October 5, 2008

Education Funding in a Recentralizing Democracy

By Sands, Joseph C Jr

ABSTRACT The administration of President Fernando Henrique Cardoso enacted measures to constrain subnational politicians in a newly democratized and highly decentralized federal system. Lacking sufficient accountability at the subnational level, the central government attempted to increase its control of educational funding and minimize local discretion over educational spending. These reforms constrained the distribution of intergovernmental transfers, but entrusted the disbursement of educational spending to local oversight. This article argues that while the constraints protected the federal treasury from predatory practices, the local oversight did not protect educational spending from mayoral discretion. This argument is based on an analysis of initial reform implementation in four municipal school systems. The mayors responded to the federal initiatives in a variety of ways, but these were based on the requirements of their own political survival. The four cases thereby become studies in how one effort to recentralize democracy is working in Brazil.

In Latin America, recent experiments with decentralized democracy and participatory practices have provided opportunities for citizens throughout the region to elect local officials and oversee public spending. Some of these new institutional arrangements have given scholars the opportunity to explain "audacious reforms" and even "quiet revolutions" (Grindle 2000; Campbell 2003). In Brazil, the 1988 Constitution established one of the most decentralized democracies in the world (Souza 1997). The charter granted autonomy to municipalities and guaranteed them access to financial resources. It also mandated the participation of civil society on oversight councils at every level of the new federal system. In this case, however, the new institutional arrangements have left scholars with the task of explaining "predatory federalism" and "democratic deadlock" (Abrucio 1994; Ames 2001).

Within a decade of the promulgation of the constitution, the administration of President Fernando Henrique Cardoso (1995-2002) enacted centralizing reform measures designed to minimize the predatory strategies of subnational politicians and increase the democratic vitality of local communities (Bresser Pereira 1999; Eaton 2004; Montero 2000; Samuels 2000). One target of reform was public primary education, where more than 50 million students were enrolled in school systems financed and administered almost entirely by state and municipal governments (INEP 2001). This unprecedented and unsolicited intervention by the central government into the preserve of autonomous subnational politicians came to be hailed as "the most successful achievement of the [Cardoso administration] in the social policy domain" (Draibe 2004, 405).

The magnitude of the education reform was great, and its impact has been widespread. An amendment to the constitution assigned municipal governments the responsibility to provide primary education. A new national funding mechanism distributed resources more equally, and a new national education law required higher minimum qualifications for all school teachers. States, municipalities, and schools created councils to control expenditures. In selected school systems, teachers and parents participated in programs supported by the World Bank to learn how to manage school resources. The Ministry of Education encouraged new pedagogical strategies designed to keep students in school. Nationwide, student enrollment expanded, drop-out and repetition rates declined, and teacher remuneration increased.

While the reform measures did produce positive outcomes, their implementation in the more than 5,500 municipalities in Brazil pitted the designs of the federal government against the autonomy of local politicians. In the absence of sufficient financial and electoral accountability at the subnational level, the federal government attempted both to increase the central control of educational funding and to minimize the local discretion over educational spending. To achieve these goals, the reforms subjected the distribution of intergovernmental transfers to a new automatic control mechanism, but entrusted the disbursement of educational spending to the vicissitudes of local oversight.

This study argues that while the centralized control over funding successfully protected the federal treasury from the predatory practices of subnational politicians, the local oversight of spending did not protect educational spending from mayoral discretion. Based on the empirical evidence from four cases of reform implementation in the period 1997-2001, this study argues that the combination of centralized control over funding and local oversight of spending did not overcome the decentralized autonomy that subnational politicians had reserved for themselves in the 1988 Constitution. Thus, modifications in the municipal provision of primary education that occurred subsequent to the enactment of the reforms came at the discretion of the mayor. While mayors responded to the federal initiatives in a variety of ways, they did so according to the constraints of local conditions. In particular, they chose to contest or co-opt, ignore or embrace the reforms according to their own immediate needs for political survival.

Researchers have paid increasing attention to municipal politics in Brazil, especially as cities have adopted strategies, such as participatory budgeting, to overcome institutionalized roadblocks to the more effective management of public resources. Participatory budgeting, a process introduced by and closely associated with the Workers' Party, allows city officials and community organizations to negotiate, implement, and enforce budgetary priorities, usually with an emphasis on expanding urban infrastructure (see, among others, Abers 2000; Avritzer 2002; Baiocchi 2003; Koonings 2004; Nylen 2003; Wampler 2007). The analysis here moves the study of municipal politics forward by going beyond novel attempts by a single political party to elicit the participation of organized civil society through unique institutional channels. It analyzes the unsolicited initiatives of the federal government that required public school communities in every municipality, regardless of its governing political party, to oversee new spending requirements that often conflicted with the established routines of municipal bureaucracies, bureaucracies recognized for their connection with partisan politics, their practice of ineffective management, and their antagonism toward the user public (Valla 1994; Bresser Pereira 1999; Da Matta 1995). The story of municipal politics told here is neither one of isolated success nor one of widespread failure. It is a story of, at times, surprising variation among the most recent local skirmishes, provoked when federal reforms elicited the participation of local communities in the longstanding struggle between centralization and decentralization, between national policymakers and subnational elites in Brazil (Hagopian 1996).

The empirical evidence in this study is based on research conducted in the municipal school systems of four cities: Sao Luis, Natal, Goiania, and Campo Grande. These cities rank among the 25 most populous in Brazil, and they also number among the 23 metropolitan areas selected by the Ministry of Education to participate in the initial round of Fundescola, one of the various programs aimed at enhancing participation. Each of these cases represents one of four possible categories related to the context for reform implementation. Throughout the country, municipalities either lost or gained revenue through the new control on funding, and municipal school systems either had or did not have prior experience with community participation in school management (see table 1).

Among the four cases, the municipal governments in Sao Luis and Natal both lost revenue to the new controls on intergovernmental transfers, while those of Goiania and Campo Grande gained revenue. In the early 1980s, the school systems in Goiania and Natal initiated community participation in school management, while those in Sao Luis and Campo Grande did not.

