Eastern Caribbean Currency Union: Selected Issues

The Eastern Caribbean Currency Union (ECCU) countries’ economies are heavily dependent on the United States for foreign direct investment, mainly in the tourism sector. The Selected Issues paper discusses economic development and policies of the ECCU. About one-third of the stayover tourists to the ECCU countries are from the United States., the top tourist-source country. The flow of remittances is also an important channel of influence, reflecting the significant proportion of Caribbean migrants living in the United States.

Abstract

The Eastern Caribbean Currency Union (ECCU) countries’ economies are heavily dependent on the United States for foreign direct investment, mainly in the tourism sector. The Selected Issues paper discusses economic development and policies of the ECCU. About one-third of the stayover tourists to the ECCU countries are from the United States., the top tourist-source country. The flow of remittances is also an important channel of influence, reflecting the significant proportion of Caribbean migrants living in the United States.

V. A Risk Analysis of Public Debt in the ECCU: A Fan Chart Approach 1

A. Introduction

1. Public debt-to-GDP ratios in the ECCU countries are among the highest in the world. Public debt in the ECCU has risen significantly since the mid-1990s, reflecting the region's proneness to natural disasters, other exogenous shocks, macroeconomic volatility, and large fiscal deficits. Public debt averaged about 100 percent of GDP at end-December 2008 and is set to increase during the current economic downturn.2 While the Monetary Council of the ECCB has set an objective for all countries to achieve a debt-to-GDP ratio of 60 percent by 2020, national fiscal policies have not yet been linked to this medium-term benchmark. This paper is motivated by the rising indebtedness of countries in the region. The objective is to illustrate the risks surrounding public debt dynamics related to macroeconomic volatility, and the role of strong fiscal policy adjustment that is responsive to shocks.

2. This paper applies a stochastic simulation algorithm to analyze debt sustainability in the ECCU. The standard framework used by the IMF and the World Bank to conduct debt sustainability analysis assesses uncertainty surrounding future macroeconomic conditions and fiscal policy through the use of single shock “bound tests”.3 An important shortcoming of this approach is the deterministic approach to assessing risks. It ignores both the correlation among shocks and the joint response of macroeconomic variables relevant for debt dynamics. A number of recent studies have attempted to address these shortcomings by introducing uncertainty into the analysis. Given the ECCU countries' proneness to natural disasters and macroeconomic volatility, this paper uses the stochastic simulation algorithm proposed by Celasun et al. (2006). The virtue of this approach is that it exploits the historical dynamics observed in the data to generate more likely shock scenarios, and recognizes the probabilistic nature of assessing long-term debt sustainability.

3. A necessary condition for debt to be sustainable is that primary fiscal balances react to the accumulation of public debt, with higher past debt leading governments to adjust by running larger primary balances. We find that, with the exception of Dominica and Grenada, fiscal reaction functions for countries in the ECCU are not well behaved in this sense.4 The estimated reaction functions also indicate strong fiscal policy inertia. Two alternative well-behaved fiscal reaction functions (estimated from different country samples) are used to generate scenarios for public debt risk profiles, illustrating how the responsiveness of fiscal policy to past debt affects prospects for debt sustainability in the ECCU. The first scenario utilizes an estimated fiscal reaction function for 14 countries in the Caribbean, and the second uses coefficients estimated by Celasun et al. (2006) for 5 middle-income countries (Argentina, Brazil, Mexico, South Africa, and Turkey). The results, together with the strong presence of policy persistence, underscore the need to improve the sensitivity of fiscal policy to public debt, in order to put medium-term debt dynamics on a path to achieve the debt-to-GDP ratio of 60 percent by 2020.

4. The paper is organized as follows. Section B surveys the literature and Section C describes the basic structure of the model. Section D illustrates the model's key simulation properties, including the fan charts. Section E provides concluding remarks.

