Protesta popular y redes clientelares en la Argentina: el reparto federal del plan trabajar (1996-2001)*

German  Lodola

Department of Political Science,  University of Pittsburgh. [4L01 Wesley W. Posvar Hall / Pittsburgh, PA 15260 / USA / E-mail: <gjl16@pitt.edu>]

Translated by Marta Ines Merajver
Translation from Revista Desarrollo Económico, Buenos Aires, v.44, n.17, p.515-536, Jan./Mar. 2005.

SUMMARY

This paper presents a statistical analysis of the determinants that ruled the federal allocation of the Argentinean Plan Trabajar for the period 1996-2001. The overall results indicate that this emergency employment program was created as a mechanism to deal with increasing unemployment. However, the federal distribution of resources was later associated to the frequency of popular protest and the parallel evolution of the so-called “picket movement”. The political variables included in the model suggest that the Peronist Party succeeded in making a better political use of federal allocation of funds than did the Unión Cívica Radical/Alianza.

Descriptors: Political Science, Social Policies, Social Movements, Popular Protest, Clientelar Net, Plan Trabajar, Argentina.

RESUMEN

Este trabajo presenta un análisis estadístico de los determinantes distributivos del Plan Trabajar a las provincias argentinas en el período 1996-2001. En general, los resultados indican que la creación de este plan de emergencia ocupacional obedeció a la decisión de luchar contra el creciente desempleo. La distribución federal de recursos, no obstante, pasó luego a estar también asociada con la frecuencia de la protesta popular y el desarrollo paralelo experimentado por el denominado movimiento piquetero. Por su parte, las variables políticas incluidas en el modelo sugieren que el Partido Justicialista ejerció mayor influencia sobre la asignación federal de fondos que la Unión Cívica Radical/Alianza.

Descriptores: Ciencia política, Políticas sociales, Movimientos sociales, Protesta popular, Red clientelar, Plan Trabajar, Argentina.

Introduction

The case of Argentina in the  90s, however, illustrates a completely different pattern from that proposed by the neopopulist theory of compensations. Even when well-oiled clientelar practices characterized the political nets of the ruling party (Auyero, 2001; Levitsky, 2003), Carlos Menem’s Peronist administration failed to strengthen the political feasibility of the neoliberal project through focalized compensation². Indeed, it was only after the structural transformation process was well under way that the government launched an ambitious, nationwide employment program called the Plan Trabajar. This emergency employment program was aimed at allaying the costs of sweeping unemployment and controlling increasing social mobilization. Moreover, the implementation of this program was accompanied by an escalation (rather than by a decrease) of protests led by unemployed and informal workers gathered in the so-called Movimiento Piquetero (Picketer Movement). Thus, Argentina  witnessed a delayed compensatory intervention that neither quieted social unrest nor conditioned the organization of popular sectors.

This study presents a statistical analysis of the role played by popular protest and party politics in the distribution of resources from the Plan Trabajar among Argentine provinces during the period 1996-2001. The overarching results indicate that the plan was designed as a way to fight unemployment. The federal allocation of funds is also related to the frequency of popular protests and the concomitant development of picketer organizations. In addition, the political variables included in the model suggest that the Partido Justicialista [Justicialist Party] (PJ) exercised a greater influence on the federal allocation of resources than did the Unión Cívica Radical / Alianza.

Unfortunately, theorits of social movements in Latin America have tended to focus on the rise and evolution of movements throughout time while paying little attention to their impact on the policy-making process³. This paper aims to identify the causal effect of popular protest in the region. Likewise, by analyzing the role played by executives, governors, and political parties in the federal distribution of public resources, this study attempts to contribute to the growing debate about the functioning of clientilistic politics in the context of Argentinean federalism⁴.

This paper has been organized as follows. The first section presents descriptive evidence of the evolution of social expenditure and employment policies in the 90s to show that the Menem administration lacked compensatory policies. The second section develops the hypotheses of analysis to test the main determinants of the distribution of the Plan Trabajar. The third section, explains the statistical models and discusses the results. The last section concludes. 

1.Focalized social expenditure and employment policies in the 90s

The profound process of productive restructuring and redefinition of the labor system that accompanied programs of economic liberalization entailed high costs for the working classes⁵. The reform of rigid labor institutions in force since postwar times involved the flexibilization of labor relations while drastic changes in the development model particularly affected new workers, informal workers and the unemployed. The case of Argentina was not an exception. The labor movement joined the governing coalition at an early stage, a fact which allowed trade union leaders to delay and/or to curb the scope of the labor reform when it compromised the interests of trade unions as organizations, therefore affecting the interests of unionized workers as well (Etchemendy, 2001; Murillo, 1997).

Latin American governments, however, did not respond in the same way to the social impact brought about by economic adjustment. Some understood that structural reform was bound to exclude a significant sector of the society, and so devised safety nets in the form of focalized social assistance⁶. Others sought to counteract the distributive costs of economic transition by actively generating employment policies⁷. Overestimating the market’s capacity to adapt to the new scenario, the Menem administration did not contrive actions to temper the negative effects of changes or redirect the labor market in accordance with the process of economic reconversion. In effect, as we will see below, the Menem administration did not strengthen social expenditures specifically targeted the poor, nor did it allocate large resources to protect or to generate employment. Rather, the government encouraged workers to join plans of voluntary retirement. While it can be argued that this kind of “market compensations” had been resorted to by most European countries when they adjusted their economies in the 80s (Kohli et al., 1991), the Argentinean government implemented such measures without providing incentives for the retiree to find a new job or promoting access of younger, more qualified workers to replace old ones. 

Social expenditure and focalized policies

It is worth noting that Argentina’s social expenditure included in the “social service” category is more addressed to the poor than that included in the “social security” category, which requires recipients to be employed or to have been employed in the formal economy⁹. On average, the social services expenditure, which amounted 7.4% of the GDP in 1984-1989, increased to 8.6% in 1990-1999, while expenditure on social security grew from 8.8% to 11.6%¹⁰. However, the increment in social services was not applied to build focalized programs, whose growth barely averaged 0,2%. In addition, about 85% of resources of social services were allocated to education and public health, universal categories that do not benefit the poor exclusively, but generate positive externalities for middle sectors as well. Finally, three quarters of the increase observed in social security programs were transferred to the retirement and spouse pension systems¹¹.

