This Selected Issues paper focuses on the fiscal position of Sri Lanka. The standard analysis shows that, prior to the adjustment announced in the 2002 budget, fiscal policy was clearly unsustainable, leading to a rising debt-to-GDP ratio. The paper looks at external debt and complements the analysis of the public debt dynamics. The baseline scenario shows that debt ratios decline significantly in the medium term, as a result of strong growth founded on renewed peace and political stability, far-reaching structural reforms, and stable macroeconomic conditions.


This Selected Issues paper focuses on the fiscal position of Sri Lanka. The standard analysis shows that, prior to the adjustment announced in the 2002 budget, fiscal policy was clearly unsustainable, leading to a rising debt-to-GDP ratio. The paper looks at external debt and complements the analysis of the public debt dynamics. The baseline scenario shows that debt ratios decline significantly in the medium term, as a result of strong growth founded on renewed peace and political stability, far-reaching structural reforms, and stable macroeconomic conditions.

III. External Debt Sustainability6

A. Introduction

1. This chapter examines Sri Lanka’s external debt-as a complement to the preceding chapter on fiscal sustainability as well as the next chapter analyzing current account sustainability.7 The issue is important for Sri Lanka for several reasons, including concerns over public debt sustainability, for gauging the scope and terms of external borrowing to finance reconstruction of the North and East, and the vulnerability of the balance of payments to external shocks given the openness of the economy. Such assessments of debt sustainability involve judgments about whether the current account can be financed through private and official flows, medium-term balance of payments and associated debt projections, assessments of the appropriate level of the exchange rate, the degree of adjustment that is politically and socially feasible, as well as the balance sheets of several interlinked sectors.8

2. The new framework that was recently developed by Fund staff is used to gauge whether Sri Lanka’s external debt dynamics over the medium term are sustainable. Section B summarizes the baseline medium term macroeconomic scenario, and the assumptions on which the external projections are based. Section C presents a short overview of Sri Lanka’s external debt history and structure. It then outlines and assesses external debt sustainability using a framework recently proposed by the staff.9 This assessment decomposes the external debt dynamics of the baseline scenario and incorporates a standard set of sensitivity tests around this projection using alternative assumptions about the variables affecting the ability to service debt. These results are compared with levels that empirical data show have been associated with debt problems or corrections. Finally, since the results from these scenarios and the debt burden need to be viewed in the context of the debt structure, notably its concessionality, Section D discusses the present value of Sri Lanka’s debt, comparing the debt burden with a framework-the HIPC Initiative-which is used for assessing the external debt sustainability of low-income countries.10 Section E concludes with a summary of the external debt sustainability results and policy implications.

B. Medium Term Balance of Payments Prospects—the Baseline Scenario

3. The baseline macroeconomic scenario assumes the conflict does not escalate, and that far reaching structural reforms and the maintenance of macroeconomic stability raise the growth rate.11 The staff’s central scenario also assumes a significant reduction in the political uncertainty that has plagued policy making in the recent past. The main elements of the government’s strategy are elaborated in its poverty reduction strategy paper (PRSP),12 which identifies reforms of investment policy, the regulatory environment, labor market, financial and public sectors, as well as governance as key for liberalizing the economy and supporting private sector led growth.

4. Central to the baseline macroeconomic framework is the achievement of low single digit inflation and strong fiscal consolidation. The overall fiscal deficit of the central government is forecast to be more than halved over the four years ending 2006, to 4 percent of GDP. Improvements, in the form of a simplified tax system and strengthened administration, are likely to raise total revenue by 2½ percent of GDP, while expenditure measures to reduce defense spending, reduce public employment, rationalize public departments and transfers, and improve expenditure control, account for almost 2 percent of GDP. The baseline scenario assumes that economic growth will increase steadily to 6½ percent over the medium term, driven by higher investment, exports and consumption.