B. Literature Review

5. The literature contains several operational definitions of debt sustainability. In the classical literature, if the real interest rate is less than the economy's growth rate, government deficits could continue ad infinitum without an increase in the debt-to-GDP ratio.5 By contrast, if the government borrows at an interest rate that exceeds the growth rate, the debt will rise unless compensated for by the primary surplus.6

  • The concept of government solvency is closely related to debt sustainability. For a government to be solvent, it must be able to service its debt obligations in perpetuity without explicit default. This, in turn, requires that the current debt not exceed the present discounted value of future primary surpluses. 7

  • The solvency concept is not very demanding, however, in that a government could satisfy the condition by running large primary deficits for some time, but promising a sharp adjustment and series of primary surpluses in the future, which may not be feasible.

  • Thus, the concept of debt sustainability typically requires more; i.e., that the government's intertemporal budget constraint is satisfied without an unrealistically large future correction in the primary balance, given the cost of financing (IMF, 2002). Operationally, sustainability assessments determine whether a projected plausible path of future primary balances implies a stabilizing or declining debt-to-GDP ratio over time.8

  • A third operational concept relates to debt thresholds, which could be of two types: (i) a debt-to-GDP ratio beyond which debt distress, default or crises are likely; and (ii) national or regional targets for debt-to-GDP ratios deemed to be sustainable, such as the ECCB's target of 60 percent debt-to-GDP ratios for ECCU countries by 2020.

6. In recent years, several studies have generated multivariate stochastic simulations of future debt trajectories, based on econometric models. 9 Garcia and Rigobon (2004), Hostland and Karam (2005), Celasun et al. (2006), Penalver and Thwaites (2006), and Tanner and Samake (2006) generate explicit probability distributions for projected debt paths that take into account the interaction among key variables that influence debt. One of the most prominent papers on debt thresholds is Reinhart, Rogoff, and Savastano (2003), who identify thresholds beyond which countries are susceptible to debt crises, which vary from country to country and depend importantly on history. To our knowledge, Di Bella (2008) is the only previous study to have used this type of approach in the Caribbean. He extends Reinhart, Rogoff, and Savastano's (2003) model to estimate a country specific debt threshold for the Dominican Republic.10

C. Methodology

7. This paper uses a stochastic simulation algorithm developed by Celasun et al. (2006). The methodology randomly generates a large sample of bound tests covering a range of likely shock combinations from which frequency distributions of the debt-to-GDP ratio are derived for each year of a projection, permitting an explicitly probabilistic assessment of debt sustainability. Projected paths and shocks to the economic variables are combined with a model-based projection for the primary balance, where the primary balance responds to economic shocks and to past debt levels. The virtues of the probabilistic approach are twofold. First, it depicts debt paths under realistic shock configurations (to growth, interest rates, exchange rates), derived from country-specific estimation of the correlation across shocks and the joint responses. Unlike standard DSAs, it recognizes the persistence of shocks through time, reflecting historical dynamic relationships. Second, the primary balance is endogenous, based on estimated fiscal reaction functions. For this application to the ECCU, the fiscal reaction functions play a different role than in Celasun et al. (2006), illustrating how the assumption of alternative fiscal policies would affect the probabilistic projection of debt paths.

8. The methodology encompasses three building blocks. 11

  • First, a fiscal policy reaction function is estimated. In line with the literature, it is specified as:

    bi,t=a0+αsi,t1+γygapi,t+Ci,tλ+ηi+ϵi,t,t=1,T,i=1,N(1)

    where bi, t is the primary fiscal balance-to-GDP ratio in country i and year t, si,t-1 is the stock of public debt-to-GDP ratio at the end of period t-1, ygapi,t is the output gap, Ci,t is a vector of control variables, ηi is an unobserved, country fixed-effect, and εi,t is an error term.12

  • Second, the joint distribution of shocks and forecasts of the nonfiscal determinants of public debt are calibrated based on historical country-specific data. For each country, the following unrestricted VAR model is estimated:

    Yt=γ0+ΣγkYtk+ξt(2)

    where Yt=(rtf,rt,gt,zt),γk

    article image
    is a vector of coefficients, and rf,r,g,z, and ξ, denote the real foreign interest rate, the real interest rate, the real GDP growth rate, the log of the real effective exchange rate, and a vector of well-behaved error terms: ξ ∼ N (0, Ω). Thus, as shocks occur each period, the VAR produces joint dynamic responses of all elements in Y.

