Recent Experience and Lessons for Latin America
Edited by Charles Collyns and G. Russell Kincaid
©2003 International Monetary Fund
April 10, 2003
Charles Collyns and G. Russell Kincaid
As 2001 came to a close and during the opening days of 2002, Argentina slipped into a crisis of unprecedented scope, magnitude, and complexity. The government made a forced exit from the long-standing currency board regime, imposed a comprehensive freeze on bank deposits, and defaulted on a significant portion of its sovereign debt. This crisis had enormous costs—social, economic, and political—for Argentina, and it sent shock waves that were felt throughout Latin America, most acutely in neighboring Uruguay and Paraguay, but eventually reverberating throughout the region during the course of 2002. By early 2003, at the time when this overview was finalized, the situation in the region as a whole had stabilized to some degree, as most governments were able to take effective measures to contain the crisis, but economic activity continued to be depressed, the outlook remained uncertain, and many underlying weaknesses had yet to be resolved.
This Occasional Paper seeks to draw lessons from the IMF's experience in handling financial crises around the world over the past ten years that are relevant to the challenges faced by countries in Latin America. A central theme is that, like other recent crises in emerging market countries, Latin America's problems have been related to underlying balance sheet weaknesses that exacerbated the magnitude of the crisis and greatly complicated the task of stabilization.1 In contrast to the Asian crisis, where public sector finances were sound, weak public sector finances—high levels of public debt and continuing fiscal deficits—prevailed in all the crisis countries of Latin America.
In such circumstances, attempts to run a prudent monetary policy were jeopardized by severe pressures to provide financing for the public sector—fiscal dominance—with the issuance of quasi-currencies by many Argentine provinces being an extreme example. At the same time, bank and corporate balance sheets, which previously had been perceived as quite robust in most cases, came under significant stress from wide swings in the exchange rate and interest rates, and a sharp drop in activity levels.
These homegrown balance sheet weaknesses, together with a continuing overall dependence on external financing, left these economies highly vulnerable to sudden shifts in investor confidence—which could be sparked by a reassessment of fundamentals, political uncertainties, or policy miscues—that could suddenly cut off access to emerging market financing and spur capital flight by domestic depositors. Transmission of such confidence effects across countries proved to be a critical channel for contagion across the region, in a process that turned out to be very difficult to predict or anticipate.
In much of Latin America, the imbalances and the task of resolving them were further complicated by widespread informal dollarization—a legacy of many years of macroeconomic volatility. In a highly dollarized economy, the usual benefits from an exchange rate adjustment can be overwhelmed by the negative impact on corporate, and hence bank, balance sheets as domestic earnings fall relative to dollar liabilities. Moreover, the central bank's inability to print U.S. dollars means that it cannot act as a fully credible lender of last resort to respond to bank runs, further destabilizing expectations.
In such circumstances, experience suggests that there is no quick or easy fix: effective solutions depend on developing a comprehensive strategy combining the full range of policy instruments. Critical elements of such a strategy must include fiscal measures to deal with immediate financing needs as well as provide a realistic path to medium-term sustainability; monetary and exchange rate policies consistent with quickly reestablishing a credible nominal anchor; a banking strategy to restore confidence in the domestic financial system; and an approach to debt restructuring to achieve viable public and corporate balance sheets and revive normal creditor relations.
Another key policy lesson is that, in moving beyond the immediate task of crisis resolution, it will be important to start laying the foundations for a more robust institutional framework that will be more crisis-proof and resilient in the future. Key tasks include bolstering central bank independence to provide a more robust nominal anchor; developing more effective financial supervision to identify and control balance sheet vulnerabilities; implementing fiscal reforms to underpin revenue buoyancy, expenditure prioritization, and transparency; introducing enhanced public debt management to lessen exposure to refinancing risks; and developing a more effective and predictable legal framework to support economic activity.
