ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

The 1999 Brazilian Financial Crisis

The path to the 1999 financial crisis in Brazil started with a massive surge in inflows, but the scene was soon dominated by the high domestic interest rates, initially necessary for price stabilisation but later becoming permanent to avoid "another Mexico" and to respond to continuous external shocks. But these high interest rates soon created massive domestic financial fragility in the banking sector and in state government finances, leading to an increase in public debt through continuous private banking and state government rescue activities. And this public debt exploded due to high interest rates, which became systematically higher than both the growth in public revenues and the returns on reserves. In the meantime, the real economy imploded because of these rates, but high interest rates became even more necessary as a (poor) substitute for missing public sector reforms and as a price for political stalemate. It then did not take much (the Russian devaluation and default in August 1998 and a relatively minor internal political crisis at the beginning of January 1999) for Brazil to end up in a major financial crisis.

The 1999 Brazilian Financial Crisis

‘Macho-Monetarism’ in Action

The path to the 1999 financial crisis in Brazil started with a massive surge in inflows, but the scene was soon dominated by the high domestic interest rates, initially necessary for price stabilisation but later becoming permanent to avoid “another Mexico” and to respond to continuous external shocks. But these high interest rates soon created massive domestic financial fragility in the banking sector and in state government finances, leading to an increase in public debt through continuous private banking and state government rescue activities. And this public debt exploded due to high interest rates, which became systematically higher than both the growth in public revenues and the returns on reserves. In the meantime, the real economy imploded because of these rates, but high interest rates became even more necessary as a (poor) substitute for missing public sector reforms and as a price for political stalemate. It then did not take much (the Russian devaluation and default in August 1998 and a relatively minor internal political crisis at the beginning of January 1999) for Brazil to end up in a major financial crisis.


I Introduction

his paper attempts to understand the Brazilian financial crisis mainly from an “endogenous-failure” perspective. It argues that the general mechanisms that led to this financial crisis were in essence endogenous to the workings of an economy facing sudden liberalisation, a surge in capital inflows, ineffective regulation and weak governance. This paper will also argue that within this general framework, there is a very specific “Minskian” feature to the Brazilian crisis, which made it different from other financial crises both in Latin America and in east Asia: how a particularly radical monetary policy led to a major financial fragility in the private sector and state finances, and to an unmanageable Ponzi finance in the accounts of the federal government.

In Brazil, the absorption of inflows and the dynamic that it generated were uniquely conditioned by an environment characterised by an excessively aggressive monetary policy. These were the result of the Brazilian government’s (“machomonetarist”) attempts to sterilise inflows and defend the exchange rate and its stabilisation programme at any cost from continuous external shocks, especially those of the Mexican, east Asian and Russian crises. These high interest rates – sometimes the Brazilian authorities chose to set deposit rates over 20 percentage points above international interest rates plus country risk – and the peculiar (and self-defeating) way in which the federal government and central bank dealt with the financial fragilities of the private sector and state governments that these high interest rates created, meant that among the Latin American and east Asian crisis-countries of the 1990s it was only Brazil that saw its public sector sleepwalking into a major Minskian “Ponzi”.1

From this perspective, the Brazilian financial crisis seems to contradict one of the key propositions of the “moral hazard” literature. As is well known, one of the characteristics of this literature is to blame artificially low interest rates for recent financial crises in developing countries. Low rates would have created a general macro-micro dynamic which ultimately rendered both lenders and borrowers unable to assess and price their risks properly – leading them to accumulate more risk than was privately (let alone socially) efficient. It was not that under a special set of circumstances international and domestic lenders were intrinsically unable to assess and price their risk properly. It was that artificially low interest rates took away most of the incentive to do so. These low rates were the result of the combination of moral hazards (created mostly by government guarantees on deposits), and of the near-certainty in international financial markets that, as in every old movie Western, the cavalry, in the form of a vast international rescue operation, could be counted upon to arrive in the nick of time should the “natives” threaten to default or close their capital account. In short, if only these moral hazards had not existed, interest rates would have been higher, the over-lending and over-borrowing would have been averted, and the financial crisis could have been avoided.2 What the case of Brazil shows us is the dangers of the alternative scenario: excessively high interest rates can just as easily lead to financial fragility and Ponzi finance.

This paper concentrates on the period between the beginning of the Brazilian experiment with financial liberalisation and economic reform proper and the outbreak of the financial crisis

– i e, between the Cardoso “Real Plan” of mid-1994 and the financial crisis of January 1999. However, some attention will also be given to the period between the 1990 “Collor Plan” and the 1994 “Real Plan”, since it was with the Collor Plan (or “New Brazil” programme) that economic reforms first began to be implemented. Throughout the paper, the mid-1994-January 1999 period will be compared and contrasted with similar periods (between financial liberalisation and financial crisis) in Latin America and east Asia, in particular the cases

Figure 1: Brazil: Quarterly GDP Growth, 1993-1999

(Per cent variation with respect to the same quarter in the previous year)

a b c d e Per centPer cent 12 10 8 6 4 2 0 –2 –4 12 10 8 6 4 2 0 –2 –4 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1993 1994 1995 1996 1997 1998 1999

Notes: [a] = Beginning of the Real Plan; [b] = Mexican crisis; [c] = East Asian

crisis; [d] = Russian crisis; and [e] = State of Minas Gerais’ default and

the January 1999 crisis.


Source: Macrometrica.

of Chile (1975-82), Mexico (1988-94), Korea, Malaysia and Thailand (1988-97) and Argentina (1990-2001).

II Brazilian Reforms

Brazil’s long pre-1980 economic growth period was brought to an abrupt end by the 1982 debt crisis. The country first experienced a severe recession, made more acute by inexorably rising inflation and a heavy debt burden; it then expanded again, but this time only slowly and erratically until the end of the decade. As events of the 1980s placed the model of state-led development under considerable strain, in March 1990 the incoming president, Collor de Mello, announced a set of reforms in his New Brazil programme: the removal of subsidies for exports and phased reductions in tariffs.3 Also, a privatisation programme was begun in 1991, the fuel market was deregulated (ending years of state support for ethanol production), and the dissolution of the coffee and sugar trading boards was announced. The most contentious measure was the temporary freezing of virtually all financial assets (with limits of US $ 1,000 on bank and savings account withdrawals), but by the middle of the year this was subject to increasingly corrupt evasions and was subsequently abandoned.

Despite the political weakness of Collor de Mello’s short presidency, its unsure handling of macroeconomic policy and the humiliating way in which he was thrown out of office, the initiatives he launched radically changed the direction of Brazilian economic development.

In 1992, escalating corruption charges and several extraordinarily bizarre private scandals led to the forced resignation of president Collor de Mello. His deputy, Itamar Franco, became acting president in September and president in December. Economic policy continued largely unchanged during the first year of president Franco’s government, with its principal objectives being the liberalisation of most prices, control over public expenditure and a strict monetary policy. However, political and economic uncertainty continued after the change of president and inflation increased in late 1992, reaching almost 2,500 per cent in 1993.

