Information about Sub-Saharan Africa África subsahariana

9 Assessing the Efficiency of Mechanisms for Dealing with the Debt Problems of Low-Income Countries

Zubair Iqbal, and S. Kanbur
Published Date:
September 1997
  • ShareShare
Information about Sub-Saharan Africa África subsahariana
Show Summary Details

It is now a decade and a half since Mexico nearly defaulted on its external obligations and the “debt crisis” hit began. There was, of course, more than one type of debt crisis, and while moves to reduce the burden of commercial obligations have helped bring middle-income countries to more sustainable debt situations, the position of many low-in-come countries, concentrated particularly in sub-Saharan Africa, has deteriorated. This paper assesses the mechanisms that have been used to tackle the external debt difficulties of poor countries against a set of efficiency criteria, with a view to identifying ways in which the mechanisms might be improved. The recent HIPC Debt Initiative, which brings multilateral lenders more fully into debt relief processes, makes it especially timely to undertake such an assessment, as also to find long-term indicators of the persistence of debt difficulties.

Some indication of the effectiveness of past mechanisms is provided by presentation of some key debt indicators for the HIPCs. Three widely used ratios are given in Table 1. The first year recorded, 1980, shows the situation before the debt crisis struck. If severe indebtedness is conventionally taken to mean ratios of debt to GNP exceeding 80 percent,1 the table shows that, on average, this group of countries was not in great difficulty at the beginning of the 1980s. By 1985, however, the average ratio had nearly doubled and actual debt servicing was absorbing more than a quarter of export earnings, despite extensive arrears on official and commercial debts. The necessity for action became accepted by creditors and rescheduling procedures were put in place within the Paris and London clubs. However, by 1990 the ratio of debt to GNP had risen further to 175 percent, almost three times the figure a decade earlier.2 The ratio of debt service to exports had begun to fall, however, reflecting continuing arrears, the writing-off of some past loans, increasing substitution of concessional for nonconcessional debt, and numerous rescheduling agreements. During the period 1990–94, these trends continued. Despite an intense period of further rescheduling, the nominal values of the debt stock and the ratio of debt to GNP were still growing, while the ratio of debt service to exports recorded a further small fall. The main impact of debt relief appears to have been on lowering this latter ratio, but this result is only obtained when measuring actual debt service; the average scheduled debt-service ratio of HIPCs in 1994 was 56 percent of exports (IMF, 1995, p. 75, Table A8), with the difference reflecting the extent of arrears.

Table 1.Debt Indicators for Heavily Indebted Poor Countries
Total external debt stock (in billions of dollars)62.8110.8210.9236.1
Average ratio of actual debt service to
exports of goods and services (in percent)17.026.423.721.4
Average ratio of external debt stock to
GNP (in percent)59.2102.5174.8220.6
Source: World Bank (1993, 1994b).Note: Time series figures for the fiscal burden were not available at the time of writing. More details are provided on the fiscal angle in the text.
Source: World Bank (1993, 1994b).Note: Time series figures for the fiscal burden were not available at the time of writing. More details are provided on the fiscal angle in the text.

The Efficiency Criteria

There are thus prima facie grounds for questioning the efficiency of past approaches to the debt difficulties of poor countries. But how might we go about evaluating the extent of efficiency? Four main criteria are utilized in this paper:

  • (1) Adequacy: The extent to which past mechanisms have provided relief adequate to meet, but not to exceed, needs.
  • (2) Productivity: The extent to which the relief provided has been able to discriminate efficiently between debtor countries that are, or are not, likely to put the additional resources to productive use.
  • (3) Cost minimization: The extent to which there are avoidable transactions costs in the negotiation and implementation of country debt relief arrangements. The degree of complexity of the mechanisms in place are an important aspect of this.
  • (4) Transparency: Agreements, and the processes by which they are reached, should be transparent and predictable, with a known set of common rules applicable to all cases.
  • The following pages elaborate on these criteria.

Adequacy of Relief

Since the initial premise is that the levels of debt held by some countries is, in some sense too high, the corollary is that relief must be sufficient to bring remaining obligations within reasonable estimates of their debt-servicing capacities. It is necessary, therefore, to define what is too high in order to determine which countries should benefit and how much relief they require.

This consideration raises the question of eligibility rules. These are best based on country-specific economic indicators, of which three are widely accepted:3

  • The GNP of debtor countries, both per capita (as a crude indicator of poverty and social needs) and relative to the stock of debt.
  • Balance of payments indicators, permitting estimates of the foreign exchange that may be available for servicing external debts, after consideration of import and other competing needs.
  • Fiscal indicators, allowing an assessment of government revenues that would be absorbed in servicing external public debt, relative to the tax base and to competing categories of government spending.

Each criterion shows, from a different angle, a debtor’s ability to pay in cash-flow terms, and may be projected to provide an assessment of future debt-servicing capabilities. However, to the extent that ability to pay is projected into the future, provision needs to be made for the possible supervention of external shocks. Such shocks are a particular feature of low-income countries, with generally more fragile economies and reliance on an export basket heavily weighted to a narrow range of commodities with large price variability. However, shocks can also be positive, improving economic performance. Arrangements should further distinguish between temporary and long-lasting shocks. A debtor suffering, say, a natural disaster chiefly requires short-term financing, whereas a persistent terms of trade shock will call for assistance with structural adaptation. The higher the probability and severity of external shocks, and the greater the uncertainties, the more flexible the mechanisms will need to be. A similar consideration applies to the adaptation of mechanisms to uncertainties about the required level of future debt relief. Since these estimates are based on forecasts, often about unpredictable variables, arrangements should embody enough flexibility to accommodate exogenous deviations from expected values.

A further issue related to the adequacy of relief is the extent to which arrangements are comprehensive, including all creditors and categories of debt. The major omissions under either heading will tend to undermine the sustainability of relief arrangements. This argues in favor of mechanisms that include both multilateral and bilateral creditors; among bilaterals, for both Paris Club and non-Paris Club creditors; for Paris Club creditors, that the division between pre- and post-cutoff debt is important; that so too is the degree to which mechanisms include previously rescheduled debt, as well as obligations still held on their original terms; and that both concessional and nonconcessional loans be included. A related consideration is the extent to which the process is sufficiently inclusive to facilitate a comprehensive view of a debtor country’s economic situation and a package of relief designed in the context of other likely financial flows.

Finally, an efficient process should not result in levels of debt relief in excess of those necessary to restore viability to beneficiaries’ debt-servicing situations, bearing in mind creditor governments’ other objectives in their aid programs to these countries of economic development and poverty reduction. In other words, relief mechanisms should permit creditors to recover as much of their past loans as debtors’ economic situations permit, after balancing this goal against those of creditors’ policies regarding official development assistance.

Productivity of Resource Use

Debt relief releases financing that would not otherwise be available (except to the extent that the debt would not otherwise be serviced). Moreover, since the demand for such resources outstrips the supply, there is an opportunity cost. If these resources are not well employed, relative to alternative uses, they will undermine the objectives of restoring debtor countries’ economic situations and, more specifically, of so improving future debt-servicing capabilities that debt viability can be attained and the threat of never-ending relief avoided. We therefore need to inquire about the extent to which debt relief mechanisms discriminate in favor of debtor countries that are likely to put the resources to good use. Various factors influence the likely effectiveness of this form of aid but the one we wish to emphasize here—and which the creditors have always emphasized—is the quality of the policy environment.

There is little controversy about the importance of this factor. Contrasting performances between otherwise similar economies seem largely explicable in terms of differing policy environments.4Reynolds’ (1985) survey of long-run development experiences concludes that “political organisation and the administrative competence of governments” is the single most important explanation of variations in developing countries’ growth records, and Bleaney (1996, p. 476) is among several researchers who have found macro-economic instability to retard economic growth, which, he suggests, occurs by reducing the productivity of capital formation. On similar grounds, Isham and Kaufmann (1995, p. 22) describe the positive effects of macroeconomic stability on the productivity of investment as “the forgotten rationale” of policy reform.

