Abstract
Public sector unions are increasingly becoming the hegemonic contemporary labor actor in terms of membership and militancy in both advanced and emerging economies. However, political economy lacks a unified theoretical approach to study mobilization by state unions. The analysis of public sector union politics has been largely separated by regional (United States vs. Europe vs. Global South) and disciplinary (American politics vs. comparative politics/political sociology) divides. We contend that though both public and private workers belong to the subaltern classes, public sector union politics and mobilization have different foundations than in the private sector. Unlike private unions, state labor mobilization is essentially driven by what we call the “reverse economic cycle” (militancy increases in bad—rather than good—economic times), by the legal enforcement of bargaining rights (which in contrast to the private sector substantially varies across and within democracies), and by the likelihood of a political exchange between labor and the government. Argentine teachers between 2006 and 2019 provide an ideal laboratory to test this argument through a multilevel (i.e., national and subnational), mixed-methods strategy, which includes comparative and statistical assessments.
Modern public sector unions are a lightning in the landscape of quiet and shrinking labor movements in both advanced and emerging economies. While private sector unions are generally in retreat, public sector labor is a consistent force in most of the democratic world. It is widely argued that structural trends that undermine nonstate labor amid globalization, such as enhanced international economic competition, widespread flexibilization of employment conditions, and retrenchment in traditional manufacturing strongholds, have affected much less, or largely sidelined, public unions. 1 During the last fifteen years, public sector unions and teachers (the largest state workers’ group in almost every country) have staged prominent strike waves in places as diverse as the United States, South Africa, Israel, the Nordic countries, China, and Argentina. 2 The president of Peru elected in 2021 built his career in a teacher union, and state labor and teachers became the organizational backbone of the Tunisian revolution that sparked the Arab Spring in 2011. Ruth B. Collier has recently stated that across Latin America “the labor movement has come to be centered in the public rather than the private sector, with teachers emerging as the largest organized group.” 3 Public sector unions have been the main force behind massive strikes in the United Kingdom and France in the first half of 2023, the largest in at least two decades. In brief, most twenty-first-century big labor battles are, essentially, public sector union clashes.
Given the relevance of public sector conflict in contemporary political economies, it is surprising that we lack a general theoretical approach to study state unions. For some scholars, they tend to be more combative because they are less challenged by price competition and enjoy employment protection. But others insist that they tend to be more regulated/controlled by the government, and composed by white collar and educated workers that are prone to moderation and “yellow” unionism. 4 In short, it is not clear why and in what ways public sector unions differ from private sector labor organizations. Though the current rise of public unions is reflected in various aspects such as union density and contract coverage, this article focuses on a key contemporary feature: mobilization capacity.
The increasing primacy of public unions has been noted both in advanced and developing economies alike. However, the study of public sector union politics is largely separated by regional (United States vs. Europe vs. Global South) and disciplinary (American politics/economics vs. comparative politics/political sociology) divides that scarcely speak to each other. In the United States, a “rent-seeking,” neo-utilitarian approach critically emphasizes the growth of public unions. 5 It holds that state labor has mostly used its gained legal rights to stage workplace and political mobilization, and protect their rents in the form of wages and expanded employment. The (mostly European-based) comparative politics/political sociology subfield largely sees rising state unions as part of larger labor movements that are national political actors. In this literature, public unions can affect general wage moderation because of their noncompetitive nature. Thus, the key issues are the conditions (i.e., common government-union political identity and unified labor movements) in which the “political exchange” between government and state labor is more likely. 6 Finally, scholars in Latin America and the Global South have pointed out how, in those cases where the union movement is relatively strong at a national level (such as Argentina, Brazil, South Africa, and Tunisia) public unions increasingly dominate national labor confederations and social contestation. The question here is whether party-union ties and varieties of top-down state-corporatist arrangements (which generally date from the predemocratic period) gave way to more autonomous patterns of mobilization that can truly benefit state labor and workers. 7
In this study, we aim to build the first steps for a general theoretical approach to the study of public sector unions. We argue that a more unified view of public sector union politics should begin by disentangling the similarities, and the key differences, between public and private sector trade unionism. In a capitalist society, both public and private labor constitute a “fictitious commodity” that is sold for a price, 8 and they are thus part of the subaltern classes. In both cases workers can try to compensate in the political realm for their “natural” disadvantage of being deprived of the means of production/subsistence, especially under democratic governance. However, we contend that public unionism radically differs in (a) the contemporary key role of state regulation in collective bargaining and (b) the labor market foundations for collective action, as public labor is less threatened by international and domestic competition, and by retaliatory layoffs (in view of its statutory protections).
As we will elaborate below, these differences shape distinct conditions for collective action. Put differently, private and public workers belong to the same social class but have different mobilization patterns. First, against established private sector–based economic theories of strike mobilization, we put forward a “reverse economic cycle” hypothesis for state unions. We argue that public sector workers tend to protest less in the upper part of the economic cycle, when the economy grows, unemployment is low, and budget constraints are less stringent. Second, we contend that in a free union organizing environment, mobilization declines when legal collective bargaining (at the national and/or subnational levels) is mandatory and thus labor disputes become institutionalized. Finally, in the line of European and Global South corporatist theories, we claim that public sector militancy weakens when government leaders share a party identity with relatively unified labor unions, and a political exchange of moderation for wage and nonwage payoffs unfolds. Conversely, contention will be stronger in times of fiscal adjustment and high unemployment, when public sector activists who generally cannot be fired will fight for scarce resources. Militancy will also increase when collective bargaining is hindered by state regulation (which, unlike in the private sector, is frequent under democracies), when government and labor leaders belong to different political parties, and labor is divided internally and/or among many unions.
Argentine teachers provide an ideal laboratory to test our argument. They are the paradigmatic example of a public sector union more powerful after the economic liberalization of the last quarter of the twentieth century than before. Argentina is a robust federation in which teacher mobilization and the legal status of bargaining rights vary dramatically across the twenty-four provinces. It also harbors the strongest labor movement on the continent, which is politically close to one of the mainstream parties, the Peronist movement. These aspects allow us to test the effects of both legal and political exchange factors on public sector union militancy. Argentina is a democracy, public unions are autonomous actors, and elected politicians fear labor strife. Teachers are the largest public sector union in most countries and play a central role in the politics of education. 9
This study will analyze teachers’ mobilization in Argentina from 2006 and 2019, that is, under the populist-left, Peronist governments of Néstor Kirchner (2003–7) and Cristina Kirchner (2008–15), and the center-right administration of Mauricio Macri (2015–19). Although we focus on strikes as the paradigmatic weapon of labor mobilization, our qualitative analysis includes other forms of collective action by public sector workers, such as lobbying and participation in policy formulation, street demonstrations, and road blockades. As Roberto Franzosi argues in his monumental mixed-methods study on labor conflict in postwar Italy, purely quantitative work tends to downplay the most fundamental aspect of strikes: the strategic interaction between state and social actors with specific interests and capacities. 10 Therefore, our research design recurs to a multilevel (i.e., national and provincial), mixed-methods approach. 11 We use, first, the comparative method, in its diachronic, most-similar-cases design variant, to assess national teachers’ mobilization under the Kirchner and Macri administrations. Second, drawing upon strike information provided by the Ministry of Labor and Social Security, and unique data on 331 yearly teachers’ wage bargaining rounds in the twenty-four provinces between 2006 and 2019, we estimate a series of negative binomial statistical models that adjust for clustering in the errors.
