IV. An Assessment of Public Debt Sustainability ln Tanzania After the hipc Initiative42
97. On November 27, 2001 Tanzania became the fourth country to reach the completion point under the enhanced framework of the Heavily Indebted Poor Countries (HIPC) Initiative. External debt-service relief under the enhanced HIPC Initiative from all of Tanzania’s creditors is expected to amount to approximately US$ 3 billion over time (US$ 2,026 billion in net present value (NPV) terms). As a result of HIPC Initiative assistance, the NPV of Tanzania’s total external debt has been reduced by some 54 percent.
98. Debt-service payments have been cut substantially—as a percentage of government revenue, debt service has been reduced from 19 percent in 2000/01 (July-June) before HIPC Initiative assistance to an average after HIPC Initiative relief of 7.7 percent over the subsequent few years. Resources made available by this debt relief are to be used to fund antipoverty programs, which are outlined in Tanzania’s poverty reduction strategy paper (PRSP).
99. One key objective of the HIPC Initiative is to put a recipient country on a trajectory of external sustainability that does not require further rounds of debt relief. The (external) debt sustainability analysis (DSA) that underpins the debt relief provided to Tanzania under the enhanced HIPC Initiative was prepared by the staffs of the International Monetary Fund and the International Development Association and is detailed in the joint completion point document of November 8, 2001.43 In the baseline scenario, which assumes average economic growth of 5.7 percent between 1999/2000 and 2008/09 and 6 percent between 2008/09 and 2017/18,44 the staffs’ judgment is that, after full debt relief under the enhanced HIPC Initiative, Tanzania’s external debt position would be sustainable over the medium term. This conclusion is based on a number of benchmarks agreed under the HIPC Initiative that signal whether external debt is sustainable. For example, the NPV of the debt to exports ratio was expected to remain well below the HIPC Initiative target of 150 percent over the projection period. In the sensitivity analysis of the DSA, two pessimistic scenarios were considered: scenario 1 assumed lower economic growth and exports, and scenario 2 assumed lower external grants. Under both scenarios, the debt indicators deteriorate considerably but remain sustainable—the impact of lower growth on the debt indicators is considerably larger. This underscores the need for Tanzania to implement structural policies that are growth enhancing, and to maintain access to external grants and highly concessional financing.
100. One potentially important class of public debt that is not considered in the DSA is public domestic debt, including government guarantees. In the medium term, depending on the evolution of future government policy regarding the mix of domestic versus foreign borrowing, domestic debt may become an important source of additional stress on the government’s fiscal position. Moreover, if at any time the international donor community were to reduce its support to Tanzania, the government may have to borrow more from domestic sources at market-determined interest rates to allow for a gradual adjustment of its fiscal position. Otherwise, a large compression in government expenditures would be needed to close the fiscal gap, resulting in adverse consequences for the wider economy and the program of poverty-reducing expenditures.
101. This section extends the traditional DSA analysis in two ways: it includes domestic debt in the analysis; and the framework analyzes the implications for debt sustainability of the willingness of creditors to provide new financing to Tanzania. The focus of the framework is on the primary fiscal balance (excluding grants) that the government would be required to run to maintain a sustainable path for government debt, subject to the behavior of creditors, and the macroeconomic assumptions underpinning the projections.
102. The degree of access to highly concessional financing to help finance Tanzania’s priority expenditures has major implications for debt sustainability in Tanzania because of the large grant element provided by this type of financing. Different assumptions about the growth in concessional financing could have a large impact on the primary balance required for debt sustainability. One of the scenarios considered in this section also sheds light on the implications for fiscal policy of a withdrawal of donor support.
103. We focus here on the path of the primary balance that is consistent with debt sustainability-defined as a public debt-to-GDP ratio that is constant or declining over time. The growth in concessional loans is assumed to be driven by the willingness of external donors to provide additional concessional financing, and the growth in domestic debt by the rate of growth in nominal U.S. dollar GDP. The primary fiscal balance (nominal fiscal balance, excluding interest payments) can serve the government as an operational target to ensure that its debt profile remains sustainable, given assumptions about the availability and terms of future sources of financing If fiscal policy were on an unsustainable path, the government would need to implement changes in taxes and expenditure to bring about the required adjustment in the primary balance, or make efforts to secure additional concessional financing.
B. The Analytical Framework
104. Following Edwards (2002),45 the level of nominal public debt is disaggregated into two types: (i) concessional debt (DC) granted by multilateral agencies and other donors, and (ii) debt issued on commercial terms (DD). For simplicity, it is assumed that all new external debt taken up by Tanzania’s public sector after the HIPC Initiative completion point is on concessional terms (specifically, IDA terms), and all new domestic debt is contracted at market-determined interest rates in the domestic capital market.
