Information about Western Hemisphere Hemisferio Occidental

Chapter 4. Global Large Debt Reduction: Lessons for the Caribbean

Charles Amo Yartey, and Therese Turner-Jones
Published Date:
May 2014
  • ShareShare
Information about Western Hemisphere Hemisferio Occidental
Show Summary Details
Charles Amo-Yartey and Joel Chiedu Okwuokei

The global financial crisis has led to renewed interest in the issue of debt reduction for many governments. Low economic growth, low budgetary revenues, and stimulus spending to prop up economic activity have resulted in a sizable accumulation of debt, especially by the developed world. For instance, the ratio of general government debt to GDP increased from 50 percent in 2007 to 90 percent during the crisis in advanced economies. In the Caribbean, the ratio of public debt to GDP increased by about 15 percentage points between 2008 and 2010.

Rising debt not only raises the risk of a fiscal crisis but also imposes costs on the economy by keeping borrowing costs high, discouraging private investment, and constraining fiscal flexibility. Empirical evidence points to a nonlinear relationship between public debt and growth, suggesting that public debt beyond certain levels can have negative effects on economic activity. Greenidge and others (2012) address the use of threshold effects between public debt and economic growth in the Caribbean. Their results show that, at debt levels lower than 30 percent of GDP, increases in the debt-to-GDP ratio are associated with faster economic growth. However, the effect on growth diminishes rapidly as debt rises beyond 30 percent of GDP, and in fact beyond 55 percent of GDP debt becomes a drag on growth.

This chapter investigates the determinants of global large debt reduction using a data set that spans more than four decades for a large sample of developed and developing economies. The analysis uses a panel data set of about 160 countries to estimate the probability that a large debt reduction will be initiated using the logit regression approach. More specifically, the chapter attempts to answer the following questions: How have large debt reductions occurred in practice? What factors determine the probability of a large debt reduction? And, what lessons can we draw for future large debt reduction in the Caribbean?

Most studies in the literature have examined debt reduction within the context of fiscal consolidation (Giavazzi and Pagano, 1990; and Alesina and Perotti, 1995). According to this approach, only public debt reductions resulting from sizeable improvements in either the primary balance or the cyclically adjusted primary balance were considered. This approach ignores other potential determinants of public debt reductions, such as business cycle developments and the magnitude of debt servicing costs. Using a sample of 15 European Union countries for the period 1985–2009, Nickel, Rother, and Zimmermann (2010) analyzed the determinants of large debt reductions. Their main findings were that major debt reductions are primarily driven by strong GDP growth and decisive and lasting fiscal consolidation efforts focused on reducing government expenditure, particularly by cuts in social benefits and public wages. This chapter extends the work of Nickel, Rother, and Zimmermann by analyzing the factors determining large debt reductions using a large data set of 160 countries for the period 1970–2009.

Our result shows that global large debt reduction is associated with robust economic growth, decisive and lasting fiscal consolidation, and a favorable external environment characterized by strong global growth. Fiscal rules are found to be positively associated with a higher probability of debt reduction because they increase fiscal discipline and the credibility of fiscal policy and help secure the gains of fiscal consolidation. The initial level of debt and debt servicing cost appears to play a disciplinary role by enhancing the incentives of governments to consolidate aggressively. The results are robust to alternative estimation methodologies and alternative thresholds for identifying large debt reduction episodes. We conclude that future large debt reduction programs in the Caribbean need to be based on credible fiscal plans to increase primary balances, fiscal rules to enhance fiscal credibility, and structural reforms to promote growth.

Literature Review

Theoretically, countries have a number of options to reduce their debt levels, including economic growth, fiscal consolidation, inflation, debt restructuring and defaults, and privatization. Their pluses and minuses can be summed up this way (IMF, 2003):

  • Reducing debt through economic growth would generally be the preferred option of policymakers, but growth is virtually nonexistent in the Caribbean at present.
  • Reducing debt through explicit defaults entails reputation costs that could influence future borrowing and constrain fiscal policy.
  • High inflation has enormous growth and welfare costs, while privatization, though debt reducing, does not change the net worth of the government.
  • Reducing debt through fiscal consolidation maintains the credibility of the government, but it is often politically difficult, and the gains need to be maintained over a long period of time.

The Size and Duration of Debt Reductions

What constitutes a large debt reduction episode? There is no standard definition of what constitutes a large reduction in the public-debt-to-GDP ratio. Existing literature explaining the dynamics of these reductions has analyzed episodes in terms of the magnitude of debt reduction over a specific period of time. Nickel, Rother, and Zimmermann (2010) define an episode as one in which the debt-to-GDP ratio declines by more than 10 percentage points in 5 consecutive years. In their analysis, Finger and Sadikov (2010) consider cases associated with debt-to-GDP ratio reduction of more than 20 percentage points. The IMF (2003) identifies cases in which this debt ratio was reduced by at least 18 percentage points over a three-year period. In a more recent study, the IMF (2012) analyzes cases in which the ratio exceeded the 100 percent threshold and discusses the debt dynamics 15 years following the event. Bandiera (2008) considers debt reductions on the order of 30 to 190 percentage points of GDP measured in net present value (NPV) terms. Baldacci, Gupta, and Mulas-Granados (2012) examine a debt-to-GDP ratio decline from a high level to a prudent threshold of 60 percent of GDP for advanced economies and 40 percent for emerging economies.

Examining a sample of 15 European Union countries during 1985–2009, Nickel, Rother, and Zimmermann (2010) find that the total debt reduction per country, on average, stood at about 37 percent of GDP over a relatively long period, ranging from 5 to14 years. In a large sample of countries covering 1980–2010, Baldacci, Gupta, and Mulas-Granados (2012) identify 104 episodes of public debt reduction, excluding debt relief cases lasting 2 to 13 years. Half of these episodes achieved a debt-to-GDP reduction of at least 20 percentage points, while in more than a third debt reduction was higher than 40 percentage points of GDP. The IMF (2012) identifies 26 episodes in advanced countries spanning almost 100 years; it finds that more than half of the 22 countries studied had at least one high-debt reduction episode between 1857 and 1997.

