution industrialization. In return, AD administrations provided workers with social and labor benefits, and union leaders with political influence.

In 1988 AD union leaders endorsed the populist Carlos Andrés Pérez in the primaries of AD and in his campaign for the presidency. Pérez extended state intervention, established a minimum wage, and made dismissals more difficult during his first administration in the 1970s. After his inauguration in February 1989, however, he surprised foes and followers with his announcement of the "Great Turnaround." The Great Turnaround included trade liberalization, macroeconomic adjustment, and structural reforms of the state. The response of Perez's union allies to his reforms differed from that encountered by his Argentinean and Mexican counterparts. The AD-controlled CTV responded to his policy shift by organizing the first economic general strike in Venezuelan history, followed by a series of demonstrations and other strikes that boycotted most of Pérez's reforms in the labor and social sectors. Opposition by the CTV resulted in diverse concessions related to labor-market and social-security reforms until the government reformist intentions receded under the pressure of social protests and two failed military coups.

The Policy Turnaround and the Challenge for Labor Unions

These convergent policy turnarounds in Argentina, Mexico, and Venezuela resulted from the failure of strategies based on state intervention and import substitution industrialization. The 1982 debt crisis highlighted their strategic limitations while it worsened fiscal deficits and balance-of-payment difficulties. Macroeconomic instability and recession followed during the 1980s, the "lost" decade. By the end of the eighties, incumbent populist labor-based parties in Argentina, Mexico, and Venezuela had shifted toward market-oriented reforms. ⁵ According [End Page 140] to Cukierman and Tommasi, labor-based parties have a comparative advantage in implementing market-oriented policies that bring uncertainty to their constituencies because they enjoy their trust. ⁶ That is, they are more credible when they claim that it was the exogenous shock rather than ideology that had induced them to implement market-oriented reforms. Yet, to bring capital back into their economies, labor-based parties in government have to overcome business's distrust of their previous populist character. Their policies, therefore, should be more drastic to show their new commitment to the market.

Drastic market reforms, though, had costs for labor unions and workers whose influence had developed based on state expansion, protectionism, rigid labor markets, and political clout. Trade liberalization increased differences among workers across and within sectors, making it harder to organize labor unions based on horizontal solidarity. International competition and privatization also provoked labor restructuring and layoffs in sectors that had been among the most highly unionized in the past, thus reducing the relative influence of unions. Higher unemployment hurt union bargaining power and increased job instability for union constituencies. ⁷ Stabilization policies that relied on wage restraint and international competition further reduced their wage bargaining power. ⁸ Moreover, the reform of social and labor regulations challenged institutions that had provided unions with legal and political clout--from appointments on social security boards to monopolies of representation--which they would not have been able to achieve based solely on their industrial power. More importantly, market [End Page 141] reforms introduced a high degree of uncertainty about the future positions of union constituencies in the labor market, which often induced them to reject the changes. All these effects were more acute in the sectors that had previously enjoyed high levels of protection from competition (such as the public and the manufacture sectors).

In spite of the effect of market-oriented reforms on unionized labor, incumbent labor-based parties and their allied unions in Venezuela, Mexico, and Argentina had strong incentives to bargain with each other. Both sides preferred to avoid costly militancy and to be governed by labor-based parties rather than by other parties implementing the reforms. Yet, while some of the unions discussed in this study collaborated with their allied parties in government, others engaged in costly militant activities, and some of them undermined the governance of labor-based administrations. What conditions explain labor loyalty to or betrayal of long-term party allies? Furthermore, the success of labor unions in obtaining their objectives was varied, whether they chose restraint or militancy. Why did government officials grant concessions in some cases and not in others?

Interest, Institutions, and Comparative Analysis

Various bodies of literature provide important insights into these questions. Interest-based theories have focused on classes, factors, and sectors as the unit of analysis. These demand-driven theories derive the policy preferences of different actors from their economic interests, explaining policies as a result of these demands. ⁹ Some of them focus on the effect of globalization and market-oriented reforms on the economic interest of diverse sectors, such as exposed and protected or public and private. ¹⁰ According to these theories, the policy preferences of [End Page 142] employees (and employers) in the internationally exposed (or protected) sectors are based on their economic interests in international markets or their dependence on state subsidies. These theories provide a robust account of the origin of union preferences, based on union interests vis-à-vis economic liberalization. These theories, however, present some empirical limitations for explaining the patterns of interaction in my case studies because I found variation within sectors that share a common economic interest. For instance, although both Mexico and Venezuela are oil exporters and oil production in both is monopolized by the state, their oil workers' unions reacted to management attempts to increase labor productivity very differently.

Interest-based theories overlook the effect of institutions on shaping social demands. Institutional variables influence the bargaining power of unions and their relationship with governments. In particular, explanations based on the economic interest of unions alone cannot account for the influence of populist legacies or of organizations on members who are uncertain about the effect of market-oriented reforms on their future interests. Institutional analyses focus either on "formal-public institutions" that influence the responses of governments to social demands or on "socioeconomic institutions" that shape the distributional pressures from social organizations. ¹¹ Regarding formal-public institutions, the recent literature on the politics of economic liberalization centers on policymakers and looks at the capacity of governments to control social demands that may derail market-oriented reforms. ¹² Regarding socioeconomic institutions, the literature on corporatism emphasizes the effect of macrolevel organizational variables, including union internal dynamics and long-term partisan alliances, on union [End Page 143] preferences and strategies vis-à-vis governments. ¹³ The West European version of corporatism assumes open economies and societal corporatism. The Latin American version assumes closed economies and puts greater emphasis on the use of state institutions to control labor organization. ¹⁴

While interest-based theories disregard institutions, macrolevel institutional analysis does not sufficiently explain variation in union behavior within the same country. ¹⁵ My case studies, though, show diversity in the union-government interaction within each country. Hence, following a path set by Pizzorno and Crouch ¹⁶ and reexamined by the recent comparative literature on union politics and the effects of globalization or market reforms on different sectors and across different countries, ¹⁷ I shift the focus of analysis to unions as organizations and to their interactions with governments in a variety of contexts. I propose [End Page 144] a partisan theory of union-government interactions that can be applied both to national confederations across countries and to individual unions in different sectors within the same country by focusing on the interaction of variables that can be singled out in different contexts. The comparison across countries and sectors allows me to isolate the effect of interest at the sector level and of institutions at the national level.

Patterns of Union-Government Interaction

The interaction between unions and governments involves either union militancy or restraint and the capacity to obtain concessions from the government through these means. Militancy here means union-organized protest that affects labor relations; it is the most usual measurement of union behavior. Militancy can disrupt production and undermine governance, especially of labor-based parties who are in power because part of their electoral appeal is based on their control of labor. Militancy is usually measured by counting the number of strikes, their duration, and their scope. Repertoires of protest vary, however, depending on institutional and cultural legacies. ¹⁸ Other means of protest include demonstrations, boycotts, sabotage, hunger strikes, sit-ins, and the like.

The interaction does not end with union militancy or restraint because the government can respond by granting or refusing concessions. Since militancy is costly for unions, union leaders prefer to threaten industrial action rather than actually to exercise it. If the union is strong, this threat should suffice to obtain concessions from the government because conflict also involves costs for the government. Yet, some strong unions choose militancy to achieve their demands despite the cost. Indeed, even weak unions that have more to lose and less to win from labor conflict sometimes opt for "heroic defeats," to borrow Golden's metaphor. ¹⁹

To understand this apparent irrationality in the behavior of unions as well as its effect (if any) on the government granting them concessions, I combined the reactions of both actors into four possible interactions. [End Page 145] Four categories resulted from the combination: successful militancy or "opposition," unsuccessful militancy or "resistance," successful restraint or "cooperation," and unsuccessful restraint or "subordination." The classification is presented in Table 1.

I applied this classification to the main national labor confederations in Argentina, Mexico, and Venezuela, as well as to unions in five economic sectors deeply affected by market reforms: automobile, education, electricity, oil, and telecommunications. ²⁰ The national labor confederations were the Argentine CGT, the Mexican CTM, and the Venezuelan CTV. ²¹ The individual unions organized sectors that were under state management, with the exception of the automobile sector, which enjoyed preferential protection in all three countries. ²² These sectors were chosen because their previous conditions strengthened their bargaining power during the period of protectionism but made them more vulnerable to economic liberalization and state retrenchment. This research design facilitates comparisons across countries, sectors, and even different levels of union organization and facilitates the isolation of the common effect of the independent variables on the dependent variables in these diverse contexts. Since about half of the unions experienced a change in strategy during the period under discussion, [End Page 146] two observations are reported for each union, thus doubling their number (N=18x2=36) to show the continuity or change in union strategies. ²³ Table 2 summarizes all studied interactions by country and sector. The arrows show the passing of time between the first and the second observation. The indicators for militancy and concessions are provided in the Appendix.

