Jubilee 2000: What will it cost to cancel unpayable debt?

There is no simple answer to how much debt should be cancelled and what it would cost. Jubilee 2000 believes that the amount and mechanism of debt cancellation cannot be imposed by creditors and must be negotiated between creditor and debtor though a fair and transparent process. Thus the manner of calculating “unpayable” debt will vary between countries and regions.

However, in three international conferences in Accra, Rome and Tegucigalpa the international members of Jubilee 2000 have set out very clear demands on what kind and levels of debt must be cancelled.

This paper is an attempt to show the implications of those demands, as well as of demands by other groups, and also to look at the implications of historic debt settlements (particularly in Indonesia and Germany) where development consideration were taken into account. It is not intended to define policy in itself – rather as a supporting document to the proposals that have been advanced within the Jubilee 2000 international movement.

This paper can only give estimates. Without a World Bank full of economists and details of all outstanding loans, we cannot be fully detailed. Thus we make a number of simplifying assumptions to allow us to use published figures. Furthermore, published figures, even in different official sources, vary widely. Therefore our estimates may not be valid for each individual country, but we feel they give a reasonable picture.

This paper only looks a debt service and not at debt stock. This is both political – we believe debt service payments give the only valid picture of the impact of debt – and practical – because it reduces the calculations. We look at debt service both due and paid.

We look at 93 countries. These are countries with a 1995 GDP at PPP less than $4500, plus middle income countries which are severely & moderately indebted. We exclude Europe and the former USSR (because they have a different kind of problem), Eritrea (which has no debt), and non-reporting countries (Afghanistan, Cuba, D R Korea, Namibia and Iraq). Full data for all 93 countries is given in table 7. [Download xls (excel) file]

Debt service (TDS) criteria based on exports (XGS) can be calculated and require no simplification.

We want to look at debt service instead of debt stock, which means we need to convert present value of debt (NPV) to debt service. On average for our countries [1], TDS/NPV = 14%, so we use that.

Many debt service criteria are based on government income (DBR), but consistent data is very hard to find. Indeed, there is also a wide range of revenues accepted by the IMF in ESAF programmes. According to World Development Indicators 1998, table 4.12, for the limited number of countries for which data is available, government revenue ranges from 10% of GDP for low income countries to 23% for upper middle income countries.

Thus, we prefer to use criteria based on gross domestic product (GDP), in particular TDS/GDP. We note:

Two historic debt settlements – Germany in 1953 and Indonesia in 1970 – are particularly important because they were both based on development criteria. Unlike HIPC, which defines sustainability purely in terms of ability to pay without defaulting [4], these settlements defined sustainability on the basis of how much was needed for development and how much was left over for debt repayment. The European economic miracle for the past 45 years shows the wisdom of this decision.

The German and Indonesian settlements were based on annual cash payments, but they can be converted into present-day criteria (see Table 8 for details [Download xls (excel) file]) by taking the average for the first two years (which was also the largest payments):

A number of past agreements contained what are known as “Bisque clauses” which give the debtor government the right to unilaterally suspend or defer payments under certain conditions. [5] Indonesia, for example, was given the option of deferring principal repayments during the first 8 years, [6] Colombia in the early 1940s was able to limit its debt service payments to 3% of government revenue [7], and this is the basis of Jubilee 2000's Tegucigalpa declaration (see below).

One bisque clause is particularly important and will be included in the tests in this paper. The 1953 London agreement on Germany's debt specifically says that Germany should repay its debts only out of a trade surplus, and could halt payments if there was an inadequate surplus. This was based on the realisation that the German debt crisis in the 1930s was caused in part by Germany taking new loans to repay old debts. It explicitly assumes that the creditors will buy German exports in sufficient quantity to allow it to repay debts. This is particularly important in terms of poor country debts, because limitations in market access and deteriorating terms of trade mean that most debtor countries also run a trade deficit. Such a bisque clause in poor country debt might be useful to open northern markets.

If we take the highest end of the World Development Indicators range, DBR/GDP = 23%, we find the same level as was agreed for Indonesia, TDS/GDP = 0.7%. If we take a somewhat lower level, TDS/GDP = 15%, we get the German level of TDS/GDP = 0.4%. So these two historic examples serve as a good model for the Jubilee demand.

