Abstract
The goal of this paper is to suggest that municipalities can facilitate the central government's ability to carry out its desired policy goals. Using three institutional variables that provide internal evidence for each case, I will argue that within some institutional configurations the center may seek uniformity of outcome by trying to circumvent governors and/or regional intermediation and forge a direct relationship with municipalities. Based on research from the area of social protection policy in Brazil and Argentina, I suggest that direct national-local collaboration contributed to the ability of the Brazilian central government to bypass governors and evenly spread non-contributory welfare goods across its territory and alleviate poverty. I argue that such policy collaboration is less likely to be institutionally feasible in a federal system like Argentina's, where such equivalent collaboration is impeded by the ability of its provinces to directly capture local units of government and undermine national policy outcomes and executive preferences.
Introduction
Much progress has been made by academics in explaining the varying outcomes of decentralization in Latin America at both the national level, (Fox 1994; Abrucio 1998; Faletti 2003; Mainwaring and Samuel 2004; Stepan 2004) and the subnational level (Eaton 2004; Gibson 2005; Ward et al. 2006; Montero 2008; Durazo 2010 in this issue). Less attention has been paid to the municipal level (Nickson 1995; Gómez 2003), although increasingly this level of government is being included in what González aptly describes as “the power distributional approach” (2008: 212) of understanding Latin American federations. 1
The author would like to thank Al Montero, Carlos Gervasoni, and three anonymous reviewers for their insights and helpful feedback.
In three out of four of Latin America's federations, Mexico, Argentina, and Brazil, central governments have been unable to resolve their challenges of governing and spreading democracy territorially without interacting with one or both of the subnational levels of government. 2 During the most recent democratic periods in all three of these countries, it has not been so much the case that subnational actors were dependent on the federal center for pushing reforms and ensuring policy outcomes, but that the central government has in fact been dependent on subnational actors both to pursue national executive preferences and to carry out national policy objectives. In short, federalism in Latin America has more often than not enabled subnational levels of government to either constrain or free-ride the central government's preferred options (Abrucio 1998; Mainwaring and Samuels 2004; Stepan 2004).
Venezuela will not be referred to herein because of its decreasingly democratic context.
Brazil and Argentina are both highly decentralized strong federal systems. However, varying political, institutional, and structural factors have shaped the ability of each central government to implement their preferred policy goals. Most evident within a paired comparison of these two countries is the emerging capacity of Brazilian federalism to avoid governors (Dickovick 2006; Borges 2007; Fenwick 2009a; Cheibub, Figueiredo, and Limongi 2009). The goal of this paper is to suggest that municipalities can facilitate the central government's ability to carry out its desired policy goals, particularly in social arenas. I will argue that within some institutional configurations the center may seek uniformity of outcome by trying to circumvent governors and/or regional intermediation and forge a direct relationship with municipalities. Based on research from the area of social protection policy in Brazil and Argentina, I suggest that national-local collaboration contributed to the ability of the Brazilian central government to bypass governors and evenly spread non-contributory welfare goods across its territory. In addition, it has been argued that the successful distribution of poverty alleviation benefits contributed to the ability of the Worker's Party (Partido de Trabalhadores) to triumph in the traditionally conservative and neo-patrimonial Northeastern States of Brazil in 2006 (Hunter and Power 2007; Nicolau and Peixoto 2007; Zucco 2008, 2010; Borges 2007; Marques et al. 2009; see also Montero 2010 in this issue). But for this to occur here, or elsewhere, the federalization of national policies must be directly implemented at the municipal level without being captured by powerful governors. I argue that such policy collaboration is less likely to be institutionally feasible in a federal system like Argentina's, where such equivalent collaboration is impeded by the ability of its provinces to directly capture local units of government and undermine national policy outcomes and executive preferences.
This paper suggests that three key institutional variables shape the feasibility of direct national-local collaboration. The three variables I propose and the corresponding propositions relating to the institutional feasibility of direct national-local policy collaboration are the following:
Variable 1: The Constitutional Recognition of Subnational Levels of Government
Direct national-local policy collaboration only becomes de jure feasible when municipalities are recognized within a national constitution as being distinct from the states and/or provinces. For example, although many features of a majoritarian party system lend itself towards the ability of the intermediate levels of government 3 to capture local units within their territory, a priori in many federal systems, local units are constitutionally subordinate. By nationally recognizing municipalities as autonomous federal units that are distinct, three of the Americas’ six federations in their Constitutions go de jure beyond Riker's supposition of two levels of government: Brazil (1988: Article 18), Venezuela (1999: Article 168), and Mexico (1983: Article 115). While Argentina (1994: Article 5), the USA and Canada, continue to subordinate municipalities to provincial constitutions, seeing them as creatures of either the states or provinces There are nuanced, yet important differences across federal systems. Weak governors can become stronger when united with mayors, and strong governors can become weaker when mayors are bolstered by a stronger central government. By distinguishing municipalities as a separate actor, a dynamic intergovernmental relationship is represented; indeed, there is shown to be a power struggle between multiple levels of government and their electorates, not just a dichotomously framed two-level game of the national level versus the subnational.
Intermediate levels of governments in this paper always refer to states and/or provinces.
Variable 2: Political Dynamics
A decentralized federal system requires intergovernmental (and interparty) cooperation to carry out a successful national policy objective that contributes to “spreading clusters of rights across its territory.” 4 Each federal system's political dynamics contributes to the amount of political will across levels of government that can be expected to achieve such a policy goal. Using Lijphart's strategy of creating a dichotomous overarching typology, I reduce and synthesize many party and electoral system variables into this single category and refer to them as majoritarian or non-majoritarian political dynamics. I use this strategy to highlight the existence of contrasting governing logics among federal systems, not to contrast between federal and unitary systems. It is a very broad categorization that reduces many complex variables into a simple dichotomous framework. I believe, however, that it offers a successful analytical strategy, particularly when comparing two presidential federal countries whose differences can be quite subtle.