Table 2 shows that the case studies provide an array of observations of different union-government interactions with no apparent national or sector-level pattern. It illustrates the need to go beyond country and sector-level variables. The bottom row of the table shows that each of the countries presented at least three different types of interactions despite common national contexts (defined in terms of political institutions and macroeconomic conditions). In turn, the data in the right-hand columns of the table also show diversity in the interactions within each of the economic sectors despite the similarity of interests. In the next section I propose a hypothesis to account for this variation.

A Partisan Theory of Union-Government Interaction

Union members have different preferences for wages, work conditions, and job stability derived from their diverse labor-market alternatives. [End Page 147] Union leaders thus aggregate a particular combination of such preferences into specific union demands. Union leaders also organize collective action (either strikes or restraint) and make possible the control of workers' behavior for the intertemporal exchange of current restraint for future benefits. ²⁴ Hence, union leaders, as workers' representatives, are the agents of any exchange with government officials.

Unions, however, are not perfect agents, and the preferences of union leaders need not be the same as those of workers. ²⁵ Union leaders also have their own objectives in the exchange besides those of their constituencies. For example, union leaders may seek ideological or material rewards while acting as workers' representatives, or they may prefer to maximize the long-term rather than short-term goals of workers. It is possible, however, to assume that they want to maximize leadership survival. That is, whatever the objectives of union leaders, their primary constraint is to "remain in power because otherwise they would not be able to pursue their objectives." ²⁶ Hence, while their objectives can be perceived as a cost for their agency role in the exchange, union leaders are constrained by their constituencies' preferences to the extent that they want to avoid being replaced as workers' agents. ²⁷ Union leaders want to avoid replacement by internal or external rivals. They can be replaced by new leaders who propose a different set of union demands (partisan or leadership competition), or their members can leave the union for other unions whose banners have become more attractive for them (union competition). ²⁸

Previous attempts at leadership survival have led union leaders to build long-term alliances with political parties to complement industrial muscle with political influence. Political parties channeled labor demands through the state and helped mobilize support from other sectors. ²⁹ Partisan alliances built on historical exchanges created loyalty [End Page 148] bonds between parties and unions, thus influencing both their ability to bargain and the attitudes of union leaders toward incumbent politicians. Thus, partisan loyalty implies that, all things being equal, union leaders should be more willing to restrain their militancy when their allied parties are in the government and to increase it when those parties are in the opposition. ³⁰ It follows that if allied union leaders are replaced by others associated with opposition political parties, the change will affect not only the union leaders but also the terms of union-government interactions.

Strategic politicians consider the effect of partisan loyalty on union attitudes. Loyal unions can facilitate the implementation of government policies and provide electoral support. Incumbent politicians should prefer to reward their loyalty rather than support unions that have no attachment to the governing party ³¹ --in particular, nonallied union leaders, who have no ideological and electoral attachments to the government and who, therefore, have fewer incentives for restraint. The identity of the party in government thus affects the organizational dynamics of labor unions. Partisan loyalty may facilitate fulfilling constituencies' demands at the time of the original alliance. If, however, market reforms implemented by labor-based parties increase the uncertainty of workers about their future, partisan loyalties can become contradictory with constituencies' demands, thereby affecting leadership survival and the incentives of labor leaders in their interaction with governments.

Partisan or Leadership Competition

Diversity in labor partisan affiliations implies that various parties appeal to organized labor. In Europe these parties tend to be the Communist, the Socialist, and the Christian Democratic Parties. In Latin America strong populist labor parties have historically competed with leftist parties for union influence. The higher costs of leadership competition in a less democratic context also increased the value of subsidies provided by political parties in Latin America. Leadership competition between rivals associated with different political parties could take place within a single union in the form of partisan competition or across diverse unions combining partisan and union competition [End Page 149] for membership. A partisan monopoly exists when the union leadership affiliated with the labor party faces no competition, such as in the Mexican telephone union. Partisan competition occurs when activists of diverse parties compete or even share leadership positions through proportional representation in the same organization, such as in the Venezuelan CTV.

In some cases, diverse factions of the same party compete for union leadership. Such competition still implies the risk of the allied union leaders being replaced as representatives. For the incumbent labor-based parties, partisan competition implies the threat of allied union leaders being replaced by hostile activists associated with the electoral opposition. Leadership competition among factions will only imply such a threat if the winning faction is likely to defect from the party and join the opposition.

If allied union leaders perceive that leadership competition grows (for example, by losing local union elections), they will try to recover the following of their constituencies to avoid replacement. If they believe that rival leaders are taking advantage of their restraint vis-à-vis market reforms, their incentives for militancy will increase to avoid replacement. Calls for militancy aim to show their constituencies that they have not "sold-out." Leadership survival increases incentives for militancy--even if it is not the best bargaining strategy considering their available information--because if labor leaders are replaced, they will not be able to continue bargaining.

Union Competition

Union competition, or organizational fragmentation, is the rivalry among unions for the representation of workers in the same sector. Where a union monopoly exists, a single union represents all the workers in the sector. In Mexico, for example, the teachers' union was the only union in the education sector, and teachers had no option but to join. Where there is union competition, however, several organizations vie for the membership of the same workers. In Venezuela, for instance, thirteen federations competed for the representation of teachers by 1991.

Union competition introduces the need to coordinate the action of different unions to organize collective action, whether militancy or restraint. The larger the number of unions competing for the same members, the harder it is to coordinate them for collective action. ³² Union competition makes coordination and its enforcement more difficult by [End Page 150] creating incentives to attract members from rival organizations, thereby increasing coordination costs for unions engaging in collective action. ³³ Unions trying to attract members from rival organizations are more likely to differentiate themselves by breaking coordination to become the most appealing to potential members. This situation is heightened if unions have diverse partisan identities that generate not only different attitudes toward the government but also different bargaining costs among competing unions.

In addition to coordination costs, each union is also weaker than a single monopolistic union because each leader controls the militancy or restraint of only a share of the involved workers. As a result, government officials put less value on the exchange with individual unions because they have to enter multiple interactions in their negotiations with the sector, thereby increasing bargaining costs. Thus, although government officials prefer to reward loyal unions, they should be less likely to make concessions to competing unions. Government response to loyal unions, therefore, is more related to the strength of the union than to its militancy.

The Partisan Theory and Union-Government Interactions

The effect of union competition and leadership competition on the interaction between union leaders and government officials varies according to the party in power. When labor-based parties implement market-oriented reforms, allied union leaders are willing to collaborate despite the uncertainty and distress of their constituencies due to their loyalty to long-term allies. Loyal union leaders are predisposed to collaborate and can expect some concessions in return. Moreover, they have better information about the commitments and constraints of their partisan allies and the need to implement these policies.

Yet, when labor-based parties shift their policy, they leave void the populist space of those who disagree with market-oriented policies. Opposition political parties from the left or disgruntled politicians splitting from the governing party (on ideological grounds) can take advantage of the policy shift to occupy that space. This movement may facilitate the growth of militant union activists allied with political parties or partisan factions opposed to market-oriented reforms and may [End Page 151] enhance leadership and partisan competition. Indeed, in Mexico, Venezuela, and Argentina, new political parties occupying this populist space emerged. ³⁴ Furthermore, this movement more likely turns factional competition into partisan competition. Partisan competition, in turn, increases the likelihood of militancy. These effects are more likely in sectors where the high uncertainty of workers about the sudden shift from public to private or from protection to exposure is likely to prompt them into militancy. Additionally, market-oriented reforms can sharpen union competition by provoking splits over how to respond or by aggravating the distributive struggle for shrinking resources among rival unions. In their attempts to attract members from rival organizations to increase their representation in these disputes, unions have more incentives to boycott coordination efforts as a strategy of differentiation, thereby weakening all the unions within the sector.

These explanatory conditions are not fixed. The party in power may change with elections. Union competition and partisan competition within the union movement can also change. Governments, though, cannot usually manipulate these variables in the short-term, although their policies can influence their changes. Market reforms in particular can affect all these variables by improving or damaging the electoral opportunities of incumbents, thus affecting the likelihood of challenges to labor leaders and making it harder for unions to coordinate their strategies. If these changes in the explanatory conditions occur, changes in the dependent variable should follow, thereby explaining variations in the union-government interaction during the period under discussion. Table 3 summarizes the expected union-government interactions when a labor-based party implements market-oriented reforms.