According to this framework, it could be expected that the city of Goiania would provide the most positive context for reform impiementation, since it benefited from an increase in revenue and the school system possessed ten years' experience in participatory management. The case of Campo Grande provides an intermediate context, because it also benefited from increased revenues, though it had no prior experience with participatory management. Likewise, the city of Natal represents an intermediate context. It lost revenue, but it did have ten years of participatory management. Sao Luis, a city that suffered a loss of revenue and had no experience with community participation in school management, represents the least positive context for reform implementation.

Of course, a comparative study at the municipal level of analysis calls for caution. For one thing, a detailed study of municipal politics necessarily demands a restriction in the number of cases. The implementation strategies highlighted here are limited to cases of urban school systems located in three of the five administrative regions of the country. A municipal level of analysis also demands caution about the confidence level of government data on, for example, municipal budgets, department spending, employment conditions, and even classroom space. More than one administrator whispered to the author that "in Brazil all budgets are fictitious." Over the course of 18 months in 1999-2000, the archival research was supplemented with 100 openended interviews of federal, state, and municipal administrators; principals; teachers; parents; students; and community leaders. The research also involved attending various conferences for educational professionals, workshops for council members, and union meetings (Sands 2004). Given the limited number of cases and the limited level of confidence in available data, the findings presented here are only an initial, tentative attempt to understand the impact of these federal educational reforms on municipal political arrangements. Future research will be able to analyze the longer-range impact of these centralizing reforms on the discretionary use of public resources by municipal politicians.

The article first examines the factors that allowed Brazilian decentralization to sustain predatory strategies and produce democratic deadlock at the municipal level. It describes how the design of the reforms of public primary education put forth by the Cardoso administration attempted to correct these conditions. Local reactions, implementations, and outcomes are analyzed on the basis of evidence from the four cases. Finally, policy alternatives for the next round of both political and educational reform in Brazil are suggested.


By establishing autonomous municipalities that both had access to guaranteed financial resources and provided spaces for the expression and incorporation of local demands, the subnational politicians who composed the Constituent Assembly (1987-88) seemed to embrace decentralized strategies that would provide local officials with both the mechanisms to minimize clientelist political pressures and the means to improve the delivery of local public services (Bradford 1994; Borja 1993; Hipsher 1998; World Bank 1997). Yet while the constitution guaranteed resources, in the form of intergovernmental transfers, to municipal governments, it refrained from assigning these governments the responsibility for the provision of any public service, with the single exception of regulating intra-urban transportation. By withholding greater responsibilities from municipal officials, the constitution allowed municipal governments to become, in the words of Timothy Power, arenas of "institutional permissiveness" characterized by the absence of "any and all barriers to the personal latitude of individual politicians" (Power 2000, 4). This lack of barriers produced negative consequences for both the federal treasury and municipal school systems.

Guaranteed access to financial resources, municipal officials often refrained from enforcing or even enacting a property tax. By the time Cardoso took office, almost 90 percent of all municipalities relied on intergovernmental transfers to provide them with 95 percent of their revenue (Mendes 2001). Some state-level elites, moreover, availed themselves of a strategy that entitled them to an increased share of the federal treasury: they created more municipalities. The state of Maranhao, under the leadership of Governor Roseana Sarney of the Partido Frente Liberal (PFL) (1997- 2000), created 95 new municipalities simply by dividing already existing ones. All these new municipalities, however humble, received intergovernmental transfers. During the first ten years of the democratic regime, state governments created more than 1,300 municipalities. This explosion in the number of subnational jurisdictions left the federal government with an unsustainable fiscal burden. By 1995, it was distributing almost 25 percent of its revenues to subnational governments through intergovernmental transfers (IBGE 2002).

How did subnational politicians make use of these financial resources? Just as intended, according to David Samuels's analysis. The subnational politicians who wrote the constitution created intergovernmental transfers to increase their own capacity to reward their supporters and, in, so doing, sustain their electoral ambitions (Samuels 2003). These ambitions were regulated partly by an open-list, proportional representation electoral system. An open- list ballot allows electors to cast their votes for individual candidates. Proportional representation in the legislative body ensures that political parties are represented according to the proportion of votes received. This system, in use for legislative elections since the 1940s, was preserved by congressional vote shortly after the inauguration of a civilian president in 1985 (Ames 2001).

In Brazil, this electoral system and its accompanying rules serve to disaggregate the demands of the electorate and minimize the participatory capacity of citizens at the most basic level of democratic participation: local elections (Mainwaring 1999). The rules designate each municipality as a single electoral district, allow hundreds of candidates to seek election to dozens of city council seats, make voting obligatory, and restrict voters to casting a ballot for only one candidate. Under these rules, candidates win by receiving merely a thin slice of the vote (Ames 2001).

While the vast majority of voters cast their single vote for a losing candidate, a small minority of voters are able to elect the entire membership of a city council. In the 1996 elections, among the four cases in this study, more than 1,000 candidates competed for fewer than 100 city council seats. The highest vote getter received just 2.5 percent of the vote. One-third of all the winning candidates received less than 1 percent of the votes cast. The 21- member city council of the city of Sao Luis was elected with just 16 percent of the vote. Eighty-four percent of the electorate had cast their ballots for losing candidates (TSE 1996).

The lack of both financial and electoral accountability in Brazil has led scholars to describe the nation's governance "not as a tyranny of taxation without representation, but as a farce of representation without taxation" (Mendes and Gall 2001). The arrangement allows candidates to neglect even to promise collective goods, such as primary education. They have little incentive to promise collective goods because they win votes in exchange for particular goods, such as a place on the city payroll, a space in a school classroom, a hot meal, or a pair of sandals (Desposato 2001; Geddes 1994). Under these conditions, inclusion in the political community is bestowed not through casting a vote but by trading it (DaMatta 1995). In the context of scarcity, trading votes for access to immediate benefits reduces the voter's "uncertainty of survival ... to a minimally tolerable level" (Avelino Filho 1994). Since most voters do not pay municipal taxes, they have little interest and less incentive to hold municipal administrations accountable for the provision of collective goods or long-term remedies (Dagnino 2003).

