papers - ddci research

Papers	Education - Training	Exhibition	Country profiles	Economic Literacy
DDCI Research Conference Papers Policy Submissions Briefing Papers Articles

Can the World Bank and IMF Cancel 100% of poor country debts?
Written and Researched by Sony Kapoor

Jubilee Research at the New Economics Foundation
For
Debt and Development Coalition Ireland

Foreword

Debt and Development Coalition Ireland runs Ireland's campaignign for developing country debt cancellation. We have collaborated with Jubilee Research at the New Economics Foundation on this report in order to challenge IMF and World Bank opposition to 100% cancellation of debts owed to them by the poorest countries. As the IMF and World Bank are the most significant creditors of the poorest countries they are a major obstacle to a just and speedy end to the two decade old debt crisis.

In 2002 the Irish government launched its 'Policy on Developing Country Debt' which supports 100% cancellation for the poorest countries. This put Ireland in a leading position internationally on the debt crisis. The policy was welcomed by Debt and Development Coalition Ireland as a fitting response to the hundreds of thousands of people around Ireland who signed the Jubilee 2000 petition for debt cancellation. However, while recognising the urgent need for debt cancellation, the government uncritically accepts IMF and World Bank arguments against using their own reserves for this purpose. By taking this position, the government is tying its hands and making it impossible to effectively promote its debt policy internationally.

We hope that this report will fuel the debate on IMF and World Bank debt cancellation. In particular we hope it will encourage the Irish government to follow through on its debt policy and challenge the IMF and World Bank to look to their own resources to cancel the debt owed to them by the poorest countries.

Debt and Development Coalition Ireland
September 2003

Executive summary

The international community has rallied behind the Millennium Development Goals, and widely accepted that debt-servicing capacity should be assessed relative to the country's need for achieving the goals. Despite this, conditions in most HIPCs (Heavily Indebted Poor Countries) continue to deteriorate. Rather than spend all possible resources on basic needs such as health and education, the HIPCs are still servicing unsustainable levels of debt.

While the G7 nations have promised to cancel debt owed to them, both the International Monetary Fund (IMF) and the World Bank have pleaded poverty in a bid to abrogate their responsibility towards the HIPC nations. They have said that any additional debt cancellation through the use of their resources would seriously endanger their financial soundness and sustainability.

However, using rigorous financial analysis this paper shows that both the IMF and the World Bank have ample resources to cancel all the HIPC debt. We show that they could finance this debt cancellation without in any way jeopardizing their normal operations.

Specifically this report finds

• That the IMF has largely identified sources of funds for the $2 billion (NPV) of the HIPC debt that it has already committed to cancel. Of this, more than half has come directly or indirectly[1] through donor country contributions and the balance through the investment income from off market gold sales.

• That the World Bank has only identified sources to fund less than its existing commitment of $6.4 billion[2] (NPV) to the HIPC debt cancellation effort. Of these resources, more than half have come through donor country contributions and the balance through income transfers from the International Bank for Reconstruction and Development (IBRD).

• That the IMF has enough resources to fund the cancellation of all of the $5 billion (NPV) of additional HIPC debt owed to it. In fact, the IMF can afford the cancellation of even more debt.

• That the IBRD has enough resources to fund the cancellation of all of the $13 billion[3] (NPV) additional HIPC debt owed to the World Bank group.

While trying to highlight their self proclaimed paucity of resources, the IMF and the World Bank have sought to underplay their considerable financial strength, which is underpinned by their distinctive political and financial structure and their special role within the international financial system. They have incorrectly made implicit comparisons with the private sector to highlight their 'poverty'.