"Cooperation," or successful restraint for concessions, is more likely in the absence of union and partisan competition--when only one union organizes all workers and is affiliated with the governing party. Partisan loyalty reduces the incentives for militancy and facilitates bargaining while union monopoly boosts the bargaining power of the union because government officials want the collaboration of a strong and loyal union.

"Opposition," or successful militancy, can more likely be expected in the presence of growing partisan competition and union monopoly-- [End Page 152] when leaders affiliated with different parties compete for the control of a single union. Growing partisan competition based on protesting the consequences of the policy shift increases the incentives for "irrational" militancy ³⁵ because allied union leaders are afraid of being replaced and the sections controlled by contenders have already turned militant. Since the union is strong, government officials are more likely to grant concessions so that allied union leaders can show a better record than their contenders.

"Subordination," or unsuccessful restraint, more likely results from competition among different unions affiliated with the governing party. While partisan loyalty facilitates restraint, union competition weakens all unions despite their loyalty. Government officials may also choose to reward only the most compliant unions. This selection should also prompt competing unions to comply to avoid losing resources and becoming less attractive for members than other competing unions.

"Resistance," or unsuccessful militancy, more likely happens when partisan competition and union competition overlap--competing unions affiliated with different parties. Union competition weakens all unions and, together with partisan competition, makes coordination more difficult. Unions associated with opposition parties protest to differentiate themselves from cooperative loyal unions. If they succeed in attracting members due to their bellicose stance, loyal unions will turn more militant to avoid losing members although union competition makes them weak and unlikely to be successful. [End Page 153]

Applying the Partisan Theory in Argentina, Mexico, and Venezuela

According to the partisan theory, union-government interactions in Argentina, Mexico, and Venezuela should have felt the effects of partisan ties because the incumbent parties were labor-based. In this context, leadership competition should have increased the likelihood of labor militancy, and union competition should have decreased the likelihood of concessions for unions. This section provides a brief description of the dynamics of the case studies, relating them to the predicted background conditions for each interaction, which are summarized in the Appendix.

In Venezuela, although AD historically controlled the CTV, other parties competed for leadership and displaced AD briefly in the 1960s. Also the system of proportional representation reinforced the pluralism of the CTV in its leadership, resulting in the inclusion of minority parties in the executive committee. ³⁶ In 1989, after Pérez announced the reforms, urban riots signaled the general discontent of the population. Union leaders of AD, afraid of losing control of the CTV to opposition parties protesting the reforms, called the first economic general strike in Venezuelan history for May 1989. The tension between partisan loyalty and leadership survival divided the AD union leadership between those with positions in the CTV who were more favorable to the strike and those with party appointments who were more reluctant about it. ³⁷ Pérez's concessions included emergency wage hikes, suspension of layoffs, and retaining price controls for basic staples. In 1990 and 1991, however, Causa R, a new political party, grew rapidly in the union movement by rejecting reforms, especially among public sector workers, further increasing partisan competition. As a result, the CTV called additional protests despite Perez's move to halt the reform of the severance payment system and social security and his acceptance of union demands for the resignation of the labor minister. Hence, union monopoly and increasing partisan competition explain the CTV opposition.

In spite of the national tendencies in union behavior, Venezuela experienced variation in union-government interactions. In the case of the privatization of the state-owned monopoly of telecommunications, [End Page 154] the combination of union monopoly and partisan monopoly exercised by the AD union leadership resulted in cooperation. Labor restraint was exchanged for concessions that included employee-owned stock, job stability, and labor directors. Subsequent worker discontent with the privatization, however, resulted in Causa R winning the union elections in the main local union (Caracas) while growing in other regions by protesting market reforms and privatization. As a result, a new AD leadership took control of the union and increased the militancy of the national union against labor restructuring to show their responsiveness to the discontented rank and file. In spite of attempts at broad restructuring, the company could not change work conditions from the privatization agreements. In this case, the increase in partisan competition combined with union monopoly to move the union into opposition. ³⁸

Very similar dynamics explain the shift from cooperation to opposition in the cases of the electricity workers' union of the state-owned electric company (CADAFE) and of the state-owned oil company (PDVSA). Union monopoly and an unchallenged AD leadership facilitated cooperation when the Pérez administration decentralized CADAFE. Concessions to the union included handsome monetary incentives for transfers, job stability, and wage increases for workers in the interior. Yet, the subsequent discontent of workers also gave the election of the Caracas's union to Causa R, who had led the local union into a wildcat strike. Subsequently, the AD leadership increased the militancy of the union and successfully boycotted industrial restructuring. Again, the increase in partisan competition combined with union monopoly resulted in opposition. In PDVSA, the combination of union monopoly and no partisan competition for the AD leadership favored cooperation in 1991. In return, the union retained hiring prerogatives. Afterward, workers' discontent resulted in increasing partisan competition, not only from Causa R but also from other left-wing parties--the Movement to Socialism (MAS) and the Electoral Movement of the People (MEP). As a result, the AD leadership took a more combative stance against management proposals in 1993 and tried to gain legitimacy by decreasing its discretionary use of hiring prerogatives and obtaining a halt to the reform of commissaries. ³⁹ Thus, partisan competition, along with union monopoly, explains the shift of the union to opposition.

Contrary to the unions in public enterprises, the two remaining cases show a different pattern. The union of Ford Motors, a local union [End Page 155] based in Carabobo, opted for cooperation. The union restrained and accepted layoffs and new working rules in return for participation in the selection of those laid off, reincorporated, and trained. In this case, cooperation resulted from the combination of union monopoly and partisan monopoly. In a personal interview, the labor relations' manager said that cooperation for restructuring and training was possible because the AD union leadership did not face competition from left-wing parties. ⁴⁰ In education, the overlap of union competition and partisan competition in thirteen federations affiliated with diverse political parties and groups resulted in militant but unsuccessful resistance to the administrative reform initiated by Minister Gustavo Roosen. While coordination was difficult, the militancy of the rank and file rewarded the free riding of bellicose organizations. This situation brought the vice minister of education, Francisco Castillo to complain, "I think that there is a competition among unions, where the one that strikes the most is the one that is the most supported by the rank and file." ⁴¹ In spite of the differences between these two cases, both sectors contrast with telecommunication, electricity, and oil in that the levels of union competition and partisan competition did not change during the period studied, and neither did the studied interactions.

In Argentina, ⁴² Peronism's policy shift divided the CGT. All three competing factions, though, remained within the ranks of Peronism. The Menem administration was unwilling to make concessions, in particular to the most "populist" faction, although it provided selective incentives (such as guaranteed monopolies of representation and executive [End Page 156] appointments) to the most compliant. Peronist labor unions were unable to halt a new law that introduced temporary hiring and decrees that established wage restraint and deprived them of collecting and administering union health-fund fees. As a result, many important unions moved out of the populist faction, and in 1992 all three factions decided to unify under a leadership dominated by the most cooperative union leaders. Thus, from 1989 to 1992 the combination of partisan monopoly and union competition subordinated Peronist unions to the administration. After its unification, the CGT recovered its union monopoly and maintained its Peronist loyalty, moving from subordination to cooperation. As a result, it was rewarded with concessions that included changes in the reforms of pensions, social security, subsidies for health funds, and the maintenance of legislation on collective bargaining and labor organization.

Argentina also shows diversity in union-government interactions although national institutions and conditions remain the same. Menem also privatized the state-owned monopoly of telecommunications, ENTEL. Unlike its Venezuelan counterpart, the union opposed privatization by increasing its militancy. A combination of union monopoly and leadership competition explains this outcome. Leadership competition in the telephone workers' union increased when a populist faction that opposed market reforms and privatization won control of the union in Buenos Aires (the largest in the federation) and increased union militancy. ⁴³ A monopolistic union, nonetheless, was able to obtain concessions, including the administration and representation of employee-owned stock, handsome retirement packages, and subsidies for the union health fund. After privatization, the loyal Peronist leadership won control of the rebellious locals and curtailed leadership competition. Hence, with the absence of union and partisan competition, the union cooperated with the new private management in discussions of labor restructuring. Concessions at this stage included union participation in training, contracts for union-organized companies, and voluntary retirement packages. The privatization of state-owned electric companies also resulted in initial opposition due to a combination of union monopoly and leadership competition. The local unions, which the populist militant faction controlled and which competed with the [End Page 157] loyal leadership, joined the CTA in 1992. Since the national federation was a member of the CGT, the dissident unions were expelled. As a result, leadership competition receded. Lacking union competition and partisan competition, the union began to cooperate. In the first period, union concessions included employee-owned stock, subsidies for the union health fund, and contracts for union-organized cooperatives of former employees. In the second period, they included subsidies for union participation in the privatization of companies--public utilities, an energy transmission company, and the concession of a coal mine--as well as voluntary retirement packages.