Recent experience with crisis has also had important implications for the IMF's work in assessing crisis vulnerabilities. IMF surveillance work has been strengthened, and initiatives introduced to increase its independence from IMF lending operations, responding to perceptions that the IMF could have been more forceful in questioning Argentina's policy framework in the 1990s, particularly the currency board arrangement.2 A more objective framework has been developed for the assessment of the sustainability of a country's external and public debt positions, and this approach continues to be refined.3 In addition, this experience has helped to shape discussions on the IMF's access policy in capital account crises and proposals to enhance the process for sovereign debt restructurings. The IMF's efforts related to compliance with standards and codes as well as on vulnerability assessments and early warning techniques have also been reinforced and deepened. Further lessons will no doubt be drawn in the future, especially as the IMF's Independent Evaluation Office has decided to review the IMF's role in Argentina as part of its work program for FY2004.4
The Argentine Context
Because the sections of this volume have their origins in the Argentine crisis, it may be helpful to provide a short reminder of the events that unfolded in late 2001 and early 2002, before offering a brief summary of the main findings of the Occasional Paper's individual sections. We will not explore the many interesting questions related to various policy options open to the Argentine authorities during the course of 2001 and whether and how this cataclysm could have been avoided, leaving that story to be written by others.5 It is also too early to write about the dialogue during 2002 between the IMF staff and management and the Argentine authorities to reach understandings on a comprehensive economic program to achieve lasting nominal stability, to restore fiscal sustainability, and to regain external viability.6
The story begins on December 1, 2001, when the Argentine government imposed wide-ranging controls on banking and foreign exchange transactions that effectively brought to an abrupt end the open capital account and fixed exchange rate regime of the currency board. The new controls included a weekly limit of Arg$250—at the time equivalent to $250—on withdrawals from individual bank accounts (el corralito), a prohibition on banks granting loans, and foreign exchange restrictions on travel and transfers abroad. These draconian measures were introduced to combat withdrawals of private sector deposits that threatened to topple the banking system. These withdrawals totaled some Arg$16 billion during 2001, or a decline of nearly 21 percent. The heavy deposit losses reduced gross international reserves of the central bank by nearly half, to $15 billion in December 2001—or equivalent to about one quarter of private sector deposits in the banking system. The loss of confidence, soaring interest rates, and a tight squeeze on private credit contributed to a sharp contraction in economic activity, with GDP declining an estimated 4 percent in 2001.
Demonstrations and riots greeted these measures, and President Fernando de la Rua resigned on December 20, 2001. Meanwhile, the complete disruption of financial intermediation had an immediate adverse impact on the already weak economy; in December, industrial production was down 8 percent from its level one year previously, construction was down 36 percent, and imports were cut in half. On December 23, a new President, Adolfo Rodriguez Saa, declared a default on government debt except for debt to international financial institutions and debt that had been subject to a voluntary, market-based debt exchange in October 2001—the so-called Phase 1 debt (because a second debt exchange operation was to follow, although it was never launched). The debt affected by this default amounted to about $70 billion out of sovereign debt held by the private sector of $130 billion. Total public sector debt of Argentina was estimated at nearly $167 billion at the end of 2001.
Civil unrest continued, and President Rodriguez Saa was soon replaced by President Eduardo Duhalde. On January 3, 2002, President Duhalde formally announced the abandonment of the currency board—the convertibility regime—and replaced it with a dual exchange rate system with an official exchange rate of Arg$1.4 per U.S. dollar—a 40 percent devaluation—for public sector and most trade-related transactions, while all other transactions would take place at the exchange rate prevailing in the private foreign exchange market. At the same time, export surrender requirements were introduced, foreign transfers abroad were blocked unless explicitly authorized by the central bank, and the limit on monthly deposit withdrawals was raised to Arg$1,500, which was then equivalent to less than $1,110 at the official rate of exchange. This experiment with a dual exchange rate regime did not last long, owing to continued pressure on gross international reserves, the wide spread between the official and floating exchange rates, and the leakages in the exchange controls. Consequently, in early February 2002 the government unified the foreign exchange market at the floating exchange rate of Arg$1.8 per U.S. dollar.
Also in early February, the government announced the asymmetric "pesification" of commercial banks' assets and liabilities. Bank loans to the private sector were converted from U.S. dollars into pesos at a rate of Arg$1 per U.S. dollar to lessen the repayment burden on debtors. (Loans to the public sector received a conversion rate of Arg$1.4 per U.S. dollar.) Deposits were converted at a rate of Arg$1.4 per U.S. dollar and indexed to inflation; this represented a 38 percent increase in real terms, but an initial decline of 25 percent in U.S. dollar terms.
Asymmetric pesification was a further blow to commercial bank balance sheets, which were already under heavy pressure because of the banks' substantial holdings of restructured Phase 1 government debt. In addition, the real increase in the money supply, together with the collapse in the demand for deposits in the Argentine banking system, resulted in a large monetary overhang.
Even with continued limits on deposit withdrawals, there was a general flight from pesos into U.S. dollars, and the exchange rate rapidly depreciated to Arg$3–4 per U.S. dollar during the period March–April 2002. The central bank had limited options to mop up liquidity, since it had lost the use of open market operations with government debt, owing to the partial government default, and had to rely on other instruments and develop new ones. At this point there was a clear danger of returning to hyperinflation, already suffered by Argentina several times in its history.