This is Edward L. Gibson's definition of territorial democratization (2008).
Majoritarian political dynamics, in the sense in which I am using them, are party-centered. They are characterized by a concentration of executive power, single-party majority cabinets, executive-legislative relations where the Executive is dominant, party systems where the number of effective parties hovers at around two, and/or, exhibits an integrated federal party. An integrated federal party or party system is essentially one in which “politicians at one level of government bear an organizational relationship to politicians at other levels”; or the essential commodity between them is a shared party label (Ordeshook, Filippov, and Shvetsova 2004: 190). In a decentralized federal system, the strength of majoritarian political dynamics varies across levels of government. Logically, however, a country's political dynamics will be more majoritarian when electoral rules and the party system encourages political parties and candidates from multiple levels of government to rely on each other for their survival and success.
Non-majoritarian political dynamics push in the direction of consensus. They are characterized primarily by executive power-sharing in broad multiparty coalitions, an executive-legislative balance of power, a high number of effective parties, and local autonomy (i.e., minority veto). With these political dynamics, elite political actors are not so concerned with office-seeking but with policy-seeking (Strom 1990). Based on electoral motives, policy success allows them to personally credit-claim within their territories. Since politicians have to win the support of such an array of groups, the nature of these dynamics, particularly when combined with open-list electoral rules, means that “the survival of most politicians depends on their ability to deliver goods to the regions they represent” (Mainwaring 1992: 682). This holds even if it means cooperating with a party to whom they are not officially affiliated. In such a context, intergovernmental policy cooperation is facilitated by a loosely maintained coalitional logic of governance which decreases the ideological or partisan ownership of policy ideas. This allows for both the “designer” and the “deliverer” of policy outputs to mutually credit-claim among their constituents. It is in this way that non-majoritarian political dynamics can constitute a win-win game for key political actors at different levels who can mutually choose to deliver a successful policy.
Variable 3: The Rules Regulating Subnational Finances
The rules regulating subnational finances are an integral part of understanding the politics of the federal game. They help determine what fiscal incentives are present for intergovernmental policy collaboration. Political economists specializing in fiscal federalism are most concerned with the consequences for overall macroeconomic performance. The effects of recent processes of fiscal decentralization on economic performance in Argentina and Brazil and elsewhere in the Americas have been well documented (Tommasi Sebastian, and Sanguinetti 2001; Dillenger and Webb 2002; Giugale, Trillo, and Oliveira 2000; Rodden and Wibbels 2002; Haggard and Webb 2004). Less attention has been paid to how the rules regulating subnational finances affect the political will of subnational levels of government to implement policy.
Ideally, the ability of these governments to manage their spending should be regulated by market discipline. In practice (particularly in less developed countries), this may not be such an ideal solution. Market-based regulations are known as “soft budget constraints.” 5 They refer to the responsibility of the subnational governments to cover their budget expenditure out of their allotted central government transfers and own source revenues. If they fail to cover the expenditure of their budgets, they cannot survive without outside fiscal intervention. This intervention may come from either the central government or from accessing external credit. Beyond influencing macro-economic performance, soft budget constraints shape political behavior and intergovernmental relations. Political motives often induce a level of government to extend fiscal support to a lower level of government in need because of already-given political support, or of patronage that it would lose if it did not. Three federal countries in the Americas, Argentina, Brazil, and Canada, attempted this mechanism during the 1980s, at significant social cost by the end of the 1990s (Guigale, Trillo, and Oliveira 2000; Rodden and Wibbels 2002).
The term and theory was coined by Janos Kornai (1980).
There is an alternative, rule-based approach to using market discipline to monitor the capacity of lower levels of government to pay, and stay out of excessive, debt. Referred to as a situation where subnational levels of government will not receive outside support to cover their excessive spending and will thus be obliged to reduce or terminate an activity if the deficit persists (Kornai, Maskin, and Roland 2002), known as “hard budget constraints.” The advantage of hard budget constraints over soft budget constraints within the context I am using them is that it makes it harder for governors to hold mayors fiscally hostage. It weakens the interdependent relationship between the intermediate and local levels of government. Such an approach has been implemented, to varying degrees, in Brazil and the USA, and among other federal countries outside of the Americas. In practice, it normally entails the central government limiting subnational borrowing capacity, regulating public sector spending, enforcing bankruptcy and fiscal responsibility laws, and private risk-rating. This budgetary regulation approach has the advantage of improving transparency and impartiality, and encouraging the minimalization of political bargains and discretionality (Giugale, Trillo, and Oliveira 2000: 257). Most importantly, hard budget constraints provide a fiscal incentive for subnational levels of government to implement nationally financed public policy.
Insights from Brazil and Argentina
Using the variables broadly outlined above, I will show how direct national-local policy collaboration was institutionally feasible in Brazil, but not in neighboring Argentina. These insights emanate from research I conducted on two nationally defined social programs designed to alleviate poverty in Brazil and Argentina. The outcomes of the two most recent national social programs I analyzed have been varied. Brazil's newest national program, Bolsa Família, was very successful in terms of numbers, territory and social groups covered. By 2006, poverty alleviation benefits were being delivered to more than 11 million households in all 5,564 Brazilian municipalities. By contrast, Argentina's latest national initiative, Programa Familias, has been less successful in terms both of territorial coverage and its total distribution. By 2006, it delivered benefits to only 372,000 households in 232 of 1,930 units of local government. 6 This specific policy sector has been selected to serve as a springboard for analyzing how the federalization of social policy, directly implemented at the municipal level, was institutionally feasible in the context of strong federalism where governors notoriously protect their territorial boundaries.