The cases of the restructuring and privatization of the state-owned oil company (YPF) and of the decentralization of education differed from the previous two examples. In the case of the YPF, the Peronist union leadership faced no competition and had a very close relationship with Menem that facilitated bargaining for concessions, which included subsidies for the union to buy the oil fleet, contracts for union companies employing laid-off workers, and voluntary retirement packages. Thus, leadership monopoly and union monopoly resulted in cooperation. Unlike the oil company, the education sector, as in Venezuela, was fragmented into various unions. Many of them gathered into a national confederation, CTERA, whose leader, Mary Sánchez, was one of the Peronist union leaders who had left the party and the CGT to found the FREPASO and the CTA in protest of the policy shift. Yet, many others unions did not belong to the national confederation and competed with CTERA unions in every province. Many of the non-CTERA unions were Peronist. In this case, the overlapping of union competition and partisan competition coincided with resistance or unsuccessful militancy against the decentralization of education to the provincial level enacted by Congress in 1992. In spite of the restraint of oil workers and of the militancy of teachers that accounted for more than a third of total strikes in 1991 and 1992, in both cases the levels of union competition, partisan competition, and union-government interactions remained the same during the period under discussion.

Lastly, the national union of automobile workers moved from opposition to cooperation in 1991. In this case, union monopoly with no partisan competition made the union's initial opposition unexpected, although the subsequent cooperation coincides with the theory. The increasing militancy of the union until 1991 is related to the rejection of trade liberalization in a heavily protected sector. The consequence of this militancy together with lobbying by business was to persuade the government to grant an industrial regime of protection and gradual [End Page 158] opening for the sector after 1991. The industry was thus exceptionally successful in shielding itself from the conditions created by a sudden trade opening. After the regime was granted, the government became again an important partner, and partisan loyalties moved the union toward cooperation.

The Mexican cases highlight the importance of contextualizing the explanatory variables within the constraints created by the political regime on the means of expressing militancy and the costs of partisan competition. ⁴⁴ Yet at the same time, even in the most constrained environment, where the regime is not totally democratic, the interaction between unions and the government varies. This variation cannot be explained by traditional theories based only on the features of the Mexican regime. The Mexican CTM subordinated to Salinas's policies. The only important concession that it was able to obtain was the prevention of labor-code reform. The same variables--partisan monopoly combined with union competition--explain the confederation's subordination and its exception to cooperation in relation to the labor code. ⁴⁵ The CTM competed with other national labor confederations also associated with the monopolistic PRI. While partisan loyalty reduced their incentives for militancy, government officials manipulated union competition for scarce resources among the various PRI-related national confederations and rewarded the most compliant of them with selective incentives (such as public recognition and favorable treatment in arbitration boards) to weaken CTM claims. The only exception was when the CTM unified with all other PRI-confederations and reduced union competition to boycott the reform of the labor code successfully. The same institutional mechanism that provided a common forum for all of them to unify, an umbrella organization called the Congress of Labor, had previously facilitated union competition because decision making was based on consensus or unanimity. ⁴⁶ [End Page 159]

As in Venezuela and Argentina, the state-owned monopoly of telecommunications, Telmex, was privatized in Mexico. In this case, cooperation in relation to both privatization and restructuring corresponds to the overlap of union monopoly and leadership monopoly, which remained unchanged during the entire period studied. The latter, held since 1974 by a charismatic PRI union leader, Francisco Hernández Juarez, who had co-opted the opposition, successfully provided benefits for his constituencies and developed a close relationship with the president, who often used this union as an example of modern unionism. Moreover, since this union was not affiliated with the CTM, it had more room to maneuver and even organized the Federation of Goods and Services Workers (FESEBES), a new labor confederation competing with the CTM. ⁴⁷ The union obtained employee-owned shares and a labor director in the private company, wage increases, job stability, and the maintenance of work conditions during privatization. Afterward, concessions included participation in joint committees with management for training and restructuring and increases in wage and nonwage benefits.

The union of electricity workers of the Company of Light and Power, which was not a CTM-affiliate either, also joined the FESEBES. Its charismatic union leader, Jorge Sanchez, also had a close relationship with the president. In this case, the absence of union and leadership competition led to cooperation between the union and management. Cooperation led the government to bail out the company, establish guarantees for union survival, and develop joint union-management committees to discuss productivity and financial issues. Unlike Hernández Juarez, though, the loyal leader lost the union elections in 1993 and was replaced by an independent leadership, breaking the loyalty strings. As expected, this situation increased union incentives for [End Page 160] militancy (although less than if the leadership was associated with an opposition party). It also, however, reduced the partisan incentives of the government to grant concessions to the union, which nonetheless included a reduction of productivity targets and the maintenance of job stability. In this case, the tension between partisan loyalty and leadership survival was broken by the replacement of the loyal leader and his succession by an independent and more militant union leadership.

In Mexico, the education sector was not as fragmented as in Argentina and Venezuela. The National Union of Education Workers (SNTE) which also was not a CTM-affiliate, had a union monopoly and was controlled by PRI union leaders. A discontented rank and file, organized by dissidents in the National Coordinating Committee of Education Workers (CNTE), had toppled the previous PRI leadership. The new PRI leader, Elba Ester Gordillo, therefore, had to win legitimacy to avoid replacement. As a result, the SNTE became more bellicose after the change in leadership while including the dissidents within the executive committee through proportional representation and increasing internal debate. ⁴⁸ In 1991 when the decentralization law was discussed, a leak from the secretary of education mentioned the division of the centralized union. The SNTE not only became more militant but also joined forces with the dissident CNTE. The government responded by granting centralized national work conditions and earmarked budgets for the states as well as salary hikes and nonwage benefits for the union. Hence, the combination of union monopoly and partisan competition resulted in opposition.

The unions of oil workers and Ford automobile workers in Mexico, both CTM-affiliates, provide interesting cases to illuminate the limits of my explanatory framework. In such cases, partisan competition may be punished by a nondemocratic government that can resort to coercion. Both unions faced sharp restructuring due to the opening of their sectors to private capital and international competition. Both attempted to resist the changes and were forced into subordination by the repression of the CTM and the state. The workers of the Ford Motors plant at Cuatitlán rejected restructuring and chose an independent leadership (linked with left-wing parties), which broke the tension between partisan loyalty and leadership survival and brought the local union to resist industrial restructuring--albeit unsuccessfully--in 1988 and 1992. The [End Page 161] local union not only increased its militancy but also attempted to withdraw from the CTM-affiliated national union. The CTM, upheld by the government and the company, responded to the militant leadership and their supporters with violent repression, forcing the local and the national union into subordination. ⁴⁹ In the case of the oil workers' union, the PRI union leader had supported the opposition presidential candidate, Cuahutemoc Cárdenas, ⁵⁰ by delivering votes in the oil regions, although PRI local candidates carried the elections. As a result of the partisan threat, the state gave a fatal blow to the union by putting the union leader in prison under fabricated murder charges and by bringing the union into subordination, thereby shifting the issue of leadership survival from workers to state officials. ⁵¹ In both cases, the restrictions on political pluralism of the regime limited partisan competition in the unions and tipped the balance in favor of partisan loyalty by dramatically raising the costs of noncompliance with government officials. ⁵² The Mexican political regime restricted partisan competition in the union movement, thus limiting the explanatory power of this variable for certain unions. Yet, other non-CTM unions, such as the teachers', telephone workers', and electricity workers' unions--and even the CTM itself--retained their relative autonomy, thus permitting the use of the explanatory variables based on union dynamics to illustrate their interaction with the government.

Table 4 shows the high correspondence between the patterns of union-government interaction presented in Table 2 (and summarized in the Appendix) and the explanatory conditions defined by my partisan theory and presented in Table 3.