5. Structural reforms and infrastructure development would allow the economy to adjust to the shock from the removal of textile quotas and increased regional competition. Improved business and investor confidence in response to the economic reforms and peace is likely to encourage a shift toward higher value added garment exports, a growth of other industrial exports, and a resumption of agricultural and fisheries exports from the North and East. From 2005, however, export volume growth tapers off as total apparel and textile exports are programmed to remain flat in value terms due to the adverse impact on smaller scale garment producers-from the loss of “secure access” and heightened competition from more competitive producers-of the elimination of quotas on textiles and clothing exports under the Agreement on Textiles and Clothing.13 Growth of services is expected to average 5½ percent between 2002-06, led by increased tourism resulting from peace and further infrastructure development. As a result of these developments, and fairly stable income account and private transfers as a share of GDP, the external current account deficit is forecast to remain around 2½ percent of GDP into the medium term.

6. The capital and financial account is expected to grow in line with GDP as donor flows and FDI respond to the reform effort. The capital and financial account recovers in 2002 based on the inflow of program loans and delayed asset sales, as well as foreign direct and portfolio investment buoyed by increased investor confidence and private sector participation in infrastructure and utilities. Capital inflows are projected to remain at these levels over the medium term as the authorities implement their reform agenda and accelerate project implementation, with the large majority assumed on concessional terms. Gross reserves would increase steadily to just over $3 billion by 2006, equivalent to 4 months of import cover, with this cushion being justified by the narrow export base and vulnerability to shocks, and the dependence on official capital inflows.

C. External Debt Sustainability

Evolution and Structure of External Debt

7. After a rapid increase over the first half of the 1990s, Sri Lanka’s external debt has remained fairly stable, as creditors have contained their exposure. External debt increased sharply between 1990-95 from around $6.5 billion to $9.7 billion but net inflows have been limited over the last six years (Table III.1).14 Reflecting the predominance of the state in the economy and its failure to raise per capita incomes above low income country thresholds, the increase has largely been in public and publicly guaranteed debt associated with concessional disbursements from IDA, the AsDB and Japan.15 In terms of the debt burden, the external debt-to-GDP and debt-to-exports ratios averaged close to 80 percent and 250 percent respectively over the first half of the 1990s, reflecting low growth, inefficient public investment and poor export performance through the late 1980s and early 1990s16. Debt ratios, however, declined noticeably over the second half of the decade, with the debt-to-GDP and debt-to-exports ratios averaging 63 and 170 percent respectively. This was mainly the result of very limited net debt inflows, rather than exceptionally strong growth and export performance. The concessionality of the external debt and lengthy grace periods have moderated debt service ratios, which averaged 15 percent over the 1990s.

Table III.1.

External Debt Outstanding, 1990–2001

(In millions of U.S. dollars)

article image
Source: Data provided by the Sri Lankan authorities.

With government guarantee.

Without government guarantee.

Includes acceptance credits of Ceylon Petroleum Corporation, trade credits and short-term borrowings from FCBUs. Excludes nonresident foreign currency deposits.

Including use of Fund credit.

Includes banking sector liabilities.

8. At end-2001, total external debt was around 60 percent of GDP but the majority was concessional public sector liabilities to official creditors. Over 80 percent was public external debt, with the large majority being claims on central government. In terms of maturity, about four fifths of the total debt stock is medium-and long-term, although the proportion of short term debt declines to around 7 percent if banking system liabilities are excluded. The large majority of public debt is held by multilateral and bilateral official creditors, who each account for around 40 percent of total liabilities. IDA and the AsDB together account for almost one half of central government debt and Japan, which is the largest official creditor, a further one third. Sri Lanka’s eligibility for low income country facilities is reflected in the concessionality of the debt structure with over 90 percent of public external debt and between 65-75 percent of total external debt (depending on whether banking sector liabilities are included) on concessional terms. In NPV terms the total and public debt stock is estimated to be around 51 and 39 percent of GDP respectively, reflecting a grant element of approximately 20 percent (see Table III.3).17 Total private sector and public corporation debt which is not covered by a government guarantee (mainly trade credits to BOI companies) was less than 3 percent of the total debt stock, rising to 14 percent if banking sector liabilities are excluded.

Table III.2.

External Debt Sustainability Framework

article image
Source: Data provided by the Sri Lankan authorities and staff projections.
Table III.3.

Net Present Value of External Debt

article image

Includes IMF debt.

Includes new disbursements on existing and new debt.

Uses the three year moving average of exports of goods and services ending in the current year.

Uses central government revenue excluding grants converted at end-2001 exchange rates.