Categories ii)-iv) require a debt-by-debt analysis which cannot be done here. Category iv) is dealt with in a Jubilee 2000 paper “Dictators and debt” which suggests that up to a quarter of all poor country debt is based on loans given to oppressive and military regimes, and which might in international law be considered “odious”. [8] This paper looks only at unpayability, but it must be kept in mind that some debtor countries will separately demand cancellation of a range of debts on the grounds that they are not valid.

The first category, defining unpayability in terms of impact on the poor, is taken up in the development and debt section below.

There is much discussion of the cost of the current HIPC programme. But if all 41 countries on the current HIPC list [9] were to repay their debts at the HIPC “sustainable” level [10], then the creditors would profit by $1.2 billion per year [11]. This is because HIPC countries are only paying 55% of their debt service due:

In all, 20 countries could benefit from HIPC and would reduce their debt service payments by $2 bn per year. But 19 countries would have to pay a total of $3.2 bn per year more than they are paying now to reach the HIPC sustainable level.

Thus the “cost” of HIPC is purely notional, because it is a cost of writing off debts which would never be paid. In reality, the creditors would profit.

Table 2 shows that the “break even” point is only reached when HIPC is offered to all 93 countries on our list. Of those, 33 would gain some reduction of the debt service they are now paying, but this would be fully met from other countries paying more of what they owed.

So the starting point, based on present reality rather than theory, is that HIPC as presently designed benefits the creditors.

None of the Jubilee criteria link debt service to exports, and there are growing arguments against using export criteria, both because of falling export prices and because free movement of foreign exchange means that governments cannot directly capture export earning to pay debt service. But we can look briefly at the implication of other export criteria.

Another way of looking at this is to take account of the fact that 50 of the 93 countries have debt service due greater than 20% of exports while 43 do not. This gives the following step-by-step debt reduction:

The German deal (3.5%) would benefit 91 countries (all but 2 on our list) and reduce payments due by $201 bn per year. This is a very large figure, but it is worth keeping in mind that in 1952 Germany argued that it was impossible to even pay at the Indonesian level (6%), and that this was accepted by the creditors. If that judgement remains valid, then debt cancellation on this scale will be required.

Fiscal or earnings criteria are becoming more popular than export criteria. Fiscal criteria are really based on two different criteria, XGS/DBR plus DBR/GDP. The latter figure is either set as part of the package or assumed to be agreed as part of an ESAF programme – ESAF is itself controversial and the linkage of debt relief to ESAF has been opposed by Jubilee 2000, which wants debt relief linked to locally-developed poverty reduction programmes.

We estimate both fiscal and earnings criteria by using the simple measure of XDS/GDP. As noted in the introduction, we look at four levels: HIPC (7.8%), Eurodad (2.8%), Indonesia and Jubilee high estimate (0.7%), and Germany and Jubilee low estimate (0.4%).

Only 19 countries would pay less under the HIPC level, but there are significant gains under other criteria, as this table shows:

Although debt cancellation at these levels involves large amounts of money, it must be remembered that the Indonesian and German settlements are the only ones that provide any historic precedent for having been based on development considerations rather than simply how much money can be squeezed out of the debtor.

Jubilee 2000's Rome declaration does not define what it means for debt service payments to place a burden on the poor. In this section, we make an attempt to link debt service to essential development spending as one was of looking at the burden on the poor, but this cannot be taken as an official Jubilee 2000 definition of this concept.

The starting point is the six targets adopted by the OECD Development Assistance Committee (DAC) in 1996 [13]. The best known of these is the commitment to “a reduction by one-half in the proportion of people living in extreme poverty by 2015”. This has been widely quoted and is officially backed by all donor countries.

The basic assumption is that the spending to meet this target must be made before any debt service is paid.

There are no reliable estimates as to how much spending is required to meet the DAC targets, but we have drawn on three different methods:

Cafod. The British agency Cafod estimates that a country must spend at least $28 per capita on health and education. [14]

HDR 96: UNDP's Human Development Report 1996 [15] uses an econometric exercise to estimate that “a 1 percentage point increase in the average share of GDP invested in health and education is estimated to reduce ... the child mortality rate by 24 percentage points.” One of the other DAC targets is cutting infant and child mortality by two-thirds [16], which would require a transfer of 4% of GDP to health and education. So we add an extra 4% of GDP to present health and education spending and consider this essential.