This high correspondence between the observed outcomes and the explanatory conditions provides a better account than either macrolevel or sector-level theories for the interactions studied. In a context of allied [End Page 162] labor parties implementing market reforms, partisan competition and union competition influenced the interaction between loyal union leaders and government officials in different sectoral and national contexts. Additionally the explanatory conditions in the studied interactions varied as expected over the short-term, further reinforcing the implications of this explanatory logic.

Other patterns of interaction also deserve attention. In Argentina a pattern of cooperation between unions and government tended to arise mainly from opposition but also from subordination at the national level. In Venezuela a pattern of opposition between the two arose mainly from cooperation, although arriving early in this interaction at the national level. Finally, although Mexico shows a less clear trend, it is where most cases of subordination concentrate. These trends in the interactions studied are related to national institutions that can facilitate the emergence of union competition and partisan competition, thus explaining these patterns.

In Argentina national regulation established monopolies of representation for collective bargaining except in the public administration. As a result, union competition is limited except at the peak level and in the public administration, where the cases of union competition leading to subordination or resistance emerged. The strong partisan identity of labor unions combined with their ability to obtain concessions from the government helped Peronist labor leaders retain control of their unions. In Venezuela collective bargaining rules also facilitated effective union [End Page 163] monopolies. ⁵³ The education sector was an exception, excluded from collective bargaining and faced with union competition, as in Argentina. The growth of leadership competition that increased labor incentives for militancy in many AD unions is related to the political changes brought about by the combination of market reforms and political decentralization. These changes made room for new political options not only in the union movement but also in the electoral arena. ⁵⁴ In Mexico the characteristics of the regime allowed the government to overcome the effect of the explanatory conditions in the cases of oil workers and the Ford Motors' workers of Cuatitlán. These cases do not contradict the explanatory logic but highlight its limitation. In the oil workers' union, partisan monopoly and union monopoly had previously led to labor restraint in return for sizable concessions including nonwage benefits for workers, hiring prerogatives, contracts for union companies, and fees over suppliers' contracts for the union. After its leader challenged the election of Salinas, however, the union was forced into subordination. In the case of Ford Motors, when alternative leaders won local elections, breaking loyalty ties and increasing labor militancy, the government and the CTM also forced the union into subordination. Thus, in Mexico the opportunities for leadership or partisan competition were limited in most cases by a regime that had also curtailed political pluralism in the electoral arena.

National patterns also had important policy consequences. The prevalence of opposition in Venezuela weakened the pace of market reforms until they finally were suspended. The predominance of cooperation in Argentina facilitated the rapid implementation of market reforms and the adoption of union strategies more akin to a market economy. In Mexico, although subordination eased the implementation of reforms, it also slowed the adaptation of union strategies to the new environment. In spite of national trends, though, some cases in each country entered into different interactions explained by other combinations of union competition and partisan competition.

At the sector level there were less clear patterns. The cases in the oil and automobile sectors exhibited a large variation in interactions. Telecommunications and electricity do not have a single predominant interaction, but unions in both sectors successfully obtained concessions through cooperation or opposition (and enjoyed union monopoly). [End Page 164] In the education sector, both resistance and opposition interactions are associated with higher militancy. This finding is consistent with predictions of higher militancy in public-sector workers ⁵⁵ and with the politicization of the education sector ⁵⁶ that facilitates the emergence of partisan competition. While partisan competition in Argentina and Venezuela had been an enduring feature of teachers' organizations, in Mexico it took an internal rebellion to allow for increasing partisan pluralism within the union. ⁵⁷ Thus, although some sector-level patterns emerged, they are insufficient to account for the studied interaction.

Conclusions

This article attempts to complement the focus on policymakers that has predominated the literature on the political economy of market reforms in developing countries by providing some perspective on labor organizations, which have been neglected. The findings underscore the effect of partisan identities on creating a tension for union leaders when their allies implement market-oriented reforms. In my case studies, union leaders, on the one hand, were pulled by their partisan loyalties--based on long-term goals or on their own personal gain. On the other, they had to respond to their constituencies--based on the short-term goal of political survival. Partisan loyalty to governing allies on the part of union leaders increases incentives for restraint, while leadership or partisan competition pushes them toward militancy if demanded by constituencies. At the same time, union competition affects the strength of the union and its ability to obtain concessions from the government. This theory explained most of the interactions in my study between union and labor-based governments implementing market-oriented reforms in Argentina, Mexico, and Venezuela. In these countries, partisan loyalties inherited from the original postwar alliance between the Peronism, PRI, and AD with labor unions had an important influence on the transition to open economies. The interaction of partisan loyalties with union competition and partisan competition, however, [End Page 165] can also influence the attitudes of unions and governments toward each other for nonlabor parties in government. The absence of partisan loyalties can make restraint and negotiations more difficult for unions and governments affecting the influence of leadership competition for militant labor leaders and of union competition in noncooperative unions. Further testing in other cases will show the extensions and limitations of this partisan theory.

If my argument is correct and political actors were aware of the effect of union competition and leadership competition, unions should have tried to affect these variables. In the previous section, I analyze the effect of legal institutions and electoral dynamics in shaping national patterns of union competition and leadership competition and thus union-government interactions. According to Collier and Collier the incorporation of labor shaped both legal institutions and the political system. ⁵⁸ When the alliance between labor unions and political parties was established, the corporatist labor legislation that emerged shaped the opportunities for union competition and leadership competition because unions sought to strengthen their bargaining power and politicians wanted to reinforce labor loyalties. In Argentina, Mexico, and Venezuela, the labor legislation included restrictions to exit (ranging from permission to close shops to monopolies of representation) and facilitated the controls over leadership selection by incumbents. ⁵⁹ At the same time, the partisan loyalties derived from the alliance were reinforced by the delivery of wage and social benefits when labor-based parties were in government. ⁶⁰ National patterns thus resulted from the diverse emphasis on these variables that emerged from the original alliance. Union competition, however, was "stickier" than partisan competition because it was harder to modify on an individual basis without a legislative reform. This stickiness explains why leadership competition varied more frequently than union competition in the cases studied, making changes in militancy levels more likely than changes in the ability of unions to obtain concessions in the short term.

Although market reforms affect both conditions as analyzed above, changes are not easy to manipulate in the short term. Politicians cannot easily control electoral dynamics unless they resort to repression. Indeed, they are usually subject to the electoral dynamics unleashed by their shift toward the market and by the new alliances that this critical [End Page 166] juncture allowed in the union movement. ⁶¹ They could reform the rules for labor organization, but these regulations have remained unchanged in all three countries during the period studied. Reforms to the regulations on labor organization affecting union competition would have also risked changing the partisan ties between labor unions and governing politicians on which incumbents counted to implement their policies and remain in power. ⁶²

The conclusions of this study resonate with Michels's claim that although organization is the weapon of the weak in their struggle with the strong, "it is the organization which gives birth to the dominion of the elected over the electors, of the mandataries over the mandators, of the delegates over the delegators." ⁶³ He considers, however, that leadership competition should limit the iron rule of oligarchy in organizations because the "ascent of the new leaders always involves the danger, for those who are already in possession of power, that they will be forced to surrender their places to the newcomers. The old leader must therefore keep himself in permanent touch with the opinions and feelings of the masses to which he owes his position." ⁶⁴ Coincidentally with Michels and Hirshman, this study shows that although union competition provides exit options for workers, it weakens unions. Instead, leadership or partisan competition provides voice options to increase representation but without weakening their bargaining power to the same extent. ⁶⁵

Finally, this study suggests that future research should address the interaction between different levels of analysis. Such research would not only look to recover forgotten actors (such as individual unions) from oblivion but also to move beyond the national-level bias to include units of analysis defined at the level of sectors, subnational units, and even organizations. The multilevel research design of this study combined cross-country and within-country comparisons to facilitate [End Page 167] the testing of alternative explanations within a relative small N. In turn, a small N made possible the collection of the data necessary to test causal mechanisms based on organizational dynamics. Theoretically this research design demonstrates the possibility of holding national institutions and macroeconomic conditions constant in comparisons within countries as well as sector-level variables constant in comparisons within sectors. Empirically it provides a better picture of the complex reality of the countries studied and the organizational dynamics of each case while advancing comparative analysis.

M. Victoria Murillo is Assistant Professor of Political Science at Yale University. She is the author of Divided Loyalties: Labor Unions and Market Reforms in Latin America (forthcoming). She has also contributed to numerous academic journals and edited volumes on themes such as teachers' unions and the decentralization of education in Argentina and Mexico, and market-oriented reforms and Argentine corporatism.