On the fiscal side, the Argentine authorities had not been able to deliver a promised zero fiscal deficit in the second half of 2001. The consolidated deficit of the public sector rose to nearly 7½ percent of GDP in 2001, as the primary balance moved to a deficit of nearly 2 percent of GDP. The debt ratio reached 62 percent of GDP at end-2001. However, the massive depreciation of the real exchange rate following the end of the convertibility regime led the public sector debt ratio to almost double—to nearly 119 percent of GDP in 2002.
Two additional elements further complicated fiscal management. First, provincial governments, faced by a collapse in their revenue base, expanded issuance of their own quasi-monies to pay wages and suppliers. Some of these quasi-monies were accepted by the federal government in lieu of federal tax payments, contributing to the broad acceptance of these quasi-monies and further complicating monetary management. Second, the government faced the need to issue new government securities to recapitalize the banking system following the asymmetric pesification. These new securities would place more immediate pressure on the fiscal deficit and add to the challenge of restoring fiscal sustainability over the medium term.
At this point, with revenues hit hard by the weak economy, the government had little choice but to apply very tight control over public spending, notwithstanding that under better circumstances the cyclical weaknesses of the economy would have argued for fiscal stimulus. The overwhelming necessity of avoiding hyperinflation placed severe limits on the monetary financing for the fiscal deficit. Monetary financing was the only financing available aside from the accumulation of arrears because the government had lost the ability to market its own debt with its partial default. In addition, even allowing for a restructuring of public debt to the private sector, restoring fiscal sustainability (and credibility) required generating a substantial primary surplus over the medium term.
Discussions on macroeconomic and structural policies with the transitional government of Argentina to begin the orderly resolution of the deep economic crisis lasted all of 2002. A total of 22 IMF staff missions visited Argentina in 2002, while numerous discussions with the authorities were held at IMF headquarters. Finally, on January 24, 2003, the IMF's Executive Board approved a Stand-By Arrangement in support of the authorities' short-term economic program (January–August 2003) to provide a policy framework for Argentina's political transition—presidential elections were scheduled for April 2003, with a new government expected to take office in late May.7 This economic program focused on core fiscal, monetary, and banking policies, while progress would also be made in preparing reforms of the tax system and intergovernmental relations, enhancing the application of the insolvency law, reviewing the financial situation of privatized utility companies, and collaborating with foreign creditors. This transitional program represents a key first step toward building the sustainable and comprehensive policies that are needed by Argentina to restore its economic well-being. With this IMF-supported program, one chapter in Argentina's economic history draws to a close even as another chapter begins.
Contagion and the Macroeconomic Implications of Regime Change
As Argentina slipped into crisis in late 2001 and early 2002, a key question that arose was whether contagion would spread to other countries in the region as it had during the Asian crisis. Section II of this volume investigates the various channels by which contagion could spread, including real linkages via trade flows, depreciations, or fiscal tightening; financial spillovers through common creditors and portfolio effects; a "wake-up call" leading to increased investor risk aversion or confidence loss by depositors; balance sheet fragilities; and policy paralysis.
After reviewing various early warning models, the authors of Section II conclude that predicting the spread of crisis remains more an art than a science. The signal-to-noise properties of such models are still not high enough, implying that these models remain incomplete and that they may overlook factors that may generate future crises. Indeed, by their nature crises are hard to predict and typically reflect vulnerabilities that markets and policymakers have previously overlooked. Once crisis triggers have been identified and assimilated, then an early response by markets and policymakers is more likely to head off that particular risk. However, the authors warn against complacency, pointing out that the experience of previous crises suggests that one should expect the unexpected and look for new sources of vulnerability.
With these caveats in mind, several Latin American countries were seen as particularly vulnerable, on the basis of a variety of standard leading indicators of crisis. No country was viewed as immune to adverse effects from economic turmoil in its neighbors, but the nature and severity of spillover effects depended primarily on a country's initial vulnerabilities. These vulnerabilities were seen as exacerbated in some cases by policy paralysis related to electoral cycles or social divisions and to policy rigidity, especially associated with crawling-peg exchange rate regimes and "fear of floating."
Against this background, Section III examines the macroeconomic consequences of a disorderly regime change such as that which took place in Argentina, drawing on the experience in the wake of crisis of nine emerging market countries. All countries, except one, experienced a substantial contraction in output in the first year of the crisis; the severity of the output contraction varied markedly, ranging from 5 percent to 13 percent. The output contraction mainly reflected a sharp curtailment of domestic demand, particularly investment. Beyond a general loss of confidence, domestic demand was hit by the adverse impact of large changes in the exchange rate on balance sheets of the household, corporate, and banking sectors with heavy foreign exchange exposures. The situation was accentuated by loss of access to external financing coupled with a domestic credit crunch reflecting a weakened banking system.