In its earlier study, the IMF (2003) identifies about 26 episodes of public debt reduction in emerging market economies during 1970–2002, of which 19 were associated with a default or restructuring. Excluding the restructuring cases, the IMF further finds that the median decline in the public debt ratio was 34 percent of GDP over three years. According to Baldacci, Gupta, and Mulas-Granados (2012), it took about 6 years, on average, during which public debt was reduced by about 29 percent of GDP. A key message that emerges from these experiences is that debt reduction takes time. It is even more difficult, and sometimes impossible, for countries emerging from a crisis. Reflecting higher uncertainty and the difficulty of fiscal adjustment in a postcrisis environment, Baldacci, Gupta, and Mulas-Granados (2010) find that lowering public debt takes about 10 years, on average, for a successful episode.

Factors that Contribute to Debt Reduction

There are diverse views in the literature about what has helped reduce debt substantially in various countries. Evidence suggests that large debt reductions are associated with a combination of factors, and the contribution of the respective factors to debt reduction differs. Findings in the literature point to the following factors: real GDP growth, fiscal adjustment, inflation or hyperinflation, interest rate, real exchange rate changes, debt restructuring or default, debt relief, and sale of government assets. On top of these, some studies have argued for an appropriate policy mix and deep-rooted structural reforms to boost competitiveness and growth.

In general, findings give prominence to robust real growth and fiscal consolidation.

The Role of Growth

Growth is mostly helpful as it allows countries to easily “grow their way” out of debt. However, the growth factor is often missing, especially in many advanced economies today, and in such cases, countries have relied on others factors to reduce public debt significantly. In emerging market economies, Baldacci, Gupta, and Mulas-Granados (2012) find that in addition to higher growth, the bulk of debt reduction in the period 1980–2010 was driven by lower interest rates. Over the period 1970–2002, strong growth performance averaging 8.5 percent of GDP and fiscal consolidation largely through expenditure restraint contributed to debt reduction in these economies (IMF, 2003).

In advanced economies, the correlation between growth and debt reduction is less clear (IMF, 2012). However, excluding hyperinflation cases, a relatively stronger growth performance is associated with debt reduction. Finger and Sadikov (2010) note that while favorable debt dynamics played a lesser role in reducing public debt in advanced economies, countries with already high debt levels tended to rely more on favorable macroeconomic conditions than on fiscal adjustment, and the episodes were less sustained.

In a study of middle-income countries, Reinhart, Rogoff, and Savastano (2003) find only one debt reduction case solely associated with real GDP growth. Badiera (2008) finds that GDP growth was the main factor in all large debt reduction episodes in seven low-income countries. Results from the World Bank (2005) indicate that primary fiscal surpluses and real GDP growth contributed to the reduction in the average public-debt-to-GDP ratio in all 15 countries covered.

The Role of Fiscal Consolidation

A number of studies have attributed large debt reductions to fiscal consolidation, implying that any serious plan for a major debt reduction must include a credible fiscal consolidation strategy. Indeed, evidence suggests that fiscal adjustments have been needed to reduce debt-to-GDP ratios to prudent levels and that these past episodes serve as a lesson for countries facing high debt levels today. Nickel, Rother, and Zimmermann (2010) and Amo-Yartey and others (2012) find that fiscal consolidation efforts must be decisive and call for reductions in government expenditure, in particular in wages and transfers. IMF (2012) lends support to this, arguing that successful debt reduction requires fiscal consolidation and a policy mix that supports growth.

Evidence from Baldacci, Gupta, and Mulas-Granados (2012) indicates that about half of the decline in public debt has emanated from stronger primary balances, especially in Africa. In oil exporting countries, the primary balance contribution to debt reduction was greater than the contribution from favorable debt dynamics, reflecting revenue booms following improvements in terms of trade (Finger and Sadikov, 2010). Additional findings from Finger and Sadikov suggest that middle-income countries have relied more on favorable debt dynamics than on fiscal consolidation. Yet significant fiscal adjustment has been needed to set off the debt reduction episode. In their analysis of major episodes of large debt changes in a sample of 19 advanced economies over 1880–2009, Ali Abbas and others (2011) find that debt reduction was accounted for by the primary balance and the growth-interest differential components in roughly equal proportions.

While fiscal consolidation is favored in the absence of growth, Bandiera (2008) notes that it was not a key factor for the seven cases he studied. No country was able to run fiscal surpluses consistently over time. Similarly, most countries covered in Finger and Sadikov (2010) were noted to have run primary deficits of about 2 percent of GDP on average over the debt reduction episode. In these cases, the countries involved benefited from debt restructuring and relief that contributed substantially to debt reduction. For the low-income cases, the debt relief option appeared unsustainable in the absence of fiscal and structural reforms and a good debt management policy, as the countries soon found themselves back in rapidly increasing debt, which reinforces the earlier finding by the IMF (2003) that unless accompanied with sound macroeconomic policies, default is not a long-term solution for a large debt problem. Indeed, evidence suggests that debt reduction via restructuring or default is a common phenomenon even in emerging markets and several advanced economies, particularly for external debt (Reinhart and Rogoff, 2011).

Other Factors

Turning to other factors, research on the relationship between inflation and debt reduction is inconclusive. One view is that generally, inflation is not associated with debt reduction, apart from exceptional cases of hyperinflation (IMF, 2012). Citing the experience of the United States after World War II, the IMF finds that high rates of surprise inflation combined with low nominal interest rates, largely reflecting financial repression, helped the United States to reduce debt significantly. Similarly, Finger and Sadikov (2010) finds no evidence of systemically higher inflation for countries with a lower share of foreign-currency debt. However, the IMF (2003) finds that in some emerging market cases, moderate inflation and exchange rate appreciation were helpful in reducing debt levels. For the serial debt defaulters, two-thirds of the cases involved real exchange rate appreciation (Reinhart, Rogoff, and Savastano, 2003).

The literature on debt reduction has also emphasized the appropriateness of the policy mix, sound macroeconomic policies, pro-growth structural reforms, and the need to have the right environment for a sustained debt reduction, such as supportive monetary conditions and a friendly external environment. An inappropriate policy mix—severe fiscal austerity and tight monetary policy—hurts growth and raises debt, as the IMF (2012) observes in the case of the United Kingdom after World War I. The view is that debt reductions are larger when fiscal measures are permanent or structural rather than temporary and when they are anchored in fiscal frameworks, including fiscal rules. The findings of Baldacci, Gupta, and Mulas-Granados (2012) also reflect the importance of combining supply-side structural reforms with policies to ameliorate debt dynamics.