Appendix

Notes

* Earlier versions of this paper were presented at the twenty-first meeting of the Latin American Studies Association, Chicago, September 24-26, 1998, and at the conference "Space, Place, and Nation: Reconstructing Neoliberalism in the Americas," University of Massachusetts, Amherst, November 20-21, 1998. I would like to thank Arun Agrawal, Ernesto Cabrera, Ernesto Calvo, Javier Corrales, Miguel Glatzer, Miriam Golden, Pauline Jones-Luong, Robert Kaufman, Steven Levitsky, James McGuire, Nicoli Nattrass, Phillip Oxhorn, Héctor Schamis, Jeremy Sikkins, and four anonymous reviewers for helpful comments to previous versions of this paper, and Antonieta Mercado for her research assistance in Mexico. I would also like to thank Alberto Alesina, Robert Bates, David Cameron, John Coatsworth, Jorge Domínguez, Geoffrey Garrett, Peter Hall, Ian Shapiro, Juan Carlos Torre, and Deborah Yashar, who provided useful comments to clarify the ideas presented in this paper. I gratefully acknowledge the financial support of the Mellon Foundation, the Harvard Academy for International and Area Studies, and the David Rockefeller Center for Latin American Studies at Harvard University, as well as the Institute for the Study of World Politics. I also want to acknowledge the institutional support of the Department of Political Science at Yale University; the Instituto Torcuato Di Tella in Argentina; the IESA in Venezuela; the Colegio de México, Flacso, and ITAM, and Intelmex in Mexico.

1. Salinas was president of Mexico from the end of 1988 to the end of 1994, and Carlos Andrés Pérez was president of Venezuela from the beginning of 1989 to mid-1993. To hold the international conditions constant in the comparison with Mexico and Venezuela, I analyze the first administration of Carlos Menem in Argentina, which ran from mid-1989 to mid-1995.

2. Ruth Berins Collier and David Collier, Shaping the Political Arena (Princeton: Princeton University Press, 1991).

3. The politics of Argentina, Mexico, and Venezuela have a strong effect in Latin America and the Caribbean. By 1995 they made up 32 percent of the regional population, 43 percent of the regional gross domestic product, and 48 percent of regional exports. Inter-American Development Bank, Economic and Social Progress in Latin America (Baltimore: Johns Hopkins University Press, 1996), 357-61.

4. Import substitution industrialization (ISI) was a development strategy adopted by most Latin American countries after the Great Depression. They originally raised tariffs to compensate for the shortage of foreign exchange produced by the crisis, but this policy gradually evolved into active protectionism that included subsidized exchange rates for importing inputs with closed markets. ISI created few incentives for developing internationally competitive industries and, thus, exporting.

5. Market-oriented reform included short-term stabilization measures, fiscal restraint, tax reform, financial liberalization, competitive exchange and interest rates, trade liberalization, privatization, and deregulation of most markets, including the labor market. See John Williamson, Latin American Adjustment: How Much Has Happened? (Washington, D.C.: Institute for International Economics, 1990).

6. Alex Cukierman and Mariano Tommasi, "Credibility of Policymakers and of Economic Reform," in Federico Sturzenegger and Tommasi, eds., The Political Economy of Reform (Cambridge: MIT Press, 1998).

7. In Argentina, according to the permanent household surveys of the National Institute of Statistics and Censuses (INDEC), unemployment increased from 6.5 percent in 1988 to 18.6 percent in 1995. In Venezuela unemployment rose from 6.9 percent in 1988 to 9.6 percent in 1989 and 10.4 percent in 1990, although it fell to 6.5 percent in 1993. See Keila Betancourt, Samuel Freije, and Gustavo Márquez, Mercado labora: Instituciones y regulaciones (Labor markets: Institutions and regulations) (Caracas: IESA, 1995), 5. In Mexico, according to the International Labour Organisation, (ILO), unemployment measurements are highly contentious, but open unemployment peaked in 1983 and 1984. See ILO, Yearbook of Labour Statistics (Geneva; ILO, various years). In addition, the combined official rate of open unemployment and underemployment grew from 6.8 percent in 1989 to 8 percent in 1994. See Carlos Salinas de Gortari, VI Informe de Gobierno (VI State of the union) (Mexico City: Presidencia de la Nación, 1994).

8. In Argentina, hyperinflation cut manufacture real wages by 36.3 percent between January 1989 and March 1991. Even after the success of stabilization, manufacture real wages fell by 12 percent between April 1989 and June 1995. Consejo Ténico de Inversiones, La economía argentina:. Anuario 1997 (The Argentine economy: Yearbook 1997) (Buenos Aires: CTI, 1997), 65. In Venezuela the real industrial wage fell 35 percent in the 1989-93 period. Unido Industrial Statistics Database, 3-digit (1998). In Mexico real wages in manufacturing had dropped by almost 40 percent between 1982 and 1988, and despite improvements during the Salinas administration, they did not recover their 1982 level. ILO (fn 7).

9. Jeffry Frieden, "Invested Interests: The Politics of National Economic Policies in a World of Global Finance," International Organization 45 (Autumn 1991); Peter Gourevitch, Politics in Hard Times: Comparative Responses to International Economic Crises (Ithaca, N. Y.: Cornell University Press, 1986); and Ronald Rogowski, Commerce and Coalitions (Princeton: Princeton University Press, 1989).

10. For example, Frieden presents a general argument about sector-driven policy preferences and applies it to unions and exchange rate policies. Frieden (Ibid.); and idem, "Labor and Politics of Exchange Rates: The Case of the European System," in Sanford Jacoby, ed., The Workers of Nations: Industrial Relations in a Global Economy (New York and London: Oxford University Press, 1995). Peter Swenson focuses on cross-class coalitions based on sector-level preferences with regard to collective bargaining decentralization and state adjustment. Swenson, "Bringing Capital Back In, Or Social Democracy Reconsidered," World Politics 43 (July 1991). The Latin American literature uses a similar logic to explain populist coalitions between urban workers and industrialists producing for the domestic market based on the transfer of resources from exporting to protected sectors for import substitution industrialization. Fernando Henrique Cardoso and Enzo Faletto, Dependencia y desarrollo en América Latina (Dependency and development in Latin America) (Buenos Aires: Siglo XXI, 1969); and Guillermo O'Donnell, Modernization and Bureaucratic-Authoritarianism (Berkeley: University of California, Institute of International Studies, 1973).

11. Geoffrey Garrett and Peter Lange, "Internationalization, Institutions, and Political Change," in Robert Keohane and Helen Milner, eds., Internationalization and Domestic Politics (Cambridge: Cambridge University Press, 1996).

12. Volumes by Joan Nelson, Steven Haggard and Robert Kaufman, and Dani Rodrik assume that liberalization has concentrated costs and diffused benefits, making it difficult to organize collective action in support of reform. Nelson, ed., Economic Crisis and Policy Choice: The Politics of Adjustment in the Third World (Princeton: Princeton University Press, 1990); idem, Fragile Coalitions: The Politics of Economic Adjustment (New Brunswick, N.J.: Transaction Books, 1989); Haggard and Kaufman, eds., The Politics of Economic Adjustment (Princeton: Princeton University Press, 1992); idem, The Political Economy of Democratic Transitions (Princeton: Princeton University Press, 1995); and Rodrik "Understanding Economic Reform," Journal of Economic Literature 36 (March 1996). Therefore, the management of reform implies the thwarting of societal resistance and the insulation of reformist policymakers. Catherine Conaghan and James Malloy and Carlos Acuña and William Smith also point out the combination of repression, co-optation, and insulation by skillful policymakers for the implementation of market reforms for Latin America. Conaghan and Malloy, Unsettling Statecraft: Democracy and Neoliberalism in the Central Andes (Pittsburgh: University of Pittsburgh Press, 1994); Acuña and Smith, "The Political Economy of Structural Adjustment: The Logic of Support and Opposition to Neoliberal Reform," in Smith, Acuña, and Eduardo Gamarra, eds., Latin American Political Economy in the Age of Neoliberal Reform (New Brunswick, N.J.: Transaction Publishers, 1994).