105. In the model simulations, all debt is denominated in foreign currency (U.S. dollars).46 The stock of public debt at time t is then DCt plus DDt. The net increase in this stock at time t is given by the following expression:
106. This bounded property ensures that the primary balance, pbt, required to prevent a financing gap (net of foreign and domestic loans) converges over time to a steady state level. The expression for the dynamic path of the primary balance that is consistent with debt sustainability is then the following (see the appendix for the derivation):
where ddo and dco are the initial (time t=0) ratio of domestic debt and concessional debt to GDP, respectively (in U.S. dollars).48 Specifically, this is the initial stock of debt after full HIPC Initiative relief, including additional relief by bilateral creditors.49
107. The simulations track over time the primary balance required to maintain debt sustainability under the assumption that Tanzania achieves its growth target of 6 percent and inflation target of 4 percent implied by the macroeconomic assumptions (see Table IV.1).50 Three cases are considered in the simulations:
|Case 1||Case II||Case IE||Case IV|
|Net growth in foreign loans, 9 (in percent)||0||8||8||-5|
|Nominal stock of foreign loans||Remains at initial level (foreign donors provide for a rollover of maturing debt)||Grows at 8 percent||Grows at 8 percent||Falls by 5 percent each period (net repayment each period)|
|Net growth in domestic borrowing, β (in percent)||8||8||8||8|
|Initial real exchange rate overvaluation (in percent)||10|
|GDP growth, g(in percent)||6||6||6||6|
|US dollar inflation, π*(in percent)||2||2||2||2|
|Domestic inflation, π(in percent)||4||4||4||4|
|Interest rate on concessional loans,||0.75||0.75||0.75||0.75|
|Interest rate on domestic debt, ||8||8||8||8|
Case I: recourse to domestic financing
108. This is the baseline scenario and assumes that the donor community is not willing to provide (net) new concessional loans, so that θ = 0, and that any additional concessional financing from donors is sufficient only to roll over the existing stock of foreign obligations of the government as they become due. In the simulation, additional domestic borrowing increases the stock of domestic debt at the same rate as nominal U.S. dollar GDP ((β=g+π*=0.1), and so the ratio of domestic debt to GDP remains constant at its initial level, while the ratio of external debt to GDP declines over time to a steady state of zero.
Case II: external borrowing on highly concessional terms
109. This simulation assumes that, on top of the domestic financing available under case I, additional new loans are available to the government from external donors on highly concessional terms, so that the ratio of external loans to GDP remains constant at its initial level. By comparing case II with case L we can assess the impact of the concessional financing on the government’s required primary balance. In this case, θ and β are equal to growth in nominal GDP (g+π*), and so the ratio of the aggregate stock of debt to GDP remains constant at its initial level over the simulation horizon. This case is analogous to the traditional steady state analysis of debt sustainability, which asks what level of the primary balance would stabilize the current level of the debt-to-GDP ratio, given the projections of the growth rate of GDP and the real interest rate.
Case III: real exchange rate overvaluation
110. This case incorporates the impact of changes in the real exchange rate on the sustainable primary balance, while the other macroeconomic assumptions are the same as in case 17. The real exchange rate is assumed to be overvalued by 10 percent at the completion point, and in each period the gap between the real exchange rate and its equilibrium value is assumed to close by 10 percent. A partial adjustment equation, such as expression below, then governs the path for the real exchange rate over the projection horizon:
111. The impact of the real exchange rate on the sustainable fiscal position translates into the following expression, which is expression (1) augmented with changes in the real exchange rate (see the annex for the derivation):
where ∆rer is the change in the real exchange rate.
Case IV: net repayment of foreign loans
112. This case highlights the risk of foreign creditors—withdrawing from Tanzania. It assumes that foreign creditors reduce their exposure gradually, so that the Tanzanian government is required to make a net repayment of its foreign loans as they fall due—the net repayments amount to a reduction in the nominal stock of outstanding foreign debt of 5 percent in each period. This contrasts with case I, in which donors provide enough external financing to roll over Tanzania’s external debt obligations.
D. Fiscal Effort Under Alternative Cases
113. The least amount, of fiscal effort is required under case II (Figure IV.1).51 Under this case, the government has access to highly concessional financing from foreign sources; this allows for a rapid expansion in its resource envelope, while maintaining a fiscal position that is sustainable. Of the three cases, growth in the total stock of debt is highest in case II because net new financing from foreign sources grows at the rate of nominal GDP. The large grant element in the foreign financing provides the government with some additional resources, and, because the rate of interest paid on concessional debt is considerably lower than the rate of growth in GDP, the debt dynamics under case II are very favorable.