The chief way in which creditors have sought to build this factor into debt arrangements has been by linking debt relief to the adjustment programs of the Bretton Woods institutions, particularly those of the IMF, with eligibility for Paris Club relief being conditional on the debtor having an agreement in force with the Fund (or some other Fund seal of approval for its macroeconomic policies). The World Bank has also become more involved and is firmly entrenched in the mechanisms associated with the HIPC Initiative. Fund and Bank programs largely seek to influence the quality of policies by means of conditionality, so we will need to enquire into the effectiveness of this.

The IMF-World Bank conditionality is further intended by creditors to protect against moral hazard, that is, against the danger that debt relief, by providing additional resources, will diminish governments’ incentives to undertake (often politically risky) economic policy reforms. So long as conditionality can be enforced, that possibility should be ruled out. However, moral hazard may result not only from a lack of will to reform, but also from inadequate incentives to do so. Suppose, for example, that a debtor government is strongly committed to meeting all obligations but is forced by external circumstances to fall into arrears. In the absence of a negotiated reduction in arrears, a large part of any future improvement in economic performance is then likely to accrue to external creditors rather than to investment in internal development. In such circumstances, as Sachs (1989) and others have argued, a government may become disillusioned with the benefits of continuing to enforce stringent policy measures.

On the creditor side, an efficient mechanism must ensure consistency between the use of development aid and other resource transfers. These should be viewed together: debt arrangements should not have the effect of diverting ODA intended to promote economic development and/or the reduction of poverty into debt servicing. It is important that both the ODA and finance arms of creditor countries act in concert: the aid arm to ensure that its composition is consistent with the long-term need for fiscal and external balance in the debtor country; and the finance arm to ensure that the terms of debt relief are not such as to subvert its government’s ODA objectives. This set of considerations raises the question of whether institutional arrangements, internationally and within creditor governments, are such as foster the harmonization of debt and ODA objectives.

Minimization of Transactions Costs

Efficiency implies the absence of avoidable transactions costs. Substantial costs may be incurred by all parties during negotiations, however. These may be real-resource costs, with specific fiscal implications, or they may be more intangible. The more complex the debt negotiation mechanisms, the heavier the costs are likely to be. Three particular types of transactions costs may be identified: (1) the direct costs of negotiating with multiple agencies; (2) the cost of collecting information that is required in negotiations; and (3) the opportunity cost resulting from the diversion of scarce skills into negotiations that could otherwise be spent on more productive policy-analysis and implementation work. Needless to say, the burden of such costs is potentially far more serious for the governments of low-income countries, with weak public administrations and limited expertise.

Similar considerations point to the superiority of mechanisms that provide for an end to the need for negotiations—an exit—rather than apparently envisaging an indefinite series of time-bound relief agreements dealing with only a slice of outstanding debts. Frequent returns to the negotiating table not only increase the costliness of the exercise but also increase uncertainty, with potentially adverse effects on future economic performance and, therefore, debt-servicing capabilities.

Transparency and Predictability

For maximum efficiency in the debt assessment process, both sides must know what options are available and how a debtor can qualify for relief. If the options are clear and the process of assessment transparent, informational requirements will be largely known and negotiations can be kept short, reducing transactions costs. Predictability and transparency will mean that debtors will have a clear, confident idea of the actions required of them, increasing the probability that they will undertake these while at the same time avoiding unnecessary measures (“overkill”) and the costs that may be associated with those.

There is also the possibility that greater predictability will reduce the so-called overhang effect, whereby the uncertainties associated with a large debt stock are alleged to depress private capital formation, particularly investment from abroad.5 There is some debate over the relevance of the debt overhang hypothesis to the circumstances of low-income countries. The paper in this volume by Elbadawi and others gives evidence for taking it seriously. For whatever reason, it is clear that the boom in foreign direct investment to developing countries of recent years has largely missed many of the poorer countries.

A further consideration is that the transparency of debt relief mechanisms, and the rules they are based on, are likely to have a positive influence on the fairness of treatment of different debtor countries, in the sense that any two debtor countries in comparable situations should receive like amounts of relief. This is not only intrinsically desirable but may also enhance efficiency by reinforcing debtor commitment to their obligations.

However, by setting up our four ideal criteria of adequacy, productivity, cost minimization, and transparency, we are in danger of appearing to set impossible standards, against which any actually feasible arrangements are bound to score poorly. We should therefore acknowledge now that to these criteria can be added a fifth: the political acceptability of debt relief mechanisms to all major parties, particularly to the chief creditor governments. There is a potentially acute tension here, with technically optimal solutions being politically unacceptable, and the politically acceptable falling far short of what would be technically efficient. We will return to this theme.

An Assessment of Existing Mechanisms

Following the sequence of the criteria in the last section, we will now assess past and existing mechanisms. First, however, the various mechanisms that have operated since the mid-1980s will be briefly described, to put the present situation into its historical context.

Mechanisms Described

The early-1980s response of official and private creditors to the debt difficulties of developing countries was to reschedule the interest and principal payments over a longer period. Arrears in payments were capitalized, at first at nonconcessional interest rates, so that rescheduling increased the debt stock and debt-servicing obligations. As suggested by the figures in Table 1 above, this approach contributed to the rise in the debt stock of HIPCs during the 1980s.

A subsequent response was to refinance existing debts with new loans at more concessional interest rates. The Paris Club introduced this principle in its 1988 Toronto terms. The IMF’s ESAF and the World Bank’s IDA resources have also both been used to refinance past credits provided on more onerous financial terms.

The Brady Plan was designed to encourage more commercial creditors to discount their loans to developing countries and made a significant contribution to improving the debt position of many Latin American countries. Following this. World Bank and certain bilateral donor resources have been used to fund buybacks of commercial debt in low-income countries, usually at a heavy discount relative to the face value of the debt. The Brady Plan, by increasing the securitization of debt, also brought about more debt swaps, whereby the creditor exchanged debt for other assets (including local currency) or for government policy actions. Thus, among many other transactions6 debt has been exchanged for equity in companies (often in connection with privatization programs), or in return for government action to preserve the environment. In low-income countries, where equity investment opportunities are limited, swaps have been used to increase expenditures on social services.

Debt forgiveness has also been encouraged in respect of past bilateral concessional (mainly ODA) credits to low-income countries, as a result of which a substantial volume of loans to African and other poor countries has been written off. Many aid donors have subsequently adopted the principle that additional ODA transfers to such countries should only be in the form of grants, rather than debt-increasing loans.

More recent developments for dealing with the continuing debt difficulties of low-income countries have been aimed at reducing the debt stock. Until the Paris Club introduced its Naples terms in 1994, stock reduction was not one of the options available (although the Toronto terms were intended to apply progressively to the entire stock as each tranche came up for rescheduling, provided a need could be demonstrated). There was previously a possibility for creditors to choose to cancel one-third of the eligible slice of debt, a proportion that was increased to 67 percent under the Naples terms. However, the big break-through with Naples was a commitment, albeit qualified and conditional, to consider a reduction in the entire stock of a debtor’s eligible debt, not just a slice of it.

The most recent development has been to bring multilateral tenders within debt relief (as distinct from refinancing) mechanisms. A central feature of the HIPC Initiative is the creation of a Trust Fund (administered by the World Bank) to be used for the purchase of debt (or payment of debt service) for qualifying countries. Use of this Fund is allowed only after a Paris Club deal has been agreed to, with an increase in the concessionality offered by it to an 80 percent reduction in the NPV of the debt. Since one of the Naples options is a reduction in the eligible debt stock, rather than simply in interest payments, there is a possibility of further stock reduction in addition to the reductions in multilateral debts that may be financed by the Trust Fund. Moreover, a precondition for utilization of trust fund resources is that debtors must seek “at least comparable” treatment from other bilateral non-Paris Club and commercial creditors. Thus, the Initiative opens the possibility of action on the debt stock from three different directions.