In May 1993, after several changes of finance minister, president Franco appointed senator (and well-known sociologist) Fernando Henrique Cardoso to the position. Together with a group of economists, including Edmar Bacha and Pedro Malan, Cardoso devised an all-encompassing stabilisation plan, which came into operation on July 1, 1994 (the Real Plan, which took its name after the new currency, which it introduced, the real

– the fourth in nine years!) The main characteristic of this new plan was that, as opposed to most of its predecessors, it intended to avoid “shock treatments”, price freezes or surprise announcements. The plan took a long time to be prepared and was announced in detail several months in advance of its implementation. It was an attempt to reduce prices gradually by reducing both inflationary expectations – through the real being “pegged” to the US dollar at a rate of around one real to US $ 1 (but allowed to move in a narrow band) – and inflationary “inertia” (indexation); simultaneously, there was the aim of a progressive achievement of internal and external macroeconomic equilibrium. One of the main strengths of this new plan was the fact that it succeeded in gathering an overwhelming degree of consensus and public support. Its initial successes in mid-1994 significantly helped Cardoso’s own campaign for the presidency.

In terms of growth, after the 1982 crisis Brazil first experienced a severe recession, and then expanded slowly and erratically until the beginning of the Real Plan, when growth briefly accelerated rapidly, reaching a rate of 10.4 per cent in the first quarter of 1995 (vis-à-vis de same quarter the year before). The main growth-stimulus was brought about by the massive reduction of the “inflationary tax” that followed the implementation of the new stabilisation plan (Figure 1).

However, as growth accelerated with price stabilisation, the balance of payments deteriorated rapidly; at the same time, the “Tequila effect”, which followed the Mexican crisis of December 1994, began to bite. This forced the government to curtail aggregate demand rapidly. The annualised growth-rate of the economy fell by nearly 12 percentage points between the first quarters of 1995 and 1996. As monetary and fiscal conditions then eased, growth began to recover again in the second quarter of 1996, but after the east Asian crisis, monetary policy was tightened again and growth fell once more. This fall accelerated in 1998 as a result of the repercussions of the Russian devaluation and default; finally, the political crisis following the default declared by the newly-elected Minas Gerais state governor (former president Itamar Franco) on the state debts to the central government brought Brazil into its January 1999 financial crisis.

Thus, one of the most important peculiarities of the January 1999 crisis is that repeated external and internal shocks brought down the growth rate long before the financial crisis. In fact, this crisis is unique among those in Latin America and east Asia during the 1990s not only in that it took place during a period of recession, but also in that growth actually picked up almost immediately after the crisis – in fact, in 1999 Brazil posted an overall small positive growth figure for the year (0.8 per cent).5 As will be discussed below, the continuous external shocks and the resulting downward growth trajectory of Brazil are directly associated with the other peculiarities of Brazil’s experiment with economic reform and liberalisation, especially with its much higher interest rates, its growing public sector deficit and, in particular, its public sector Ponzi finance.

Initially, the Real Plan achieved a budget surplus in 1994 (equivalent to 1.1 per cent of GDP), but by 1995 even the deficit

Figure 2: Brazil: Net Private Capital Inflows and TheirFigure 3: Latin America and East Asia: Real Effective Rate ofComposition, 1970-1999 Exchange between the Beginning of Financial Liberalisation(US $ (1999) billions) and Respective Financial Crisis 60

60 120







40 3030 2020 1010 00 –10–10

-12 FDI portfolio “Other”

1970 1974 1978 1982 1986 1990 1994 1998

Cl Mx Ar Br KMT

1975 1981 1988 1990 1992 1994 1996 1998 2000 Base year = 100 in each country

Notes: In March 1990 the incoming president, Collor de Mello, began the

process of reforms with his New Brazil programme; also around that

time, the crucial Brady initiative was taking shape in many Latin

American countries.8 Source: World Bank (2004a). See this source for the definition of the three

components of capital inflows.

that was acknowledged had already turned into a massive figure equivalent to 4.9 per cent of GDP.6 This deficit increased to 5.9 per cent in 1996, 6.1 per cent in 1997 and 8 per cent in 1998. In the meantime, total net public debt (that is, total debt minus international reserves and other financial assets of the public sector) nearly doubled during this period, from 28.5 per cent to 50 per cent of GDP.7 This amount, although not excessively large as a share of GDP compared with other countries, became unmanageable due to the remarkably high interest rates. As the “primary” accounts of the federal government were never in any significant deficit throughout this period, (Figure 11), this large increase was exclusively a consequence of what was happening in the financial accounts of the public sector; and these reflected the cost of the financial fragilities created by the policy of high interest rates. In particular, it was the huge cost to the public sector of the repeated rescue operations of the domestic banking system (both private and public) and of state governments by the federal government and central bank via the absorption of large amounts of bad debt. Contrary to what is usually understood, it was these continuous rescue operations – rather than the cost of sterilisation, the acknowledgement of financial “skeletons” from the past or the financial cost of the previous constitutional reforms – that led the Brazilian authorities to sleepwalk into their Ponzi finance (see the table).

As mentioned above, the main peculiarity of the Real Plan is that it was implemented during a period in which Brazil had experienced three major external shocks, which had severe implications for its external accounts (and fiscal position). As a direct result of the Tequila effect that followed the sudden devaluation of the Mexican peso, 1995 saw a reduction in capital inflows into Brazil. The stock exchange suffered a similar problem – in the first three months there was a net outflow of US $ 2.0 billion. The figure for the same period in 1994 had been a net inflow of US $ 5.0 billion. Foreign reserves also fell by US $ 7 billion in early 1995.

However, these problems did not last long, and from the second quarter of 1995 there was a rapid return of private foreign capital (following Clinton’s US $ 50 billion Mexican rescue package







Notes: (i) A decrease in the index signifies a revaluation. (ii) [Cl] = Chile;

[Mx] = Mexico; [Ar] = Argentina; [Br] = Brazil; and [KMT] = average of

Korea, Malaysia and Thailand.

and the remarkable increase of domestic interest rates; (Figure 13). This continued in 1996, aided by changes in the legislation regulating foreign investment, as well as the ending of state monopolies (such as in the petroleum and telecommunications sectors) and the privatisation of other public assets (particularly in the electricity sector). In this year, foreign direct investment nearly trebled, to US $ 9.1 billion (Figure 2).

As is often the case in Latin America (but not in east Asia), a capital surge of this magnitude tends to “crowd out” national saving and to revalue the real rate of exchange (Figure 3).

The rapid revaluation of the real exchange rate was exacerbated by the economic authorities’ use of the nominal rate of exchange as one of their main anti-inflationary mechanisms (i e, as a price “anchor”). In fact, the real exchange rate fell by nearly one half between mid-1992 and mid-1996. In 1997 and 1998 this trend began to be reversed, but at a rate that eventually proved to be too little, too late.