Local units of government are not referred to as “municipalities” in all Argentine Provinces.
The three broadly conceived institutional variables provided internal evidence for each case. The variables I proposed relating to the institutional feasibility of direct national-local policy collaboration are as follows for each case.
Evidence from Brazil
Increasingly in post-1995 Brazil, locally designed anti-poverty programs and primary health and basic education initiatives were nationalized, but continued to rely on municipal implementation. In these policy areas, successful examples of direct national-local policy collaboration have contributed to the ability of the central government to deliver on their social policy promises. It has also contributed to the decreasing ability of governors to undermine or free-ride the central government by either blocking or not carrying out national policy objectives. Dickovick (2006) argued that decentralization to the municipal level in Brazil was both a governing strategy and a second-best option to further recentralization. In my own research (Fenwick 2009b), I assert that municipalization was the unintended consequence of the radical decentralization measures of the 1988 National Constitution, which post-1995 were reined in more at the state level, yet continued to flourish at the local level in the context of a more stable national political and economic context. Whether in fact it was intended as Dickovick posits, or unintended, it has slowly contributed to weakening the ability of governors to control their territorial boundaries.
Gibson defines “boundary control” as a situation where the leaders of provincial/state regimes maximize their influence over the politics of their territory and deprive opposition-based municipalities of access to national allies and resources (2005: 107). The success of municipalization in Brazil, slowly contributed to making democracy territorially more even by ensuring uniform social policy outputs. Moreover, it is because federalized social programs which rely on municipal implementation like Bolsa Família provide small economic inputs to municipalities (Marques 2004), as well as an opportunity for mayors to mutually credit-claim for its success, that it provides mayors direct access to both national allies and resources.
Municipalization in Brazil did not however automatically occur following the constitutional reforms of 1988; rather, it emerged following a process of gradual muddling through that spanned more than a decade. In terms of political decentralization, the most significant change brought about by the 1988 Constitution is that in Article 1, for the first time in Brazilian history, it recognized municipalities as official constituent units of the federation. This gave them de jure the same status as the 26 states and the União, the federal government. The political determinants of this change are well known to have been linked to the members of the Constituent Assembly (1987-1988) who were Legislators with strong bases in both state-level and municipal-level politics (Souza 1997).
The political power of governors post-1988 nevertheless remained stronger than that of mayors because of their strong representation in Congress and because of the ability of the states to continue exercising “extensive control over local finances” (Abrucio and Samuels 2000: 47). The constitutional recognition of municipalities did begin a process of democratization at the local level that enabled new actors to both innovate and experiment. Although the heterogeneity of municipalities, in terms of size and capacity, could not be resolved through political decentralization (which in fact limited the ability of smaller municipalities to exercise their newly attributed political power), the 1988 Constitution did increase the decision-making responsibility of municipalities regarding the provisioning of public services (Souza 2005). By strengthening the position of mayors, it also served to gradually increase the attractiveness of holding office at this level.
For a political position to appear attractive to key stakeholders beyond simply the chance of electoral victory, it must be meaningful. Prior to democratization in Brazil, open and direct elections were held at the municipal level in 1982, excluding state capitals and zones of national security. Although these elections were direct, party competition was still controlled through restrictive electoral laws. In the 1988 municipal elections (which were both less restrained and direct), the effective number of parties (ENP) elected was 4.5. By the time of the 2000 municipal elections, the ENP represented at this level had increased to seven. In 2004, it increased to 8.9. This increase in the ENP elected at the municipal level indicates that electoral competition at this level increased substantially post-1988, parallel to the growing importance of holding power at this level because of increasing political, policy, and fiscal authority.
Policy and fiscal authority are however very difficult to measure because of the overlapping nature of the Brazilian 1988 Constitution. There has been a shift towards municipal authority over social policy delivery in Brazil, although state and municipal government continue to share federal taxes, municipalities share some state taxes, and the federal government designs and finances programs in key municipal policy authority areas such as health, education and social assistance (Souza 2005). Intergovernmental cooperation is therefore the key to successful governance in Brazil. Post-1995, such cooperation became gradually more evident in key policy sectors because of the increased use of centralized fiscal matching grants, which have decreased state and municipal competition over policy implementation. Moreover under President Fernando Henrique Cardoso's two administrations between 1995 and 2002, municipalities were liberated from the states through new administrative laws and reforms that regulated the competencies of subnational governments. Those included the “Law of Concessions,” the “Law of Guidelines and Bases for Education,” the “Reform on Public Administration,” and the “Statute of Cities” (Arretche 2007). While in theory these laws increased the centralization of policymaking in Brazil, in practice, they contributed to decreasing the opportunistic behavior that had previously occurred between all three levels of government following decentralization in 1988.
A large part of this behavior also emanated from each level of government trying to maximize their access to government revenue. The 1988 Constitution had increased the level of revenue transfers, primarily to the states, by 12.5 percent. Expenditure levels however did not increase parallel with the increase in revenues, leaving the central government in a position of fiscal weakness, with fewer resources yet still ultimately responsible for financing social policy. This practice was gradually altered over time by the hardening of subnational budget constraints and the tightening of the rules regulating subnational finances. Beginning with the fiscal stabilization plan implemented on July 1, 1994 which both lowered inflation and secured Cardoso's electoral victory, the central government managed to shut down state banks and tie the hands of governors to match their revenue and expenditure. Most importantly, the effect of this process on municipalities was not the same as on states. The municipal level began spending far more than their revenue base (calculated in the table below after constitutional transfers) permitted them post-1988. After the recentralization implemented during Cardoso's two terms (1995-2002), municipalities’ revenues increased 13.4 percent from 1988, and their expenditure decreased 7.8 percent (see Table 3).
