VI Public Debt Dynamics and Fiscal Adjustment
Author:
Mr. Richard Hemming
Search for other papers by Mr. Richard Hemming in
Current site
Google Scholar
Close
and
Mrs. Teresa Ter-Minassian
Search for other papers by Mrs. Teresa Ter-Minassian in
Current site
Google Scholar
Close

Abstract

Paying careful attention to public debt dynamics can provide advance warning of potential solvency and liquidity problems. The IMF’s new debt sustainability framework can make an effective contribution to crisis prediction. Fiscal tightening will usually be required to help resolve a crisis, and will be unavoidable when financing dries up. While no clear evidence was found that fiscal tightening has contributed significantly to the depth of recessions during crises, the quality of fiscal adjustment should receive more attention. In particular, adjustment measures should be consistent with medium-term objectives for tax reform and expenditure restructuring, and effective social safety nets need to be established. Institutional reforms to strengthen fiscal management will also help to resolve crises and prevent future crises. However, prudent fiscal policies aimed at achieving fiscal surpluses and low levels of debt in good times are the key to crisis prevention. Establishing a sound fiscal position will also permit a more flexible fiscal policy response should a crisis hit.

Paying careful attention to public debt dynamics can provide advance warning of potential solvency and liquidity problems. The IMF’s new debt sustainability framework can make an effective contribution to crisis prediction. Fiscal tightening will usually be required to help resolve a crisis, and will be unavoidable when financing dries up. While no clear evidence was found that fiscal tightening has contributed significantly to the depth of recessions during crises, the quality of fiscal adjustment should receive more attention. In particular, adjustment measures should be consistent with medium-term objectives for tax reform and expenditure restructuring, and effective social safety nets need to be established. Institutional reforms to strengthen fiscal management will also help to resolve crises and prevent future crises. However, prudent fiscal policies aimed at achieving fiscal surpluses and low levels of debt in good times are the key to crisis prevention. Establishing a sound fiscal position will also permit a more flexible fiscal policy response should a crisis hit.

This section is concerned with fiscal sustainability in crisis countries and opens by describing how judgments about the sustainability of public debt can be made using the IMF’s new debt sustainability framework (see IMF, 2001).1 It is concluded that the framework provides a useful basis for making such judgments, but that it requires strengthening and supplementing. The analysis then turns to issues concerning the fiscal adjustment required to restore sustainability once a crisis hits or to ward off a potential crisis. While significant fiscal adjustment is the norm, there seems to be no clear evidence that fiscal tightening has contributed significantly to the depth of recessions during crises. That said, the quality of fiscal adjustment needs to receive increased attention, and effective social safety nets have to be established. Finally, institutional reforms to strengthen fiscal management are discussed. These reforms would not only make fiscal adjustment more durable, but would also leave vulnerable countries better placed to prevent or respond to a crisis. Case studies of seven debt crisis episodes—in Argentina (2001), Brazil (1998), Indonesia (1998), Mexico (1995), Russia (1998), Turkey (2001), and Ukraine (1999)—are used to illustrate the main points.2

Debt Sustainability

The key to making judgments about debt sustainability and medium-term solvency is to determine how public debt dynamics are likely to evolve under different circumstances. This is an exercise that has long formed part of economists’ analytical tool kit, and of IMF program and surveillance work, but it is one that is being refined to better serve the objectives of crisis prediction, resolution, and prevention.

The IMF’s New Debt Sustainability Framework

The IMF’s traditional approach to analyzing debt sustainability involves projecting public debt by using certain assumptions about future primary surpluses and macroeconomic parameters (i.e., interest rates, real GDP growth rates, inflation, and exchange rates) that reflect current and announced policies, market reactions, and in some cases the recognition of previously hidden liabilities of government.3 Judgments about sustainability are made by reference to the ratio of debt to GDP and its trend in a baseline projection, the particular aim being to identify when unsustainable debt dynamics are a threat to macroeconomic stability, limit the growth potential of the economy, or both. Alternative policy scenarios are then constructed to assess the size and timing of the fiscal adjustment required to restore sustainability by placing the debt on a stable or downward path over some specified time period.4

The principal innovations of the IMF’s new debt sustainability framework are the use of a standard template for projections of both public and external debt, and an emphasis on stress testing—based on the historic behavior of key macroeconomic variables—to produce bands akin to confidence intervals around the baseline projection.5 A key objective is to impose increased discipline on the process of assessing sustainability.6

For the present analysis, the new framework has been applied retrospectively to crises in Argentina, Brazil, Mexico, and Turkey, to see whether actual outcomes fall within the stress test range. Because it is based on IMF program projections, the framework can most usefully be applied in this way to Argentina and Turkey, where programs had been formulated well ahead of the crises in 2001. In Brazil, the program was formulated after the 1998 crisis had hit, while in Mexico the 1995 crisis followed shortly after the program was formulated.

Figure 6.1 shows that, even though program fiscal targets were in most cases complied with, public debt generally exceeded baseline program projections, and in some cases exceeded the upper bound of the stress test range. More specifically:

Figure 6.1.
Figure 6.1.

Debt Sustainability Analysis

(Debt-to-GDP ratio)

Source: IMF staff projections.
  • The 2001 debt ratio in Argentina turned out to be slightly above the upper bound of the stress test range. The deviation from the baseline program projection mainly reflected unanticipated recession and weaker-than-programmed fiscal adjustment. If the projection had been extended to 2002, the debt ratio would have moved far above the upper bound of the stress test range, in part owing to the unanticipated securitization of contingent liabilities, but mainly because of a larger-than-expected real exchange rate depreciation.

  • For Turkey, the outcome for 2001 was well above the upper bound of the stress test range, despite the beneficial impact of stronger fiscal adjustment and lower real interest rates than expected. The deviation from the baseline program projection was not only the result of a much larger real depreciation and sharper recession than anticipated; in addition, the securitization of contingent liabilities added unexpectedly and significantly to the debt.

  • The post-crisis debt ratio in Brazil was at the upper end of the stress test range, despite a better-than-expected post-crisis fiscal adjustment and lower-than-expected real interest rates. An unanticipated large real depreciation was the main contributor.

  • In Mexico, the outturn in 1995 was above the one-year-ahead program baseline projection, because of a larger real depreciation, higher real interest rates, and slower growth than expected, and because of the unanticipated securitization of contingent and unfunded liabilities. Nevertheless, the outturn was within the stress test range. However, by 1997 the debt ratio had moved into the middle of the stress test range because initial fiscal adjustment was not maintained as planned.

While outcomes tended to be less favorable than baseline program projections, the stress test ranges broadly capture most of the outcomes. This being the case, the new debt sustainability framework provides a useful starting point for assessing the risk that public debt dynamics will become unstable. As such, it has the potential to provide valuable input into efforts to predict, resolve, and prevent crises. However, there is a need to strengthen the framework, in part through technical refinement—for example, by improving the statistical properties of the stress tests. But more critically, sources of sizable errors in the projections have to be addressed if the stress tests ranges are to encompass most of the plausible outcomes. These improvements are discussed in more detail immediately below, as is some complementary work that should be done to increase the usefulness of the new framework in informing judgments about sustainability.