External Sustainability Framework

9. The framework for assessing external debt recently developed by Fund staff seeks to build on existing best practices in the assessment of sustainability.18 Of the two main elements of the framework, the first is the centrality of the staff’s baseline medium-term scenario. The historical and projected debt dynamics resulting from this scenario are decomposed, allowing identification of whether the stability or otherwise in the debt ratio arises from the behavior of interest rates, growth rates, inflation or real exchange rate movements or through adjustment in the trade balance. In terms of debt dynamics, a country is sustainable if it satisfies the present value budget constraint (i.e., the present discounted value (PDV) of its current and future primary spending is no greater than the PDV of current and future income, net of initial indebtedness) without a major correction in the balance of income and expenditure given the costs of financing it faces in the market. Formally, the decomposition of the change in the external debt ratio is based on the debt dynamics equation:

Dt+1 = (1+r)Dt - TBt+1

where D is end-period debt in U.S. dollars, r is the external interest rate, and TB is the debt-creating component of the noninterest current account of the balance of payments. Expressed in terms of the change in the external debt ratio, the debt dynamics equation can be written:


where the change in the net debt ratio is a function of real GDP growth (g), the external interest rate (r), the growth rate of the U.S. dollar value of the GDP deflator (p), the external debt-to-GDP ratio (d), and the debt-creating component of the noninterest current account as a percent of GDP (tb).

10. The second part of this framework consists of a set of standard sensitivity tests around the medium-term scenario, examining the implications of alternative assumptions about policy variables, macroeconomic developments and the costs of financing. The first sensitivity test sets the key parameters to their historical averages, thereby assessing the realism of the baseline projection given past outturns. The other tests consider adverse two-standard deviation shocks lasting two years to each of the key parameters-the current account, growth, interest rates, and dollar GDP-and a one-standard deviation combined shock to each of these variables. Finally, since the volatility of the real exchange rate may, historically, be low under a crawling band regime of the type operated by Sri Lanka prior to 2001, a scenario in which there is a 30 percent real exchange rate depreciation is also considered.

11. While the focus is on the debt dynamics, these need to be viewed against the level of the debt ratio. It is difficult to establish any definitive danger levels for the external debt-to-GDP ratio given the importance of country specific factors. Existing cross country empirical work suggests that an external debt ratio of about 40 percent of GDP provides a useful benchmark, with an appreciable increase in the likelihood of a debt crisis or debt correction at ratios above this level.19 Moreover, debt ratios have to be viewed in terms of the concessionality of the debt, which is typically an important consideration for low income countries such as Sri Lanka. In this regard, the thresholds under the HIPC Initiative framework provide a reference point, where a country is considered unsustainable if the NPV of public debt-to-exports is above 150 percent or, for very open economies, the NPV of public debt to government revenue exceeds 250 percent.20

Baseline Scenario

12. Under the baseline scenario, the external debt-to-GDP ratio declines gradually through the forecast horizon. The debt ratio rises modestly in 2002 to 63 percent of GDP and then declines by 12 percentage points of GDP over the forecast horizon through 2007 (Table III.2). With the exception of 2002, the decline is driven by a fall in net debt creating external flows, which more than offsets any increase in gross foreign assets. In 2002, the increase in the debt ratio is primarily the counterpart of the large accumulation of international reserves by the central bank and the reconstitution of commercial banks’ net foreign assets. This build up of foreign assets is primarily supported by official capital inflows (albeit on concessional terms), which outweigh the impact of higher growth, a recovery in foreign direct and portfolio investment, relatively low interest rates on external debt, and favorable projected impact of changes in the U.S. dollar value of the GDP deflator on the debt ratio. From 2002 onwards, however, there is a significant and continuous decline in the external debt ratio that averages about 2 percentage points each year. While continued reserve accumulation over the medium term contributes to debt accumulation, albeit at a slower pace, the decline in the external debt-to-GDP ratio is primarily being driven by lower net debt creating inflows. This is despite a current account deficit (net of interest) of around 1 percent of GDP throughout the projection period. The key variable driving this reduction is the growth rate, and to a lesser extent continued high foreign direct and portfolio investment and the impact of the exchange rate (and domestic prices) on dollar GDP.

