I

Do democracies enjoy an advantage when negotiating financial assistance from international financial institutions (IFIs)? To date, Latin American and Caribbean countries have collectively received a total of 5,103 loans from the World Bank (WB), the world’s leading development institution which provides developing nations loans of critical importance. Yet little is known about how and why the type of regime in the borrowing nation affects WB lending patterns. This question is particularly relevant for Latin American countries, which have been governed by both democratic and dictatorial regimes that have employed a range of political models and economic policies. According to former WB executives, while WB staff may have initially held personal preferences on the matter, their work with developing nations convinced them of the importance of assessing each regime on a case-by-case basis. Gerald Alter, in charge of WB–Latin America relations in the 1960s and 1970s, stated that working with unstable countries like Argentina, Brazil and Peru taught Bank staff that ‘the old saw that it takes a dictatorship to resolve [economic] problems just turned out to be completely false’.Footnote ¹ Furthermore, Robert McNamara, the WB’s president from 1968 to 1981, explained the WB’s assistance to dictatorial regimes in Nigeria, Korea and China as follows: ‘I’m not arguing [that] governance isn’t important. What I’m arguing is … [that] there is no clear correlation in my mind between economic social advance on the one hand and democracy on the other’.Footnote ² For McNamara, development could occur under various regimes. Therefore, displaying a neutral and pragmatic stance, the WB was to fulfil its mission regardless of regime type. This declared neutrality, however, enabled the Bank to support illegitimate regimes, often displaying indifference to the social costs of regressive economic measures and systematic human rights violations.

A handful of historical studies have analysed the stance that the WB and/or the IMF adopted toward democracies and dictatorships in Argentina (Kedar Reference KEDAR2018a; García-Heras Reference GARCIA-HERAS2018a, Reference GARCIA-HERAS2018b), Brazil (Altamura and Kedar Reference ALTAMURA and KEDAR2021), Chile (Kofas Reference KOFAS2002; Kedar Reference KEDAR2017a), Colombia (Kofas Reference KOFAS2002) and Peru (Brands Reference BRANDS2007, Reference BRANDS2010). Yet most of the literature on WB lending to different regime types has been produced by social scientists such as Banjak (Reference BANJAK2010); Carnegie and Samii (Reference CARNEGIE and SAMII2019); Houghton (Reference HOUGHTON2019); Roberto (Reference ROBERTO2021); Teichman (Reference TEICHMAN2004); and Winters (Reference WINTERS2010), to name just a few. This is surprising as the assistance that the international financial community granted to dictatorships in Latin America is a recurrent theme in historical accounts of those regimes (Veigel Reference VEIGEL2009; Harmer Reference HARMER2013; Verbitsky and Bohoslavsky Reference VERBITSKY and BOHOSLAVSKY2016). No less striking, as detailed in the following section, economists offer contradicting interpretations regarding the alleged preference of certain IFIs in lending to different political regimes. Informed by these theoretical debates, this article asks whether all democracies and all dictatorships in Latin America received equal treatment from the WB. Drawing on never-before-consulted documents from the WB archives and additional primary sources, it examines the Bank’s lending to Latin America from its first loan to the region in 1948 until 1988, when most nations were once again under democratic rule. In 1988, the US-promoted Brady Plan replaced the ‘Program of Sustained Growth’ or Baker Plan, which since 1985 had enabled an unprecedented boom of WB lending to the region. The article focuses on the WB’s four largest borrowers in Latin America – Brazil, Mexico, Colombia and Argentina. Particular attention is given to Brazil, the most populous country and largest economy in Latin America, and the region’s heaviest borrower from the WB. Like Argentina, but unlike Colombia – which remained democratic – and Mexico – which became democratic only in the 1990s after decades of civilian-authoritarian rule – Brazil experienced both democratic and dictatorial regimes in the second half of the twentieth century.

The article advances three interrelated arguments. First, regime type played a marginal role in WB lending decisions – it granted and denied loans to dictatorial and democratic regimes alike. Adhering to its self-imposed principle of ‘political neutrality’ and striving to expand its operations in developing nations, the WB consistently prioritised purportedly impartial economic criteria over political considerations. However, maintaining political neutrality proved particularly difficult during the Cold War, when a country’s real or perceived non-alignment with US Cold War interests posed obstacles to securing WB loans. The WB’s refusal to lend to João Goulart’s administration (1961–4), which many in Washington erroneously portrayed as communist, illustrates these difficulties. Second, the terms ‘democracy’ and ‘dictatorship’ are, by themselves, insufficient for evaluating WB lending. As known, there is an ongoing debate about polity characterisation. By prioritising different variables, scholars offer various categorisations. For instance, the Democracy-Dictatorship (DD) dataset (Cheibub et al. Reference CHEIBUB, GANDHI and VREELAND2010) categorises democracies as parliamentary, mixed or presidential, and dictatorships as monarchic, military or civilian. Given that democracy entails more than the sole presence of elections, the Varieties of Democracy (V-Dem) dataset identifies five types of democracy: electoral, liberal, participatory, deliberative and egalitarian (Coppedge, Edgell, Knutsen and Lindberg Reference COPPEDGE, EDGELL, KNUTSEN and LINDBERG2022). The Polity5 Project, for its part, points to a spectrum of governing authority that spans from fully institutionalised autocracies through mixed, or incoherent, authority regimes to fully institutionalised democracies. The ‘Polity Score’ captures this spectrum on a 21-point scale ranging from -10 (hereditary monarchy) to +10 (consolidated democracy).Footnote ³ The Democracy Index highlights Brazil’s political changes in the period discussed here. In 1949, when Brazil received its first WB loan, its democracy index was 5, rising to 6 between 1958 and 1960. Following the military coup of 1964, the democracy index plummeted to -9, improving to -4 in 1975 and -3 in 1983. With the transition to democracy in 1985, the index rose to 7 and has remained at 8 since 1988.Footnote ⁴

Notwithstanding the undeniable importance of the above-mentioned categorisations, this article argues that the WB exhibited a clear tendency toward pragmatism and ‘political indifference’. As WB documents reveal, its management and staff showed no interest in the distinctions between various democratic administrations and between the different phases of the 21-year-long military dictatorship in Brazil. This political indifference should not surprise us. During the Cold War, with few exceptions – such as the Carter administration (1977–81) – the US government was ready to support any regime, democratic or dictatorial, as long as it was anti-communist. The WB, for its part, systematically refrained from discussing political issues. Instead, the Bank focused on evaluating a regime’s economic policies, its level of political and economic stability, and, to a lesser extent, its Cold War alignment, particularly its stance toward Cuba, its ties with the Soviet Union, and the promotion of economic measures like expropriations that many in Washington perceived as a key Cold War concern. This fact inevitably leads to the third argument – the Cold War mattered. In effect, the WB’s indifference toward the regime type was pivotal in enabling it to continue lending to countries like Brazil and influencing its economic policies and development plans, thereby ensuring that the country remained aligned with the capitalist principles of the Western bloc.