Priorities for Future Work

The unanticipated securitization of previously hidden contingent and unfunded liabilities of government has been a significant source of error in the projections. For example, bonds issued in connection with bank restructuring in Turkey added about 16 percentage points of GNP to the debt ratio in 2001. In other countries, bond issues to cover a combination of contingent financial sector and corporate sector liabilities, as well as unfunded public pension liabilities, occurred over a number of years (including in the pre-crisis period).7 These totaled 19 percent of GDP in Mexico during 1995–97, 14½ percent of GDP in Argentina during 1999–2002 (mostly in 2002), and 8½ percent of GDP in Brazil during 1996–2000.

Unanticipated real exchange rate movements have also been important. These were especially significant in Turkey, where they added 13 percentage points of GDP to the debt ratio in 2001. An exchange rate depreciation of a magnitude that would have been impossible to predict was almost entirely responsible for the increase in debt—65 percentage points of GDP—in Argentina in 2002.8,9

Regular reporting of unfunded and contingent liabilities, reporting of quasi-fiscal activities (which are often a source of contingent liabilities), and assessments of the fiscal risks they pose are good practices of fiscal transparency that would provide a basis for more realistic debt sustainability analysis. In this connection, the development of a methodology for evaluating the full range of explicit and implicit contingent liabilities would be a priority.

A better understanding of real exchange rate movements would also strengthen debt sustainability analysis. In the medium term, the focus should be on the equilibrium real exchange rate rather than on its current level, since the latter is often subject to overshooting, especially with a shift from a fixed to a flexible exchange rate regime. In the short term, however, the ability to service external debt is affected by the level of the real exchange rate, and therefore its behavior during crises is also important. In addition, the debt dynamics are sensitive to interest rates, real GDP growth, and inflation. The impact of these variables is typically assessed on the assumption that they move individually in ways that have similar implications for the debt dynamics (e.g., the real exchange rate depreciates, interest rates increase, growth falls, and inflation is lower; as a result, the primary balance also deteriorates). The inclusion of model-based links between key macroeconomic variables in the new framework would allow a more precise assessment of the implications of the comovements in macroeconomic variables.

Even with improved debt projections, stress testing only points to the range of possible outcomes. While this assessment is useful, it falls short of indicating whether the debt dynamics suggest an increased likelihood of a crisis. As is well known, economic theory does not provide much practical guidance about when a debt ratio becomes excessive. Nevertheless, there is obvious appeal in a “magic number,” and it would be useful to know whether experience points to a threshold for the debt ratio that would signal a need to be more vigilant. Unfortunately, early warning system (EWS) models of debt crises suggest that the debt ratio is in fact a very poor predictor of debt crises. Short-term debt and especially foreign currency debt are far better predictors, although country-specific thresholds vary widely. That said, existing EWS models of debt crises are not well developed, and there is a need for a methodology that better links debt sustainability concerns with debt crises.10

Finally, it is critically important to pay attention to short-term liquidity as well as to medium-term solvency, either within the debt sustainability framework or as a supplement to it. More often than not, although solvency concerns are an important factor in limiting market access, it is a loss of liquidity that precipitates a debt crisis. Moreover, financing can dry up on a global or regional basis. The debt structure in crisis countries has a significant influence on liquidity, in that it is typically characterized by a sizable component that is linked to the exchange rate, together with a domestic component that is mainly short-term and in many cases indexed to interest rates. This structure is a particular source of vulnerability when real exchange rates and short-term interest rates are extremely volatile, especially since interest rate risk premia tend to rise, and average maturities to shorten, as longer-term solvency concerns build up.11 Under such circumstances, it is important to monitor rollover requirements relative to the availability of short-term financing with a view to being alert to possible liquidity crises.12

Fiscal Adjustment in Crisis Countries

Crisis countries routinely undertake large fiscal adjustments. This is what markets have come to expect, and indeed require, as a basis for restoring investor confidence. Yet the scale of fiscal adjustment has become one of the more contentious aspects of IMF involvement in crisis countries, with its rationale, impact, and quality being questioned.

Why Fiscal Adjustment?

Once a crisis hits, fiscal adjustment will usually be unavoidable, especially when crises have their origin in market perceptions that large government deficits and high debt are the root causes of a country’s economic problems. This is most obviously the case where failure to address severe fiscal imbalances poses a serious risk of insolvency and eventual default. A sizable fiscal adjustment is then urgently needed to return public debt to a sustainable path. Fiscal imbalances were a clear source of vulnerability in Argentina, Brazil, Russia, and Turkey, where initially large overall or primary deficits or high debt (or both) were a source of solvency concerns. Details are provided in Table 6.1, which shows the pre-crisis, crisis, and post-crisis fiscal positions for the seven crisis countries, centered on the country-specific crisis years.13

Table 6.1.

Overall Balances, Primary Balances, and Public Debt1

(In percent of GDP)

article image
Source: IMF staff estimates and projections.

Coverage is the consolidated public sector for Brazil, Mexico, and Turkey; general government for Argentina, Russia, and Ukraine; and central government for Indonesia.

Projections based on the proposed budget for 2003.

Program projections.

Net debt for Brazil, Mexico, and Turkey; gross debt for Argentina, Indonesia, Russia, and Ukraine (central government only).

The crisis itself can also have a sizable adverse fiscal impact, which can exacerbate solvency concerns. This impact can reflect the effect of changes in macroeconomic variables on fiscal aggregates, and in particular the impact of large real interest rate and exchange rate changes on the costs of servicing short-term and foreign currency debt, and on the size of the latter (in domestic currency).14 Crises can also expose hidden government liabilities, and securitizing them adds to the debt and to interest costs. In addition, crises can give rise to the need for additional primary spending, most notably on social safety nets, as in Argentina and Turkey.

Solvency concerns, whether they are the source or the consequence of a crisis, may lead markets to demand a high price—in terms of interest rate risk premia—for even very limited access to new financing, in which case a government can quickly face a liquidity crisis. Under such circumstances, fiscal adjustment takes on greater urgency, given the need to restore market access on reasonable terms, and its extent will be dictated largely by the financing constraint. Short-term liquidity has been a problem in most debt crisis countries, most notably in Argentina, Russia, and Turkey. But it was also a problem in Ukraine, despite initially low debt, because of a reversal of market sentiment following the crises in Asia and Russia.

Fiscal Adjustment Experience

As shown in Table 6.2, the average increase in the primary balance has typically been modest in crisis years (about ¾ of 1 percent of GDP), but it then becomes substantial in the immediate post-crisis year (about 2¼ percent of GDP) before being scaled back in the following year (to less than 1 percent of GDP). However, the change in primary balance describes the overall extent of fiscal tightening (and loosening).15 The modest increase in the primary balance in crisis years is actually the net effect of a loosening that reflects the operation of automatic stabilizers in the context of a sharp decline in growth, and an offsetting tightening due to discretionary adjustment measures. To distinguish these two components, Table 6.2 therefore also reports estimated changes in cyclically adjusted primary balances for a subgroup of crisis countries—Mexico, Turkey, and Ukraine— where comparable estimates are readily available. These estimates, which must be treated extremely cautiously because of methodological difficulties,16 reveal significant discretionary tightening on average during crisis years. However, about two-thirds of this tightening is offset by cyclical loosening. Discretionary tightening is eased substantially in the immediate post-crisis year, and there is a shift to significant discretionary loosening in the subsequent year. With post-crisis recovery, cyclical tightening adds to the former and offsets the latter (again by about two-thirds).17

Table 6.2.