We show that Peronist governments’ sanction of mandatory collective bargaining rights for a federal minimum wage, and their political exchange with the dominant teachers’ confederation CTERA (Confederation of Education Workers of Argentina), substantially lowered nationwide strikes relative to the Macri center-right administration. Our statistical results also substantiate that the reverse economy cycle (i.e., fiscal adjustment), provincial “duty to bargain” laws, and a common political identity between the governor and unified teacher unions, largely explain variation in subnational teachers’ strike activity. The first part of the article presents evidence on the rise of state labor in the twenty-first century in the United States, Europe, and the Global South. Subsequently, we lay out our theoretical approach to public sector union mobilization in democratic countries. The second part deploys our multilevel, mixed-methods analysis of Argentina: the comparative assessment of the politics of teachers’ national mobilization under the Kirchner and Macri governments, and the statistical analysis of strike activity in the twenty-four provinces. We end with a discussion of both qualitative and quantitative findings, and the implications of our empirical analysis for the broader study of public sector unions in contemporary political economies.
The Rise of Public Unions in the Twenty-First Century: Evidence and Theoretical Perspectives
Public sector unions have displaced private unions as the locus of labor power in many democratic countries. This is evident, for example, in union density trends within each sector. In advanced countries for which we have systematic historical data, union membership in the private sector has been consistently declining since the last quarter of the twentieth century. By contrast, union density in the public sphere has been continually growing in the 2000s. This phenomenon is broadly similar in union-weak Liberal Market Economies (e.g., United States, United Kingdom) and in union-strong Coordinated Market Economies (e.g., Germany, Norway, Sweden). See Figures 1 and 2. For example, in the United States the trajectory is eloquent. Private union density declined 78 percent between 1970 and 2017 (from 29 percent to 6.5 percent). By contrast, public union membership increased by 60 percent in that same period (from 19.1 percent to 48.7 percent). In Norway, a traditional labor stronghold, density in the private sector went from 43 percent in 1970 to 37 percent in 2013, whereas in the public realm it climbed from 35 percent to 55 percent. In the United Kingdom, density in the private sector plummeted from 45 percent in 1995 to 13 percent in 2016. By contrast, public union membership increased from 40 percent in 1979 to 57 percent in 2016. 12
Union density in the private sector (1970–2017), selected advanced countries. Source: J. Visser, ICTWSS Database, version 6.1 (Amsterdam: Amsterdam Institute for Advanced Labour Studies, University of Amsterdam, November 2019).
Union density trends in the public sector (1970–2017), selected advanced countries. Source: J. Visser, ICTWSS Database, version 6.1 (Amsterdam: Amsterdam Institute for Advanced Labour Studies, University of Amsterdam, November 2019).
We lack longitudinal data to assess membership trends in the Global South. Yet the International Labour Organization (ILO) has collected sectoral data on labor conflict since 2010. Figure 3 compares the share of strike days (2010–20) in relation to the share of total employment in 2017 for both private and public sectors in developed and developing (Latin America plus South Africa) democracies. The data show that there is a “public sector conflict premium”: in both groups of countries, state unions are more contentious than what their share of employment would predict. In developed democracies, the difference is of about 8 percentage points. But in Global South countries, the gap is impressive: public sector workers share, on average, only 13 percent of employment, but they explain 51 percent of strike days. The figure also reveals that private sector unions, while often in retreat, still seem to hold in advanced countries, and wage the majority of strikes. 13 Conversely, their demise in the Global South countries is clearly evident.
Public sector workers’ share of total strike days (2010–20) and of total employment (2017) in democracies. Source: Strikes ILO, % employment, World Bank Bureaucracy Database.
In sum, the increasing relevance, and often preeminence, of the public sector in the labor movement is evident in the United States, Europe, and in many union-strong countries of the Global South. In the United States, in 2009 and for the first time, the number of workers that belong to state unions surpassed those in the private sector. 14 Until the recent 2023 strike wave, public sector disputes were in practice the only major type of labor conflict in the contemporary US economy, including memorable “wars” between public unions and Republican governors in Wisconsin, Ohio, and other states, and a mass teacher strike wave in 2018. 15 Two camps have dominated the study of public unions in the United States. A “rent-seeking,” neo-utilitarian perspective largely sees state labor as a special interest, and bargaining laws (which make contracts legally binding, and in practice grant union monopoly) as a major force behind public sector unionization and mobilization. 16 A more critical perspective underscores the “divided” (i.e., public vs. private) nature of the American labor movement. 17 For this literature, public unions were crucial ambits for the emergence of the “social movement” unionism that revitalized American labor more recently. 18 Public sector unions benefit from the territorial dispersion of state offices, which can provide social mobility for women and nonwhite workers, 19 and constitute an ideal vehicle for the expansion of community-linked labor demands.
Public sector mobilization has substantially risen in Europe since the early 2000s, outpacing private conflict in many countries. 20 Unlike in the United States, by the mid-twentieth century public sector unions and bargaining were legal, and state unions were part of encompassing labor movements. 21 Political exchange/neo-corporatist theories, the dominant approach in Europe in the 1980s and 1990s, 22 hold that the incidence of strikes declines when labor-based or social-democratic parties control governments and their associated unions are rewarded through stable patterns of income distribution, social policy, or organizational advantages. In the famous Korpi and Shalev dictum, “class conflict moves from the industrial to the political arena.” 23 Political exchange has two general prerequisites. 24 The first one is low union fragmentation. Monopoly unions that cannot be outflanked mitigate intraclass distributional struggles and have more incentives to cooperate. The second is a common identity between labor-based parties and unions: left or popular governments reassure unions that their restraint will be compensated in the future. In the 1990s, a new round of works in the neo-corporatist vein argued that in the nontradable economic activities—and specially in the state—in which there is nothing “akin to a world market price,” 25 governments are potentially free to set wages. Thus, a rational public worker has more incentives to push up militancy and wages. In view of statutory protections and the absence of price competition, the subordination of public unions to the party-labor alliances of Coordinated Market Economies (as opposed to Liberal ones) becomes even more important than in the private sector. 26
Finally, public sector unions are also currently key players, or simply dominant in national labor confederations, in countries of the Global South such as Brazil, Argentina, Tunisia, South Africa, and (only in the case of teachers) Mexico. 27 In all these cases public unions and teachers have been part of historical alliances between labor and national-popular or leftist parties in the twentieth century—though generally in a less central role than private unions. These coalitions were often shaped by “state corporatism,” that is, the top-down regulation and monitoring of unions by the party or the state. 28 In Brazil, scholars have noted how public unions have displaced historically dominant metalworking unions, which gave birth to Lula's PT (Worker's Party). 29 In South Africa, the ANC (African National Congress)-affiliated COSATU (Congress of South African Trade Unions) has become “an almost exclusively public sector trade union federation.” 30 In Tunisia, public sector unions form the backbone of the powerful UGTT (Tunisian General Labor Union), which became the organizational fulcrum of the revolt that toppled Ben Ali's authoritarian regime in 2011. 31 Except for Mexico, 32 in all these countries state labor (outside the public business and industrial sector) arguably became relatively stronger after economic internationalization consolidated in the late 1990s.