The article is divided into four sections. Section II surveys the scholarly debate regarding the democratic advantage. Sections III and IV examine WB lending to its four largest borrowers in Latin America, particularly to Brazil, and Section V offers several conclusions.

II

Scholars are divided as to whether international lenders prefer to engage with authoritarian or democratic regimes. Some posit the existence of a ‘democratic advantage’ while others claim that a ‘democratic disadvantage’ is evident. Regarding the former, Schultz and Weingast (Reference SCHULTZ and WEINGAST2003) argue that democratic governments – being constrained and held accountable by stakeholders – are better positioned than authoritarian regimes to make credible commitments to repaying their debts. Hence, potential creditors, perceiving democracies as more creditworthy, are more inclined to lend to them than to nondemocratic regimes. The WB’s lending patterns to Latin America, however, do not indicate that it viewed democratic regimes as more likely to repay their debts. Moreover, the notion that democratic sovereign borrowers can be held accountable by lenders is mainly applicable to developed countries wherein sovereign borrowing often involves domestic financial institutions. Most Latin American borrowers, by contrast, rely on foreign lenders as well as on foreign credit rating agencies whose evaluations inform lenders’ decisions. According to Biglaiser and Staats (Reference BIGLAISER and STAATS2012), these agencies take regime type into account when assessing a country’s creditworthiness. In this regard, Hansen (Reference HANSEN2023) distinguishes between democracies with high debt levels and low foreign reserves – common in Latin America – which experience worse credit outcomes, and democracies with low economic vulnerability that experience better outcomes due to their liberal democratic credentials.

Other scholars, like Przeworski (Reference PRZEWORSKI1991) and Simmons (Reference SIMMONS1997), propose a ‘democratic default’ explanation, arguing that the adjustment programmes requested by the IMF are politically costlier for democracies than for authoritarian regimes. However, this argument does not fit the Latin American reality in which authoritarian leaders, such as Castelo Branco in Brazil, Augusto Pinochet in Chile, and Jorge Rafael Videla in Argentina, prioritised economic stability as a means to retain power. In this regard, DiGiuseppe and Shea (Reference DIGIUSEPPE and SHEA2015) claim that sovereign credit affects leadership survival in nondemocratic regimes more than in democratic regimes.

In his analysis of borrowers in developing nations, Saiegh (Reference SAIEGH2005) advocates a ‘democratic disadvantage’ approach. He argues that because democratic governments are more prone to default or reschedule foreign debt compared to non-democratic regimes, they receive fewer loans at higher interest rates. Like Saiegh, Kim and O’Neill (Reference KIM and O’NEILL2017) posit that democratic regimes are more inclined to default than dictatorships. However, unlike Saiegh, they assert that IFIs have been more willing to reschedule debts of democratic regimes than those of their dictatorial counterparts. This concession that IFIs supposedly extend to democracies does not stem from the potential for punishment by citizens, as suggested by Schultz and Weingast, but rather from the transparency inherent to democratic regimes. The authors further argue that the transparency and credibility of democratic administrations allow them to default without fear of retaliatory refusal to lend on the part of the IMF and the WB. In short, the views brought by social scientists are mixed and ambiguous.

The question of WB bias regarding regime-type is intertwined with the Bank’s principle of ‘political neutrality’, as outlined in its Articles of Agreement, approved in 1944. Art. III, Section 5, establishes that ‘the Bank shall make arrangements to ensure that the proceeds of any loan are used only for the purposes for which the loan was granted, with due attention to considerations of economy and efficiency and without regard to political or other non-economic influences or considerations’. Art. IV, Section 10 stipulates that ‘the Bank and its officers shall not interfere in the political affairs of any member; nor shall they be influenced in their decisions by the political character of the member or members concerned. Only economic considerations shall be relevant to their decisions, and these considerations shall be weighed impartially …’. Scholars have questioned the capacity of the Bretton Woods institutions to act neutrally (Swedberg Reference SWEDBERG1986; Thacker Reference THACKER1999; Barro and Lee Reference BARRO and LEE2002). Yet, this scholarship has left aside the potential impact of regime type on neutrality, instead emphasising the impact of pressures exerted by the US administration on WB’s and IMF’s decisions. This is significant as the United States has always been the only member-state with de facto veto power on the Bank’s and Fund’s executive boards, the only bodies with the authority to approve or reject loan proposals.Footnote ⁵

Studies that examine Washington’s influence on the WB have yielded mixed findings. Babb (Reference BABB2009), for instance, highlights the tensions between the US Treasury and US Congress over Washington’s policies vis-à-vis the WB. Gwin (Reference GWIN, Kapur, Lewis and Webb1997) argues that the US government and Washington-based non-governmental actors sought to influence WB policy only when US strategic interests were threatened. While many studies conclude that the United States and other donors (Britain, France, Japan and Germany) indeed exerted an impact on who received WB and IMF aid, and how much they received (Peet Reference PEET2003; Andersen et al. Reference ANDERSEN, HANSEN and MARKUSSENA2006; Fleck & Kilby Reference FLECK and KILBY2006; Woods Reference WOODS2006), others argue that both institutions are more responsive to the economic needs of recipient states than to the strategic interests of donors (Burnside & Dollar Reference BURNSIDE and DOLLAR2000). Another group of scholars suggest that the IMF and the WB were able to advance their own policies and insulate themselves from external political pressures (Staples Reference STAPLES2002; Helleiner Reference HELLEINER2014). As shown below, the WB, with few exceptions, tended to be politically neutral and seemingly indifferent to the type of regime. Yet a country’s political instability was among the few political factors that the WB openly discussed while evaluating loan requests.