The IMF is an apparent exception, continuing its past refinancing strategy by providing new credits. However, these credits will be from an enlarged and permanently established ESAF offering either grants or longer grace and maturity periods than at present. It is understood that grant terms are likely to be the normal case in this context, in which case the effect is very similar to debt forgiveness (but without the quasi-legal complications that might arise from forgiveness).

A common feature of all these mechanisms, to be taken up in more detail below, is the prerequisite that debtors must first agree to a high-conditionality IMF program (usually ESAF for poor countries). Bank conditionality is also included in the package of measures under the HIPC Initiative, with debtors required normally to have a track record of implementing policy reforms during six years. We turn now to applying the four efficiency tests described above.

Adequacy of the Relief

The essential point under this heading has already been made: that the recurring need for progressively more concessional debt relief terms—from the post-1982 short-term nonconcessional reschedulings, to Toronto, to “enhanced Toronto,” to Naples, to “HIPC Initiative”—is sufficient testimony to the inadequacy of earlier attempted settlements. The deteriorating debt ratios recorded in Table 1 drive the point further home, as does the continued large-scale accumulation of arrears on past loans, chiefly to bilateral creditors.

Past debt relief measures can be viewed as reflecting a continuous attempt by creditors to marry their desire to recoup as much of their original loans as possible with the reality that many debtors are unable to honor the entire obligation. This has resulted in a series of convulsions, which result in enough creditor governments becoming convinced of the need to go beyond what is at that time available that they, in turn, are able to carry the rest (sometimes kicking and screaming) with them to the next level of concessionality.

The question for creditors has thus been, what is the minimum amount of relief that must be granted to debtors such that the remaining debt-service burden can be paid without recourse to yet further relief? This is a standard problem of optimization under a constraint. However, it has proved a far from simple task to define the parameters of the problem, let alone calculate its solution. Past mechanisms have all suffered from wishful thinking, underestimating the severity of the constraint; none have given sufficient relief, which is why the debt burden today remains so onerous and why pressures became strong enough to overcome the heavy institutional and political resistance to the HIPC Initiative.7

The sources of difficulty may be divided under four main headings: (1) eligibility criteria for debtors; (2) the types of debt included; (3) debt sustainability; and (4) the forecasting of relevant economic variables.

Eligibility has been defined in successive debt relief mechanisms according to income and to the level and nature of the debts owed. Table 2 shows the eligibility criteria for four of the more recent initiatives. The Brady Plan identified eligible countries according to the ratio of debt to GNP and to exports, and of accrued debt service and interest payments to exports. More recently, ceilings for per capita income have been lowered and minimum ratios of debt to GNP have been raised, restricting eligibility at the risk of excluding debtor countries in genuine distress. In the HIPC Initiative, coverage is confined to countries eligible to borrow from the World Bank only on IDA, not “blend” or IBRD terms, and that are not expected to achieve debt sustainability on the basis of preexisting (i.e., Naples) terms. Exclusion of “blend” countries—not necessarily possessing much creditworthiness on financial markets—may give rise to difficulties, especially bearing in mind the inevitable arbitrariness of the dividing line between the IDA-only and blend categories and the way various countries have shifted between these categories.

Table 2.Eligibility and Sustainability Criteria for Some Recent Debt Initiatives
Brady PlanED : GNP > 50 percent

ED : XGS > 275 percent

ADS : XGS > 30 percent

Al: XGS > 20 percent

Three out of four eligibility criteria to be satisfied to qualify of relief.
Toronto termsLow-income

(< $610 per capita in 1990).

Persistent debt-servicing problems and poor balance of payments prospects.

Naples termsGNP per capita < $500.

NPV of ED : XGS > 350 percent.

If the first condition is not satisfied, lower debt relief may still be granted.
HIPC DebtAs for Naples. IDA only.NPV of ED :

XGS = 200–250 percent. EDS : XGS = 20–25 percent.
Other factors, Initiative such as the fiscal burden, may also influence sustainability assessments.
Sources: World Bank (1990, 1993, 1996a).Note: The following abbreviations have been used: ED = external debt; XGS = exports of goods and services; ADS = accrued debt service; Al = accrued interest; NPV = net present value; EDS = external debt service; GNP = gross national product.

All mechanisms have required eligible debtors to be undertaking IMF structural adjustment programs or

Sources: World Bank (1990, 1993, 1996a).Note: The following abbreviations have been used: ED = external debt; XGS = exports of goods and services; ADS = accrued debt service; Al = accrued interest; NPV = net present value; EDS = external debt service; GNP = gross national product.

All mechanisms have required eligible debtors to be undertaking IMF structural adjustment programs or

A problem also remains with the exclusion of certain types of debt. Over the years, the composition of HIPC debt has become more and more concessional, while also becoming more concentrated in loans from multilateral institutions, particularly the IMF, the World Bank, and the African Development Bank.8 Meanwhile, the Paris Club has fought hard to defend the principle that only pre-cutoff-date debt will be considered for relief. Given the preferred creditor status of the multilaterals and the long history of rescheduling negotiations with the Paris Club, the proportion of pre-cutoff debt eligible for rescheduling or cancellation has steadily diminished. Even if 80 percent debt relief were granted by the Paris Club, the effect on the obligations of some HIPCs would be small so long as post-cutoff debt remains excluded.9 Some members of the club remain wedded to this principle, while others accept the inevitability of making concessions, but it appears that the club has signaled a willingness to consider inclusion of post-cutoff debt in particular cases. Another important principle that the club is understood to have conceded is the inclusion of previously rescheduled debt, as part of the HIPC package.

There is also the position of non-Paris Club creditors. These, mainly bilateral official creditors from the former Eastern Bloc countries (of which Russia is much the most important), continue to lie outside the negotiation process. It is estimated that the amount owed to non-OECD creditors is about a quarter of total HIPC debt, and more than one-half for countries such as Angola, Ethiopia, Mauritania, and Mozambique (UNCTAD, 1996), although in practice these are not being serviced. In the past, these have been rather left outside established processes, perhaps because (1) nobody has known the full extent of debt from Eastern Europe and the countries of the former Soviet Union, (2) rapid exchange rate changes have left the actual valuation of the debt in doubt, and (3) the OECD’s focus has been concentrated on the processes of transition within these economies and with dealing their own financial obligations. The HIPC Initiative tries to tackle this group by requiring that debtor countries should “seek treatment on debt owed to other bilateral and commercial creditors on terms at least comparable with those agreed with the Paris Club” (World Bank, 1996b), but this is a retreat from an earlier position that required debtors to obtain equivalent relief. To seek is not necessarily to obtain, so it appears that lack of progress on that front would not be allowed to hold up negotiations with Paris Club creditors. There nonetheless remains a need for such creditors to be brought either within the Paris Club or into a parallel structure that will conform to the same rules.

Should the Paris Club’s 80 percent relief in fact fall short of needs, due to the exclusion of major categories of debt, the resulting burden on the IMF and the World Bank—essentially standing as debt relief providers of last resort—could be significant, especially bearing in mind that the Fund and the Bank originally urged the Paris Club to agree to go as far as 90 percent relief. The HIPC Initiative requires contributions by Paris Club creditors to the Trust Fund such that their share of the total is proportionate to the level of their outstanding claims. The success of the Initiative therefore hinges on the extent to which Paris Club creditors will contribute to the Trust Fund, raising questions about the volume of resources that creditors are willing, in total, to devote to HIPC debt relief. It is also contingent on maintaining the level of IDA lending by the World Bank, an issue that is currently in question because continuing arrears on payments into IDA-10 by the United States are undermining the agreement arrived at for IDA-11 and creating doubts among some other donors about the possibilities of negotiating an IDA-12 agreement.

