Joseph Hanlon (j.hanlon@open.ac.uk)
Development Policy and Practice, Technology Faculty, Open University
Milton Keynes MK7 6AA, England

Publication: This is a preprint of a paper published in the Journal of International Development 12, 877-901 (2000). ( http://www.interscience.wiley.com )

Abstract: Developing country debt now exceeds $2.4 trillion and has become a major international political and economic issue for debtor governments, creditor governments, the IMF and World Bank, and campaigning organisations such as Jubilee 2000. For the poorest countries, debt has become unpayable and debt service an obstacle to development. This paper argues that debt crises and substantial debt cancellation are part of the normal economic cycle, and that an unusual aspect of this cycle has been the unwillingness to cancel debt. Recent historic precedents suggest that at least $1 trillion in debt would need to be cancelled. The paper then uses a 'rights-based approach to development' to estimate that more than $600 billion in debt must be cancelled to release sufficient funds to meet internationally agreed development targets and thus satisfy human rights. Finally, the paper argues that lenders must take responsibility for illegitimate, corrupt and odious loans. Debt cancellation is the norm, not the exception, and the only question is how much.

1. INTRODUCTION:

Developing country debt increased ten-fold in the 1970s. This was followed by a debt crisis in the early 1980s, in which many countries were unable to meet repayments. In 1982, Mexico and Brazil defaulted. Initially, attempts to resolve the crisis involved rescheduling of debts, normally allowing payments over a longer period. But this proved insufficient.

In the period 1984-1991, developing counties paid $209 billion to northern creditors (net – that is, they gave creditors $209 bn more in interest payments and principal repayments than they received in new loans). Yet total debt doubled again in the 1980s, not because of new lending but because old debts were being rolled over. Thus, although there was a net flow of funds from South to North, the total debt increased. (See Tables 1 and 2; World Bank 1991, 1994, 1997, 2000a) For the 38 countries defined by the World Bank as 'severely indebted low income', total debt rose from 5% of GNP in 1970 to 31% of GNP in 1980 to 139% of GNP in 1990, and were at the same level in 1999. Many of the poorest countries pay more in debt service than they spend on education or health.

There were two responses to this problem an increase in new money and a realisation that some debt needed to be cancelled. During the early 1980s new lending had fallen, but from 1987 to 1994 lending doubled and in the same period aid more than doubled.

The summit of the Group of Seven (G7) industrialised countries in Toronto in 1988 agreed a new form of rescheduling that in practice included a slight reduction in the debt of the poorest countries. G7 summits in 1991 and 1994 increased the amount of debt of the poorest countries which could be cancelled. Meanwhile, by the mid-1980s banks were beginning to accept that some debt would not be paid, and a secondary market in developing country debt was created; in some cases countries bought back their own debt at a discount. (World Bank 2000a vol 1 appendix 2 and 3)

Nevertheless, it was becoming clear that deeper debt cancellation would be needed. At the instigation of the G7 leaders, the World Bank and International Monetary Fund developed the Heavily Indebted Poor Countries Initiative (HIPC) to cancel some debt for up to 41 countries, and this cancellation was extended by the G7 at its 1999 summit in Köln. This affects only a small portion of poor country debt, however. (See Table 3) Meanwhile, the international Jubilee 2000 campaign “to cancel the unpayable debts of the poorest countries” had become, according to World Bank spokesman Anthony Gaeta, 'one of the most effective global lobbying campaigns I have ever seen'. (PR Week 16 April 1999)

Although the millennium ended with a new recognition that some debt had to be cancelled, little debt had actually been written off and total poor country debt was still increasing.

This opens by arguing that debt crises and cancellation are both normal parts of economic cycles. It then looks at the recent history of the present debt crisis. Next, we look at debt cancellations of the 1930s-1960s and their implications, and make a first attempt to see how much debt would need to be cancelled if existing pledges for poverty reduction are to be met. Finally, the paper looks briefly at a set of related moral and political issues, including lender liability, odious debt, and the use of funds released by debt cancellation.

2. CYCLES, LOAN PUSHING AND DEFAULT

Debt and default are not new. The present problem of countries not able to repay debts is very similar to the 1930s - and the 1880s and 1840s.

In the 19th century, the United States was a developing country, and British banks were lending to the new US. 'London saw it as a very unreliable developing country, with a black record of embezzlement, fraudulent prospectuses and default,' wrote Sampson (1981 pp 29-31 and 54-56). But the bankers made loans. In 1842, 11 states including Maryland, Pennsylvania, Mississippi and Louisiana defaulted.

During and after the First World War, the US was relatively wealthy; Europe was poorer and borrowed from the US. In the 1930s Europe defaulted. The United States Treasury (1997) reported that Britain still owes $14.5 bn to the United States for loans during the First World War; only three countries in sub-Saharan Africa have larger debts. France owes $11.8 bn, and the rest of Europe owes another $6.2 bn to the United States. None are repaying this money. [1]

Latin America was a major borrower in the 1920s, mainly paying old loans by taking new ones. But with the1929 crash, new loans stopped and exports fell dramatically as the industrialised countries became protectionist; by 1934, Latin America had defaulted on 85% of its dollar bonds (Drake 1989 p 50). Default was a sensible response to falling commodity prices and unpayable debts; a World Bank study showed that those Latin American countries 'which opted for default recovered more successfully from the ravages of the Great Depression' than those which struggled to pay their debts (Eichengreen and Portes 1989a).

Debt crises are linked to the business cycle. Economist Charles Kindleberger (1978 pp 6-23) wrote a book whose title, Manias, panics and crashes, captures his view of the cycles. He points to the tulip mania of 1634, the South Sea bubble of 1720, the cotton and railway booms of the 1830s, and so on. Kindleberger says that in each cycle, there is a period of growth involving a rise in profits and a rapid expansion of bank credit. Eventually money growth outstrips production and money goes into speculation. This is often linked to fraud and swindles, as speculators look for ever more profitable investments. This is the period of bubbles or what Kindleberger calls 'manias', and usually involves international lending. Eventually the bubble bursts, prices fall, and investors try to sell or to collect on their loans. This is the period of 'panic' as investors all rush for the exit. The panic feeds on itself, leading to the 'crash'.

Economists point to both long and short cycles, but it may be useful to look at four long cycles (each of which had short cycles within them): [2]

1) growth 1780-1820, mania 1820s, crisis 1830s & 1840s.

2) growth 1850s, mania 1860s, crisis 1870s & 1880s.

3) growth 1893-1913, mania 1920s, crisis 1930s.

4) growth 1948-1967, mania 1967-1979, crisis since.