  • Third, for each simulated constellation of shocks, quarterly VAR projections are annualized. The corresponding debt path is then calculated recursively using equation (1) and the conventional stock-flow identity:

    st[(1+rtf)1+gt](1+Δzt)st1*+[1+rt1+gt](1+rt)s~t1bt+ht,(3)

    where the total debt-to-GDP ratio, st, is the sum of foreign-currency, st*

    article image
    and domestic currency, st
    article image
    denominated debt and depends on the real cost of borrowing in foreign (rtf)
    article image
    and domestic (rt) currency, real GDP growth, gt, the real rate of currency depreciation, Δzt, primary fiscal surplus-to-GDP ratio, bt, and below-the-line expenditures (stock-flow adjustments) in percent of GDP, ht. 13

D. Data Analysis

Debt dynamics in the Caribbean

9. Governments concerned with solvency would be expected to run higher primary surpluses if debt in previous years has been increasing, in order to ensure that public debt does not explode. Bohn (1998) shows that a positive and significant coefficient on lagged debt in a regression explaining the primary surplus implies the consistency of fiscal policy with long-run solvency, and ensures that the debt ratio will revert to some long-run steady state value. This condition is necessary but not sufficient for ensuring debt sustainability, since as noted above, solvency in present value terms could rest on a promise of a future large adjustment.

10. How does fiscal policy in the Caribbean respond to past accumulation of public debt? This section first examines unconditional correlations between primary balances and public debt. These relationships may be misleading, however, as they do not control for key determinants of the primary balance such as output cycles and institutions. Estimated fiscal policy reaction functions can be used to explore how the endogenous response of fiscal policy—controlling for a range of determinants and country-specific effects—affects projected debt dynamics, and the risks to debt sustainability.

11. A casual examination indicates that the primary balance has been related to changes in the public debt-to-GDP ratio in the six ECCU countries. The scatter plot in Figures 1 shows that periods of high primary surpluses (deficits) are associated with decreases (increases) in public debt ratios. The association between the two variables is stronger for St. Kitts and Nevis and weaker for St. Vincent and the Grenadines, perhaps reflecting significant stock flow adjustments in the latter country. Figures 2 plots the path of lagged debt-to-GDP ratios against primary fiscal surplus-to-GDP ratios for each country, showing that in several of the countries (Dominica, Grenada, St. Kitts and Nevis) fiscal policy appears to have been responsive to developments in the debt-to-GDP ratio in recent years. The high variability in the data, especially for Antigua and Barbuda, indicates that both variables are subject to various shocks (e.g., fluctuations in GDP growth, and below-the-line movements). Next, econometric methods will determine whether the relationship holds conditioning on other determinants of the primary balance.

Figure 1.
Figure 1.

Change in Public Debt and Primary Surplus

(In percent of GDP)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A005

Sources: ECCB; and authors' calculations.
Figure 2.
Figure 2.

Primary Surplus and Lagged Public Debt

(In percent of GDP)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A005

Sources: ECCB; and authors' calculations.
Figure 3.
Figure 3.

Public Debt Under the First-Case Scenario, 2008–12

(In percent of GDP)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A005

Sources: Country authorities; and authors' calculations.
Figure 4.
Figure 4.

Public Debt Under the Best-Case Scenario, 2008–12

(In percent of GDP)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A005

Sources: Country authorities; and authors' calculations.