Changes in Primary Balances

(In percent of GDP)

article image
Source: IMF staff estimates and projections.

Projections based on the proposed budget for 2003.

Program projections.

The variation across countries is large in terms of the size and timing of fiscal tightening. Thus, there was significant tightening in response to crises in Brazil (taking into account the year immediately following the crisis), Mexico, and Turkey (where there had already been significant tightening in the year preceding the crisis). This tightening was the key to a return to debt sustainability and market access. This contrasts with Argentina and Russia, which were unable to deliver needed fiscal adjustment and defaulted on their debt. However, a sizable post-crisis fiscal adjustment in Russia contributed to a return to debt sustainability and a restoration of market access.18 Initial fiscal adjustment was also inadequate in Ukraine, which was forced to restructure its debt under threat of default. While the debt in Indonesia increased from a relatively low pre-crisis level to one of the highest post-crisis levels, this was a consequence of high bank recapitalization costs and would not have been much easier to bear if there had been more aggressive fiscal adjustment.

The Demand Impact of Fiscal Tightening

In a simple Keynesian world, fiscal tightening has a negative impact on domestic demand, and as such would contribute directly to the depth of recessions. Thus, fiscal policy should balance concerns about debt sustainability against cyclical considerations. Recent criticism of the IMF approach to fiscal policy during crises (most notably by Professor Joseph Stiglitz) is that too much attention has been paid to the former and too little to the latter. However, in practice, immediate liquidity constraints and concerns about longer-term solvency generally leave crisis countries with little alternative but to tighten fiscal policy, even if its impact is procyclical (i.e., it leads to deeper recessions). Moreover, recessions during crises are also a response to high real interest rates and a lack of financing for the private sector, and not solely a consequence of fiscal tightening. Indeed, Figure 6.2 shows that, on average, output growth is negative in crisis years, and this is accompanied by sharp increases in real interest rates and spreads. But, as already noted, overall fiscal tightening has typically been modest in crisis years. By the time fiscal policy becomes seriously tighter in immediate post-crisis years, real interest rates and spreads are falling, and real GDP growth is recovering.

Figure 6.2.
Figure 6.2.

Primary Balances, Growth, and Interest Rates

(Averages)

Source: IMF staff projections.1Excluding Russia.

Notwithstanding its obvious limitations in terms of establishing causality, and the fact that averages can mask quite different country experiences, it is not self-evident from Figure 6.2 that fiscal tightening is responsible for deep recessions during debt crises. Ghosh and others (2002) reach a similar conclusion for a different set of capital account crisis countries on the basis of an examination of cyclically adjusted fiscal balances. Of course, this does not mean that fiscal adjustment is not in general contractionary during crises or, in other words, that recessions would not have been milder or subsequent recovery faster if solvency concerns or financing constraints (or both) had not precluded a looser fiscal policy. As is so often the case, the counterfactual is unknown.

There is also a possibility that fiscal tightening, even if it is normally Keynesian, may be non-Keynesian during debt crises. This is because public borrowing is reduced with fiscal tightening, and more generally there is a boost to confidence, which can lower interest rates and in turn stimulate consumption and investment. If this effect outweighs the direct contractionary impact of fiscal tightening, fiscal policy ends up being non-Keynesian and actually contributes to recovery.19 However, the scope for such an outcome to emerge through the interest rate channel may be limited in countries that have lost market access, and probably also in those countries that are undergoing banking crises, although fiscal adjustment that signals an end to a period of policy paralysis could, however, have a similar non-Keynesian effect. Moreover, where fiscal policy appears to be non-Keynesian, other factors, and most notably an accompanying real exchange rate depreciation, may be a more important source of a favorable output response.

Although fiscal tightening is the norm during crises, there are exceptions where fiscal policy has been loosened with a view to stimulating the economy. Thus, the Asian crisis countries tightened fiscal policy when the crisis hit, in response to external developments as well as to pay for financial sector restructuring. But with output contracting faster than anticipated and external constraints easing, fiscal policy was subsequently loosened in Korea and Thailand (although institutional weaknesses made it difficult to quickly reverse the direction of fiscal policy). Similarly, although Mexico tightened when the 1995 crisis hit, fiscal policy was loosened in the immediate post-crisis years. There is evidence that discretionary fiscal policy was indeed Keynesian in effect during this period (World Bank, 2001). In Argentina, the IMF-supported program allowed higher deficits at end–2000 to respond to recession, albeit unsuccessfully. In Brazil, the program fiscal target was temporarily eased in early 2000 to accommodate higher public investment against the background of an improved macroeconomic environment. But this easing had to be reversed later in the year in response to external and domestic shocks.

Tax and Expenditure Policies in Crisis Countries

Significant progress has been made in reforming the direct tax system in Brazil since the mid-1990s, but the federal indirect tax system relies heavily on cascading turnover taxes, which are detrimental to the competitiveness of exports, and a large share of subnational indirect tax revenue comes from the value-added tax (VAT) levied on utilities and business inputs. Widespread earmarking of revenue has reduced flexibility on the expenditure side, while minimum spending floors for several programs, particularly education and health, are an additional source of rigidity. However, progress has been made in reforming social spending, leading to considerable improvements in several social indicators; expenditure programs have become more output-oriented with the implementation of longer-term planning; and local governments are providing an increasing range of public goods and services.

Following the crisis in Russia, tax reform sought to simplify the tax structure, broaden the tax base, reduce rates, eliminate nuisance taxes, and end tax offsets. Thus, a flat 13 percent rate replaced a progressive income tax; a unified social tax consolidated payroll taxation; the VAT was converted to the destination principle, except for energy exports to the Commonwealth of Independent States (CIS) countries, and extended to individual entrepreneurs; excises were broadened to cover fuel products; tax avoidance provisions enhanced excise collections; the profit tax rate was reduced from 35 percent to 24 percent; and the punitive rate of turnover tax on gross sales has been gradually phased out. While there have been achievements in some areas on the expenditure side (such as the elimination of un-funded federal mandates and the establishment of more effective spending controls through the federal treasury), plans for civil service and public administration reforms introduced soon after the crisis have been of limited effect so far. More generally, there has been little progress in establishing expenditure priorities.

In Ukraine, budget rigidities, including earmarking and reliance on tax offsets, have been reduced, and some improvements have been made to the tax structure by shifting the VAT to an accrual basis and reducing payroll taxes. However, the frequent granting of tax amnesties and new tax preferences, together with failed attempts to index excise taxes, have eroded the tax base. At the same time, high income and payroll tax rates continue to provide incentives to a reportedly large shadow economy. Expenditure controls have been strengthened. However, the structure of government spending has not improved significantly, across-the-board expenditure compression during the crisis having since been gradually reversed without significant restructuring. Hence, despite a rapid increase in social spending in 2001–02, pension and social benefits remain poorly targeted. In addition, measures to reduce implicit subsidies in the energy and communal services sectors have not been pursued with determination.