A second indicator of the government’s weak commitment towards social focalization resides in the elasticity of social expenditures¹². Figure 2 (see following page) plots annual changes in the logarithms of public and social expenditures by category for the 1990-1999 period. It can be appreciated that social expenditure invariably decreases when there is a drop in public expenditure, and that it falls to lower, even negative levels, when public expenditure increases. Likewise, social services expenditure was reduced regardless of whether the country experienced periods of fiscal contraction (1995, 1996, and 1998) or expansion (1991, 1992, and 1994). It was only in 1990 and 1993 that expenditure on social programs directly related to the poor behaved in an inelastic fashion, as predicted by the neopopulist theory of compensations¹³.

In short, the information presented indicates that focalized compensation occupied a residual place in president Menem’s agenda (see also Cortes and Marshall, 1999; Repetto, 2000). In sharp cpntrast to neopopulist theorists, therefore, social focalization was not a necessary condition for the advancement of neoliberalism in Argentina¹⁴. We will now see that the employment system also lacked compensatory features quite until the Plan Trabajar was launched.  

Employment Policies

With the exception of the Plan Trabajar, employment policies were limited in time, depended on scanty resources and suffered endless modifications in their design and implementation. In 1991, Congress passed the National Law of Employment, which created an unemployment insurance funded by labor taxes¹⁵. However, strict legal provisions conditioned the scope of this policy and favored its redirection to formally employed middle-class workers. Firstly, only salary-earning workers registered in the social security system with a minimum of twelve contributions a year over the three years prior to redundancy were entitled to unemployment insurance. Secondly, temporary and informal workers, rural workers, hired domestic help, building industry workers, and public workers subjected to the state’s rationalization process were excluded from the benefit.

Indeed, no Latin American country has developed a universal unemployment insurance system. Nonetheless, it should be noted that the Menem administration allocated fewer resources to this policy than did the governments of Brazil, Chile, and Uruguay (Marquez, 2000). Resources devoted to this area barely reached 0.14% of the GDP, to cover only between 7% and 8% of the workforce (Government of Argentina, 2000)¹⁶. The rate of unemployment insurance lay well below the ones in other reformist countries, and did not go up as unemployment mounted. Etchemendy (2004), for example, reports that 37.8% and  51.6% of the unemployee were benefited by the insurance system in Chile and Spain respectively. In Argentina, the number of unemployed workers tripled between 1990 and 1996, but unemployment insurance rose only by 22%.

The national government also allocated modest resources to the active generation of employment policies. Although between 1993 and 1996 the Ministry of Labor (MTSS) launched a set of programs to promote temporary public employment, labor training, and subsidies for the private sector, this battery of policies proved shallow and lacked compensatory direction. The most important measure in the period was the Programa Intensivo de Trabajo [Intensive Labor Program] (PIT), which employed low-income workers to develop communal projects. Yet in the year when unemployment reached its peak, PIT provided cover for only 2.2% of the unemployed (Marshall, 1997). On the other hand, labor training programs engaged a monthly average of only 2,000 to 12,000 workers, while subsidy programs for the private sector averaged just 25,000 beneficiaries per month (MTSS 1997, 1998). Taken together, these programs were not really significant, since they amounted 0.1% of the GDP to cover 9% of the urban unemployed (Government of Argentina, 2000). Proportionally, and in comparison with other countries in the region, the Argentinean government allocated fewer resources to the creation of genuine employment that did its counterparts in Mexico, Brazil, Uruguay, Peru, and Costa Rica (Marquez, 2000).

In 1996, immediately after violent riots unleashed in some cities of the interior, the government launched the Plan Trabajar¹⁷. Devised by the MTSS together with World Bank (WB) economists (an agency that co-funded the program since 1997), the Plan was in operation until 2001, spending an average of 130 million pesos a year to cover about 20% of the unemployed. The Plan Trabajar engaged unemployed workers who had no social security coverage and paid a monthly salary of 200 pesos over a six-month period (with limited possibilities of renewing their contract) in exchange for communal tasks. The national government took upon itself the hiring costs of unskilled workers, while the municipalities or NGOs in charge of executing infrastructure projects supplied the materials and hired skilled workers.

The distribution process consisted of three stages. First, the Executive transferred the resources to the provinces. Then, Governors decided on the municipalities to which the funds would be allocated. Finally, Mayors were responsible for the selection of beneficiaries and decided on the execution guidelines. Within this institutional framework, unemployed organizations (particularly those with a long tradition) established direct relations with the Town Halls, made their own lists of beneficiaries through a previous registration of unemployed workers, and adopted roadblocks as a strategy to claim for their inclusion in the Plan Trabajar. This resulted in an increase of the volume of organizational resources and fostered new memberships, including those social actors that had links with the PJ punteros’^** net¹⁸.

When Fernando de la Rúa took office, the Alianza administration changed rules of distribution. Paradoxically, these changes strengthened the larger picketer organizations and contributed to the growth of smaller groups. With the purpose of constraining the clientelar control of funds made by the PJ, it was decided that the plans were to be managed by authorized NGOs through projects submitted to and approved by the MTSS. The picketers reacted by establishing their own NGOs, thus formalizing their access to the distribution chain. Towards the end of 2000, the government created a system of local institutions caleed Consejos de Emergencia [Emergency Councils], with the purpose of monitoring the distribution and preventing further social protest¹⁹. Far from fulfilling their objectives, the Councils returned power to Peronist municipalities so that these local authorities could handle the allocation of employment plans, and institutionalized the presence of picketers in the arenas of negotiations.

Numerous assessments on the performance of Plan Trabajar acknowledge the importance of clientelar practices in the distribution (World Bank, 2000; Jalan and Ravallion, 1998; Ravallion, 2002; Ronconi, 2002; SIEMPRO, 1997, 1998). Likewise, essays on the picketer movement point out that the plan was a condition of possibility for its development²⁰. However, none of these studies has undertaken a systematic analysis of the causal effect of party politics and popular protest on the program’s distributive patterns. The following section, then, presents a series of hypotheses to empirically test the distributive determinants of Plan Trabajar.

2. Hypotheses of analysis

Which were the determinats of the intergovernmental allocation of resources from the Plan Trabajar? I present a set of hypotheses that explore the role played by three factors: material needs, social conflict, and party politics.