The reversal of the rapid revaluation of the real began with the new upheaval in international financial markets owing to the crisis in many of the Asian economies; this brought the net quarterly inflow into Brazil down from US $ 10 billion (in the third quarter of 1997) to just over US $ 0.7 billion (in the fourth quarter). However, the sharp fall in net inflows at the end of 1997 was drastically reversed in the first quarter of 1998, when total net inflows reached US $ 22.5 billion.9 This was followed by another large net inflow in the second quarter of 1998, amounting to US $ 10.6 billion, but then, due to the impact of events in Russia, the figure for the third quarter dipped to a massive outflow of US $ 15.9 billion, followed by another net outflow in the fourth quarter. As these net outflows proved unsustainable, the government had no option but to devalue the real in January 1999. Thus, 1998 posted both the all-time record for net inflows (first quarter), and for net outflows (third quarter)! This exemplifies the difficulties confronted by economic authorities in the implementation of their macro-policies when they voluntarily operate with a liberalised capital account in a world of highly volatile flows, a high degree of “contagion”, and asymmetric information. Chile’s and Colombia’s experiences in the 1990s with controlling short-term capital flows via price mechanisms, or Malaysia’s

Figure 4: Latin America and East Asia: Domestic Real DepositFigure 5: Latin America and East Asia: Domestic Real Lending Rates between the Beginning of Financial Liberalisation andRates between the Beginning of Finanical Liberalisation and Respective Financial Crisis Respective Financial Crises

-25% Cl Mx Br Ar KMT Per centPer cent 40 30 20 10 0 –10 40 30 20 10 0 –10 Cl Ar Mx KMT Br 301% Per cent 125 100 75 50 25 0 125 100 75 50 25 0 Per cent

1976 1980 1988 1990 1992 1994 1996 1998 2000

Note: [Cl] = Chile; [Mx] = Mexico; [Ar] = Argentina; [Br] = Brazil; and [KMT] = average of Korea, Malaysia and Thailand.

short experiment with quantitative inflow-controls in 1994 seem to be among the few options open to policy-makers wishing to minimise this kind of instability but still operate with a relatively open capital account [Ocampo and Palma 2005]. Initially, the Brazilian monetary authorities tried to play down their reaction to the east Asian crisis, increasing their deposit interest rates by just one percentage point between July and October (from an annualised level of 20.98 per cent to 21.99 per cent, respectively). However, as the fundamentals deteriorated rapidly, they switched tactics later in November and increased the deposit rate to a remarkable 43 per cent – and as the consumer price index in the following 12 months only reached 1.3 per cent, from this perspective deposit rates in real terms reached a similar exorbitant level. The government also significantly increased some import tariffs and took measures to facilitate the inflow of foreign capital.

The short-term effect of this monetarist shock was to decrease the pressure on the exchange rate and help the return of foreign capital. However, short-term successes often led to complacency. By mid-1998, particularly due to the renewed inflows of foreign capital in the first half of the year, the east Asian crisis appeared to have been forgotten; and so was the large cost to the public sector of yet again having to rescue the heavily indebted state governments from the negative effects of the soaring interest rates, and the domestic banking sector from the effects of their rapidly growing non-performing portfolio – not surprisingly, few banking assets could perform after November when lending rates for working capital (in this practically inflation-free economy) reached 70.4 per cent and for consumer lending 82.7 per cent.

In fact, even before the Russian devaluation, Brazil already needed to borrow at a higher interest rate than many other Latin American countries while coping with reduced demand for its exports from Asia and Argentina. Its public sector accounts were also already out of control. Furthermore, 1998 was a year of presidential, state government and municipal elections. Not surprisingly, little was done, especially in relation to sorting out the public accounts – in fact, the political manoeuvres to negotiate the constitutional reform that would allow the president to run for a second term were already responsible for a long political stalemate and huge subsidies to state governments (in particular those controlled by opposition parties whose support was needed to approve the constitutional reform that would allow a president to run for re-election). This political and economic inaction of

1975 1980 1988 1990 1992 1994 1996 1998 2000

Notes: [Cl] = Chile; [Mx] = Mexico; [Ar] = Argentina; [Br] = Brazil; and

[KMT] = average of Korea, Malaysia and Thailand. Sources: For Latin America, ECLAC Statistical Division; for Korea, Malaysia

and Thailand, World Bank (2004b).

the government and huge subsidies to state governments may have helped Cardoso’s re-election, but did little to strengthen the Brazilian economy; as a result, the Russian crisis of mid1998 hit it very hard. The central bank reacted the only way it seemed to know how to react, and increased deposit rates from

19.2 per cent (August) to 41.6 per cent (October); in this inflationfree economy this led lending rates for working capital to jump to an annualised rate of 60 per cent and for consumer credit to an staggering 117 per cent.

Once again it became evident that no matter how large the levels of reserves, and no matter how high interest rates, they can never be large enough and high enough to withstand a sudden collapse in confidence and withdrawal of funds by restless international fund managers in an economy with a liberalised capital account. By the end of the year the central bank had acknowledged the loss of half its reserves and had a substantial proportion of those left committed in forward operations to support the real. Therefore, the government had little option but to devalue.

To contain the subsequent crisis, the central bank initially reacted as it had before, increasing deposit rates yet again to 43.3 per cent. However, by then it had become rather obvious that a monetary policy that reacts to external shocks in this exuberant way almost inevitably leads to an interest rate trap and in particular to banking sector fragility and an unsustainable public sector Ponzi-type finance. Furthermore, the president of the central bank – and later his successor too – had to resign in the midst of corruption scandals. To calm the markets Cardoso appointed Arminio Fraga (Soro’s right-hand man) to the presidency of the bank. Almost immediately, he decided to abandon the policies of “exuberant monetarism” that had characterised monetary policy since the beginning of the Real Plan for a softer monetary policy mostly aimed at maintaining the stability of the exchange rate.

III Dynamics of How Brazil Sleepwalked into a Public Sector ‘Ponzi’ Finance

If we compare the financial crises in Chile (mid-1982), Mexico (December 1994), Argentina (beginning of 1995 and December 2001), east Asia (mid-1997) and Brazil (January 1999), what most

Figure 6: East Asia and Latin America: Credit to Private Sectorbetween the Beginning of Financial Liberalisation andRespective Financial Crisis (Per cent of GDP) Figure 8: Latin America and East Asia: Annual Stock MarketIndices between the Beginning of Financial Liberalisation andRespective Financial Crisis (US $ terms, base year = 100 in each country) Cl KMT Mx BrPer cent 125 100 75 50 25 0 Per cent 125 100 75 50 25 0
2,214 1,907 Cl Mx Ar Br [86% p a] [57% p a] [64% p a] Ko Ma Th 1,500 1,250 1,000 750 500 250 0 1,500 1,250 1,000 750 500 250 0

1975 1981 1988 1990 1992 1994 1996 1998

1975 1980 1988 1990 1992 1994 1996 1998 2000

Notes: [Cl] = Chile; [Mx] = Mexico; [Br] = Brazil; and [KMT] = average of Korea, Malaysia and Thailand.

Notes: [Cl] = Chile; [Mx] = Mexico; [Ar] = Argentina; [Br] = Brazil; [Ko] = Korea; [Ma] = Malaysia; and [Th] = Thailand.

Figure 7: Latin America and East Asia: Real Estate Price

For the clarity of the graph, the Argentinean data are shown as a

Indices between the Beginning of Financial Liberalisation andthree-year moving average. The percentages shown are averageRespective Financial Crisis annual growth-rates; Chile’s rate refers to 1975-1980, and Argentina’s(local currencies, 3-quarter moving averages) to the growth-rate between 1988 and 1994. Sources: DataStream and IFC (2002).



amounts of foreign capital; however, their absorption had led (mainly via increasing liquidity, high expectations, and declining

Mx Ma Th Br [63% p a]



interest rates) to a Kindlebergian mania-type credit expansion, and an unsustainable consumption boom and asset bubbles.10 As a result, the Brazilian authorities (whose successful price stabilisation and its financial liberalisation coincided with the 1994 Mexican crisis) became determined to prevent the massive inflow of foreign capital end up being transformed into a credit


400 boom to the private sector. Figure 6 shows how successful they

were in preventing this, mostly via the sterilisation of foreign inflows and high interest rates.