The issue of eligibility criteria is fairly closely tied up with that of sustainability. By implication, countries whose debts are large enough to bring them within the eligibility rules are regarded as being in an unsustainable situation. Table 2 indicates that most past packages have not explicitly considered sustainability. It was not until the adoption of the Naples terms that it was accepted that debt relief mechanisms ought to offer debtors the possibility of an eventual exit from negotiations through debt stock reductions. However, there was a crucial ambiguity here, for, having willed the end, it did not appear that the creditors would will the means, that is, to provide sufficiently generous settlements to offer debtors realistic prospects of escaping the treadmill of repeated debt renegotiations. To secure a debt stock reduction under Naples terms, debtors had to be able to convince creditors that this would, indeed, be their final request for relief—a task that, however, appeared objectively impossible for many of them given the major classes of debt excluded from the Naples terms. On an uncharitable reading, the exit provisions under Naples could have been designed as a device for limiting the number of countries receiving debt stock reductions.

The HIPC Debt Initiative much improves on this situation by widening the coverage of eligible debt to include multilateral debt, by setting out explicit criteria for sustainability, and by offering a process through which relief may be afforded to eligible debtor countries such that their debt is reduced to these sustainable levels. In addition, specific targets within the sustainability ranges set out in Table 2 are to be determined “in the light of country-specific vulnerability factors, such as the concentration and variability of exports, and with particular attention to fiscal indicators of the burden of debt service” (World Bank, 1996b).

Fiscal indicators (an important element in the criterion of adequate relief identified earlier) have not been featured in previous packages, and their inclusion thus moves mechanisms further toward adequacy, even though they are in addition to, not in replacement of, the conventional indicators listed in Table 2. It has become more and more acknowledged that the heavy concentration of HIPC external debt held by the public sector places a burden not only on the balance of payments but on the public finances, and IMF research (IMF, 1995) has shown that there is a less than perfect correlation between debt-related difficulties on the balance of payments and public finances. In 1994, scheduled external debt service was equivalent on average to about 70 percent of HIPC government current expenditures (IMF, 1995, Table A7).

Assessments of sustainability have been complicated by the weakness—and systematic tendency to overoptimism—of forecasts of debt-servicing capabilities, especially relating to export growth. Such fore-casts are essential for judging future financing gaps and resulting needs for debt relief. However, some past forecasts have been massaged so that financing gaps accord with creditors’ negotiating positions, rather than to a more objective view of debtors’ economic prospects (Martin, 1991, p. 259). This has caused programs to move off course almost before they have begun. Killick (1995, p. 14) shows, in the context of their multilateral indebtedness, how the number of countries classified as in need of relief varies greatly according to chosen assumptions about export growth. Past papers by the IMF and World Bank themselves have given widely varying numbers of countries at risk, depending on the criteria employed (see also Martin, 1996, Table 1, and the paper by Elbadawi and others in this volume). Improved realism is desirable in making forecasts of future ability to pay, even though the short-term (but not long-term) effect will be to increase the resources required for debt relief.

Quite apart from the tendency to systematic bias, the accuracy of forecasts for countries vulnerable to external shocks will always be imperfect, and the likelihood of inaccuracy will be compounded the longer the period of the forecast. This brings into question the use of sustainability classifications based on projections of ability to finance debt during ten years or more, as used in the latest Initiative (Martin, 1996, p. 6). Not only are such classifications highly sensitive to the parameters of the model used, but they are also vulnerable to manipulation, to the introduction of small changes in the parameters to include or exclude debtors. Involvement of debtors themselves in producing sustainability forecasts has so far been minimal. Here again, the HIPC Initiative moves in the right direction, by including debtor country officials in the preparation of the debt sustainability analyses that will determine countries’ eligibility for the HIPC terms. This improvement will not alter the inherent uncertainties surrounding such forward-looking exercises, however.

The adequacy of debt relief arrangements will be partly determined by their success in inducing additional inflows of private capital, as discussed above, although it is easy to exaggerate the likely effect of the existing debt overhang on new flows. Few would suggest that removal of the debt burden would lead to a flood of foreign investment.10 Additional factors, such as the state of the economy (which may or may not be affected by the debt overhang), resource endowments, and political stability, will also be important. In the case of ODA, budgetary conditions within donor countries will be critical. Unfortunately, past debt relief does has not obviously generated large amounts of new finance for low-income countries. Net flows of foreign direct investment remain low for SILICS relative to other developing countries, at only $2.2 billion in 1994 (World Bank, 1996d, p, 220), although this has risen since 1992. Most inflows in recent years have been in the form of ODA. In total, however, this is in decline, with 14 out of 21 OECD-Development Assistance Committee donor countries recording reduced volumes of ODA in 1994–95 (Development Assistance Committee, 1997, Table 6a). This trend is apt to add to the difficulties of IDA described above.

Given the weak prospects for private capital inflows, HIPCs will continue to rely on official bilateral and multilateral sources to finance future adjustment gaps and development expenditures. The HIPC Initiative effectively recognizes this, and the World Bank has committed itself to providing positive net transfers to eligible countries (World Bank, 1996c). The success of the Initiative will partly hinge on the extent to which it results in new money, through the proposed extension of ESAF and the Trust Fund, over and above previous lending through ESAF and IDA. Unfortunately, the amount of new money so far in prospect under the Initiative appears modest. It seems more likely than not that the IMF will eventually be allowed to sell a proportion of its gold reserves and utilize income generated by investment of the proceeds for new softer ESAF lending. It may well also be the case that more resources for this group of countries will be made available by the World Bank than it would otherwise have done. Both these sources will be genuine new money, but their envisaged scale appears limited. However, Martin (1996, p. 21) makes the point that disbursement of aid through the HIPC Initiative is likely to be quicker and more reliable than is normally the case with conventional program and project aid, so that available ODA may become more fully utilized.

A final factor in considering the adequacy of present arrangements concerns the way debt burdens are calculated. Under the HIPC Initiative, these are calculated in NPV terms. Using NPVs is appropriate when a valuation is desired that reflects different degrees of loan con-cessionality. However, what matters to debtor governments’ financial managers is the magnitude of real-money current servicing commitments relative to available fiscal and foreign exchange resources. Through its commitment to maintain positive net transfers, the World Bank implicitly recognizes the cash flow need. But it is only one of a number of players, and not all the others may take the same position. Indeed, the IMF has long been a net recipient of financial transfers from low-income countries, and it is not known to have entered into any commitment to reverse that situation. The position of the AfDB should also be mentioned. Its past loans bulk large in the debt problems of some African HIPCs, but the financial problems of that cash-strapped institution remain unresolved, and this remains a source of weakness in an Initiative intended to be comprehensive.

To sum up on adequacy, it is self-evident that earlier mechanisms offered inadequate relief, which is why a succession of progressively more concessional packages have had to be devised. Creditors have been reluctant to face up to the financial implications of debtors’ parlous economic situations and prospects, only gradually and partially coming to terms with these. The HIPC Initiative is best viewed as the latest of such moves, albeit one of particularly large and positive significance. However, there remain questions about the adequacy of even this and doubts about whether it offers realistic prospects of an end to HIPC debt problems. A great deal hangs upon the spirit in which the Paris Club creditors choose to implement the agreement, for they retain much scope for interpretation, and upon the volume of supporting resources that is forthcoming. For all the progress that it marks, there remains a real risk that the Initiative will be found wanting and will need to be succeeded by a yet more inclusive and concessional scheme.

The Productive Use of Relief Resources

We argued above that the quality of policies in debtor countries was a key determinant of the productivity of the additional resources made available through debt relief, expressed in terms of developmental and poverty-reduction objectives, as well as in terms of strengthening future debt-servicing capabilities. A key issue, therefore, is the extent to which debt relief mechanisms have been able to discriminate effectively among debtor governments according to the appropriateness of their policies. This is an important issue because the position of the World Bank is that there is virtually no relationship across countries between the level of external debt and current policy performance.

As was pointed out, conditionality has been the creditors’ chosen modality for achieving such discrimination, with debt relief conditional upon debtor agreements with the IMF (which, in turn, are often associated with World Bank structural adjustment programs). By this means, creditors have attempted to overcome the moral hazard problem. The utilization by bilateral creditors of the policy-design and -monitoring capacities of the Fund and Bank is certainly more efficient than each creditor setting its own policy conditions and undertaking its own monitoring, but unfortunately the adjustment programs of the Fund and Bank have had an unconvincing record in achieving their objectives in the circumstances of poor countries.