Variables That Impede or Encourage National-Local Collaboration
Source: Own elaboration.
The Variables that Impede or Encourage National-Local Collaboration in Brazil and Argentina
Source: Own elaboration.
Brazil's Revenue and Expenditure by Level of Government after Decentralization and after Recentralization
Source: Willis, Garman, and Haggard (1999); Afonso (2005).
This process of recentralization, following a process of decentralization, was not unique to Brazil, and was a re-equilibration of power between the center and state governments. Recentralization in Russia and Argentina was contextualized by a strong party system, where partisan dynamics enabled rapid but volatile top-down centralizing economic reforms. Most importantly for the argument I am making, beyond the expected consequences of recentralization for macroeconomic stability, recentralization in Brazil contributed to the unintended consequence of strengthening the position of municipalities, vis-à-vis the states.
Following the Cardoso years, the predatory nature of Brazilian federalism that had begun in 1988 changed accordingly. Fiscal recentralization post-1995 was motivated by necessity, yet successfully replaced some of the burden-shifting characteristics that were apparent post-1988. Fiscal incentives to share power in areas of overlapping constitutional authority increased the potential for “good fiscal accounts,” and thus for claiming credit, at all levels of government. Additionally, the Fiscal Responsibility Law of 2000 (a hard budget constraint), placed limits on the expansion of both expenditure and debt by criminalizing it with penal consequences. Within this context, mayors, who did not have the ability to access “special funds” from Congress as did governors, had an incentive to take on social policy execution responsibilities. The result has been a synthesis of two pre-existing elements, decentralization and recentralization, resulting in a new model of “collaborative federalism” that has been distinguished, broadly speaking, by partnerships between the central and municipal governments over the delivery of social policy. This has forced state governors into an unexpected federal game which pushes participants towards a consensus policy-based direction. The remaining question however is whether there has been an electoral effect at the subnational level emanating from this reconfiguration of power.
Several authors have attributed the observed shifts in the voting patterns of the 2006 presidential and gubernatorial elections to the success of President Luiz Inácio Lula da Silva's social policies, in particular Bolsa Família (Hunter and Power 2007; Nicolau and Peixoto 2007; Borges 2007; Zucco 2008, 2010; Marques et al. 2009). It is evident, that the biggest losers in Bolsa Família were the 26 states that the central government cut out by building on and expanding poverty alleviation initiatives without their involvement. The states cannot credit claim for the design of this program or its implementation. They can also not intermediate the distribution of its funds to local beneficiaries. The municipalization of social policy has not prevented governors from turning to clientelism per se, it does however entail that their ability to utilize social expenditure as source of revenues to funnel into their own patrimonial machines has decreased because of Brazil's unique institutional configuration. It is this configuration that permits the federal government to avoid them, instead relying on municipalities to be their prime agents.
As this debate is in its infancy, information (and elections) are still lacking to conclusively evaluate whether (and how much) municipalization has affected democracy at all levels of government in Brazil. We do however have increasing evidence that municipalization has strengthened the central government's role vis-à-vis the states, and that the era of governor politics in Brazil may slowly be coming to a close.
Evidence from Argentina
Beyond the Brazilian case above, the key question to ask is whether such direct national-local policy collaboration is feasible in other federal countries that are also robust, but where a varying institutional configuration is present. Argentina makes an interesting case to contrast with Brazil because of the relative strength of its own governor politics. Evidence from Argentina suggests that the institutional feasibility of national-local collaboration is limited by an archaic National Constitution that reflects the origins of Argentine federalism and its original interprovincial conflicts (see Gibson and Faletti 2004), strong majoritarian political dynamics, and the politicized “soft” nature of the rules regulating subnational finances. All of these factors have contributed to impeding equivalent national-local collaboration in Argentina and the federalization of social services.
Similar to Brazil, Argentina is a highly decentralized federal country, with 23 provinces, and one autonomous city, Buenos Aires, constituted in 1994. The relative success of democratic federalism in Argentina has historically been linked in the literature to the following factors: the success of a single, integrated, “adaptable” federal party (Levitsky 2003), a two-tier party system (Gibson 1996; Jones 1997; Malamud and de Luca 2005), a moderate to high level of party discipline in the Legislature (Jones and Hwang 2003), and “the nation's federal framework [that] reduces the winner-takes-all nature of politics by providing areas of subnational autonomy for opposition parties” (Jones 1997: 261). There has been very little dissent or variation in the views expressed in the literature covering the years 1983 to 2000.
Unlike in Brazil, Argentina's third tier of federal government, municipalities, are creatures of the provinces. This means that in Argentine federalism the concept of “subnational” remains, both in theory and in practice, an aggregation of both provincial and local power. 7 The federal game in Argentina is therefore played predominantly between two politically powerful levels of government. Relations between these two levels are extremely partisan-based because of Argentina's majoritarian political dynamics which have been traditionally dominated by two political parties, the Radical Civic Union (UCR) and the Justicialist Party (PJ).
“Local” refers always to municipal power or the lowest federal unit, and not, as is sometimes the case in other works about Argentina, provincial or regional power.