However, their unique status based on explicit guarantees from donor (mostly G7) countries makes them highly resource rich. We compare both the IMF and the IBRD with the private sector after suitably factoring in these guarantees. We find

• That both the IMF and the IBRD derive significant financial benefits from the explicit guarantees provided by donor countries which makes them far more resource rich than their closest private sector counterparts

• That factoring in the benefits of these explicit guarantees, both the IMF and IBRD are overcapitalised and hence inefficient in relation to levels of financial prudence in the private sector

• That credit analysts and other debt market professionals have privately confirmed this finding

As the two institutions have enough resources to afford 100% HIPC debt cancellation, we recommend specific mechanisms to finance this. These mechanisms successfully address all the concerns (some genuine) that these institutions have raised about the impact of 100% debt cancellation. We recommend

• That the IMF should sell some of its gold reserves directly into the market as recommended in the report and use the proceeds to bankroll the full cancellation of the HIPC debt owed to itself

• That the IBRD mobilize its internal resources by using a combination of retained earnings and future income allocations as recommended in this report to fully bankroll the total cancellation of the HIPC debt owed to the World Bank group

In response to the growing call for 100% debt cancellation, the IMF and the World Bank have written a joint paper[4] criticising the concept and claiming how it would seriously jeopardize their finances. This report rebuts these arguments.

The IBRD has often implied that if it allocates additional resources towards HIPC debt cancellation, its much-valued AAA credit rating would be threatened. It has also implied that allocation of these additional resources would also drive up its borrowing costs significantly. However through the analysis of the serious problems at the African Development Bank (AfDB) throughout the 1990s and their limited market impact, we show that the real anchor behind the creditworthiness of multilateral development banks such as the IBRD is political in nature[5].

In the 1990s the AfDB was beset with financial and operational problems of such a serious nature that if it were a private institution, its viability would definitely have been questioned. However, because of the unique nature of the political and financial guarantees that the AfDB enjoys (similar to the ones that the IBRD has), it managed to hold on to its AAA credit rating with Moody's, one of the most respectable credit rating agencies in the world. Also, its cost of borrowing in the private capital markets did not go up significantly[5].

Sitting atop billions of dollars of untapped resources, the IMF and the World Bank are amongst the most prosperous financial institutions in the world. It is rather astonishing then, to see them plead poverty over the issue of cancelling 100% of the debts of countries they have already identified as having "unsustainable" debts - the HIPCs. Not only have they refused to use their resources for 100% debt cancellation, but they have also been niggardly about paying for their existing commitments.

Under current plans, donor country taxpayers are financing the lion's share of debt cancellation and the illusion of equal burden sharing among creditors, the centrepiece of the HIPC initiative, has been shattered by the refusal on the part of the Bank and the Fund to contribute their fair share.

As a result the HIPC countries will be left with unsustainable levels of debts - debts that can only be repaid at great human cost to their citizens. We firmly believe that creditor rights should not supersede human rights and the Fund and the Bank should muster the internal funding necessary to fulfil what we hope is more than just the 'spirit of millennial rhetoric'.

Altruism and idealism aside, the debt of the HIPCs is also uncollectable; they cannot afford to pay, and public lenders must do what private lenders do in these circumstances, get out of denial and face reality.

The IMF and the World Bank have for years written a self-delusional record of never making a bad loan by refusing to ever recognize a formal loss. They have a deep-rooted system of masking default by rolling over bad debts each year, with enough added to cover new interest costs. These are financially unsound practices that would invite regulatory reprimand in any private sector environment.

Fortunately, as HIPC debt has mounted, so have IMF and World Bank reserves. These have accumulated on account of the special structure of the IMF and the World Bank, which offers large capital and credit guarantees. The World Bank, for instance, generates upwards of $2 billion each year by investing its equity capital and through arbitrage between the favourable borrowing rates it enjoys (owing to guarantees of its donor members) and the higher yields on the market instruments in which it reinvests.

In this report, using rigorous financial analysis, we show that the Bank and the Fund are richly endowed with resources. They hold a pre-eminent position in the international capital markets, which will in no way be compromised by the allocation of a part of their ample resources to 100% debt cancellation.

What are the Millennium Development Goals and why are they vital to HIPCs?

