Recent scholarship on Latin America has focused on the contribution to economic development of state bureaucracies, operating independently of narrow interests. Most agree that bureaucratic independence is necessary for the implementation of any state-sponsored development strategy. In spite of their importance, we know little about what fosters the creation and institutionalisation of such bureaucracies.
The search for answers has led several to re-examine the remarkable experience of Brazil during the presidency of Juscelino Kubitschek (1956-61).(1) One of the country's most successful development plans, the Targets Plan was implemented during this period. It consisted of 41 production objectives in six broad categories: energy, transportation, foodstuffs, basic industries, education, and the construction of a new capital, Brasilia. A few highly professional and independent bureaucratic agencies formulated and implemented these projects. The most important of these agencies was the Banco Nacional de Desenvolvimento Economico (BNDE), which provided the technical expertise and financial resources for this ambitious investment programme.
Two recent studies have credited the Targets Plan's success to Kubitschek's efforts to protect the BNDE and other government agencies from encroachment by political and economic elites.(2) These studies seek to explain temporal variation in the degree of insulation enjoyed by such bureaucratic agencies by reference to changes in presidential behaviour. According to one account, the benefit of presidential support for bureaucratic independence was felt only in the short run. The opposing view contends that presidential action in support of bureaucratic independence had a lasting impact on the process of institution-building within the Brazilian state.
Closer examination of the BNDE during the 1950s suggests a different approach to factors fostering agency independence within a politicised bureaucracy. The independence of the BNDE was not determined by presidential preference alone. Presidential influence has been exaggerated to the exclusion of other influences. This analysis shows how innovative bureaucrats undertook specific measures to maintain the Bank's independence and to ensure its institutionalisation, even in the context of shifting and uncertain support from the chief executive.(3) Variation in the degree of presidential support affected the financial capacity of the BNDE and, at times, limited the scope of its lending. But it had little impact on its independence, as measured by two other criteria: specific loan decisions and personnel recruitment. Thus, despite fluctuations in presidential support, the BNDE's independence from political and class interests became well established. This independence made it one of the most important governmental institutions involved in promoting Brazilian development during the 1950s and beyond.
To make this argument, the present article draws from internal BNDE documents, minutes from meetings of Bank directors, oral histories, and interviews. I conducted open-ended interviews with current and former staff members and administrators who served in the BNDE during the period under study. In addition, I have used interviews conducted by Bank personnel as part of an oral history project, Projeto Memorial
Sources of Bureaucratic Independence
To be `independent', a bureaucratic agency must be both `autonomous' and `insulated'. States are autonomous when they pursue goals that do not simply mirror the interests of elite groups or classes in society.(4) Autonomous bureaucrats can pursue policies that conflict with, or fail to support, the interests of such groups. Bureaucrats are insulated when they enjoy freedom from partisan or patronage pressures.(5) This is particularly important in countries lacking a merit-based civil service, where clientelism turns state offices into sinecures and diverts public resources towards private uses. Such a patronage-based system was a pervasive feature of the Brazilian federal government during the 1950 and 1960s. It represented a major impediment to the implementation of economic policy.
Recent literature on development has established the vital linkage between bureaucratic independence and the capacity of the state to carry out development policies.(6) Without independence, bureaucratic agencies are unlikely to possess the capacity to implement policy. But we know relatively little about the conditions under which it is created. What are the organisational characteristics of states that make it possible for officials within them to make policy choices without external interference?(7) The degree of independence, to be sure, can vary within any governmental system over time.(8) Moreover, only certain agencies within the government will enjoy high levels of independence. This is particularly true when a state has been effectively penetrated by patron-client networks or interest group pressures.(9)
The organisational characteristics of the Brazilian state in the 1950s did not appear propitious for the creation of autonomous development agencies. Patron-client networks pervaded the bureaucracy and the state was highly fragmented. Earlier attempts by the dictator Getulio Vargas to professionalise the civil service in the 1930s and 1940s had failed.(10) The return to democratic rule in 1945 had greatly weakened the power of the central government to make and implement policy. These organisational weaknesses were aggravated by a series of political crises that began with the contested election of Getulio Vargas to the presidency in 1950, and did not end until the military coup in 1964. Such crises threatened the legitimacy of civilian presidents and interfered with their ability to carry out consistent development policies.