High interest rates and the successful control of credit expansion, together with a more moderate policy of trade liberalisation, also succeeded in restraining the rate of growth of imports of consumer goods.11 Also, in Brazil high interest rates avoided a Kuznets-type cycle led by the construction sector; meanwhile in Chile, Mexico, Malaysia and Thailand, large inflows of foreign capital, the liberalisation of domestic finance and low interest rates did set in motion a construction mania boom, led by rises in real estate prices. Figure 7.

Finally, and despite large inflows and a radical privatisation policy, Brazil also avoided the other asset bubble (stock market) that characterised the similar period in Chile and Mexico (as well as those in Malaysia and Thailand). In fact, the only period of rapid stock-prices growth between the first quarter of 1996 and the second quarter of 1997 was almost entirely led by just one sector (telecommunications).

Although high interest rates were able to check the development of a Kindlebergian mania (via credit expansion leading to a consumption boom and asset bubbles in construction and the stock market), which in particular characterised other experiments with financial liberalisation in Latin America, this “success” came at a huge cost: high interest rates created an “interest rate trap”, which equally led to a financial crisis but via a different route. Therefore, the Brazilian policy-making

1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 88 89 90 91 92 93 94 95 96 97 98

Notes: [Mx] = Mexico; [Br] = Brazil; [Ma] = Malaysia; and [Th] = Thailand. The percentage shown in the graph is Mexico’s average annual rates of growth.

Sources: DataStream for Mexico and east Asia and ‘Jonas Lang LaSalle Index’ for Brazil.12 The Brazilian average is a mixture of some priceincrease in Rio, price-stagnation in São Paulo and a price-fall in Brasilia. Neither index provides data for Chile between 1975 and 1981; however, Chilean central bank statistics, though using a different methodology, show an increase similar to that of Mexico (Chile 1988; an analysis of these data can be found in Palma 2000b).

distinguishes the Brazilian period between its own financial liberalisation and financial crisis in terms of economic policy is its already mentioned high domestic interest rates, both for the deposit and lending rates. Figures 4 and 5.

The Brazilian authorities certainly used extremely high interest rates not only to defend their exchange rate and stabilisation programme from the continuous external shocks which they were experiencing, but also to avoid the worst excesses that had characterised previous experiences of financial liberalisation in Latin America. In short, previous experiments with financial liberalisation and economic reforms had also attracted massive

Figure 9: Latin America: Spread between Lending and DepositFigure 10: Latin America and East Asia: Public Sector BalanceRates between the Beginning of Financial Liberalisation andbetween the Beginning of Financial Liberalisation andRespective Financial Crisis Respective Financial Crisis

(as per cent of deposit rates, 3-year moving averages) Cl Ko Th Mx Ar Br Ma (Per cent of GDP) Cl KMT Mx Ar Br Per cent 200 160 120 80 40 0 Per cent 200 160 120 80 40 0 Per cent 10 5 0 –5 –10 Per cent 10 5 0 –5 –10

1975 1981 1988 1990 1992 1994 1996 1998 2000

Notes: [Cl] = Chile; [Mx] Mexico; [Ar] = Argentina; [Br] = Brazil; [Ko] = Korea; [Ma] = Malaysia;15 and [Th] = Thailand.

during this period created a financial crisis precisely by trying to avoid one!13

In relation to the domestic banking system, tight monetary policy led to such remarkably high lending rates that hardly any banking asset could perform – in fact, in this practically inflationfree economy, interest rates towards the end of 1998 for consumer credit peaked at an annualised rate of 123 per cent and those for credit for working capital at 63 per cent. That is, high interest rates may have helped to reduce the growth of the credit exposure of the financial system to the private sector, but they certainly did little to help improve the quality of this exposure; so it is hardly surprising that the banking system had problems with nonperforming assets. In fact, the ease with which the government could finance its domestic debt was due primarily to private banks falling over themselves to buy public paper, as this was just about the only asset that could perform in such a high interest rate environment. As for the rest of their portfolio, private banks tried to increase profitability by the self-defeating policy of ever increasing spreads.14

In fact, spread-data on their own are able to indicate in which countries (in both regions) the crisis of the domestic banking system came before the overall financial crisis, and was a major component that led up to it (Brazil, Argentina and Thailand), and in which it only came after the crash (when bank-portfolio became non-performing due to falling incomes and asset-price deflation, at the same time as banks’ foreign liabilities were soaring due to sharp devaluations).

As the Brazilian government chose continuously to rescue private and state banks in difficulties (adding an additional component of moral hazard by not making any serious investigation into how these banks had got into trouble in the first place – not even when the bank in question belonged to a senior minister),16 this policy contributed substantially to public debt. High interest rates were also at the core of the worsening of the other components of the public sector finance. In fact, while Chile and Mexico had at least managed a significant improvement in their public accounts, being either in equilibrium or in surplus at the time of their financial crisis (as well as east Asia), Brazil shows a growing deficit from practically the very beginning17 (Figure 10).

1975 1981 1988 1990 1992 1994 1996 1998 2000

Notes: [Cl] = Chile; [Mx] = Mexico; [Ar] = Argentina (consolidated public sector);

[Br] = Brazil; [KMT] = average of Korea, Malaysia and Thailand. Sources: ECLAC statistical division for Chile, Mexico and Brazil; Damill,

Frenkel and Juvenal (2003) for Argentina; and World Bank (2004b)

for Korea, Malaysia and Thailand.

Obviously, the key question in Brazil’s case is whether what was happening in the private and the public sector balances were interrelated phenomena, or whether the growing imbalance in the public sector was simply the result of a typical weak, “populist” government unable or unwilling to keep its house in order.

One of the difficulties that those working under the “Washington Consensus” paradigm have had in explaining other (non-Brazilian) financial crises is that these crisis-countries, particularly in Latin America, had both opened up their economies and implemented economic reform and done it by the book (including the small print). Furthermore, their public sector had managed to reach balance and their accounts had then been kept in order. Under these conditions it was not easy to explain following the neoliberal creed why the private sector had run wild, creating such large private macroeconomic imbalances, and accumulating so much financial risk, as to make a financial crisis almost inevitable. As a result, the Washington Consensus type of crisisexplanation seems to have had little choice but to fall back onto well-rehearsed arguments of “exogenous” market interference by governments and international institutions, thus switching the whole debate towards issues such as “moral hazards” and “cronyism”.