"We will spare no effort to free our fellow men, women and children from the abject and dehumanising conditions of extreme poverty, to which more than a billion of them are currently subjected. We are committed to making the right to development a reality for everyone and to freeing the entire human race from want".[6]
United Nations Millennium Declaration

In the year 2000, the world's leaders met in the United Nations General Assembly to set out a new global vision for humanity. They agreed to worthy goals, subsequently known as the Millennium Development Goals[7] - to eliminate world poverty by the year 2015; to achieve universal primary education; to promote gender equality and empower women; to reduce child mortality; improve maternal health; to combat HIV/AIDS and other diseases; and to ensure environmental sustainability.

Since then, these have been adopted by all major donor agencies as guiding principles for their strategies for poverty eradication. The OECD 'confirmed their commitment to reducing poverty in all its dimensions and to achieving the seven International Development Goals.[8] The IMF and World Bank too claim to have coordinated their efforts behind this set of goals. More importantly, the adoption of the targets has motivated a fundamental shift within development thinking - away from a narrow focus on inputs, towards a fundamental concern with outcomes for the poor of the world.

Amongst the countries of the world farthest away from meeting these goals are the 42 HIPCs (heavily indebted poor countries), mostly concentrated in sub-Saharan Africa. They are abjectly poor and are becoming worse off. Average per capita real income is about $300 a year, down from about $400 in 1980. Unwholesome debt-to-export ratios are matched by an unhealthy life expectancy of 51 years.

Though their situation is desperate and the suffering of their people dismal, there is a tiny ray of light at the end of the dark tunnel. This ray is the hope of collective conscientious action by the international community to pool their resources towards the implementation of the MDGs in the HIPCs.

How much would it cost the HIPCs to implement Millennium Development Goals and why is 100% debt cancellation so important?

Ernest Zedillo, in his report of the High Level Panel for Financing for Development, has assessed that total additional resources of $50 billion per year will be needed to meet these targets worldwide. Additionally, it has been estimated that even if all the debts of the HIPC countries are cancelled, they will need an additional $30 billion in aid each year if there is to be any hope of halving poverty and hunger, while for meeting the other goals, an additional amount of $16.5 billion will be needed.[9]

As the first step towards implementing the MDGs for the HIPCs, there is an urgent and immediate need to cancel all of their crippling outstanding debts.

Together with other debt campaigns, we firmly believe that debt cancellation as envisioned under the HIPC initiative must not be based on arbitrary debt to- export ratios, but instead should be linked to the resources the country needs to meet the MDGs.

This definition of debt sustainability has gained wide acceptance - not only in the non-governmental community, but also in the United Nations, amongst African governments, HIPC Finance Ministers, and even northern governments such as Ireland. The latest Human Development Report produced by the United Nations Development Programme (UNDP), for example, argues that 'debt servicing capacity should be assessed relative to the country's needs for achieving the Goals. For many countries this will require full debt cancellation'.

The Monterrey Consensus on Financing for Development, agreed in March 2002 states that 'future reviews of debt sustainability should also bear in mind the impact of debt cancellation on progress towards achievement of the development goals contained in the Millennium Declaration.' African leaders, in the New Partnership for Africa's Development (NEPAD), have argued that their long-term objective is 'to link debt cancellation with costed poverty reduction outcomes' - in other words, the MDGs.

Poor countries prepared to commit resources to meeting the basic needs and economic rights of their populations should not be prevented from doing so because of the need to pay back debts to rich creditor countries and institutions.

What is the amount of HIPC debt cancellation required and who is paying for it?

Collective HIPC debt has now spiralled to more than $200 billion in nominal terms, but the cost of a write-off is less demanding than it appears. As, the majority of the loans carry a concessionary interest rate a cash offset equal to the net present value (NPV), of roughly 70% of the nominal amount of the debt is all that is required.

Thus the nominal $176 billion of official loans, which represents 82% of the total due, translates into some $80 billion (NPV) in effective bilateral (mostly G7) obligations and $45 billion (NPV) owed to the 27 multilateral agencies, of which the World Bank and the IMF largest hold a dominant 60% share.

The G7 creditor nations have already promised a near total write-off of HIPC debts owed to them.

Have the IMF and the World Bank committed themselves to the HIPC Initiative?

