A. Current Practices in Setting External Debt Limits

2. All PRGF-supported programs include a performance criterion (PC) limiting nonconcessional external borrowing (of both government and government-guaranteed debt).² Concessionality is typically defined as debt with a grant element of 35 percent or more, although three programs require a higher grant element (the Kyrgyz Republic, Niger, and Rwanda with minimum grant elements of 45, 50, and 50 percent, respectively). Most of these PCs preclude any nonconcessional external borrowing. However, about one quarter of the PCs on debt limits included in PRGF-supported programs permit contracting of some debt with lower grant elements up to a specified nominal ceiling. The nonconcessional debt allowed under these PCs does not generally correspond to a single project.³ Among those programs which incorporate nonconcessional external borrowing, about half permit such borrowing only from development banks. Only one PRGF- supported program refers explicitly to commercial borrowing.

3. For most members, there is no ceiling on medium- and long-term external concessional borrowing and thus on overall debt. PRGF-supported programs do not generally include conditionality on the net present value (NPV) of total external debt. Also, while PRGF-supported programs systematically include a ceiling on the domestic financing of the deficit, only about half of PRGF-supported programs have a PC that extends to the overall fiscal deficit.⁴ Even where PCs on the overall deficit are used, most do not set a firm upper limit on the total external borrowing because they exempt concessional borrowing for specific purposes (generally capital projects) from the deficit limits either directly or through program adjusters.

4. Thus, current Fund conditionality does not systematically address countries’ risk of debt distress. While debt sustainability is important in PRGF program design, conditionality has focused on limiting the contracting of nonconcessional debt. Experience has shown, however, that borrowing on concessional terms does not guarantee that debt sustainability will be maintained, particularly if there is a large amount of such borrowing or the funds are used to finance projects with low or uncertain returns, or if returns are unlikely to be in the form of increased taxable activity. This could lead, and may in some cases already be leading to, unsustainable levels of debt.

B. Options for Incorporating the DSA Framework

5. Three aspects of Fund conditionality in PRGF arrangements could be modified to address debt sustainability concerns better. These are: (i) introducing limits on NPV of debt ratios, (ii) taking a more flexible approach to the minimum grant element permitted for concessional external borrowing and to establishing permitted levels of borrowing deemed nonconcessional, and (iii) broadening fiscal limits to capture external borrowing more systematically. Given the limited experience with conditionality designed to address debt sustainability concerns directly, the use of indicative targets rather than PCs in Fund-supported programs would generally be more appropriate at the outset unless debt difficulties are substantial. These indicative targets could be used to guide program review discussions. As experience accumulates, the form of conditionality for limits on NPV ratios and broader overall fiscal deficit limits can be revisited depending on the specific circumstances of the country.

6. The most direct approach to incorporating debt sustainability analysis into Fund conditionality would be to establish limits on NPV ratios themselves. A limit on the NPV of disbursed external debt would help focus PRGF-supported programs on external debt sustainability considerations. However, because PCs can only be set on variables over which the member has control, it would be difficult to set a PC on disbursed external debt as the authorities of the borrowing country do not directly control the timing of disbursements on contracted external debt. There are sometimes long lags between the contracting and disbursement of external debt. Also, NPV of external debt ratios incorporate elements beyond the authorities’ ability to control or reliably predict, including current exports and national income levels, as well as industrial country discount and exchange rates. Program discount and exchange rates and historical export and GDP data would need to be used if the limits were to be formulated as PCs.⁵ Finally, the calculation of overall NPV levels would be more complex than the existing grant element calculations and, hence, using a PC on the NPV of debt could pose a challenge for countries with serious capacity constraints and give rise to the risk of misreporting and noncomplying purchases of Fund resources.⁶ Recognizing the controllability issues, a limit on the NPV of external debt would be better formulated as an indicative target rather than as a PC, and the interpretation of performance against the target would be guided by consideration of the extent to which divergences are attributable to factors beyond the control of the authorities.

7. Performance criteria on minimum levels of concessionality of newly contracted de bt would still be needed in PRGF arrangements in light of the difficulties in setting PCs on NPV ratios. While an NPV ceiling would provide an incentive for a country to seek highly concessional terms and maximize the volume of financing, it would not preclude nonconcessional financing. An increasing share of nonconcessional financing may be appropriate in countries in the process of graduating from the World Bank’s concessional IDA financing or the Fund’s PRGF lending. However, excessive commercial or nonconcessional borrowing could lead to a reduction in the overall lending volume under the NPV ceiling by crowding out borrowing on more favorable terms, and raise concerns that official creditors providing concessional assistance would be cross- subsidizing commercial creditors. In this context, official creditors might not be willing to continue such subsidization over the medium term, particularly if the nonconcessional financing is associated with governance concerns. Finally, because the contracting of nonconcessional external debt is under the direct control of the authorities, debt limits of this sort can be set as PCs.⁷

8. Minimum levels of concessionality in standard external debt limits could be adapted more flexibly to country circumstances, as could the amount of nonconcessional borrowing.

• ?? Regarding concessionality thresholds, the staff proposes to introduce some flexibility to reflect the degree of debt distress, per capita income, and also the availability of financing. For example, the required level of concessionality could be raised above the 35 percent norm in countries experiencing debt distress.

• ?? Setting tighter concessionality thresholds could affect a country’s financing envelope, as some creditors may find it difficult to engineer even a small increase in the required grant element. Thus, depending on the degree of debt distress and prospects of a graduation from IDA and PRGF lending, on a case-by-case basis, programs could accommodate specified amounts of debt that falls outside the concessionality ceilings for specific high-priority and high-return projects. Absent debt distress, for IDA/IBRD and PRGF/EFF blend countries the norm would be to accommodate limited amounts of such nonconcessional debt in Fund-supported programs.

9. Performance criteria on the contracting of nonconcessional loans should continue to be applied on a loan-by-loan basis, rather than on a weighted average of borrowing.⁸ Using averages would raise the complexity of calculations and could pose problems for countries with capacity constraints with possible implications for misreporting.

10. Conditionality could also be strengthened by including an overall fiscal deficit limit more systematically in PRGF-supported programs. Many PRGF-supported programs already have PCs or indicative targets on the overall fiscal deficit. However, these have not been used systematically to reflect debt-sustainability considerations. Fiscal limits could be modified to capture external borrowing more systematically, but these limits would be used flexibly; broader fiscal limits would usually be binding only for countries in debt distress, or at high risk of it, to protect them from an unsustainable build-up of debt, while in other countries, borrowing options would be wider. Limits should be set on the fiscal deficit including grants as a revenue item (i.e., with grants excluded from the financing captured by the deficit limit). Also, given the difficulty in predicting disbursements for project loans, indicative targets or PCs on the overall fiscal deficit and their adjusters would need to take into account the ability of the authorities to control and monitor the timing of external loan disbursements. The use of such adjusters in Fund-supported programs allows for greater flexibility. They are already used in the context of current domestic fiscal financing limits and grant inflows. Their application prevents undue compression of expenditures in the event of shortfalls in budgetary grant financing to the extent that this is consistent with macroeconomic stability and debt sustainability, while at the same time allowing higher-than-expected grant financing of the budget to be used for additional poverty-related or other priority spending.