Despite these inauspicious circumstances, a few agencies achieved a high level of autonomy and insulation in this period. One author has appropriately described these as `pockets of efficiency' within the otherwise politicised Brazilian state.(11) The BNDE was the most important and influential pocket of efficiency to emerge at this time. It played a particularly important role in Kubitschek's Targets Plan. All accounts of the period credit the Bank with providing the technical expertise and institutional capacity required to formulate and implement this ambitious investment programme which provided the foundation for the country's industrial development.(12) These accounts also attribute the impressive capacity of the Bank to the high level of insulation and autonomy it achieved within Kubitschek's government.
Recent explanations for the remarkable autonomy and insulation that characterised the Bank's action during the Kubitschek period have emphasised the explanatory importance of presidential behaviour. This offered a needed corrective to earlier approaches that stressed the role of structural features of the economy, with little regard for the role of human agency in promoting development. A focus on presidential behaviour as an explanatory variable makes sense in the context of Brazilian politics during the 1950s. Presidents wielded enormous influence over the bureaucracy. Although presidents had many power resources at their command, three proved particularly critical in shaping bureaucratic independence. First, Brazilian presidents had the power to appoint and dismiss the heads of government ministries and agencies. There existed some formal limits to this power, but presidents always reserved the right to appoint all the top officers within the government. Second, presidents possessed an array of legal instruments and prerogatives that permitted them to intervene directly in the administration of an agency. For example, through the power of the decree law, they could alter the ministerial location of an agency or redefine its functions. Ministerial location was important because agencies associated with powerful ministries were accorded greater independence than those in weak ministries.(13) Presidents could thus affect the funding available to agencies dependent on budgetary allocations. Third, they could completely undermine the financial viability of an agency by diverting its resources, slowing the distribution of budgeted funds, failing to collect relevant revenues, or declining to support the agency's bid for increased budgets from the legislature.
Presidential control over these important resources has led many to conclude that bureaucratic insulation and autonomy could not be achieved without active presidential support and protection. According to Geddes, the independence attained by the Bank varied with its presidential support.(14) In her account, Kubitschek actively protected and promoted the independence of the BNDE from external pressures. The lack of such support greatly reduced the Bank's independence under Getulio Vargas, she contends. Although Vargas established the BNDE during his administration, he politicised the Bank by appointing a political crony as head director and diverting legislatively mandated resources toward political ends. By contrast, Kubitschek cultivated the Bank's independence by introducing merit examinations for personnel recruitment, expanding its funding resources, and giving its directors primary responsibility for implementing the Targets Plan. Geddes concludes that this independence endured only as long as Kubitschek's presidency, and disappeared with his successor, Joao Goulart.
Kathryn Sikkink also emphasises the decisive role played by presidents in ensuring the insulation and autonomy of the BNDE during the 1950s.(15) However, she does not share Geddes's view that the degree of bureaucratic independence changed in rapid response to presidential preference. Institution-building occurred despite fluctuating presidential support, she contends. Sikkink credits Vargas with establishing the Bank in the first place. She agrees with Geddes's view that Kubitschek did more to promote its independence, however. She attributes his support for the Bank to two factors. First, both managers and tecnicos shared the main assumptions and goals of Kubitschek's developmentalist philosophy.(16) Second, he chose to make the Bank the primary institutional vehicle for carrying out the main component of his development policy, the Targets Plan, because it had already acquired the necessary expertise. Sikkink points out that without this prior development of independent, technically expert agencies, the Kubitschek government would not have had the institutional capacity to implement the Targets Plan. With his backing, the Bank was able to maximise its institutional capacity.