Neither the 1853 Constitution, nor the 1994 constitutional reforms, recognized municipal autonomy, although it did identify this third tier as an integral part of federalism. The Constitution formally affirms two levels of government, the central and the provinces. Municipalities are simply mentioned in Article 5, which states the principle that provinces must create municipal regimes (1853/1860). This means de facto that municipalities are creatures of the provinces, although they have at least the constitutional right to exist. Provinces in Argentina are institutionally heterogeneous, with independent authority to write their constitutions, make their electoral laws, set their election dates, and design their municipal regimes. Because of the ambiguity about municipalities in the Constitution,
the degree and extent of municipal autonomy, in the end, depends on the responsibilities/obligation that each provincial constitution, starting from their own respective organic laws, delegates or confers on local entities (Clemente and Smulovitz 2004: 104).
This means that only municipalities within territories that so choose have the right to be considered as an official third tier of federal government. Presently eight out of the 23 provinces do not formally recognize municipal autonomy. 8
They are Buenos Aires, Chaco, Entre Rios, La Pampa, Mendoza, Santa Fe, Santa Cruz, and Tucumán.
The dependent nature of municipalities on the provinces creates a two-level federal game between the provinces and the central government, where the distribution of power, revenue and expenditure fluctuates back and forth between these two dominant players in a debilitating manner. This dichotomous relationship creates a non-cooperative game not easily accessible by either outsiders or newcomers, particularly at the local level. Moreover, it allows the provinces to both overawe and free-ride on municipalities. The central government however, cannot directly overawe municipalities, because they are protected by the provinces. In this sense, one of the goals of federalism is constitutionally assured.
Another however, the ability of provincial actors to block national objectives through either direct action or policy apathy is not easily resolved. Equally, national-level political elites do not always have an incentive to use their power to impose policy uniformity at the subnational level. National leaders may even pursue “heterogenizing” strategies to deliberately create subnational deviations from their policy preferences (Snyder 2001: 208). This is frequently because the existence of asymmetric territorial representation in Congress and the institutional requirement within Argentine democracy to achieve single party majorities provides the central government with an incentive to turn a blind eye to what often occurs at the subnational level.
Governors have considerable leeway to protect their boundaries. Before 2006, governors had the constitutional freedom to set the dates for provincial and municipal elections and to decide on their own set of electoral laws. Thus they had an institutional opportunity to manipulate electoral laws and electoral calendars in order to guarantee their own continued dominance (Calvo and Micozzi 2005). It is because of provincial-biases such as this one that the power of governors remains considerably stronger than that of mayors. De jure and de facto, they have both an elective incentive and the institutional ability to both capture and subordinate the local level. Beyond the evident limitations to local autonomy, it is the powerful majoritarian-bias of many provincial electoral systems which further limits the political autonomy of mayors and reduces the attractiveness of occupying this political position.
The low levels of electoral competition are eye-opening. Gibson and Suarez said of the late 1990s that “partisan realignments modified this two-party scheme somewhat” (2007: 20). At all events, it is worth noting that, for example, in the Province of Buenos Aires the ENP competing for provincial governor rose from a decade average of 3.06 to 4.26 in 2003 (Calvo and Escolar 2005:111), but at the municipal level for that same year they remained a low 1.9. Provincial partisan realignments did not affect the two-party scheme at the municipal level in most of the country. Using interparty competition as a gauge to indicate the desire of political leaders to want to fill local positions, it is evident that beyond the electoral authority that within Argentina's two majoritarian parties (the PJ and the UCR plus their allies) can lead to political career-making, a position of mayor in and of itself does not offer many inducements to ambitious career politicians, unless it is a key electoral district like in much of the Bonaerense. Moreover, unlike local elections in Brazil, which use open-list voting systems for legislative positions at all levels, Argentina's party-based nomination procedures for selecting electoral candidates provide few incentives for provincial and local elites to provide constituency services to bolster their personal popularity directly with citizens. The ability of Argentine voters to reward individual candidates is reduced by each party's candidate nomination procedures and closed-list electoral rules.
Only two national parties have had consistent access to centralized state resources, the UCR and the PJ thanks to their alternating monopoly of the National Executive. This has contributed to the complete failure of any new national “third” party to make inroads at the municipal level (Jones and Hwang 2005). Even more significantly, many studies have suggested that when the provincial and local levels are both dominated by the Peronist (PJ), the chances of municipal subordination are even higher, given the characteristics of Peronism and its dependence on localized networks (Auyero 2000; Calvo and Murillo 2004; Brusco, Nazareno, and Stokes 2004). Even though local government was assumed to be accountable by virtue of the fact that it was elected (Stoker 1988), during a period of economic boom, falling out of favor with the dominant party can equate to political alienation at both the provincial and municipal level. 9
Interview with Julio Pereyra, President of the Argentine Federation of Municipalities, Buenos Aires, 2006.
The constitutional ambiguity of municipalities, coupled with their limited political autonomy, has not begun a process of democratization at the local level that enables new actors to innovate and experiment with public policy, as was seen in post-1988 Brazil. 10 As municipal autonomy was never seriously considered during the 1994 constitutional reforms that granted autonomy, among other measures to the City of Buenos Aires, the political decentralization that has occurred in post-1983 Argentina has not increased the decision-making responsibility of municipalities regarding the provision of public services.
There are exceptions to this generalization, particularly in the major urban cities as the Autonomous City of Buenos Aires and Rosario under the administration of Mayor Hermes Binner (1995-2003).