A key question in Global South labor studies is whether public sector unions and teachers, traditionally allied with (and to some extent controlled by) the national-popular parties, achieve autonomy and mobilization capacity in more open polities. In Tunisia and Mexico, the UGTT and the SNTE managed to sever ties with the formerly autocratic parties and became autonomous forces (though of a very different kind) in postauthoritarian politics. 33 In South Africa, Brazil, and Argentina, public sector union alliances with the largest popular parties, the ANC, the PT, and Peronism, are still solid. Indeed, Murillo has laid out a version of the political-exchange framework for Latin America. In her formulation, public union protest will be lower when competition among unions, and party rivalry within labor organizations, are both low. 34
The Nature of Public Sector Union Politics and Mobilization
As Paul Johnston has argued in his analysis of government unions in the United States, our conceptual tools for thinking about labor relations are shaped by their origins in the private sector. 35 Thus, the quest for a more unified approach to state unionism should start by sorting out the essential features that public and private unions (and their workers) have in common, and those in which they differ. In a Marxist and Polanyian vein, both public and private workers are deprived from the means of production/subsistence. They sell their labor to formal employers as a commodity in the marketplace. Indeed, in both cases workers need to turn an individual wage relation into a collective form of negotiation to increase power. As individuals who sell their labor, private and public sector workers essentially demand better wages and working conditions. 36 Classical contemporary sociological approaches to class formation and power from both Marxist and Weberian bents neither fundamentally differentiate private and state workers 37 nor conceptualize public workers under capitalism as a different class. 38 Moreover, as corporatist perspectives in both Europe and the Global South convey, political exchanges to compensate for labor structural disadvantages may be an option for both private and public workers.
We contend that public and private sector unionism crucially diverge in two fundamental aspects that affect unions’ capacity and workers’ incentives to mobilize. The first refers to the role of state regulation in collective bargaining. Democratic regimes essentially accept free (and generally, in practice, mandatory) collective bargaining with employers when private sector workers are organized. By contrast, public sector bargaining is much more regulated. Second, government unions are substantially different in that (a) they are much less vulnerable to price competition and outsourcing employer strategies and (b) they enjoy statutory protections against layoffs, while private unions do not. These theoretical similarities and differences are summarized in Table 1.
Theoretical Assessment: Labor Relations in the Private and Public Sectors.
Our central claim is that these patterns of convergence and divergence bring about distinctive conditions and incentives for contentious collective action. Public workers, and their unions, are part of the subaltern classes in the traditional, “structural” sense. The key difference with private workers is the combination of factors that shape their mobilization capacity, which are laid out in the following hypotheses.
Public labor militancy is weaker (stronger) when the economy grows (shrinks), unemployment is low (high), and there is a fiscal surplus (deficit).
As we noted before, both the European-based studies and the neo-utilitarian American politics approach stress the noncompetitive nature of state unions as a paradigmatic nontradable actor, and their potential to drive up wages unconstrained. Public workers enjoy statutory protection and the state, unlike private firms, cannot go bankrupt and cease to exist. Most public services (trash collection, security, water sanitation, education) cannot be totally replaced or outsourced. However, much of this literature forgets too hastily that public union militants may not be restrained by international economic competition, but are severely conditioned by state budgets. As R. Freeman puts it when he analyzes US public sector unions, “The budget serves as a disciplinary device in the public sector as does market demand in the private sector.” 39 We therefore propose that actors’ incentives in the public sector are the opposite of those in labor disputes in the private economy: in good times rational politicians benefit from delivering resources, which (unlike in the case of business) are not self-owned or need to be reinvested, to a potential disruptive constituency. In bad times, however, the scarcity of budget resources exacerbates wage disputes. Private unions essentially extract surplus in periods of economic growth when low unemployment prevents retaliation. 40 Conversely, public unions tend to be “rationally” appeased by politicians in good economic times but, in view of their protection against layoffs, fight more for limited resources in recessions.
Incentives for public sector labor mobilization decline with the institutionalization of “duty to bargain” rights at the national and/or local provincial levels. Conversely the absence of mandatory bargaining rights at both levels increases labor militancy.
As stressed in the American politics literature, variations in state regulations at the national and subnational levels are key to understand public sector unionism. 41 Indeed, public sector collective bargaining around the globe, even under democracies, has been historically severely limited. These limitations have multiple origins, among them the conception of public workers as providing “essential services,” and the reluctance of politicians to bargain with their own workers. 42 Restrictions remain in place today specially outside the core union-strong countries of Europe: while 70 percent of contemporary democracies restrict collective bargaining in the public sector in some way, only 27 percent does so in the private sector (and in this case restrictions tend to be minor). 43
Public sector wage bargaining regulations at the national and (especially in federal countries) subnational levels range from no legislation or formal prohibition, permission but no obligation, “right to meet and confer clauses” (which exclude mandatory wage bargaining), and exclusion of certain types of public service workers (i.e., police), to employer duty-to-bargain provisions. 44 They may by complemented by the regulation of strikes. For the sake of simplicity, in this study we draw the broad distinction between government “duty to bargain” clauses, in which state bureaucrats must negotiate wages with public unions free from any formal impediment, versus situations in which state employers can decide wages unilaterally. 45 Following classic labor institutionalization theories, 46 we hypothesize that when duty-to-bargain rights are consolidated, incentives to strike mobilization in the public sector will tend to be lower. When labor has formal institutional mechanisms to advance its interests in the policy process, the probability of direct action diminishes.
Public sector militancy declines when monopoly unions, in which internal leadership competition is low, share a partisan identity with both national or subnational governments. Conversely, militancy grows when a fragmented union movement, with high internal leadership competition, does not share a partisan identity with national or subnational governments.
As is the case of private labor, public sector unions are essentially political creatures. They are obviously interested in short-term rents such as wages and employment. Yet, often public sector union leaders are militants immersed in the broader political struggle, and they value the fate of the parties to which they belong, and of friendly administrations—a fact particularly underscored in the European and Global South corporatist literatures. As such, they may have incentives to exchange short-term restraint for nonwage benefits (e.g., social policy or future pensions) and/or organizational payoffs (e.g., specific union privileges in bargaining, or the administration of social services or skill formation programs) in the long run. Thus, building on the neo-corporatist tradition and the work of Murillo, 47 we argue that the political exchange between governments and labor, and its organizational prerequisites (i.e., low interunion fragmentation and internal competition, and common partisan identity), are decisive to explain public union militancy. While in the United States, public sector unions tend to obtain exclusive rights of representation, this is not the case in many European and Latin American countries, where many unions (often of different political loyalties) may compete in state administrations.
In the light of the above hypotheses, Figure 4 draws “ideal” theoretical paths for both high and low public sector union militancy. Democracy and the legal existence of unions and strikes function, in methodological terms, as our scope conditions—that is, the domain in which causal processes are presumed to be homogenous. 48 When sustained grievances or wage demands are manifested, unlike in the private sector, an economic expansive context of fiscal leeway and low unemployment builds a more favorable scenario for state labor accommodation. Beyond economic conditions, mandatory bargaining rights set the legal/institutional path for labor moderation. Thus, in our stylized model duty-to-bargain rights serve as an antecedent condition that facilitates the political exchange: when bureaucrats can set wages unilaterally, government-labor settlements, based on common political identity and unified public sector unions, though not impossible, are less likely.
Paths to high and low public sector union conflict.
Conversely, when economic recession and fiscal adjustment hit hard and unemployment goes up, state unions, whose activists (unlike those in the private sector) generally cannot be fired, have stronger incentives to fight for scarce resources. Plus, when no bargaining is allowed in the public sector, the possibilities for an institutionalized political exchange between elected politicians and public sector union leaders strongly diminish. Furthermore, if the labor movement is fragmented, if unions are internally competitive, and if unions do not belong to the government party, then the motivations for overall union restraint are even weaker. Naturally, these alternative paths constitute theoretical opposites. Reality should be messier.