12. Fiscal reaction functions are estimated using panel data techniques, given the relatively short data period. The function relates the primary fiscal balance to lagged public debt, the output gap, real oil prices, and a measure of institutional quality, and in some specifications controls for inertia in fiscal policy by including the lagged primary balance.14 The equations are estimated using generalized least squares (GLS) random effects, and system generalized method of moments (GMM). The latter controls for endogeneity of the output gap and lagged debt, and corrects for the bias introduced by the lagged primary balance variable in the presence of country fixed effects.15

13. The estimation results indicate that fiscal policy in the ECCU as a whole does not respond systematically to the public debt-to-GDP ratio. The results are not consistent with the requirement for long-term debt sustainability—lagged debt is negatively and significantly correlated with the primary fiscal balance, suggesting that debt would not converge to a steady-state value (Table 1). However, Ordinary Least Squares (OLS) estimates show that Dominica and Grenada have behaved differently from the rest of the ECCU, perhaps reflecting the fact that they undertook debt restructuring in the early 2000s (Table 2).16 Information from the ECCU fiscal reaction functions cannot be used as an input to the stochastic fan-chart algorithm, since the necessary response of the primary balance to lagged debt is absent. As discussed below, this result and the evidence of strong fiscal policy inertia supports the case for institutional mechanisms to link national fiscal policies to the 60 percent of GDP regional target for public debt.

Table 1.

ECCU: Fiscal Reaction Function, 1984–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.
Table 2.

Antigua and Barbuda: Fiscal Reaction Function, 1986–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.
Table 2a.

Dominica: Fiscal Reaction Function, 1984–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.
Table 2b.

Grenada: Fiscal Reaction Function, 1989–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.
Table 2c.

St. Kitts and Nevis: Fiscal Reaction Function, 1984–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.
Table 2d.

St. Lucia: Fiscal Reaction Function, 1984–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics.
Table 2e.

St. Vincent and the Grenadines: Fiscal Reaction Function, 1990–2007 Dependent Variable: Primary Fiscal Surplus in Percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics.

14. Results are mixed for a fiscal reaction function using a wider sample of 14 Caribbean countries. In the GLS model the primary surplus exhibits a positive and statistically significant response to the lagged debt ratio, suggesting that fiscal policy conforms to this condition for long-term debt sustainability (Table 3). The coefficient on lagged debt remains positive in the system-GMM model, but loses significance, and the lagged primary surplus effect indicates strong policy inertia.

Table 3.

Caribbean Countries: Fiscal Reaction Function 1/

Dependent Variable: Primary Fiscal Surplus in percent of GDP

article image
Source: Authors' calculations.Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.

In addition to the ECCU, the sample includes The Bahamas, Barbados, Belize, Dominican Republic, Guyana, Jamaica, Suriname, and Trinidad and Tobago.

15. There is some support for the hypothesis that the fiscal response to past debt is stronger when the debt is below a particular threshold. An alternative specification considered in the literature includes a nonlinear reaction term to determine whether the primary balance reacts positively to lagged debt at low debt levels, but the relationship weakens or disappears at higher debt levels. The results of the nonlinear function including a debt spline indicate that in the GLS random effects estimation, the primary balance responds positively and significantly to lagged debt when debt is below 60 percent of GDP, but this relationship is not present for debt above this threshold17.

16. While there is no evidence for counter-cyclical fiscal policy, other potential determinants of the primary balance have the expected effects in the wider Caribbean sample. The estimated coefficient on the output gap depicts an acyclical behavior—the primary fiscal balance does not seem to react in any significant way to GDP shocks.18 Real oil prices have a significant negative effect, suggesting that higher oil prices translate into lower primary surpluses, possibly through higher government spending on fuel price sensitive goods and services. Estimated results suggest that, everything else being equal, countries with strong political institutions (low corruption, high bureaucratic quality, efficient law enforcement, government stability and high democratic accountability) generate higher primary surpluses.

17. In contrast to the ECCU results, fiscal reaction functions estimated in the literature find that emerging market countries tend to improve the primary balance when debt has increased. Celasun et al. (2006) estimate fiscal reaction functions for a set of five emerging market countries and find a positive response of primary surpluses to public debt in all specifications (Table 4).19 The magnitude of the coefficients is much higher than in the Caribbean country sample, indicating that fiscal policy responds more aggressively, tending to make corrections to moderate debt.