III

During the long 1970s, a wave of military coups swept across Latin America, overthrowing populist, reformist and socialist regimes and replacing them with right-wing dictatorships. The new military rulers, like democratic leaders, viewed economic growth and stability as crucial in gaining legitimacy both domestically and internationally. To achieve their goals, each regime promoted its own mix of reforms, which included varying degrees of financial and trade liberalisation, privatisations and austerity plans. Through unprecedented financial and technical assistance, the WB (and the IMF) backed these efforts. This fact has contributed to the widespread, albeit inaccurate, view that the Bank, as a rule, favoured authoritarian regimes over democracies. This article argues that the Bank’s own organisational interests, a country’s creditworthiness and political stability, its readiness to adopt the WB’s policy recommendations, its proximity or enmity with the US administration, and its alignment with the US-led bloc during the Cold War, constituted the crucial factors in WB lending decisions in this period. To be sure, in most cases, the type of political regime of the borrower country hardly influenced the WB’s decisions.

The rise in WB lending during the mid to late 1960s and 1970s, when most Latin American nations were under dictatorial rule, is indicative, perhaps more than anything else, of the Bank’s own evolution. During George Woods’ presidency (January 1963 ‒ March 1968), the Bank’s membership increased from 81 to 106 member-states; the list of borrowing nations surged from 21 to 51; and the annual count of approved loans nearly tripled, rising from 44 in 1959‒60 to 103 in 1968‒9, with 47 of those loans granted to Latin American nations (Kapur et al. Reference KAPUR, LEWIS and WEBB1997, p. 186). During McNamara’s tenure (April 1968 – June 1981), the Bank’s staff expanded rapidly, credits to low-income countries quintupled, and loans to middle-income countries, such as those in Latin America, increased from $847 million in 1968 to $8 billion in 1978. From 1948 until the end of McNamara’s term, the WB granted 735 loans to Latin American nations: 122 loans before Woods assumed office (an average of 7.5 loans per year), 76 loans under Woods (approximately 15 loans annually), and 537 loans under McNamara (nearly 41 loans per year).

WB lending to Latin America (Figure 1) underwent three major cycles. The timing of these cycles varied slightly across countries, as seen in the cases of Argentina (which joined the Bretton Woods institutions in 1956, twelve years later than all other Latin American countries) and Brazil (with one of the longest dictatorships in the region, between 1964 and 1985). For Latin America as a whole, the first cycle began in 1948 and continued until the early to mid 1960s, encompassing crucial events such as the Cuban Revolution (January 1959) and the launch of JFK’s Alliance for Progress (May 1961). Among the most significant developments during this cycle was the launch of WB operations in Argentina in 1961, during the democratic regime of Arturo Frondizi (1958‒62). Colombia, under democratic rule, was another emblematic case, becoming a poster-child of the Alliance for Progress and, consequently, an important WB loan recipient (Offner Reference OFFNER2019). The second cycle reached its peak in the 1970s, following two oil shocks and the UN declaration of the New International Economic Order (NIEO) in 1974, when developing nations, critical of the Bretton Woods institutions, demanded an end to financial discrimination on political or economic grounds.Footnote ⁶

Figure 1.

World Bank loans per year ‘Big Four’, 1948–88

Source: Data from World Bank website, Projects by Country (elaborated by author).

The third cycle commenced in the aftermath of the Latin American Debt Crisis of 1982, particularly with the launch of the Baker Plan in October 1985, which revived credit flow to the region. The year 1985 also marked the introduction of the Structural Adjustment Programmes (SAPs) that led to a substantial increase in IMF and particularly WB lending. The ‘Cartagena Group’, an informal forum of the region’s largest debtors – Brazil, Mexico, Argentina, Venezuela and Colombia – along with Chile, Peru, Ecuador, Uruguay, Bolivia and the Dominican Republic, looked to the Bank to resolve their $360 billion debt problem. In late 1985, the group proposed a 20 percent increase in loans from the WB and other IFIs over the next three years.Footnote ⁷

Below, we examine the Bank’s lending to its four largest borrowers in Latin America. Two of them, Brazil and Argentina, experienced periods of military rule, while Mexico was under a civilian-authoritarian regime until the 1990s, and Colombia was the only one that remained democratic. As Figure 1 shows, WB lending followed similar patterns across the region despite substantial differences in regime type. The analysis is based on number of loans rather than loan amounts for three reasons. First, seeking to remain active and advance their organisational interests, WB staff aimed to increase the volume of lending. Each loan generated its own set of missions, sector studies, negotiations, and surveillance procedures, meaning that granting several small loans rather than a single large one was more advantageous for this purpose. Second, given that each loan agreement contained specific conditions and targeted different economic sectors, the more loans the WB granted, the deeper its involvement and influence became. Third, an additional sign of the importance that the WB has always attributed to the number of loans is that its loan officers are rewarded for the quality of their work. As Burke Knapp, a senior WB official explained, that quality ‘may be expanding loans and finding new opportunities for loans …, but it may also be denying loans for the right reasons. … [B]ut if you asked me who in the end, …, really got – who were the right boys [among WB’s staff], I’d have to say that the bright boys, …, were those that produced the volume [of loans]’.Footnote ⁸

IV

To date, Brazil has entered into 791 loan agreements with the WB, 123 of them during the last dictatorship (31 March 1964 – 15 March 1985). This article represents an initial attempt to provide an overview of the Bank’s first 40 years of operations in Brazil, with the type of political regime as its focus. As Figures 2 and 3 show, after a modest beginning, WB lending to Brazil increased almost steadily in both the number of loans and the amounts granted, reflecting a broader regional trend indicative of the Bank’s evolution. Except for a few loan-less years, Bank financing became a constant in the Brazilian economy, irrespective of the political regime in power or the economic policies being pursued. The longest break in WB lending (1959‒65) exemplifies this pattern, covering five administrations: the democratic administration of Juscelino Kubitschek; the brief presidencies of Jânio Quadros and Ranieri Mazzilli; the populist period of João Goulart; and the first 10 months of military rule.