Explaining Teachers’ Mobilization: A Multilevel and Mixed-Methods Strategy
Public Sector and Teacher Unions in Argentina
Argentine public sector unions and teachers are a model case story of a recent rise to political preeminence. They were minor actors during the inward-oriented economic model in place until the 1990s. The initial corporatist incorporation of unions by Peronism in the mid-forties was based on private sector industrial unions, not state unions, and excluded teachers. 49 Public unions only obtained collective bargaining rights in 1998 (central state administration), almost fifty years later than their private counterparts, and in 2008 in the case of teachers at the national level. 50 CTERA, by far the largest national teachers’ organization and close to the left-wing of the Peronist movement, only emerged as a national actor in the early 1970s. It is the umbrella confederation for base teacher unions in the twenty-four Argentine provinces. 51 Other unions also organize teachers at the provincial level, and the radical (mostly Trotskyite) left is the main challenger to Peronist dominance both within and outside CTERA provincial base unions. Militancy is led by public education (around 70 percent of all Argentine nonuniversity teachers), 52 especially at the primary level. Teachers are by some distance the most combative unions in the country. They share 42 percent of total individual workdays lost in strikes in the period under study, twice the second sector (noneducation state employees). 53
Argentina is a robust federal country in which education at the initial, primary, and secondary levels is run by the provinces. Hence, it lacks a national framework that regulates all public wage negotiations. Teachers’ main wages and working conditions are decided at the provincial level. It is worth stressing that, though hypotheses 1–3 above refer to militancy in general, in this study we test them quantitatively with data about the paradigmatic weapon of labor action: strikes. Subnational variation of teachers’ strike activity in Argentina is impressive: it ranges from more than 175 workdays lost per teacher in the Patagonian provinces of Santa Cruz and Neuquén between 2006 and 2019 (almost a year of classes), to 4.58 days in the northern province of Tucumán (Table 2).
Total Workdays Lost per Teacher (2006–19).
National-Level Comparative Assessment
Our multilevel, mixed-methods strategy begins with an assessment of the two Kirchner (2003–15) and Macri (2016–19) governments based on the most similar, diachronic comparative method. The general context of within-country comparison provides essential controls—such as culture, type of political system (in our case presidential and federal), and the national organization of teachers’ unionism. As Gerring argues, the diachronic comparative method, as long as the two periods are fairly close together, facilitates comparability. It is also especially suitable when X and Y are closely linked in time (such is the case in this study), and the purported causes are not historically “removed” from the outcome of interest. 54
Néstor Kirchner, leader of a leftist faction of the traditionally populist Peronist party, won the 2003 presidential elections embodying a discourse that repudiated the prior neoliberal decade. Only eight days after he took office, the government intervened in a harsh, monthslong teacher strike and wage conflict in the province of Entre Ríos. It provided federal funds and reached out to the local and national union for a final settlement, 55 which signaled a future mark of the Kirchnerista governments of 2003–15: the convergence with the teacher unions in general and with largest confederation CTERA in particular. Between 2003 and 2007, the national government passed a series of laws and decrees that cemented this confluence for the years to come. Among the most important, the Education Finance Act (26.075) established a 6 percent GDP goal of spending for 2010—which would be fulfilled. The National Education (26.206) and Technical Education (25.864) laws strengthened the role played by the national state in education planning and content formulation, and reconfigured the system of vocational training at the secondary and tertiary, nonuniversity levels. Critically, the Education Finance Act mandated the creation of the National Wage Board (NWB) to discuss a federal minimum income and labor conditions with national teacher unions. The government also established a National Compensation Fund to finance the provinces that could not afford that wage floor. The bipartite NWB met for the first time in 2008. At the same time, the Ministry of Education included CTERA and its base unions in the Federal Network of Teachers’ Skill Formation and designed a new scheme to prop up pensions in the eleven provinces still covered by the national social security system. 56
Analysts have labeled the confluence between the national teacher unions led by CTERA and the Kirchner governments as an “alliance,” 57 one that resulted in “collaborative policymaking.” 58 The importance of the NWB as a coalitional fulcrum with CTERA cannot be underestimated. First, the legalization of bargaining rights at the national level pushed up the salaries in the less developed provinces through the National Compensatory Fund. Second, in the rich provinces, the federal minimum wage percentage increase served as a parameter (for many local unions as a baseline) for the subsequent subnational bargains. Of course, the system was not without tensions, especially when the economy stalled after 2012. Governors distrusted a policy that pushed teachers’ wages up but left the rest of provincial state workers with no national coverage or financial support. When the economy slowed in 2012 and 2013, peak negotiations at the NWB tuned sour. CTERA rejected government wage offers. President Cristina Kirchner openly criticized the union, and the government (after negotiations at the NWB collapsed) set the minimum federal wage unilaterally. Consensus over the teachers’ minimum wage between the government and CTERA, however, returned in the 2014 and 2015 rounds of negotiations at the NWB. Overall, the above legislation and wage centralization at the NWB, together with the 2002 economic stabilization, set the stage for an 83 percent real growth in national education investment between 2003 and 2007, and a 60 percent teacher real wage growth between 2003 and 2010. 59
The center-right front Cambiemos took power in 2016 and, led by President Macri, initially developed a conciliatory stance with CTERA leaders—by that time strongly aligned with Kirchnerismo. In a context of rising inflation and currency depreciation, the incoming government closed a relatively good deal for teacher unions in the 2016 National Wage Board. However, tensions mounted when the Macri administration uncovered initiatives to promote what it considered quality or performance education reforms, openly rejected by CTERA. Moreover, in 2017 Cambiemos suppressed by decree both the National Wage Board and the Compensatory Fund. Though the National Education and Technical Education acts were not formally revoked, in practice, the Macri government largely rolled back the education policy recentralization of the previous years. It closed most national programs or transferred them to the provinces. The national FONID (Teachers Incentive Fund), a federal supplement to all Argentine teachers’ salaries notably strengthened in the prior decade, was frozen in a context of rising inflation. In short, wage financing and the management of teachers’ labor relations were again entirely in the hands of the provincial administrations. In April 2017, police repression of CTERA activists who wanted to set up an “Itinerant School” in front of the national Congress demanding the return of the NWB signaled the end of the national settlement with the teacher unions.
In sum, the national teachers’ confederation CTERA mutated from being a “resistance” actor in the 1990s, at the forefront of the struggles against education decentralization and neoliberal adjustment, to being a neo-corporatist actor in the 2000s, thoroughly involved in education policy formulation and the recentralization of wage bargaining. After 2015, however, education policy and financing were again largely decentralized to the provinces under the Macri administration, and the national government-labor political exchange unraveled.
Figure 5 depicts national strike trends in Argentina between 2006 and 2019 (i.e., nationwide or interprovincial strikes organized by CTERA). It is difficult to assess our reverse economic cycle hypothesis with descriptive data for only fourteen years/observations—we test it quantitively below. Figure 5 suggests, however, that national strike mobilization begins its upward trend in 2012, precisely when the economy began to stall after almost a decade of sustained GDP growth in the context of a commodity boom led by soy exports. Indeed, after Argentina's 2014 currency devaluation, erratic growth, economic recession, and fiscal stringency became a constant in the years to come. Correspondingly, public sector teachers’ national mobilization increased steadily except for one year (2015). Figure 5 seems much clearer about how strike trends were heavily shaped by the legalization of national bargaining rights and the collaborative policymaking in skill formation and pensions, that is, the legal (hypothesis 2) and political-exchange (hypothesis 3) dimensions. After the institutionalization of the NWB in 2008, confederate-level organized strikes almost disappeared for three years—a rarity in Argentina. 60 It peaked again in 2018–19 just after President Macri revoked national bargaining rights. More generally, both the number of strikes and workdays lost skyrocketed in relative terms under Macri, suggesting the vital role played by the national-level political exchange between labor and government during 2003–15.
Teachers’ individual workdays lost (left axis, in thousands) and number (right axis). National strikes (2006–19).