The lack of financial and electoral accountability increases the cost that mayors incur to maintain electoral allies and ensure legislative cooperation. Once elected, mayors govern by trading particular, immediate resources for legislative support (Santos Junior 2001). The cost of securing this support increases in proportion to the fragmentation of the local political party system. The mayor of Sao Luis, for example, could count only two copartisans on the 21-memher city council. In Goiania, 33 city council seats were distributed among 15 political parties. The volatility of even these fragmented partisan alignments provokes an ever-increasing demand for access to patronage and other immediate goods (Rezende 1998). In such a context, as electoral competition increases, so does the need to secure access to more immediate, individual goods, such as patronage. One consequence of this increasing need is the deterioration in the provision of collective goods, such as primary education (Ross 2001; Scott 1972).

In the 1970s, public primary education in Brazil was recognized for its excellence (Draibe 2002). Over the following two decades, services deteriorated as demand increased. By the mid-1990s, 25 percent of school-aged children were not even enrolled in school (World Bank 1997). In 1994, UNICEF reported that only 39 percent of Brazilian children completed the fifth grade (Gadotti 1997).

In 1996, the Cardoso administration attempted to reorder intergovernmental relationships and reinforce mechanisms of local accountability. The administration sought to diminish the drain on the federal treasury and limit the permissiveness of subnational politicians. In particular, increased demand for a more educated workforce and and increased incentives from international financial institutions prompted the Cardoso administration to attempt a reform of public primary education.


The Cardoso administration introduced various reforms of primary education. The foundational reform was the new Basic Law of National Education of 1996 (.Lei de Diretrizes e Bases da Educacao Nacional, LDB), which assigned responsibility for the delivery of primary education to municipal governments. (For a study of how the LDB was passed, see Rosar 1995.) The law also established new minimum academic qualifications for primary school teachers. By 2007, all teachers of fifth- to eighth-graders would be required to have earned an undergraduate degree. The Ministry of Education promoted various additional reforms, such as accelerated class instruction, which were designed to diminish the student repetition and drop-out rates. The administration also launched reform measures to control educational financing and encourage local oversight of spending, on which this analysis focuses. Through these measures, the government in Brasilia wanted to restrict the amount of funding that the central government provided to subnational politicians, yet it also wanted these politicians to increase the amount of money spent on primary education. These measures were the first federal constraints on the distribution and expenditure of the constitutionally guaranteed intergovernmental transfers. A constitutional provision that required 25 percent of all receipts to be spent on education had generally been ignored for almost a decade.

Controls on Financing

To better control educational financing, the Cardoso administration created the Fund for the Maintenance and Development of Primary Education and the Valorization of the Teaching Profession {Fundo de Manutencao e Desenvolvimento da Educacao Fundamental e de Valorizacao do Magisterid), known by its Portuguese acronym, FUNDEF, and henceforth as the Primary Education Fund. This fund achieved two purposes: it created hardly any additional expense for the federal treasury, and it rationally distributed resources on a per student basis.

Funding for the Primary Education Fund came from four existing intergovernmental transfers: Impostos sobre Circulacao de Mercadoria e Servicos, ICMS; Fundo de Participacao dos Estados, FPE; Fundo de Participacao dos Municipios, FPM; and Impostos sobre Produtos Industrializados, IPI. Fifteen percent of the value of these four transfers was withheld from state and municipal governments and deposited into a Primary Education Fund. More accurately, there were 27 Primary Education Funds: one for each of the 26 states plus one for the autonomous Federal District. Money from each fund was then distributed to the state government and all municipal governments in that state on a per student basis. Although all the receipts from the fund were to be spent on primary education, 60 percent was to be spent exclusively on teacher compensation. While the fund did not mandate salary increases, it did require that all municipal governments update the salary scales and promotion guidelines of public school teachers. Each municipality was also required to establish a council to oversee these expenditures.

Incentives for Participatory Oversight

Other reform measures also created opportunities for new educational funding and new oversight of educational spending. The Money Directly to the School Program (Programa Dinheiro Direto na Escola) and the School Development Plan (Piano de Desenvolvimento da Escola) also distributed monies on a per student basis. Funds from these sources were deposited directly into an account controlled by a school oversight council.

The Money Directly to the School Program received its funding from the Ministry's National Fund for the Development of Education (Fundo Nacional de Densenvolvimento da Educacao), which distributed all discretionary federal funds designated for primary education throughout the country. This program earmarked funds to buy pedagogical materials and cover the cost of small repairs. To qualify for these funds, each school was required to create a school oversight council composed of teachers and parents. The ministry deposited the funds directly into an account of this council. The funding thus bypassed all levels of the municipal bureaucracy: the mayor, the Department of Finance, the Department of Education, and even the school principal.

The School Development Plan was one aspect of a more extensive reform program called the Fund for Developing the School (Fundescola). This plan cost US$1.3 billion, half of which was financed by a loan from the World Bank. It coordinated and financed strategic planning in individual schools in 23 metropolitan areas located in 3 of the country's 5 administrative regions: North, Northeast, and Center-West (FNDE 2008). During a school year, professional educators hired by Funclescola, an autonomous subdivision of the Ministry of Education, accompanied each participating school through a planning process. This process consisted of regular meetings throughout the school year. The various school constituencies met by cohorts. In their respective groups they decided on managerial and pedagogical goals, created a set of objectives to meet the goals, and designed strategies to achieve the objectives. They then divided the strategies into assigned tasks. Representatives of all of the groups then met to adopt their plan and a timetable for implementation.