Table 4.

Selected Emerging Markets: Estimates of the Fiscal Reaction Function, 1990–2004

Dependent Variable: Level or Change in Change in the Primary Fiscal Balance

article image
Source: Celasun, Debrun, and Ostry (2006).Note: Brackets denote t-statistics; *, **, *** denote significance at 10, 5, and 1 percent respectively.

Stochastic Debt Fan Charts

18. The next step in the methodology estimates VAR models for the ECCU countries, to produce a joint distribution of shocks for the fan charts. Four-variable VAR models (domestic and foreign interest rates, real GDP growth, and real effective exchange rate) are estimated using quarterly data over the period 1986:Q1 to 2007:Q4. Table 5 provides the estimated VAR coefficients and correlation matrices of shocks for each country. The joint distribution of all nonfiscal shocks for the fan charts is calibrated using the estimated covariance matrix of the VAR residuals.20 The fiscal shock is assumed to be orthogonal to the other variables and is consistent with the standard error of the OLS regressions reported in Table 2. New macroeconomic disturbances occur every quarter and feed into VAR forecasts of the nonfiscal variables. After annualizing the results, the public debt and primary balance projections are determined recursively using equation (3). Annual frequency distributions for public debt are calculated on a sample of 1,000 repeated simulations.

Table 5.

Antigua and Barbuda: VAR Model

article image
Source: Authors' calculations.
Table 5a.

Dominica: VAR Model

article image
Source: Authors' calculations.
Table 5b.

Grenada: VAR Model

article image
Source: Authors' calculations.
Table 5c.

St. Kitts and Nevis: VAR Model

article image
Source: Authors' calculations.
Table 5d.

St. Lucia: VAR Model

article image
Source: Authors' calculations.
Table 5e.

St. Vincent and the Grenadines: VAR Model

article image
Source: Authors' calculations.

19. Fan charts summarize risks to the debt dynamics—the frequency distribution of the 1,000 debt paths generated. Shocks to interest rates, growth and the exchange rate affect the economy over the projection period (2008–12).21 Because the panel estimates of the fiscal reaction function for the ECCU are not well behaved, Scenario 1, or “no policy change” uses the estimated fiscal reaction function for the Caribbean as a whole (Table 3), and in Scenario 2, or “best-case,” fiscal policy is allowed to adjust to relevant macroeconomic shocks according to the pattern observed in the five middle-income countries (Table 4). The responsiveness of the primary surplus to public debt is stronger in the latter case.

20. The fan charts illustrate the significant uncertainty surrounding public debt projections, giving a better idea of the overall risks (upside and downside) to public debt projections in the ECCU. Three conclusions can be drawn:

  • For all six ECCU countries, both scenarios generally show the median debt path to be either stable or falling. Fiscal policy that makes adjustments in response to past public debt accumulation, as assumed in the scenarios, would contribute to putting debt on a downward path in the ECCU.

  • The overall risk profile obtained for the six countries reflects the idiosyncrasies of their respective economies. Given that the sensitivity of fiscal policy is the same for each country under a given scenario, the widths of the confidence intervals reflect each country's past volatility as identified by the VAR. Comparing the ninth to the first decile indicates that less volatile economies—such as Dominica and St. Lucia—exhibit narrower confidence intervals (Table 6). Grenada and especially St. Kitts and Nevis exhibit the widest confidence intervals for the public debt risk profiles, reflecting greater past volatility. Compared to estimates for South Africa in Debrun (2006), the uncertainty associated with public debt profiles in the ECCU, measured by the gap between the ninth and the first percentile, is two to ten times wider, reflecting the region's extreme past macroeconomic volatility, particularly for growth.