In most cases, external adjustment occurred rapidly as a strong growth of net exports led to dramatic swings from current account deficits to surpluses. However, the external strengthening had its basis mainly in a severe compression of imports, reflecting domestic economic weakness, with a relatively small contribution from exports (even in Asian economies with relatively large export sectors).
Typically, the output decline was short-lived, with reasonably robust recoveries occurring in the second year after the crisis erupted. In some countries, particularly in Asia, once financial markets stabilized and confidence began to return, there was scope to ease policies, particularly monetary policy, to support recoveries.
A common feature of crises of the 1990s was the very sharp initial depreciation of the nominal exchange rate, ranging from around 50 percent to 375 percent. In almost all cases, the real exchange rate seemed to have overshot its fundamental equilibrium levels. The experience with subsequent inflation varied considerably among the countries that suffered currency crises. Most Asian countries (and also Brazil in 1998) saw only a mild pickup in inflation. On the other hand, exchange rate pass-through to prices after one year for a range of other countries—including Ecuador, Mexico, and Russia—was much higher. Observers have suggested that the greater price stability for the Asian countries reflected their longer records of relatively low and stable inflation as well as more flexible labor and product markets.
Section III (prepared in early 2002) then seeks to draw implications for the unfolding crisis in Argentina. The authors conclude that no previous crisis was quite like the one in Argentina, emphasizing its complexity and multiple dimensions. While a handful of mitigating factors are cited, aggravating factors—such as the deposit freeze and forced asymmetric pesification, unilateral sovereign debt default and associated closure of private international finance, loss of a credible nominal anchor and confidence in the local currency, and the disregard for property rights—are viewed as predominating and likely to exacerbate the downturn in economic activity, depreciation of the currency, and the inflationary surge. The inflation outlook was seen as more difficult to predict, depending on the future stance of monetary policy, efforts to restore confidence in the currency, and the exchange rate path. (Based on information available in early 2003, activity is now estimated to have contracted by 11 percent in 2002. Consumer prices rose cumulatively by 41 percent over the 12 months through December 2002.)
Reestablishing a Credible Nominal Anchor
In the wake of the Mexican and Asian crises, in which "soft" exchange rates pegs proved untenable in the face of market pressures, it became widely believed that the only viable exchange rate regimes were at the two extremes, either a "hard" exchange rate peg—supported by institutional arrangements such as a currency board or formal dollarization—or a lightly managed or floating exchange rate.8 The collapse of the currency board in Argentina has seemingly toppled one pillar of this emerging paradigm for exchange rate regimes. Or at a minimum, it has reemphasized the demands placed on the economy—in terms of labor and product market flexibility and fiscal discipline—if a currency board is to be sustained, and the need for a viable exit strategy, including the possibility of exercising that option at a time of the authorities' choosing and from a position of strength, rather than after a collapse of policy credibility.
Against this background, Section IV examines how countries have found ways to reestablish a credible nominal anchor after suffering financial crises. The authors observe that attempts to retain a soft peg after a controlled devaluation in response to a speculative attack have generally ended in failure. Thus, countries have had only two real choices for their exchange rate regimes—a float or a very hard peg—with most post-crisis countries opting for a float.
In most post-crisis cases in which countries floated, the nominal and real exchanges stabilized within a few months after the onset of the crisis, and the initial overshooting was usually reversed within a year. Typically after an initial spike in nominal interest rates, real interest rates returned to their pre-crisis levels in a few months, and output effects of the tight monetary policy were limited. In practice, the countries most successful in ending quickly the period of exchange rate volatility and limiting inflation pass-through were those that tightened monetary policy early and sharply, and then sustained very high real interbank interest rates until stability had clearly been restored.
Perhaps surprisingly, the two countries that chose hard pegs—a currency board (Bulgaria) and dollarization (Ecuador)—experienced higher inflation and saw it decline more slowly than in countries that floated. Nominal and real interest rates fell more rapidly in countries that chose hard pegs, however. The relatively incomplete disinflation for hard peggers was attributed to their constrained nominal currency appreciation compared with countries with floating exchange rates. Consequently, pressures for real appreciation following the exchange rate overshooting could only be released via higher inflation among the hard peggers. The output decline (and ensuing recovery) was similar for hard peggers and floaters. However, the small sample size (10 cases) limits the confidence one can place in these results.