The Quality of Fiscal Adjustment

When a crisis hits, and speed is of the essence, administrative and political constraints mean that the quality of adjustment usually suffers. Whether adjustment is primarily on the revenue side or the expenditure side, and despite some exceptions, fiscal measures that can be implemented quickly and with a significant impact often have negative structural consequences. This is the case, for example, when there is increased reliance on distortionary taxes, such as export taxes and taxes on financial transactions. Russia levied the former, Brazil the latter, and Argentina both. One-off tax measures are an additional problem; for example, tax amnesties in Argentina and Turkey have increased revenue in the short term but are a longer-term disincentive to tax compliance.

Across-the-board expenditure compression has also been common. This approach may have some merit in that it begins to tackle areas of spending that are traditionally resilient, such as public sector wages in Argentina and subsidies in Ukraine. However, it can also result in an accumulation of spending arrears, which was a major consequence of cash sequestration in Russia. There is also a tendency for public infrastructure to bear the brunt of the adjustment. The evidence for Mexico is that such cutbacks contributed to a reduction in output in the immediate post-crisis years (World Bank, 2001).

A trade-off between the speed and quality of adjustment may be inevitable, and low-quality adjustment an unavoidable consequence of having to act quickly. However, adjustment measures judged by markets to be of low quality are unlikely to restore credibility on a sustained basis. There is thus a risk that capital outflows could resume and financing constraints reemerge. Therefore, when quality of adjustment has been sacrificed in the heat of a crisis, in the interests of size and speed, it is crucial to quickly substitute high-quality for low-quality measures as the crisis abates. This is primarily a policy issue, the most pressing need being to improve the efficiency of the tax system and the effectiveness of spending programs. However, although a number of crisis countries have implemented significant tax or expenditure reforms (or both), weaknesses remain that constitute a source of continuing vulnerability. This is illustrated in Box 6.1 (on preceding page), which discusses tax and expenditure policies in Brazil, Russia, and Ukraine, and points to damaging distortions and inflexibilities that continue to limit the capacity of fiscal policy to respond effectively to crises.

Social Safety Nets

Brazil has a broad array of social insurance and assistance programs and spends over 20 percent of GDP on social programs. However, only a small share (approximately 1 percent of GDP) of public social spending is devoted to social assistance. Excluding pensions and other social security benefits, unemployment insurance is the main safety net program in Brazil, but it does not reach the poor in the informal sector. Most social safety nets (for instance, income-support programs, rural pensions, and old-age and disability benefits) are targeted and in general pro-poor. Some subnational governments have their own social assistance programs. In November 1998, 22 core social programs were identified in coordination with the Inter-American Development Bank (IDB) and the World Bank to be preserved from cuts in the ensuing period of fiscal adjustment. More recently, social assistance policies are being integrated into broader human development initiatives, consisting of focusing social policies and outlays on existing social programs in the states and municipalities with human development indices below the national median.

Prior to the crisis, the record of poverty reduction in Indonesia was impressive—the poverty rate declined from 60 percent in 1970 to below 8 percent in mid-1997. The crisis delivered a sharp setback to this trend. In the depths of the crisis, the poverty rate is estimated to have risen to more than double its pre-crisis level. To protect the poor from the effects of the crisis, the government launched a series of social safety net programs, including the subsidized rice scheme, the school scholarship scheme, a community-based public expenditure program, and an employment-generating public works scheme. The performance of these schemes was mixed, with the rice subsidy and scholarship schemes being the most successful. However, evidence indicates that informal social safety nets (that had been operating for years, albeit never under such difficult circumstances) played a more important role during the crisis than government-sponsored programs. While the poverty rate has declined from the level reached during the crisis, poverty vulnerability is still an important issue in Indonesia.

The social safety net in Turkey was underdeveloped when the crisis hit. However, several mechanisms were put in place to reduce the social impact of the crisis. In late 2000, an unemployment insurance program was launched, and payouts commenced in early 2002; direct income support for small farmers was introduced, cushioning the blow from reductions in untargeted subsidies; and spending on education and health care for the poorest members of society was increased through the Social Risk Mitigation Project, partly financed by a World Bank loan. Overall, social spending held up well, and at 15.7 percent of GNP both exceeded the target in the World Bank program (15.3 percent of GNP) and was 1.1 percent of GNP higher than in 2000.

An important area of expenditure reform relates to social safety nets, which take on added significance during crises when deep recessions have a disproportionately large yet unpredictable impact on the most vulnerable. For this reason, close attention has been paid to establishing social safety nets or strengthening existing ones—as illustrated in Box 6.2, which describes the experience in Brazil, Indonesia, and Turkey—and to protecting social spending. However, much remains to be done before effective social safety nets can be claimed to be in place.

Institutional Reforms

Improving the quality of fiscal adjustment also has an important institutional dimension, since fiscal policy is likely to be most effective when it is bolstered by reforms to strengthen fiscal management capacity.

Fiscal Policy Formulation and Implementation

The success of tax and expenditure policy reform is closely linked to the quality of revenue administration and budget management, and few countries have made significant progress with reforming policies without undertaking fundamental institution building. Brazil and Russia have benefited from moving on both fronts simultaneously. This contrasts with Argentina, where political consensus on needed policy and institutional reforms has not been achieved, as illustrated in Box 6.3.

One area where a number of crisis countries have made significant strides is in strengthening the framework for fiscal policy formulation and implementation. Particularly notable in this regard is Brazil, which has adopted a rules-based framework for fiscal responsibility that emphasizes medium-term fiscal targets (see Box 6.3). However, to be effective, rules have to be durable, which means that fiscal policy flexibility has to be limited to what is provided under the rules, rather than being the result of bending or breaking them. Thus, Argentina could not live by its fiscal responsibility law of 1999, or achieve its zero deficit plan of 2001, and both had to be abandoned.

Fiscal Policy Formulation and Implementation

The efficiency of tax administration in Argentina has declined sharply, and tax evasion has increased in recent years. This has reflected in part the impact of the economic crisis, but it has also been encouraged by the repeated use of tax amnesties and has been compounded by a proliferation of tax exemptions and special provisions in the tax system. A substantial simplification of the tax system is a prerequisite for effective tax administration reform, including a drastic reduction in tax exemptions, starting with remaining competitiveness plans. Other significant deficiencies of the tax administration also need to be addressed, with the emphasis on strengthening audit and the provision of full support for enforcement by the government and the judiciary. The budget process is complicated by the large share of revenue and expenditures taking place at the provincial level. There is no serious medium-term budgeting process. Subsidies to special interest groups have been maintained, and social spending, although significant, is poorly targeted.

Since the 1998 crisis, Brazil has achieved a high level of fiscal transparency and has increasingly relied on rules-based fiscal policymaking and management. The Fiscal Responsibility Law, enacted in May 2000, sets out fiscal rules aimed at ensuring medium-term fiscal sustainability for all levels of government, as well as stringent transparency requirements. The federal budget has also become more comprehensive and transparent and now includes most quasi-fiscal activities, a list of tax expenditures, an analysis of fiscal risks, and a list of ongoing and planned operations to securitize previously unrecognized liabilities. There are no extrabudgetary funds. Facilitated by very extensive use of the Internet, fiscal information is widely disseminated, including statistics, legislation, and administrative regulations on tax and budgetary matters, particularly at the federal level.