Only when a participating school had agreed on a plan was it eligible for direct funding from Fundescola. Again, the amount of funding depended solely on enrollment. Fundescola deposited the funds in an account established especially for the school. The school community designated a group of participants in the planning process to report to Fundescola on how the funds were spent.


How did mayors respond to modifications in the level of guaranteed, unearmarked funding? How did they respond to the school community's expanded role in spending oversight?

The Primary Education Fund produced an immediate impact on municipal governments throughout the country. The automatic withholding of transfers and their centralized distribution according to a new criterion modified the receipts of every municipality. The distribution from the fund resulted in revenue decreases for some municipalities and increases for others, since the equalizing per student formula regulated the distribution of an already existing, stable, and fixed amount of funding. For municipalities that depended on intergovernmental transfers as practically their only source of revenue, the withholding of 15 percent of anticipated transfers represented an immediate loss of almost 15 percent of total anticipated revenue. Among these municipalities, those that enrolled relatively few primary school students received a relatively small amount of revenue from the fund. Thus, those municipalities that both relied heavily on transfers and enrolled a relatively small number of students lost revenue to the Primary Education Fund. In the fund's first two years, approximately 40 percent of municipalities fell into this category (MFC 2002). Those municipalities that did not depend so heavily on intergovernmental transfers and enrolled a relatively large number of students received a relatively large amount of revenue from the fund. When this amount was greater than the amount withheld, the fund produced an increase in municipal revenue.

The modification in municipal receipts provoked various responses from the municipal officials in the four cases in this study. In Sao Luis, when the city lost revenue, the mayor chose to make equivalent cuts in the budget of the municipal Department of Education. In both Sao Luis and Natal, the diminishment of revenue provoked each mayor to increase the number of enrolled students. Municipal school systems across the country also reported increases in enrollment. However, there were very few controls to ensure that these reports were accurate, and even fewer to ensure that newly enrolled students had desks to sit in, teachers to instruct them, or even schools to attend. As the cases will demonstrate, thousands of students were squeezed into shorter instruction times, or not assigned to any school at all.

In municipalities that gained revenue, it was the mayor who decided how to spend the windfall. An increase in revenue did not obligate the mayor to increase education spending. The mayor of Campo Grande, for example, built new schools to accommodate new students; while in Goiania, the mayor kept the school system budget at existing levels, and the Primary Education Fund served to give him greater discretionary use of the city's unearmarked revenue.

Although local governments were required to spend the earmarked revenue on primary education, the enabling legislation did not explicitly itemize what qualified as such an expenditure. The mechanism established to monitor spending was the local Primary Education Fund oversight council, whose members were appointed by the mayor. At times, citizens did question the legality of certain expenditures. In the first years of the fund, the Ministry of Education and the federal police investigated hundreds of mayors for allegedly diverting funds to cover expenses in areas other than primary education (Folba de Sao Paulo 2000a, b).

The only clear stipulation on spending was that a minimum of 60 percent of the fund was to be spent on teacher compensation. This provision was designed to encourage local officials to increase teacher salaries. In all the cases in this study, which were relatively large urban school systems, the payroll costs of the school system exceeded the entire amount of revenue provided by the Primary Education Fund. In not one of these cases did salaries increase, even when teachers went on strike to demand it. In some municipalities, the mayor abided by the 60 percent requirement by hiring more primary school teachers, though not necessarily by following constitutionally mandated civil service hiring practices. Thus mayors met the fund's guidelines and increased patronage opportunities at the same time.

While the Primary Education Fund deposited money directly into the hands of the mayor, other new funding initiatives deposited money directly into accounts of oversight councils. The response to these initiatives varied according to the existing relationship between the mayor and the school system. One important factor in this relationship was democratic management reforms. These reforms had been introduced in various cities in the 1980s by leaders of the then-democratic opposition, who looked on public schools as a privileged site for participatory democracy (Dourado and Costa 1998; Mello 1995). The reforms introduced elections to select principals and members of school councils. In these school systems, the elections broke the tradition of selecting principals through patronage appointments. Greater autonomy from mayoral intervention in selecting school principals allowed educational professionals- teachers-to assume a greater leadership role in public school administration. In some cases, the reform measures included direct funding of schools to cover such expenses as maintenance and supplies (Amaral Sobrinho et al. 1994; Saviani 1999). These reforms of the 1980s modified the relationship between the school system and the mayor to a limited degree. Although the reforms diminished the direct connection between school management and patronage politics, even with the minimal direct financing, few school systems possessed any financial autonomy within the municipal bureaucracy.

Ten years after democratic leaders trusted in community participation as an empowering strategy for resisting authoritarian policies, the Cardoso administration, backed by the prevailing wisdom of the development community, trusted in community participation as an instrumental strategy for improving educational indicators (Borja 1993). This new approach immediately encountered resistance from teachers in the two school systems in this study where democratic management practices were in place.

In Natal, the teachers' union campaigned against the mayor's implementation of this federal intervention in school management. It viewed the measures as top-down, outside interference that would provide additional funding to a savvy partisan enemy, the current mayor. The union accused the mayor of wanting to end the local experience of democratic management. It also expressed skepticism not only about the capacity of the new oversight mechanisms to monitor expenditures, but also about the federal government's promise actually to provide additional funding. In Goiania, the mayor, who benefited from millions of dollars of additional revenue from the Primary Education Fund, demonstrated no interest in the additional sources of funding. Without mayoral interest, school leaders expressed minimal interest in incorporating the new initiatives into existing management practices.

In the two school systems with no experience in community participation, the new initiatives received a creative welcome by the respective mayors. In Sao Luis, the mayor introduced municipal legislation that minimized the oversight capacity of school councils. For example, principals were designated to preside over council proceedings even though their performance was to be supervised by the council. In Campo Grande, the mayor enthusiastically embraced the federal measures and invited school leaders to assumer a greater oversight role in school management. In this case, the schoolteachers and parents throughout the school system refused his invitation. Their refusal was based not on their desire to defend local democratic management practices but on their perceived need to protect themselves from anticipated recriminations from school principals.