As early as 1996, the Committee of the Board of Governors of the IMF

"warmly endorsed the program of action proposed by the Fund and the Bank to ensure that the heavily indebted poor countries (HIPCs) that have shown a sound track record of economic adjustment can attain a sustainable debt situation over the medium term. It endorsed the conclusions by the Executive Board on financing the continuation of ESAF (the Enhanced Structural Adjustment Facility) and the Fund's participation in the Initiative to assist the HIPCs to which all members are committed…It also reaffirmed the importance of the Fund's preferred creditor status". (Our italics)

James D. Wolfensohn, President, World Bank. September 1996.

"The HIPC debt initiative was proposed by the World Bank and IMF and agreed by governments around the world in the fall of 1996. It was the first comprehensive approach to reduce the external debt of the world's poorest, most heavily indebted countries, and represented an important step forward in placing debt cancellation within an overall framework of poverty reduction." (Our italics)[10]

Both the IMF and the World Bank are at the forefront of the HIPC initiative and are the primary institutions involved in the execution of the actions required by the initiative. They have kept a tight control over the HIPC debt cancellation mechanisms and implicitly profess ownership of the Initiative.

The World Bank and the IMF are also the HIPCs' main creditors among multilateral institutions and significant contributors to the HIPC Initiative in terms of intellectual and technical resources. However, they both have been very wary of cancelling any HIPC debts owed to them.

We believe, that both the Bank and the Fund should cancel all the HIPC debt owed to them. However, they have reacted negatively to this suggestion and have sought to argue that any further mobilization of their own resources towards the debt cancellation effort would seriously jeopardize their financial soundness and curtail their ability to undertake their normal operations.

Are the IMF and the World Bank contributing to the HIPC initiative?

It now seems that those who sought the spotlight are now retreating to the end of the queue when the costs of the HIPC initiative need to be paid. The IMF and the World Bank are saying: not relief for all 42 countries, not total relief for any country, not now, and, certainly, not us.[11]

Pleading scarcity of resources, they plan to write off a miserly $8.4 billion (NPV), less than one-third of the amount owed to them, and that too only after every other avenue has been exhausted, and even then mostly with funds from donor countries. They have committed only limited amounts from their own resources for the funding of these cancellations. They will still be the two biggest creditors of HIPCs even after all cancellation has been provided through the original and enhanced HIPC initiatives.

But do the IMF and the World Bank have enough resources?

The IMF and the World Bank hold a wealth of resources on their own balance sheets - about $500 billion in effective capital and $40 billion in provisions for loan losses and reserves.

They can easily marshal internal resources for total debt cancellation as it represents just 5% of their effective capital and 65% of provisions for losses and reserves. These are accumulated for the day when borrowers cannot pay and they would not be disabled by such an insignificant drop in equity. Contrary to their claims, the viability of the IMF and the World Bank is not threatened.

What can we do to convince the IMF and the World Bank to cancel 100% of the HIPC debt?

This paper is an attempt to draw attention to the resources that the IMF and the World Bank have at their disposal and also to counter the arguments that they have made against the wisdom of using these resources for the purpose of 100% debt cancellation.

In this report, we demonstrate how significant additional resources could be mobilized in an efficient and financially prudent way by both the institutions. We begin by offering concrete proposals on how these resources could be mobilised to meet the commitment to sustainability.

Next, we rebut, point by point, the IMF's and the World Bank's opposition to 100% cancellation, outlined in a recent paper.[12] Finally, in the appendices we back up these claims mainly through the use of prudential comparisons with the private sector as appropriate.

The HIPC countries owe the IMF a total of about $7 billion (NPV) of debt, of which the Fund has already cancelled or promised to cancel about $2 billion (NPV). This means that for all of the HIPC IMF debt to be cancelled another $5 billion (NPV) must be mobilized.

The IMF, like other multilateral creditors, is not writing off debt outright. Instead, in efforts to preserve its "preferred creditor" status, it is keeping the HIPC loans on its balance sheet. However, non-repayment on these loans will be substituted for by payments from the HIPC account, as and when the repayments become due.