  • Except for Dominica, the policy response to shocks under both scenarios proves too weak to prevent growing debt ratios in the two upper deciles of the charts. For St. Kitts and Nevis, for example, the distribution of the debt ratio is skewed towards the upside, indicating that there is at least a 20 percent chance that combinations of adverse shocks may lead debt to exceed 220 percent of GDP by 2012. Unsurprisingly, skewness seems to be associated with high volatility. Given the probabilities of an increasing debt ratio, a more aggressive response of fiscal policy to the public debt-to-GDP ratio than in the two scenarios is warranted to contain upside risks to debt dynamics.

Table 6.

Antigua and Barbuda: Public Debt Risk Assessment 2008–12

article image
Source: Authors' calculations.
Table 6a.

Dominica: Public Debt Risk Assessment 2008–12

article image
Source: Authors' calculations.
Table 6b.

Grenada: Public Debt Risk Assessment 2008–12

article image
Source: Fund staff calculations.
Table 6c.

St. Kitts and Nevis: Public Debt Risk Assessment 2008–12

article image
Source: Authors' calculations.
Table 6d.

St. Lucia: Public Debt Risk Assessment 2008–12

article image
Source: Authors' calculations.
Table 6e.

St. Vincent and the Grenadines: Public Debt Risk Assessment 2008–12

article image
Source: Authors' calculations.

21. The public debt risk profiles can also provide debt sustainability indicators. As discussed in Section B, there are a number of operational methods for assessing the sustainability of public debt. For the ECCU, the ECCB Monetary Council has set a target of 60 percent debt-to-GDP by 2020, a threshold that the region has determined would be sustainable. For each country, the probability that public debt will decline towards the 60 percent target during the projection period, and the probability of upside risks to the debt ratio (debt exceeding 90 percent of GDP) are summarized in Table 6. If the probability of debt below 60 percent is very low by the end of the projection period (2012), then reaching the region's sustainability target by 2020 would also seem unlikely without assuming a future radical improvement in fiscal policies.

22. Even with the assumption of strong primary surplus responsiveness, the probability of public debt declining toward 60 percent of GDP in the medium term is low. St. Lucia, Dominica, and St. Vincent and the Grenadines are the only countries where there is greater than a 50 percent probability that debt declines to 60 percent of GDP in the best-case scenario. These probabilities are below 20 percent for Antigua and Barbuda and St. Kitts and Nevis. There are also significant differences across the countries in the extent to which a stronger responsiveness of the primary surplus to public debt (best-case scenario) substantially increases the probability that debt will decline to the 60 percent sustainability target by 2012. This reflects the significant distance a number of the countries are from the threshold in either scenario. The scenario assumptions do matter for Dominica, where the probability is 0.3 percent in the no-policy change scenario and 0.88 percent in the best-case scenario.

E. Concluding Remarks

23. This paper applies a probabilistic approach to debt sustainability analysis for countries in the ECCU. For all but two countries in the region, fiscal policy does not react adequately to past debt accumulation to provide a foundation for debt sustainability. Public debt risk profiles are derived using fiscal reaction functions which do respond to past debt, combined with estimated ECCU country-specific macroeconomic shocks. For half of the countries, there is only a low probability in the next five years of being on the right path to achieve the region's 60 percent debt-to-GDP target by 2020.

24. Stronger responsiveness of primary fiscal balances to past public debt, particularly in volatile macroeconomic environments such as in the ECCU, would increase the likelihood of placing debt on a downward path and limiting the upside risks. These results, coupled with the findings of strong fiscal policy inertia, underscore the need for formal mechanisms to operationalize the ECCB's debt sustainability benchmark, by linking national fiscal policies to the benchmark.22 This would involve setting primary balance targets consistent with achieving the 60 percent debt-to-GDP ratio, and adjusting those targets in response to macroeconomic and debt developments.