As the cases demonstrate, the institutional permissiveness that characterized the subnational political relationships proved to be a greater restraint on initial reactions and outcomes at the municipal level than the constraints and incentives of the reform measures. The federal reforms placed greater control over the distribution of educational funding, but they failed to enact sufficient constraints on educational spending. Without greater external controls on spending, the mayor could implement the new federal mandates within the discretionary parameters afforded him by the Constitution of 1988.


In Sao Luis, the conditions for implementation of the reforms were precarious. The city lost revenue to the Primary Education Fund and, absent a participatory management legacy, the user public was ill-prepared to hold elected officials and school principals accountable even for those funds that were deposited directly to the school.

The case of Sao Luis demonstrates the limits of federal mandates and the expansiveness of local permissiveness. In this city, the delivery of primary education was, among the four cases, the most intimately connected to partisan politics. Historically, the sole criterion for the hiring of principals was their capacity to deliver votes. The implementation of the reforms by Mayor Jackson Lago of the Partido Democratico Trabalhista (PDT) was marked by a dramatic increase in enrollment, accompanied by a dramatic decrease in spending, a combination that proved prejudicial to the enrolled students. Although the city allowed the establishment of school councils, it regulated this structure to such an extent that principals remained accountable to the mayor alone. The reform implementation did, however, help to produce a positive outcome for Mayor Lago. The mayor managed to use the mandates of the Primary Education Fund to expand not only the municipal payroll but also his capacity to reward political allies. In 2000, he was re-elected.

Primary Education Fund

The city government of Sao Luis was a loser in terms of the Primary Education Fund. Since federal and state transfers accounted for 86 percent of municipal receipts, the fund withheld from the city a relatively large amount of anticipated revenue (Sao Luis 1996- 2000). Since the municipal school system enrolled only a relatively small number of students, the fund returned to the city only a fraction of its withholding. The mayor responded by increasing student enrollment by more than 50 percent over the course of just two years while reducing the budget (Sao Luis 1997-99). In the second year of the reforms, the Department of Education's per capita spending amounted to less than 60 percent of that of the previous year (Budget Director 2000).

How were newly enrolled students accommodated as the budget dramatically decreased? When the school system increased enrollment by more than 17,000 students in the course of a single school year, the local press investigated the whereabouts of these new students. It reported that the municipal school system had enrolled thousands of children but had "forgotten" to find space for them. "Without spaces and sufficient structures to meet the exacerbated demand of students . . . the alternative was to place them in any 'hole' where they might be able to study, even amid an environment of precarious conditions" (O Imparcial 1998). These holes were "community schools," which the Secretary of Education described as "having no relationship with any community and having nothing to do with education" (Soares Pflueger 1999).

Community schools, established by neighborhood organizations in cities throughout Brazil in the years following World War II, spared municipal governments the expense of accommodating in their own school systems a rapidly expanding urban population. By the 1990s, however, most community schools in the city of Sao Luis had become little more than patronage-based child care programs for school- aged children. Over several decades, neighborhood groups had come to depend more and more on the resources of elected officials, members of the city council, or the state assembly to sustain community schools. These politicians, in turn, used the schools to reward electoral clients with teaching positions, enrollment slots, or even access to maintenance supplies. Community schools, moreover, were not accredited by the state government. Lacking the permission to award diplomas, the schools merely provided attendance certificates. Consequently, their connection both to local communities and to any educational project was precarious (Alberquerque 2000).

At the start of the 1998 school year, the Department of Education in Sao Luis annexed 122 of city's 400 comunity schools. This annexation increased both the number of students and the number of teachers in the municipal school system. The enhanced enrollment increased the educational funding the city received. The expansion of the city payroll, however, connected the school system more closely to the practice of electoral clientelism. The 1,839 newly hired community school teachers were contracted as temporary employees with no claims to either employment benefits or career promotion (Sao Luis 1999b). Their inclusion on the city payroll provided the mayor the opportunity not only to expand his patronage base, but also to meet the federal requirement that 60 percent of the Primary Education Fund be spent on teacher remuneration.

In implementing the reforms, Mayor Lago did not initiate any measures to update the teachers' pay schedules. All salaries in the municipal school system were held constant, even though they remained at only 60 percent of the region's average salary for teachers (MEC 2002; Sao Luis 1985). The local teachers' union, however, did not mobilize for improved employment conditions during the Lago administration.

Participatory Oversight

The municipal school system had no practice of citizen oversight of public resources before the introduction of the federal reforms. In Sao Luis, the mayor routinely selected school principals not for their affinity for education but for their connection to political party leaders. The accepted practice was that the mayor would distribute the appointment of principals among city council members according to capacity of each member to deliver votes to his electoral victory. Although, when elected, Mayor Lago specifically proposed to introduce the election of school principals, he never acted on the proposal (Sao Luis 1997). After two years in office, he replaced his secretary of education, who had introduced various initiatives, including the selection of school principals based on merit (Pires Costa 2000). The school system did respond positively to the various incentives of the Ministry of Education to create such avenues for community participation and oversight. In particular, it participated in both Money to Schools and the School Development Plan. School councils were established and given juridical status; they received funding from the Ministry of Education. Under the constraints of the enabling municipal legislation, however, the councils were structured to prevent parents from effectively reviewing the expenditures of the direct funding. Parents were confined to a minority of seats on the board, and only the school principal could preside over the school council (Sao Luis 1999a). According to one administrator in the Department of Education, the city council had reasonably decided that since the principal was an appointee who had the confidence of the mayor, no other individual could suitably review municipal expenditures (Director of Direct Deposit Program 2000). The School Development Plan was introduced very slowly; and after two years, fewer than half of the 62 eligible schools in the system participated in the program (Fundescola 2002).