Figure IV.1.Required Primary Balance for Debt Sustainability
114. One important consequence that emerges from a comparison of cases I and II is that the resource envelope of the Tanzanian authorities grows more rapidly under case II. If higher growth in the resource envelope were to have a positive effect on economic growth, possibly through a more rapid expansion in growth-enhancing expenditures, then the gap between the required primary balance under the two cases could be even larger, primarily because the expected growth in GDP might then be higher in case II. The effect of fiscal policy on growth is subject to considerable uncertainty However, as an illustration, if the higher-priority expenditures in case II were to boost GDP growth by 1 percentage point to 7 percent, the permissible primary deficit would rise from an average of 4 percent over the first ten years of the simulation period to 4.2 percent. This provides an illustration of the potential for a virtuous cycle in which higher-priority expenditures boost growth, with this, in turn, providing more room for additional growth-enhancing expenditure.
115. Based on recent developments in the primary deficit and projections for 2002/03 and 2003/04, the simulations indicate the debt situation in Tanzania to be sustainable, once account is taken of the level of grants that Tanzania is expected to receive. Under case I, the required primary surplus declines over time from an initial level of 0.4 percent of GDP (before grants), while in case II, the Tanzanian government can comfortably run a primary deficit (before grants) of just over 4 percent of GDP. For fiscal years 2002/03 and 2003/04, grants are projected at 5.7 percent and 4.9 percent of GDP,52 which would allow the authorities to accommodate, in 2002/03, a primary fiscal deficit of 5.3 percent under case I, and of 9.7 percent of GDP under case II.
116. The favorable outlook for grant financing and its positive implications for public debt sustainability reflect recent experience in Tanzania From 1999/2000 to 2001/02, the Tanzania government had a primary deficit of 0.6 percent of GDP (inclusive of grants), while grants averaged about 4 to 5 percent over the last two fiscal years. Overall, the projections for the primary deficit, together with the simulations presented above, suggest that Tanzania’s projected primary balance is consistent with debt sustainability, so long as the government continues to have access to substantial donor financing on concessional terms.
Fiscal effort when the real exchange rate is overvalued
117. The previous cases assume that the exchange rate is constant over the simulation horizon Case III takes the macroeconomic assumptions of case IT and incorporates the impact of a real exchange rate depreciation toward its equilibrium value—the initial overvaluation of the real exchange rate is assumed to be 10 percent A depreciation of the real exchange rate adversely affects the sustainable primary balance through a deterioration in the relative purchasing power of the Tanzanian government’s tax base (which is in local currency) vis-à-vis its external debt obligations in U.S. dollars. Moreover, if an overvalued real exchange rate is an indication of an imbalance in domestic savings versus investment (domestic absorption is higher than output), then a tightening of fiscal policy would help address the problem by improving the public savings balance.
118. In case III, a steady depreciation in the real exchange rate toward its long-run equilibrium level unambiguously increases the fiscal effort required to maintain a sustainable debt position (Figure IV.2). The impact of a 10 percent overvaluation on the required primary deficit is small but temporary—on average, over the simulation horizon, the required primary deficit falls from 4 percent in case II to 3.9 percent in case 111.53 Eventually, the two cases converge as the real exchange rate converges toward its equilibrium value—the rate of convergence is a function of the speed with which the real exchange rate returns to its equilibrium value. The largest part of the adjustment in the real exchange rate occurs in the early part of the simulation. At first, there is a small reduction in the required (permissible) primary deficit, but once the gap between the real exchange rate and its equilibrium is sufficiently small, the favorable debt dynamics dominate, and the required primary deficit increases—this gives the backward-bending profile for the required primary deficit under case III.
Figure IV.2.Impact of a Real Exchange Rate Depreciation on Case II
The impact of a gradual withdrawal of foreign donors
119. While the simulations under cases I and II suggest that Tanzania does not have a near-term problem with debt sustainability, for the longer term, the risk remains that foreign donors could withdraw their support. Cognizant of this risk, case IV simulates a gradual withdrawal of foreign donors from Tanzania to assess the impact on the required primary balance.54 In terms of the model’s behavioral parameters, 0 is set at -5 percent, which is a net repayment in each period by the Tanzanian government of 5 percent of its outstanding stock of external obligations as they fall due. The impact of even this gradual withdrawal of foreign donors on the required primary surplus is significant—the average primary surplus required over the first ten years of the simulation to maintain debt sustainability increases from 0.3 percent in case 1 to 1.9 percent in case IV. This suggests that even a gradual withdrawal of donor funds would require the Tanzanian government to significantly tighten fiscal policy.