13. This decomposition of the external debt dynamics highlights the crucial role of higher growth and FDI for both a continuous decline in the debt ratio and securing a more comfortable reserves cushion. In turn, this growth clearly depends on the returns on foreign borrowing and FDI. In this regard, it is imperative that the government implement its reform agenda to increase the marginal efficiency of investment and total factor productivity more generally, and for donors to align their assistance behind these reforms.21 This requires liberalizing product and factor markets where extensive regulations misallocate resources by distorting economic incentives, and a major reduction and reorientation of the role of the public sector. Macroeconomic stability also needs to be secured-indeed, this is inextricably related to the excessive role of the state in the economy, resulting in large fiscal deficits that reflect large public claims on output.

Sensitivity Analysis

14. If the new policy agenda is not implemented vigorously, debt ratios could rise well into unsustainable territory. The baseline can be compared with various shocks (see Figure III.1).22 The implications of the status quo are illustrated by comparing the difference between the baseline with simulations based on historical values of the key variables (Stress Test 1). Under these assumptions, the debt ratio increases steadily from 2001 to 69 percent of GDP by 2007, or 17 percentage points of GDP above the baseline. The main factors are the higher current account deficit and the lower growth rate. More specifically, the baseline assumes growth will average 5.8 percent or about 1 percentage point above the historical average. As for the current account, the baseline assumes that the deficit will be 1-1½ percentage points lower than the average of the past ten years, driven primarily by an improved trade deficit. The stress test makes explicit the assumptions underlying the baseline, highlighting the urgency of increasing growth-through openness, good governance, investing in and rewarding human capital, and building sound institutions-if the poverty reduction goals in the PRSP are to be realized. Indeed, to the extent that historical performance overestimates future prospects in the absence of reform-in the sense that market access was guaranteed in the past and regional competitors such as China and India were at earlier stages of reform-the opportunity costs of delaying structural reform will be greater.

Figure III.1.
Figure III.1.

Sri Lanka: External Debt Sustainability

Citation: IMF Staff Country Reports 2002, 208; 10.5089/9781451823516.002.A003

15. Other stress tests highlight that the sensitivity of Sri Lanka’s external debt dynamics to macroeconomic shocks is closely related to the type of shock. The interest rate shock (Stress Test 2) would have relatively little impact, raising the external debt-to-GDP ratio by only 1 percentage point of GDP, due to the low historical volatility and level of interest rates. A large shock to the growth rate in 2002 and 2003 (Stress Test 3) that resulted in no growth in those years would see the debt ratio increase to 66 percent by 2003-some 10 percentage points above the baseline-but the resumption of high growth through the rest of the projection period would prevent an explosion of the external debt-to-GDP ratio. A large deterioration in the current account deficit in 2002 and 2003 to 7¾ percent of GDP (Stress Test 5, with the noninterest current account deficit at 6 percent of GDP) would force the debt levels to 71 percent in 2003, only returning to 2001 levels by 2006. If Sri Lanka experienced a combination of growth, interest rate, current account and exchange rate shocks-perhaps reflecting a breakdown in peace talks, stalling of structural reform, and downturn in the global economy-debt ratios would be between 10-13 percentage points above the baseline (Stress Test 6). A large real exchange rate depreciation, such as a 30 percent GDP deflator shock, which might follow a major escalation of the civil conflict and large scale reversal of the reform momentum, would translate into an explosive increase in the debt ratio in 2002 (to 90 percent of GDP), with this remaining over 20 percentage points of GDP above the baseline scenario five years later (Stress Test 7).

D. Concessionality

16. In addition to analyzing the debt dynamics, assessment of the level of the debt ratio can be meaningful for low income countries once allowance is made for the concessionality of the debt. To this end, a simple estimation was undertaken to determine the NPV of the total and public and publicly guaranteed external debt, and how this is projected to evolve over the medium term under the baseline scenario. Although actual repayment profiles are used for the three largest creditors-Japan, IDA and the AsDB-these estimates should be interpreted with caution since they are based on aggregate debt stocks by groups of creditors (as opposed to a loan by loan decomposition, with loan specific terms applied), and stylized assumptions on the terms of new borrowing.23 The results are summarized in Table III.3.