Restoring credibility is made easier by establishing a new monetary policy framework to help anchor public expectations. While a hard peg—either a currency board or dollarization—can gain credibility quickly if it is bolstered by adequate international reserves and consistent macroeconomic policies that command popular support, the long-run advantages are less clear, and exiting can be costly. A new hard peg is not a viable option for a country that has just abandoned another hard peg.
Countries with floating exchange rates can choose, in principle, to adopt a monetary-targeting or inflation-targeting framework for conducting monetary policy. In practice, countries have rarely adopted a monetary-targeting framework because of the unpredictability and instability of money demand. Even if a money target is met, the exchange rate may still be subject to wide swings, and the low short-run interest elasticity for money demand tends to result in high interest rates. Nevertheless, monetary aggregates still play a useful role, providing an objective "trip wire."
A full-fledged inflation-targeting framework can sometimes be put in place quickly, as in Brazil, and gain credibility swiftly. However, most monetary authorities have not followed this route because they fear that in the highly uncertain post-crisis environment it would be difficult to ensure satisfactory inflation results, and the credibility of the new regime can be quickly eroded.
In most crisis cases, the framework for monetary policy has not been clearly defined, leaving the authorities to exercise a fair amount of discretion and to rely on a wide range of high-frequency indicators (e.g., exchange rate, inflation, wages, industrial production, reserves, and monetary aggregates). However, winning market credibility for a new anchor may be difficult in such circumstances, particularly where there are strong pressures on the central bank to expand credit to either the government to meet a fiscal financing or because the banking system is illiquid. Although some delay in choosing a nominal anchor is understandable in the aftermath of a crisis, the authorities should therefore devote considerable effort to developing a clear monetary policy framework and then communicating with the markets.
Institutional reforms can also contribute to the attainment of credibility for monetary policy. Changes in the central bank law, such as assigning the central bank a clear mandate to stabilize prices and providing it the operational independence to carry out this task, are widely recognized as a prerequisite for successful inflation targeting. In the context of a simultaneous banking and currency crisis, it will also be important to put in place a credible strategy for financial sector rehabilitation and a strengthening of banking supervision to convince markets and the public at large that monetary policy will not be dominated by the liquidity needs of the financial system. Where the failure of the previous monetary regime was dramatic, full-scale, upfront institutional reform of the central bank has often been implemented (e.g., Argentina in 1991, Bulgaria in 1997, and Ecuador in 2000). However, it is important not to undertake institutional reforms myopically or partially.
Controllable operating targets and appropriate instruments are also essential to conducting an effective monetary policy. In a floating exchange rate regime, open market operations are the most efficient and effective means to influence the level of domestic interest rates. In cases where domestic money markets are seriously disrupted, however—for example, by a sovereign debt default (Russia in 1998)—other instruments must be relied on, such as reserve (or liquidity) requirements, direct credit controls, and unsterilized foreign exchange intervention. In the extreme case of Argentina in 2002, the central bank—which had continued to service its own obligations—began to issue its own paper, as a means both to absorb excess liquidity and also to send signals to the market about appropriate interest rate levels.
A more controversial issue relates to the contribution of foreign exchange intervention. Although large-scale intervention to defend an exchange rate level can be self-defeating, Section IV makes the case that relatively small-scale, intermittent intervention in the foreign exchange market may help to quiet the highly uncertain post-crisis environment if it is complemented by appropriately tight monetary policy. As confidence returns and capital inflows reappear, sterilized intervention can also help to facilitate a rebuilding of gross international reserves, as occurred in a number of the post-crisis Asian countries.
Dealing with Banking Systems Under Pressure
Section V examines the particularly difficult problem of how to address banking system problems in the context of a dollarized economy. As already emphasized, policies to resolve systemic insolvency are vital to restore confidence in the banking system and alleviate the threat that the central bank's lender-of-last-resort function will overwhelm its pursuit of price stability. This task is especially difficult in a dollarized banking system, where the policy options available to address bank runs and to rebuild a sound banking system are highly limited.
In most banking systems, deposit insurance and a credible lender of last resort ready to act in an emergency can reassure depositors and help to forestall deposit runs. However, dollarization limits the operation of both forms of protection, exposing depositors to greater risk of both illiquidity and, ultimately, loss. This greater risk will tend to make depositors more willing to rush to the door at the first sign of trouble. Thus, runs in highly dollarized financial systems may be more likely to occur, and once under way are harder to stop. These circumstances can make such financial systems more vulnerable to shocks, and demand for bank deposits highly unstable, thus jeopardizing the pursuit of price stability.