13. A number of scholars have analyzed the effect of union density and concentration, centralization of wage bargaining, and partisan affiliation on union behavior. See, for example, Phillipe Schmitter, "Still the Century of Corporatism," Review of Politics 36 (January 1974); David Cameron, "Social Democracy, Labor Quiescence, and the Representation of Economic Interest in Advanced Capitalist Society," and Peter Lange, "Unions, Workers, and Wage Regulation: The Rational Bases of Consent," in John H. Goldthorpe, ed., Order and Conflict in Contemporary Capitalism (Oxford: Clarendon Press, 1984); Lars Calmfors and John Driffill, "Centralization and Wage Bargaining," Economic Policy 3 (April 1988); R. Michael Alvarez, Geoffrey Garrett, and Peter Lange, "Government Partisanship, Labor Organization, and Macroeconomic Performance," American Political Science Review 85 (June 1991); and Peter Lange and George Tsebelis, "Strikes around the World: A Game Theoretic Approach," in Jacoby (fn. 10).

14. In Schmitter's original definition, organized interests in "societal" corporatism emerged more autonomously from the state than in "state" corporatism. In the comparative Latin American literature, Ruth Berins Collier and David Collier emphasize state "incorporation" of labor, and Francisco Zapata stresses the deeply political character of union activity resulting from the high degree of state intervention. Country studies confirm this view. In particular, these three countries are classified by both Zapata and Collier and Collier as having historically high levels of state intervention and legal benefits for formal workers. Collier and Collier (fn. 3); idem, "Inducement versus Constraints: Disaggregating 'Corporatism,'" American Political Science Review 73 (December 1979); Zapata, El conflicto sindical en América Latina (Labor conflict in Latin America) (Mexico City: El Colegio de Mexico, 1986); and idem, Autonomía y subordinación en el sindicalismo latinoamericano (Autonomy and subordination of Latin American unionism) (Mexico City: Fondo de Cultura Económica, 1993).

15. Katrina Burgess's dissertation, an institutionalist work on the reshaping of the alliance between organized labor and labor parties, focuses on the external costs created by political parties on the decision making of unions at the national level. Instead, I propose to analyze the internal dynamics of unions and the effect of competition for leadership or for members on their relations with political parties. Burgess, "Alliances under Stress: Economic Reform and Party-Union Relations in Mexico, Spain, and Venezuela" (Ph.D. diss., Princeton University, 1998).

16. Allesandro Pizzorno, "Political Exchange and Collective Identity," in Colin Crouch and Pizzorno eds., The Resurgence of Class Conflict in Western Europe since 1968, vol. 2 (New York: Holmes and Meier Publishers, Inc., 1978); and Crouch, Trade Unions: The Logic of Collective Action (Cambridge: Fontana Paperback, 1982).

17. Volumes edited by Miriam Golden and Jonas Pontusson and by Christopher Candland and Rudra Sil provide a nice sample of new work in this direction for the developed and the developing world respectively. Golden and Pontusson, eds., Bargaining for Change: Union Politics in North America and Europe (Ithaca, N. Y.: Cornell University Press, 1992); and Candland and Sil, eds., Industrial Relations in the Age of Globalization (Oxford: Oxford University Press, forthcoming).

18. Due to the effect of institutional and cultural constraints, militancy was measured using diverse forms. For instance, the meaning of a general strike in Argentina, where they have occurred often in the past--even under Peronist administrations--is different from that of a general strike in Venezuela, where there were no antecedents of such means of protest for economic demands. See James W. McGuire, Peronism without Peron: Unions, Parties, and Democracy in Argentina (Stanford, Calif.: Stanford University Press, 1997); and Steve Ellner, Organized Labor in Venezuela, 1958-1991 (Wilmington, Del.: Scholarly Resources, 1993).

19. Miriam Golden, Heroic Defeats: The Politics of Job Loss (Cambridge: Cambrige University Press, 1997).

20. National confederations are multisectoral economy-wide organizations to which industry-specific unions adhere.

21. In all three countries, I compare the responses of the main national confederations to a set of policies: stabilization, privatization, trade liberalization, social security, changes in the regulations for labor organization, and labor market flexibility.

22. The unions involved in the study in Argentina were the Union of Automobile Workers (SMATA), the Federation of Light and Power Workers (FATLyF), the Federation of Telephone Workers (FOETRA), the Union of State-Owned Oil Workers (SUPE), and the Argentine Federation of Teachers (CTERA) with its rival unions. In Mexico, the unions in the study were the Mexican Union of Telephone Workers (STRM), the Mexican Union of Oil Workers (STPRM), the Mexican Union of Electricity Workers (SME), the National Union of Education Workers (SNTE), and the local union of Ford Motors Workers at Cuatitlán. In Venezuela, the unions studied were the Federation of Telephone Workers (FETRATEL), the Federation of Electricity Workers (FETRAELEC), the Federation of Oil Workers (FEDEPETROL), the Ford Motors' Section of the Federation of Automobile Workers (FETRAUTOMOTRIZ), and the multiple unions in the education sector. In all these cases, I analyzed the process of industrial restructuring and reform that involved bargaining with the specific union studied.

23. Changes in the independent variable from one category to another were used as cut-off points when possible. When there were no changes in the independent variable, the cut-off was based on two diverse reform initiatives (for example, privatization and restructuring) when possible.

24. Pizzorno (fn. 16), 278.

25. Crouch (fn. 16), 161. The imperfection of the union agency may be a desirable goal for workers that select union leaders not just to carry their demands but also to articulate them and to calculate the benefits of intertemporal exchanges.

26. Henry Farber, "The Analysis of Union Behavior," in O. Ashenfelter and R. Layard, eds., Handbook of Labor Economics (Amsterdam: North-Holland, 1986), 1080.

27. Farber argues that the democratic constraints on the leadership range from cases where the limits are so loose that the leadership can maximize their objective function without regard to the constraints of the political process (dictatorship) to cases where the leadership is severely hampered by the political process and the need to answer the rank and file. Yet, he shows that the possibility of insurgency constrains leaders even in imperfect democracies. Ibid.

28. In a Hirschmanian sense, replacement by alternative leaders can be assimilated to "voice" within the same organization while the abandonment of the union by members is similar to his concept of "exit." Albert Hirschman, Exit, Voice, and Loyalty (Cambridge: Harvard University Press, 1970).

29. J. Samuel Valenzuela, "Labour Movements and Political Systems: Some Variations," in Marino Regini, ed., The Future of Labour Movements (Newbury Park, Calif.: Sage Publications, 1992).

30. Walter Korpi explains this pattern as the result of a trade-off between industrial and political resources. Korpi, The Working Class in Welfare Capitalism (London: Routledge and Kegan Paul, 1978). In the Latin American context, Zapata explains the empirical regularity of political strikes by the influence of the state on industrial relations. Zapata (fn. 14, 1986).

31. The cost of bargaining with nonallied union leaders is higher due to the lack of mutual trust while part of the agency costs may indirectly feed the coffers of opposition parties.

32. Mancur Olson, The Logic of Collective Action (Cambridge: Harvard University Press, 1971), 48.

33. Miriam Golden shows that coordination in wage bargaining is most likely when union monopoly is high because the competition for members "provides a strong incentive for unions to try to maximize their wage gains in order to retain members or to attract them away from competitors." Golden, "The Dynamics of Trade Unionism and National Economic Performance," American Political Science Review 87 (June 1993), 441.

34. Market reforms created a new critical juncture in the electoral politics of these countries (together with a simultaneous process of democratization or decentralization) that resulted in the emergence of new parties or divisions in the incumbent labor-based parties. Argentina experienced the emergence of FREPASO or Front for a Country with Solidarity (originating in a splintering of Peronism). Mexico saw the organization of the PRD or Party of the Democratic Revolution (also emerging from a division within the PRI). In Venezuela, Causa R grew to become a national political party and was followed by an array of new political options that reshaped the traditional two-party system.

35. Since union monopoly makes the union strong and the government can observe this strength, it seems unnecessary to enter a conflict to show it in terms of the external interaction, but partisan competition prompts union leaders to protest for reasons linked to their internal power.

36. Ellner (fn. 18).

37. Author's interviews with AD union leaders of different factions confirmed the account of Steve Ellner. See Ellner, "Organized Labor's Political Influence and Party Ties in Venezuela: Acción Democrática and Its Labor Leadership," Journal of Interamerican Studies and World Affairs 31 (Winter 1989).

38. Union leaders, government officials, and company managers, interviewed by the author, Caracas, June-July 1994, May-July 1996.

39. Ibid.

40. Author's interview with Ford's labor relations manager and AD union leader was confirmed by Consuelo Iranzo, Luisa Bethencourt, Hector Lucena, and Fausto Sandoval Bauza. See Iranzo, Bethencourt, Lucena, and Bauza, "Competitividad, Calificación y Trabajo: Sector Automotriz Venezolano" (Competition, qualifications, and work: Automobile sector in Venezuela) (Manuscript, Cendes, 1996 ).