The capacity to formulate and implement fiscal policy in Turkey has been weak, primarily because significant fiscal activity is undertaken off-budget; the budget process lacks a medium-term context; budget management and revenue administration are highly fragmented; and the payments system is complex. Steps have been taken to address some of these issues, mainly in the context of improving fiscal transparency. In particular, budgetary and extrabudgetary funds were brought into the budget; quasi-fiscal operations were significantly scaled back (or directly compensated by the budget); a new law on debt management provides much greater power to control contingent liabilities; disclosure of contingent liabilities became the norm; accounting reforms, which would allow timely presentation of consolidated general government accounts, were advanced to the pilot stage; and efforts were made to place budget preparation within a better medium-term macroeconomic context. Further reforms to the budget process are proposed in a draft budget systems law, which will be considered by parliament in 2003. However, implementation of many reforms needs to be advanced more rapidly, and although revenue administration is being upgraded, fundamental reorganization has yet to be addressed.

Moreover, any rules-based framework must set a high standard for fiscal transparency. Indeed, independently of whether there are rules or not, transparency can reduce fiscal vulnerability in important respects. Transparency regarding contingent liabilities, quasi-fiscal activities, and off-budget operations in particular can reduce the chances of sudden adverse shifts in market perceptions of fiscal solvency and liquidity, and could play a significant part in encouraging longer-term lending to government. Many countries have taken steps toward, or are committed to, improving fiscal transparency—often in the context of meeting the requirements of the IMF’s Code of Good Practices on Fiscal Transparency (available on the Internet at http://www.imf.org/external/np/fad/trans/code.htm).

Fiscal policy performance can also be undermined by institutional fragmentation in budget management and revenue administration. A key consequence of fragmentation in Turkey has been that high-quality, durable adjustment measures have proved difficult to put in place. Thus, there has been a reliance on temporary revenue measures and across-the-board expenditure compression. Moreover, with a need for new measures to be put in place each year to maintain steady progress toward fiscal sustainability, there is a heightened risk of adjustment fatigue. To a lesser extent, fragmentation has also been a problem in Russia and Ukraine, leading some powerful ministries and politically influential large industries to avoid much of the burden of adjustment measures in the former case, and the preparation of unrealistic budgets in the latter case.

Intergovernmental Fiscal Relations

If fiscal federalism worked as theory would suggest, subnational governments would provide only those goods and services for which local provision is allocatively more efficient than central provision. Tax competition and financial market discipline would ensure cost efficiency. However, the reality is that these mechanisms do not work well in many emerging market economies with federal structures, mainly because subnational governments tend to rely heavily on central government transfers and do not face hard budget constraints. Hence, the capacity of central government to control general government finances, and therefore fiscal policy in general, is often compromised. A major challenge is getting subnational governments to contribute to required fiscal adjustment by the general government.20 In some cases, this can be achieved by reducing transfers to subnational governments, as in Russia, and by setting enforceable limits on subnational borrowing, as in Brazil (see Box 6.4). In the latter case, fiscal discipline at the subnational level is now grounded in the framework of fiscal responsibility legislation that applies to all levels of government. Federal transfers to subnational government in Mexico are now also strictly rules-based, which would allow the burden of adjustment to be shared in the event of a crisis.

Intergovernmental Fiscal Relations

Problems with intergovernmental fiscal relations are at the heart of the fiscal shortcomings in Argentina. Considerable pressure on public payrolls occurs at the provincial level. However, the constitutional autonomy of the provinces makes it impossible for the federal government to require them to adopt needed structural spending reforms. Moreover, under the current system, any one of the 23 provinces or the City of Buenos Aires can veto initiatives to reform revenue-sharing arrangements. That said, a reduction in federal transfers and firmly enforced limits on access to financing would force provinces to adjust and would have the virtue of leaving each to select an autonomous adjustment strategy. But this would only be effective if provinces could be prevented from issuing subcurrencies or bonds (a national agreement and law would be required) and if penalties were enforced for running arrears. A nobailout commitment by the central government would be a natural adjunct, but this would become credible only over time. The reduction in federal transfers would need to be enshrined in a permanent reform of the revenue-sharing system and accompanied by supporting institutional reforms.

Brazil is a very decentralized federation, and the subnational jurisdictions (states and municipalities) enjoy considerable autonomy in expenditure and tax policy (and their administration). Revenue-sharing arrangements between the central and subnational governments are formula-based and transparent but have not focused on the equalization of expenditure capacity among the subnational jurisdictions, particularly the states. Due to their autonomy in tax policy, there has been growing predatory tax competition among the states (and the municipalities, more recently) to attract investment, leading to significant erosion of their tax bases. The municipalities have increasingly played a more active role in service delivery, particularly in the social area. Subnational finances have improved remarkably since the mid-1990s, when debt-restructuring agreements were signed by the states and most large municipalities with the national treasury. Institutional reform, particularly after the enactment of the Fiscal Responsibility Law in 2000, has contributed to strengthen rules-based fiscal discipline at the sub-national level.

The major tax reform following the crisis in Russia led to a significant change in tax assignment for the federal and subnational governments. By 2001, VAT revenue accrued exclusively to the federal budget, while the personal income tax was fully assigned to sub-national governments. More recently, changes in tax sharing were extended to the profit tax, resource taxation, and excises. In the process, resources were increasingly being shifted away from subnational governments to the federal government. To make up for some of the resulting loss in revenue, subnational governments introduced a new regional sales tax and explored other local tax sources. While revenues were centralized, some federal expenditure responsibilities were devolved to subnational governments. To force adjustment at the subnational level, where expenditure programs were generally regarded as inefficient, the increase in federal budgetary transfers to regions was kept at a minimum. Instead, in support of the new arrangement, federal programs were put in place to help strengthen fiscal management and treasury operations at the subnational level. For instance, under a World Bank–supported program, subnational governments committed to fiscal reform stand to benefit from technical and financial assistance, while those that fail to adjust adequately could forfeit their control of local treasury operations to the federal government.

More often, however, given constitutional provisions and the realities of democratic government, the contribution of subnational government to fiscal adjustment has to be negotiated. This has been the case in Argentina, where repeated fiscal pacts have had to be negotiated to promote a degree of fiscal adjustment at the provincial level.

Debt Management

Improved debt management should be a component of a strategy to reduce the risk of debt crises, and in particular to ease liquidity concerns. Lengthening the average maturity of the debt and issuing more domestic currency debt are key to reducing vulnerability, albeit at the cost of higher contractual debt service, at least in the short term (which in effect can be considered as an insurance payment). The achievement of these objectives is likely to require significant time, since the market’s appetite for longer-term domestic debt can be expected to develop only after substantial policy credibility and confidence in fiscal solvency have been established. Moreover, wider reforms to develop financial markets and increase domestic saving may also be required. Success in these regards has been achieved in Brazil and Mexico, as explained in Box 6.5.