Figure 2.

Brazil: number of World Bank loans per year, 1948–88

Source: Data from World Bank website, Projects by Country, Brazil (elaborated by author).

Figure 3.

Brazil: amounts of World Bank loans (in millions of US dollars) per year, 1948–88

Source: Data from World Bank website, Projects by Country, Brazil (elaborated by author).

As elaborated below, WB lending to Brazil until 1988 can be divided into three cycles: 1948‒64; 1965‒85; and 1986‒8.

The first cycle. The WB’s need to redefine its role was a major factor driving the beginning of financial activities in Brazil in 1949. When the Bank began operations in June 1946, its focus was on providing reconstruction assistance to post-war Western European countries. However, the launch of the Marshall Plan in March 1948 disqualified those nations from receiving further assistance from the WB and the IMF while the plan was in effect. Deprived of their initial clientele, both institutions began seeking new borrowers, with Latin American countries – the largest group of members at the time – emerging as their primary focus. The WB granted its first two development loans to Chile in March 1948, and the next ones, to Brazil and Mexico, in January 1949. Brazil’s vast development needs encompassed sectors such as agriculture, mining, energy and transportation.Footnote ⁹

Brazil’s first loan agreements with the WB, for telephone and hydroelectric projects in what the Bank defined as ‘underdeveloped regions’ of Brazil, were signed in January 1949 and May 1950Footnote ¹⁰ by President Euricio Gaspar Dutra (January 1946 ‒ January 1951). Running as a candidate from the centrist Social Democratic Party, Dutra was elected with the support of the Brazilian Labour Party. With the onset of the Cold War, he aligned with Washington, severing diplomatic ties with the Soviet Union and outlawing the Communist Party. His economic policy was in line with monetarist precepts promoted by the Bretton Woods institutions: he eliminated exchange rate controls, promoted the liberalisation of foreign trade, and implemented anti-inflationary measures. His developmentalist strategy prioritised four sectors: health, food, transport and energy – the last two comprising the bulk of the WB lending portfolio at least until the early 1970s.

When voting on its first loan to Brazil, the WB’s executive board had little information concerning the country’s economy and financial sector. The Bank’s staff saw the lack of reliable data, combined with exchange restrictions and ‘perhaps excessively nationalistic policies toward oil and other mineral resources’, as an obstacle to economic growth. Additional problems included the shortage of a trained workforce, a deficient tax system and the absence of a central bank. However, the staff assessed that Brazil’s political situation was ‘increasingly stable’, which could ‘accelerate Congressional action on a number of economic matters’.Footnote ¹¹ Hence, Dutra’s administration was perceived as stable enough to implement a consensual economic plan. Under these circumstances, Brazil emerged as a strategic WB client.

Mutual relations continued to evolve under President Getúlio Vargas, a former dictator (1937‒45) and leader of the Brazilian Labour Party, who was in power, for a second term, between January 1951 and his suicide in August 1954. Vargas’ regime faced fierce opposition and political instability. Nevertheless, it successfully promoted economic development by founding, among other institutions, the National Bank for Economic Development (BNDE) in 1952 and the state-oil company Petrobras in 1953 – a project that was supported by left- and right-wing parties alike. Vargas sought to solve acute problems in the balance of payments that could help attract foreign investments and facilitate the import of capital goods needed to foster an ambitious industrialisation plan. Facing inflation and consequent social unrest, Vargas implemented a mixed economic policy. On the one hand, his pragmatism led him to accept the rules of international finance and seek WB assistance; on the other, he maintained a developmentalist and nationalist stance (Skidmore Reference SKIDMORE2007).

The WB, like the IMF,Footnote ¹² saw Vargas as a reliable partner for expanding its operations in Brazil, granting his administration eight loans for energy and transportation projects. It should be noted that Brazil’s energy sector, which attracted most of the Bank’s loans during this cycle, was largely dominated by foreign companies.Footnote ¹³ In any event, during Vargas’ administration, Mexico received two WB loans, Colombia five, and Peronist Argentina, which until September 1956 was not allowed to join the Bretton Woods institution due to a US veto, received none (Kedar Reference KEDAR2012). Under Vargas’ regime, by July 1953, Brazil became the Bank’s chief Latin American borrower, receiving $152.8 million, equivalent to 40 percent of a total of about $384.53 million that the Bank lent to all its borrowers that year.Footnote ¹⁴ The WB assistance was accompanied by US loans. By June 1952, the United States Economic Commission had already promised $137 million to his administration.Footnote ¹⁵ In January 1953, US and Brazilian representatives completed drafts for 19 projects worth $493 million, while half of that sum was to be covered by the WB and the Export-Import Bank of the United States (Eximbank).Footnote ¹⁶

In the 18 months following Vargas’ death, Brazil had three different presidents. Instability, one of the few political issues that Bank management and staff openly discussed, negatively impacted Brazil’s creditworthiness and thus its eligibility for new loans. Between the approval of the last loan to Vargas in February 1954 and February 1965, when lending was resumed, the WB approved only three loans to Brazil, all during Juscelino Kubitschek’s democratic administration (1956–61) which, to be sure, was much less fiscally prudent than Vargas’ government: two in 1958 and one in 1959. This is significant given that since October 1958, a sharp economic deterioration, strikes, and riots against the high cost of living hampered his image. Notwithstanding the drop in WB lending between 1954 and 1965, bilateral interactions were hardly affected. Disbursements for previously approved loans continued, several WB missions visited Brazil and Brazilian delegations made visits to Washington. New sector studies were also agreed upon, setting the stage for future projects.Footnote ¹⁷ Waving the flag of ‘political neutrality’, the WB justified its meagre lending to Kubitschek based on economic criteria, especially his refusal to follow IMF and WB advice regarding inflation and utility tariffs – two recurrent themes in WB relations with Brazil and other Latin American borrowers (Marx and Turdo Reference MARX and TURDO2007). Ultimately, Kubitschek, who, like the US administration, held that development was needed to contain communism, found himself negotiating wage increases with powerful unions, which, in turn, drove up prices and disqualified Brazil for WB lending.