A first argument suggests that the programs’s transfer system followed a public policy logic compensating provinces according to some social equity criteria. This notion found wide support among World Bank (WB) officials and it is implicit in the neopopulist theory of compensations. For WB officials , the central government is a kind of social planner in charge of redistributing public resources to favor regions with the lowest  social and demographic indicators. For the neopopulist view, distribution does not depend on normative principles but rather it is guided by political needs, namely, the need to earn acquiescence from excluded sectors. Along these lines of thought, we may expect the following:   

Hypothesis 1: The greater the “material need” in province i, the greater the probability that this province may receive a proportionally higher number of resources from the Plan Trabajar.

A second argument stems from the classic works by Piven and Cloward (1977, 1992, 1993) about the effects of popular protests on the expansion of the welfare state in the US between 1930 and 1970. On the basis of a series of case studies, these authors put forward a compelling thesis: disruption rather than organization is the most powerful resourse for poor organizations to achieve their goals. Ever since then, specialists in urban riots and social movements in developed countries have analyzed whether disruption or moderation is functional to the success of social protest. Empirical evidence is not conclusive: while some scholars contend the efficacy of disruption and violence (Jennings, 1979; McAdam, 1999; Gamson, 1990; Tarrow, 1994), others argue that disruption is effective only under particular circumstances (Schumaker, 1978; Koopmans, 1993).

It is important to note that, unlike countries such as Bolivia and Venezuela, where neoliberal policies gave rise to generalized popular protests, Argentina implemented its market-oriented  reforms in an atmosphere of relative social peace. Indeed, on the one hand, acceptance by the union movement drastically reduced labor conflicts and confine trade-union activism to public sector workers affected by the rationalization and decentralization of state services. On the other hand, social protest was initially sporadic and restricted to a few peripheral cities especially affected by economic reconversion processes. As increasing unemployment began to erode the relationship between the unemployed and trade unions, social conflict grew steadily and roadblocks (the modality of accion used by picketer organizations) replaced strikes as the crucial manifestation of social discontent. Figure 3 (see following page) shows this change.

If protests led by the picketer movement affected the federal distribution of the Plan Trabajar, then those provinces that had been more exposed to social conflicts should have received a proportionally higher share of funds than those suffering less. Consequently, the second hypothesis states:

Hypothesis 2: The higher the frequency (or number) of protests in province i, the higher the probability that this province will receive a proportionally higher number of resources from the Plan Trabajar. 

Finally, a third argument focuses on the role played by party politics. Here the assumption holds that the government will favor provinces where the president’s party is dominant, measured in terms of political control of the municipalities and governor’s partisanship. Conversely, it will penalize provinces that do not comply with these requirements. Thus, the third and fourth hypotheses state:  

Hypothesis 3: The larger the percentage of municipalities controlled by the President’s party (PJ or UCR/Alianza) in province i, the higher the probability that this province will receive a proportionally higher number of resources from the Plan Trabajar. Conversely, the larger the percentage of municipalities controlled by the opposition (PJ or UCR/Alianza), the smaller that probability.

Hypothesis 4: Provinces governed by the President’s party (PJ or UCR/Alianza) receive a proportionally higher number of resources from the  Plan Trabajar. Conversely, provinces ruled by the opposition (PJ or UCR/Alianza) receive a proportionally smaller number of resources.

Ever since the return to democracy in 1983, the PJ enjoyed a considerable electoral advantage over other political parties. Indeed, the PJ controlled the Senate (even when it lost the presidential election) and the two legislative chambers when it came into office (1989-1999). The UCR, in contrast, never succeeded in controlling the Senate and only held the majority in the Lower Chamber between 1983-1987 and 1999-2001. During the period under study, the PJ governed  14 (1995-1999) and 15 (1999-2001) provinces and about 40% and 50% of the municipalities. Peronist electoral supremacy, added to the PJ’s decentralized organization, results in forceful structures of provincial power whose cooperation is vital to govern effectively (Levitsky, 2003). We can then expect that provinces in which the PJ is the dominant political force would receive a preferential treatment by the national government, even when the Presidency is not in Peronist hands. In other words, the federal distribution of employment plans depends on who are in power and in the opposition at the national level. Such reasoning leads us to the fifth hypothesis:

Hypothesis 5: Provinces politically dominated by the PJ receive a proportionally higher number of resources from the Plan Trabajar even when the President is not a member of the PJ.  

Finally, the last hypothesis refers to the provinces’ and governors’ ability to influence the federal distribution of resources through their national legislators. First, the President may grant material benefits to certain provinces in the hope that their representatives will support his/her proposals in Congress. Thus, when analyzing the role of Congress in the national budget, Jones (2001) demonstrates that the higher the proportional contribution made by a given province to the President’s legislative block, the larger the funds the province is likely to receive. Second, legislators may direct part of their legislative work to divert resources toward their provinces if their political careers depend on governors’ political strength.Argentine legislators are indeed vulnerable to the interests of provincial political leaders, since they control party nominations to national candidatures and the position of candidates on the party lists (Jones, Saiegh, Spiller and Tommasi, 2002; De Luca, Jones, and Tula, 2002).. Therefore, the sixth hypothesis tests the argument about the exchange of votes for favors in the legislative arena, and examines the “indirect” influence of governors on the intergovernmental allocation of employment plans.

Hypothesis 6: The greater the proportional contribution of province i to the President’s coalition in Congress, the greater the probability that this province will receive a proportionally higher number of resources from the Plan Trabajar.

3. Specifications of the model and analysis of results

In order to test the hypotheses presented in the previous section, I estimated two separate panel regression models for the Menem and de la Rúa administrations respectively. The choice for running separate models is based on solid theoretical reasons given that I postulate that the federal distribution of resources differs depending on who govers at the national level: the PJ or the UCR/Alianza²¹. All estimates are OLS with robust standard errors to correct for potential heteroskedasticity. The unit of analysis is the 24 Argentine provinces ²². The dependent variable is the yearly amount (in 1991 million pesos) from the Plan Trabajar allocated to each province²³. These data were supplied by the Secretariat of Labor Relations,MTSS.