Global Large Debt Reductions: Stylized Facts and Anatomy

In our analysis for this chapter, we define a large debt reduction episode as occurring if the debt-to-GDP ratio declines by at least 15 percent of GDP over five consecutive years. Using this definition, we recorded about 206 episodes of large debt reductions around the world between 1970 and 2009 (Figure 4.1). These are among the key findings:

  • About 100 (48 percent) of the debt reduction episodes were achieved through debt restructuring or default.
  • About 106 (52 percent) of the debt reduction episodes were achieved through higher GDP growth, higher inflation, or fiscal consolidation (see Appendix 4.1).
  • Of the debt reduction episodes achieved through fiscal consolidation, about 25 percent were preceded or accompanied by the existence of fiscal rules.
  • Most of the large debt reduction episodes lasted over a relatively long period of time, ranging from 4 years in Panama to 18 years in Australia. The average duration of large debt reduction episodes not achieved through debt restructuring was about 7 years.
  • The average decline in the debt-to-GDP ratio was 35 percent of GDP.

Figure 4.1Number of Global Debt Reduction Episodes by Year and Nature, 1970–2009

Source: Authors’ calculations.

Factors Behind the Decline in the Public-Debt-To-GDP Ratio

How have large public sector debt reductions occurred in practice? To answer this question, we used data for advanced, emerging market, and developing economies for the period 1970–2010. We identify cases where public debt was reduced over a five-year period by 15 percentage points, dropping cases in which the debt stock at the end of the period was still above the level three years prior to the event.

This section illustrates the potential drivers of large debt reduction. In particular, we examine the behavior of macroeconomic factors, as depicted in Figures 4.2 to 4.7, such as primary balance, GDP growth, government spending, government revenue, inflation, and the composition of public spending before and after the onset of a large debt reduction. Among our key findings are the following:

  • In the 106 cases in which large debt reduction was not due to a restructuring, the median decline in the debt-to-GDP ratio was 26.4 percent over a five-year period (Figure 4.2).
  • A strong economic performance seems to have contributed significantly to the reduction in the debt-to-GDP ratio. Real GDP growth starts to pick up one year before the event and averaged about 5 percent per year during the first five years of the debt reduction episode.
  • A strong fiscal effort appears to have played an important role in the debt reduction. The primary balance starts to improve significantly at least two years before the large debt reduction episode, and the improvement continues to be sustained during the first five years of the episode (Figure 4.3).
  • The fiscal improvements were due to a combination of revenue-enhancing measures and expenditure restraints. The median decline in the ratio of government spending to GDP was 3 percentage points of GDP over the five-year period (Figure 4.4).
  • The reduction in total spending came mainly from cuts in current spending, with capital spending remaining broadly flat over the five-year period (Figures 4.6 and 4.7).
  • Moderate inflation also contributed to the decline in the debt-to-GDP ratio, with inflation averaging 5 percent over the five-year period.

Figure 4.2Public Debt, 106 Global Study Casesa

(percent of GDP)

Source: Authors’ calculations.

a Consist of cases where public debt was reduced over a five-year period by 15 percentage points or more without restructuring.

Figure 4.3Primary Balance, 106 Global Study Cases

(percent of GDP)

Source: Authors’ calculations.

Figure 4.4Government Spending, 106 Global Study Cases

(percent of GDP)

Source: Authors’ calculations.

Figure 4.5Government Revenues, 106 Global Study Cases

(percent of GDP)

Source: Authors’ calculations.

Figure 4.6Current Spending, 106 Global Study Cases

(percent of GDP)

Source: Authors’ calculations.

Figure 4.7Capital Spending, 106 Global Study Cases

(percent of GDP)

Source: Authors’ calculations.

Public Debt Reduction: Tales From Seven Countries 1

This section analyzes the experience of seven countries that have succeeded in reducing their debt levels substantially during different periods of time: Brazil, Canada, Denmark, Lebanon, New Zealand, South Africa, and Vanuatu. The aim is to further examine the factors that promote debt reduction, including the specific measures implemented by these countries over the course of their debt reduction episodes. As with the analyses for the full sample, we analyze for these country cases, the behavior of public debt, growth, primary balance, government revenue, government spending and the composition of public spending before and during the debt reduction episodes.

Brazil, 2002–08

In a space of five years, covering 2002–07, the public-debt-to-GDP ratio fell by 15 percentage points from its peak of 80 percent in 2002. Over the debt reduction episode, real GDP expanded by 4 percent on average, while the primary-balance-to-GDP ratio was consistently above 3 percent of GDP (Figure 4.8). Brazil’s story reflects the country’s adherence to a well-established macroeconomic policy framework based on a high primary surplus objective, skillful debt management, inflation targeting, and a flexible exchange rate regime.

Figure 4.8Brazil: Public Debt, Real GDP Growth, and Primary Balance, 1999–2007

(percent of GDP)

Sources: National authorities; and authors’ calculations.

Debt reduction was also boosted by a favorable external environment: improved commodity prices and enhanced access to foreign financing. The country successfully implemented the 2002–05 IMF-supported programs that helped restore confidence and improve market conditions following its 1998–99 currency crisis.

Fiscal adjustment was essentially revenue based. The record high primary surpluses reflect strong revenue performances mainly through higher tax revenue at both the federal and subnational government levels. Expenditure measures included efforts to strengthen the social safety net and pension reform to reduce generous benefits. Current spending came down in 2002 following the start of consolidation but started rising in 2005, partly reflecting entitlement spending, while capital spending became flat. Fiscal adjustment appeared to have been achieved at the expense of a high tax burden and limited public investment.

A budget guideline law introduced in 2007 maintained the primary surplus target, eliminated ceilings on government revenues and expenditure set a year earlier, and outlined targets for reducing current expenditure. There was a debt restructuring agreement between the federal and the subnational governments and legislation limiting personnel expenditures and debt levels at all levels of government, paving the way for a comprehensive fiscal responsibility law.