41. Cite from El Nacional, January 9, 1991. Former Minister Gustavo Roosen and union leaders of FetraMagisterio and Fenatev confirmed this account derived from a press chronology, in interviews with the author, Caracas, June 1996.

42. The information on the Argentine cases is derived from a press chronology, union annual reports and other documents, interviews with union leaders of all factions, three ministers of labor and other government officials, as well as labor relations managers in the involved companies. The Argentine press chronology was elaborated in the archives of the newspaper Clarin and includes newspapers such as Clarin, La Nacion, La Rozon, Cronica, ElCronista Comercial, Pagina 12, and Ambito Financiero. Union sources included the internal constitution, annual minutes, and balances for the 1988 to 1994 period for SMATA, FATLyF, CTERA, FOETRA, and SUPE. Documents included collective bargaining contracts signed by the unions and approved by the Ministry of Labor, bill proposals, parliamentary information on introduced, modified, and passed bills. Interviews included SUPE and CGT union leader Antonio Cassia (Buenos Aires, 1993, 1995), FATLyF union leader Carlos Alderete (Buenos Aires, 1993), CTA union leader Victor De Gennaro (Buenos Aires, 1995), CTERA union leader Marta Marffei (Buenos Aires, 1995), ministers of labor Armando Caro Figueroa (Buenos Aires, 1994), Rodolfo Diaz (Buenos Aires, 1992, 1995), and Enrique Rodriguez (Buenos Aires, 1992, 1995), managers of labor relations, such as YPF's Roberto Teglia (Buenos Aires, 1995), and Telecom's Juan Giar (Buenos Aires, 1995).

43. The policy-shift of Peronism encouraged the formation of a splinter group, which together with some left-wing parties, formed, in turn, a new opposition political party in 1992 that would became the FREPASO in 1994. Among its core organizers was a group of union leaders that had broken ties with the Peronist union movement and organized a new small confederation, the CTA (Congress of Argentine Workers) in 1992.

44. State imposed limits on strike activity discard the usefulness of using strikes as a measure of militancy but do not imply that militancy does not take place. Mexican workers have held illegal strikes, used strike petitions as a threat to show their militancy, and resorted very often to sit-ins, demonstrations, and even extreme measures such as going naked or on hunger strikes. On repertoires of protest, see the account of Maria Lorena Cook, Organizing Dissent (University Park: Pennsylvania State University Press, 1996). For a historial analysis of the evolution of union strategies, see Alberto Aziz Nassif, El estado mexicano y la CTM (The Mexican state and the CTM) (Mexico City: Ed. La Casa Chata, 1989); Ilán Bizberg, Estado y sindicalismo en México (State and unionism in Mexico) (Mexico City: El Colegio de México, 1990); and Kevin Middlebrook, The Paradox of Revolution (Baltimore: John Hopkins University Press, 1995).

45. Graciela Bensusán confirms interviews with union leaders and government officials. Bensusán, "Los determinantes institucionales de la flexibilización laboral" (Institutional influences on labor flexibility), Revista Mexicana de Sociología 1 (1994).

46. For instance, the Revolutionary Confederation of Workers and Peasants (CROC) and the Revolutionary Confederation of Mexican Workers (CROM) explicitly boycotted CTM protest against wage ceilings in the Congress of Labor and were publicly rewarded by the government. Subsequent exit of unions from the CTM into preferred organizations, most notably the CROC, increased the pressure on its leadership. See my press chronology, which derives from the archives of Entorno Laboral and which includes newspapers such as La Jornada, Excelsior, El Sol de Mexico, Uno mas Uno, and Reforma.

47. Ruth Berins Collier and James Samstad analyze the development of the FESEBES and the "new unionism" in "Mexican Labor and Structural Reform: New Unionism or Old Stalemate?" in Riordan Roett, ed., The Challenge of Institutional Reform in Mexico (Boulder, Colo.: Lynne Rienner, 1995). Interviews with union leaders, government officials, and company managers confirmed the numerous accounts on the strategy of this union. For telecommunications, see Enrique De La Garza, "Sindicato y restructuración productiva en México" (Union and labor restructuring in Mexico), Revista Mexicana de Sociologia 1 (1994); and De La Garza and Javier Melgoza, "Reestructuración y cambio en las relaciones laborales en la telefonía mexicana" (Restructuring and change in labor relations in the Mexican telephone industry), in Jorge Walter and Cecilia Senen González, eds., La privatización de las telecomunicaciones en América Latina (Privatization of telecommunications in Latin America) (Buenos Aires: Eudeba, 1998). For electricity, see Javier Melgoza, "El SME y la productividad: Los saldos de la negociación" (SME and productivity: Outcomes of bargaining), Polis 93 (1994).

48. Although my account derives from a press chronology and interviews with union leaders, government officials, and experts, the process of modernization and democratization of the union has been widely studied. See, for instance, Cook (fn. 44); and Joe Foewaker, Popular Mobilization in Mexico, the Teachers' Movement 1977-87 (Cambridge: Cambridge University Press, 1993).

49. Interviews and press releases are confirmed by Marisa Von Bulow, "Reestructuración productiva y estrategias sindicales. El caso de la Ford en Cuahutitlán 1987-1993" (Production restructuring and union strategy: The case of Ford in Cuahutitlán 1987-1993) (M.A. thesis, Facultad Latinoamericana de Ciencias Sociales-Sede México, 1994); and Jorge Carrillo, "La Ford en México: Restructuración industrial y cambio en las relaciones sociales" (Ph.D. dissertation, El Colegio de México, Centro de Estudios Sociológicos, 1993).

50. Cárdenas split from the PRI because he rejected the policy shift and failed to be selected as its presidential nominee. He launched a new coalition, later to be renamed as the PRD, and he delivered the strongest electoral blow to the PRI by dramatically reducing its share of votes (and even claiming victory stolen by fraud) in 1988.

51. Fabio Barbosa gives a graphic account of the process that confirms interviews and press releases. Barbosa, "La reestructuración de Pemex" (The restructuring of Pemex), El Cotidiano 46 (March-April 1992).

52. These two cases related to Burgess's argument about the PRI imposing external costs on union leaders' decisions. Burgess (fn. 15).

53. International Labour Organization, Relaciones de Trabajo en Venezuela (Labor relations in Venezuela) (ILO report, 1991).

54. Not only did Causa R grow in the unions and in the 1994 elections, but also many political outsiders, including current President Chávez, challenged the traditional political parties after that election.

55. Geoffrey Garrett and Christopher Way, "The Rise of Public Sector Unions, Corporatism and Macroeconomic Performance, 1970-1990," in Barry Eichengreen and Jeffry Frieden, eds., The Political Economy of European Integration (New York: Springer-Verlag, 1995).

56. Ivan Nuñez, "Sindicatos de maestros, Estado y Políticas Educacionales en América Latina" (Teachers unions, state and education policy in Latin America), in M. L. P. B. Franco and D. M. L. Zibas, eds., Final do Seculo: Desafíos da educacao na América Latina (End of the century: Challenges for education in Latin America) (Sao Paulo: Cortes Editora, CLACSO REDUC, 1990).

57. Cook (fn. 44).

58. Collier and Collier (fn. 2).

59. Collier and Collier (fn. 14, 1979).

60. Zapata (fn. 14, 1986).

61. The new political parties emerging during the policy shift--FREPASO in Argentina, PRD in Mexico, and Causa R in Venezuela--built alliances with labor unions. Labor leaders, however, were more aware of the costs of corporatism once the state started its retrenchment, and the terms of new associations tended to be more fluid than in the past, thus affecting the extent of future "loyalties."

62. Maria Victoria Murillo, "The Corporatist Paradox: Labor Parties and Labor Reform in Latin America" (Paper presented at the conference "Institutional Reforms, Growth, and Human Development in Latin America," Yale Center for International and Area Studies, April 16-17, 1999).

63. Roberto Michels, Political Parties: A Sociological Study of Oligarchical Tendencies in Modern Democracy, trans. Eden and Cedar Paul (New York: The Free Press, 1962), 365.

64. Ibid., 172.

65. These implications are consistent with Hirschman's claim on the superiority of voice over exit as a mechanism for improvement in certain contexts when exit is not easily available or could provoke the demise of the organization.

[an error occurred while processing this directive]