120. Based on the macroeconomic assumptions made about growth, inflation, and the availability of external concessional and domestic financing, the simulations indicate that the Tanzanian government can afford to run a primary deficit that is consistent with debt sustainability, if it continues to enjoy access to highly concessional external financing.
121. Recent experience suggests that Tanzania does not face debt sustainability problems, but over the longer term the risk remains that concessional financing from donors will dry up. The stylized simulation in case IV indicates that the impact of even a gradual exit of foreign donors would have a significant adverse impact on Tanzania’s required primary balance, implying significant fiscal tightening.
122. Given the pressures to spend resources on social priorities such as health and education, the scope for reducing expenditures is probably limited. Consequently, over the medium term there is a need for the government to put in place revenue-generating measures that raise government revenue as a share of GDP, and to begin to put Tanzania on a fiscal path that is less reliant on debt financing, including external concessional financing. Encouragingly, the simulations indicate that, with access to external concessional borrowing, the Tanzanian government has time to bring about a positive change in revenue-generating capacity from domestic sources.
123. The framework presented in this section suffers from a number of limitations that should be borne in mind when interpreting the results. First, the growth in 8 (the per-period growth in the nominal stock of foreign concessional loans) and P (the per-period growth in the nominal stock of domestic debt) is assumed constant and exogenously given. However, it can be argued that these may change over time to reflect developments in donor attitudes to Tanzania, as well as the evolution of domestic debt markets. Second, the rate of growth of debt financing, external and domestic, may have the desired impact of enhancing growth in the medium to long run, especially if resources are well utilized, thereby improving the debt position. However, notwithstanding these limitations, the framework goes further than the concept of external debt sustainability, and highlights the interrelationships among fiscal policy, debt policy, and the mix of external versus domestic debt for financing poverty-reducing government expenditures.
Upper-case expressions are nominal values, and lower-case expressions are ratios to GDP. Begin with the first expression presented in the main text:
Note that in expression (3) PB, is expressed as expenditure minus revenue, and so to express it in the usual way of revenue minus expenditure, everything is multiplied by-I:
Expressing all elements in terms of ratios to GDP, we get the following:
The key behavioral equations that govern the evolution of the ratio of debt to GDP are the following:ddt-1= (1 +β) ddt and dct-1= (1 + β) dct-1,. Consequently, setting t = 1 gives:
Including the real exchange rate changes into the expression amounts to augmenting the interest on domestic and concessional debt with the change in the real exchange rate. Then, expression (8) becomes the following:
where Δrert is the change in the real exchange rate at time t
Prepared by Ashok Bhundia.
Available at www.imf.org.
The interest rate, exchange rate, and macroeconomic assumptions are detailed in Tables 8 and 9 of the completion point document.
Sebastian Edwards, “Debt Relief and Fiscal Sustainability,”NEBR Working Paper 8939 (Cambridge, Massachusetts: National Revenue of Economic Research, 2002).
In subsection C, the benchmark model is extended to include valuation effects from changes in the real exchange rate.
Since both the domestic and foreign debt is converted into U.S. dollars, the nominal growth in U.S. dollar GDP is equal to the sum of the real growth in U.S. dollar GDP plus U.S. dollar inflation.
The initial ratios are those prevailing after full enhanced HIPC Initiative relief as at June 2001 Negotiations continue between Tanzania and its non-Paris Club creditors, which complicates the treatment of this debt as at June 2001. To simplify the analysis, and without knowledge of the outcome of these discussions, all debt is assumed to have been provided relief in line with the enhanced HTPC Initiative.
The link between this model and the DSA framework is clear if we assume the exports-to-GDP ratio remains constant at is initial level because then the evolution of the debt-to-GDP ratio over the simulation horizon pins down the evolution of the debt-to-exports ratio, which is an important indicator of external debt sustainability under the DSA framework of the HIPC Initiative.
The primary balances implied by the simulations are conditional on the macroeconomic assumptions being satisfied. Therefore, using this framework, it is not possible to provide insights into what fiscal adjustment would be required, and the form it should take, if, for example, there is a negative shock to GDP (or if there is a shock to 9 or P) in one year during the simulation period. In this sense, the simulations are deterministic.
Fiscal effort here is defined as the balance between noninteresl expenditure and revenue, and not the revenue-GDP ratio. A greater fiscal effort means a smaller primary deficit/larger primary surplus.
These are IMF staff projections as of October, 2002.
This gap would increase for a larger overvaluation of the real exchange rate.
The precise details of the simulation would depend on the nature and speed of the withdrawal of foreign concessional financing. However, the argument here is that a gradual withdrawal, as captured in the net repayment of 5 percent of outstanding obligations at each period, represents a fairly well-managed and phased exit, thus limiting the damage.