Canada, 1997–2007

On the back of robust growth and fiscal consolidation, the public-debt-to-GDP ratio in Canada came down by 35 percentage points over an episode that lasted for 10 years starting from 1997 (Figure 4.9). Macroeconomic policy measures implemented since the early 1990s put the economy on a strong footing and underpinned the buoyancy of economic activities. The country has a fiscal framework that targets a budget balance over a rolling two-year period based on conservative fiscal assumptions.

Figure 4.9Canada: Public Debt, Real GDP Growth, and Primary Balance, 1993–2001

(percent of GDP)

Sources: National authorities; and authors’ calculations.

The idea is to build sufficient savings to stem future pressures on public finances associated with an aging population. This approach proved quite successful as it appeared to have provided the authorities with the flexibility to respond to changing circumstances while delivering exceptionally high fiscal surpluses and sustaining the social consensus to reduce the debt level.

Decisive fiscal consolidation, which started in the mid-1990s, was sustained. It saw government expenditure declining substantially, largely on account of lower public consumption, reduced unemployment benefits, less capital spending, and lower wage bill and transfers to provinces. Thus, with tax rates already high, fiscal adjustment had an expenditure focus, complemented by some revenue measures including higher excises and a broadening of both the personal income tax and corporate income tax bases. There were structural reforms in many key areas, including employment insurance, a public old-age pension scheme, education, and trade.

Fiscal consolidation was extended to lower-level governments. Within the period, provinces raised education and health fees and excises and broadened the corporate income tax base. With the return to fiscal surpluses, however, the government launched a five-year tax reduction plan in 2000, which significantly lowered statutory tax rates on personal income and corporate income, and it took steps to increase the contribution limits for tax-deferred retirement savings plans and cut capital gains taxes. Some provinces also lowered marginal income tax rates as well for both households and businesses.

Denmark, 1998–2007

Denmark cut the public-debt-to-GDP ratio by almost 40 percentage points over a period of nine years starting from 1998, reflecting buoyant economic activity and fiscal consolidation. Real GDP growth averaged 2 percent during 1994–2002, while primary surpluses averaged 4 percent of GDP (Figure 4.10). Prior to the debt reduction episode, the economic fundamentals seemed strong. The memory of bad times and the gains from earlier consolidation in the 1990s helped build a nationwide consensus about the importance of prudent fiscal policies. Fiscal adjustment consisted of a mix of revenue and expenditure measures, with emphasis on expenditure control, especially on transfers given the very high expenditure-to-GDP ratio. Caps on expenditure growth in real terms led to a gradual reduction in the expenditure-to-GDP ratio. To contain spending overruns, the counties were legally required to comply with budget targets beginning in 2003. At the same time, capital spending virtually disappeared.

Figure 4.10Denmark: Public Debt, Real GDP Growth, and Primary Balance, 1994–2002

(percent of GDP)

Sources: National authorities; and authors’ calculations.

As in the case of Canada, the government was committed to a strategy of running fiscal surpluses and lowering public debt to avert future spending pressures owing to changing demographics. To this end, fiscal policy has been explicitly guided by a medium-term fiscal framework since the early 1990s, initially motivated by the need to meet the European Union’s deficit targets, debt ceilings, and convergence programs.

There were also labor market reforms, including reductions in labor taxes and flexibility of hiring and dismissal, which resulted in significant increases in employment while maintaining high standards of social security for the unemployed. In 2006, the political parties reached a welfare agreement and included various measures aimed at later retirement and increasing labor supply in the short term.

Lebanon, 2006–10

Economic activity has been adversely affected by civil conflicts since 1990. However, real GDP growth rebounded in 2007 and registered 8½ percent during 2006–11. This was facilitated by a new and ambitious fiscal reform and financial support from donors for reconstruction, permitting the government to raise primary surpluses in the range of 1½ to 3 percent, thereby reducing the public-debt-to-GDP ratio by 45 percentage points in four years from a very high level in 2006 (Figure 4.11). The country benefited from Emergency Post-Conflict Assistance (EPCA) from the IMF in 2007 and 2008, which helped in improving macroeconomic stability and public finances.

Figure 4.11Lebanon: Public Debt, Real GDP Growth, and Primary Balance, 2003–11

(percent of GDP)

Sources: National authorities; and authors’ calculations.

Under a medium-term fiscal adjustment program adopted in 2007, referred to as the Paris III Agenda, a number of measures were planned for implementation in phases. On the revenue side, the value added tax (VAT) rate increased from 10 percent to 12 percent in 2008, and then to 15 percent in 2010 bringing it in line with the regional average. A global income tax was introduced in 2008, in addition to raising gasoline excises to their pre-capping levels of 2004. Tax on interest income was also raised, from 5 percent to 7 percent, in January 2008.

Other revenue measures included improvement in revenue from government properties and reforms in property tax administration. Revenue administration reforms included the introduction of a medium-sized taxpayer office and new audit procedures, and changes to the property tax evaluation system. On the expenditure side, measures included reforms that aimed at a nominal freeze in the wage bill and structural reforms in the energy and social sectors to curb fiscal leakages. Efforts to strengthen public financial management involved introducing medium-term planning. In 2008, a public debt directorate was created and a single treasury account was proposed.

However, some of the planned fiscal consolidation and structural reforms have been delayed repeatedly due to the difficult security situation and political tensions between the government and the opposition party, especially since 2010. Progress in introducing a global income tax, increasing the VAT rate, and implementing a single treasury account has been slow. Notably, however, the cabinet has approved the reform of the energy sector. Legislation has been passed for investments in the electricity sector, and reforms in tax administration and public financing management have been making progress. Despite the challenging circumstances, the government appears committed to pursuing fiscal consolidation and reducing debt further, including by implementing the key measures of the Paris III Agenda.

New Zealand, 1992–2007

Public debt fell from about 65 percent of GDP in 1992 to 17 percent in 2007 (Figure 4.12). During 1992–97, real GDP grew by about 4 percent on average while the primary balance averaged 5 percent. The economic expansion can be traced to radical structural reforms that commenced in 1985—institutional reforms that effected a strong medium-term orientation to monetary and fiscal policies to achieve macroeconomic stability; privatization and labor market deregulation to enhance competition and efficiency; and efforts to raise the productivity of core government services.