In response, policymakers have tried to increase the resilience of dollarized financial systems by inviting in foreign banks with parents that have deep pockets to operate domestically, organizing prearranged credit lines, establishing elaborate liquidity requirements coupled with tight restrictions on foreign exchange mismatches, and maintaining sizable, liquid international reserves at the central bank. In some cases, countries have earmarked foreign exchange for liquidity provision, excluding such assets from usable reserves for purposes of exchange rate management.
However, in the case of systemic runs in dollarized financial systems, these safeguards may not be sufficient, or deeper weaknesses can be revealed. For example, liquid foreign exchange assets of banks may include short-term U.S. dollar-denominated debt of the domestic central government. To gain U.S. dollar currency to satisfy depositors, banks must turn to the central bank to sell these bonds for U.S. dollars, draining international reserves unexpectedly. Prearranged credit lines may also be unilaterally cut by correspondent banks in time of the most urgent needs.
Depositor protection, in particular through a blanket government guarantee, has proven an effective, albeit expensive, tool to address systemic bank runs. In a dollarized financial system, funding constraints emerge similar to those discussed above for liquidity support. A guarantee of deposits in foreign currency can be effective only to the extent that U.S. dollar backing is available. A blanket guarantee extended in local currency will provide liquidity, but it is unlikely to instill depositor confidence that their U.S. dollar deposits are safe, and it may actually worsen the situation. The situation is further complicated when the government itself is perceived as bankrupt—as in the case of Argentina—implying that securities issued to depositors in return for bank claims cannot be automatically assumed to have full face value.
As the two key instruments to cope with bank runs are constrained, authorities with highly dollarized financial systems tend to resort earlier and more frequently to ad hoc administrative measures that restrict the availability of deposits or seek international support on a large scale. In considering the pros and cons of ad hoc administrative measures, the authors consider two main options.
A bank resolution strategy is equally important for a dollarized financial system to regain credibility and avoid a recurrence of bank runs. Dollarization can complicate, however, the assessment of the magnitude of the problem in the banking system. For example, currency mismatches in banks' balance sheets combined with massive depreciation of the exchange rate can cause sudden and substantial declines in banks' net worth that are difficult to detect and manage. Dollarization may also increase credit risk in the loan portfolio if households and corporates that have borrowed in U.S. dollars do not possess natural foreign exchanges hedges. The quantification of this additional credit risk is also extremely difficult to estimate in the short run.
Fiscal Policy and Economic Crisis
Notwithstanding the increasing relevance of banking issues in many recent crises, fiscal policy has remained a key source of crisis and also an unavoidable element of the policy response to crisis. Section VI examines two key fiscal issues: the assessment of when a fiscal position is unsustainable and likely eventually to trigger a crisis, and the role of fiscal policy in responding to crisis.
As regards assessing debt sustainability, in the past there has been a tendency, including in the IMF, to base the analysis on excessively optimistic assumptions about the economy's likely medium-term performance and the external environment. The IMF has therefore recently developed a new framework to provide a more disciplined approach to preparing projections by making explicit their underlying assumptions and by systemically subjecting these assumptions to sensitivity tests based on the historical experience of shocks.9
In Section VI, this new framework for debt sustainability is applied retrospectively to Argentina, Brazil, Mexico, and Turkey to see whether actual outcomes were within the ranges of the stress tests. In three of the four cases, the outcomes exceeded the upper end of the range for the stress tests for the ratio of public debt to GDP. One important reason for this deviation was unanticipated securitization of previously hidden contingent and unfunded liabilities of the government—in particular, the cost of recapitalizing the domestic banking system, which added on average nearly 15 percentage points of GDP to the public sector debt ratio. A second important source of underestimates for public debt was that actual movements in the exchange rate were substantially greater than allowed for in stress tests, again implying a jump in the ratio of public debt to GDP to the extent that debt has been denominated in foreign currencies. These results provide relevant pointers to the continuing refinement of the new framework.
Turning to experience with the conduct of fiscal policy in a crisis, the authors examine the experience with fiscal tightening in the aftermath of the crisis and consider the relationship with economic activity. Typically, the primary balance has improved modestly in the crisis year itself, while the adjustment becomes more substantial in the post-crisis years, totaling almost 4 percentage points of GDP over three years. The analysis divides the change in the primary surplus into two components: the cyclically adjusted component, reflecting fiscal policy decisions, and the cyclical component, reflecting the impact of the business cycle on the fiscal balance. The authors estimate that the improvement in the cyclically adjusted primary balance for four crisis countries has been on average around 3 percentage points of GDP in the first year when financing was limited and policy credibility was low, leaving the country with little alternative but to pursue procyclical fiscal tightening. (An appendix to Section VI presents in greater detail the techniques utilized in the case of Argentina to estimate potential output and to derive cyclically adjusted primary balances.)