The IMF has already identified financing to meet most of its current share of debt cancellation from internal resources. This includes

• Earnings on investment profits from the off-market gold transactions undertaken especially for this purpose

• Contributions from its members and other sources[13]

IMF gold reserves

The IMF holds the largest reserves of gold in the world after the United States and Germany. It's gold reserves amount to 103.4 million ounces and the IMF holds them at a book value of only SDR 35($48) per ounce as opposed to the current market price of $376 per ounce. Hence, there is a lot of latent value locked up in the IMF's gold reserves. We believe that this can be put to productive use such as providing for HIPC debt cancellation through one of the mechanisms suggested here.

Through off market gold transactions

We believe that the IMF could use off-market gold transactions of the type it conducted in 1999-2000 to mobilize resources.

In December 1999, the executive board of the IMF authorized off-market transactions in gold of up to 14 million ounces to help finance the IMF contribution to the HIPC initiative. By April 2000, the IMF had carried out transactions involving 12.9 million ounces of gold with Mexico and Brazil both of which had obligations to the IMF falling due.

As a first step the IMF sold gold to Mexico and Brazil at the prevailing market price and placed the profits from the proceeds in a special account. As a second step, it immediately accepted back the same amount of gold at the market price in lieu of the financial obligations falling due. The net effect of these two steps was to leave the amount of physical gold with the IMF unchanged.

That off-market gold transaction raised a sum of $3.9 billion, the interest from which is financing a part of the Fund's share of the HIPC initiative. That transaction involved only about 13% of the IMF's gold holdings, which in turn represent only about 10% of the world's total gold holdings for monetary purposes. If the IMF re-values all of its gold holdings through similar off-market transactions, it has the potential to generate up to $30 billion in income. In comparison, the IMF only needs $5 billion to provide for the total cancellation of debts owed to it by the HIPCs.

We suggest that the IMF engage in off-market transactions and revalue about a third of its gold reserves. This would generate an income of $10 billion of which, $5 billion can be appropriated towards HIPC debt cancellation instantly. The rest of the $5 billion should be invested and the income from this investment should be used to offset the loss of income resulting from the decrease in interest free reserves.[14]

Alternatively, the IMF could revalue all of its gold reserves and invest all of the $30 billion proceeds to generate income. A part of this income could then be used towards providing 100% HIPC debt cancellation and the rest of it could be used to offset the loss of income resulting from the decrease in interest free reserves.

All that would be needed would be for IMF members to authorize the use of gold reserves in the same way as the previous off-market transactions. There is a precedent for this solution, discussed above, and used by the IMF as recently as 2000.

Through outright gold sales

We believe that it is actually preferable to raise the amount needed for 100% HIPC debt cancellation, through straightforward market sales, than through the tortured "off-market" procedure of 1999-2000. However, that may prove the politically more difficult option due to the fears that exist about the adverse impact on gold prices.

It is of course right that the IMF should, as far as possible, avoid any disruption to the functioning of the gold market especially insofar as such a disruption may destabilize the international financial situation. However, as we argue later, we have good reason to believe that the impact of the sale of gold on the market would be minimal.

As a first step the IMF would need to estimate the amount of gold it would need to sell to raise enough resources for HIPC debt cancellation. It would be prudent to take a conservative estimate for the price at which it can sell the gold. The current price (as on 1st September 2003) of gold is $376 per ounce and the average price since 1980 has been $370 per ounce. The IMF being one of the most financially astute institutions in the world should be able to lock in a good price for its gold through the use of derivative[15] transactions.

Even so, this price would almost certainly be somewhat lower than current levels as the IMF unloads more gold into the market. Hence, the IMF can take a conservative estimate, say at a 20% discount to the current market price and assume that it will be able to get a price of $300 per ounce for its gold. Of the money that the IMF gets from the gold sales, the book value of SDR 35($48) per ounce will go to the GRA (General Resources Account) and the excess of $252 per ounce will be allocated to the SDA (Special Disbursement Account). For a current financing need of $5 billion, this would mean that the IMF would need to sell about 20 million ounces of gold (or less than a fifth of its total gold holdings) into the market.