However, Brazil’s financial crisis (and later Argentina’s in 2001) seems to be an exception; apparently, these are crises easy to “explain” and easy to dismiss – the growing public sector deficit provided a familiar way out. Under these conditions reforms are not “credible”: in a financially liberalised economy within a growing globalised world, either governments understand that they have to be serious about the way in which they implement their reforms and make the necessary efforts to do so – i e, keep their accounts at least in balance – or the neoliberal “model” will have little chance of success.18

Nevertheless, the logic that led to Brazil’s 1999 crisis is (as is so often the case) more complicated than a typically simplistic Washington Consensus one: the obvious issue that so far is missing from the traditional explanations is precisely an understanding of the economic and institutional dynamics that led to this growing public deficit, and its relationship to the

Figure 11: Brazil: Financial Requirements ofFigure 12: Brazil: Net Internal Debt of the Federal Governmentthe Public Sector, 1993-98 and Central Bank and State Governments, Municipalities and(Per cent of GDP) Public Corporations, 1992-99 10

(Per cent of GDP)



1 2




1 2 3

1993 1994 1995 1996 1997 1998






Per centPer cent

Per cent

Per cent











Notes: [1] = Interest payments on public debt; [2] = public sector deficit/

surplus; and [3] = “primary” balance. A positive figure is a deficit, a

negative is a surplus.

radical-monetarist economic policies devised to keep the credit in the private sector, the private sector consumption and the asset prices under control. The first point to note is that the growing public deficit was almost entirely the result of interest payments on the public debt.

Other than for a small deficit in 1997, Figure 11 shows that the “primary” accounts of the public sector were in surplus or balance throughout the period, and that it was only the growing interest payments that brought the public sector into deficit. This seems to be the obvious “other side” of the high interest rates coin. However, the problem is more complex than that.

First, the true primary balance was not anywhere as healthy as the government statistics suggested. Not only were large privatisation receipts – US $ 58 billion up to the January 1999 crisis – accounted as “ordinary” public revenues,19 but also the huge costs of the bailouts of bank and state finances were disguised using a long lag between the actual bailouts and the proper acknowledgement of the associated costs in the public sector accounts. Regarding the latter, although the government did register the debt issued for the bailouts of banks as new liabilities on the balance sheet of the central bank, it registered at the same time the non-performing debt it had absorbed as an increase in “assets”. In this way, from an accounting point of view, there was initially very little immediate “expenditure” to register in the accounts of the federal government and central bank – leading to members of the government and journalists in some of the friendly financial press at the time even boasting that there were no significant fiscal costs associated with the bailouts! Of course, these “assets” never performed, a fact that was only gradually acknowledged as public “expenditure” later on by slowly reducing the value of these “assets” in the central bank balance sheet through provisions for “credit losses”.20

In this way, the government not only made the cost of bank bailouts less transparent, but was also able to transfer the fiscal impact of the bailouts to the future: at the time of the October 1998 elections, the central bank accounts showed “assets” specifically associated to the bailing out of private banks equivalent to US $ 24 billion, and “assets” associated with both the


1992 1993 1994 1995 1996 1997 1998 1999

Notes: [1] = State governments, municipalities and public corporations; and

[2] = Federal government and central bank.

bailout of state banks and other state debts equivalent to US $ 72 billion.21

Second (and as distinct from the 1982 debt crisis) in the case of Brazil in the 1990s it was the service of the public domestic debt that accounted for most of the interest payments paid by the public sector. Thus, while the external net debt of the government fell by more than half between 1993 and 1998 (from

14.4 per cent of GDP to 6.3 per cent), the internal net debt (despite the under-reporting of the bailout costs via “asset inflation”) doubled – from 18.5 per cent to 36.1 per cent.

Third, within the internal net debt, it was only the federal government and central bank component of this debt that was booming (Figure 12).

The reasons for the growth of the net debt of the federal government and central bank are apparently not completely straightforward. First, as they had a relatively low stock of debt to start with at the time of price stabilisation and financial liberalisation – in July 1994, the net domestic debt of the federal government and the central bank (although already rising) was equivalent to just 8 per cent of GDP – it is difficult to blame the effect of high interest rates hitting the initial debt for this extraordinary growth. Obviously, in order to get into an uncontrollable Ponzi situation one needs to have large debts to start with! Second, throughout the period between financial liberalisation and financial crisis there were (apparently) no significant “primary” deficits in the public accounts that needed to be financed. So, the crucial issue that needs to be explained is where did the Ponzi take-off? Why did the stock of debt of the federal government and central bank increase in the first place?

For many Brazilian analysts, the usual suspect to blame is the high cost of the policy of sterilisation followed by the central bank in the hope of avoiding “another Mexico”. Others blame the “skeletons” from the past – such as the need to recognise previously unconsolidated public liabilities, and the undercapitalisation of public sector banks (especially the Banco do Brasil).22 Finally, it has also been common to blame the new constitution for bringing about a delayed increase in public expenditure.23

Figure 13: Brazil: Interest Rates Paid for Internal Public Debtand Growth of Public Revenues, 1994-1999

1 2 a b c d Per cent 70 60 50 40 30 20 10 0 Per cent 70 60 50 40 30 20 10 0

8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 1994 1995 1996 1997 1998 1999

Notes: [1] = Interest rate paid by the government on its domestic paper

(annualised monthly rates); and [2] = growth of public revenues (annual


[a] = Interest rate increase following the Mexican crisis; [b] = following the east Asian crisis; [c] = following the Russian crisis; and [d] = following the State of Minas Gerais’ default and the January 1999 crisis.

However, although the cost of sterilising foreign inflows and of acknowledging “skeletons” from the past was high, as is shown in the table, the bailout of banks and of state governments made a far larger contribution to the increase in the public domestic debt. These constant rescue activities may not have been properly accounted for in the “primary” balance, but there was no way of hiding their impact on the gross stock of debt of the federal government and the central bank (which is where the public sector Ponzi finance took off.) The table presents an attempt to breakdown the increase of domestic debt of the federal government and central bank into its main components between the beginning of the Real Plan and the month before the January 1999 financial crisis.

The key issue that those working under the Washington Consensus type of economics do not seem to understand, that in a financially liberalised economy the government constantly has to face “damned-if-you-do, damned-if-you-don’t” types of choices in relation to interest rates. For example, as in Brazil during this period, a rapid surge in inflows tends to overvalue the currency; then, as long as this overvaluation is expected to continue (commitment to a “peg”), this becomes an incentive to substitute borrowing in domestic currency with borrowing in foreign currency. Soon there are domestic exposures that require interest rates to come down, but there are also external exposures that require that the currency remain overvalued (i e, interest rates need to stay up). In other words, from the point of view of the domestic financial system, interest rates are soon stuck between the needs of banks’ foreign-currency liabilities, and those of the banks’ domestic assets. In the case of Brazil, policy-makers opted for trying to avoid bankruptcies on banks’ foreign-exchange exposures, at the cost of bankruptcies on banks’ domestic assets exposures – and then they opted to foot the bill for the latter. In this interest rate “trap”, it is the choice between one type of bankruptcy and the other, (and given the attitude of the government) between one type of government rescue operation and the other. Either way,

(Annualised monthly rates, domestic currency)

Figure 14: Brazil: Interest Rates Paid for Internal Public Debtand Received for Foreign Exchange Reserves, 1994-1999*

1 2 a b c d Per cent 70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0 Per cent

8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 1994 1995 1996 1997 1998 1999

Notes: [1] = Interest rate paid by the government on its domestic paper; and

[2] = interest rate received for foreign exchange reserves (assumed equivalent to returns on US Treasury Bills.) 1999 is not included due to large monthly variations caused by sharp exchange rate changes that followed the January crisis. Domestic currency. Dates as above.