HDR 97: UNDP's Human Development Report 19967 [17] estimates that an extra $80 billion per year must be spent on the 1.3 billion people living in poverty in order to meet the DAC targets. This is an additional $62 per person per year which must be spent on everyone with an income under the international absolute poverty line, $1 per person per day at purchasing power parity (PPP) at 1985 prices [18]. We add this to present health and education spending and consider it essential.

All three of these approaches have methodological problems. But at present nothing else is available to allow an estimation of the costs of meeting the DAC targets. Therefore, to reach at least a rough estimate of essential social spending (ESS), we simply average the results of these three approaches. [19]

The next step is to ask how much tax a country can be expected to raise. Here we follow the Cafod method [20]. All countries have a tax threshold – a level of income below which no tax is paid. We take the tax threshold at the absolute poverty line, $1 per day, and assume all income below that is not taxed [21]. We assume that above that level, tax can be 25% of income [22].

We assume that essential social spending (ESS) must be taken out of that tax first. But social spending is not the only essential spending. The United Nations Conference on Trade and Development (UNCTAD) [23] uses work by Jeffrey Sachs to argue that, in addition to health and education, a government should be spending 2% of GDP on “public administration”, 3% of GDP on “expenses for police and defence”, and 5% on infrastructure such as rural roads which are “much harder to finance through the market”. UNCTAD and Sachs argue that this is also essential spending which must be made before debt service is paid.

Using this, we present two different calculations of how much tax revenue is available for debt:

Cafod: We take out ESS first and assume 70% of what is left is for other essential spending, so 30% of what is left is available for debt service.

UNCTAD: We first take out ESS and 10% of GDP, and then assume all of the rest is left for debt service. [24]

Using these methods, we find that many countries cannot even raise enough tax revenue to meet the DAC targets and thus not only can pay no debt service but also need substantial aid. These countries would need total debt cancellation:

Under the Cafod method, another 55 countries would need debt reduction. Under the UNCTAD method another 21 countries need debt reduction.

Even accepting that these are very rough estimates, this makes clear that the commitment to reaching the DAC targets will require debt cancellation far beyond anything suggested by the creditors so far.

It is often argued by the creditors/donors and the international financial institutions that it is also necessary to take aid flows into account. This can be done building on the previous section.

We use the UNCTAD method for essential expenditure and tax revenue available, and we add to that half of aid in the form of grants (excluding loans and technical cooperation) as a fair estimate of aid which is available for essential spending and debt service. [25]

Taking aid into account, we find that 28 countries need full debt cancellation and need an extra $12 billion in aid:

Country needing cancellation	Extra aid $ mn
Bangladesh	197
Burkina Faso	102
Burundi	59
Cambodia	56
Chad	30
Congo DR (Kin)	1170
Ethiopia	1293
Gambia	11
Haiti	53
India	4597
Kenya	368
Lesotho	12
Madagascar	269
Malawi	132
Mali	142
Mozambique	161
Nepal	234
Niger	164
Nigeria	1021
Rwanda	2
Sierra Leone	87
Somalia	160
Sudan	366
Tanzania	461
Togo	25
Uganda	124
Viet Nam	153
Yemen	339

The proposal of the British Chancellor Gordon Brown to increase aid levels by $12 bn per year would only meet this need if all the debt was cancelled for 28 countries.

What this table makes clear is that to meet the DAC targets we need at least $12 bn per year more in aid for 28 poor indebted countries in addition to writing off debt service of $75 bn per year. Of course, there can be endless discussions about the validity of our assumptions and estimates, but what cannot be contested is that they show a need for debt cancellation going far beyond anything offered by creditors so far.

And it is not a case of either/or. If we are to fulfil our commitment to the DAC targets, we need increased aid and a much deeper level of debt cancellation.

Perhaps the most important aspect of the 1953 London agreement on German debt was the recognition that Germany could only repay its debts if it had a trade surplus, which would only occur if its creditors bought its products. This may seem obvious, but it is a point which is ignored in all of the present debt discussions. And Table 6 shows why. Only 3 or the 93 countries have a trade surplus large enough to meet the debt service due. Seven others can meet part of their debt service. But the remaining 83 countries ran a trade deficit in 1996 or 1995, and export commodity prices have fallen since them.

In all, out of $239 bn per year due, only $16 bn can be paid out of trade surpluses. The rest must be paid out of aid or from loans and investment – forms of funding which have a future cost. [26]

Perhaps the time has come to insert a bisque clause into any debt settlement which ensures that debt service can only be paid out of trade surpluses, to force the creditors to also buy poor country exports.