The Bank shied away from Kubitscheck’s administration for other reasons as well. During the first half of his presidency, Kubitscheck refused to negotiate a new and controversial stand-by arrangement (SBA) with the IMFFootnote ¹⁸ and, in June 1959, unilaterally interrupted interactions with the Fund. This partial detachment from the IMF also hindered US financial assistance. Indeed, Washington made its loans conditional on IMF agreements (Conforto de Oliveira Reference CONFORTO de OLIVEIRA2023). Kubitschek’s ties with the Eisenhower administration were ambivalent. On the one hand, he fully aligned with the US anti-communist stance; on the other, he was critical of the volume of US financial assistance to the region, publicly urging Eisenhower to increase aid to Latin America and adopt a ‘continental approach’ to address widespread and ‘dangerous’ communist unrest.Footnote ¹⁹ Against this background, the three WB loans that Brazil received in 1958 and 1959 actually stand out as a WB seal of approval for the Kubitschek administration.

Brazil’s situation vis-à-vis the Bank deteriorated during the constitutional but unstable presidencies of Jânio Quadros (January‒August 1961), whose independentist foreign policy worried the Kennedy administration, particularly his defence of Cuba’s right to self-determination – one of the most contentious issues in Latin America’s Cold War;Footnote ²⁰ Ranieri Mazzilli (August‒September 1961); and João Goulart (September 1961 ‒ April 1964), all of whom belonged to different political parties. During their terms, Brazil did not receive a single WB loan (Loureiro Reference LOUREIRO2017). The IMF, by contrast, provided two loans during this period: an SBA to Quadros in May 1961, and a non-conditional and automatically approved Compensatory Financial Facility (CFF) loan to the Goulart administration in May 1963.Footnote ²¹

Goulart, a former protégé of Vargas, encountered political weakness,Footnote ²² rising inflation, a slowdown in GDP growth, problems in the balance of payments, and difficulties attracting foreign capital and accumulating domestic savings. In tandem, the US embassy in Brazil, concerned about ‘Goulart’s past associations with Communists and his anti-US positions’, recommended being ‘particularly slow in entering into new aid commitments’, besides those already undertaken as part of stabilisation efforts and the Alliance for Progress.Footnote ²³ Although Goulart, seeking to avoid an overt crisis with the US administration, rejected the option of entering into economic accords with the USSR, his alliance with powerful unions, economic and political instability, and the expropriation of IT&T properties in Rio Grande do Sul in early 1962Footnote ²⁴ posed an obstacle to WB lending. The expropriation of US assets became a central Cold War issue in the late 1960s and posed significant difficulties to WB lending. For instance, in 1968, when left-wing Peruvian military dictator Juan Velasco Alvarado (1968–75) expropriated the assets of the US-owned International Petroleum Company (IPC) and other foreign firms, the WB suspended lending to Peru. Lending resumed only in 1971 after a meeting between Presidents Nixon and Velasco Alvarado, during which Peru’s president agreed to initiate secret negotiations on the IPC dispute and reached settlements with other expropriated companies (Brands Reference BRANDS2010). In reality, as revealed by an internal memorandum circulated among the WB’s executive directors, the Bank shared the US government’s concern about an imminent Marxist takeover in Brazil (Altamura and Kedar Reference ALTAMURA and KEDAR2021), even though Goulart was not a Marxist, as many in Washington feared. Nonetheless, such ungrounded concern did not necessarily have to translate into non-lending to Goulart. WB loans could have been used as leverage to encourage Goulart to more fully align with capitalist principles and strengthen Brazil’s dependence on Western financing sources, as the Bank did when commencing operations in Latin America, or as the IMF did during most of the Cold War. As Akerman, Weller and Pessoa (Reference AKERMAN, WELLER and PESSOA2022) illustrate in their analysis of IMF loan agreements between the 1970s and 1990s, the IMF imposed fewer conditions on loans to member-states more vulnerable to the global spread of communism than on countries where communist influence was less likely. It thus seems that Goulart’s case provides a clear example of the WB’s alignment with US Cold War policy and interests. The US anti-Goulart stance was so strong that the WB could not escape it without endangering its own ties with the US administration, on whose financial contributions it depended.

In 1962, Minister of Planning Celso Furtado launched a three-year stabilisation plan based on CEPAL’s developmentalism. The plan, which aimed to reduce inflation to 10 percent by 1965 and reverse external disequilibria without compromising growth, also advocated land and tax reforms (Bastian Reference BASTIAN2012). The plan’s success however, depended on foreign aid. In April 1962, the Eximbank, the United States Agency for International Development (US-AID), the US State Department and the Treasury agreed to extend to Brazil $35 million to support the stabilisation programme. The IMF for its part, agreed to postpone Brazil’s repayment of a $20 million debt owed to it.Footnote ²⁵ Nevertheless, in light of Goulart’s refusal to negotiate a new SBA and the failure to attract foreign credit, the plan was abandoned in July 1963, when inflation rose to 82 percent and output stagnated (Monteiro and Fonseca Reference MONTEIRO and FONSECA2012; Loureiro Reference LOUREIRO2010, Reference LOUREIRO2017). Known for his strong ties with organised labour and his autonomist foreign policy, Goulart aroused the suspicions of Brazil’s right-wing and conservative elites, who ultimately convinced the US administration to support a military coup to depose him (Rabe Reference RABE1999, pp. 64–7, 196–7). Goulart was overthrown on 31 March 1964. Discussing in retrospect this 11-year period with merely three loans, Burke Knapp, WB’s vice president of operations at the time, insisted that the WB’s main criterion was Brazil’s poor creditworthiness. In light of Brazil’s macroeconomic disorder, he claimed, the Bank suspended lending ‘even though there were projects which financially and even economically looked very rewarding’.Footnote ²⁶ Explanations aside, the WB’s non-lending to Goulart’s administration aligned with and served US Cold War interests in the region, marking an exception to its usual ‘neutral’ stance.