The “kings” of public policy-making in Argentina remain governors (Braun and Aradanaz 2008), who are responsible for providing the majority of social public goods. This seriously affects the ability of the national government to federalize social programs and motivate municipalities to implement their policy initiatives in a uniform manner throughout the country. Because of the high levels of provincial autonomy in public-service provision, the implementation and quality of provincially-delivered social policy remains uneven. For example, while the province of Buenos Aires spent 39.3 percent of its 1999 budget on health, 32.3 percent on education and 13 percent on administration, the province of La Rioja spent 9.2 percent on health, 23.7 percent on education, and 40.2 percent on administration (Rezk 2000). This ability of the provincial authorities to set their own budget priorities and control the amount of revenue they transfer to municipalities within their territories has impeded any significant move towards municipalization and creates great policy heterogeneity at the provincial level. There exists no uniform mechanism to distribute financial resources to municipalities. This means that where municipal decentralization has occurred, it does not reflect a territorially even process throughout the federation. Clemente and Smulovitz confirm the limited role of municipalities in delivering services by saying:
While the average of resources per capita of national origin that the provinces receive is $827, the average of resources per capita that the municipalities receive from the provinces is $159.9. The difference between these two averages indicates that a significant part of the resources that the Nation transfers to the provinces is retained at a provincial level (2004: 113).
The national director of consolidated social expenditure in the Ministry of Economy further added that “tributary per capita is calculated by each province using their own variables.” He claims that, without the legal responsibility of provinces to pass municipal fiscal data to the federal government, “there exists no disaggregated expenditure for the municipal level in Argentina” 11 making it almost impossible to use either constitutional authority or fiscal expenditure at this level to measure policy decentralization. The provincial allocation of policy responsibility in the National Constitution and the current system of fiscal revenue transfers to the provinces, coupled with a weakly institutionalized mechanism to distribute from the provinces to municipalities, have impeded the feasibility of direct national-local collaboration and halted any meaningful move of decentralization towards the municipal level.
Information obtained from an interview with the Director of Social Expenditures in the National Ministry of Economy, September 19, 2006 [Buenos Aires]. All interviews cited from here forth were taken from my Ph.D. fieldwork from September to November 2006 in and around the City of Buenos Aires.
The decentralization of social services was also impeded in Argentina, as in Brazil from 1988 to 1995, because of macroeconomic instability. Social programs in both federal democracies could only be advanced in a permanent fashion when the central government could ensure stable economic policy (Alston et al. 2008). The decentralization of social services to the provinces without revenues was not however a direct cause of rising inflation and national macroeconomic instability as in Brazil, but was ironically its consequence. According to Faletti, the decentralizing administrative reforms undertaken at the beginning of 1992 can best be characterized by the central government offloading expenditures to the provinces without transferring resources (2003: 170). Inflation was eventually controlled, and although the fiscal pacts of 1992 and 1993 recentralized some revenues, these fiscal reforms did not have an overall negative effect on revenues available to the provincial governments. The provincial share of revenues actually increased, while their expenditures were held constant. As governor power re-emerged post-1995, so did their revenue transfers. The share of municipal expenditure however decreased relative to that of the provinces. The small piece of the pie they had during the 1990s effectively shifted back onto the central government (see Table 4).
Revenue Collection and Expenditure by Level of Government in Argentina post-Decentralization and post-Recentralization
Source: Willis, Garman, and Haggard (1999); Porto (2004).
Many explanations have been offered for how President Carlos Menem (1989-1999) managed to bring about fiscal recentralization in the early 1990s in a country where governors are kings. Jones (1997) asserts Menem's success was determined by a two-party dominant political system and high levels of party discipline. Gibson and Calvo posit that it was not just partisan control of Congress and the governorships that mattered, but the fact that many of them came from overrepresented peripheral provinces. They assert that political support for the reforms was achieved by postponing public sector unemployment cuts in peripheral provinces and increasing subsidies to some provincial coffers (Gibson and Calvo 2000). Beyond the subtle differences in these arguments, there is a general consensus in the literature that partisan factors and party-system characteristics facilitated Menem's ability to build a reform-supporting coalition that would pass administrative decentralization in 1991, fiscal reforms in 1992 and 1993, and constitutional reforms in 1994, all affecting the territorial distribution of power between levels of government.
However, these reforms were all weakly institutionalized because of the party-centric way in which they were negotiated. This has resulted in inconsistent implementation (many of the 1994 reforms were never implemented) and low enforcement (fiscal revenues began to flow again towards the provinces post-1995). The Federal Executive's intention in transferring expenditure responsibility to the provinces without revenues was to force them to reduce public-sector employment and use their un-earmarked transfers to pay for social services (Eaton and Dickovick 2004). In fact, it led to a provincial debt crisis that gave the Federal Executive an unintended opportunity to privatize provincial banks (a recentralization measure). Shortly thereafter, with Menem's decreasing popularity post-1995 in the lower chamber of Congress and a worsening fiscal situation in Argentina caused by the (now recognizable) flaws of the convertibility plan, governors successfully re-emerged as the key power-brokers in Argentina's federal game.
According to Tommasi and Spiller (IDB 2008: 100), provincial governors in Argentina have only a secondary interest in national public goods and in the quality of national policies; “their primary interest, on the basis of which they grant or withdraw support to national governments and their policies, is access to common-pool resources.” The strategy of provinces exchanging votes for cash did not enable the central government to ensure fiscal stability, or to ensure that available revenues were being used at the subnational level to provide badly needed social services. Moreover, the central government lacks a political incentive to regulate subnational finances using hard budget constraints or to enforce fiscal reforms on the subnational public sector. United PJ governors continued to extract further revenue increases from President Fernando de la Rúa from 1999 to 2000, despite the decreasing levels of centralized taxation and an evident national currency crisis. De la Rúa continued hopelessly to play this game in exchange for their political cooperation, to an eventual great cost.