Because of its recent colonial history and because of the slave trade, the industrialised countries owe a substantial historic debt to Africa which is probably much larger than Africa's current debts to the north. This leads to the demand by the Jubilee 2000 Afrika campaign for the total cancellation of African debt.

Table 6 shows that the African countries in our list are currently paying $24 billion.

Table 5 shows that all but 6 of the African countries are on the list of countries needed some debt cancellation on development grounds, even if aid is taken into account. Indeed, 22 of the 28 countries needing total debt cancellation plus more aid are in Africa. Only 4 African countries have trade surpluses.

This suggests that after other debt cancellation mechanisms are used, it would cost relatively little to cancel the rest of Africa's debt.

All of our estimates have shown the necessity of massive debt cancellation, on a level demanded by the debtors but so far not proposed by the creditors. Is this “realistic”? Arguably, this is the wrong question – it is too late; we are already committed.

First, we are all committed to the Universal Declaration of Human Rights which guarantees everyone the right to an adequate standard of living and to health and education. Therefore, we must assume it to be a violation of human rights that children in Africa do not go to school so that debts can be repaid. Thus major reductions in debt service payments are demanded on human rights grounds.

Second, the creditor countries are committed to the DAC targets and to the promise of halving the number in absolute poverty by 2015. Commitments to DAC targets and human rights are absolute; neither was restricted by “if the rich countries feel they can afford it”.

Third, the international community has already accepted that money at this level needs to be mobilised. The 20:20 Initiative [27], endorsed by the World Summit for Social Development in Copenhagen in 1995, is explicitly based on the need to mobilise an additional $40-80 billion per year. [28]

UNDP, in its Human Development Report 1998 [29], stresses that on a global scale, the amount of money needed is small. In an article which says that “debt is one of the biggest obstacles to human development,” the report points out that when the Asian crisis loomed, “$100 billion [was] raised in a few months”. All that is needed is “political leadership”.

UNDP goes on to give some comparative figures for world consumption. World military spending is $800 bn per year. World spending on narcotics is $400bn a year. OECD countries subsidise agriculture with $350 bn a year (which, in passing, reduces poor country exports and makes it harder for them to pay debts). Europe alone spends $155 bn a year on cigarette and alcohol.

Finally, there is the possibility of a Tobin tax on foreign exchange transfers, which now probably exceed $500 trillion per year. A tax of only 0.04% (that is $4 per $10,000) would raise enough to pay the debt service of all 93 countries on this list.

In conclusion, a lot of money is needed. But we are already committed to it. And the money can be found, if we have the will to fulfil promises already made.

Jubilee 2000 Coalition
PO Box 100
London SW1 7RT
tel: +44 171 401 99 99 ext 22
fax: +44 171 401 39 99
email: jhanlon@jubilee2000uk.org

Footnotes

[1] For the 41 HIPC countries, TDS/NPV = 9.3%. But for the 3 HIPC countries likely to use the fiscal criteria, TDS/NPV = 13%. Thus we take 14% and use it for all countries.

[2] Proposed at the HIPC Consultation Meeting at the Commonwealth Secretariat, London, 5 March 1999.

[3] Joseph Hanlon, “We've been here before”, Jubilee 2000, London, 1998. It should be noted that there were a number of other settlements which simply involved massive new lending to repay old loans; we assume here that that solution has already been shown to be a failure.

[4] The official definition under HIPC is that debt is “sustainable” if “a country is able in all likelihood to meet its current and future obligations in full without resorting to rescheduling in the future or the accumulation of arrears.”

[5] Edward Dommen, “Lightening the debt burden”, UNCTAD Review, Vol 1 No 1 p79, 1989.

[6] Chandra Hardy, Rescheduling developing country debts, Overseas Development Council, 1982, p61.

[7] Jon Kofas, Foreign debt and underdevelopment, University Press of America, 1996, p95.

[9] Nigeria was deleted last year and Malawi added by the IMF this year; we use this revised list.

[10] TDS/XGS < 20% except for the fiscal criteria for one country named by the IMF as likely to benefit – Congo (Brazzaville) – and two others already judged by the fiscal criteria – Cote d'Ivoire and Guyana. Table 1 shows the rigidity of the fiscal criteria; we calculate at least 7 other countries could benefit from the use of the fiscal criteria.