To conclude, the first cycle of WB lending coincided with the WB’s shift from Europe to Latin America, from reconstruction to development. Although during this formative period in the WB’s history its operations were modest, Brazil emerged as a prominent venue of Bank activity. This was true regarding a series of democratically elected administrations that adopted developmentalist economic approaches, implemented heterodox plans and gradual strategies to curb inflation, experienced political instability, and adopted diverse positions towards the United States and the IMF, whose loans were considered a precondition for WB lending. The sense is, as McNamara acknowledged several years later, that Brazil was such a vast country with such urgent needs that the Bank could not renounce it.

The second cycle. Following Goulart’s deposal, the military installed a right-wing dictatorship that lasted for 21 years, until March 1985. During this period, five different administrations were in place, each with a different economic policy and a distinct approach to Brazil–US relations and sensitive Cold War issues. President Castelo Branco (April 1964 ‒ March 1967) was clearly pro-US while encouraging foreign investment. Costa e Silva, and to a certain extent also Emílio Médici, adopted a more nationalist economic approach. When, in the mid 1970s, Presidents Ford and Carter began incorporating human rights among the criteria for foreign aid, President Ernesto Geisel (1974‒9) assumed a more contentious stance towards Washington, signing a controversial nuclear power agreement with West Germany in June 1975 (Cameron Reference CAMERON2018). The Johnson administration recognised Brazil’s military regime two days after the coup that it had supported, and promised to augment financial assistance to Brazil (Simoes Reference SIMOES2010, pp. 41–2). Marking a clear departure from Goulart’s foreign policy, Castelo Branco cut diplomatic ties with Cuba in 1964 and supported the US invasion of the Dominican Republic in 1965. However, the Johnson administration delayed providing financial support to the dictatorship because the State Department wanted to ensure that it first adopted what it perceived as the right policies (Wiesebron Reference WIESEBRON2016).

Reaching a compromise with the civilian elite and recruiting to governmental positions economists, urban planners, academics and other professionals (which the WB applauded), the regime propelled a long-desired modernisation (Klein and Vidal Luna Reference KLEIN and VIDAL LUNA2017, pp. 76-7). A reduction in real wages, which was aggravated by price liberalisation, was further facilitated by the repression of the labour movement and other opposition groups – a problem to which WB economists had been consistently indifferent, in Brazil and elsewhere. Simultaneously, the regime embarked on far-reaching reforms, including creating a modern central bank and establishing new financial entities. The IMF was quick to recognise this policy shift. A mission arrived in Brazil in May 1964Footnote ²⁷ and a first SBA was approved in December 1964.Footnote ²⁸ The WB, however, adopted a more cautious approach. Its financial exposure in Brazil was already too high to continue lending without waiting to see the regime’s capability to implement a ‘right’ economic plan.

As was the case under other administrations, dictatorial or democratic, the WB renewed lending to Brazil only once the government launched a coherent stabilisation plan. The Plano de Ação Econômico do Governo (PAEG) (November 1964 – March 1967), aimed to reduce government spending through institutional reforms, budget cuts and wage controls. However, it adopted a gradualist approach to inflation, similar to Goulart’s triennial plan (Kearney Reference KEARNEY, Galvan and Sil2007). Encouraged by the PAEG’s initial results, including a slowdown in inflation rates, the compliance with the financial targets stipulated for 1965 and the improvement of Brazil’s foreign exchange position, the WB ultimately deemed Brazil creditworthy.Footnote ²⁹ In February 1965, the WB granted Castelo Branco’s administration two loans for hydroelectric projects. WB President Woods explained that the Brazilian government, as part of a general debt rescheduling, had requested that the Bank also restructure a portion of the payments of its loans. Given that the Bank, as a rule, was opposed to rescheduling debts owed to it, Woods decided that ‘the Bank would be prepared to lend to Brazil for suitable projects in the amount necessary to offset the net repayments of principal (i.e. gross repayments minus projected disbursements) on the Bank’s existing loans to Brazil over the period 1965 through 1970’. As additional justification for lending, he stressed that the IMF would support Brazil’s economic plan.Footnote ³⁰ Also in 1965, the Inter-American Development Bank, US-AID and the Eximbank granted Brazil around US$650 million (Ribeiro Reference RIBEIRO2006). Despite partial disagreement over economic policy, relations between the WB and Castelo Branco were fruitful, securing him six additional loans in 1966, all for hydroelectric projects.

By means of wage cuts, implementing new taxes, modernising the financial sector, establishing a system for inflation adjustments and encouraging the inflow of foreign capital, the regime laid the groundwork for the so-called ‘Brazilian economic miracle’ of 1968‒73. Although the precise rate of GDP growth in Brazil during this period remains a topic of debate, the country’s economic performance undeniably improved. Conventional estimates suggest an annual growth rate of about 11 percent, while more recent studies, such as Bacha et al. (Reference BACHA, TOMBOLO and VERSIANI2023), place it closer to 7.1 percent. Industry expanded rapidly, with a significant increase in output. While the ‘miracle’ quelled social unrest, the presidencies of Costa e Silva (1967‒9) and Emílio Medici (1969‒73) brought unprecedented repression of workers, students, politicians and guerrilla organisations (Frieden Reference FRIEDEN1987). In 1974, when the ‘miracle’ dissipated, annual inflation rates rose to 35.2 percent, later decreasing to 22.1 percent in May 1975. In tandem, a rapid deceleration of industrial growth from 14 to 0.8 percent became politically intolerable for the government, especially after the parliamentary elections of November 1974 signalled that the regime would face a combative opposition in Congress for the first time since the coup (Carneiro Netto Reference CARNEIRO NETTO1986). Despite the authoritarian nature of the regime, WB lending to Brazil continued to expand. This became particularly pronounced after the first oil shock of 1973, which severely affected Brazil, a major oil importer. The lending surge was further bolstered by the implementation of the Second National Development Plan in 1974, which remained in place until 1979 and relied on a growth-cum-debt strategy (Batista Reference BATISTA1987).