A decentralized structure of federal government is coordinated by a given set of institutional mechanisms, which provide either incentives or deterrents to intergovernmental cooperation. Similarly to most federal countries, each level of government in Argentina wants to maximize its portion of revenue. In a system of intergovernmental fiscal relations where the provincial and municipal governments are completely separated from taxing and revenue decisions, an expansionary bias is built into subnational spending (Jones et al. 2002). Such fiscal relations work as a deterrent to greater interprovincial and vertical intergovernmental cooperation. In fact, “if you raise your capacity to collect taxes, you are less likely to benefit from future bailouts” (Tommasi, Sebastian, and Sanguinetti 2001: 14). The fiscal situation in Argentina is made worse by the near total discretion of the provincial governments over social policy and budget priorities, and the lame-duck status of most municipalities. Menem had believed in 1992 that transferring policy responsibilities without financing would force lower levels of government to “mop up” (Falleti 2003) their poorly used automatic transfers. It did not work: the intergovernmental fiscal scheme allowed both municipal and provincial governments to blame poor social policy performance on a lack of funds emanating from a higher level of government.
Within a context of soft budget constraints, where limited electoral incentives for high quality subnational policy outputs exist, and municipal autonomy is ambiguous, an incentive is created for provinces to compete with each other for revenues, which leads to the infamous federal “race-to-the-bottom” in local tax rates. According to Treisman, it resulted in Argentina (and federal Russia) “not just in a race-to-the-bottom, but a race to escape – or defraud – the top” (2004: 820). Within this context, political, fiscal and policy decentralization did not lead to greater subnational democratization or local policy innovation; neither could the national government build a direct collaborative relationship with municipalities in order to pursue the second-best option of municipalization. This does not mean, however, that it did not try.
For example in 2003, the central government under the leadership of President Néstor Kirchner (2003-2007) (PJ) launched a new national social program, Programa Familias, to alleviate poverty that was intended to bypass provincial intermediation. It was suppose to reduce the political manipulation of programs in this policy sector which occurs at that level. This new initiative would shut down the previous implemented emergency agenda called Program for Unemployed Households (PJJHD), a plan that was heavily discredited by the public and in the media for its manipulations. 12 The majoritarian nature of Argentina's political dynamics impeded the national policy goal of Familias in three ways.
Rebecca Weitz-Shapiro (2006) also argues that beyond partisanship social protest plays a role in the distribution of benefits.
First, mayors had no incentive to carry out this policy goal in collaboration with the national government as long as it was not supported by their provincial government, even though the provinces were not officially involved in the design of this program. Whether through action or inaction, Familias in 2006 had not been implemented in any municipality in five provinces (Neuquén, City of Buenos Aires, La Pampa, Catamarca, and Santiago del Estero). Following an established electoral logic, it does not make sense for local governments to participate in the implementation of a national program if they are not aligned, through partisanship, with the ruling national political elites. Additionally, in the context of these party-centered political dynamics with closed-list candidate nominations, mayors are generally more concerned with office-seeking rather than with policy-seeking.
Second, mayors cannot personally credit-claim for the successful implementation of Familias within their territory, both because of its rigid design, and because of high levels of voter identification in Argentina. Within the design of Familias, “municipalities just sit back and watch.” 13 This impedes their political willingness to cooperate with the program's implementation. Essentially, within Argentina's long established political dynamics and the importance of most voters’ partisan identity, the only layer of government that could credit-claim for Familias from 2003 to 2007 were the Kirchners (Nestór, President and his sister Alicia, National Minister of Social Development).
Information obtained from an interview with the ex-director of social development, Municipality of La Matanza. September 29, 2006 [San Justo].
Thirdly, Familias gained very little public attention and acknowledgment, because of the continued existence of ex-interim President Eduardo Duhalde's benchmark program, PJJHD. De Jure, PJJHD had been closed and was to have been actively replaced by the alternate federalized program Familias. In reality, the importance of PJJHD to provincial and municipal power-brokers for maintaining localized patronage networks, together with Kirchner's need of the support of as many provinces as he could muster, made it almost impossible for his new program to de facto replace PJJHD.
The intention of Familias to bypass provinces in its implementation goes against the logic of federalism in Argentina. Municipalities are creatures of the provinces, and as such, have little agency to behave without either, the approval or, the involvement, of the province to which they are institutionally subordinate. Provinces have not uniformly provided constitutional autonomy to their local units, and when they have, de jure, because of Argentina's provincially-organized majoritarian political dynamics, the de facto policy effects of municipal autonomy within provincial constitutions remains unclear. The lack of municipal autonomy coupled with the decreased political incentives of national-local collaboration impedes the delivery of social services at the municipal level.
Therefore in the case of Famílias, low levels of constitutional autonomy and decision-making authority at the municipal level prevented the central government from achieving policy success through direct national-local policy collaboration. If the province was not willing to play by the technocratic rules of the program set by the center (i.e., accept the candidates selected to implement and monitor the program at the municipal level), they simply blocked adhesion. Obviously, provinces are larger entities than a single local unit, hence the effect, of an entire province blocking and/or being blocked on the program's distribution was detrimental to the program's success. Essentially, it impeded the central government's ability to deliver welfare goods evenly across the nation in either a decentralized or centralized fashion. Furthermore, because of the ambiguous status of municipalities within the National Constitution, provincially-based political actors control the municipal level and provide resources for municipal leaders to engage in clientelism to maintain a solid base of supporters (Jones 2008). This characteristic politicizes the territorial distribution of collective goods in Argentina. A de facto two-level federal structure set within Argentina's provincially-based majoritarian political dynamics provides governors with the ability to constrain the power of mayors, and the power of the President, simultaneously. To the detriment of many citizens, it compromises the ability of the central government to federalize important social initiatives and to ensure their delivery at the municipal level, without provincial intermediation.