The first independent variable is the percentage of unemployed poor people per province (estimates issued by the World Bank). The models were also estimated using the index of Basic Unsatisfied Needs (NBI) from the 1991 and 2003 censuses, and the provincial unemployment rate recorded by INDEC. The second independent variable –popular protests –is operationalized as the frequency (number) of roadblocks in province i over a particular year. This information was provided by the Centro de Estudios Unión para la Nueva Mayoría, which generate data from Gendarmería Nacional [National Gendarmerie] and nationwide newspapers²⁴. The third independent variable measures the power of the President’s party and opposition forces in province i as the percentage of municipalities governed by the PJ and the UCR/Alianza respectively.²⁵ These data are taken from  Cao (1999). Following Jones (2000), the fourth independent variable is the sum of the percentage of legislators that province i contributes to the President’s coalition in both the Senate and the Lower Chamber divided by two. The information used to build this variable was supplied by the Programa de Estudios Ejecutivos y Legislativos [Program of Executive and Legislative Studies] (PEEL-UTDT). The models also include a dummy variable coded 1 for Peronist governors in 1996-1999 and 1 for UCR/Alianza governors in 2000-20001. Moreover, models control for population through its logarithms and assume a contemporary effect of the independent variables²⁶. Tables 1 and 2 report descriptive data of the main variables of interest.

Tables 3 and 4 show regression results of models for Menem and de la Rúa administrations. Overall, results confirm most hypotheses, although some unexpected outcomes are observed. Concretely, distributive patterns are different under the two administrations. This bears out the notion that federal distribution of employment plans depends on who holds the Presidency. In addition, results indicate that the effect of unemployment keeps constant throughout the period, yet the causal impact of protests only occurs during the Alianza government and  party politics becomes ore apparent during the Menem administration.

Let us proceed to analyze these results. First, and in consonance with Hypothesis 1, the unemployment variable is positive and statistically significant under both administrations. This suggests that, other factors being equal, the distribution of the Plan Trabajar compensate provinces with larger percentages of unemployed poor people. For example, the 97.43 coefficient under the Menem administration (see Table 3, model 1) indicates that a 1% increase in the percentage of the unemployed poor in province i  is - ceteris paribus – related to 974,300 more pesos. Coefficients for this variable are slightly lower under the de la Rúa administration, where  a 1% increase in the percentage of poor unemployed people is related to 743,700 more pesos (see Table 4, model 1). 

What proves most interesting is the effect of popular protest and party politics on the distribution of funds among provinces. In the first place, although the protest variable has the hypothesized sign, social conflict did not make a significant impact during the Menem administration. But it does prove positive and significant when the Alianza ruled the country²⁷. Coefficients for the period 2000-20001 suggest that each protest amounted to 5,000 more pesos for the province where it took place. Although the value of these coefficients is relatively low, it is substantive considering that each employment plan amounted 200 pesos. The province of Buenos Aires, for instance,  witnessed 452 roadblocks in 2001 and should have received about 2,25 million more pesos. Similarly, a province that shows the median for roadblocks would receive about 400,000 more pesos.

While it cannot be denied that the “puebladas”[social uprisings] in 1996-1997 influenced the government’s decision to launch the Plan Trabajar (notwithstanding the fact that roadblocks were always related to claims for employment programs), the causal effect of picketer protests seems to have passed unnoticed until the de la Rúa administration. Yet this finding should not come as a surprise since it is consistent with the cycle of popular mobilization and the concomitant development of the picketer movement. Between 1996 and 1999, protests constituted an incidental phenomenon, were restricted to some social uprisings in the interior and a few collective claims for food headed by territorial organizations in La Matanza (the biggest electoral district of the country, located in the west of the Buenos Aires province). Encouraged by these foundational events, a new repertory of collective action (the roadblocks) became consolidated as a modality of action and a mechanism to claim for employment. Roadblocks gradually spread out, involving both the picketer organizations led by Luis D’Elía (the Federación de Tierra, Vivienday Hábitat) and Juan Carlos Alderete (the Corriente Clasista y Combativa) and a number of radical and autonomous groups, and picketer organizations more associated to leftist parties²⁸. But it was not until  2000 that the multiple branches of the movement escalated to national politics thanks to generalized and successful roadblocks led by groups from La Matansa and the south of the province. The upward cycle reached its peak in the demonstrations held by the Asamblea Nacional Piquetera [National Picketer Assembly] in July and September 2001. These events demonstrate that although the movement did not respond to a unified political leadership, it was able to engineered coordinated actions called ”national days of roadblocks ²⁹

Regarding the role played by party politics, there is partial evidence that provinces with a larger number of municipalities ruled by the President’s party were awarded. On the one hand, as it is stated in Hypothesis 3, the variable “municipalities PJ” is positive and significant during the Menem administration³⁰. In accordance with previous findings, this suggests that initially the local nature of protests led to a distribution of funds that rewarded Peronist mayors in a way they could reinforce their political structures, mobilize activists and thus prevent new social turmoils. On the other hand, the coefficients for “municipalities UCR/Alianza” are positive during de la Rúa administration, but lack statistical significance. The Alianza administration, then, did not reward provinces with a larger proportion of municipalities ruled by its politicans. One possible explanation for this result may lie in the fact that with the Peronism in the opposition, the national government’s prioritized stopping the picketer escalate rather than using resources with a partisan goal, in the hope that governance might remain unaffected by the social conflict.

Hypothesis 3 also holds that the Executive will penalize provinces whose municipalities are ruled by mayors from the main opposition party. Results reported in Table 3 confirm this idea. As it can be seen from the negative and statistically significant coefficient of  the municipalities UCR variable, the Menem administration showed a clear tendency toward penalize provinces that elected a proportionally higher number of UCR/ Alianza municipal governments. Concretely, the coefficient –3.57 indicates that a province where municipalities governed by UCR/Alianza mayors exceeded by 10% those ruled by Peronist, received 357,000 fewer pesos³¹. Estimates in Table 4, however, also seem to indicate that the Alianza benefited Peronist municipalities. But although the variable municipalities PJ has a positive sign, coefficients are not statistically significant. In sum, while the Peronist government penalized the UCR , the Alianza resorted to a more appeasing strategy in an attempt to preserve its future years in office.