Both the 1870s and the 1930s saw major international depressions.

Each mania involved an international lending boom which overlapped with the end of growth in domestic economies of the then developed countries, as lenders were forced to look abroad for higher profits. After each cycle, there have been retrospective complaints of reckless lending and of 'loan pushing' – banks and lending agencies encouraging foreign governments to take loans they do not need, and encouraging borrowers to live beyond their means. Toward the end of the mania, borrowers are encouraged to take new loans simply to repay old ones. With the panic, lending suddenly stops, borrowers cannot repay, and they default.

Francis White, the US Assistant Secretary of State for Latin American Affairs in the early 1930s, commented that 'in the carnival days from 1922 to 1929, when money was easy, many American bankers forsook the dignified, aloof attitude traditional of bankers and became, in reality, high pressure salesmen of money, carrying on a cut-throat competition against their fellow bankers, and once they obtained the business, endeavoured to urge larger loans on the borrowing countries' (Drake 1989, p 43).

3. THE 1980s DEBT CRISIS

The 1970s can be seen as the mania phase of the current cycle [3]. It seems likely that the 1979 oil price rise, followed by the 1982 Mexican default, burst the bubble and caused the panic; so far, there has been no depression in the industrialised countries.

'The more recent debt crisis was very much a rerun of that in the 1920s,' noted Derek Aldcroft (1997, p 121). 'The present wave resembled earlier experiences more closely than most participants realised. The onset of the crisis in 1982 again looked hauntingly familiar to those who have read the history of debt crises. It is the official response to the crisis since 1982 that stands in stark contrast to the past,' wrote Peter Lindert (1989 p 227).

As we show below, the mania phase was very similar, but in the response to the panic there are two differences between the 1930s and 1980s-1990s that seem important:

• In the 1930s it was the industrialised countries which suffered, while Latin America was hardly harmed – in part because most countries simply defaulted. By contrast, in the 1980s amd 1990s, the industrialised and creditor countries have not suffered very much, whereas the crisis in the south has been very serious.

• Governments and creditors have worked in a much more coordinated way both to prevent default by individual countries, and to bring collective pressure on debtors.

The 1970s definitely looked like the 1920s, especially in terms of surplus capital and banks 'pushing' loans on developing countries. 'The main explanation' for the sharp rise in lending to developing countries was the 'failure of demand for loans from borrowers in developed countries to keep pace with the expansion of credit availability', noted US Federal Reserve Governor Andrew Brimmer in 1973. This caused 'the Eurocurrency banks (especially in London) ... to push loans to the developing countries with considerable vigor.' Indeed, it has been argued that the Eurodollar market ballooned precisely because there were no regulations and thus riskier loans were possible (Darity and Horn 1988, p 8, 9, 15). Kindleberger (1996 p183) notes that in the 1970s, 'Brazil, Mexico, South Korea, Zaire, Peru and others found themselves courted by the Eurocurrency banks, just as in the “good old days” of 1927.'

Brimmer cited a particular form of loan pushing, which involves a drastic softening of terms, and this actually worsened after his warning. Darity and Horn (1988 pp 11, 15) note that between 1975 and 1979, real interest rates fell and repayment periods increased, and banks became 'aggressive credit promoters ... literally covering the globe to find new borrowers.' According to the World Bank (1990 p 15), real interest rates [4] fell to a negative 3% in 1975 and stayed negative until 1978. In other words, developing countries were told they could borrow for projects and repay less than they borrowed.

a) Changes after 1979

There was a dramatic change in the global economy at the end of the 1970s. It had the effect of transferring resources from the poor countries of the south to the industrialised countries of the north, through higher interest rates and lower commodity prices, in a way that may have helped to avoid a 1930s style depression.

Real interest rates jumped from -1% in 1978 to +9% in 1982 (World Bank 1990 p 15). In the 1980s, countries simply borrowed to repay old debts, rather than for new projects. Arrears reached $111 billion by 1990 and total debt had doubled in the decade. Nevertheless, for the entire nine-year period 1983-1991, for all developing countries, debt service (interest plus principal repayments) was greater than new loans; $209 billion was transferred from debtor countries to creditor countries (See Table 1; World Bank 1991, 1994, 1997). At the same time, commodity prices fell sharply; between 1980 and 1986, terms of trade for sub-Saharan Africa fell by more than 30%. So developing countries were earning less but paying more, and getting deeper in debt.

Higher interest rates and lower commodity prices caused a growing debt crisis, marked by the Mexican default in 1982; between 1983 and 1987, Mexico rescheduled $125 billion in commercial loans. Virtually every developing country had to reschedule its debts.

Eventually aid and new lending began to increase, and in 1991, for the first time in a seven years, aid was larger than the gap between debt service payments and new loans. In other words, for the first time since 1984, there was a net transfer of money from the north to the south.

Why did the developing countries agree to send so much money to the industrialised countries, and not default as they had in the 1930s? 'In the 1980s, creditor-country governments, motivated by the desire to protect their banking systems, have exerted their greatest pressure on the debtor countries, urging full repayment and macroeconomic adjustment,' note Eichengreen and Lindert (1989 p 7). They have done so largely though their control of the International Monetary Fund, which was only established in 1944 and is, perhaps, the single most important difference between the 1980s and 1930s.

Huw Evans (1999), UK Executive Director of the IMF and World Bank 1994-97, wrote recently that the main area of concern in the 1980s was 'the dangers to many of the world's largest banks, and the banking system. Exposure to the debtors by many banks, especially in the United States, but also to a lesser extent in the UK and the rest of Europe, was several times total bank capital. The debt strategy of the 1980s bought time for the banks to rebuild their capital.' In 1989 the US Treasury Secretary Nicholas Brady created the 'Brady bonds' for Latin American countries which were mainly a way of drawing in the US government and IMF and World Bank to bail out US banks. Thus there was a slow but steady run down of debt to banks and its replacement with official debt to international agencies and governments, thus ensuring that the banks were not penalised for unwise lending.

b) Official lenders

An important difference from the 1920s is that today only half of debt is actually private debt – usually bank loans and usually guaranteed by government. Private debt constituted 62% of long term developing country debt in 1980, falling to 49% by 1990 and remaining at that level through the 1990s.

There are three other kinds of debt:

• bilateral non-concessional debt (usually export credits guaranteed by an export credit agency, where they buyer has failed to pay and the debt has been nationalised by the lender country),

• bilateral concessional debt (aid was often given as soft loans in the 1970s, and Germany and Japan continued to do this into the 1990s), and

• multilateral debt (from the IMF, the World Bank, and regional development banks)

The IMF and World Bank, which soon took charge of managing poor country debt repayments, insisted that they be paid preferentially. Bilateral creditors came next, while commercial banks tended to be last in the queue. The Paris Club was set up in 1956 as an informal group of creditor governments, with a permanent secretariat in the French Treasury, and after 1980 debtor governments increasingly were forced to go to the Paris Club to negotiate extensions on their debts.