Of equal importance, municipalities from 2004 to 2007did not have a fiscal incentive to carry out Familias. Municipalities remain dependent on the provinces for their revenues, which emanate from the secondary distribution of the coparticipación (a tax-sharing agreement between the federal and provincial governments). Furthermore, their borrowing capacity is determined ad-hoc by provincial legislatures which makes the soft-budget syndrome all the more salient between the two subnational levels. Municipalities are neither dependent on the central government for resources, nor is their social expenditure hard-wired to either the provincial or national budget, as is the case in Brazil. Governors have considerable influence over the execution of public policy and of its financing. Therefore, municipalities that are formally aligned with the provincial incumbent are delivered resources and public policy projects to fuel local patronage and clientelism from the province downwards. The small economic inputs that come from a national social program designed to alleviate poverty do not provide sufficient fiscal incentives to entice a municipality to give up the potential fiscal benefits it gains by directly cooperating with particularized provincial interests – even in areas of health, education, and social protection policy. The threat of fiscal punishment from a strong governor to a municipality that attempts to bypass the province's political interests and directly collaborate with the central government is large. 14 It equates to quite simply “fiscal suicide.” 15
Information confirmed in interviews with various government officials from the Municipalities of La Matanza, Rosario, the Argentine Federation of Municipalities and the National Congress. Interviews took place in and around the City of Buenos Aires in September 2006.
Information obtained from interviews with officials in the Argentine Federation of Municipalities [November 2, 2006] and with Juan Pablo Cafiero, ex-National Minister of Social Development [October 15, 2006].
As a means-tested program, Familias has not performed very well either in terms of popularity or territorial coverage. Recently attributed public perceptions of poverty alleviation programs in Argentina tend to support this interpretation. Based on available public opinion surveys only two out of ten Argentines surveyed felt that current programs were good, and 88 percent of respondents thought the plans were used for political motives (CEDLAS 2007). These opinions are in stark contrast to the public's evaluation of Bolsa Família in Brazil – especially given the two programs were similarly designed and implemented, and originally received the technical and financial backing of external organizations such as the World Bank and the Inter-American Development Bank (IDB). Their performance is far better explained by their varying political, institutional, and structural variables.
The goal of unraveling this story about non-contributory social protection policies in Argentina has been to show that the discontinuation of heavily discredited emergency workfare programs such as PJJHD and the institutionalization of new national-local poverty alleviation initiatives such as Famílias were impeded by governors, who can constrain the policy goals of the central government by their ability to control mayors located within their territory. The credibility of federalism in Argentina is therefore compromised by a lack of robustness at the municipal level, which, I suggest, fortifies de facto the strength of the provinces. There exists a dichotomy between disconnectedness and subordination at the provincial and municipal level (Medina and Ratto 2005), because of the ability of provinces to threaten municipalities with punishment if they do not cooperate with key provincial stakeholders. This not only reduces the incentives of local elites to provide constituency services to their communities, but it also reduces the incentives of mayors to cooperate in the implementation of national policy goals that may alienate them from forging much politically and fiscally needed provincial allies.
Conclusions
In this paper, I have explored the role that municipalities can play in facilitating the central government's ability to carry out its desired social policy objectives in two strong federal systems. Through direct national-local policy collaboration, municipalities can enable a more even territorial distribution of collective goods. However, the ability of municipalities to make this contribution depends on the varying institutional configuration of each federal system and the incentives it provides.
By using the policy area of non-contributory social protection as a springboard for analyzing how varying institutional configurations affected the development and implementation of a uniform non-contributory social safety net, it became evident, that the unique institutional configuration of Brazil's federal system enabled successful national-local policy collaboration. According to the common wisdom of comparative politics, the key political institutions required for high quality policy performance are existent in neighboring Argentina. The central government here, ought to be able to overcome the ability of provincial-based actors to constrain its actions by using a party-centered coordinative mechanism. In practice however, the problem of using a partisan-based mechanism to ensure the kind of intergovernmental cooperation required to uniformly deliver national collective goods is that it leads to a zero-sum game of winner-take-all. Moreover in Argentina, the game and its winners operate from the provincial level, which further jeopardizes the national public good and the quality of democracy at all levels.
This paper therefore also calls into question the usage of a party-based mechanism to ensure intergovernmental cooperation within a decentralized federal context. In federal countries like Brazil, or even Canada, where majoritarian political dynamics are absent, the unevenness of the territorial distribution of social goods and moreover democracy, is not the product of an enforcement problem like in Argentina or Mexico, but is rather the product of a transmission problem. Where the ability to deliver public goods like health, education, and social protection is linked to a transmission problem, a strategy of multi-level governance which is based on municipalization may be an appropriate solution. An interesting area of future research would be to further explore the potential feedback effects of this strategy.
In federal countries like Mexico and Argentina where majoritarian political dynamics prevail, this is currently politically, fiscally, and administratively unfeasible. This is not to conclude that such systems cannot resolve their governance challenges, but rather it suggests that their central governments will have to give a lot of power away before it can start claiming it back. Giving power away and then slowly taking it back represents a process of simultaneous decentralization and recentralization that can lead to political and institutional change, change that does not often take place because of a crisis but rather reflects a slower evolutionary process. Critical junctures, in the form of political, social and economic crises, have not sufficed to bring about institutional change in Latin America's federations, no matter how tragic they have been. This research posits that in these countries electoral and constitutional reform will ultimately be required to bring both municipalities and voters back into the political process.