In fact, the two main organizational prerequisites for neo-corporatist convergence, common party identity and union leadership monopoly, operated under the Peronist governments in 2003–15. First, CTERA had been, since its birth in 1973, permeated by a populist-left ideology, which the Kirchners embodied in the 2000s. The Peronist faction “Lista Celeste” had consolidated its dominance of the union in the early 2000s. 61 Second, though it is commonly argued that Argentina has a plural teacher union structure, in the 2000s that was not the case at the national level in practice. By the late 1980s CTERA, as a confederation of subnational unions originating “from below,” had active organizations in almost all Argentine provinces. 62 Thus, CTERA benefited vis-à-vis other unions when education (and teachers’ labor relations) was completely decentralized to the provinces in the 1990s. Its already-developed base provincial unions started to gain even more representativeness. Wage bargaining decentralizations undermined other nationally organized teacher unions like UDA (Union of Argentine Teachers) and AMET (Association of Technical Education Teachers), which were deprived of their national bargaining ambit, and lacked a solid provincial-level structure. The national hegemony of CTERA was implicitly recognized in the design of the NWB in 2007, in which CTERA took the majority of labor spots (five of nine) according to “its representation coverage.” The Kirchner governments knew that at the national level CTERA was the “only game in town,” that is, the only union in a position to launch nationwide strikes. On the other hand, CTERA, in practice the only federal teacher union, did not fear being outflanked by a more militant national organization.
In short, Argentina's national story between 2003 and 2019 is close to the ideal theoretical paths to public sector conflict sketched in Figure 2: contention was low when bargaining rights were bestowed in a context of fiscal profligacy, and a political exchange between the Kirchner governments and CTERA—in practice, a monopoly national union of Peronist identity—unfolded. When, after 2012, fiscal conditions worsened, bargaining rights were revoked, and the political exchange collapsed, nationally organized militancy resumed virulently. Yet, so far this assessment leaves the persistent subnational teacher strikes unexplained—80 percent of total workdays lost in the period under study.
Quantitative Analysis
Data and Hypotheses
In order to explain subnational variation in teacher union militancy, we conduct a statistical analysis of strike occurrence and intensity in the twenty-four Argentine provinces. We collected and generated unique data on teacher strikes, provincial unemployment and local fiscal conditions, bargaining legal frameworks, and teacher unions’ organizations, for each province in the 2006–19 period. The classic labor literature typically distinguishes two measures of strikes: frequency (or the number of strikes) and severity. 63 Scholars in this tradition have documented that measures capturing severity, such as magnitude (i.e., number of strikers) or duration (i.e., working days lost per worker), provide a better indicator to assess patterns of contestation and the political effects of strikes than their simple occurrence. This is so for at least two reasons: first, because longer and massive strikes are naturally harder to sustain than one-day, small strikes; and second, severity better captures the role played by labor in general (as only strong organizations can sustain massive and long strikes) and public unions in particular.
Based on information collected by Argentina's Ministry of Labor and Social Security, we consider two dependent variables that measure these different but complementary dimensions of strike activity. We assess strike frequency by means of the annual number of primary, secondary, and tertiary/vocational (nonuniversity) provincial teachers’ strikes. We account for strike severity by measuring the total number of individual teacher's working days lost per year. The appendix provides details on the coding of all variables used in the statistical analysis.
For the operationalization of our key explanatory variables, we constructed an original data set of 331 yearly teachers’ wage bargaining rounds in all Argentine provinces between 2006 and 2019. 64 We test the “reverse business cycle” (hypothesis 1) by including in the statistical models three independent variables. First, the state of the local economy is measured by means of the annual provincial unemployment rate. Second, we assess the impact of local fiscal conditions in two ways. We created a provincial fiscal balance variable measured as total provincial revenues minus total expenditures, relative to total expenditures. We also include a variable measuring the annual per capita federal transfers received by a province as a proxy of additional provincial fiscal resources. Since the bulk of intergovernmental transfers to the provinces in Argentina are not earmarked, they automatically increase provincial budgets and provide governors ample discretion to use federal money to target social and material needs. 65 As stated by our theory, and contrary to the mostly private sector–based business cycle literature on strike patterns, we anticipate more labor conflict during bad economic times when unemployment grows and fiscal conditions worsen. The reason lies in the public sector nature of teachers and their job stability. Thus, we expect the unemployment variable to be statistically significant and positive, and provincial fiscal balance and federal transfer variables to be significant and negatively signed. Finally, all estimation models incorporate as a control variable the percentage annual change in teachers’ real earnings in the previous year, with the expectation that negative changes will induce labor militancy.
We coded for the first time in Argentina the mechanisms of wage negotiation and the legal existence of teachers’ bargaining rights at the subnational level. We test the effect of legal duty-to-bargain rights (hypothesis 2) by including two dummy variables in the regression models. The first dummy variable indicates whether or not a given province in a given year employed a legal framework that made bargaining mandatory for the provincial government and regulated who could negotiate. According to our theory, formal bargaining rights diminish strike action, so we expect this variable to be negatively related to teachers’ militancy. The second legal dummy variable incorporated into the regression models refers to the existence of a wage bargaining law at the national level, where 1 denotes a year in which the federal government and teachers’ confederations formally discussed a minimum income for education workers (2008–16), and 0 otherwise (2006–7 and 2017–19). In our multilevel analysis, national bargaining rights should be of little use for explaining variation in teachers’ strike action across provinces. Only national unions negotiated salaries in the NWB. The national teacher unions and confederations, we argue, are unable to control plural and competitive provincial union structures, and the federal minimum wage approved in the NWB is only taken by provincial unions as a baseline to bargain salaries in their own districts. Thus, we expect this variable not to be statistically indistinguishable from zero.
To assess the impact of political exchange (hypothesis 3) we first focus on union leadership monopoly, which has two dimensions that independently affect incentives for contestation: organized labor fragmentation, and internal union competition. Recall that, in contrast to union monopoly, fragmentation increases coordination problems and interunion competition, and thus hampers consensus. In order to account for this factor, we created a continuous variable that identifies all relevant teacher unions that formally or informally participated in a collective negotiation (some of them eventually signed an agreement) with the provincial governments. The variable ranges from a minimum of 1 teacher union to a maximum of 17, with a mean of 4 and a standard deviation of 3.5.
The second dimension of labor leadership monopoly is internal union competition. While in Europe organized labor is often split among Communist, Socialist, and Catholic factions, in Latin America the main division tends to be between populist and left parties and identities. Indeed, in Argentina the main opposition to dominant (generally Peronist) teacher union leadership comes from different varieties of the Trotskyite left. Consequently, we measure internal union competition by means of a dummy variable coded 1 if the radical left has a strong organizational presence in a given province during a given year (either within or outside CTERA base unions), and 0 otherwise. In our data set, nearly 63 percent of the time the left is coded as strong and in two-thirds of these cases it is powerful within CTERA. We expect that teachers’ militancy will increase when the left competes with Peronist leaders for the allegiance of workers, regardless of whether they belong or not to CTERA.
To complete the analysis of the political exchange framework, we also evaluate the potential effect of party identities, as conveyed by prior research on Argentina on the propensity of provincial teacher unions to strike. 66 We created a dummy variable that measures political alignment between provincial governors and teacher leaders, coded 1 when the governor and the leadership of the majoritarian teacher union in the province share the same party identity (54 percent of our observations), and 0 otherwise. The standard neo-corporatist approach holds that common party-union identities reduce uncertainty for labor and diminish union militancy, so we expect this variable to be statistically significant and with a negative sign.