What was the impact of provincial governors? In order to test Hypothesis 4, I estimated models including a dummy variable coded 1 for PJ governors and 0 otherwise between 1996 and 1999, and 1 for UCR/Alianza governors and 0 otherwise for the period 2000-2001³². If intuitions stated in the fourth hypothesis are confirmed, then coefficients for both variables should be positive and greater than zero. Evidence found in models 3 is discouraging. First, as we expected, the sign of the variable governors PJ points to the predicted direction, but coefficients lack statistical significance. Second, against our expectations , the sign of the variable governors UCR/Alianza is negative but not statistically significant. Whereas the first result indicates that the Menem administration did not differentially reward Peronist governors, the second suggests that the Alianza distributed fewer resources among the provinces ruled by its own coalition. This does not mean, however, that PJ governors were rewarded with larger funds when the Peronist party was in the opposition. Contrary to Hypothesis 5, models for 2000-2001 including the variable governors PJ clearly show that the national government did not allocate more funds to Peronist governors. Therefore, even though the Alianza needed the PJ’s support to implement pending structural reforms and a package of fiscal adjustment measures in the provinces –the so-called “Deficit Cero” policy – demanded by international credit agencies³³, the Plan Trabajar does not seem to have played a role in trading legislative votes for favors.

Finally, how influential were national legislators? Recall that Hypothesis 6 tests both the political power of province i on the president’s legislative coalition, and the “indirect” effect of governors over national legislators. None of the results indicates that the national governments allocated resources of the Plan Trabajar through Congress: note that the variable president’s coalition lacks statistical significance in all models and it oddly has a negative sign for the Menem administration³⁴. A possible interpretation, is that the PJ built distributive channels resorting to its party mayors. The Alianza, on the other hand, may not have used Congress partly due to the PJ’s cooperation in issuing crucial laws for the government and due to the absence of conflicts between the government and the UCR/Alianza governors. These local politicians supported (and depended on) the continuity of the coalition for their own political survival. Regardless of these nuances, the overall analysis indicates that politics played a major role in defining who got what in the intergovernmental distributive game.

Final considerations

At a time where there is a large number of studies on the picketer movement, it is alarming that none of them attempt a systematic analysis of the impact of unemployed organizations on public policies. While it is undeniable that existing studies contribute to our understanding of the phenomenon, results presented here indicate the need to abandon discussions on the origin and development of the movement to begin exploring the political scope of its actions. Moreover, evidence on the effects of party politics suggests the insufficiency of undertaking a shallow study of the efficacy of social expenditure to allay or neutralize the distributive costs of economic changes. Obviously, this was not the sole concern underlying the implementation of  the Plan Trabajar. Nor was it the single factor that defined who obtained what. There were many other motives: to contain social conflict, to build up clientelar nets at the subnational level, and to promote political interchange between officialism and opposition. In short, the officials responsible for the plan’s distribution did not merely act as technicians or social planners but actively engaged in politics.

This study also presents some other issues that deserve closer attention. The first and most obvious is that existing data impede a reliable conclusion about the ultimate destination of resources. This could be solved with information at the municipal and/or  individual levels. The causal connection between protests and resource distribution entails a more substantive issue. This analysis has demonstrated the impact of protests on the federal allocation of funds, but further studies investigate analyze whether there is also an inverse relation. Programs like the Plan Trabajar might encourage conflict rather than prevent it, or at least stimulate contested actions by the most underprivileged sectors when it comes to the distribution of resources. Finally –and this is a more general concern –observers should analyze the impact of different institutional settings, party-policy alignments, and the economic independence of governors and mayors in the face of clientelar nets such as the ones built up in federal systems like Argentina’s.

Bibliography

Acuña, Carlos, and Smith, William (1994): “The Political Economy of Structural Adjustment: The Logic of Support and Opposition to Neoliberal Reforms”, in W. Smith, C. Acuña and E. Gamarra (comps): Latin America Political Economy in the Age of Neoliberal Reform, New Brunswick, Transaction Books, pp. 17-66.

Auyero, Javier (2001): La política de los pobres. Las prácticas clientelistas del peronismo, Buenos Aires, Manantial.

Auyero, Javier (2002): La protesta: Retratos de la beligerancia popular en la Argentina democrática, Buenos Aires, Libros del Rojas.

Banco Mundial (2000): “Implementation Completion Report. Republic of Argentina. Second Social Protection Project”, Washington D.C.

Blake, Charles (1994): “Social Packs and Inflation Control in New Democracies”, Comparative Political Studies 27 (3), pp. 381-401.

Cao, Horacio (1999): Elecciones y reelecciones en el nivel municipal argentino: 1983-1999, Buenos Aires, Grupo Provincias/Jefatura de Gabinete de Ministros.

Calvo, Ernesto, and Murillo, María V. (2004): “Who Delivers? Partisan Clients in the Argentine Electoral Market”, American Journal of Political Science  48 (4): 742-57.

Conniff, Michael (1999): “Introduction”, in M. Conniff (comp.): Populism in Latin America, Tuscaloosa, University of Alabama Press, pp. 1-21.

Cortés, Rosalía, and Marshall, Adriana (1999): “Estrategia económica, instituciones y negociación política en la reforma social de los noventa”, Desarrollo Económico 39 (154), pp. 195-212.

Delamata, Gabriela (2002): “De los estallidos provinciales a la generación de las protestas en Argentina”, Revista de Ciencias Sociales,  Universidad de Quilmes, 14.

De Luca, Miguel, Jones, Mark, and Tula, María Inés (2002): “Back Rooms to Ballot Boxes?: Candidate Nomination in Argentina”, Comparative Political Studies  35 (4), pp. 413-36.

Díaz-Cayeros, Alberto, Magoli, Beatriz, and Weingast, Barry (2002): “Democratization and the Economy in Mexico: Equilibrium (PRI) Hegemony and its Demise”, mimeo.

Etchemendy, Sebastián (2001): “Construir coaliciones reformistas: la política de las compensaciones en el camino argentino hacia la liberalización”, Desarrollo Económico  40 (160), pp. 675-706.

Gamson, William (1990): The Strategy of Social Conflict, Belmont, Wadsworth.

Gasparini, Leonardo (1999): “Incidencia Distributiva del Gasto Público”, mimeo, FIEL.

Giarraca, Norma (2001): La protesta social en Argentina: Transformaciones económicas y crisis social en el interior del país, Buenos Aires, Alianza.

Gibson, Edward (1997): “The Populist Road to Market Reforms: Policy and Electoral Coalitions in Mexico and Argentina”, World Politics 49 (3), pp. 339-70.

Gibson, Edward, and Calvo, Ernesto (2000): “Federalism and Low-Maintenance Constituencies: Territorial Dimensions of Economic Reforms in Argentina”, Studies in Comparative International Development 35 (3), pp. 32-55.