By the late 1980s, official creditors – governments and the international development banks – realised that simply delaying or reducing payments on commercial loans was not adequate, and that something would have to be done about official loans. The IMF, for example, in 1987 created a new Enhanced Structural Adjustment Facility to give softer loans to poorer countries. But former IMF Executive Director Evans (1999 p 271) says that 'in reality, the initiative was ... aimed at ensuring that existing IMF loans to these countries – arrears on which were beginning to mount – could be refinanced. ... This initiative confirmed the need for action to reduce debt, but postponed until the mid-1990s the pressure on the IMF to join in effective and comprehensive debt relief.'

Leaders of the Group of Seven (G7) industrialised countries dealt with this issue twice, in Toronto in 1988 and London in 1991, and in each case agreed that the Paris Club of official bilateral creditors would reschedule more debt at concessional interest rates – an admission that full repayment could not be expected.

By the late 1960s, the Paris Club had made all reschedulings conditional on having an agreement with the IMF (Eichengreen and Portes, 1995, pp 23-24). As Evans (1999 p 276) notes, the IMF only lent 'modest' amounts of money, but all 'other official flows – aid and export credit and Paris Club rescheduling – were dependent on an IMF programme being in place, which gave these countries big incentives to sign up to IMF programmes.'

IMF programmes always involved cuts in domestic spending and an opening to the global market. They have been widely criticised in the South for being too restrictive and increasing poverty, but they did ensure that poor countries continued to service their debts, at least partly. Thus there was no 1930s style default. Throughout the 1980s and 1990s, the main way of treating the debt crisis has been to lend more money. But simply rolling over loans was creating an impossible problem because of compound interest – interest payments were deferred, but the borrower was charged interest on the unpaid interest.

Debt continued to increase, as countries continued to borrow more money in an effort to repay old debts. This resulted in a huge churning of money. In 1999, for example, all developing countries borrowed $246 billion, but of this $214 billion went back immediately to repay the principal of old loans. Developing countries had to find $135 billion to pay interest on old loan, meaning there was a net transfer to the creditor countries of $103 billion. In the same year, sub-Saharan Africa borrowed $11 billion but repaid $15 billion ($10 billion in principal repayments and $5 billion in interest payments). Thus, nearly half of the $10 billion in aid grants to sub-Saharan that year stayed in the north to help repay old debts (World Bank 2000a, vol 1, pp 238, 250).

4. HISTORIC CANCELLATIONS

Although the emphasis remained on rescheduling and giving new loans to repay old ones, the 1990s saw growing discussion about cancellation. Debt cancellation is not new and is often political.

There was extensive debate on foreign debt in the 1920s, including a discussion of the unpaid 1840s bonds of the US southern states and a suggestion, rejected by the US, that they be swapped for World War I bonds issued by the US. Discussion became more intense after the 1929 crash, and in 1930 the US announced the formal repudiation of the 1840s bonds. This was finally accepted by British bondholders in the 1950s, more than a century after the bonds had been issued. (Ugarteche 2000)

In the early 1940s, creditors were still negotiating with Latin America over the 1930s defaults. In 1945 Peru reached agreement with its US bondholders that unpaid interest between 1931 and 1945 would be cancelled, and the remainder of the debt would be repaid by a maximum of 3.5% of the national budget. (Ugarteche 1999) By 1945 creditors and bond holders had accepted settlements which took the stock of Latin American debt down to one-third of its former value (Díaz Alejandro 1983 p 27). During World War II the United States was anxious to gain support from Mexico, and it intervened in the negotiations with the bond-holders. In 1942, it was agreed that 90% of Mexican debt would be written off (Aggarwal 1989 p 148).

In 1952 in London a West German team headed by Hermann-Josef Abs negotiated with representatives of 24 creditor governments for nearly a year to try to settle pre- and post-war debts. The context was highly political. The Berlin blockade of 1948-49 had been followed by the creation in 1949 of the Federal Republic of Germany from the territory occupied by the Western allies. The Cold War had begun in earnest and the West was anxious to give West Germany an economic boost in its competition with East Germany. Also, the allies remembered that one cause of World War II was that after World War I Germany had been saddled with heavy debt and reparations payments. Thus the London negotiators for the first time based a debt settlement not on how much the country could be forced to pay, but on how much it could afford to pay if it was allowed to spend for reconstruction and development. Under the London agreement signed 27 February 1953 Germany acknowledged its liability for all pre-war debt, and the allies wrote off most of it. Payments scheduled for 1953 were expected to be 3.5% of exports and 0.4% of GDP, and never rose above this level. The agreement included the unprecedented clause that repayments would only be made out of German trade surpluses – if the allies wanted debt repaid, they had to buy German goods (Hanlon 1998a, 1999).

The next two major debt settlements were also political. In August 1965 Indonesian President Sukarno withdrew from the World Bank and IMF. The army overthrew Sukarno, put General Suharto in power and massacred up to 750,000 alleged communists. Suharto rejoined the IMF and World Bank. In December 1996 the Paris Club of Western creditors met and rewarded Suharto with a four year moratorium on all debt service payments. Hermann-Josef Abs was sent to negotiate a new debt service agreement. The final deal said that payments would resume in 1970, but were in effect limited to less that 6% of export earnings and 0.7% of GDP (Kuhn Loeb Lehman Brothers 1979).

In 1991 the United States wrote off $10 billion of Egypt's debt in exchange for Egyptian cooperation in the Gulf War. This was by far the largest single debt cancellation of the 1990s, and accounted for one-third of all developing country debt cancellation in the six years 1990-95.

5. CANCELLATION AND HIPC

In the 1980s, creditors pretended that poor countries would eventually be able to repay because they refused to 'admit to themselves and the world that the debts were no longer worth their face value,' according to Evans (1999 p 268). Thus they continued to roll over and reschedule loans. The first change came when private lenders, often under pressure from their own country's central banks, began to admit that the chances of collecting some of these loans was small. On their own books they began to write down the value while not actually cancelling that part of the debt; the country was still expected to pay the full amount. That, in turn, led in 1983 to the setting up of a secondary market in Third World debt. Like junk bonds and other risky investments, banks and traders bought and sold Third World debt – at prices that took into account the high probability that the debt would never be repaid (Garrett and Travis 1999).