Debt Management

Brazil has followed a cautious debt-management strategy since the floating of the real in 1999. In an environment of considerable fiscal stress, the government’s debt-management strategy has aimed primarily to reduce refinancing risks and develop a secondary public debt market. Macroeconomic volatility, particularly since mid-2001, has hampered the government’s efforts to lengthen maturities and reduce the share of debt that is denominated in foreign currency and/or indexed to the exchange rate, which are perceived as the main sources of vulnerability in Brazil’s public debt dynamics. Debt management has also become more transparent. The treasury started in 2001 to publish its annual debt-management plan, with explicit targets for the overall domestic and external public debt stocks, average maturities, and the currency composition of new debt issuances. Auctions are now preannounced, less liquid security series have been consolidated and/or discontinued, and, since May 2002, as required by the Fiscal Responsibility Law, the central bank no longer issues debt in the domestic market. A code of conduct has also been issued for debt managers in an effort to strengthen governance.

Public debt management in Mexico has been successful in reducing the rollover and interest rate risks of domestic debt, as well as in lowering the share of external debt. Further accomplishments of external debt management include the development of a yield curve for sovereign bonds in international capital markets; improvements in the amortization profile; and the conduct of liability-management operations such as the Brady-bond buybacks and swaps. With regard to domestic debt, the authorities have succeeded in deepening the market for domestic debt instruments through regular issuances and the adoption of a system of market makers. As a result, the average maturity and duration of domestic securities has been extended, and the share of fixed-rate instruments has increased. Moreover, no new exchange-rate-linked domestic debt was issued after the phasing out of the Tesobonos, which has led to a very significant decline in the share of holdings by nonresidents. However, domestic debt remains subject to considerable refinancing and interest rate risks deriving from the still relatively large share of short-term and indexed debt (which account for 80 percent of the total).

Concluding Remarks

The preceding discussion suggests the following main conclusions.

  • It is clear that there has been little room for fiscal policy maneuver in emerging market economies when crises have hit because of concerns about solvency and liquidity. Such concerns have meant that the trade-off between adjustment and financing has inevitably had to emphasize the former.

  • Close attention should be paid to assessing debt sustainability on an ongoing basis, with emphasis on monitoring contingent and unfunded liabilities, and on understanding developments in, and the relationships between, the real exchange rate and other key macroeconomic variables. At the same time, there is a need to better link assessments of sustainability with those of vulnerability to solvency and liquidity crises.

  • Good fiscal policies are a must, the objective being to ensure that prudence during good times guards against the need for fiscal policy to be procyclical during bad times. Starting with a cyclically adjusted fiscal surplus and a low level of public debt would leave more room to be flexible in determining the fiscal stance during crises. This is desirable since the complexity and uncertainty that characterize many emerging market crises make it hard to predict their consequences. In addition, smoother fiscal policy responses are better for longer-term growth.

  • Where fiscal tightening is unavoidable, more attention should be paid to the quality and durability of adjustment measures. Short-term measures should be guided by clear medium-term objectives for tax reform and expenditure restructuring.

  • Although it is unclear whether fiscal tightening deepens recessions during crises and slows recovery, it is important to establish or strengthen social safety nets to mitigate the effects of both recession and adjustment measures on vulnerable groups.

  • Finally, attention should also be paid to the full range of institutional reforms that limit fiscal vulnerability and permit a more effective fiscal policy response in the event of a crisis. These relate to the capacity for fiscal policy formulation and implementation, including revenue administration and budget management capability; the soundness of intergovernmental fiscal relations; and the quality of debt management.

Appendix 6.1. The Output Gap and Fiscal Policy: A Case Study of Argentina21

When an economy experiences a deep and prolonged recession, it is particularly useful to disentangle those changes in the fiscal position that are a direct consequence of a low level of activity from those that reflect the discretionary fiscal policy of the government. This appendix tries to empirically distinguish cyclical from discretionary changes in the fiscal position in Argentina, using different techniques for measuring the size of the output gap and alternative fiscal policy indicators. Independently of the approach taken, a sizable discretionary fiscal tightening in 2002 more than offset a loosening due to the operation of automatic stabilizers at a time when the economy was contracting sharply.

The Output Gap

In recent years, Argentina has experienced an economic contraction of unprecedented magnitude. At the time of writing (early 2003), the indications were that there was an economic contraction of about 11 percent in 2002, which would be the sharpest single-year output decline in Argentina’s history and the fourth consecutive year of recession. On a cumulative basis, output would have contracted by about 18½ percent over the period 1998–2002, similar to the contraction experienced during World War I and twice as deep as that during the Great Depression of the 1930s.

Four different methods are used to measure the output gap implied by the recent contraction: a log-linear trend; a moving average; a Hodrick-Prescott filter; and Okun’s Law. The results for 2002 can be summarized as follows:

  • Log-linear trend. With an estimated long-term growth rate of about 3 percent, the output gap was 35½ percent in 2002 (Table 6.3 and Figure 6.3, top panel). However, the uneven growth rates of the past, and a high degree of autocorrelation, suggest that the assumption of constant growth may not be statistically valid.

  • Moving average. A (centered) moving average was used to allow for changes in trend growth. Using this technique, the output gap was estimated to be 10½–12 percent in 2002, depending on the number of years used for the moving average (Table 6.3 and Figure 6.3, middle panel).

  • Hodrick-Prescott (HP) filter. The HP filter, which minimizes the variation of output subject to a penalty for variations in the change of the trend component, yielded an estimated output gap of 14–15 percent in 2002, depending on the smoothing parameter used (Table 6.3 and Figure 6.3, bottom panel).

  • Okun’s Law. 22 With an unemployment rate estimated at 22 percent in 2002, and assuming a natural (or structural) unemployment rate of 8 percent (similar to the unemployment rate during the late 1980s), “excess unemployment” would be about 14 percent. Using an Okun coefficient of 2.3 (see footnote 22, and Figure 6.4, top right panel), eliminating this excess unemployment would generate additional output of about 32 percent, implying an output gap of about 24½ percent (i.e., [1/1.32]–1) in 2002 (Table 6.3 and Figure 6.4, top left panel). Sensitivity tests were carried out to see how estimated output gaps would be affected by different assumptions. Estimated output gaps for 2002 widen as the number of years used for the moving average is increased; the smoothing parameter (λ) for the HP filter is increased; and the Okun coefficient (θ) is increased (Table 6.3 and Figure 6.4, bottom left panel). A natural unemployment rate of 10 percent, as opposed to 8 percent, would reduce the Okun’s Law output gap to 21½ percent (Figure 6.4, bottom right panel).

Table 6.3.

Argentina: Measures of the Output Gap, 2002

(In percent)

article image
Source: IMF staff estimates and projections.

Projections based on the proposed budget for 2003.

Program projections.

Figure 6.3.
Figure 6.3.

Argentina: Potential Output and Output Gap

Sources: Instituto de Estudios sobre la Realidad Argentina y Latinoamericana; Argentina Ministry of Economy; UN, Economic Commision for Latin America and the Caribbean; International Monetary Fund, World Economic Outlook, and IMF staff estimates.
Figure 6.4.
Figure 6.4.

Argentina: Output Gap According to Okun’s Law

Source: IMF staff estimates.