Gibson, Edward, Calvo, Ernesto, and Falleti, Tulia (2004): “Reallocative Federalism: Legislative Overrepresentation and Public Spending in the Western Hemisphere”, in E. Gibson (comp.): Federalism and Democracy in Latin America, The Johns Hopkins University Press, Baltimore and London, pp. 173-96.

Giugni, Marco (1998): What is Worth the Effort? The Outcomes and Consequences of Social Movements”, Annual Review of Sociology 24, pp. 371-93.

Giugni, Marco (1999): “How Social Movements Matter: Past Research, Present Problems, Future Developments”, in M. Giugni, D. McAdam y C. Tilly (comps): How Social Movements Matter, University of Minnesota Press, Minneapolis, pp. xiii-xxxiii.

Gobierno de Argentina (1999, 2000, 2003): Caracterización y Evolución del Gasto Público Social, Buenos Aires, Secretaría de Programación Económica y Regional.

Graham, Carol (1994): Safety Nets, Politics, and the Poor. Transitions to Market Economies, Washington D.C., The Brookings Institution.

Haggard, Stephen y Kaufman, Robert (1992): The Politics of Economic Adjustment, New York, Oxford University Press.

Haggard, Stephen and Webb, Steven (1994): “Introduction”, in S. Haggard y S. Webb (comps.): Voting for Reform, New York, Oxford University Press, pp. 1-36.

Kohli, Martin, Rein, Martin, Guillemard, Anne-Marie, and van Gunsteren, Herman (1991): Time for Retirement. Cambridge, UK: Cambridge University Press.

Koopmans, Ruud (1993): “The Dynamics of Protest Waves: West Germany, 1965 to 1989”, American Sociological Review  58, pp. 637-58.

Jalan, Jyotsna, and Ravallion, Martin (1998): “Income Gains to the Poor from Workfare. Estimates for Argentina’s Trabajar Program”, Washington D.C., Banco Mundial.

Jennings, Edward (1979): “Civil Turmoil and the Growth of Welfare Rolls: A Comparative State Policy Analysis”, Policy Studies Journal 7, pp. 739-45.

Jones, Mark (2001): “Political Institutions and Public Policy in Argentina. An Overview of the Formation and Execution of the National Budget”, in S. Haggard and M. McCubbins (comps), Presidents, Parliaments, and Policy, Cambridge, Cambridge University Press.

Jones, Mark, Saiegh, Sebastián, Spiller, Pablo, and Tommasi, Mariano (2002): “Amateur Legislators-Professional Politicians: The Consequences of Party-Centered Electoral Rules in a Federal System”, American Journal of Political Science 46 (3), pp. 656-69.

Knight, Alan (1998): “Populism and Neo-populism in Latin America, especially Mexico”, Journal of Latin American Studies 30 (2), pp. 223-48.

Leaman, David (1999): “Populist Liberalism as Dominant Ideology”, Studies in Comparative International Development 34 (3), pp. 98-118.

Levitsky, Steve (2003): Transforming Labor-Based Parties in Latin America. Argentine Peronism in Comparative Perspective, Cambridge, Cambridge University Press.

Lodola, Germán (2002): “Social Protests under Industrial Restructuring: Argentina in the 1990s”, Paper submitted to the V CALACS Conference, Montreal, October.

Lodola, Germán (2004): “Neopopulismo y compensaciones a los perdedores del cambio económico en América Latina”, Diálogo Político 2, pp. 11-37.

Marquez, Gustavo (2000): “Labor Markets and Income Support: What Did We Learn from the Crisis?”, Working Paper 425, Washington D.C., BID.

Marshall, Adriana (1997): “State Labor Market Intervention in Argentina, Chile and Uruguay: Common Model, Different Versions”, Employment and Training Papers 10, ILO.

McAdam, Doug (1999): Political Process and the Development of Black Insurgency, 1930-1970, Chicago, Chicago University Press.

MTSS (1997): Informe de Coyuntura Laboral, Buenos Aires.

MTSS (1998): Programa de la Seguridad Social. Segundo Semestre, Buenos Aires.

MTSS (2003): Boletín de Estadísticas Laborales, Buenos Aires.

Murillo, María Victoria (1997): “La adaptación del sindicalismo argentino a las reformas de mercado en la primera presidencia de Menem”, Desarrollo Económico 147 (37), pp. 419-46.

Nelson, Joan (1990): Economic Crisis and Policy Choice: The Politics of Adjustment in the Third World, Princeton, Princeton University Press.

Oviedo, Luis (2001): Una historia del movimiento piquetero. De las primeras coordinadoras a las asambleas nacionales, Buenos Aires, Ediciones Rumbos.

Oxhorn, Philip (1998): “The Social Foundations of Latin America’s Recurrent Populism”, Journal of Historical Sociology 11 (2), pp. 212-46.

Piven, Frances, and Cloward, Richard (1977): Poor People’s Movements: Why They Succeed, How They Fail, New York, Vintage.

Piven, Frances, and Cloward, Richard (1992): “Normalizing Collective Protest”, in A. Morris and C. Mueller (comps.): Frontiers in Social Movement Theory, New Haven, Yale University Press, pp. 301-25. 

Piven, Frances, y Cloward, Richard (1993): Regulating the Poor: The Functions of Public Welfare, New York, Vintage.

Przerworski, Adam (1991): Democracy and the Market: Political and Economy Reforms in Eastern Europe and Latin America, Cambridge, Cambridge University Press.

Ravallion, Martin (2002): “Are the Poor Protected from Budget Cuts? Evidence from Argentina”, Journal of Applied Economics 5 (1), pp. 95-121.

Remmer, Karen, and Gélineau, Francois (2003): “Subnational Electoral Choice. Economic and Referendum Voting in Argentina”, Comparative Political Studies 36 (7), pp. 801-21.

Remmer, Karen and Wibbels, Erik (2000): “The Subnational Politics of Economic Adjustment”, Comparative Political Studies, 33 (4), pp. 419-51.

Repetto, Fabián (2000): “Gestión pública, actores e institucionalidad: las políticas frente a la pobreza en los ‘90”, Desarrollo Económico 39 (156), pp. 597-617.

Roberts, Kenneth (1995): “Neoliberalism and the Transformation of Populism in Latin America: The Peruvian Case”, World Politics 48 (1), pp. 82-116.

Roberts, Kenneth, and Arce, Moisés (1998): “Neoliberalism and Lower-Class Voting Behavior in Peru”, Comparative Political Studies 31 (2), pp. 217-246.