The World Bank and the governments of the industrialised countries intervened to bail out the international banks by taking over some of the private debt, usually at a discount, and sometimes even giving countries loans to buy their own debt on the secondary market. The World Bank estimates that between 1989 and the end of 1997, $56 billion in debt had been cancelled – about 10% of developing country commercial loans. Nicaragua and Ethiopia paid as little as 8% of the face value to buy back their own debt, and this ranged up to 91% for Venezuela (world Bank 1998a vol 2 pp 83-85). But that cancellation must be compared with an increase in total debt over the same period of $972 billion.

The G7 meetings in Naples in 1995 and Lyon in 1997 agreed, for the first time, to the actual (rather than defacto) cancellation of a small amount of Paris Club bilateral debt. And in 1996, with the Heavily Indebted Poor Countries (HIPC) Initiative, the World Bank and IMF accepted for the first time that they, too, would have to cancel some debt. The Bank and Fund identified 41 HIPCs, and they used a new concept to define when a debt was 'sustainable'. This was not based on development considerations, as were used in the 1953 German agreement, but rather on how much money could be squeezed out of a poor country. A debt was to be considered sustainable if 'a country is able in all likelihood to meet its current and future external obligations in full without resorting to rescheduling in the future or accumulation of arrears' (World Bank 1998a, vol 2, p 55). All creditors agreed that for these 41 countries, they would cancel debt proportionately to reach the 'sustainable level'.

The World Bank and IMF then decided to assess 'sustainability' in terms of export earnings, rather than the more logical measures of GDP or government budget, and it chose to call debt 'sustainable' at a level that only two year before it had said was 'unsustainable'. [5] A HIPC 'sustainability' criterion defined in 1996 was that annual debt service should be between 20% and 25% of export earnings. The HIPC process proved slow and confusing, and for most countries simply cancelled the debt that was not being serviced, so poor countries which qualified for HIPC gained little. When they met in Köln (Cologne) in 1999, the G7 leaders agreed to lower the 'sustainability' level to, effectively, an annual debt service of 15% of export earnings. [6]

After the Köln G7 meeting, it was estimated that about $100 billion [7] in debt of 33 countries would be cancelled by a combination of the Lyon Paris Club agreed terms and the new HIPC terms. The HIPC countries are currently supposed to pay $13 billion per year in debt service, but actually pay about $9 billion; the $100 billion in debt cancellation would cut debt service by about $6 billion per year, but since $4 billion is already not being paid, this would cut debt service of these poor countries from $9 billion to $7 billion per year. In late 1999, the World Bank estimated that eight countries which it defines as 'heavily indebted poor countries' would not have any debt cancelled; other poor and indebted countries, such as Bangladesh and Morocco, remained ineligible even for consideration.

The HIPC debt cancellation process is based entirely on reducing debt to a level to which it can be repaid. It asks no questions about development, poverty, or post-war reconstruction.

6. HOW MUCH DEBT SHOULD BE CANCELLED?

Historic debt settlements tended to be on an individual basis. But it is possible to look back at them and use the HIPC methodology to set alternative cancellation criteria.

HIPC is the first time that export earnings have been formally used as a basis for defining sustainable debt, but it is possible to calculate comparisons. In 1952 the victorious allies originally demanded that Germany pay the equivalent of 10% of export earnings in debt service, and Hermann-Josef Abs said this was intolerable (Hanlon 1998a). The allies agreed, and Germany only had to pay 3.5% of export earnings. Indonesia was only asked to pay 6% of export earnings. This is compared to 15% of export earnings under the improved HIPC – two-and-a-half times the Indonesia level and more than four times the German level.

Although exports can be seen as one way of measuring a country's income and ability to pay, the level of exports also depends on the size and location of the county. GDP is often considered a better measure. Germany was required to pay 0.4% of GDP and Indonesia 0.7%. In terms of GDP, HIPC country exports are 28.5% of GDP, so debt service of 15% of exports corresponds to debt service of approximately 4.3% of GDP. This is 10 times the level seen as acceptable for Germany in 1953 – 0.4% – and 6 times the level seen as acceptable for Indonesia in 1970 – 0.7%.

Government revenue is an alternative measure, and was the one selected by Peru in 1945. For most countries, government revenue is similar to exports and runs at 15-30% of GDP.

Finally, and most importantly, Germany was expected to pay only when it had a trade surplus -- the creditors had to buy German exports if they expected to be repaid.

Using these criteria, we can estimate how much a country can pay in debt service and compare that to the amount that it owes. We look at 93 of the poorest and most indebted countries. [8] In 1998, these 93 countries had a debt of $1700 billion, out of a total developing country debt of $2465 billion (See Table 3). In 1998, they owed $226 billion per year in debt service (interest and principal repayments) and actually paid $209 billion of that.

Table 4 sets out the results of applying historic criteria. The implication is that between 75% and 93% of debt would need to be cancelled -- which is approximately the proportion of debt which was cancelled after the previous economic cycle. By contrast, HIPC will only cancel 4% of debt.

Looking at export criteria, only 3 of our 93 countries could pay all their debt if the German criterion applied, and only 5 if the Indonesia criterion was used. HIPC only applies to a small group of countries, but if its 15% were extended to all countries in the sample, only 34 would be able to pay all their debts.

7. HUMAN RIGHTS & DEVELOPMENT GOALS

All of the previous discussion has been about simple fiscal measures of possible debt service payments, based on fixed percentages of GDP, exports or revenue. In this section, we make a first attempt at defining a human rights and development approach to debt cancellation.

'A rights-based approach to development ... has been rising up the agenda',' notes Simon Maxwell (1999), director of the Overseas Development Institute. This approach 'sets the achievement of human rights as the objective of development. It uses thinking about human rights as the scaffolding of development policy. It invokes the international apparatus of human rights accountability in support of development action.'

Most importantly, the new 'rights-based approach' strives to put economic and social rights on the same plane as civil and political rights. And it draws directly on the Universal Declaration of Human Rights, adopted by the United Nations more than 50 years ago. Two of these rights are particularly important:

• 'Everyone has the right to a standard of living adequate for the health and well being of himself and his family, including food, clothing, housing and medical care and necessary social services.' (Article 25)

• 'Everyone has the right to education. Education shall be free, at least in the elementary and fundamental stages. Elementary education shall be compulsory.' (Article 26)

'The new rights agenda runs alongside an agenda derived from the international development targets, which focuses on poverty and human development,' Maxwell notes.

a) DAC targets

The most widely cited development targets are the 'DAC targets', adopted in 1996 by the 21 donor countries which are members of the OECD Development Assistance Committee (DAC). [9] These donors committed themselves to a set of six of what they called 'ambitious but realisable goals':

• a reduction by one-half of the proportion of people living in extreme poverty by 2015,

• universal primary education by 2015,

• eliminating gender disparity in education by 2005;

• reduction by two-thirds in infant and child (under 5) mortality and three-fourths in maternal mortality, all by 2015,

• access to reproductive health services for all individuals of appropriate ages by 2015, and

• a end to the loss of environmental resources by 2015.