Both authors have added a great deal to our understanding of this important period in Brazil's development. But the BNDE's evolution as a pocket of efficiency cannot be explained by focusing exclusively on presidential behaviour. Empirical evidence shows that the levels of independence and institution-building attained by the BNDE were largely attributable to the strategies employed by innovative bureaucrats occupying leading positions in the Bank's administration. These bureaucrats were aided in their efforts by the existence of two public investment programmes, a favourable pattern of elite recruitment into the state bureaucracy, a sense of shared mission among bureaucrats, and the relative weakness of other political and economic elites.
The analysis that follows examines the actions of these civil servants during the nine years that followed the BNDE's creation. Bank independence is gauged by its control over three activities: personnel recruitment, loan allocation, and generation of funds. The recruitment of personnel is a particularly good indicator of the insulation of a government agency, since patronage hiring poses the most significant challenge to this form of independence. Decisions about the recipients and conditions of loans provide a reliable measure of both the autonomy and insulation of a state-owned financial institution. The degree of control a state agency exercises over the sources of its funding generally measures how autonomously it operates from other political actors.
The Bank's experience shows that in personnel recruitment and loan allocation, Bank bureaucrats proved highly effective at protecting themselves from outside encroachment. Despite proposing innovative strategies for establishing independent sources of funding, these bureaucrats were less successful in maintaining control over the generation of funding resources. This lack of control did have an adverse impact on the independence and institutional development of the Bank. More specifically, the failure to create an independent source of funding limited the scope of BNDE lending and made it more vulnerable to the vicissitudes of Brazilian politics. However, the high degree of independence achieved in the areas of personnel recruitment and loan allocation helped to counteract the negative effects of continued funding problems and permitted the Bank to emerge as an independent institution within the Brazilian economic bureaucracy.
The CMBEU and the Creation of the BNDE
The first major attempts to build a modern state in Brazil occurred during the period of the Estado Novo (1937-1945). Using dictatorial powers, Getulio Vargas established many new government agencies based on a model of modern administrative values such as rational decision making and merit in personnel recruitment.(17) Vargas aimed to create a part of the state that would be under his control and not influenced by the patronage practices that permeated most of the bureaucracy. This great experiment in modern administration failed as newly created agencies such as the Departamento Administrativo do Servico Publico (DASP) reverted to traditional practices following Vargas's removal from power.(18) Aside from a small number of state-owned enterprises operating at both the national and state level, there existed no pockets of efficiency capable of carrying out a national development programme.
Vargas was given a new opportunity to pursue his professed goal of improving the efficiency and rationality of governmental intervention in economic development after his election to the presidency in 1950. By the 1950s the earlier so-called `unconscious' phase of industrial development had ended, and it was thought that the next phase of industrial growth would require more conscious direction and support from government agencies. Of particular concern were `infrastructural bottlenecks' in the transportation and electric energy sectors which threatened to curtail further industrialisation. The elimination of infrastructural bottlenecks required the Brazilian government to undertake large-scale investments in railroads, roads, ports, and dams. It was within this economic context that the Banco Nacional de Desenvolvimento Economico was created.
The BNDE was an outgrowth of the work of the Comissao Mista Brasil-Estados Unidos de Desenvolvimento Economico (CMBEU) which operated in Brazil between 1951 and 1953.(19) This joint effort was the third, and most important, of three study missions to visit Brazil since the early 1940s.(20) Unlike the previous missions, the CMBEU made a significant and lasting contribution to fostering industrial development. Following a close study of the economy, the CMBEU proposed the implementation of over 40 infrastructural projects which later became the basis of Vargas's National Plan for Economic Reconstruction. This plan became known as the Lafer Plan after it was announced to the public by Vargas's Minister of Finance, Horacio Lafer. The CMBEU estimated that these investments would cost about eight billion dollars. The World Bank and Export-Import Bank offered, in principle, to cover 500 million dollars of foreign currency costs of these projects. As a condition for releasing these …