Results: The General Logic of Subnational Public Sector Union Conflict
Our dependent variables—both the number of individual teacher's class days lost, and the number of teachers’ strikes—are nonnegative integers or counts. In addition, both are not normally distributed, but truncated at zero and their distribution is skewed to the right. Therefore, OLS and the linear model are not appropriate specifications. The first option would be to estimate a Poisson count model, which assumes that the conditional mean of the outcome variable equals the conditional variance. However, a likelihood-ratio test indicates that our dependent variables have signs of overdispersion, that is, the variance largely exceeds the mean. Under this condition, the truncated Poisson estimator continues to be consistent but it is inefficient, thus obtaining biased standard errors. Following the canonical statistical approach, we thus run a negative binomial model, which accounts for overdispersion in the data. 67 We run models with robust standard errors. Because observations are not truly independent, we also corrected standard errors for the clustering nature of our data belonging to each province. Naturally, there is a problem of autocorrelation within groups as some explanatory variables do not change in many provinces over the period (i.e., political alignment) or they do so very sporadically (i.e., strong left). Results with cluster standard errors are very similar, particularly in models with class days lost as the dependent variable, which, as we have stressed above, better captures labor contestation. All models include dummy variables for year (not reported) to account for unobserved year-specific factors that may affect teachers’ militancy.
Table 3 shows the results of our negative binomial estimations for both dependent variables controlling for robust (models 1 and 2) and cluster (models 3 and 4) standard errors. The results are consistent with most of our expectations. Contrary to private sector–based, business cycle approaches to the study of strikes, the provincial unemployment variable is often positive and strongly significant. This indicates that teacher union conflict increases when the economy is doing bad. Given that models include a control for real wage change, which critically never reaches statistical significance, this finding points to the advantages (such as job stability) that public workers have over private ones to strike in harsh economic times, in which the latter may feel more threatened. Also, as predicted by hypothesis 1, both the provincial fiscal balance and the federal transfers variables are negatively signed and statistically significant. These findings indicate that, unlike in the private sector, teachers’ protests tend to increase with fiscal imbalances, precisely when there are less resources to distribute. Likewise, strike activity, particularly its severity (model 2 and model 4), declines when governors have access and discretion over the use of sizable federal fiscal transfers free of local taxation so as to placate social discontent.
Teachers’ Militancy in the Argentine Provinces, 2006-2019. Negative Binomial Model
**** p<0.001; *** p<0.01; ** p<0.05; * p<0.1
Moving to the institutional determinants of public labor conflict, results substantiate that national collective bargaining rights (and the allocation of salary compensation funds to the less developed jurisdictions) do not reduce teachers’ militancy in the provinces, which as we expected respond entirely to subnational dynamics. However, against our initial expectations, institutionalized provincial “duty to bargain” rights do not seem to decrease (at least by themselves) teachers’ incentives to mobilize (hypothesis 2). In all models, both the dummy and categorical versions of this variable never reach statistical significance.
To account for a potential conditional effect of provincial bargaining rights, we run the same models with the addition of an interaction term between this variable and the predictor of union fragmentation (models 5 to 8 in Table 3). No matter how we estimate these models, there is evidence that teachers’ strike activity varies (indeed decreases) with the degree of union fragmentation. Note that the estimate for provincial bargaining rights is often negative and the interactive term is often positive, both reaching statistical significance at reasonable levels. This result documents that the negative impact of institutionalized bargaining diminishes (or strike activity is more likely) as union fragmentation increases. In substantive terms, the deterrent effect of provincial bargaining rights over teachers’ propensity to strike is relatively moderate as it is only observed in bargaining rounds with one or two teacher unions (concretely, 1.7 union or less). Thus, wage-setting institutionalization matters to contain teachers’ conflict in the Argentine provinces to the extent that the structure of provincial unions is essentially unified. When many unions exist and frequently outflank each other, the institutionalization of formal bargaining rights became less relevant in restraining militancy. We did the same exercise (not reported) interacting bargaining rights with union internal competition (in our case, presence of a strong left in teacher provincial unions). The interaction is never statistically significant, which appears to indicate that the role of subnational collective bargaining institutions seems to be weakened by inter-union competition rather than by intra-union factionalism.
As regards the political exchange framework (hypothesis 3), the estimations indicate that the three variables included in the statistical analysis are significantly related, in the hypothesized direction, to teachers’ strike militancy. In effect, political alignment between union leadership and incumbent governors reduces the likelihood of conflict, while union fragmentation and union internal competition increase it. 68
Substantively, the estimates in Table 3 are not immediately interpretable because negative binomial coefficients are expressed in logs that need to be exponentiated. A more intuitive interpretation indicates that, all other things being equal, a one-unit increase in union fragmentation would increase the number of class days lost by a considerable 65 percent, and the number of teachers’ strikes by 12.5 percent. Under high (left) internal competition, the number of class days lost would notably increase by 147 percent, and the number of strikes by 33 percent, relative to a jurisdiction where internal competition (left presence) is low. In a political exchange framework, the substantive impact of union organizational variables appears to be somehow more prominent than the influence of political alignment: political copartisanship between the governor and leaders of the majoritarian teacher unions in the provinces is expected to decrease by 46.5 percent the number of days lost and by 25.5 percent the number of strikes.
The substantive impact of economic variables also deserves some attention. In a nutshell, the effect of unemployment (arguably a measure of how the real economy is working) is moderate in comparison with that of fiscal conditions: whereas a one-unit increase in unemployment is expected to increase the number of strikes and days lost by 8.4 percent and 17 percent, respectively, an additional point of fiscal deficit almost duplicates the frequency and intensity of teachers’ militancy. These findings are in line with our theoretical statement. As they cannot be easily fired, unemployment is not a key matter of concern for public workers (at least in comparison with private ones), and fiscal constraints generate incentives to mobilize for obtaining scarce resources and oppose state adjustments.
Alternatively, to assess the magnitude of the effects produced by our political variables of interest, we can display the incidence rate ratio (i.e., the exponentiated coefficients) or the discrete change in the expected value of the number of class days lost and the number of strikes for a unit of change in the regressor computed. Our results indicate that, all other things equal, a one-unit increase in union fragmentation would augment the number of class days lost and the number of strikes in a given province by a factor of 1.65 and 1.12, respectively, and these effects are compounded. So a three-unit change (about 1 standard deviation) would increase it by a factor of 4.95 and 3.36, respectively. On the other hand, provinces where the left enjoys representativeness among education workers have incidence rate ratios of days lost and strikes that are 2.48 and 1.33 times those in provinces with no radicalized unions, holding the other variables constant. Finally, in terms of magnitude, provinces where teacher unions are politically aligned with the governor would decrease days lost and strikes per year by a factor of 0.53 and 0.74, respectively.
Interpreting Qualitative and Quantitative Results: National and Subnational Levels
Our mixed-methods strategy—comparative-national and statistical-subnational—for the study of teacher unions’ militancy in a democratic and robust federal polity like Argentina helped us to understand labor dynamics and patterns of contentious activity at each level of government, and the interaction (or lack of it) between them.
The national picture closely represents the sequence argument in our explanatory model depicted in Figure 4 above: in an initial context of fiscal abundance, the progressive Peronist goverment passed a duty-to-bargain clause that consolidated its alliance with the hegemonic and Peronist left teachers’ confederation (CTERA). This alliance was cemented by real wage increments concerted in the NWB, and the allocation of labor and organizational benefits to teachers’ confederations. In exchange, nationally organized strikes substantially lessened. Under the Macri administration, once national bargaining rights were revoked, in a context of opposing government-labor partisan identities, the political exchange collapsed and national conflicts resumed virulently.