Figure 4.12New Zealand: Public Debt, Real GDP Growth, and Primary Balance, 1989–1997

(percent of GDP)

Sources: National authorities; and authors’ calculations.

Fiscal consolidation focused on limiting expenditures. For example, expenditure reduction accounted for approximately 40 percent of the improvement in the primary balance of about 1.4 percent of GDP in 2003. Caps on current expenditure led to a gradual reduction in the expenditure-to-GDP ratio (Figure 4.13). During the period, the government reiterated its commitment to the principles of medium-term budgeting and emphasized the need for higher savings in the light of future pension and health care obligations. Lower interest rates significantly reduced the cost of debt servicing, and a debt management office established in 1998 helped manage the public debt portfolio. As a result, the share of foreign-currency-denominated debt fell from 58 percent in 1992 to 22 percent in 2001. Moreover, a new monetary policy framework brought inflation down from an average of 8.3 percent during 1986–91 to 1.9 percent during 1992–97.

Figure 4.13New Zealand: Government Revenue and Spending, 1989–1997

(percent of GDP)

Sources: National authorities; and authors’ calculations.

The productivity of the public sector was substantially enhanced by reforms that placed department heads on performance contracts in return for flexibility to manage financial and human resource inputs. Building on the comprehensive public sector reforms, a Fiscal Responsibility Act was introduced in 1994, emphasizing the accountability and transparency of government fiscal operations. The tax system was also overhauled to broaden the tax base, lower marginal rates, and shift tax incidence from income to consumption. Most tax exemptions were abolished. The personal income tax brackets were reduced from five to two brackets, with the top rate of 33 percent the same as the company income tax rate. A VAT of 12½ percent replaced a range of indirect taxes.

The social welfare system was streamlined, while targeting was improved; eligibility requirements were tightened and the level of benefits was lowered for a broad range of programs. The national pension scheme was made partly income dependent, and its eligibility age was set to increase gradually from age 60 to age 65 over a 10-year period. A major labor market reform introduced a legal framework for a highly decentralized wage-bargaining system, so that individual contracts became prevalent in many sectors. This enhanced labor flexibility and remuneration, and moderated the spillover of wage pressure across firms and sectors.

In 2005, the incumbent party was reelected with a mandate for continued fiscal consolidation. Continuing with the medium-term budget framework that it had introduced in 1994, for example, after the elections in 2005 the government set out explicitly its long-term fiscal objectives, including reducing net public debt on a sustained basis to remain between 20 percent and 30 percent of GDP, restoring government net worth to positive levels, and reducing expenditure to below 30 percent of GDP.

South Africa, 1999–2008

The public-debt-to-GDP ratio came down by 19 percentage points over an episode that lasted for nine years commencing in 1999 (Figure 4.14). Real GDP growth averaged 3 percent from 1996 to 2004. Over the same period, the primary balance averaged 2.6 percent of GDP. South Africa undertook a wide range of structural reforms beginning in the early 1990s, which laid the basis for improvement in macroeconomic performance and in public finances many years later, including the period of debt reduction. Key structural reforms adopted at various times included trade liberalization, tax base broadening and lower rates for income taxes, revenue administration reforms, a medium-term budget framework, and expenditure planning and financial management. Income tax rates and import duties were cut as incentives for investment and job creation.

Figure 4.14South Africa: Public Debt, Real GDP Growth, and Primary Balance, 1996–2004

(percent of GDP)

Sources: National authorities; and authors’ calculations.

Revenue administration reforms covered the creation of a Revenue Authority in 1997 accompanied by a reorganization of revenue administration, modernization of the information system, and strengthening of audit and collection capacity. Public finance management legislation introduced in 2000 boosted financial accountability and improved internal controls at the national and provincial levels. On top of these reforms, there was an improvement in treasury control, and a centralization of personnel spending at the provincial level.

Fiscal consolidation focused on expenditure, with deliberate efforts to reduce the overall fiscal deficit. Expenditure measures targeted cuts in the wage bill and in subsidies and transfers and allowed a marginal increase in capital expenditures. Following fiscal adjustment, current and total expenditure were reduced while revenue improved, although the initial gains from expenditure reduction were later reversed.

The reduction of budget deficits over the years was accompanied by a reorientation of spending toward poverty reduction and social projects. The public financial management framework helped strengthened national and provincial capacity through provisions for multiyear budgeting and strategic planning. Beginning in 2000, state-owned enterprises, especially the four largest, which accounted for about 91 percent of assets of the top 30, restructured to increase their efficiency through improved governance and competition.

Vanuatu, 2002–07

After years of sub-par economic performance, Vanuatu’s economy is currently one of the fastest growing among all the small islands in the Pacific. Real GDP growth averaged 6 percent over the period 2003–2007. The strong growth performance can be explained by a booming tourism sector combined with foreign direct investment in real estate. At the same time, the country lowered its public debt by 23 percentage points of GDP in the five years following 2002 (Figure 4.15). Political stability, largely absent since the early 1990s, boosted confidence and has enabled the government to undertake structural and fiscal reforms. Macroeconomic management has improved and has created the right environment for economic activity to thrive.

Figure 4.15Vanuatu: Public Debt, Real GDP Growth, and Primary Balance, 1999–2007

(percent of GDP)

Sources: National authorities; and authors’ calculations.

The strong fiscal position in recent years is credited to the buoyant economic activity and important reforms, including a widening of the tax base, improvement in tax compliance, and strengthened expenditure control. Other reforms have included the introduction of a VAT, capacity strengthening at the tax office, strengthened regulation of the state-owned utility company, and fiscal transparency through more frequent financial reporting. In addition, the government liberalized the telecommunication sector in 2006 by offering a second telecommunication license. Furthermore, technical assistance to the central bank, especially from the Pacific Financial Assistance Centre, has helped build capacity and enhanced the transparency of the country’s monetary policy operations.