Fiscal Policy Indicators

Over the past seven years, Argentina’s fiscal position has fluctuated within a narrow range, from a primary surplus of 0.6 percent of GDP (1998) to a primary deficit of 1.4 percent of GDP (2001). For 2002, notwithstanding the sharp contraction in economic activity, the government aimed to achieve a primary balance that would prevent an uncontrolled inflationary spiral in the absence of noninflationary financing. This implied a tightening of the primary balance by about ½ percent of GDP relative to 2001. However, this tightening is made up of two components: a loosening due to the operation of automatic stabilizers (primarily lower revenue) as the economy contracts, and a more-than-offsetting discretionary tightening. To fully understand fiscal performance in 2002, and more generally, a range of fiscal policy indicators can be used to distinguish between these two components.

Fiscal Stance and Fiscal Impulse

The fiscal stance and fiscal impulse indicators are commonly used to abstract from cyclical variations in the fiscal position.23 They do so by asking how the current fiscal position compares with the fiscal position at a time in the past when the fiscal situation was considered to be relatively normal, and the economy was operating close to potential (neither overheating nor in a recession). For the purpose of this exercise, 1996 was chosen as the base year.24 A neutral primary balance also has to be estimated for each year. The neutral primary balance is the difference between neutral revenue and neutral noninterest expenditure. Neutral revenue is calculated by assuming that revenue grows at the same rate as actual GDP from the base year, so that the revenue to GDP ratio remains constant at its 1996 level (i.e., 22.1 percent of GDP). Neutral noninterest expenditure is assumed to grow at the same rate as trend output (3 percent a year), so that the ratio of noninterest expenditure to trend GDP remains constant at its 1996 level (i.e., 23.2 percent of GDP) but changes relative to actual GDP.

The neutral primary balance is reported in Table 6.4. By construction, the neutral primary balance is the same as the actual primary balance in the base year, 1996 (i.e., –1.1 percent of GDP). The neutral primary balance improved somewhat during 1997–98 as the economy was growing faster than trend while noninterest expenditure grew with trend. This pattern is reversed in 1999–2002, when the neutral primary balance deteriorated continuously as the economy contracted. By 2002, the neutral primary balance had fallen to –8¼ percent of GDP.

Table 6.4.

Argentina: Fiscal Policy Indicators, 1996–2002

(In percent of GDP)

article image
Sources: Argentina, Ministry of Economy; and IMF staff estimates.

Positive indicates fiscal loosening and negative indicates fiscal tightening.

In percent of potential GDP.

Measured by the Hodrick-Prescott filter (λ = 100).

The fiscal stance is the difference between the neutral primary balance and the actual primary fiscal balance. When the primary balance is greater than the neutral primary balance, there is discretionary fiscal tightening. Conversely, when the primary balance is less than the neutral primary balance, there is discretionary fiscal loosening. As reported in Table 6.4, the fiscal stance is zero in 1996, since by construction the primary balance and the neutral primary balance are the same. For the whole 1997–2002 period, the actual fiscal balance is greater than the neutral primary balance, implying discretionary fiscal tightening.

A problem with the fiscal stance indicator is that it is a function of the choice of a base year, and it is therefore not a very robust fiscal indicator. The fiscal impulse, which is the difference in the fiscal stance, is generally regarded to be more robust. Table 6.4 points to a positive fiscal impulse in 1996 (as the economy was coming out of the Mexico crisis of 1995) and a negative fiscal impulse for the period 1997–2002. Figure 6.5, top left panel, shows a scatter diagram of the fiscal impulse and GDP growth. Redefining the origin of the coordinates to be the point of no fiscal impulse and trend growth, the northwest and southeast quadrants (with respect to that new origin) are associated with countercyclical fiscal policy, whereas the northeast and southwest quadrants are associated with procyclical fiscal policy. The negative fiscal impulse for the period 1997–2002 was significantly procyclical, and strongly so in 2002 when the economy was contracting sharply. The fiscal impulse estimate for 2002 is fairly robust to changes in underlying assumptions (Table 6.5).

Table 6.5.

Argentina: Sensitivity of Fiscal Policy Indicators, 2002

(In percent of GDP)

article image
Source: IMF staff estimates.
Figure 6.5.
Figure 6.5.

Argentina: Fiscal Impulse and Discretionary Fiscal Policy, 1996–2002

Source: IMF staff estimates.

Automatic Stabilizers and Discretionary Fiscal Policy

An alternative approach to decomposing changes in the fiscal position is to estimate automatic stabilizers directly, by assuming that tax revenue is related to income via a fixed tax elasticity, whereas expenditure is independent of income (i.e., there are limited unemployment and other social benefits that are related to income). Discretionary fiscal policy is then a residual.

The results based on an assumed revenue elasticity of 1.1 are shown in Table 6.4 and Figure 6.5, top right panel. In the absence of significant discretionary measures, the fiscal position in 2002 would have deteriorated by about 3¼ percent of GDP owing to the revenue lost because of economic contraction. The improvement in the primary balance for 2002 required offsetting discretionary measures amounting to about 4½ percent of GDP. A scatter diagram of discretionary fiscal policy and GDP growth is shown in Figure 6.5, bottom left panel; this is very similar to the top left panel of Figure 6.5, which points to the consistency of this approach with that based on the fiscal impulse. Although the calculations are sensitive to the choice of tax elasticity, assuming different tax elasticities does not alter the fact that automatic stabilizers and offsetting discretionary fiscal policy were both large in 2002 (see Table 6.5).

Full-Employment Primary Balance

The full-employment primary balance is the primary balance when the output gap is closed, and as such it also filters out cyclical variations in the fiscal position (Figure 6.5, bottom right panel). The estimates in Table 6.4 assume the output gap estimates of the HP filter, that revenue adjusts to the closing of the output gap with an elasticity of 1.1, and that expenditure does not respond to income. Table 6.4 indicates persistent full-employment primary deficits over the period 1996–2001. However, the full-employment primary balance for 2002 moved to a surplus of about 3 percent of full-employment GDP, indicating a discretionary tightening of fiscal policy by about 4 percentage points of full-employment GDP compared with 2001. The estimate of the full-employment primary surplus for 2002 varies with the output gap, but the surplus remains significant (see Table 6.5); a higher output gap would shift the full-employment primary balance line upward in Figure 6.5, bottom right panel.

References

  • Alesina, Alberto, and Roberto Perotti, 1997, “Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects,” IMF Working Paper 96/70 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Blejer, Mario, and Adrienne Cheasty, 1991, “Analytical and Methodological Issues in the Measurement of Fiscal Deficits,Journal of Economic Literature, Vol. 29 (December), pp. 164478.

    • Search Google Scholar
    • Export Citation
  • Burnside, Craig, Martin Eichenbaum, and Sergio Rebelo, 2001, “On the Fiscal Implications of Twin Crises,CEPR Discussion Paper No. 2918 (London: Centre for Economic Policy Research).