Ronconi, Lucas (2002): “El Programa Trabajar”, Centro de Estudios para el Desarrollo Institucional-Fundación Gobierno y Sociedad, Documento 63.

Schumaker, Paul (1978): “The Scope of Political Conflict and the Effectiveness of Constraints in Urban Politics”, Sociological Quarterly 19, pp. 168-84.

SIEMPRO (1997): Evaluación de Programa Trabajar I, Buenos Aires.

SIEMPRO (1998): Evaluación de Programa Trabajar II, Buenos Aires.

Smulovitz, Catalina (2003): “The (Unforeseen) Effects of the Trabajar Program on Empowerment in Argentina”, mimeo, UTDT.

Stokes, Susan (comp.) (2001): Public Support for Market Reforms in New Democracies, Cambridge, Cambridge University Press.

Svampa, Maristela, and Pereyra, Sebastián (2003): Entre el barrio y la ruta. La experiencia de las organizaciones piqueteras, Buenos Aires, Ediciones Biblos.

Tarrow, Sidney (1994): Power in Movement: Social Movements, Collective Action, and Politics, Cambridge, Cambridge University Press.

Torre, Juan Carlos (1998): El proceso político de las reformas económicas en América Latina, Buenos Aires, Paidós.

Weyland, Kurt (1996): “Neo-populism and Neoliberalism in Latin America: Unexpected Affinities”, Studies in Comparative International Development 31 (3), pp. 3-31.

Weyland, Kurt (1998): “Swallowing the Bitter Pill. Sources of Popular Support for Neoliberal Reform in Latin America”, Comparative Political Studies, 31 (5), pp. 539-68.

Weyland, Kurt (1999): “Neoliberal Populism in Latin America and Eastern Europe”, Comparative Politics 31 (4), pp.379-401.

Weyland, Kurt (2002): The Politics of Market Reforms in Fragile Democracies. Argentina, Brazil, Peru, and Venezuela, Princeton, Princeton University Press.

Weitz, Rebecca (2003): “The Politics of Workfare Distribution in Argentina”, mimeo, Columbia University.