[11] We should not be surprised by this. In a press conference on 17 March 1998, World Bank Vice President for Africa Jean-Louis Sarbib stressed that HIPC is “just making sure these countries ... pay their debt.”

[13] Adopted by the 34th High Level Meeting of the Development Assistance Committee, held on 6-7 May 1996, and published in the 1996 Development Cooperation Report, OECD, Paris, 1997.

[14] H Northover, K Joyner & D Woodward, “A human development approach to debt relief for the world's poor”, Cafod, London, 1998.

[16] The full target is: “A reduction by two-thirds in the morality rates for infants and children under age 5 and a reduction by three-fourths in maternal mortality, all by 2015”

[18] Note that there are strong reservations about the use of this figure, which is said by some experts to underestimate the level of poverty. See, for example, Michel Chossudovsky, Journal of International Affairs, fall 1998 (vol 52 no 1).

[19] We assume that if a development-based approach were used to decide on what level of debt service is sustainable, that this would be negotiated on a country-by-country basis using local information. The use of rough estimates to give a general picture cannot be seen as projecting possible debt service for any individual country.

[21] This is a relatively complex calculation. The $1 per day poverty threshold is conventionally taken at purchasing power parity (PPP) of 1985 US dollars. Accounting for inflation, this is equivalent to $526 current PPP dollars. For those people earning more than $1/day, we just discount this portion of their earnings. The poorest people who earn below the tax threshold have a share of non-taxable income which is less than the equivalent of $1 per day. By definition, the “poverty gap” is the amount their earnings fall below $1 a day (in effect, the amount the poorest do not earn and which thus cannot be subtracted as non-taxable from GDP). The share of GDP excluded from taxation becomes:

[22] Clearly not just income tax. This will include all revenue sources including VAT and sales taxes, duties on alcohol and imports, etc.

[25] We also assume that all countries on our list have already been given HIPC level (TDS/XGS=20%) debt relief if they need it (this already affects 39 countries), and define DB as the debt service then due. Thus:

[26] David Woodward, “Drowning by numbers”, Bretton Woods Project, London, 1998. This warns of the dangers involved in becoming dependent on flows of new investment to meet old debts, and suggests that this could be the next debt trap.

[27] The 20/20 Initiative is a donor-poor country compact, under which governments agree to allocate 20% of their budgets to basic social services and donors agree to increase their aid to a level such that 20% is given to basic social services, and that this should be sufficient to provide universal coverage. The original document, however, stresses that the 20% cannot be applied rigidly, and should reflect national development priorities. It also stresses that “20/20 is not proposed as new form of conditionality, nor should it become one.” Finally, it stresses the importance of debt relief as part of the package.

[28] These figures were first published in the joint UNDP, UNESCO, UNFPA, UNICEF and WHO 1994 paper “The 20/20 Initiative”, and draw on World Bank, UNDP and UNICEF figures. The paper estimates the needs at $25 bn per year for health, $6 bn per year for education, and $9 bn per year for water and sanitation. This is expanded in the UNDP's Human Development Report 1997 (p 112) and 1998 (p 37) to include a $40 bn income transfer to the poorest.

Due	$ 15.7 bn
HIPC sustainable [12]	$ 9.9 bn
Paid	$ 8.7 bn

	TDS/XGS	$ bn	Number of countries which gain compared to debt service paid
Debt service due		239
Debt service paid		216
Debt service possible:
HIPC sustainable	20%	216	33
HIPC improved	10%	108	61
Indonesia	6%	65	73
Germany	3.5%	38	81

	$ bn	Number of countries which benefit – compared to debt service due
Debt service due	239
Reducing to TDS/XGS=20% for countries with higher TDS	163	50
TDS/XGS = 10%	108	75
TDS/XGS = 6%	65	85
TDS/XGS = 3.5%	38	91

	TDS/GDP	$ bn	Number of countries which save compared to debt paid
Debt service due		239
Debt service paid		216
Debt service possible:
Eurodad	2.8%	127	62
Indonesia & J2	0.7%	32	87
Germany & J2	0.4%	18	90

	Number of countries which cannot meet essential expenditure from taxes	Shortfall
Cafod	16	$ 6 bn
UNCTAD	48	$33 bn

	Number of countries	Cost $ bn/yr
Can pay debt service	27
Need reduction but not to HIPC levels	4	7
Reduction to HIPC	39	32
Further reduction	62	36
TOTAL DEBT SERVICE CUTS NEEDED		75