The implications of changes in the fiscal position on activity are difficult to disentangle. In a simple Keynesian framework, fiscal tightening is procyclical, although the implications for aggregate demand also depend on the scale of countervailing automatic stabilizers. But one also has to take account of the role of fiscal tightening in easing immediate cash constraints on the government and in restoring confidence in longer-term fiscal sustainability. Thus, fiscal policy tightening will bring benefits in terms of lower real interest rates and interest rate spreads. Although these effects are hard to quantify, experience suggests that while real interest rates rise sharply in the crisis year, rates fall quickly in the two post-crisis years, contributing to an output recovery. Moreover, the experience of past crises shows that private sector consumption and investment—not public spending—were the main drivers for both the collapse of output and its subsequent recovery. No clear evidence emerged that fiscal tightening has contributed significantly to the depth of recessions during crises. Of course, recessions could have been milder if sustainability and financing constraints had been less binding on fiscal policy.
Section VI concludes with the authors developing the case that the quality of fiscal adjustment is also a relevant consideration for the impact of fiscal policy. High-quality measures, such as enduring reforms in tax structures and spending programs, are more likely to bring enduring improvements in the fiscal balance and, therefore, to have a greater impact on fiscal sustainability and confidence, thus mitigating Keynesian effects on aggregate demand. Fiscal consolidation is also likely to be more durable if it is accompanied by institutional reforms to strengthen fiscal management, enhance fiscal transparency, and bolster intergovernmental fiscal relations. Finally, it will be important to strengthen social safety nets to protect those most deeply affected by a financial crisis and to sustain political support for needed adjustment measures.
Twin banking and currency crises typically place the corporate and household sectors under enormous strain. Such crises usually involve a rapid and substantial change in relative prices, principally due to a depreciation of the real effective exchange rate, and simultaneously depressed domestic demand. Both of these developments place a heavy burden on the domestic (nonexport) corporate sector. Profitability and balance sheets are adversely affected, while access to external financing by domestic corporations is likely to be curtailed by deteriorating economy-wide financial conditions, weakening of the sovereign's creditworthiness, and by tightening of exchange controls.
Country authorities, therefore, need to develop strategies to cope with potentially widespread domestic debt restructurings for the corporate and household sectors and to handle relations with external creditors that hold both sovereign and nonsovereign claims as defaults loom or materialize. A two-track approach for dealing with both sovereign debt and household/corporate debt is often necessary because such debts have been issued under both domestic law and foreign law. The final two sections of this Occasional Paper address, first, issues pertaining to corporate debt restructuring under domestic law and, second, issues relevant to sovereign debt management, particularly restructuring under foreign law.
Household and corporate debts in developing countries are mainly held by the domestic banking system. Consequently, the burden of widespread debt restructuring by the nonfinancial private sector falls on commercial banks, weakening their balance sheets. A case-by-case (or debtor-by-debtor) approach pursued by banks could be highly time-consuming and requires a strong legal and institutional framework, which may not be available. In such circumstances, asset stripping by debtors is not uncommon and is inefficient, weakening both the creditor and debtor. Government intervention—across-the-board support—would result in a significant shift toward public ownership of the corporate and banking sector, at a heavy cost to public sector debt sustainability.
Section VII draws on experience of nine systemic corporate crises during the past two decades in Asian and Latin American countries to propose a hybrid approach. This hybrid approach combines elements of a creditor-driven, debtor-by-debtor process with a government-driven, across-the-board approach that could be fine-tuned to specific requirements of each individual country case. A government finance agent—an asset management company (AMC)—could engage in debtor-by-debtor resolution, using market-based incentives to encourage restructuring of viable firms. The in-court insolvency law and legal system may need to be strengthened and made more efficient, which will tend to encourage out-of-court negotiations to take place "in the shadow" of the court system. This will speed up the restructuring process. The authors also suggest that the government could, in some cases, provide foreign exchange cover to help the corporate sector to settle external arrears to foreign suppliers and to facilitate refinancing; such mechanisms could be modeled after the Fideicomiso para la Cobertura de Riesgos Cambiarios (FICORCA) scheme employed by Mexico in 1983.
Section VIII focuses on issues related to managing relations with external creditors during periods of increasing market tension and, in extremis, default. Based on a review of experience, it discusses the application of the Prague Framework, which is the operational framework endorsed at the 2000 Annual Meetings of the IMF and World Bank in Prague. This framework relies as much as possible on market-oriented, voluntary solutions, but in extreme circumstances it recognizes that a comprehensive debt restructuring may be unavoidable.