The DAC targets can be seen as a way of reaching the goals of the Universal Declaration of Human Rights. [10] The international community is now committed to both of these.

Since the late 19th century, it has been recognised that human rights take precedence over debts, however those debts were acquired. Most countries now have bankruptcy laws. If a company goes bankrupt, no one would expect the children of the owners or employees of that company to drop out of school to go to work to pay the company's debts. Similarly, if a man runs up huge debts through drinking or gambling, no one any longer expects his children to leave school to work to pay the debts.

Yet, these principles are not applied at the level of nations within the global economy. For example, Mozambique is one of the poorest countries in the world, yet even after some debt cancellation it spends as much repaying debts as on health. President Joaquim Chissano has appealed for 100% debt cancellation but this has been rejected by the international community. [11] Tens of thousands of Mozambican children will be denied access to school and to health services because money goes to repay the World Bank and other creditors instead of being spent on poverty reduction.

Simon Maxwell (1999) notes that one aspect of a rights-based approach to development is that 'because rights are universal, the wider international community has at least a moral duty to support rights, including financially, in partnership with states; this moral obligation may extend to non-state actors, particularly international financial institutions, TNCs, and NGOs'.

Thus if a creditor takes money in debt service that is needed by a poor country to fund basic health services, it can be argued that the creditor is violating the human rights of the people of that country – as well as going against the international commitment to the DAC targets. Thus it might be argued that the IMF and World Bank are violating human rights.

Following Maxwell's lead, we use progress toward the DAC targets as a way of measuring satisfaction of human rights. To link this to debt, we estimate how much essential spending is needed to meet the DAC targets, and argue that this must be made before any debt service is paid.

b) Defining essential spending

In this section, we make a first estimate of how much needs to be spent to meet the DAC targets and how much money is available. Assuming that essential spending takes priority over debt service allows an estimate of how much debt cancellation is required to meet promises the industrialised countries have already made by agreeing the human rights declaration and DAC targets.

There are apparently no published estimates of the cost of meeting the DAC targets for 2015, nor are there any agreed figures as to what is 'essential spending', nor are there agreed estimates of how much tax revenue a country can be expected to raise. Indeed, the latter two issues have been at the heart of the debate of the past two decades.

Lack of data and lack of agreement on tax and spending policies means that we must make a series of often heroic assumptions, and can only make a very rough initial estimate of essential spending. [12] Below, we try to define essential expenditure, set out plausible estimates of income, and then see what is left for debt service. Our 93 countries have 4.2 billion people, of whom 1.2 billion have an income under the international absolute poverty line, $1 per person per day at purchasing power parity (PPP) at 1985 prices. [13] .

We first attempt to define Essential Social Spending (ESS). With no published estimates of the cost of meeting the DAC targets, we draw on three different methods:

Cafod extended. The British agency Cafod estimates that a low income country must spend at least $28 per capita per year on health and education. (Northover et al. 1998) This, in turn is based on World Bank (1993 p 66) estimates that low income countries must spend $12 per capita per year on 'public health and minimum essential clinical services', at 1990 prices. To this Cafod adds $12 for education and $4 for inflation and other expenditure, to reach $28. Noting that the World Bank also says that middle income countries must spend $22 per capita on the minimum health package, and we simply extend this to $52 for a Cafod-like package for middle income countries. This gives essential social spending of $136 billion per year.

'20-20': As part of the '20-20 Initiative', the United Nations system (UNDP et al 1998 p 20) estimated 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. We add this to present health and education spending [14] and consider it essential. This gives essential social spending of $309 billion per year.

HDR 96: UNDP's Human Development Report 1996 (p 113) 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 DAC targets is cutting infant and child mortality by two-thirds, 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. This gives essential social spending of $400 billion per year.

All three of these approaches have methodological problems. But as nothing else is available to allow an estimation of the costs of meeting the DAC targets, and there is no strong reason to choose between them, we simply average the results of these three approaches as at least a rough estimate of essential social spending (ESS).

Social spending is not the only essential spending. The United Nations Conference on Trade and Development (UNCTAD 1998 p 130) 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.

Therefore, we assume that ESS plus 10% of GDP is 'required spending' (RS), which must be made first before any debt service is paid.

c) Available revenue

The next step is to ask how much tax a country can be expected to raise. Here we follow an estimate first made by Cafod (Northover et al. 1998). All countries have a tax threshold – a level of income below which no tax is paid. Cafod takes the tax threshold at the absolute poverty line, $1 per day, and assumes all income below that is not taxed. The $1 per day poverty threshold is conventionally taken at purchasing power parity (PPP) of 1985 US 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. [15] By definition, the 'poverty gap' is the amount their earnings fall below $1 a day – in effect, the 'average income shortfall of the poor' (Gordon and Spicker, 1999, p 103) – and below-poverty-line incomes are taken as $1 per day reduced by this amount, which is then subtracted as non-taxable from GDP. [16]

We assume that above that level, tax can be 25% of income. [17] Clearly, however, this is not just income tax and will include all revenue sources including VAT and sales taxes, duties on alcohol and imports, corporation taxes, etc.

For the poorest countries, aid is a major source of revenue. We assume that technical cooperation, food aid, loans, etc are not available for debt servicing. Thus we only consider what the World Bank in its annual Global Development Finance defines as 'grants'. Further, we assume that half of that is allocated by the donors and half is general budget support which can be spent on debt servicing. [18]

Thus we assume that possible government income is possible tax revenue (defined as 25% of income above the poverty line) plus half of grant aid.

d) Possible debt service

Although much of the debt cancellation literature talks about the volume of debt, what is relevant to a country is the actual debt service (interest plus principal repayments) that it pays. Debt service due [19] is used rather than debt service actually paid, because many of the poorest countries pay only half of what they owe, but the debt remains. The 41 countries defined as Heavily Indebted Poor Countries (HIPC) by the World Bank and International Monetary Fund paid only 71% of their debt service due in 1998; they owed $13 billion but in fact paid only $9 billion of this.

Essential social spending, required spending, and possible income [20] are calculated as set out above. The results are given in tables 5-7 which split the 93 countries into three groups.