Although this city also lost revenue to the Primary Education Fund, the case differs from that of Sao Luis in that the school system in Natal had been under democratic management for ten years. Mayor Wilma Faria (Partido Socialista Brasileira, PSB), one of the highest vote getters in the state of Rio Grande do Norte, was on her way to becoming governor in 2002. Faria had developed her electoral base when she served as the state secretary of labor and social welfare during the administration of her nephew, Governor Jose Agripino Maia (1982-86). From this position, she created and co- opted community councils in various neighborhoods on the periphery of Natal through her distribution of federal funds supplied by the then-authoritarian regime (Andrade 1987).

In this case, an organized teaching faculty, the front-line implementers of policy, did not support the reform implementation, which they perceived as a threat, an imposition, and a ruse through which the mayor and her political allies would be the primary financial beneficiaries. The experience of democratic management served to galvanize a capable faculty against the implementation of the initiatives inspired by international financial institutions. The union had to accept, with little preparation and less willingness, increases in student enrollment and adaptations of pedagogical practices. The union's resistance led to a work stoppage, an unsuccessful demand for a salary increase, and an equally unsuccessful demand to stop the expansion of the School Development Plan and its new management practices.

Primary Education Fund

In the city of Natal, the Primary Education Fund produced a decline in municipal receipts. In 1996, the intergovernmental transfers accounted for 78 percent of these receipts. Consequently, the fund withheld a large portion of municipal revenue. Because the school system was small, the fund disbursed a relatively small amount to the city (Rio Grande do Norte 1996-2000). To increase its receipts from the fund, the city, over the course of Mayor Faria's term, increased enrollment by 20 percent.

Although the city implemented various strategies to accommodate this increase, it could not find a space for every student. The school system created partnerships with a handful of private schools. This allowed the city to claim the students as its own while each of the schools retained its autonomy, especially in regard to faculty hiring. The school system did provide financial resources to the schools, including the money to pay teacher salaries. The school system also converted space in four rented buildings into classrooms. In a third strategy, 19 of 60 schools in the system began to operate a fourth shift of classroom instruction. These schools squeezed in an "intermediate" shift, from 10:50 A.M. to 2:30 P.M. (Natal 2000). In 2000, despite these adjustments to both space and time, one thousand enrolled students of the municipal school system had no school to call their own (.Diario do Norte 2000).

Although the Primary Education Fund was designed to improve teachers' working conditions, in Natal it resulted in extra demands placed on them. Schoolteachers had to face not only increased enrollment but also the demands of new pedagogical strategies, such as acceleration and cycles. In 2000, the teachers of the municipal school system, having received no salary increase in the previous six years, sought a 45 percent increase. According to the Ministry of Education, between 1996 and 1998, the salaries of teachers employed in municipal school systems in the Northeast region had increased, on average, by 50 percent. When Mayor Faria refused to increase salaries at all, the teachers went on strike. After 41 days without entering the classroom, the teachers finally returned when the union called off the strike without having received any of its demands.

Participatory Oversight

Amid a contentious relationship with the mayor, the union viewed the implementation of the School Development Plan as a threat not only to its ten-year experience with democratic management but also to the job security of its members. The union responded to this threat by attacking the design, the methodology, and the anticipated results of the program.

First, the union alleged that the program's designers, whom the union identified incorrectly as "the IMF," were interested not in forming competent local oversight but in creating self-reliant service users. The design thus betrayed the true agenda of the program: to lay the foundation for the eventual privatization of public schools. The union further argued that the plan's methodology, "focusing on the student," would be measured solely according to quantitative indicators of academic achievement and student promotion rates. The union was dedicated not to improving test scores but to educating capable citizens. Moreover, the "student-centered" program made no acknowledgment of the broader societal factors that affected student performance. Consequently, the teachers feared that the program's less-than-optimal results would be explained by and blamed on the incompetence of teachers. The union also alleged that the mayor was interested in the School Development Plan only for the money it would bring to the school system. The union understood the implementation of the program as a political struggle over anticipated resources-resources that the mayor undoubtedly would channel away from public schools (SINTE-RN 2000).

In 1999, a year after the School Development Plan was introduced in 11 schools, the city's secretary of education unilaterally modified the scheduled seventh biannual elections for principal and school council members. She cancelled elections in one-fourth of the system's schools; those newly established, those undergoing construction, and those participating in the School Development Plan. It would be particularly unfair to those 11 schools, she argued, to risk a change of leadership in the midst of this multiyear planning process.

The teachers' union, however, argued in favor of proceeding with the scheduled elections. For the union, the elections would serve not only as a way to maintain the continuity of the democratic management process, but also as a referendum on the School Development Plan itself. The education secretary ignored the union's position and conducted elections in only three-fourths of the schools. She then used the results of the elections to question the nature of the schools' democratic management. The evidence: in 75 percent of the schools, only one candidate had run for the office of principal. In two-thirds of the schools, the incumbent was re- elected.

At the conclusion of the 2000 academic year, the secretary of education invited the remaining 45 schools in the system to participate in the School Development Plan. The union lost its last battle to prevent this program from expanding when all schools chose to accept the invitation (Fundescola 2002).


The conditions for implementation in Goiania seemed to suggest that the reform measures would benefit the municipal school system. First, the city gained revenue from the Primary Education Fund. Second, the school system had its ten years of experience in democratic management, which provided schools with relative autonomy from mayoral interference. Moreover, among the four mayors of this study, only Nion Albernaz, of the Partido Social Democratica Brasileira (PSDB), did not harbor electoral ambitions. He had already decided that at the end of his term he would retire from public life. Reform implementation, however, produced minimal changes in the administration and financing of the school system. Despite revenue gains, Albernaz chose not to increase the budget of the Department of Education; and despite incentives for increased participation, department administrators displayed little enthusiasm for either the Money Direct to Schools Program or the School Development Plan.