17. Even after the concessionality of the external debt is taken into account, Sri Lanka’s debt sustainability depends on an aggressive modernization of the economy. Results from the NPV calculations indicate a total debt-to-GDP ratio of around 50 percent, or 11 percentage points of GDP below the nominal debt ratio, implying a grant element of around 20 percent. For publicly and publicly guaranteed debt, the ratio is 39 percent of GDP.24 Under the baseline scenario, the NPV public external debt ratio declines steadily through the projection period to 34 percent by 2006, reflecting the strong growth assumed over the medium term. At the same time, however, the NPV public external debt to exports ratio increases from 100 percent in 2001 to 127 percent by 2006. In part this reflects the gradual decline in the exports to GDP ratio into the medium term due to the slowdown in export volumes of textiles and clothing. It is also being driven by the significant increase (almost 30 percent) in the NPV of public debt over the projection period, entirely reflecting the rapid accumulation of new (mainly public) external debt.25 While the ratios to GDP and exports remain well below the averages for HIPCs, the fiscal burden of external debt is notable in that ratios to central government revenues are initially in the unsustainable range under the HIPC Initiative, although they decline rapidly thereafter.26

E. Conclusions

18. The analysis of external debt sustainability suggests the following conclusions:

  • First, Sri Lanka’s external debt dynamics are favorable under the baseline scenario and external debt ratios (based on nominal debt stocks) move toward sustainability when structural reform efforts are reinvigorated and macroeconomic stability is maintained. However, the sensitivity analysis shows clearly how dependent this outcome is on attaining higher growth rates than the recent past and the implications of failing to adjust to the type of shocks to which Sri Lanka is vulnerable.

  • Second, although most public debt is on concessional terms, the grant element in the debt stock is not particularly high (given low international interest rates). Over the medium term, the debt ratios (in present value terms) border levels that empirical work has identified as cause for concern. This is primarily on account of the large volume of new borrowing, even if assumed to be on concessional terms.

  • Finally, the policy implications of these conclusions highlight the critical importance of accelerating structural reforms that support innovation and diversification of the export base, of fiscal consolidation and of a prudent debt management strategy. Contracting of non concessional debt should be strictly limited and grant financing of the PRSP sought as far as possible. Establishing the conditions that encourage non debt creating inflows is imperative.


  • Agency for International Trade Information And Cooperation (ATTIC), “Textiles and Clothing: Implications for Less-Advantaged Developing Countries,June 1999.

    • Search Google Scholar
    • Export Citation
  • Detragiache, E and Spilimbergo, A,Crises and Liquidity: Evidence and Interpretation,IMF Working Paper, WP/01/2

  • Francois, J., Spinanger, D.,Market Access in Textiles and Clothing,Egyptian Center for Economics Studies, May 2002.

  • Government of Sri Lanka, “Connecting to Growth” Sri Lanka’s Poverty Reduction Strategy, Draft, June 2002

  • IMF, “Assessing Sustainability,SM/02/166

  • Patillo, C, Poirson, H, and Ricci, L,External Debt and Growth,IMF Working Paper, WP/02/69


Prepared by Sukhwinder Singh (PDR).


This chapter does not examine financial sector issues linked to external and fiscal sustainability, for example contingent claims on the government. Financial sector vulnerabilities were examined under the recently completed Financial Sector Assessment Program (FSAP).


Assessments of the exchange rate are important to assess external sustainability, for example the risks from an overvalued exchange rate in the presence of foreign currency denominated debt. This chapter only considers the impact of an exchange rate shock on debt ratios.


See Assessing Sustainability, SM/02/166.


The comparison with the HIPC Initiative is purely illustrative and in no way intended to suggest Sri Lanka is eligible for HIPC Initiative relief. To the contrary, the country’s exemplary debt service record and income level clearly sets it apart from HIPC countries.


Uncertainties associated with the peace process mean the baseline does not assume fiscal outlays or external borrowing associated with reconstruction in the North and East. Clearly, medium term external debt sustainability depends on the magnitude of any such expenditure as well as the terms of its financing.


See “Connecting to Growth” Sri Lanka’s Poverty Reduction Strategy, Government of Sri Lanka’s draft presented to the Development Forum in Colombo in June, 2002.


Under the Agreement on Textiles and Clothing (ATC), WTO members that maintain import restrictions on textiles and clothing must progressively reduce the restrictions and integrate the products concerned into GATT rules. All products have to be integrated by the end of a ten-year transition period (2005). Implementation of the integration process has been slow so far, suggesting a large impact from 2005 (see AITIC, 1999). Although estimates of the simultaneous effects of ATC quota elimination and China’s WTO accession on Sri Lanka are not available, general equilibrium estimates suggest a dramatic shift in textile and clothing trade toward China and India and lost market share for most other developing countries (see Francois, T. and Spinanger, D., 2002).