Yet, the duty-to-bargain rights for a minimum wage institutionalized in the federal NWB, with the consequent real wage growth for teachers it helped to spur nationwide, and the alliance between national teacher confederations and the central government in 2003–15 did nothing to restrain provincial-level teacher mobilization. Indeed, teacher militancy in the provinces responded entirely to local factors identified by our theory: the “reverse business cycle” (i.e., low unemployment and sound state finances), the formal existence of “duty to bargain” rights in scenarios of low union fragmentation, 69 and the general conditions for the political exchange (namely, labor unity and common partisan identities between local unions and provincial governors). Thus, up to this point, our analysis leaves the causal mechanisms operating at the subnational level (and some paradoxes, such as the fact that higher wages for Argentine teachers, especially in the 2006–15 period, coexisted with high levels of labor contestation in many provinces) somehow unexplored. Given Argentina's mosaic of subnational labor institutions and union organizational structures, additional qualitative evidence about teacher politics in the provinces may serve this purpose.
For example, Neuquén and Mendoza, two bordering provinces located in the southwest of the country, are roughly similar in a number of relevant contextual characteristics that could potentially affect teachers’ militancy. First, both are above the national average in socioeconomic (human development) indicators. Second, in both cases the state share of formal employment is relatively low, at least in comparison to that of other provinces. Key thriving private sectors (tourism and oil/gas in Neuquén, wine and food processing in Mendoza) provide alternative jobs to state-controlled patronage networks that are prevalent in other provincial jurisdictions. 70 Despite these similarities, Neuquén and Mendoza constitute almost polar opposites in terms of teachers’ strike mobilization: 179 and 13.5 workdays lost per teacher between 2006 and 2019, respectively, that is, the second and seventeenth most contentious in the country (see Table 2). 71
In Neuquén, ATEN (Education Workers’ Association of Neuquén) is, in practice, the single teacher union in the province. Yet the radical left exerts a powerful influence in the organization. Furthermore, provincial legislation did not bestow formal bargaining rights on teacher unions until 2012. In view of the electoral hegemony of the traditional provincial party MPN (Movimiento Popular Neuquino, which never lost a gubernatorial election until 2023), and the alternative control of ATEN by Peronist militants and the left, Neuquén never witnessed a government-union alliance. Two big strike waves demanding better wages were staged by ATEN in the period under study. The Peronist faction of the union led the 2007 strike. That year, tensions mounted when repressive and violent measures by local police to clear a road resulted in the death of a teacher activist, Carlos Fuentealba. ATEN reacted with big demonstrations joined by the entire labor movement and provincial civil society organizations. Moreover, CTERA called for a national strike that paralyzed public education in all of Argentina. 72 Significantly, since the strike began, the leadership of ATEN, progressive-Peronist and close to the national CTERA, was thoroughly challenged from below by a variety of radical left groups that controlled several sections within the union, including the largest in the provincial capital with more than 40 percent of all teachers in the province. 73 The second major wave of strikes occurred in 2013, when the Trotskyite left won control of the union. Both strike waves, in polarized contexts in which the governor could (and often refused to) negotiate wages, lasted more than forty days. They were combined with several other contentious strategies such as road blockades and street mobilizations. Interestingly, the provincial government's attempt to institutionalize wage negotiations in 2012 turned out to be futile in restraining conflict in view of the continuous strength of varieties of radicalized left in the union. 74 In short, Neuquén represents an ideal case of how high internal union competition and the strength of the militant left, 75 the absence of formal bargaining rights, and the lack of government party-labor common political identity all induced high teacher militancy.
Neuquén's northern neighbor, Mendoza, by contrast, fits well the causal path to low public sector union conflict. SUTE (Single Union of Education Workers, the CTERA provincial union) holds the informal monopoly of teachers’ representation and is arguably the most important union in the province. The radical left, though present, was relatively weak during most of the study period. The Celeste, the Peronist faction, commanded SUTE for twelve of the fourteen years analyzed. SUTE formed a solid alliance with Mendoza's Peronist governments that ruled the province between 2008 and 2015. 76 This corporatist convergence was cemented by important payoffs: SUTE obtained participation in the governmental housing program for popular sectors, received more compensation for teacher's seniority, were conferred the permanent hiring of replacement teachers, and got subsidies for teachers’ working elements. 77 In the 2012 gubernatorial elections, SUTE publicly supported the Peronist candidate Celso Jaque for governor, and his eventual choice for minister of education. These coalitional dynamics operated in a legal context where bargaining rights covering all public workers had been institutionalized by a provincial decree early in 2004, which provided the teacher union with a valuable road to channel its demands. Wage-setting institutionalization indeed reinforced union monopoly in Mendoza. The decree privileged SUTE as the single teacher representative, and relegated the two other minority unions (UDA and SADOP) to have voice, but no vote, at the bargaining table.
Case study analysis also uncovers the mechanism behind this “paradox” of higher real wages for all Argentine teachers, and provincial-level conflict unfolding simultaneously especially in 2006–15. In some contentious cases, such as Neuquén, and others not reported here such as Santa Cruz, Tierra del Fuego, Jujuy, Catamarca, and Chaco, the CTERA Peronist leadership that negotiated the federal minimum wage with the national government simply does not control the hegemonic subnational teacher union. Local unions (whether affiliated with CTERA or not) in these cases were frequently in the hands of the radical left or independent combative groups. In some other larger districts like Buenos Aires, the City of Buenos Aires, Santa Fe, and Entre Ríos, CTERA national leaders do govern powerful, dominant provincial-level unions. However, those leaders are challenged by a competitive “internal” Trotskyite left, with its tradition of maximalist demands. For example, in Santa Fe the Trotskyist left has controlled for many years the most important union section in Rosario, by far the largest city, despite the Peronist dominance of the provincial union. Varieties of the radical left have also led key SUTEBA (Single Union of Education Workers of the Province of Buenos Aires, the provincial union affiliated with CTERA) sections of the province, such as the large districts of La Matanza and La Plata. In sum, in those provinces in which the CTERA leadership (despite being, in practice, a monopoly at the national level) is thoroughly outflanked on the ground, it was very difficult to replicate the national political exchange and restrain provincial militancy.
Theoretical Implications
Are public sector unions of a different nature? Paul Johnston, who has made perhaps the most systematic attempt to assess the differences between private and public trade unionism in the United States, argues that state unions are different in that they focus on “a single organization,” while private unions often organize across a variety of employers. He also states that public unions, more than private ones, posit their demands as “public interest issues.” 78 Richard Freeman, on the other hand, observes that public sector unions, more so than private labor, “can influence the employer behavior through the political process.” 79 The work of these scholars is important to understand the particular nature of public unions. Yet we ultimately disagree with these views. As the Argentine case just analyzed shows, public unions/confederations also often face the dilemmas of facing multiple employers, both horizontally in the same subsector (e.g., teachers and nurses) or vertically (across different levels of government). We also know that private sector unions, especially when they are relatively powerful, may also tend to frame their demands in terms of the public interest. Finally, unlike Freeman's view (probably centered too much on the United States), there are plenty of examples of private unions that strongly influence employers through the political process, particularly in the coordinated economies of Europe.
In this study, we have argued that private and public labor share essential similarities as part of the subaltern classes. However, we contend that public sector unions are fundamentally different from their private counterparts in two key dimensions: (a) the role played by government regulations that induce (or not) employers to bargain and (b) the labor market conditions for collective action. These general differences, we contend, configure alternative foundations for public sector labor mobilization. Our multilevel empirical analysis of conflict in the largest state union of Argentina brings some insights for the general discussion on public sector union politics.
First, as conveyed by our “reverse business cycle” hypothesis, the logic of public sector union conflict challenges the prevalent economic theories of strike activity, all of them based on the private sector. 80 This approach essentially explains militancy as a result of economic growth, tight labor markets, and low unemployment. Protest is more likely when employers find it difficult to replace rebellious workers, and labor activists know that alternative jobs are potentially available. However, our comparative national analysis suggests, and our estimations statistically indicate, that in the case of state unions the opposite may be in fact true. Essentially because of public sector job stability and state bureaucrats’ difficulties to outsource workers, state employees’ propensity to strike increases in bad economic times, in contexts of rising unemployment and fiscal adjustment.