Primary Education Fund

The city of Goiania was a winner with the Primary Education Fund. Since intergovernmental transfers constituted only 60 percent of municipal revenues, the fund's withholding from the city was relatively small; and because the municipal school system enrolled a relatively large number of students, the fund's disbursement to the city exceeded the withholding. In 1999, Goiania's municipal government recorded an increase of approximately US$2.25 million (Goiania Dept. of Education 2000b; Office of Budget 2000). In 2000, the increase was estimated at US$4.25 million. Yet despite the increase in municipal receipts, the mayor, over the course of his term, neither increased the budget of the Department of Education, expanded enrollment, nor increased teacher salaries (Goiania Dept. of Education 2000a). In 2000, the teachers conducted a 30-day work stoppage in the second month of the school year. The union held both a vigil in front of City Hall and a sit-in at the mayor's office. When the city agreed to bargain with the union, the teachers went back to the classroom. But after a month of negotiations that gained no concessions from the mayor, the union simply called off the strike.

Participatory Oversight

With its own experience of democratic management and its own efforts at administrative and pedagogical reform already in place, the school system did not have much interest in the new participatory strategies offered by the Ministry of Education. Democratic management in the municipal school system in Goiania had already put control of the schools in the hands of the teachers. This corporatist command was achieved partly through the "index of proportionality" that governed elections for school principals. According to this index, the ballots of teachers were weighted such that their collective value equaled the collective value of all the ballots cast by the other eligible voters, all parents and all students above the age of 12 (Goiania Dept. of Education 1999, 57- 59).

The teachers who controlled the schools did not oppose the school system's participation in either the Money to Schools Program or the School Development Plan. All schools were eligible to participate in the Money for Schools program, since they all had existing school councils that could serve as executors of the direct funding. The mere existence of these councils, however, did not ensure oversight of expenditures. One school system administrator acknowledged the difficulty of creating a culture in which the school councils actually monitored the activity of the school principal. Most school councils, in this administrator's estimation, existed "just for show" (Director of School Councils 2000).

The School Development Plan was introduced in the school system at the end of the 1999 school year. By the end of the following year, three-fourths of the system's one hundred schools were participating (Fundescola 2002). In Goiania, the overall response to the program was tepid, as administrators and teachers found its strict discipline very demanding. Most administrators and teachers were happy to receive the accompanying funding but uninterested in changing day-to-day operations in the schools. Many principals who started with the program were "simply waiting for [it] to fail and come to an end" (Director of Planning 2000).


In Campo Grande, the local conditions for implementation were mixed. The city gained revenue from the Primary Education Fund, yet the school system had no experience with participatory oversight. Mayor Andre Puccinelli, of the Partido Movimento Democratico Brasileiro (PMDB), appointed all the school principals. School councils did not exist. During Puccinelli's term in office, his party, due to a shift in statewide political allegiances, lost the governorship to the Workers' Party (Partido dos Trabalhadores, PT). This loss was a gain for the mayor, however, as it catapulted him to the position of his party's most powerful executive in the state. He would win re-election in 2000 and move on to the governor's mansion in 2004.

The case of Campo Grande highlights the positive role that local discretion played in the implementation of the federal reforms. Both the mayor's political strength and his direct relationship with the school system allowed him wide discretion in this process. He increased department funding, constructed schools, and established school councils. He did not increase salaries, but did use educational funds to pay the salaries of temporary teachers. He vigorously supported the School Development Plan. Campo Grande's experience emphasizes that electoral certainty and increased funding can bring about improvements in the provision of public primary education.

Primary Education Fund

For Campo Grande, the Primary Education Fund resulted in a net gain for the city treasury, for the municipal school system, and for the enrolled students. The city's dependence on federal and state transfers represented only 57 percent of municipal receipts, the lowest percentage among the four cases. Enrollment in the municipal school system was relatively high. Consequently, the city received a large disbursement from the fund. Mayor Puccinelli used this increase in revenue to increase the budget of the municipal school system. This is the only case in this study in which per student spending actually increased, even as enrollment increased by 28 percent between 1998 and 2000 (Campo Grande Dept. of Education 1999a). The city also provided space for thousands of new students by opening 16 new schools, 13 in newly constructed school buildings and 3 in rented facilities.

With the expanded enrollment, the school system also sought to improve its efficiency. In 1998, it initiated a program of acceleration in order to diminish the high repetition rate. In the first three years, almost 4,000 first- through fourth-graders benefited from the program. When the program expanded to include fifth- through seventh-graders, more than 15,000 students joined its ranks (Campo Grande Dept. of Education 1999b). The school system recruited teachers for the acceleration program who were "receptive to changes and new challenges, committed and enthusiastic about teaching, and who believed in the acceleration project" (Campo Grande Dept. of Education 1998).

The school system's search for committed and enthusiastic teachers became more difficult over time, however, as the teachers' union recognized that the mayor's policies did not include improving their employment conditions. The mayor refused to increase teachers' pay scales; the salary paid by the municipal school system, on average, only reached 65 percent of that earned by teachers in the Center-West region (MEC 2002; Campo Grande 1999). In 1998, 400 teachers followed the decision of their union to strike. They demanded a 10.8 percent pay increase. In response, the mayor immediately fired 30 probationary teachers who had joined the strike. He also threatened to dock the pay of all teachers who did not report to work. The strike ended quickly without any mayoral concession (Portela 2000).

As enrollment and classroom space expanded, no civil service exams were given to fill teaching spaces in the municipal school system. In 2000, the year of his re-election campaign, the mayor hired 555 new teachers; but of these new hires, only 13 percent were given contracts. Eighty-seven percent were hired as temporary employees (Campo Grande Department of Education 2000).

The city partnered with the Federal University of Mato Grosso do Sul to establish a program to allow public school teachers to earn an undergraduate degree. More than one hundred teachers enrolled, but they were not pleased that classes in the four-year program were conducted only during summer vacations and winter holidays. Surrounded by discontent, the director of the university program tried three times to resign from her post. She described the schoolteachers as "irritable, tired, and spoiled" students who came to class to torment their instructors and who had no interest in finishing or even pursuing their course of study (Osorio 2000).