References

  • Abiad, Abdul and Jonathan Ostry, 2005, “Primary Surpluses and Sustainable Debt Levels in Emerging Market Countries,IMF Policy Discussion Paper No 05/6, (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Araujo, Juliana Dutra, 2009, “Fiscal Cycles in the Caribbean,forthcoming IMF Working Paper, (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Barnhill, Theodore and George Kopits, 2003, “Assessing Fiscal Sustainability Under Uncertainty,IMF Working Paper No 03/79, (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier J., Jean-Claud Chouraqui, Robert P. Hagemann and Nicola Sartor, 1990, “The Sustainability of Fiscal Policy: New Answers to an Old Question,OECD Economic Studies, vol.15, pp. 736.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem, 1985, “A Guide to Public Debts and Deficits,Economic Policy, vol.1, pp. 1362.

  • Carmichael, Jeffrey, 1982, “On Barro's Theorem of Debt Neutrality: The Irrelevance of Net Wealth,American Economic Review, vol.72, pp. 20213.

    • Search Google Scholar
    • Export Citation
  • Celasun, Oya, Xavier Debrun, and Jonathan D. Ostry, 2006, “Primary Surplus Behavior and Risks to Fiscal Sustainability in Emerging Market Countries: A ‘Fan-Chart' Approach,IMF Working Paper 06/67 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Debrun, Xavier, 2005, “Public Debt in South Africa: A Risk Analysis,in South Africa—Selected Issues, IMF Country Report 05/345 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Di Bella, Gabriel, 2008, “A Stochastic Framework for Public Debt Sustainability Analysis,IMF Working Paper 08/58 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Eaton, Jonathan, 1981, “Fiscal Policy, Inflation and the Administration of Risky Capital,Review of Economic Studies, vol.48, pp. 43445.

    • Search Google Scholar
    • Export Citation
  • Garcia, M., and R. Rigobon, 2005, “A Risk Management Approach to Emerging Markets' Sovereign Debt Sustainability with an Application to Brazilian Data,NBER Working Paper No 10336, (Cambridge, MA: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Hamilton, James D. and Marjorie A. Flavin, 1986, “On the Limitations of Government Borrowing: A Framework for Empirical Testing,American Economic Review, vol.76, pp. 80819.

    • Search Google Scholar
    • Export Citation
  • Hostland, Doug and Philippe Karam, 2005, “Assessing Debt Sustainability in Emerging Market Economies Using Stochastic Simulation Methods,IMF Working Paper WP/05/226 (Washington DC: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2002, “Assessing Sustainability,” Policy Development and Review Department http://www.imf.org/external/np/pdr/sus/2002/eng/052802.htm.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2003, “Public Debt in Emerging Markets: Is it Too High? Chapter III in World Economic Outlook, September (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2004, “Has Fiscal Behavior Changed Under the European Economic and Monetary Union?in Chapter II of World Economic Outlook, September, (Washington, DC: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • IMF and IDA, 2004, “Debt Sustainability in Low-Income Countries: Further Considerations on an Operational Framework and Policy Implications,” http/www.imf.org/external/np/pp/eng/2005/070105.thm.

    • Search Google Scholar
    • Export Citation
  • McCallum, Bennett T., 1984, “Are Bond-Financed Deficits Inflationary? A Ricardian Analysis,Journal of Political Economy, vol.92, pp. 12335.

    • Search Google Scholar
    • Export Citation
  • Kraay, Aart and Vikram Nehru (2006). “When Is External Debt Sustainable?,World Bank Economic Review, vol.20, pp. 34165.

  • Mendoza, Enrique G., and P. Marcelo Oviedo, 2004, “Public Debt, Fiscal Solvency and Macroeconomic Uncertainty in Latin America: The Case of Brazil, Colombia, Costa Rica and Mexico,NBER Working Paper, no. 10637 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Penalver, Adrian, and Gregory Thwaites, 2004, “Analyzing Sovereign Debt Sustainability: A New Probabilistic Approach,mimeo, (London: Bank of England).

    • Search Google Scholar
    • Export Citation
  • Reinhart, Carmen M., Kenneth S. Rogoff, and Miguel A. Savastano, 2003, “Debt Intolerance,Brookings Papers on Economic Activity, vol.1, pp. 174.