Table 5 shows countries where the maximum feasible income from taxes and aid is less than the required spending. These 34 countries need an extra $15 billion per year in aid just to meet the DAC targets. In addition, there is no hope of their meeting any of the demands for $25 billion per year in debt service which is owed. They have a total debt of $291 billion, which, under our assumptions, would need to be cancelled.

Table 6 shows 37 countries which have enough potential tax revenue to meet required spending, but not sufficient to pay their debt service. They were expected to pay $137 billion in debt service in 1996 but could only afford to pay $91 billion; $46 billion per year is unpayable. This corresponds to nearly $350 billion in debt which would need to be cancelled. [21] This group includes countries like Argentina and Haiti which could argue that some of the debts are odious (see section 8, below); the estimation that they have enough money to pay these debts does not remove the moral and political argument that they should not.

Table 7 shows that 22 of these countries can collect enough money to meet domestic spending needs and pay their debts.

These 93 countries owed $226 billion in debt service in 1998, and actually paid $209 billion of it. Based on the assumptions set out above, we estimate that they could only 'afford' to pay $156 billion; the extra $70 billion was money which should have been spent of health, education and development in order to satisfy basic human rights. And, as we argue above, debt relief will not work on its own; in addition, as table 5 shows, we estimate that $15 billion per year more aid is needed. These estimates suggest that if creditor countries are serious about their commitments to human rights and the DAC targets, they will need to cancel more than $600 billion in debt -- compared to the $100 billion agreed by the G7 at Köln.

We estimate that these countries need $85 billion in new aid and reduced debt service, based on consideration of essential social spending. The Economic Commission for Africa (1999 p 36) estimated that sub-Saharan Africa needed an extra 13.9% of GDP in additional investment resources to reach the growth rate needed to halve the number of people living in absolute poverty by the year 2015. This is $44 billion per year just for Africa, which suggests that our $85 billion for all 93 countries may be an under-estimate.

8. POLITICAL, MORAL AND PRACTICAL ISSUES

So far we have estimated debt cancellation based on ability to pay. But this is based on two assumptions: that the debt should be paid, and that the money will be spent on development to meet the DAC targets. It is beyond the scope of this paper to discuss these assumptions, but a few comments need to be made here.

Even the 'rights-based approach' only considers the amount of debt, and not its history. In this section we cite a 'normative' argument – not that the debt cannot be paid, but that some debt should not be paid.

Zaire's ruler, Mobutu Sese Seko, was one of the world's most corrupt leaders, and it was for his government that the world 'kleptocracy' was first coined. In 1978 the IMF appointed its own man, Irwin Blumenthal, to a key post in the central bank. He resigned in less than a year saying that 'the corruptive system in Zaire with all its wicked and ugly manifestations' was so serious that there is 'no (repeat: no) prospect for Zaire's creditors to get their money back' (Lissakers 1992 p 166). [22] Shortly afterwards, the IMF gave Zaire the largest loan it had ever given to an African country. When Blumenthal wrote his report, Zaire's debt was $5 billion. When Mobutu was overthrown and died in 1998, the debt was over $13 bn, due to huge loans from the IMF, the World Bank and governments.

Philippines dictator Ferdinand Marcos and his wife Imelda are said to have pocketed one-third of the Philippines entire borrowing (Adams 1991). The most notorious project was the $2.1 billion Bataan nuclear power station which was built on an earthquake fault and never used. Marcos is said to have received $80 million in commissions from builder Westinghouse. Filipinos will continue to pay for this corrupt project until 2018, using 'money that should have gone to basic services like schools and hospitals,' the national treasurer, Leonor Briones, said (Guardian, London, 7 September 1999).

Many of the dictators backed by the West during the Cold War are gone, but the creditors demand that the victims must pay the debts of their oppressors. There are two arguments – odious debts and moral hazard – that lenders have no right to demand that the victims repay these debts.

a) Odious debts

In 1898 the United States captured Cuba from Spain, which then demanded that the US pay Cuba's debts. The US refused on the grounds that the debts had been 'imposed on the people of Cuba without their consent and by force of arms.' Furthermore, the US argued that, in such circumstances, 'the creditors, from the beginning, took their chance of the investment.' The concept of 'odious debt' was upheld and formally entered international law in the 1923 judgement of US Chief Justice Taft in the case of Great Britain vs. Costa Rica (Adams 1991).

In 1982, at the height of lending to apartheid South Africa after the United Nations declared apartheid a crime against humanity, two lawyers from the First National Bank of Chicago warned their employers that a majority rule government in South African might not need to repay the loans because 'if the debt of the predecessor is deemed to be “odious”, i.e. the debt proceeds are used against the interests of the local populace, then the debt may not be chargeable against the successor' (Foorman and Jehle 1982).

Finally, the International Development Committee (1998 ¦11, 57) of the British House of Commons noted that 'the bulk of Rwanda's external debt was incurred by the genocidal regime which preceded the current administration. ... Some argue that loans were used by the genocidal regime to purchase weapons and that the current administration and, ultimately the people of Rwanda, should not have to repay these odious debts. ... We further recommend that the [UK] government urge all bilateral creditors, in particular France, to cancel debt incurred by the previous regime.'

In another study, I estimated that nearly a quarter of all poor country debt was incurred by dictators. Since this debt had been 'imposed on the people ... without their consent and by force of arms,' it can be argued that successor governments are not liable for those debts (Hanlon 1998b).

b) Moral hazard

The IMF (1998 p 8) says that 'moral hazard exists when the provision of insurance against a risk encourages a behaviour that makes that risk more likely to occur. In the case of IMF lending, the concern about moral hazard stems from perceptions that the availability of financial assistance may weaken policy discipline, encourage international investors to take on greater risks in the belief that they will only partially suffer the consequences, or both.'

If generations yet unborn have to pay for Marcos' nuclear power station that never worked, for Mobutu's palaces, and loans to a regime labelled by UN resolutions as committing a 'crime against humanity', then there is a serious problem of 'moral hazard'. Lenders need have no compunction about the morality of the loan, or even if it is a sensible loan; repayment is almost guaranteed.

The IMF and World Bank are in an invidious position and suffer the most risk of moral hazard, because they enforce debt repayments. The Democratic Republic of the Congo will want aid and HIPC debt relief, and that requires them to have IMF and World Bank programmes. That, in turn, gives the IMF and World Bank a heavy weapon to insist that the new government of the DR Congo does not suggest that loans to prop up Mobutu were odious and should not be repaid.