The teachers were particularly annoyed by the one modification the mayor easily obtained from the city council regarding employment conditions in the school system: an annual performance evaluation. The municipal ordinance described in detail the professional expectations and the evaluation procedures. According to the coordinator of the process, the evaluation was the first attempt by any government agency in the entire country to undertake a performance evaluation of its employees. Due both to the administrators' lack of experience and the teachers' depth of resentment, however, the first annual teacher evaluations were universally proclaimed "a fiasco" (Director of Evaluations 2000).

Participatory Oversight

In Campo Grande, the mayor appointed all the school principals. In support of professional competence, a new municipal law required that to be appointed principal a teacher had to have completed a course in public management that the municipality offered annually. The mayor also met with assembled groups of principals at least twice a semester to review education policy.

Yet Mayor Puccinelli was also anxious to have principals held accountable by the school community. He relied on three strategies to elicit this accountability. The system participated in the Money to Schools Program; by 2000, all 75 schools in the system were participating in the School Development Plan (Fundescola 2002); and the city's own evaluation process called on the presidents of the school councils to become more involved in an oversight role, specifically assisting in the school system's evaluation of principals. The school system created three new positions from which administrators would oversee the execution of each strategy across the entire system.

The school system established an office to provide administrative support for the oversight of funds received from the Money to Schools Program. To ensure that councils did not exist only on paper, administrators maintained regular contact with the council president of each school, who was required to be either a parent or a teacher.

The office overseeing the School Development Plan dealt directly with school principals. There, administrators confessed that they "had to demand that [principals] allow participation in their schools" (Director of School Development Plan 2000). Some principals enthusiastically embraced the demands of the School Development Plan. For one principal, the program initiated a new way of understanding her role as leader of the school. In her words, "Every business has a mission, but businesses also have clients. Here the clients are the parents and the students. In this school we thought our mission was to teach, period. But our mission is also to help the community. The school belongs to the community. I am just passing through. It is not my school. It is our school. The School Development Plan has helped me to see this" (Vieira Zaflon 2000). The school system's Office of Evaluations was charged with the task of conducting evaluations of school principals. The law specifically stipulated that a three-person committee, composed of an administrator, a teacher at the school, and the president of the school council, evaluate each principal. At an assembly of school council presidents held on October 25, 1999, the coordinator of evaluations informed the presidents of their upcoming role as evaluators. Despite her coaxing and pleading, however, one by one the council presidents spoke against assuming such a task. Parents expressed fears that the principal might retaliate against a negative evaluation by prejudicing the education of their children. Teachers, who feared that they were the next targets, expressed their categorical opposition to any municipal performance evaluations. In the face of this opposition, the department postponed the evaluations.


With the reform of public primary education, Brazil's federal government had sought to reduce the permissiveness of subnational elites in a newly democratized regime. At the municipal level, the federal government found existing financial and electoral constraints insufficient to motivate mayors to invest in public primary education. The reforms in this policy sector included the first attempt by the federal government to control the distribution of a percentage of guaranteed intergovernmental transfers. The reforms also sought the participation of local school communities in the oversight of educational spending. It proved easier to establish a central control on the distribution of funds, however, than to establish local constraints on spending. The reforms served to protect the federal treasury, but were less successful in protecting the municipal provision of public primary education from mayoral discretion.

Mayoral discretion was more vulnerable to immediate electoral pressures than to the constraints and incentives of the federal reforms. Alternative strategies that would place greater limits on mayoral discretion, especially in the provision of collective goods, include political and electoral reform. Although mayors are elected in majoritarian elections, they govern with a city council that is elected, in accordance with existing electoral rules, by a small fraction of the electorate. Measures that establish smaller electoral districts and closed voting lists, reduce the number of political parties, and enforce party fidelity could increase the capacity of local voters to hold local elected officials accountable for the provision of public goods, and also could reduce the demand on mayors to provide immediate access to individual goods. Diminished partisan volatility between the executive and the legislature could also reduce the demand for the mayor to provide immediate access to individual goods. Another policy alternative, which is even less likely to be implemented, is to enforce and increase municipal property taxes. If more people pay more taxes, they might take a greater interest in the quality of the public services that money provides.

Regarding education reform, the enforcement of existing laws that require all government employees to be hired on the basis of performance on civil service exams would prevent mayors from appointing temporary employees to fill teaching positions. Also, as the case of Campo Grande demonstrates, the appointment and training of school principals needs specific constraints. Not only did the mayor of Campo Grande direct leadership training workshops, but the staff of the School Development Plan focused largely on helping principals expand their management strategies. Thus, the case of Campo Grande also shows that when the mayor's electoral survival is not threatened, the discretionary relationship with the school system can lead to positive outcomes. In this city, the combination of low electoral uncertainty and increased educational funding served to bring improved service to students.

In future iterations of any direct deposit program, executors would fortify the capacity for oversight if the program were expanded beyond the exclusive focus on the school's internal dynamics to include the cooperation of members of the wider community. A network of associates would prevent the burdens of spending oversight and performance evaluation from falling solely on teachers and parents. Campo Grande also illustrates how these stakeholders often feel too vulnerable to possible reprisals. Under the credible auspices of Fundescola, interested electors and taxpayers, representatives of nongovernmental or professional organizations, or members of the business community could be invited to serve on school oversight councils and evaluation committees.


The struggle over centralization and decentralization is a longstanding battle in Brazil. The institutional permissiveness of the decentralized political system established by the 1988 Constitution is but one expression of that struggle. In an attempt to reduce the permissiveness of subnational elites, the federal government, among other measures, placed controls on intergovernmental transfers and promoted oversight by school communities. These reforms of public primary education produced positive outcomes, but they also revealed the weakness of the central government to restrict mayoral discretion regarding educational spending. Any modification in the municipal provision in public primary education that occurred after the enactment of the reforms came as a result of the discretion of the mayor.

Variation in the degree of reform implementation and the level of its impact across the four munici