    • Search Google Scholar
    • Export Citation
  • Tanner, Evan and Issouf Samake, 2005, “The Shocking Facts: Public Debt Sustainability in Brazil, Mexico and Turkey,IMF Working Paper 06/295 (Washington, DC: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Xu, David and Piero Ghezzi, 2003, “From Fundamentals to Spreads: A Fair Spread Model for High Yield Emerging Markets Sovereigns,manuscript, Deutsche Bank, July.

    • Search Google Scholar
    • Export Citation
  • World Bank, 2004a, “Debt Sustainability in LICs—Proposal for an Operational Framework”, (Washington, DC: World Bank).

  • World Bank, 2004b, “Debt Sustainability in LICs—Further Consideration”, (Washington, DC: World Bank).

1

Prepared by Koffie Ben Nassar and Catherine Pattillo.

2

See Eastern Caribbean Currency Union—Staff Report for the 2009 Discussion on Common Policies of Member Countries (www.imf.org). This study uses data for the six Fund-member ECCU countries: Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, and St. Vincent and the Grenadines.

3

See IMF (2002) and World Bank (2004a and 2004b) for the framework applied to debt sustainability in low-income countries.

4

Dominica and Grenada undertook comprehensive debt restructurings in the early 2000s. For Dominica, debt restructuring, coupled with prudent fiscal management, has placed public debt firmly on a declining path.

5

Theoretical models that seem to allow this possibility have been explored by Buiter (1979), Eaton (1981), and Carmichael (1982).

6

McCallum (1984) shows that it has been difficult to develop equilibrium models in which investors would continue to buy government debt when the government borrows at an interest rate that exceeds the growth rate.

7

Using postwar U.S. data, Hamilton and Flavin (2001) find support for the proposition that all that is needed for the government to issue interest-bearing debt is to promise its creditors that it will balance its budget in expected present-value terms.

8

Buiter (1985) and Blanchard et al. (1991) model the primary fiscal balance required to stabilize the public debt. The objective is to stabilize the debt-to-GDP ratio at its current level or at any other level deemed more desirable.

10

Kraay and Nehru's (2006) paper on determinants of debt distress, which depend on the quality of policies and institutions, was an important contribution to the development of policy-dependent thresholds for external debt sustainability, used in the IMF-World Bank debt sustainability analyses.

12

Specified this way, the reaction function helps disentangle whether a high debt level is the result of adverse shocks (bad luck) or of undisciplined (bad) past policies. If the reaction function indicates that the authorities have systematically reacted in a stabilizing way to debt buildups, one can conclude that a high debt-to-GDP ratio is primarily due to bad luck. In this case, a debt ratio can be very high and yet sustainable as long as bad luck does not strike repeatedly.

13

Notice that in each simulation, the primary surplus, bt, incorporates a fiscal policy shock φi,tN(0,σφi2),

article image
where σφi2
article image
is the country-specific variance of the reaction function's residuals.

14

Abiad and Ostry (2005), among others, also include the lagged dependent variable in a fiscal reaction function.

15

While panel estimation controls for country-specific effects, one caveat is that the linear estimators assume similar fiscal behavior across countries.

16

The estimated results for Dominica and Grenada are not used in the stochastic simulations, because they are conditional on a different set of regressors.

17

Results are shown in the forthcoming IMF Working Paper.

18

Araujo (2009) finds that fiscal policy in the ECCU has been procyclical, with higher public expenditure during good years.

20

These error terms are contemporaneously correlated and identification depends on the assumed ordering of the equations in the VAR.

21

Different colors delineate deciles in the distributions of debt ratios, with the zone in dark grey representing a 20 percent confidence interval around the median projection and the overall cone, a confidence interval of 80 percent.

22

See Eastern Caribbean Currency Union—Staff Report for the 2009 Discussion on Common Policies of Member Countries (www.imf.org).