Estimating the Probability of a Large Debt Reduction


To examine the determinants of a large debt reduction, we use a data set spanning over three decades for a large sample of developed and developing economies. The analysis uses a panel data set of about 160 countries to estimate the probability that a large debt reduction will be initiated using the logit regression approach:

where Y is the log of the odds ratio, or more specifically the log odds of large debt reduction. The variable i stands for the ith country and t for the tth time period, αi is an idiosyncratic fixed effect which accounts for intercountry differences as long as these differences are constant over time. The explanatory variables Xit and Gt—representing macroeconomic variables and measures of fiscal rule, respectively—are measured either at the beginning of the previous five-year period or during the previous five-year period.

The analysis uses a panel data set of 160 countries for eight five-year periods (1970–74, 1975–79 …, 2005–09) to estimate the probability that a large debt reduction would be initiated in each five-year period using the logit regression approach. The logit is interpreted as follows: the slope coefficient measures the change in Y for a unit change in any of the explanatory variables, demonstrating how the log odds change as the explanatory variables change by a unit.2

The predicted probability of a large debt reduction can be computed using the estimated coefficient of the above regression:

The probabilities for hypothetical observations can be calculated by first finding the average values for all explanatory variables for a subset of countries and taking this to represent a typical country within the subset and then using the following formula:

However, when the dependent variable is observed as a qualitative variable and there are few time series observations per cross-section units and no autoregressions, fixed-effect models gives inconsistent estimates of the slope parameter. Andersen (1973) and Chamberlain (1980) argue that for large N and a small number of observations, the maximum likelihood estimation of the fixed-effects model gives inconsistent estimates of the parameters. They recommend the use of the conditional maximum likelihood (conditioning on the fixed effects). The main principle is to consider the likelihood function to be conditional on sufficient statistics for the incidental parameter αi (Maddala, 1987). In our logit model in equation (4.1), these sufficient statistics are Σtyit for αi. Maddala (1987) argues that for the logit model the conditional likelihood approach results in a computationally convenient estimator. The conditional maximum likelihood estimator of β is consistent, provided that the conditional likelihood function satisfies regularity conditions, which impose mild restrictions on the αi.

Chamberlain (1980) demonstrates that the standard errors obtained by the usual conditional logit programs can be used as the asymptotic standard errors for the conditional maximum likelihood estimator of β. In the conditional fixed-effects logit approach, alternative sets for which Σtyit=0 or Σtyit=T are discarded because they do not contribute to the likelihood function. In order to test for a fixed individual effect, one can perform a Hausman-type test based on the difference between the conditional maximum likelihood estimator and the standard logit maximum likelihood, ignoring country differences:

The test statistics are asymptotically χ2 distributed with k degrees of freedom.

The Data

The dependent variable is the probability of a large debt reduction (Debtred). The variable takes the value of 1 if a large debt reduction occurs and a value of zero otherwise. If a large debt reduction occurs in period t and continues in t+1, the value of Debtred is recorded as missing.

The explanatory variables are measures of fiscal consolidation, macroeconomic variables, political and institutional variables, and fiscal rules.

Fiscal consolidation: Fiscal consolidation is measured by the ratio of cyclically adjusted primary balance to potential GDP.

GDP growth: Real GDP growth is expected to be important in raising government revenues.

Inflation: Higher inflation could inflate the debt away, but it also has a significant negative effective on economic growth and welfare. Lucas (2003) estimates that the gains from completely removing the inflation rate of 200 percent are in excess of 5 percent of GDP in the long run.

Global economic conditions: We use global real GDP growth as a measure of global economic conditions. Many analysts believe that global economic conditions could influence the success of fiscal consolidation and debt reduction efforts.

Interest cost: This cost is measured as interest payments as a ratio to GDP. This measure is used to determine whether interest cost has a disciplinary effect on debt. High debt servicing cost could negatively affect growth and investment.

Fiscal rules: A dummy variable is used to capture fiscal rules. The dummy takes the value of 1 if a fiscal rule exists when the episode starts or during the episode and a value of zero otherwise. The literature suggests that fiscal rules are estimated to have affected several dimensions of fiscal consolidation and that the size of fiscal consolidation was significantly larger when fiscal rules were present. We also investigate whether the type of fiscal rules matters for a large debt reduction. To this end, we examine the impact of expenditure rules, debt rules, revenue rules, and a balanced budget rule in explaining the probability of a large debt reduction.

Estimation Results

First, it is important to examine some comparative statistics. Table 4.1 presents comparative statistics for large debt reduction countries and no large debt reduction countries during the sample period. The data shows that as compared with countries that did not experience a large debt reduction during the period, those countries that did experience a large debt reduction—

  • had a much higher debt
  • had much higher growth on average
  • had much better fiscal performance
  • had much lower inflation
  • had a much higher interest cost, and, finally
  • were more likely to have fiscal rules (in whichever form).
Table 4.1Comparative Statistics: Determinants of Global Large Debt Reduction
Large Debt ReductionNo Large Debt Reduction
MeanStandard deviationMeanStandard deviation
Debt reduction1000
GDP growth4.94.113.723.6
Primary balance1.625.3−0.0046
Interest cost3.83.12.772.2
Political risk6512.566.513.6
Fiscal rules0.260.440.150.36
Expenditure rule0.110.320.050.21
Balance budget rule0.250.430.130.33
Debt rule0.
Revenue rule0.040.20.0160.13
Source: Authors’ calculations.
Source: Authors’ calculations.

It is not surprising that the main finding from these comparative statistics is that countries that experienced a large debt reduction were on average able to achieve a higher GDP growth and larger primary surpluses and were more likely to have fiscal rules than countries that did not experience a large debt reduction.

Table 4.2 provides the logit estimates for nine different models. Model 1 is the standard logit regression for the data that are pooled over time, and model 2 is the conditional fixed-effects logit model. When looking at the regression results, it is important to note that the fixed-effect estimator does not use information provided by intercountry comparisons of debt reduction. Consequently, the probability of a large debt reduction is identified by countries that change debt reduction status during the period. In fact, in the conditional fixed-effect model, all countries with unchanged outcomes drop out of the conditional likelihood function.