The authors divide countries' situations into two broad categories. The first category covers cases where policy adjustment by the affected country and official financing are combined with encouragement to private creditors to reach voluntary arrangements to overcome collective action problems. Within this category, the most commonly used instruments rely on voluntary suasion and regulatory requirements (e.g., market-based debt exchanges, commercial bank rollover agreements, private contingent credit lines, and investment requirements imposed on domestic financial institutions). However, although such approaches can ease the near-term adjustment burden, they can be expensive, raising the present value of sovereign debt and thus increasing the fiscal adjustment that may be required over the medium term to ensure sustainability.
In pursuing the voluntary approach, a crucial task is to preserve access to trade finance—import credits, export pre-finance, and working capital—to avoid a credit crunch that could produce a downward production spiral that would be damaging for both debtors and creditors. In this context, the authors stress the central role of the domestic banking system in intermediating such finance and reiterate the importance of a sound financial system. Official export credit agencies also can play a more active role, although historically their aggregate behavior has tended to be procyclical.
The second category involves the extreme situation in which countries face severe financing difficulties, sovereign debt is clearly unsustainable, and a restructuring of the existing debt profile, possibly including debt reduction, is required. The authors draw on the experience of five countries that have restructured their debt since the late 1990s to provide a step-by-step road map to debt restructuring. The experience demonstrates that countries that have restructured debt have lost access to international capital markets, with adverse consequences for the domestic economy. However, when a country faces a truly unsustainable debt situation, such a restructuring will eventually be required; the sooner the restructuring occurs, the lower the economic costs will be. In such circumstances, a debt restructuring that ensures a reduction in the sovereign debt of a sufficient magnitude to help restore medium-term viability is preferable to piecemeal restructuring of individual instruments. Section VIII ends with some observations about the latest proposals to resolve collective action problems, including the introduction of collective action clauses (CACs) more universally in sovereign debt contracts, and the Sovereign Debt Restructuring Mechanism (SDRM).
Allen, Mark, and others, 2002, "A Balance Sheet Approach to Financial Crisis," IMF Working Paper 02/210 (Washington: International Monetary Fund, December).
Fischer, Stanley, 2002, "Exchange Rate Regimes: Is the Bipolar View Correct?" paper presented to the American Economic Association (January 6, 2001); in Stanley Fischer, Speeches (Washington: JSTOR).
International Monetary Fund, 2002a, "IMF Executive Board Reviews the Fund's Surveillance," Public Information Notice 02/44 (April 18); available on the Internet at http://www.imf.org/external/np/sec/pn/2002/pn0244.htm.
_____, 2002b, "Assessing Sustainability," Staff Report (May 28); available on the Internet at http://www.imf.org/external/np/pdr/sus/2002/eng/052802.pdf.
_____, 2002c, "IMF Discusses Assessments of Sustainability," Public Information Notice 02/69 (July 11); available on the Internet at http://www.imf.org/external/np/sec/pn/2002/pn0269.htm.
_____, 2003a, "Argentina: Press Statement Following Article IV Consultation," Press Release 03/01 (January 8); available on the Internet at http://www.imf.org/external/np/sec/pr/2003/pr0301.htm.
_____, 2003b, "IMF Approves Transitional Stand-By Credit Support for Argentina," Press Release 03/09 (January 24); available on the Internet at http://www.imf.org/external/np/sec/pr/2003/pr0309.htm.
_____, 2003c, "Argentina: 2002 Article IV Consultation—Staff Report," and "Argentina: Request for Stand-By Arrangement—Staff Report"; available on the Internet at http://www.imf.org/external/country/ARG/index.htm.
_____, 2003d, "IEO Announces Work Program for 2003/2004," Independent Evaluation Office Press Release 03/01 (February 5); available on the Internet at http://www.imf.org/external/np/ieo/2003/pr/pr0301.htm.
Mussa, Michael, 2002, Argentina and the Fund: From Triumph to Tragedy (Washington: Institute of International Economics).
1For a complete description of the balance sheet approach to analyzing financial crisis, see Allen and others (2002).
2See IMF (2002a).
3See IMF (2002c).
4See IMF (2003d).
5Michael Mussa, the former economic counselor of the IMF, presents one informed, but personal, account of what went wrong in Argentina in his monograph entitled Argentina and the Fund: From Triumph to Tragedy (Mussa, 2002).
6See IMF (2003a). For an initial IMF perspective on the origins of the crisis and the ensuing policy responses in 2002, see IMF (2003c).
7See IMF (2003b and 2003c).
8See, for example, Fischer (2002).
9See IMF (2002b).