the public sector debt would swell – and, given the character of the present international financial market, from then on the economy is just one step away from speculative activity of the partly self-fulfilling, partly “truth-telling” type that tends to end in financial crisis.25

Table: ‘Deconstructing’ the Growth of Domestic Debt of the Federal Government and Central Bank, July 1994 – December 1998

(In US $ billion (December 1998 value))*

A Initial domestic debt (July 1994) 72.5
“Securitised” domestic debt in December 1998 266.9
(Increase in “securitised” domestic debt 194.4)
Total increase in domestic debt (“securitised” or not) 229.5
B Increase due to bailout of banks and of state governments 146.2
Private banks 42.8
State banks 48.0
State governments 43.6
Banco do Brasil24 11.8
Increase due to initial stock of debt 53.3
Increase due to sterilisation of reserves 30.0
Total increase in domestic debt (“securitised” or not) 229.5

Notes: * The “initial” domestic debt (July 1994) was first transformed into reais of December 1998 before being expressed in terms of US $ (using the exchange rate of that month). The domestic debt of December 1998 was changed from reais into US $ also using the exchange rate of that month. All other figures were initially taken in local currency of the time of the expenditure; then, to these figures, was added the interest payments that were actually capitalised until December 1998 (the Ponzi finance); finally, December 1998 totals in reais were changed into US $ in the same way as above. To calculate the increase in domestic debt due to the capitalisation of interest payment on the “initial” stock of debt (that of July 1994), primary surplus (deficits) were first deducted (added) to the “initial” stock of debt before adding the cost of capitalising interest rate payments; i e, it was assumed that the primary surpluses (deficit in 1997) were used to pay this part of the domestic debt (as mentioned above, these “primary” balances included privatisation receipts). From the cost of sterilising increases in foreign exchange reserves was deducted the interest received for foreign exchange reserves (expressed in local currency and assumed equivalent to returns on US treasury bills; see Figure 14). For a detailed explanation of the methodology used in this exercise, see Lopes and Palma (2005).

In sum, in Latin America a financial liberalisation (at a time of high international liquidity) with declining domestic lending rates (as in Chile, Mexico and Argentina; see Figure 5) seems to unleash a private sector credit explosion, leading to an unsustainable consumption boom and asset bubbles. The alternative case, Brazil, which tries to avoid these processes via high interest rates, tends to destabilise the domestic financial system and public finances in a way that leaves the government having to choose between a rock (absorbing large amounts of bad debt created by under-performing banking assets and deficits in the state governments due to the high interest rates they have to pay on state debts) and a hard place (allowing the collapse of state’s finances, and of the domestic financial system because of its foreign-exchange exposures). The Brazilian government chose the route of high interest rates leading to both high state government bad debt absorption, and high financial sector bad debt absorption.26

Related to this, of course, there were two crucial factors fuelling financial fragility: first, in Brazil financial liberalisation was implemented with a particularly inadequate system of regulation and supervision of the domestic financial sector. Furthermore, the federal government exacerbated this problem by overlooking and covering up cases of wrongdoing and corruption in financial institutions. Second, a weak government is a recipe for problems of the “federal government versus state governments” type; it is also bound to lead to political stalemates such as that between the government and parliament over much needed public finance reforms.

And once the stock of debt of the federal government and central bank began to swell, the high interest rate-related Ponzi finance took over (Figure 13).

Two crucial phenomena of the public sector Ponzi finance are evident in Figure 13. First, the violation of one of the most important financial “golden rules”: interest rates paid on public debt were systematically higher than the growth of public revenues (and, even worse, much higher than that of the practically stagnant income per capita). Second, each external and internal shock led to a sudden rise in this interest rate. This provoked a similar (but augmented) phenomenon in lending rates, which (in Minskian terminology) increasingly turned private sector finance from “hedge” into “speculative”, and then from “speculative” into Ponzi finance [Minsky 1982]. Finally, Figure 14 shows yet another side of the public sector Ponzi finance – the well known high cost of sterilisation.

Again, until the January 1999 crisis, interest payments on public liabilities due to sterilisation of foreign inflows were systematically larger than revenues from related assets (foreign exchange reserves).

Finally, not every aspect of Brazil’s 1999 crisis was, of course, unique. Although an internal public sector Ponzi was unique to Brazil among the 1990s crisis economies discussed in this paper, Brazil also had to face three problems common to all other crisiscountries: (i) constantly changing composition of the large private inflows; (ii) progressive shortening of the term structure of their debt; and (iii) the constant danger that in a financially liberalised economy the attack could also come from “within”.

Regarding the first, Figure 2 showed that in Brazil there was an erratic changing composition of inflows; this is also found in the other crisis-countries (but in east Asia it took a less extreme form). The changing composition made the already difficult matter of effectively absorbing massive inflows even more complicated. Regarding the term-structure of inflows, in Brazil these shortened significantly over the period – the ratio of shortterm debt to total debt jumped from less than 20 per cent in 1994 to nearly 60 per cent in 1998.27 As a result, the ratio of foreign exchange reserves to short-term debt collapsed from about 130 per cent in 1994 to 30 per cent in 1998. Obviously, this added further fragility and heightened uncertainty in an already difficult situation by making the economy extremely vulnerable to a sudden collapse of confidence and withdrawal of finance.

Finally, of course, in a financially liberalised economy, the “attack” could also just as easily come from within. Brazil (as Mexico and Korea before it) did not have significant defences against internal attacks on their exchange rates, as their “reserves-M2” ratios were also particularly low.28

In sum, the path to financial crisis in Brazil started with a massive surge in inflows, but the scene was soon dominated by the high domestic interest rates initially necessary for price stabilisation but later becoming permanent to avoid “another Mexico” and to respond to continuous external shocks. These high interest rates were successful in avoiding a repeat of Mexico, in consolidating price-stabilisation and in partially insulating Brazil from external shocks in 1995 and 1997, but soon created massive domestic financial fragility in the banking sector and state government finances, leading to an increase in public debt through continuous private banking and state government rescue activities. And this public debt exploded due to high interest rates, which became systematically higher than both the growth in public revenues and the returns on reserves. In the meantime, the real economy imploded because of these rates, affecting the growth of public revenues even further; but high interest rates became even more necessary as a (poor) substitute for missing public sector reforms and as a price for political stalemate, and to defend the peg, so as to avoid both further domestic banking crises due to high foreign-exchange liabilities, and a stampede by restless international fund managers. And the Ponzi finance in the public sector ballooned out of control as a result of this high interest rate trap. Again, it did not take much (the Russian devaluation and default in August 1998 and a relatively minor internal political crisis at the beginning of January 1999) for Brazil to end up in a major financial crisis.

IV Conclusions

The legacy left by the military regimes to their civilian successors was (at best) a complex one. The Brazil of 1985 was very different from that of 1964. Brazil’s economy had improved significantly, but the lot of the majority of its people certainly had not followed suit.29 The Sarney government proved unable to deal with the pressing problems of stagflation and debt. President Collor de Mello’s attempt to bring about a significant shift in the character of the economy and its integration into the world economy and to put a gradual end to over 50 years of stateled development failed. Short-term economic problems and the emergence of massive corruption and bizarre private scandals as major issues in 1992 revealed the fragility of his political base. This threatened not only his liberalisation programme, but also the survival of the administration itself. After Collor’s forced resignation, his successor Itamar Franco also proved unable to deal with Brazil’s major political and economic problems, but did manage to achieve some recognition in the last year of his government with the appointment of Fernando Henrique Cardoso as his finance minister. President Franco allowed Cardoso and his team freely to devise and implement a highly ambitious stabilisation programme. The Real Plan succeeded in dramatically reducing inflation and resulted in the election of Cardoso to the presidency.