The rise in WB (and IMF)Footnote ³¹ lending to the military regime was also driven by the second oil crisis in 1979, which led to a reduction in international resources from private lenders (Klein and Vidal Luna Reference KLEIN and VIDAL LUNA2017, pp. 96-8) and created a vacuum that the WB was keen to fill. During Geisel’s government, WB lending significantly increased, a remarkable fact in light of the simultaneous deterioration of Brazil’s economy and the multiplication of state-owned enterprises. Moreover, under Geisel, Brazil’s foreign policy diversified, and the government concluded bilateral agreements with European, Arab and African countries, as well as with Japan and China (Wiesebron Reference WIESEBRON2016). Thus, WB lending to Geisel’s government grew despite the declining economic situation, the support of state-owned enterprises, and the autonomist foreign policy. In June 1974, while approving the first loan to Geisel for a hydroelectric power project in Brazil’s Northeast, McNamara offered, by way of justification, the impressive economic growth of the miracle period. He explained that the Bank’s strategy consisted of ‘supporting several important objectives of the Government’s development policy’, particularly ‘the identification and support of projects designed to increase productivity and incomes in the lowest income segments of the population, … and to alleviate human misery’.Footnote ³² He also stressed the role that the Bank was playing in institutional and policy reform. When recommending an agricultural loan in November 1975, McNamara acknowledged that Brazil’s balance of payments had deteriorated but insisted that its economic problems were mainly the result of external shocks.Footnote ³³ This ‘neutral’ claim served as the main justification for lending to an economy in decline. In addition, given the WB’s significant exposure in Brazil, it had little choice but to continue its support. Allowing such a major debtor to fail was not an option. One also has to keep in mind that at the time, the WB feared that Latin American countries, critical of what they perceived as insufficient WB lending to them, would eventually follow in Fidel Castro’s footsteps and withdrew membership in the Bretton Woods institutions, as Cuba did in 1964. These concerns grew in the early 1970s, particularly after Socialist Salvador Allende was democratically elected president of Chile (1970–3). The Bank refused to grant loans to his government unless it resolved disputes with US-owned copper companies that had been nationalised under his administration (Kedar Reference KEDAR2017b).

The Bank consistently found an economic rationale to defend its lending, or lack thereof, to Brazil during the dictatorship. Although the Bank staff were aware of the regime’s political illegitimacy and human rights violations, they opted not to let these political and even moral factors influence their decisions. The government’s actions were known to all, including the passage of the First Institutional Act (AI-1) on 9 April 1964, AI-2 on 27 October 1965, and the more extreme AI-5, issued by President Costa e Silva on 13 December 1968. AI-5 granted the military the authority to close Congress and allowed the president to rule by decree. It also empowered the regime to suspend political rights for 10 years and to dismiss employees across all levels of government (Skidmore Reference SKIDMORE2007). The WB, showing significant political indifference, avoided uncomfortable questions regarding the legitimacy and morality of assisting a brutal dictatorship. For instance, in a meeting that McNamara held with national authorities during an official visit to Brazil in October 1968, Finance Minister Delfim Neto explained the draconian measures that the government was implementing to reduce the fiscal deficit, including wage controls and massive layoffs.Footnote ³⁴ McNamara, the architect of the Bank’s ‘war against poverty’, seems to have remained unmoved by the social consequences of these measures.

Despite the results, the WB’s positive stance towards Brazil’s dictatorship was not totally consensual. For instance, in a briefing to WB President Woods dated September 1965, the staff of the WB Western Hemisphere Department (WHD) admitted that the regime’s political survival depended on the state of the economy. They most likely understood that the regime was determined to repress dissent and social unrest if necessary to improve the economy and secure its own political survival. To remain active in Brazil, the WB ‘neutralised’ and ‘depoliticised’ these facts by either ignoring them or artificially detaching the economic measures that it requested as conditions for lending from their social consequences.Footnote ³⁵ In 1973, Holis Chenery, the Bank’s chief economist and one of the few executives to question the appropriateness of lending to Brazil’s dictatorship, suggested that McNamara not halt lending to the country but rather direct assistance only to ‘social justice’. This seems to imply that Chenery expected to diminish the inevitable political consequences of the Bank’s assistance somehow. However, the voices of Chenery and other critics were discarded one by one (Kapur et al. Reference KAPUR, LEWIS and WEBB1997, p. 278). The Bank’s indifference towards flagrant violations of human rights in Brazil was not unique. WB records indicate that McNamara did not consider the repression carried out by Latin American dictatorships as cause in itself for rendering those countries ineligible for WB lending. On more than one occasion, he brought to the Bank’s executive board vote loan proposals that the US administration opposed on human rights’ grounds (Kedar Reference KEDAR2018, Reference KEDAR2019 ).

Overall, relations between the Brazilian dictatorial regime and the WB were productive, albeit less harmonious than the increasing number of loans might suggest. McNamara viewed Brazil as a laboratory for his anti-poverty programmes and was determined to continue lending.Footnote ³⁶ However, he was frustrated because the military regime neglected its promises to improve income distribution, increase employment and reduce poverty rates. WB staff repeatedly criticised the regime for not fulfilling its commitments in the social arena, which was central to McNamara’s agenda. To counter criticism, the military rejected WB requests to establish a permanent office in Brazil, a move that stood in contrast to other chronic borrowers who had agreed to do so in the hope of improving lending. While WB staff called to cut lending to Brazil in response to this defiant stance, McNamara replied that the country was so big that ‘they [the Brazilian government] are going to tell us to go to hell’ (Sharma Reference SHARMA2017, pp. 128-30).

To conclude, WB activity in dictatorial Brazil significantly increased compared to previous periods. However, a similar increase took place during the same period in Argentina (which was under military rule only part of the time), Mexico and Colombia. This suggests that the increase in WB lending to Brazil derived, to a large extent but not exclusively, from the WB’s impressive growth during most of this cycle. Although the Brazilian dictatorship did not promote privatisations as other dictatorial regimes at the time did, its efforts to liberalise the economy and adopt anti-inflationary measures were welcomed in Washington making it eligible for generous WB lending.Footnote ³⁷

The third cycle. The greatest boom in WB lending to Brazil, and to the other ‘big four’, started in late 1985. This boom was strongly linked to two central developments: the advent of the Latin America’s Debt Crisis and the subsequent launch of the Baker Plan.