Table 4.2Regression Results: Determinants of Global Large Debt Reduction(Dependent variable: the probability of a large debt reduction)
GLSRandom EffectsFixed EffectsFixed EffectsFixed EffectsFixed EffectsFixed EffectsFixed Effects
Interest cost0.18100.1672***0.18100.4803**0.22840.5500**0.6233**0.2510
Global growth0.9006*0.63040.9006*1.2181**0.9749*1.3896**1.4011**1.0819**
Primary balance0.1418**0.0561**0.1418**0.1535**0.1411**0.1600**0.1476**0.1394**
Debt-to-GDP ratio0.0572***−0.00000.0572***0.0483***0.0558***0.0519***0.0455***0.0560***
GDP growth0.1131**0.04770.1131**0.1271**0.1144**0.1396***0.1382***0.1166**
Fiscal rules3.0087***
Expenditure rule1.0147
Debt rule5.3949**
Balanced budget rule5.4797***
Revenue rule1.7689
Log likelihood−52.6156−153.5913−52.6156−45.9339−51.8868−44.0946−42.6553−51.6386
Source: Authors’ calculations.Notes: Standard errors in parentheses. ***, **, and * denote significance at the 1, 5, and 10 percent levels, respectively.
Source: Authors’ calculations.Notes: Standard errors in parentheses. ***, **, and * denote significance at the 1, 5, and 10 percent levels, respectively.

In our sample, we observed 217 countries that changed their debt reduction status at least once during the period 1970–2009. It is evident therefore that the number of informative observations is substantially lower than the total sample size, since the superior properties of the fixed-effects estimators in terms of bias need to be traded for less precise estimates in terms of higher standard errors. A comparison between the standard logit model and the conditional fixed-effects logit shows that the fixed-effect model performs better. A Hausman test statistic of 23.50 with a p value of 0.0014 leads to a rejection of the model without fixed effects.

The main results from the conditional fixed-effects logit model are that global large debt reductions are positively associated with strong economic growth, with a favorable external environment, with lasting fiscal consolidation, and with weaker initial conditions. The probability of a large reduction tends to increase when initial debt levels are high, since high debt levels tend to make fiscal consolidation needs more pressing in our sample of countries.

Strong economic growth also increases the probability of a large debt reduction, as the implementation of sound polices helps countries grow themselves out of debt. The results also show that global large debt reductions are driven by decisive and lasting fiscal consolidation. As expected, inflation does not contribute to major debt reductions and is actually negative and significant in the conditional fixed-effects logit specification. Debt servicing costs also play a disciplinary role, as high debt servicing costs are positively associated with the probability of a large debt reduction.

Our results also show that fiscal rules tend to increase the probability of a large debt reduction because they help strengthen the fiscal framework and improve fiscal transparency. Which types of fiscal rules are more successful in debt reduction? We found that debt rules and balanced budget rules are important in explaining the probability of such debt reductions. Fiscal rules based on revenue and expenditure do not appear to have any significant impact on the probability of a large debt reduction.

Robustness Tests

How robust are these results? The robustness of our results is tested by the following measures. First, we use a variety of estimation techniques. Second, we restrict the sample period to the period 1990–2009. Third, we exclude oil exporters from the full sample. And fourth, we use an alternative definition of a “large debt reduction” episode. Some of the results of our robustness analysis are presented in Table 4.3.

Table 4.3Robustness Tests: Determinants of Global Large Debt Reduction(Dependent variable: the probability of a large debt reduction)
Alternative Definition of Debt ReductionExcluding Oil Exporters
Model 1Model 2Model 3Model 4Model 5Model 6
Interest cost0.504300.534380.563490.292720.310310.29679
World growth1.588451.677171.575952.154482.189902.23625
Primary balance0.142280.138330.150330.340350.340800.33940
Initial debt0.045980.045550.048320.098700.099110.10102
GDP growth0.124650.125060.132260.337410.331950.34562
Fiscal rules2.242562.33547
Budget balance rule2.642132.52412
Debt rule3.536402.48407
Log likelihood−51.23484−51.1829−51.23484−33.83596−33.90281−34.11112
No. of observations231231231189189189
Source: Authors’ calculations.Note: Standard errors in parentheses. ***, **, and * denote significance at the 1, 5, and 10 percent levels of significance, respectively.
Source: Authors’ calculations.Note: Standard errors in parentheses. ***, **, and * denote significance at the 1, 5, and 10 percent levels of significance, respectively.

We examine whether our benchmark regression results are robust to changes in the definition of large debt reduction. We then define a large debt reduction as occurring when the debt-to-GDP ratio declines by at least 10 percentage points over a five-year period. Using this definition, we can identify 12 more episodes of global large debt reduction. We then estimate our benchmark model using the new definition of large debt reduction as the dependent variable. The results show that the estimated coefficients are largely unchanged, even though the standard errors, as expected, are larger in the model that uses fewer observations. Higher primary surpluses, strong global growth, robust real GDP growth, and fiscal rules are all positive and significant factors explaining the probability of global large debt reduction. We also exclude oil exporters from our baseline regression, and find that the results are statistically similar to our baseline regression results.

Summary and Conclusion

Highly indebted Caribbean countries should generally aim to lower their debt levels in order to reduce vulnerability and to create a better platform for growth. The results of this chapter show that major debt reductions are mainly driven by decisive and lasting fiscal consolidation efforts focused on reducing government expenditure.

Our analyses also show that robust real GDP growth increases the likelihood of a major debt reduction, because it helps countries grow their way out of indebtedness. Since growth in the current environment is virtually nonexistent, significant fiscal consolidation in the Caribbean is inevitable.

Fiscal consolidation in the Caribbean needs to be credible in order to anchor market expectations about fiscal sustainability. It is essential to strengthen the fiscal framework by adopting fiscal rules and independent fiscal agencies to guide budget processes and improve fiscal transparency. The literature on fiscal rules shows that when such rules are present, the size of fiscal consolidation is significantly larger and the consolidation efforts are sustained longer. The adoption of a spending rule on top of a budget balance rule helps in the achievement and maintenance of a primary balance that is sufficient to stabilize the debt-to-GDP ratio.

Caribbean countries could create a stable general fiscal rule to strengthen the current fiscal framework by defining the rule in terms of primary deficit for general government, which could take the form of expenditure ceilings and revenue floors. It is also essential to support the fiscal rule by creating an independent fiscal council to assess macroeconomic projections underlying the budgeting process and assess the compatibility of the fiscal framework with fiscal rules and general government policies.