We suggest that the IMF sell 20 million ounces of its gold in the open market or through private transactions indexed to the market price. In order to at once get the best price and minimize the impact on the gold market, the Fund should spread the sale of this gold over a period of time. We believe that a period of 3-4 years would be a judicious compromise between timeliness and minimizing market impact. This can be achieved either through opportunistic sales in the spot market or through an appropriate use of the futures (or other derivatives) available in the market. As discussed in the next two paragraphs, an additional annual supply of about 5 million ounces of gold in the existing market would not dramatically impact the supply demand dynamics and hence would have a relatively small impact on the price.

The annual production of gold in the world has averaged around 80 million ounces over 1997-2002 and the demand has averaged well above 125 million ounces during the same period. Central bankers have been decreasing their stocks of gold for some time now (especially under the 1999 Washington agreement) and have sold about 10 million ounces annually over the past few years.

Also, the gold market is one of the most liquid markets in the world and had a turnover of almost $4.5 trillion (45% of the turnover of the NYSE) in the year up to March 2003. This demonstrates that there is ample room to bring new gold to the market without tipping the supply demand balance. The IMF's fears of seriously disrupting the gold market are unfounded, especially if it follows the suggestions made above.

Through renewed SDR allocations

George Soros has urged the IMF to revive periodic issues of the SDRs and use the proceeds to finance development. However, this actually amounts to a straightforward increase in donor aid. Its main advantage is that it solves the problem of the distribution of the financial burden of additional aid amongst donors by linking it to their IMF quotas. While we strongly favour increasing aid allocation by developed nations, in this report we focus exclusively on the resources that the IMF and the World Bank already have at their disposal.

Through increasing net income from investment earnings

The Fund is not a private institution, and does not set interest rates according to market conditions. Instead it targets a level of net income (effectively the profit it thinks is needed to fund its operations) and, because of its powerful monopoly position as a lender to countries in financial distress, sets interest rates accordingly. Hence, the Fund could potentially raise (or target) higher levels of income by setting higher interest rates to allocate to the debt reduction initiative. However, there are some problems with this.

Firstly, it raises the question of fairness. It would be unfair to expect countries, whose economies are under duress (much of IMF lending is to such countries), to finance the HIPC debt cancellation especially in a state of economic stress.

Secondly, the percentage of stand-by facilities (that are not withdrawn) has gone up significantly in recent times. The IMF can only charge a limited commitment fee for non-disbursed funds and it would be politically unpopular to hike the charges for such a facility.

Thirdly, the cost of borrowing from the IMF varies inversely as the volume of outstanding borrowing. This arises because of the targeting of net income. Big debtor countries are currently engaged in prepaying their outstanding loans to the IMF. This means that the total stock of debt on which the IMF can charge interest is set to go down and will thus further limit the possibility of generating additional net income.

Hence the generation of extra resources through targeting a higher net income funded primarily by higher interest rates is, in our view, not viable for the abovementioned reasons.

However, we do favour targeting higher net income, but funded instead by investment earnings from profits through gold sales.

That is why we believe that the IMF should undertake to sell its gold reserves in the open market. Spread over a sufficient number of years, this is unlikely to have any significant impact on the price of gold. This should generate resources of the order of $30 billion, which when invested in assets yielding an average of 6% would generate an annual income stream of $1.5 billion. This could then be used in any way that the IMF board deemed fit, including reducing borrowing costs for countries under financial duress.

If the IMF board so wished, then part of this income could also be transferred to other multilateral development banks to cover cancellation of debt owed to these by low-income countries.

Advantages of selling gold

We believe that there are several advantages to the sale of IMF gold for the purpose of funding HIPC debt cancellation.

First, it would be in the spirit of an earlier use of gold reserves. From 1976-1980, the IMF sold approximately one third (50 million ounces) of its then-existing gold holdings following an agreement by its members to reduce the role of gold in the international monetary system (Appendix 3, paragraph 23)[16].