To some extent, the very persistence of acute debt difficulties in many of these countries is enough to demonstrate this. As is well known, almost all HIPCs have undertaken IMF-World Bank supported adjustment programs during the last decade—typically several such programs each—but results have been generally disappointing. While the picture is not uniformly adverse, neither the IMF nor the World Bank has been able to show strong results from its programs in low-income countries.11 Economic growth has been little changed, in either direction, and the same is true of inflation. Programs have been associated with improved export performance (and hence a tendency to strengthen debt-servicing capacities) but, against this, they have also been associated with significantly reduced investment levels.

A dramatic demonstration of the ineffectiveness of adjustment programs in Africa is provided in an exercise published in a World Bank study, Adjustment in Africa (World Bank, 1994a), as subsequently refined and updated by Bouton and others (1994). The most startling result of the original study was that the Bank viewed only one country as having achieved even “adequate” macroeconomic policies by 1990–91 even though over the period in question this group of countries had between them had no less than 95 high-conditionality IMF credits and 93 Bank structural adjustment credits. The paper by Bouton and others threw further interesting, if unencouraging, light on the subject. The authors undertook some reclassification of countries’ 1990–91 policy stances on the basis of refinements in methods and revised data, and a reclassification based on 1992 information. Their results, which are confined to sub-Saharan Africa, are set out in Table 3, and to these we have added indications where countries had had IMF and Bank adjustment programs during 1992.

Table 3.Macroeconomic Policy Stances in Sub-Saharan Africa and the Incidence of Structural Adjustment programs
AdequateGambia, TheGambia, The
FairBurundiBurkina Faso
Malawi↑↑ Mozambique*
Burkina FasoCentral African Republic
Gabon↓ Madagascar
Very poorCongoCameroon*
CÔte d’lvoireCongo*
CameroonCôte d’lvoire*
Mozambique↓↓ Kenya*
Sierra Leone↓↓ Malawi*
Sierra Leone#
Source: Bouton and others (1994, Table 17), and World Bank data.

Countries shown in italics had a World Bank adjustment program for at least six months during 1992. Arrows indicate movements during 1992 (two arrows indicate a movement by two categories, such as from “fair” to “very poor”). Asterisks denote countries that agreed to IMF standby, extended facility, or ESAF programs.

denotes countries ineligible for IMF credits during 1992 as a result of being in arrears with payments on previous credits.

Source: Bouton and others (1994, Table 17), and World Bank data.

Countries shown in italics had a World Bank adjustment program for at least six months during 1992. Arrows indicate movements during 1992 (two arrows indicate a movement by two categories, such as from “fair” to “very poor”). Asterisks denote countries that agreed to IMF standby, extended facility, or ESAF programs.

denotes countries ineligible for IMF credits during 1992 as a result of being in arrears with payments on previous credits.

For 1990–91 the results of the earlier study were largely confirmed, with relatively minor changes in the classifications. However, comparison of the 1990–91 and 1991–92 results shows that:

  • There was an overall deterioration in policies, with more countries in the bottom group and fewer with fair policies. Eight countries moved down (2 all the way from fair to very poor), while only 2 moved up.
  • Twelve of the 15 bottom countries had Bank adjustment programs in 1992, while only 5 of the top 10 countries had programs. Furthermore, 8 of the eligible 13 bottom countries (i.e., excluding Sierra Leone and Zambia) had programs with the IMF for the same period.
  • Seven of the 8 countries whose policy stances are shown by the Bank’s criteria as having deteriorated had Bank programs in 1992, indicating that program conditionality was unable to prevent policies worsening. Moreover, of these 7, 5 also signed programs covering 1992 with the IMF.

This analysis shows the ineffectiveness of the adjustment conditionality rather starkly. It would be impossible to predict the likely direction of change in the quality of country policies from information about IMF-World Bank adjustment lending. It appears that creditors’ reliance on conditionality to ensure that debt relief will be well used is misplaced.

The essence of the problem lies with the weak implementation of many adjustment programs. The reasons for this are manifold and it would take us well beyond the boundaries of this paper to go into them. However, one aspect demands attention here, namely the problem of adverse selection: the danger that the IMF and the World Bank feel they must continue lending to poorly performing countries to protect past credits, to avoid the danger of these countries falling into arrears or defaulting altogether. In recent years this danger has become greater, as obligations to the two institutions have assumed a growing proportion of the total external indebtedness and debt-servicing obligations of low-income countries. Until the HIPC Initiative, the only way in which the IMF and the World Bank could respond to the growing debt-servicing difficulties of some of their clients was through refinancing: maintaining a sufficient flow of new lending to debtor countries to ensure that they could continue to service past credits. In the case of the IMF, that will continue to be the case even under the HIPC Initiative, with its reliance on an enlarged and softer ESAF. Moreover, the problem of adverse selection is likely to be worsened by pressures from Paris Club creditors anxious to retain the appearance that debt agreements are remaining on track, even when the reality is otherwise. This has been a potent factor in the past.

Lending driven by such motivations undermines the credibility of policy conditionality, for a borrowing government will be aware that the institutions’ anxiety to make sure it remains current in servicing past loans will allow it to renege on policy commitments with limited risk of financial penalties. The IMF and the World Bank deny that their lending decisions are determined by such considerations, but this denial is short on credibility. Indeed, Fund-Bank debt projections have shown that new flows must be skewed in favor of those with large multilateral debts for the outcome to remain manageable (IMF and World Bank, 1995). The extent to which lending is biased by debt considerations is difficult to demonstrate, but a simple test of the hypothesis that existing multilateral debt servicing drives new IMF-World Bank lending produced statistically strong results consistent with that hypothesis.12

The tendency for debt relief mechanisms to be overreliant on an ineffectual conditionality in attempting to ensure that new finance is used efficiently has been compounded by donor responses to improved debtor economic performance that have weakened incentives to undertake difficult reforms. The approach mentioned earlier, with creditors trying to minimize the amount of relief that must be granted, appears to have dominated so that, rather than rewarding improved economic performance with debt relief, Paris Club creditors have tended to harden their negotiating stance in order to appropriate a share of the gains for themselves (Martin, 1991, Chapter 3, and 1996, p. 4). To the extent that debt relief has been calculated on the basis of optimistic economic forecasts, the reward for hitting the target has been to pay more in debt servicing, coupled with requirements to avoid further refinancing and arrears. The implied marginal tax rate on improved performance may be very high.

Thus, experiences with the level of relief suggest that in determining the terms of country agreements more incentives should be given for policy reform and improved economic performance. The HIPC Initiative apparently incorporates such incentives. This may be inferred from the provision that where actual debt ratios turn out to be more favorable than those targeted “due primarily to exogenous factors” the level of debt relief could be reduced (World Bank, 1996b; our emphasis). The implication is that favorable outcomes due to improved policies will not be penalized by reduced relief, a point made more explicitly in an earlier World Bank paper (1996a, p. 2). This document also stresses that multilateral support for the Initiative should only be reduced if improvements in performance are due to exogenous factors, not if they are a response to policy changes. However, there remains a danger that, whatever the good intentions now, Paris Club creditors faced by a debtor with a better-than-projected outcome will attempt to appropriate some of the benefits for themselves.

Another criterion for efficient use of relief resources is consistency between decisions about debt relief and the allocation of ODA. It has been a feature of debt negotiations to date that they have generally excluded representatives of bilateral aid agencies, being typically dominated by treasury and central bank officials. At the same time, those responsible for Paris Club negotiations are typically poorly represented on aid coordination forums, such as consultative groups. This bifurcation has obstructed the development of an integrated and internally consistent approach to the financing needs of debtor countries, although it is a more serious weakness in some donor-creditors than in others (Japan and the Netherlands are mentioned in this connection, and it has also been a problem with the United States in the past).