Moral hazard can only be avoided if lending institutions are now forced to accept their responsibility for past bad lending. Cancellation of odious debts can thus be seen as necessary to avoid future moral hazard.

c) How the money is spent

If debt is to be cancelled because it is odious or because of moral hazard, then creditors have no right to ask how the money released will be used. And in most past debt cancellations, no questions were asked about how the money released was used. The United States does not ask Britain what it does with the money it should be using to pay its World War I debts to the US.

Nevertheless, if we are arguing that debt should be cancelled to release funds for development and to meet the DAC targets, then it may be reasonable to ask the debtor to show the existence of a mechanism to ensure that the funds are used in that way. Although there is a widespread view that there is a need to prevent corruption and prevent elites from wasting the money on their own consumption, there are sharp differences about how this is to be done. Donor agencies and international financial institutions want increased conditionality. Civil society in debtor countries argues that northern imposed conditionality has failed and that there is a need to support democratic and local control of the funds.

Both sides have problematic records. The international financial institutions and creditor countries encouraged corruption by lending to dictators they knew were putting money in Swiss banks, and more recently their record on adjustment and lending has been poor. When the World Bank in engaged in an internal battle over development policy which led to the resignation of two key officials in early 2000, it is hard to argue that the Bank should be allowed to impose conditions.

Democratisation is increasing in developing countries and civil society is growing, but institutions are still weak and there remain obvious problems of corruption and inexperience.

Neither north nor south have the moral authority or track record to impose conditions to ensure the wise use of funds released by debt cancellation. So a new and more humble partnership is called for, in which new and more democratic systems evolve.

In this context, it is important to remember that debt cancellation does not release a single block of billions of dollars. Rather, it means that payments need not be made over the term of the loan – often 20 years or more. Thus there is time available to gain experience and strengthen the local democratic structures necessary to ensure the best use of funds released for development.

9. CONCLUSION

As in previous debt cycles, reckless lending has led to unpayable debt. But in previous cycles substantial debt was cancelled; in this cycle this has not happened. This paper has attempted to show a number of different reasons for following the historic pattern, and writing off significant debt.

So far, the international community has proposed to write off about $100 billion of the $2554 billion developing country debt. A 'rights based approach' would require the writing off more than $600 billion in debt owed by 71 countries that cannot afford to pay their full debt service and still meet development and human rights targets to which the international community is already committed. Using historic precedents would require the writing off of $1 trillion or more.

It can also be argued that under international law, up to one-quarter of developing country debt is odious debt, which should not be repaid. Furthermore, lenders must be held responsible for their bad lending.

The debate about debt cancellation continues, but history and present need suggest that much more debt cancellation is likely.

REFERENCES

Adams, Patricia (1991), Odious Debts, Earthscan, London.

Aldcroft, Derek (1997), Studies in the Interwar European Economy, Ashgate.

Aggarwal, Vinod (1989), 'Interpreting the History of Mexico's External Debt Crisis', in Eichengreen and Lindert 1989.

Darity, William and Horn, Bobbie (1988), The Loan Pushers, Ballinger – Harper & Row, Cambridge (Mass USA).

Díaz Alejandro, Carlos (1983), 'Stories of the 1930s for the 1980s', in Armella, Pedro et al, Financial Policies and the World Capital Market, University of Chicago Press, Chicago.

Drake, Paul (1989), 'Debt and Democracy in Latin America, 1920-1980s', in Stallings, Barbara and Kaufman, Robert, Debt and democracy in Latin America, Westview, Boulder (Colorado, USA).

Economic Commission on Africa (1999), Economic Report on Africa 1999, United Nations Economic and Social Concil E/ECA/CM/24/3.

Eichengreen, Barry and Lindert, Peter (1989),The international debt crisis in historical perspective, MIT Press, Cambridge (Mass USA).

Eichengreen, Barry and Portes, Richard (1989a), Dealing with Debt: The 1930s and the 1980s, World Bank, Washington, Working Paper WPS 259.

Eichengreen, Barry and Portes, Richard (1989b), 'After the deluge: Default, negotiation and readjustment during the interwar years,' in Eichengreen and Lindert 1989.

Eichengreen, Barry and Portes, Richard (1995), Crisis? What Crisis? Orderly Workouts for Sovereign Debtors, Centre for Economic Policy Research, London.

Evans, Huw (1999), 'Debt relief for the poorest countries: Why did it take so long'. Development Policy Review, vol 17 no 3 pp 267-279

Foorman, James and Jehle, Michael (1982), University of Illinois Law Review, 1982 no 1.

Garrett, John and Travis, Angela (1999), Unfinished Business, Jubilee 2000, London.

Gordon, David and Spickler, Paul, eds (1998), International Glossary on Poverty, Zed, London.

Hanlon, Joseph (1998a) 'We've been here before', Jubilee 2000, London.

Hanlon, Joseph (1998b) 'Dictators and debt', Jubilee 2000, London.

Hanlon, Joseph (1999) 'What Will It Cost to Cancel Unpayable Debt?', Jubilee 2000, London.

Hanlon, Joseph and Pettifor, Ann (2000), Kicking the Habit, Jubilee 2000, London.

IMF (International Monetary Fund)(1998), World Economic Outlook 1998, Washington DC.

IMF (1999) HIPC Initiative Consultation Meeting Background Material, Washington DC.

International Development Committee (1998), Debt Relief, 3rd report, House of Commons, London.

Kindleberger, Charles (1978), Manias, panics and crashes, Basic Books – Macmillan, London.

Kindleberger, Charles (1996), Manias, panics and crashes, 3rd edition, John Wiley & Sons, New York.

Kuhn Loeb Lehman Brothers (1979), 'The Republic of Indonesia', London.

Lindert, Peter (1989), 'Response to debt crisis: What is different about the 1980s?', in Eichengreen and Lindert, 1989

Lissakers, Karin (1992), Banks, borrowers and the establishment, Basic Books, New York.

Maxwell, Simon (1999), 'What can we do with a rights-based approach to development?' , ODI Briefing Paper, 1999 (3) September, London.

Northover, H, Joyner, K and Woodward, D (1998), 'A human development approach to debt relief for the world's poor', Cafod, London

OECD DAC (Organisation for Economic Co-operation and Development – Development Assistance Committee) (1997), 1996 Development Co-operation Report, Paris.

Sampson, Anthony (1981), The Money Lenders, Coronet – Hodding & Stoughton, London.

Ugarteche, Oscar (1999), 'Where there is a will there is a way', paper delivered at Justicia y Paz conference on foreign debt, Barcelona, November.

Ugarteche, Oscar (2000), 'A fair and transparent arbitration process', paper presented at ILAS, University of London, April.