Half of this amount was sold in restitution to members at the then-official price of SDR 35 per ounce; the other half was auctioned to the market to finance the trust fund, which supported concessional lending by the IMF to low-income countries.

Second, in 1999-2000, the IMF once again carried out transactions involving 12.9 million ounces of gold in order to generate resources for HIPC debt cancellation. These precedents indicate that when the political will is there, resources can be found for debt cancellation. This will is also demonstrated by the recent spate of bilateral HIPC debt cancellation and additional aid promised by many developed economies.

Third, IMF members having long since agreed to reduce the role of gold in the international monetary system should now put their votes behind the sale of IMF gold.

Again there is a precedent here. In 1979-80, staff discussions took place regarding the establishment of a Substitution Account[17] (Appendix 3, paragraph 21). The then Fund management supported the use of a (substantial) part of the Fund's gold holdings to ensure the viability of the Account, and also the sale of a small portion of the Fund's gold and use of the profits to create an investment fund and thereby strengthen the Fund's income position. However, the question of a Substitution Account was later dropped, as was the issue of gold sale for the purpose of deriving income for the Fund. Sale of gold by the Fund was then seen to be problematic because of its possible adverse impact on the gold price and the value of central bank gold holdings.

Things have moved on since then and a sale now would just be belated recognition of the relatively insignificant role that gold now plays in the in monetary system. The gold market has become bigger, more liquid and diverse and central banks themselves have been off loading their reserves in the market for years.[18]

For all these reasons, we believe that this is the right time to re-consider the sale of IMF gold to fund 100% HIPC debt cancellation.

Hence we recommend that the IMF sell 20 million ounces of its gold in the open market or through private transactions indexed to the market price. In order to at once get the best price and minimize the impact on the gold market, the Fund should spread the sale of this gold over a period of time. We believe that a period of 3-4 years would be a judicious compromise between timeliness and minimizing market impact.

In response to repeated calls for additional (total) debt cancellation for the HIPC countries amounting to about $5 billion (NPV), the IMF has responded with an official position in the form of a joint paper with the World Bank entitled "100 Percent Debt Cancellation?"[19]

The IMF claims that any additional debt cancellation would necessarily come at the cost of other developing countries. The IMF says that the PRGF, which is close to being a permanent facility will in future be financed purely by reflows. Additional debt cancellation, says the IMF, would deplete the resources of the PRGF and force the closure of the facility leading to the withdrawal of the IMF from concessional lending.

We contest this view. We strongly believe in the principle of additionality in the context of debt cancellation and do not believe that HIPC debt cancellation should come at the expense of other developing countries. The main recommendations of this report provide clearly defined ways of mobilizing additional Fund resources for the purpose of debt cancellation without either using PRGF resources or at the expense of other developing countries. Gold sales to the tune of 20 million ounces spread over a number of years would generate enough additional resources for debt cancellation without a significant impact on gold prices. (And hence on gold producing developing nations)

The IMF argues that total debt cancellation would do serious damage by fundamentally changing its present character as a co-operative monetary institution designed to promote the stability of the international financial system and provide temporary balance of payments support for members in need.

We are puzzled by this statement especially since the IMF's charter was constituted under an exchange rate system based on the gold standard which has long since been abandoned. The IMF is diverging from its institutional purpose, and is currently involved in development and poverty reduction (regarded as the domain of the World Bank) and in detailed, prescriptive, long lasting structural adjustment programs. This is in contrast to its original purpose: to act as a "fire-fighter" for the global economy and to provide temporary balance of payments support to members facing difficulty. Given this context, the IMF's argument is completely irrelevant.

The IMF asserts that debt cancellation would impair the Fund's financial integrity and argues that its gold reserves constitute an integral part of its financial position; that the exceptional decision to use the income from investments of the profits from limited off market gold sales to finance part of its contribution to the HIPC Initiative has already proven to be a substantial cost to the institution and its members.

We rebut this statement. Even the Fund does not believe that gold reserves are indispensable (Appendix 3, paragraph 21)[20]. We discuss this point in greater detail below.