Second, while the literature on public sector unions has emphasized alternative dimensions for the analysis of contestation—government regulations in the United States, political exchanges in the European and Global South corporatist literatures—the Argentine case shows how both aspects can combine to shape contentious collective action. Duty-to-bargain laws that legalize unions and institutionalize collective bargaining and the conditions for the political exchange (copartisanship and labor unity) are decisive for shaping labor conflict in the country's public sector at both the national and provincial levels. Our study suggests that subfields and regional divides hamper a theoretical understanding of public sector unions’ politics and mobilization. For example, in some contexts, public unions might not simply constitute a “special interest” in search of short-run economic rents, as the neo-utilitarian approach in American politics suggests. They may also operate like political actors, exchanging some mobilization restraint for organizational and political payoffs, better labor conditions, and stable patterns of income, as was the case with national teacher unions under the Kirchner governments, in the province of Mendoza just analyzed, and in some European and Global South party-labor alliances. Likewise, institutional variables, particularly union fragmentation, might not be important in the United States or Mexico, where public labor generally enjoys exclusive right of representation. But in countries where union pluralism is the rule in the public sector, such as Spain, France, and Argentina, competition among fragmented unions for the representation of (the same) workers becomes more relevant as a factor that may hinder the political exchange.
Third, our multilevel analysis points to the complexities of corporatist arrangements for government unions in federal countries. In the classic neo-corporatist literature, largely based on the private sector, national or sectoral pacts that galvanize the political exchange of social peace for labor benefits are assumed to hold nationally. This is so even in federal countries like Germany. 81 In Argentina, the convergence of the national teachers’ confederation and the Kirchnerista governments had all the ingredients of a neo-corporatist pact, such as wage-setting centralization and labor organizational payoffs (i.e., joint public policy formulation and administration of skill formation programs). But as solid as this alliance was to temper teachers’ nationwide strikes, it was unable to bring about social peace in the provinces. It is frequently the case in federal countries that decisions over public budgets and wages are taken both at the national and state/provincial levels. National political exchanges may be simply difficult to sustain in the lower levels of government because key conditions that foster those neo-corporatist arrangements at the center—in this article duty-to-bargain laws, common party identities, and union monopoly—may not be present at the subnational stage.
Finally, the case of Argentina may also be illustrative about the fate of state corporatism in the Global South, once top-down union control prevalent during most of the twentieth century mutates into more democratic party-union alliances. We documented that party-union common identities at both the national and subnational levels reduce public sector labor militancy. But are these alliances in the end good for workers, beyond (frequent) organizational and inside deals with unions? Indeed, in cases such as Mexico and South Africa some scholars have suggested that the subordination of public sector unions to partisan interests may in the end demobilize unions and ultimately harm workers. 82
Our analysis of Argentina, however, reveals that, simply put, state corporatism and control from above in the public sector (as well as in the private domain) is a question of the past. Teacher unions are a rebellious force. First, teacher unions at the grassroots, provincial level (from both Peronist and Trotskyite bents), and CTERA, were very far from being co-opted by the Kirchner governments. In some provinces where collective bargaining was legally established and CTERA unions were dominant, Peronist governors nonetheless faced difficulties in preventing teachers’ strikes. Second, it is undeniable that, as mentioned above, teachers improved their real wages and working conditions, at least until 2012. Teacher unions’ combination of mobilization and political compromise paid off. A similar phenomenon can be observed in South Africa. Scholars have shown how public sector workers and teachers, who formed a solid party-labor alliance with the ANC, were clearly benefited, especially in the initial years after the transition to racial democracy. 83 In this case, as well as in Brazil and Argentina, public sector unions and teachers were permeated by social movement unionism and strong grassroots organization in their formative years. Given that original influence, it is unlikely (though not impossible) that public sector unions and teachers’ militancy will be entirely controlled from above despite the “necessary” deals when their sister parties are in government.
Conclusion
Beyond state retrenchment in the public industrial sector, countries still need civil servants, firefighters, trash collectors, teachers, nurses—in many cases increasingly more so. Indeed, public sector unions are becoming the hegemonic labor actor in terms of mobilization in many advanced and emerging economies. However, contemporary political economy lacks a unified toolkit to analyze public sector labor. Studies of state labor are separated by regional and subdisciplinary boundaries. Of course, mutual appreciation has been hampered by the divergent development of public sector unionism in the United States, Europe, and the Global South that we noted above. However, our general review of the subfields, and the theoretical approach we presented, suggest that these perspectives would gain from a conversation. Indeed, this article intends to lay out the initial basis for a more general (and less regionally anchored) view on public trade unionism, which underscores its key differences with private labor.
For example, scholars abundantly show how US public sector unions engage in electoral mobilization to help Democratic Party candidates. 84 However, curiously enough, government-union political exchanges under Democratic administrations (as opposed to Republican ones) are seldom analyzed systematically. By the same token, as democracy develops in the Global South, and state corporatism melts, it is worth studying possible neo-corporatist deals between more autonomous public unions and popular parties in places like Brazil, Mexico, Tunisia, or South Africa, similar to those that took place under Kirchnerismo in Argentina.
In the footsteps of classical works on labor mobilization by Shorter and Tilly and Franzosi, 85 we intended to go further from regression analysis because we wanted to capture the social and political dynamics behind the surge of teacher strikes. We are conscious that the national and subnational qualitative assessments in this article were essentially confirmatory of our general explanatory model. As Rohlfing asserts, such a strategy is not especially suited to find omitted variables or deal with the problem of possibly “underfitted” regression analysis. 86 Future research could analyze provinces or national instances in which the variables we flesh out do not seem to work, and thus search for alternative explanations. Nonetheless, our case analysis at both the national and provincial level fulfilled the essential logic of qualitative case studies in mixed-methods research: the illumination of causal mechanisms and the sequential process behind the onset of labor militancy.
Finally, though financial and regulatory factors surely play an important role, this study underscores the potential of political exchange factors for explaining public sector mobilization. In this perspective, the locus of labor demands and negotiations shift from the economy to politics. Social peace is constructed through labor market organization. This move has two essential prerequisites: union leadership hegemony (here measured in the two dimensions of number of unions and internal competition) and common party-labor political identity. In Argentina, at the national level, labor informal monopoly, and a common Peronist-left identity, cemented the political exchange between the Kirchner governments and CTERA. The result was very low nationally organized contention and tangible labor benefits. At the subnational level, however, these labor organizational variables presented a lot of variation, and thus produced very different patterns of public sector union mobilization.
Footnotes
Appendix
Acknowledgments
This article was presented at the Robert F. Kennedy Professorship Lecture, Harvard University; at the Comparative Colloquium of the Department of Political Science, University of California Berkeley, at the Latin American Series Seminar of the Department of Political Science, Massachusetts Institute of Technology, and at Seminars at the Universities of San Andrés and Torcuato Di Tella in Argentina. We thank very useful comments from participants in these events. We also want to thank very helpful comments to different versions of this paper from Martín Armelino, Christopher Chambers-Ju, Candelaria Garay, Tomás Gianibelli, Gabriel Vommaro, Andrés Schipani, Ben Ross Schneider, and Kathryn Sikkink. Oriana Peretti, Vicente-Fernandez Mouján, and Carmela Braconi did excellent work on data collection for both the quantitative and qualitative analysis. We also thank Oriana and David Trajtemberg for comments, and for our many debates on unions and teacher politics that helped us refine the argument
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Correction (January 2024):
Article updated to correct the row of “Interactive Term” in Table 3 and the title of Figure 2.