    • Search Google Scholar
    • Export Citation
  • Chalk, Nigel, and Richard Hemming, 2000, “Assessing Fiscal Sustainability in Theory and Practice,in Fiscal Sustainability, Banca d’Italia Research Department Public Finance Workshop; also available as IMF Working Paper 00/81 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Chand, Sheetal, 1993, “Fiscal Impulse Measures and Their Fiscal Impact,” in Mario Blejer Adrienne Cheasty, eds., How to Measure the Fiscal Deficit (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Chand, Sheetal, and Albert Jaeger, 1996, Aging Populations and Public Pension Schemes, Occasional Paper 147 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Ghosh, Atish, 2002, IMF-Supported Programs in Capital Account Crises, Occasional Paper 210 (Washington: International Monetary Fund).

  • Giavazzi, Francesco, and Marco Pagano, 1990, “Can Severe Fiscal Contractions be Expansionary? Tales of Two Small European Countries,” in NBER Macroeconomics Annual 1990, ed. Olivier J. Blanchard Fischer Stanley (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation
  • Hausmann, Ricardo, 2003, “Unrewarded Good Fiscal Behavior: The Role of Debt Structure,” in Rules-Based Fiscal Policy in Emerging Market Economies, ed. George Kopits (New York: Palgrave Macmillan, forthcoming).

    • Search Google Scholar
    • Export Citation
  • Heller, Peter, Richard Haas, and Ahsan Mansur, 1986, A Review of the Fiscal Impulse Measure, Occasional Paper 44 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Hemming, Richard, Michael Kell, and Axel Schim-melpfennig, 2003, Fiscal Vulnerability and Financial Crises in Emerging Market Economies, Occasional Paper 218 (Washington: International Monetary Fund, forthcoming).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 1998, World Economic Outlook, World Economic and Financial Surveys (Washington, October).

  • International Monetary Fund, 2002, “Assessing Sustainability,” Public Information Notice 02/69, July 11, 2002; available on the Internet at http://www.imf.org/external/np/sec/pn/2002/pn0269.htm.

    • Search Google Scholar
    • Export Citation
  • Okun, Arthur, 1962, “Potential GNP: Its Measurement and Significance,” reprinted in Joseph Pechman, ed., Economics for Policymaking (Cambridge, Massachusetts: MIT Press, 1983).

    • Search Google Scholar
    • Export Citation
  • World Bank, 2001, “Mexico: Fiscal Sustainability,” Report No. 20236-ME (Washington).

1

The section is based on material provided by Max Alier, Nigel Chalk, Adrienne Cheasty, Dale Chua, Luis Cubeddu, James Daniel, Luiz de Mello, Andrew Feltenstein, Mark Flanagan, Rex Ghosh, Eva Jenkner, Vincent Moissinac, Rolando Ossowski, and Martin Petri. The debt sustainability analysis was undertaken by Christina Daseking and Bikas Joshi. Rich Kelly and Erwin Tiongson also provided inputs.

2

Dating debt crises is not always straightforward. For Indonesia, crisis year data refer to FY 1998/99 (April 1–March 31), although the crisis started in late 1997. For Ukraine, the crisis was well under way in 1998 (reserves reached very low levels, the currency was devalued, debt was restructured, and the economy contracted). However, the threat of default came in 1999, and hence 1999 was chosen as the crisis year.

3

The last have tended to be limited to unfunded public pension liabilities, mainly in advanced economies with aging populations (see Chand and Jaeger, 1996).

4

See Chalk and Hemming (2000) for a fuller discussion and illustrations of this approach.

5

The confidence intervals are the upper and lower bounds generated by a 2 standard deviation adverse shock to key parameters, a 1 standard deviation shock to all parameters, and an assumed 30 percent real depreciation.

6

For further information about recent IMF work on debt sustainability, see IMF, Public Information Notice 02/69, July 11, 2002, available on the Internet at http://www.imf.org/external/np/ sec/pn/2002/pn0269.htm.

7

In addition to bonds issued in recognition of public pension liabilities with a shift to privatized pension systems, bonds were also issued in Argentina in lieu of unpaid pensions.

8

Based on an estimate of end-year debt valued at the average exchange rate for the year.

9

Although Indonesia is not included in the projections discussed here, it is clear that debt projections in this case would also be way off the mark, most notably because of the 58½ percent of GDP spent on bank recapitalization during 1998–2000, but also because of a sharp real depreciation and the costs of decentralization.

10

Critical thresholds for fiscal variables are derived from EWS models of debt crises by Hemming, Kell, and Schimmelpfennig (2003). An improved EWS model of debt crises incorporating fiscal indicators is being developed by IMF staff.

11

See Hausmann (2003) for a comparison of debt structures in Latin American and OECD countries, and a discussion of the influence of debt structure on fiscal policy.

12

The new debt sustainability framework reports gross financing in the baseline projection, but this falls short of what is required to signal possible liquidity crises. As regards the predictive capability of financing variables, the Goldman Sachs EWS model of currency crises includes short-term financing as an explanatory variable. Financing variables are being considered as part of the IMF work on an improved EWS model of debt crises.

13

Note that even where debt does not appear high relative to GDP compared with highly indebted OECD countries, it tends to be much higher relative to revenue given the relatively low ratio of revenue to GDP in emerging market economies. See Hausmann (2003) for an illustration and discussion of this point.

14

Note, however, that the impact of a real exchange rate depreciation could go either way, depending on the composition of revenue and primary spending; in particular, it can contribute significantly to measured fiscal tightening, as Burnside, Eichenbaum, and Rebelo (2001) find was the case in Korea during the Asian crisis.

15

The primary balance is a better indicator than the overall balance in this regard, because the latter includes interest payments that are nondiscretionary. The overall balance would be an especially misleading fiscal policy indicator where interest payments are subject to wide swings, as happens during crises. For a further discussion of fiscal balance measures, see Blejer and Cheasty (1991).

16

Precisely separating the contribution of automatic stabilizers from that of discretionary measures during crises is difficult when macroeconomic variables affecting fiscal aggregates change markedly, behavioral parameters may be shifting, and what constitutes discretionary policy can become unclear. See IMF (1998) for a discussion of these issues in the context of the Asian crisis.

17

The appendix to this section presents a detailed analysis of the output gap and fiscal policy in Argentina using different techniques for measuring the size of the output gap and alternative fiscal policy indicators.

18

Lower oil prices were a source of additional fiscal pressure in Russia in 1998, while higher oil prices contributed to the subsequent fiscal adjustment.

19

Episodes of non-Keynesian “expansionary fiscal contractions” have been singled out in certain heavily indebted advanced economies (Giavazzi and Pagano, 1990; and Alesina and Perotti, 1997).

20

This is not only a problem for emerging market economies. It was also a challenge for EU-area countries in meeting the requirements of the Maastricht Treaty in the run-up to monetary union and has been an obstacle to subsequent adherence to the Stability and Growth Pact.

21

Prepared by Alejandro Santos.

22

Okun’s Law is a statistical relationship between output growth and changes in the unemployment rate (see Okun, 1962). For the period 1997–2002 (a period of considerable variance in output and unemployment), the average ratio of output growth to changes in the unemployment rate was 2.3.

23

Heller, Haas, and Mansur (1986) and Chand (1993) describe in detail the fiscal stance and fiscal impulse methodology.

24

The year 1996 was chosen because the economy had recovered from the Mexico crisis, and it was before the Asian crisis produced major contagion.

  • Collapse
  • Expand
Recent Experience and Lessons for Latin America