Although until the end of 1998 the Real stabilisation plan had been succeeding in its inflation objective, this success had come at a growing cost. In fact, the crucial issue about Cardoso’s stabilisation programme was that it followed the oldest (“twostage”) macroeconomic law: one can only solve one macroeconomic imbalance by creating another, and then by hoping that the “invisible hand” of the market will sort out the newly created imbalance! The problem is that although the first stage of this macro-law sometimes can be achieved (particularly when international financial markets are happy to oblige), seldom is there such luck regarding the second. In Brazil’s case, although there was ample success with the Real stabilisation plan, what was needed regarding the second stage was firm government action to deal with the newly created imbalances and macroeconomic fragilities; instead, the government opted for the belief that the invisible hand and continuous access to international finance would eventually sort these problems out. As soon as inflation was tackled in mid-1994, tight and tough regulation and supervision of the private and public banking system and of state finances was needed; and as soon as inflation was conquered by the end of 1995 or beginning of 1996, a relaxation of the “quasipeg” leading to a managed real devaluation, a decrease in interest rates, an effective mechanism of capital controls on short-term flows, and the long-delayed public sector reforms were the only sensible options for dealing with the massively overvalued exchange rate, a fast growing deficit in the current account, the growing fragility of the domestic financial system, a domestic debt that was increasing at an unsustainable pace, and an evergrowing public sector deficit. However, as so often happens in politics, risk-averse inertia took over – particularly at a time when the government’s main political preoccupation was how to achieve a constitutional change that would allow the president a second term in office. As a result, a successful set of economic policies that brought inflation down from four figures to single digits in a few months (the Real Plan) was then kept in place after it had accomplished its objectives, run its full course and long passed its “sell-by” date. As a result, the same policies that had been the solution to the previous problem (hyperinflation) began to be the very problem of the new cycle; and the longer the policies were maintained, the more difficult it became to change them. In the meantime, the only thing the government seemed to be able to do was to try for high and volatile interest rates, and to cut the already low levels of public investment to absurd new lows.30

In this sense, the January 1999 financial crisis in Brazil was clearly more due to “weak governance” than in other crisiscountries of the 1980s and 1990s. In fact, the financial problems of the public sector demanded so much attention that some important new policy initiatives (in health and education, for example), some significant industrial restructuring, some major productivity growth going on in some tradable activities, and some modernisations in the public sector (e g, PETROBRAS) went almost unnoticed.

One of the strengths of the Brazilian economy in the past had been that its economic authorities (of different political persuasions) did not allow its domestic inflation, interest rates, exchange rate and external balance to become a constraint on economic growth. As this policy, after long periods of sustained growth, ended in the 1980s in “hyper-stagflation”, the Cardoso government reversed it into a new policy in which economic growth, interest rates, and the need for external balances were not allowed to become a constraint in the fight for price stability. Although inflation was successfully controlled (even after the huge devaluation in January 1999, the consumer price index for that year increased by just 8.6 per cent), this created major economic problems, not least the Ponzi finance of the public sector.

A domestic deregulated but badly supervised financial market, closely linked to a highly liquid, under-regulated and unstable international financial market (mostly made up of jittery fund managers prone to oscillating between mania and panic), coupled with a domestic economy characterised by large imbalances, a weak state, and an even weaker government coalition, made a sudden collapse of confidence and withdrawal of funds a real possibility. When in 1998 the Brazilian policy-makers finally realised that they had sleepwalked into a public sector Ponzi finance and some drastic policy changes were necessary, it was unfortunately an election year – again not the most propitious time to take major risks.

Initially, the Cardoso government had greatly benefited from the Brazilian people’s disillusionment with previous development and economic policies – in fact, this had been the ultimate condition required for the success of the Real Plan in its initial stages.31 However, by the end of 1998 the government found out that this was “political capital” from which they could not continue to depreciate endlessly.




[An earlier version of this paper was presented at a workshop organised by the Centre for Economic Policy Analysis, The New School, and at a conference on ‘Globalisation’ in Cambridge. I would like to thank the participants at both events, especially Paul Davidson José Antonio Ocampo and Lance Taylor, for their helpful comments. Edna Armendáriz, Ricardo Ffrench-Davis, Geoff Harcourt, Daniel Hahn, Jan Kregel, Arturo O’Connell and Fiona Tregenna also made very useful suggestions. The table is the result of joint work with Carlos Lopes; I am also extremely grateful to him for many lengthy discussions on the subject. The usual caveats apply.

1 According to Minsky, an agent runs into “Ponzi” finance when it is forced to borrow to keep up with payments on its existing debt obligations; i e, the agent has to capitalise interest and thereby add to its total debt. See Minsky (1982).

2 See, for example, McKinnon and Pill (1997).

3 Tariffs would halve in three years, with no nominal tariff exceeding 35 per cent by the end of that period. Of equal significance was the commitment to remove the cumbersome system of import licensing.

4 Unless otherwise stated, all the data in this paper have as source either the official government statistics (reproduced in the monthly bulletins of Macrometrica), the central bank data bank, DataStream, the statistical division of ECLAC, or the data banks of the IMF (2004a and b) and World Bank (2004a and b).

5 In the other crises the opposite was the case. In Latin America, in the year before their financial crisis Chile was growing at 4.8 per cent (1981) and Mexico at 4.6 per cent (1994); this was also true of Argentina before its first financial crisis after financial liberalisation, when its economy grew at 8.5 per cent (1994). In east Asia, Korea grew at 7.1 per cent in 1996, Malaysia at 8 per cent and Thailand at 6.4 per cent. In all these countries output fell massively after their financial crises.

6 See below for the less than transparent way in which the central bank acknowledged the real costs of the bailouts of private and state banks and state finances.

7 About 80 per cent of this net debt was domestic. 8 On the Brady plan, see Ffrench-Davis (2005); Palma (2003a). 9 Ironically, there is evidence that a substantial proportion of these new

inflows was made of funds brought into Brazil by foreign investors that had just been rescued in east Asia by the US government and the IMF after the July 1997 crisis.

10 For an analysis of this cycle see Kindleberger (1996) and Palma (2000b).

11 See Palma (2003a).

12 In Brazil, the stock price index of the construction sector actually fell in the four years between financial liberalisation and the January 1999 financial crisis.

13 There were obviously further problems with the banking sector and public finance, which there is no space to expand on here. The most important ones were the effect of rapid price stabilisation on banking profitability and on public finances, and the delicate political balance between federal and state governments’ finances.

14 On the double-edge effects of increasing spreads, see in particular Stiglitz and Weiss (1981).

15 For the special case of Malaysia, see Palma (2000b) and Ocampo and Palma (2005).

16 In 1997, for example, the Banco Bamerindus, owned by the then minister for commerce and industry, went bust in a very “untransparent” way, leaving a shortfall of about $ 5 billion. The government took over this bad debt without any investigation; it then sold what was left of the bank to the Hong Kong and Shanghai Banking Corporation.