One consequence of bifurcation is that aid receipts are commonly treated by creditors as a government revenue item, permitting the servicing of more external debt than would otherwise be affordable. Creditor governments have been taking away with one hand what they have given with the other. Often this substitutability between new aid and debt repayment has been direct and open. It has been no secret that new concessional finance by the IMF and the World Bank has been directed to repayment of commercial debt and nonconcessional IMF-World Bank debt, diverting the funds from investment and social projects. Program (structural adjustment) support has also frequently been diverted to debt repayment, seen as a more predictable and fast-disbursing approach to providing assistance (Martin, 1996, p. 21).

The HIPC Debt Initiative makes clear progress in these areas. By bringing multilateral debts into the equation and by its provisions for tripartite debt sustainability analyses based on an overall assessment of a country’s economic and financial situation, it facilitates a more comprehensive and coordinated approach, although it is still reliant on the actions of Paris Club creditors whose deliberations are still likely to suffer from the bifurcation problem. And by making a more concerted effort to provide an exit route, the Initiative will, if successful, reduce the problem of creditor stances that shift according to the economic performance of the debtor.

However, other problems remain. In particular, a central feature of the HIPC Initiative is that countries should normally have followed a program of policy reforms for six years before becoming eligible for the full relief. Only then will full Paris Club and, if necessary, ESAF and Trust Fund terms be granted. This is a tall order, given the problems with program implementation so far and the weakness of provisions to deal with exogenous shocks, and could severely restrict the number of countries qualifying. Moreover, the Initiative makes no provision for changes in the modalities of adjustment programs to facilitate recipient ownership. Reliance on a discredited conditionality, and the likelihood of adverse selection, will continue and perhaps be intensified.

Minimization of Transactions Costs

The level of debtor-government transactions costs is influenced by two main factors: the number of parties the debtor government must separately negotiate with; and the number of times the debtor must return to the negotiating table with each party. The influence of these factors is enhanced to the extent that each negotiation may require new documentation and data. The existence of institutions such as the Paris and London clubs helps to reduce the number of negotiations, since debtors do not need to negotiate relief terms independently with each creditor. However, many additional negotiations come outside the collective meetings. Killick (1993, p. 5) lists the following:

  • Accords with individual bilateral creditors, to be negotiated after an overall agreement with the Paris Club has been reached. This has been particularly onerous in the past because of the creditors’ short-leash approach, although familiarity with the process has helped to streamline the procedures.
  • IMF missions, which include annual consultation exercises, preparatory and negotiation missions for new programs, and three-to-six-monthly review missions for programs that are already in place.
  • World Bank missions. There is generally a larger number of these than in the case of the IMF because of the more numerous and varied nature of Bank lending operations.
  • There may be negotiations with creditors outside the Paris and London clubs, in particular the governments of the former Eastern Bloc countries and of OPEC, and nonbank commercial creditors.
  • Consultative Groups (or similar) for ODA coordination (usually annual).
  • ODA negotiations with individual bilateral and other multilateral donors, such as regional development banks and UN agencies.

The heavy demands created by such a plethora have, of course, been aggravated by the inadequate, convulsive nature of past approaches, as described above, for this has forced debtors to negotiate many patently inadequate “settlements” only to have to quickly return to the negotiating room for yet another temporary fix.

The total number of negotiations of the various types listed above was estimated for 30 African governments to have been in excess of 7,800 during the period 1980–92 (Killick, 1993, p. 5); updating this figure could scarcely leave it at below 10,000 negotiations. Each of these, moreover, has different, sometimes onerous, data requirements and, while the situation has improved somewhat, with assistance in debt management systems and the computerization of relevant departments, the total requirement is nonetheless extraordinary. Negotiations make huge demands on the time of the most senior and able officials, diverting them from more constructive analysis and implementation of policy options. Often they have little time for anything other than this incessant round of negotiations and with the juggling of day-to-day finances. The situation is not helped by the tendency for high staff turnover within external agencies. Much time may be spent with creditor-donor officials familiarizing them with the local realities, only for them to move on after a year or two.

The failure of past mechanisms to provide an exit route from debt negotiations has increased avoidable transactions costs. Each time a program has broken down or rescheduling has left a debt level that is still unsustainable, a new set of negotiations has had to take place. The HIPC Initiative will not initially reduce these costs. IMF and World Bank programs remain a central feature, so there will continue to be regular country missions and reviews. Should programs fail, or the initial projections upon which sustainability calculations are based prove overoptimistic, a whole new set of negotiations may be necessary. Moreover, the HIPC arrangements are extraordinarily complex (to an extent that is only partially conveyed in this paper)—a necessary price, no doubt, for bringing everyone on board—and will be apt to add to transactions costs. However, by incorporating the principle of debt sustainability, the Initiative does raise hopes that some countries will eventually be able to exit from negotiation processes, with corresponding reductions in transactions costs, albeit after a lengthy interim period of reform and intense scrutiny.

Transparency and Predictability

Past mechanisms have not always been characterized by transparency or predictability. There have been cases where a country that apparently qualified for particular Paris Club relief was not granted it, and of countries being excluded because to include them would necessitate a volume of relief that the creditors were not prepared to contemplate. Mechanisms have allowed eligibility criteria to be adjusted to include or exclude certain countries favored or disfavored by creditors. One of the consequences of the past lack of transparency is that negotiations have tended to be further prolonged, increasing transactions costs. New disbursements have been delayed by the absence of agreement, while in the interim interest has accumulated on debt in arrears (Martin, 1996, p. 16).

For all its complexity, the HIPC Initiative represents a major improvement in transparency. Procedures, criteria, and the responsibilities of the various parties are written down in considerable detail. The specification of sustainability criteria, summarized in Table 2 above, ought to make the amount of relief to be expected more predictable. However, there have in the past existed eligibility criteria for relief, but these have still left room for unequal treatment of different debtors, and such scope remains with the HIPC Initiative. There remains a good deal of creditor discretion, allowing varying interpretations. Among the principles of the Initiative, insistence on a case-by-case approach and on a debtor’s ability to put additional relief to good use, leave room for inconsistency, even though the basic sustainability indicators are established. How much policy reform will constitute a sufficient track record? While it is clear that failure to implement each and every program condition will not necessarily lead to a total breakdown, debtor governments will be uncertain about the consequences of a failure to meet particular conditions. Given the past record with adjustment programs, this problem is likely to arise quite frequently.

The retention of this degree of discretion, while it has the merit of retaining flexibility in the arrangements, creates possibilities for undermining the general eligibility rules and for the unequal treatment of debtors on political grounds, of which there have been examples in the past (Poland, Egypt) (Martin, 1991, and 1993, pp. 27–32). Moreover, if a major Paris Club creditor chooses to take a hard-line stand against a particular debtor country, that may harden the position of the whole group (the lowest common denominator problem). The case-by-case approach, while desirable on other grounds, increases the difficulties of achieving full transparency. There is hence a trade-off between the desirability of a clear set of rules that allow equal treatment of countries in similar positions and that inform debtor governments of precisely what is expected of them, and the need to retain flexibility to take into account new or country-specific factors that may affect future debt sustainability.

Conclusions and Directions of Improvement

This paper has examined existing mechanisms to deal with the external debt of low-income countries by suggesting a set of efficiency criteria and examining the mechanisms used so far against these criteria. The overall conclusion is that past debt relief mechanisms have been seriously deficient in a number of respects, that the HIPC Initiative substantially improves the overall efficiency of approaches to the debt difficulties of poor countries, but that shortcomings remain. These are such that overall debt sustainability may still not be attained, necessitating one or more further convulsions among creditors to finally bring them to terms with the limited debt-servicing prospects of the HIPC debtors.