Debt figures are inclusive of approximately $i billion in banking system liabilities.


Debt to the Fund declined over the 1990s reflecting repurchases of the S AF and limited new disbursements, but rose from 2001 when a new Stand-By Arrangement was approved.


Part of the increase in the debt stock, particularly over the first half of the decade reflects the significant appreciation of yen and SDR against the U.S. dollar. In terms of currency composition, about 90 percent of the debt is in one of three currencies (40 percent in SDRs, 27 percent in yen, and 22 percent in U.S. dollars).


Because the NPV of the debt takes into account its concessionality, it is a more accurate measure of a country’s effective debt burden. The grant element is defined as the difference between the nominal and NPV of the debt, expressed in terms of its nominal value.


Even though this approach is primarily intended to be applied to countries with significant market access, the analysis of the debt dynamics is useful for low income countries, especially when combined with an assessment of the debt’s concessionality.


A survey of the empirical work is provided in Appendix I of “Assessing Sustain ability,” SM/02/166. Specifically, the conditional probability of a debt crisis or correction rises to 15-20 percent for countries where the debt ratio is above 40 percent of GDP, from 2-5 percent where the ratio is below. Moreover, it is worth noting that the debt ratios that can be supported before a crisis occurs are likely to be higher the greater the export to GDP ratio; this ratio is over 30 percent for Sri Lanka, suggesting the 40 percent threshold may be conservative.


Recent empirical work using a large panel data set for 93 developing countries over 1969-98 finds that the average impact of debt on growth becomes negative at close to the 150 percent of exports threshold set under the HIPC Initiative (see Patillo, 2002). Moreover, this work indicates that debt levels beyond 35-40 percent of GDP (measured in either nominal or NPV terms) may be detrimental to growth.


The baseline scenario builds in such improvements in the efficiency of investment with the incremental capital output ratio (ICOR) declining from 6.6 to 4.5 between 2002 and 2007.


These shocks are applied as (i) temporary deviations from the baseline, (ii) with all other variables unchanged, (iii) without feedback effects, and (iv) with magnitudes based on historical standard deviations over the past ten years. A key assumption in the framework is that the level of asset accumulation is invariant to the shock-neither the authorities nor the banking system varies its build up of reserves or borrowing in response to the shock, but undertakes the necessary adjustment in macroeconomic policies to return to the baseline path. This may be unrealistic, but serves to highlight the debt implications of securing a “comfortable” level of foreign assets in the presence of shocks.


Key assumptions underlying the NPV analysis are: (i) use of actual projected debt service through 2010 for Japan, IDA, and the AsDB and average maturity and interest rates to calculate the debt service profile thereafter (these creditors account for 2/3 of total public debt); (ii) other bilateral and multilateral debt has a 5 year grace and twenty year maturity; (iii) publicly guaranteed debt is on IDA terms and has the same remaining maturity as IDA; (iv) private debt is repayable over 4 years; (v) public commercial debt, banking sector liabilities (whose maturity is typically less than one year) and short term debt are all at market rates and their NPV is equivalent to their nominal value; (vi) all new central government borrowing and 80 percent of gapfill financing is on IDA terms; and (vii) the OECD’s average six month currency specific commercial interest reference rates (CIRRs) for the last semester of 2001 are used as the discount rates, as per HIPC Initiative methodology.


The NPV and grant element projections are highly sensitive to the discount rate applied. For example, the NPV of Japan’s claims is higher than its nominal value since the average interest rate on the debt exceeds the CIRR of the yen, which was close to historically low levels at end-2001 (less than 1.5 percent).


The NPV of existing public debt would fall by $1.3 billion between 2001 and 2006.


The simple average of the total NPV of public and publicly guaranteed debt for a group of 24 HIPCs that had reached their decision point by February 2002 was 60 percent for NPV debt-to-GDP, 290 percent for debt to exports and 278 percent for NPV of debt to revenues. These ratios are after the full application of traditional debt relief mechanisms but before HIPC relief.