Fiscal consolidation needs to be completed by a comprehensive policy to reduce public debt, including reforming tax policy, improving the efficiency of government spending, containing contingent liabilities, rationalizing the public sector, actively managing debt, restructuring debt, and making growth-enhancing structural reforms. Key areas for reforms include increasing labor market flexibility, achieving greater regional cooperation, and creating an enabling environment for private sector development.

Appendix 4.1
Appendix Table 4.1Episodes of Large Debt Reduction without Debt Restructuring
Public Debt (percent of GDP)
CountryPeriodDuration (Years)PeakTroughChange in debt ratioFiscal rules?
Equatorial Guinea2000–2008834.40.733.7Yes
Lao PDR1990–19955202.6118.5−84.2No
Lao PDR1998–20013187.6140.2−47.4No
Lao PDR2002–20086144.958.0−86.9No
New Zealand1992–2007564.617.4−47.2Yes
Papua New Guinea1987–19881162.142.1−20.0No
Papua New Guinea2002–2007562.632.9−29.7No
Saudi Arabia2002–2008696.913.2−83.7No
Slovak Republic2003–2008542.427.8−14.6Yes
Solomon Islands1991–1997667.441.2−26.2No
Solomon Islands2002–2010887.025.7−61.3No
South Africa1977–1984745.323.2−22.1No
South Africa1999–2008945.927.3−18.7No
St. Kitts and Nevis2005–20083168.1138.0−30.1Yes
St. Vincent and the Grenadines1994–1997357.041.3−15.7No
Trinidad and Tobago2002–2008658.724.1−34.6No
United Kingdom1970–1975573.246.7−26.6No
United Kingdom1997–2005857.942.1−15.8Yes
United States1993–2001872.454.7−17.7Yes
Source: Authors’ calculations.
Source: Authors’ calculations.

    Ali Abbas, S. M., NazimBelhocine, Asmaa ElGanainy, and MarkHorton, 2011, “Historical Patterns of Public Debt—Evidence from a New Database,” Presented at a Conference on Fiscal Policy, Stabilization and Sustainability, June 6–7, Florence, Italy.

    • Search Google Scholar
    • Export Citation

    AlesinaAlberto, and RobertoPerotti,1995, “Fiscal Expansions and Fiscal Adjustments in OECD Countries,” NBER Working Paper No. 5214 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    Amo-Yartey, Charles, MachikoNarita, Garth PeronNicholls, Joel ChieduOkwuokei, AlexandraPeter, and ThereseTurner-Jones, 2012, “The Challenges of Fiscal Consolidation and Debt Reduction in the Caribbean,” IMF Working Paper 12/276 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Andersen, Erling, 1973, Conditional Inference and Models for Measuring (Copenhagen: Mental-hygiejnisk Forsknings Institut).

    Baldacci, Emmanuele, SanjeevGupta, and CarlosMulas-Granados, 2010, “Getting Debt Under Control,” Finance and Development, Vol. 47, No. 4, pp. 18–21 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Baldacci, Emmanuele, SanjeevGupta, and CarlosMulas-Granados, 2012, “Reassessing the Fiscal Mix for Successful Debt Reduction,” Economic Policy, Vol. 27, Issue 71, pp. 365–406, July.

    • Search Google Scholar
    • Export Citation

    Bandiera, Luca, 2008, “Public Debt and Its Determinants in Low-Income Countries—Results From Seven Country Case Studies,” (Washington: World Bank). Available at SSRN:

    • Search Google Scholar
    • Export Citation

    Chamberlain, Gary, 1980, “Analysis of Covariance with Qualitative Data,” Review of Economic Studies, No. 47, pp. 225–38.

    Finger, Harald, and AzimSadikov, 2010, “Lowering Public Debt,” Finance and Development, Vol. 47, No. 2, pp. 36–38 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Giavazzi, Francesco, and MarcoPagano, 1990, “Can Severe Fiscal Contractions Be Expansionary? Tales of Two Small European Countries,” NBER Macroeconomics Annual, Vol. 5, pp. 75–111 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    Greenidge, Kevin, RolandCraigwell, ChrystolThomas, and LisaDrakes, 2012, “Threshold Effects of Sovereign Debt: Evidence from the Caribbean,” IMF Working Paper 12/157 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF), 2003, “Public Debt in Emerging Markets: Is It Too High?” in World Economic Outlook (September) (Washington).

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF), 2012, “The Good, the Bad, and the Ugly: 100 Years of Dealing With Public Debt Overhangs,” in World Economic Outlook (October) (Washington).

    • Search Google Scholar
    • Export Citation

    Lucas, E. Robert, 2003, “Macroeconomic Priorities,” The American Economic Review, Vol. 93, No. 1, pp. 1–14.

    Maddala, G. S., 1987, “Limited Dependent Variable Models Using Panel Data,” The Journal of Human Resources, Vol. 22, No. 3, pp. 307–38.

    • Search Google Scholar
    • Export Citation

    Nickel, Christiane, PhilippRother, and LilliZimmermann, 2010, “Major Public Debt Reductions: Lesson from the Past, Lessons for the Future,” Working Paper No. 1241 (Frankfurt: European Central Bank, September).

    • Search Google Scholar
    • Export Citation

    Reinhart, M. Carmen, and KennethS. Rogoff, 2011, “A Decade of Debt,” NBER Working Paper No. 16827 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    Reinhart, M. Carmen, KennethS. Rogoff, and MiguelSavastano, 2003, “Debt Intolerance,” Brookings Papers on Economic Activity, Vol. 2003, No. 1, pp. 1–62.

    • Search Google Scholar
    • Export Citation

    World Bank, 2005, Public Debt and its Determinants in Market Access Countries: Results from 15 Country Case Studies (Washington).


The section draws from IMF staff reports on the respective countries.


The odds in favor of a large debt reduction initiation are the ratio of the probability of a large debt reduction to the probability of “no debt reduction” in any given five-year period. The odds ratio is written mathematically as ρ / (1 − ρ).

    Other Resources Citing This Publication