More specifically, on the adequacy of relief, we have argued that earlier mechanisms offered inadequate relief. The HIPC Initiative makes significant progress by (1) explicitly defining debt sustainability indicators and increasing the possibilities of exit, and (2) accepting the principle of multilateral debt stock reduction or further concessionality. However, there remain many questions about the adequacy of even this, particularly about the position of some of the countries excluded, and doubts about whether it yet offers realistic prospects of an end to HIPC debt problems. Much will depend on the spirit in which it is interpreted by Paris Club and multilateral creditors, and on the accompanying volume of supporting ODA.

Past productivity of use of relief resources has been marred by overreliance on IMF-World Bank conditionality, which has been unevenly implemented, and the consequential weak economic results obtained. This modality has given a false impression of providing an effective screening device, whereas it has in fact led to a rather undiscriminating provision of assistance to governments, some of whom have not made good use of it. Creditors have sometimes compounded the difficulties by hardening their negotiating stances when debtor performance improves, while little finance has been available to overcome adverse exogenous shocks. The HIPC Debt Initiative will do nothing to reduce reliance on conditionality and may even increase it—the reverse of what efficiency would require.

Past short-leash mechanisms for dealing with HIPC debt problems, the inadequacy of the relief provided, and the resulting necessity for repeated rounds of negotiations, have led to high and avoidable transactions costs, both by diverting scarce manpower from policy tasks and because of their large informational requirements. The HIPC Initiative is intended to reduce these costs by providing stock reductions and an improved exit route, but there are doubts about the extent to which this will be realized, while the Initiative itself adds new layers of complexity.

Criteria upon which decisions are made have in the past been neither fully transparent nor predictable, with suspicions that criteria have been altered or interpreted to exclude or include particular countries. While the HIPC Initiative goes much further in establishing transparent criteria, procedures, and responsibilities, its scope for discretionary interpretation remains substantial. Consequently, debtor governments will approach negotiations in some uncertainty about what is expected of them and the prospects for equitable treatment.

All in all, then, there remains a sizable gap between an ideally efficient mechanism and what is available. This largely reflects the acute tension between the “technocratic” efficiency criteria applied in this paper and the “political” imperatives of securing agreement among numerous creditors. Political constraints have resulted in an inefficient succession of inadequate relief packages that have gradually come closer to a recognition of the limited debt-servicing capabilities of the low-income debtors through a sequence of convulsions. The political constraints are such that some degree of technocratic inefficiency is probably inevitable, but creditors do also have an interest in creating efficient mechanisms. The transactions costs of past failures have probably been substantial and have by no means all fallen on debtors. No less serious, much public money has been wasted by the provision of financial support for doomed settlements. The remaining task, therefore, is to identify areas where there are particularly acute tensions and where efforts at improvement might be concentrated. The following recommendations are offered;

(1) All debt should be included. While the HIPC Initiative recognizes the need to solve the debt problem through the coordination of all creditors, the mechanisms by which this will be ensured need to be more clearly spelled out. In particular, mechanisms need to include non-Paris Club creditors, whose claims are a significant proportion of the debt owed by some low-income countries.13 Within the Paris Club, the cutoff date must be advanced (or waived), especially for countries that first began negotiations many years ago. Innovative solutions are needed to cancel the remaining ODA claims held by donors such as Japan; and to resolve the position of the AfDB. Debts should be included irrespective of their original concessionality It is necessary to include all debts because it is the entire stock that is unsustainable for many countries, and it is impossible to devise a scheme for reducing debt to sustainable levels except on that basis.

(2) Mechanisms should provide sufficient relief to provide a realistic probability of exit from negotiations. For some countries, this is likely to mean greater relief than presently available. There needs to be fuller recognition of the vulnerability of low-income countries to negative shocks, and of the uncertainties inherent in debt sustainability forecasts, through the provision of contingency relief sufficient to withstand these eventualities. At the same time, further encouragement must be given to vulnerable countries to provide themselves with a cushion against negative shocks. The payback from successful exit lies not only in the cash flow relief but also in eliminating the transactions costs associated with debt and related negotiations.

(3) Overreliance on IMF-World Bank conditionality should be avoided. As an alternative, a new model of creditor-debtor relationships is urged, based upon the principles of ownership, selectivity, support, and dialogue.14 The “ownership” principle asks that donors should take their own rhetorical support for this more seriously and abjure from supporting creditor-owned programs. “Selectivity” requires that debt relief and program aid be limited to countries with governments that have adopted for themselves policy strategies that are likely to accelerate growth, reduce poverty, and enhance debt-servicing capabilities. “Support” refers to conventional financial support for adjustment programs, extended to improved contingency financing and debt relief for reforming governments, and to technical assistance (also delinked from conditionality) for governments requesting this, to raise their own policy capabilities. “Dialogue” requires donors-creditors to reorient themselves to concentrate on the exercise of influence, and to maximize the number of means through which that might be exerted. The use of conditionality should not be excluded entirely, but creditors should be reluctant to use this unless there are specific grounds for believing that it will produce good results. Conditionality should be the exception, not the rule.

(4) Eligibility and sustainability criteria must be clearly defined and based on more rigorous analysis of the concepts. Using criteria based on past cross-country performance does not sufficiently take into account the particular difficulties that might face individual debtor countries. Criteria other than the ratio of debt to GNP and debt-stock or debt-service ratios should be used more explicitly as alternative tests. These should include fiscal indicators. However, given the developmental and poverty-reducing objectives of the creditor governments in their role as aid donors, indicators of the developmental and social consequences of debt and debt-reducing measures should also be brought explicitly into the negotiations and given weight. Sustainability calculations should include cash flow as well as NPV valuations; the six-year transition period before the full provision of relief may be too long for countries with serious immediate cash flow difficulties.

(5) Eligibility and sustainability criteria must be applied transparently and predictably. There has been a tendency in the application of past mechanisms to accept or reject relief to certain countries for reasons that are unclear. As a consequence, implementation of relief terms has appeared inconsistent. If a country that meets the stated criteria for relief is rejected during negotiations, the reasons should be clearly stated and the principle of precedent applied so that debtors in future negotiations know where they stand.

(6) Mechanisms should aim to coordinate the activities of financial and aid agencies. This would best be achieved by merging debt relief and ODA coordination processes, with both considered together, to form a comprehensive package of financial assistance against a demonstration of need.

(7) Greater effort must be made to include debtor governments at all stages of the process, particularly in devising economic reform programs, providing for greater consensus on objectives, targets and appropriate policy measures, and local administrative capabilities.15 While this may slow down the process, it is likely to improve the effectiveness of the outcome. A recent specific effort with attractive ownership properties is the multilateral debt fund established by the government of Uganda in 1995 (Government of Uganda, 1995), The possibilities of replicating this and bringing it within the HIPC mechanisms should be further studied.

Past mechanisms for dealing with the external debt of low-income countries have been flawed but have progressively edged toward a viable solution. The HIPC Initiative is a major step toward reducing the outstanding debt of low-income countries to sustainable levels and making recurring debt negotiations a thing of the past. The suspicion remains, however, that the course of events will reveal a need to go yet further.


    Bacha, E. L., 1990, “A Three-Gap Model of Foreign Transfers and the GDP Growth Rate in Developing Countries,”journal of Development Economics, Vol. 32 (April), pp. 279–96.

    • Search Google Scholar
    • Export Citation

    Bleaney, Michael F., 1996, “Macroeconomic Stability, Investment and Growth in Developing Countries,”journal of Development Economics, Vol. 48 (March), pp. 461–77.

    • Search Google Scholar
    • Export Citation

    Bouton, Lawrence, ChristineJones, and MiguelKiguel, 1994, “Macroeconomic Reform and Growth in Africa: Adjustment in Africa Revisited,”Policy Research Working Paper No. 1394 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    Development Assistance Committee, 1997, Development Cooperation 1996 (Paris: Organization for Economic Cooperation and Development).

    Government of Uganda, 1995, “Proposal for a Multilateral Debt Fund” (Kampala: Government of Uganda).

    Hewitt, Adrian P., ed., 1994, Crisis or Transition in Foreign Aid?ODI Special Report (London: Overseas Development Institute).