17 In a way, in Ffrench-Davis’s characterisation of the key role of the public sector deficit in Brazil’s 1999 crisis makes this crisis one that still has some elements of the “old” type of crises; see Ffrench-Davis (2005).

18 See, for example, the IMF’s World Economic Outlook [IMF 1999].

19 At least this method was less “post-modernist” than that of Margret Thatcher, who accounted for privatisation receipts as reduced public expenditure (rather than as increased public revenues!) As one of her election promises had been that she would reduced the share of public expenditure in national income, this novel accounting practice was a rather useful devise in that direction.

20 The exception to this accounting-trick rule was the recapitalisation of Banco do Brazil, which was acknowledged as such from the beginning (i e, there were no ghost “assets” added to the central bank accounts).

21 The central bank accounts are done in such a way that it is not possible to separate these two items.

22 See for example Goldfajn (2002), who blames almost everything on irresponsible behaviour by previous federal and state governments. Furthermore, some authors insist that despite extended recognition of previously unconsolidated liabilities, at the end of the second administration of Cardoso in 2002 there were still unconsolidated public liabilities equivalent to another 10 per cent of GDP. See J P Morgan (2002) and Favero and Giavazzi (2002).

23 For a detailed and critical analysis of these issues, see Lopes (2003); see also Sainz and Calcagno (1999).

24 The recapitalisation of the Banco do Brasil is showed separately from the cost of rescuing other public sector banks because it is usually argued that this was just an acknowledgement of a “skeleton” from the past; however, a significant part of its problems were created during the Real Plan. In particular, its large exposure to the agricultural sector became rapidly non-performing due to high interest rates and the unwillingness of the government to put pressure on farmers to pay.

25 See Kuczynski (1999).

26 The main banks that had to be rescued were (in chronological order) Econômico (1995), Nacional (1995), Banco do Brazil (1996), Mercantil, Banorte (1996), Bamerindus (1997), State Bank of São Paulo (Banespa, 1997), State Bank of Minas Gerais (1998), and State Bank of Rio Grande do Sul (1998). Other state banks whose rescue initially cost the federal government and central bank less than those above but still more than US $ 1 billion were those of Bahia, Pernambuco and Paraná. Besides these, there were 15 other state banks that needed bailing out.

27 IMF (2004b).

28 See Palma (2000b).

29 For Brazil’s remarkable degree of income inequality, see Palma (2002).

30 Between 1994 and 2000, the length of the Brazilian railways actually declined; the rate of growth of roads only reached 0.7 per cent per annum (the respective rate between 1964 and 1980 was 5.6 per cent) and of paved road just 1.7 per cent (previous rate 16 per cent); and the rate of growth of electric-power capacity was 3.6 per cent, while that in the previous period had reached 9.8 per cent. See Pinheiro, Gill, Severn and Thomas (2001).

31 Diaz-Alejandro always insisted that popular support was the most important component of a stabilisation package; see (1989).


Boletim Macrometrica, various issues,

Chang, H-J (ed) (2003): Rethinking Development Economics , Anthem Press.

Chile (1988): Indicadores Económicos, 1968-86, Banco Central.

Damill, M, R Frenkel and L Juvenal (2003): ‘Las cuentas públicas y la crisis de la convertibilidad en Argentina’, Documento de Trabajo N 4, CESPA, Universidad de Buenos Aires.

Díaz-Alejandro, C (1985): ‘Good-bye Financial Repression, Hello Financial Crash, Journal of Development Economics, Vol 19, No 4; also in Collected Essays, (edited by A Velasco), OUP, 1989.

Favero, C A and F Giavazzi (2002): ‘Why Are Brazil’s Interest Rates so High?, IGIER – Università Bocconi, Working Paper No 224, July.

Ffrench-Davis, R (2005): ‘The Conflict between Purely-Financial and Real-Economy Macrobalances: The Case of Emerging Economies since the 1990s’, mimeo, ECLAC.

Goldfajn, I (2002): ‘Há Razões para Duvidar de Que a Dívida Pública no Brasil é Sustentável? Notas Técnicas do Banco Central do Brasil, No 25, Julho.

IMF (1999): WEO, May.

  • (2004a): IFS database.
  • (2004b): WEO database. IFC (2002): Emerging Stock Markets Factbook. Kindleberger, C (1996): Manias, Panics, and Crashes: A History of Financial
  • Crises, John Wiley and Sons.

    Kuczynski, M (1998): How Asian has the Asian Crisis Been? mimeo, Adenauer Stiftung, December.

    Lopes, C (2003), ‘Public Debt and the Brazilian Crisis’, paper included in an unpublished PhD thesis, Cambridge University.

    Lopes, C and G Palma (2005): ‘The Economic Consequences of Mr Franco’, mimeo, Cambridge.

    McKinnon, R and H Pill (1997): ‘Credible Economic Liberalisations and Overborrowing’, American Economic Review, 87 (2).

    Minsky, H (1982): ‘The Financial Instability Hypothesis: Capitalist Processes and the Behaviour of the Economy’ in C Kindleberger and J P Laffargue (eds), Financial Crises: Theory, History and Ppolicy, CUP.

    Morgan, J P (2002): ‘Emerging Market Research’, March 7.

    Ocampo, J A and J G Palma (2005): ‘Dealing with Volatile External Finances at Source: The Rôle of Preventive Capital Account Regulations’ in J A Ocampo and J Stiglitz (eds), Capital Markets Liberalisation in Developing Countries, Oxford University Press, forthcoming.

    Palma, J G (1998): ‘Three and a Half Cycles of “mania, panic and [asymmetric] crash”: East Asia and Latin America Compared’, Cambridge Journal of Economics, November.

  • (1999): ‘An Overview of Brazil’s Economy and the January 1999 Crisis’, in Regional Surveys of the World: South America, Central America and the Caribbean, 8th Edition, Europa Publications.
  • (2000a): ‘The Magical Realism of Brazilian Economics’, New School, CEPA WPS III, No 18.
  • (2000b): ‘The Three Routes to Financial Crisis: The Need for Capital Controls’, New School, CEPA WPS III, No 19.
  • –(2002): ‘The Kuznets Curve Revisited’,International Journal of Development Issues, 1.

  • (2003a): The Three ‘Routes’ to Financial Crises: Chile, Mexico and Argentina [1]; Brazil [2]; and Korea, Malaysia and Thailand [3], in Chang (2003).
  • (2003b): Latin America during the Second Half of the 20th Century: From the ‘Age of Extremes’ to the Age of ‘End-of-History’ Uniformity’ in H-J Chang (2003).
  • Pinheiro, A C, I Gill, L Severn and M Thomas (2001): ‘Brazilian Economic Growth in 1900-2000: Lessons and Policy Implications’, Third Annual Conference on Global Development Network, Rio de Janeiro.

    Sainz, P and A Calcagno (1999): ‘La Economía Brasilera ante el Plan Real y su Crisis’, ECLAC, Serie Temas de Coyuntura, No 4. Stiglitz, J and A Weiss (1981): ‘Credit Rationing in Markets Wit Imperfect Information, American Economic Review, June. World Bank (2004a): Global Development Finance CD-ROM.

    – (2004b): World Development Indicators CD-ROM.

    Dear Reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Back to Top