UNCTAD (United Nations Conference on Trade and Development) (1998), Trade and Development Report 1998, Geneva.

UNDP (United Nations Development Programme) (1996), Human Development Report 1996 , Oxford University Press, New York.

UNDP (1997), Human Development Report 1997, Oxford University Press, New York.

UNDP (1998), Human Development Report 1998, Oxford University Press, New York.

UNDP (1999), Human Development Report 1999, Oxford University Press, New York.

UNDP, UNESCO, UNFPA, UNICEF, WHO and the World Bank (1998), Implementing the 20/20 Initiative, Unicef, New York.

United States Treasury (1997), 'Indebtedness of foreign governments to the United States arising from World War I as of June 30, 1997', Washington DC, 8 September 1997.

World Bank (1990), World Development Report 1990, Oxford University Press, Oxford.

World Bank (1991), World Debt Tables 1991-92, Washington DC.

World Bank (1993), World Development Report 1993, Oxford University Press, New York.

World Bank (1994), World Debt Tables 1994-95, Washington DC.

World Bank (1996), African Development Indicators 1996, Washington DC.

World Bank (1997), Global Development Finance 1997, Washington DC.

World Bank (1998a), Global Development Finance 1998, Washington DC.

World Bank (1998b), World Development Report 1998/99, Oxford University Press, New York.

World Bank (1999a), Global Development Finance 1999, Washington DC.

World Bank (1999b), World Bank Annual Report 1999, Washington DC.

World Bank (2000a), Global Development Finance 2000, Washington DC.

World Bank (2000b), World Development Indicators 2000, Washington DC.

World Bank (2000c), World Development Report 1999/2000, Oxford University Press, New York.

Footnotes

[1] Only three countries repaid their World War I debts to the United States: Cuba, Finland, and Liberia.

[2] See, for example Eichengreen and Lindert (1989 p 2).

[3] See, for example, Kindleberger (1996).

[4] London interbank offer rate – LIBOR – minus the US GDP deflator

[5] The World Bank and IMF use the concept of 'net present value' (NPV) of debt, which is the value which would have to be deposited in an interest bearing account to exactly pay off the debt – thus concessional debt has an NPV less than the face value and commercial debt an NPV more than the face value. In 1994 the World Bank (1994, vol 1, p 40) concluded that ratios of NPV debt to export earnings (XGS) in excess of 200% had 'generally proved unsustainable over the medium term'. In 1996, with no public justification, the World Bank and IMF announced that under HIPC debt would be 'sustainable' if the NPV to XGS ratio was between 200% and 250% – a level which just two years before had been said to be 'unsustainable'. For the 41 HIPC countries, debt service due averages 10% of NPV debt, so an NPV to XGS ratio between is equivalent to an annual debt service to export ratio between 20% and 25%, which was the second sustainability criterion set under HIPC.

[6] At Köln the G7 changed one 'sustainability' criterion but not other – the NPV debt to XGS ratio was reduced to 150%, while the annual debt service to export ratio was left between 20% and 25%. But because debt service due for HIPCs is normally 10% of NPV debt, this has the effect of reducing the second criterion to 15%.

[7] Face value, not NPV.

[8] We looked at countries with a 1995 GDP at purchasing power parity (PPP) less than $4500, plus those countries defined by the World Bank in 2000 as middle income countries which are severely or moderately indebted. We exclude Europe and the former USSR, Eritrea, and non-reporting countries (Afghanistan, Cuba, DR Korea, Namibia and Iraq).

[9] Included in the report 'Shaping the 21st Century: The Contribution of Development Co-operation' adopted at the 34th High Level Meeting of the Development Assistance Committee, 6-7 May 1996 (OECD DAC 1997).

[10] Note that there is a strong counter-argument that this is inadequate. The DAC targets only call for halving the number of people living in 'extreme poverty', and it is argued that the other half who remain in extreme poverty are having their article 25 right to an adequate standard of living violated. But Maxwell and others argue that a rights-based approach is an advance and involves changing performance standards. Since the DAC targets have been widely accepted, it seems reasonable to use them here.

[11] For example, Baroness Amos, the British government spokesperson for international development in the House of Lords, on 25 October 1999 in a debate on Mozambique said 'we do not support the call for total debt forgiveness' for Mozambique.

[12] Although these can only be rough estimates, we have been presenting preliminary versions of these calculations over a period of a year, including to a seminar in London on 18 March 1999 attended by representatives of the World Bank and IMF, and they have not been challenged.

[13] 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).

[14] From UNDP 1999 and World Bank 2000b, with gaps filled with estimates based on UNDP 1999 averages.

[15] From World Bank 2000b. Where data does not exist, we use the averages from UNDP 1998, which are 25.9% for medium human development and 43.8% for low human development.

[16] Accounting for inflation, $ per day is equivalent to $526 current PPP dollars per year. So the share of GDP excluded from taxation becomes: (526/(GDP per capita at current PPP))*(1 - ((poverty gap)*(% below $1 per day)))The poverty gap is taken from the World Bank (2000b, Table 2.7, pp 62-64). Where no data is given, we take 15% as an estimate.

[17] Data on government revenue and spending is available from only a limited number of countries. UNDP (1999 table 12 p 184) gives 1997 average tax revenue as 16% of GDP for 'medium human development' countries and 28% for the 'high human development' countries. In 1994, government revenue in sub-Saharan Africa was 21% of GDP (World Bank 1996 p 187).

[18] This is hard to estimate because of the complex nature of donor conditions, so this level has been chosen to be generous to donors. OECD DAC (1997, tables 2 &27, pp A4, A46) tables suggest that in 1995 only $7.5 billion was available for debt service – which is less than one-third of grants as given for that year by the World Bank (1997).

[19] Total debt service due is for 1998, the last year for which there is good data. This is taken to be long term debt service due for 1998 (World Bank 1999a) plus short term debt service actually paid in 1998 (as calculated from World Bank 2000a).

[20] GDP is 1998 GDP (World Bank 2000b) substituted with 1997 data (UNDP 1999) in a some cases where the World Bank has no data, and estimated in a few places where neither source has information.

[21] It is not precisely correct to say that the portion of debt needing to be cancelled is the same as the proportion of debt service that cannot be paid, because of the different repayment schedules of different kinds of debt. However data is not available to do a correct detailed calculation, so this method must be used as the only available one to give an estimate. If it is assumed that debt for each country must be cancelled in proportion to the debt service that cannot be paid, table 6 shows that $346 billion in debt would need to be cancelled – nearly $350 billion.

[22] The bracketed phrase is in the original memo. The publisher of the memo is Karin Lissakers, who has since become US Executive Director of the IMF.