    IMF, 1995, Official Financing for Developing Countries, World Economic and Financial Surveys (Washington: International Monetary Fund, December).

    • Search Google Scholar
    • Export Citation

    IMF, and World Bank, 1995, “Issues and Developments in Multilateral Debt and Financing for the Heavily Indebted Poor Countries—Further Considerations” (Washington: International Monetary Fund and World Bank, March).

    • Search Google Scholar
    • Export Citation

    Isham, Jonathan, and DanielKaufmann, 1995, “The Forgotten Rationale for Policy Reform: The Productivity of Investment Projects,”Policy Research Working Paper No. 1549 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    Killick, Tony, 1993, “Enhancing the Cost-Effectiveness of Africa’s Negotiations with Its Creditors,”in International Monetary and Financial Issues for the 1990s, Research Papers for the Group of Twenty-Four, Vol. 3 (New York: United Nations).

    • Search Google Scholar
    • Export Citation

    Killick, Tony, 1995, “Solving the Multilateral Debt Problem: Reconciling Relief with Acceptability,” report prepared for the Commonwealth Secretariat (London: Commonwealth Secretariat).

    • Search Google Scholar
    • Export Citation

    Killick, Tony, 1996, “Principals, Agents and the Failings of Conditionality,” paper for Development Studies Association (London: Overseas Development Institute, September).

    • Search Google Scholar
    • Export Citation

    Killick, Tony, forthcoming, Conditionality and the Political-Economy of Policy Change in Developing Countries (London: Routledge).

    Krueger, Anne O., 1987, “The Importance of Economic Policy in Development: Contrasts Between Korea and Turkey,” in Protection and Competition in International Trade, ed. by HenrykKierzkowski (Oxford: Basil Blackwell).

    • Search Google Scholar
    • Export Citation

    London Economics, 1996, “Costs and Benefits Associated with Commercial Debt Buy-back Operations in Sub-Saharan Africa,” report to the Overseas Development Administration (London: Overseas Development Administration, July).

    • Search Google Scholar
    • Export Citation

    Martin, Matthew, 1991, The Crumbling Faqade of African Debt Negotiations (London: Macmillan).

    Martin, Matthew, 1993, Official Bilateral Debt: New Directions for Action (Brussels: European Network on Debt and Development).

    Martin, Matthew, 1996, “A Multilateral Debt Facility: Global or National?” report for the Group of Twenty-Four (Geneva: UNCTAD).

    Mosley, Paul, JaneHarrigan, and JohnToye, 1995, Aid and Power: The World Bank and Policy-Based Lending (London: Routledge, 2d ed.).

    Mistry, Percy S., 1995, ‘The Multilateral Debt Problem: Lurching Towards Resolution,” paper for Oxfam Seminar on Multilateral Debt, London, September.

    • Search Google Scholar
    • Export Citation

    Mistry, Percy S., 1996, “Resolving Africa’s Multilateral Debt Problem: A Response to the IMF and the World Bank” (The Hague: Forum on Debt and Development).

    • Search Google Scholar
    • Export Citation

    Overseas Development Institute, 1995, “Poor Country Debt: A Never-Ending Story?” Briefing Paper No. 1 (London: Overseas Development Institute).

    • Search Google Scholar
    • Export Citation

    Reynolds, Lloyd G., 1985, Economic Growth in the Third World, 1850–1980 (New Haven, Connecticut: Yale University Press).

    Sachs, Jeffrey D., 1989, “Conditionality, Debt Relief, and the Developing Country Debt Crisis,” in Developing Country Debt and Economic Performance, ed. by Jeffrey D.Sachs (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    UNCTAD, 1996, Reducing Sub-Saharan Africa’s Debt to Non-OECD Official Creditors: Sharing the Burden (New York: United Nations).

    World Bank, 1990, World Debt Tables 1989–90: External Finance for Developing Countries (Washington: World Bank).

    World Bank, 1993, World Debt Tables 1992–93: External Finance for Developing Countries (Washington: World Bank).

    World Bank, 1994a, Adjustment in Africa: Reforms, Results and the Road Ahead (New York: Oxford University Press).

    World Bank, 1994b, World Debt Tables 1993–94: External Finance for Developing Countries (Washington: World Bank).

    World Bank, 1996a, The HIPC Debt Initiative—Elaboration of Key Features and Possible Procedural Steps (Washington: World Bank).

    World Bank, 1996b, The World Bank and the Heavily Indebted Poor Countries Debt Initiative (Washington: World Bank).

    World Bank, 1996c, World Bank Participation in the Heavily Indebted Poor Countries’ Debt Initiative (Washington: World Bank).

    World Bank, 1996d, World Debt Tables, 1995–96: External Finance for Developing Countries (Washington: World Bank).

The authors are grateful to Peter Mountfield for helpful comments on an earlier draft, but he is not incriminated in what appears here.


The number of countries now classified as HIPCs that were severely indebted rose between 1980 and 1985 from at least 8 to 18. A more complete analysis would look at all countries, rather than just HIPCs, since during the same period countries will have exited as well as entered this group.


These factors may be approximately related to the constraints shown in the three gap model (see Bacha, 1990).


Thus, Krueger’s 1987 comparison of the postwar economic histories of the Republic of Korea and Turkey attributes the greater economic progress of the former, starting from a poorer base, to the superior quality of economic policies in Korea. Within Africa, similar explanations have been given for the contrasting economic fortunes in the 1960s and 1970s of Côte d’lvoire and Ghana (although more recently it is a comparison that goes against Côte d’lvoire), and between Kenya and Tanzania in the same period. At a different level, we may get an idea of the importance of policy from the regular exercises undertaken by the World Bank and some other agencies for forecasting alternative future scenarios for the world economy, finding large differences in expected outcomes depending on the assumptions made about the policies adopted.


For example, see IMF (1995, pp. 57–66) and Martin (1996, pp. 2–3).


See World Bank (1993, Box 3.4) for more examples.


This is being presented as an Initiative of the IMF and World Bank. In its specific provisions and in the key role played by the president of the World Bank, this is accurate enough, but it should not be forgotten that as late as 1993–94 the Fund and Bank were still insisting that there was no significant multilateral debt problem.


The share of multilateral debt in total external debt of sub-Saharan Africa was only 19 percent in 1985, but grew to 22 percent in 1990 and stood at 26 percent by 1994. The share of external debt service paid to multilaterals by sub-Saharan African countries, exduding South Africa and Namibia, grew in the same years from 19.7 percent, to 33,8 percent, and finally to 45.6 percent (Mistry, 1996).


Estimates by the European Network on Debt and Development put the total nominal relief from an 80 percent reduction in eligible Paris Club debt for Uganda, Ethiopia, and Mozambique at 26, 28, and 26 percent, respectively.


An interesting recent study by London Economics (1996) examines the effects of commercial debt buybacks (financed by the World Bank’s Debt Reduction Facility) in Uganda and Zambia. This found that these had no significant beneficial effects on these two countries’ commercial creditworthiness, although the buybacks did bring cash-flow and debt-management benefits.


The evidence on the IMF is surveyed in Killick, 1995, Chapter 3, and on the World Bank in Killick, forthcoming. Chapter 2. On the World Bank, see also Mosley and others, 1995.


Excluding three outlier values (Guinea-Bissau, São Tomé and Príncipe, and Uganda), a simple regression based on average values for 1984–94, with the gross multilateral lending ratio as the dependent variable and multilateral debt-service ratios as the explanatory value, yielded the result:


The t-value was significant at the 99 percent level. For further details, see Killick (1995).


Martin (1996, p. 31) relates how the Consultative Group mechanism proved effective in catalyzing contributions to debt relief from non-Paris Club members.


We are here summarizing proposals set out in more detail in Killick, forthcoming. Chapter 7.


For evidence on how greater debtor involvement has been successful in the case of Uganda, see Martin (1996, pp. 29–31).

    Other Resources Citing This Publication