The Debt Crisis erupted in August 1982 when Finance Minister Jesús Silva Herzog announced that Mexico could no longer service its $80 billion debt. Sixteen Latin American countries, including Brazil, followed suit. In response, commercial banks halted new lending and focused on collecting or restructuring existing loans. However, creditors agreed to reschedule debts and offer bailout packages only to debtors who signed agreements with the IMF and implemented strict reforms. The debts of Brazil, Mexico and Argentina – the most indebted nations in the region – accounted for 89 percent of the total debt owed to major US commercial banks, threatening the stability of the US banking system.

In October 1985, at the annual meeting of IMF and WB governors, US Treasury Secretary James Baker announced the launch of the ‘Program of Sustained Growth’ that aimed to stimulate growth in debtor countries. The Baker Plan offered financial assistance in exchange for fiscal discipline along with cuts to public spending, increased tax collection, financial system and interest rate liberalisation, reduced trade tariffs, and the promotion of investments, privatisations, deregulation and expanded property rights (Míguez Reference MIGUEZ2013, pp. 35-7). In fact, assistance to overcome the crisis and avoid default was conditioned on the adoption of the ‘Washington Consensus’ package. By prioritising long-term development (the WB’s main mission) over short-term balance of payments stabilisation (the IMF’s mandate), the plan positioned the WB as the primary institution in charge of debt management,Footnote ³⁸ thereby expanding its lending and advisory tasks even at the expense of the IMF.

In the 1980s, the Structural Adjustment Lending (SAL) became the centrepiece of the WB’s response to the debt crisis (Sharma Reference SHARMA2013). First introduced in Africa, SAL became widespread in Latin America only with the introduction of the Baker Plan. SALs, which were not attached to any particular development project, were designed to encourage economic reforms while facilitating a rapid transfer of resources to debtor nations in order to solve acute problems in their balance of payments (Kraske et al. Reference KRASKE1996, pp. 230-3). A structural adjustment plan involved strict conditions, including the devaluation of the national currency, an increase in interest rates, a reduction of public expenditure, extensive privatisation, elimination of public subsidies and freezing of salaries (Babb Reference BABB2009, pp. 83-91).

With the transition to democracy and between José Sarney’s inauguration in March 1985 and the end of his term in March 1990, Brazil received 101 WB loans. In February 1986, Sarney launched a new stabilisation plan, the ‘Plan Cruzado’. Like his Argentine counterpart, the neoliberal President Carlos Saúl Menem (1989‒99), Sarney conducted widespread privatisations. With an average of 20 loans per year, WB lending to Sarney surpassed the 12 loans that it extended the dictatorship in 1983 – the year that marked the record in annual WB lending to Brazil’s dictatorship. During the term of Sarney’s successor, Fernando Collor de Mello (15 March 1990 – 29 December 1992), Brazil signed 57 additional agreements with the WB, 33 of them in 1992 – the largest number of WB loans that Brazil received in a single year between 1948 and 1999. While the new democratic regimes received an unprecedented number of WB loans, this lending boom was not so much aimed at promoting democratisation as preventing the collapse of the international creditors and salvaging the reputation of multilateral institutions like the WB that had been involved in Brazil and other Latin American countries for four decades.

V

This article showed that WB lending to Brazil, which began in 1949, was on the rise until 1988. A similar pattern was observed among other large Bank borrowers in Latin America: Mexico, Colombia and Argentina. While Brazil and Argentina were ruled by both democratic and dictatorial regimes, Mexico and Colombia remained under different types of civilian rule. Over four decades, the WB used its official principle of ‘political neutrality’ as justification to remain indifferent to the type of political regime and maintain its operations. For instance, during Getúlio Vargas’ government (1961–4), when the US administration and certain Bank economists viewed Brazil as leaning toward communism, the Bank invoked neutrality, claiming that economic factors disqualified Brazil from borrowing. However, during Geisel’s term (1974–9), McNamara used the same claim to justify continued lending, despite worsening economic conditions and human rights violations. To be sure, in some cases, the WB was keen to support US-Cold War allies in Brazil. In others, by contrast, it backed administrations that key figures in Washington – whether rightly or wrongly – viewed as vulnerable to communist influence. In any event, regime type and Cold War concerns did not necessarily intersect. Hence, rather than demonstrating a democratic advantage or disadvantage theory, the near-constant rise of WB lending to Brazil points to a combination of factors, most of them economic (e.g. the implementation of a consensual economic plan or anti-inflationary measures), and others political (e.g. Brazil’s relations with the US administration). As the case of Brazil demonstrates, political instability – widely regarded as an obstacle to economic stability – was the only political factor that the WB systematically considered and explicitly acknowledged, regardless of regime type. In addition, the Bank could not always disregard the US stance towards different regimes. Yet, unlike its acknowledgment of political instability, the WB has never admitted to being politically influenced by US positions or Cold War goals in its decision-making process, as was the case during Goulart’s term.

As was argued here, a main driver of the steady increase in lending was the Bank’s need to continue operating in developing nations. This was especially true in the case of Brazil, with its immense development needs. As shown, the WB experienced rapid bureaucratic expansion and significant diversification of its lending instruments between 1948 and 1988. By 1953, Brazil had positioned itself as the region’s top WB borrower, a status it retains to this day. Certainly, the WB loans added to Brazil’s external debt burden, creating a situation of mutual dependence: Brazil needed more loans to spur growth and repay its debts, while the Bank continued to lend in order to show the positive impact of its assistance and ensure that Brazil generated the income necessary to avoid default. All regimes turned to the WB, which, due to its multilateral composition, extracted a lower political cost than bilateral loans. Bank-supported development projects – whether in energy, roads, agriculture or education – brought economic and political benefits to all types of regimes. While an in-depth examination of the WB’s involvement in Brazil and other countries remains to be done, it is clear that the relations between Latin American borrowers and the world’s most influential development institution are much more complex than previously thought.