Footnotes

* I am grateful to the comments and suggestions made by Barry Ames, John Markoff, Aníbal Pérez-Liñán, Juan Carlos Torre, and two anonymous commentators. Preliminary versions of this study were submitted at the XXIV LASA Conference, Dallas, Texas, March 2003, and at the V LACEA-PEG Conference, Cartagena de Indias, Colombia, September 2003.
^** Name given in Argentina to political activists with an influence on inhabitants of poor neighborhoods [T.N]
1 The affinity between neoliberalism and neopopulism was first posted by Roberts (1995) and Weyland (1996). A few important works on this subject are Conniff (1999), Gibson (1997), Knight (1998), Leaman (1999), Oxhorn (1998), Roberts and Arce (1998), and Weyland (1999, 2002). From different theoretical stances, similar versions of this thesis can also be found in Blake (1994), Graham (1994), and Haggard and Webb (1994). ]]> 2 This does not imply ignoring that the government offered material compensations to potential losers in the formal sector, such as trade unions and local entrepreneurs (see Etchemendy, 2001). My argument rounds this off by pointing out that similar compensatory mechanisms were not created for the popular sectors excluded by adjustment policies. In other words, the Menem administration promoted compensation of “powerful” actors while abandoning the idea of dealing with “weak” actors on equal terms.
3 There is a vast corpus of literature about the political impact of social movements in developed countries. For excellent reviews, see Giugni (1998, 1999). 
4 See, for example, Calvo and Murillo (2004), Gibson and Calvo (2000), Gibson, Calvo and Falleti (2004), Remmer and Wibbels (2000) and Remmer and Gélineau (2003).
5 In Argentina, real salaries dropped by 20% between 1990 and 1995. The unemployment rate during the same period grew from 8.6% to 18.4%, and fell to 16.4% in 2001. Underemployment showed a steady increase, moving up from 9.3% in 1990 to 14.9% in 2001. This led to an increase of poverty levels. While in 1990 the poor and the destitute constituted about 22% and 3% of the population respectively, the percentages for 2001 raised to 28% and 8% (INDEC and MTSS 2003).
6 Emblematic cases are the Fondo Nacional de Compensación y Desarrollo [National Compensation and Development Fund] (FOCONDES) established by Peruvian President Alberto Fujimori (1990-2000); the Programa Nacional de Solidaridad [National Solidarity Program] (PRONASOL), and the Programa de Educación, Salud y Alimentación [Education, Health, and Nourishment Program] (PROGRESA) implemented in Mexico during the administrations headed by Carlos Salinas de Gortari (1988-1994) and Ernesto Zedillo (1994-2000) respectively; the Fondo Social de Emergencia [Social Emergency Fund] (FSE) launched by President Víctor Paz Estenssoro (1985-1989) in Bolivia; and the so-called Mega Proyecto Social [Social Megaproject] designed in Venezuela by Carlos Andrés Pérez’s administration (1989-1993).
7 For example, Augusto Pinochet’s regime (1973-1990) created the Programa de Empleo Mínimo [Minimal Employment Program] and the Programa Ocupacional para Hogares [Household Occupational Program] when the economic crisis that unleashed in the early 80s triggered a wave of protest led by the “pobladores” (the urban poor) in Santiago de Chile. Likewise, Peruvian President Alan García from the APRA (1985-1990), implemented the Programa de Apoyo al Ingreso Temporario [Program in Support of Temporary Income] and the Programa de Asistencia Directa [Direct Relief Program] both of which especially targeted the urban poor and provided a basis for the design of the Plan Trabajar. 
8 In this study, the years in which the control of the Executive Power was divided (i.e. 1983, 1989, and 1999) are considered years ruled by the outgoing president, since transfer of power took place in December. 
9 Gasparini (1999) reports that in 1996, about 30% of the expenditure on social services and less than 10% of the expenditure on social security reached the poorest quintile of households. 
10 Yet average social expenditure allocated to social services was somewhat lower during the Menem administration than it was during Alfonsin’s government (42.8% and 45.6% respectively).
11 There could not be a clearer difference between this experience and other experiences that go by the name of neopopulism. In effect, PRONASOL and PROGRESA yielded higher figures than the aggregated amount of all existing social assistance programs in the region. In real terms, Perú’s focalized expenditure grew by 60% between 1993 and 1994, and reached something like 90% in 1995. This was accompanied by a staggering increase in social expenditure, which increased from 3% of the GDP in 1993 to 7.8% in 1995. In Venezuela, social expenditure amounted to 32.1% of total public expenditure in 1989, and mounted to 37.1% in 1991, while the Mega Proyecto Social covered 9.7% of the national budget. Finally, resources allocated to the FSE involved more than 10% of the GDP. ]]> 12 Using a temporal series for the period 1980-1997, Ravallion (2002) shows an elastic response of social expenditure to the cutbacks in public expenditure, whereas its response to fiscal expansion is statistically insignificant. Along the same lines, the author reports that social expenditure was not adequately protected over the years of fiscal contraction.
13 Roberts (1995) coined the notion of “micro-level populism” to characterize reformist governments that implemented social policies in harsh fiscal adjustment settings.
14 The Venezuelan case under Carlos Andrés Pérez also shows that this was not a sufficient condition. Therefore, focalization of social policies should not be considered as a defining feature of neopopulism. On this issue, see Lodola (2004).
15 In 1984, the Alfonsin administration implemented an unemployment insurance system that guaranteed 70% of the minimal wages or a salary support for a period of four months to a very small number of unemployed workers.
16 The cover period (between four and twelve months) and the replacement rate (60% of the last six salaries) remained close to the average observed for the region.
17 For the origin of protests that in 1996-1998 erupted in oil  (Cutral-Có and Plaza Huincul in the province of Neuquén; Tartagal and General Mosconi in Salta) and sugar cane enclaves (Libertador General San Martín in Jujuy) depply harmed by the privatization of YPF, the State oil company, as well as by ruthless industrial reconversion processes, see Auyero (2002), Delamata (2002), Giarraca (2001), and Lodola (2002).
18 I mean the so-called “manzaneras”, women in charge of collecting information to aid food distribution provided by the Plan Vida in the province of Buenos Aires.
19 Governors, mayors, “punteros”, and delegates of the more temperate picket organizations, as well as NGOs members, used to participate in these Councils. For an analisis of their functioning, see Smulovitz (2003).
20 They mostly made use of the Plan Trabajar as an organizational resource that might help them deepen their intervention in neighborhoods, attract new members by selectively offering employment plans, and strengthen ties among militants.
21 The models were also estimated with a dummy variable coded 1 for the years when the Alianza was in power (2000-2001) and 0 otherwise, and interacting this term with all the independent variables. The results yielded by the interactive models (which are more complicated to interpret because we need to estimate conditional coefficients and conditional standard errors for each interaction) are similar to those obtained in the additive models reported below. ]]> 22 Although Capital Federal and Tierra del Fuego have different norms for the distribution of federal resources, these norms did not apply to the Plan Trabajar. Therefore, there are no reasons to exclude these provinces from the analysis.
23 As the Plan Trabajar was specifically targeted to low-income unemployed workers rather than to the general population, the dependent variable is measured in total amounts rather than in per capita amounts, controlling for population.
24 The data available starts in 1997. Data for 1996 are my own, based on information published by La Nación  and  Clarín newspapers. Existing data does not provide details about the number of participants, detainees, and casualties, all of which prevents us from building an index to capture the intensity of protests. However, studies on the role of social protests in developed countries tend to find a positive correlation between their frequency and intensity.
25 Since variables measuring municipal power show a 99% negative correlation (Pierson Watson = -0.61 for both periods), I decided to estimate separate models for the percentage of municipalities governed by the PJ and the UCR/Alianza respectively. Thus, although ideally the models should include both variables, I decided to avoid multicollinerarity problems that seriously affect results. See Díaz-Cayeros, Magoli and Weingast (2002), who use this procedure to analyze the federal distribution of the PRONASOL in Mexico.
26 There are good reasons to assume a contemporary effect of the independent variables on the federal distribution of the Plan Trabajar. As it has already been stated, the Executive concentrated the distribution of funds and allocated them to the provinces on a monthly base. This enabled the President to engineer distribution in accordance with the political priorities of the moment. In other words, the assumption (confirmed by several interviews) is that the government took advantage of their margin for discretionality to enact a political reaction in accordance with the course of events. As we will see, this does not mean that the distribution of employment plans was completely independent of guidelines and/or control mechanisms. As Weitz points out (2003, see also Ravallion, 2002) in a critique to a preliminary version of this study, the intervention of the World Bank seems to have induced a distribution of the plan that took into account its anti-poverty component.
27 There are no significant changes in the results when the models are estimated using NBI (poverty compensation) and the unemployment index (DES) rather than using the percentage of poor unemployed per province. The coefficients for the variable “protest” remain significant in 2000-2001, while those for NBI and DES are slightly lower and statistically significant.
28 For a  historiography of the picket movement, see Oviedo (2001) and Svampa and Pereyra (2003). 
29 The I Asamblea Nacional Piquetera [First National Picket Assembly] decided to launch a struggle plan that combined gradual and full roadblocks of 48 and 72 hours on the main 50 country roads. The II Asamblea Nacional Piquetera [Second National Picket Assembly] extended this strategy by blocking access to factories and calling a national strike with mobilization by state workers’ unions.
30 These results proved robust in estimates made with NBI and DES.
31 Results remain virtually the same in estimates made with NBI and DES. ]]> 32 Results remain unchanged when these variables are coded 0 for non-Peronist governors (whether or not they are from the UCR) and 0 for non-Alianza governors (whether or not they are from the PJ) respectively.
33 It should be noted that governors had frequently declared that a shrinking of federal fiscal resources would result in violent social turmoil in their territories.
34 In Gibson and Calvo’s interesting study (2000), the authors demonstrate that the success of economic reforms under the Menem administration lay mostly on the legislative support provided by peripheral provinces which, owing to the characteristics of Argentinean federalism, are overrepresented in Congress.  Along the same lines, Gibson, Calvo and Falleti (2004) show that overrepresentation in the Lower House promotes reallocation of federal funds and discretional expenses that favors small provinces. In order to test for the effect of federalism (overrepresentation) on the distribution of the Plan Trabajar, I estimated models with a dummy variable coded 0) for metropolitan provinces (City of Buenos Aires, Buenos Aires, Córdoba, Santa Fe and, alternatively, Mendoza) and 1 for peripheral provinces. In none of the cases do the results support the hypothesis that overrepresented provinces received more resources.