Having analysed the Fund's financial statements, we conclude that the off market gold sales did not have a substantial cost for the IMF. The transaction and the subsequent commitment of the investment income towards HIPC debt cancellation did lead to a decrease in the level of interest free assets (Appendix 3, paragraph 20). However this decrease could not only have been avoided; but could have been converted into an increase if the Fund had sold the gold outright rather than using off market transactions (Appendix 4, paragraph 25).

The IMF claims that holding undervalued gold provides fundamental strength to its balance sheet. It also claims that any transaction involving gold should avoid weakening the IMF's overall financial position.

We contend that the sale of IMF gold would strengthen rather than weaken its position. The gold on the IMF's books is currently held at SDR 35($48) per ounce. If the IMF were to sell the gold in the market or even replicate the recent off market transactions, it would be able to retain a substantial part (more than three fourths) of the profits on its balance sheet even after providing for 100% HIPC debt cancellation. These added reserves (estimated at well over $20 billion), which would be much more substantial than the current book value of the gold (and over and above), would actually provide the IMF balance sheet with fundamental strength.

The IMF claims that gold holdings provide the IMF with operational manoeuvrability both as regards the use of its resources and through adding credibility to its precautionary balances. It also claims that in these respects, the benefits of the IMF's gold holdings are passed on to the membership at large, to both creditors and debtors. More specifically, it argues that its gold reserves allow even conservative central bankers to treat quota increases as asset swaps[18] rather than a donation because they know that the IMF could, in the event of non-payment of some of its loans, sell the gold to make up for the loss.

We assert that this is not only inaccurate, but also misleading.

Due to the preferred creditor status of the IMF and its pre-eminent role in international finance as the central banker's central banker, there has been no default on IMF obligations in the past. There have been several very severe currency and debt crises in the past two decades and the IMF has had enough resources to cope with ample resources to spare. There is no evidence that it will not be able to cope with future crises without dipping into gold reserves especially after the most recent quota increase in 2000.

Unlike the Multilateral Development Banks the IMF does not need reserves to reassure lenders and to permit it to borrow cheaply. The only function is to reassure central bankers that their funds are safe with the IMF. The IMF's balance sheet is perfectly sound for central bankers to treat quota increases as asset swaps[21], gold reserves or not.

We believe that the claim about the benefits of gold holdings being passed on to members is especially misleading as these stated benefits are intangible and insignificant compared to the significant opportunity cost of holding undervalued gold reserves. According to our calculations[22], if the IMF had sold its gold holdings into the market gradually over a period of say twenty years from 1980, it would have had current reserves of about two times the current market value (or $60 billion) of its gold holdings, which stands at about $30 billion. If anything, IMF members have lost rather than gained because of the conservative approach to gold reserves.

It is important to recollect, that Fund staff in 1979-1980 wanted to sell the gold and invest proceeds in income generating assets but were thwarted by lack of political will and shallow markets (Appendix 3, paragraph 21). Given this earlier position and taking note that both the political will to put the Fund's gold to productive use and deep and liquid gold markets to facilitate this exist now, we are baffled by the current position of the IMF on the issue.

The IMF further states in it's policy on gold[23] that it should continue to hold large quantities of gold among its assets, not only for prudential reasons, but also to meet contingencies.

Once again, we assert that the IMF position on this goes contrary to good financial practice. The IMF has long sacrificed efficiency in its bid to be 'prudential'. However, had the IMF monetized its gold reserves either through off market transactions or through outright sales in the market, it would have had reserves that were twice the current market value of its gold reserves. So not only is holding gold inefficient, but it is also not 'prudential' as a higher level of reserves implied by the sale of gold would be more 'prudential' than a lower level of reserves implied by gold holdings.

Additionally, the IMF claims that it holds on to its gold to meet contingencies. Given its much-touted concern about disrupting gold markets, this is a strange claim. One would assume that in the event of the occurrence of a 'contingency', the IMF would need to sell some of its gold holdings. The very occurrence of a contingency would imply the existence of some form of global macroeconomic stress. Any way one looks at it, selling huge quantities of gold in such a stressed global environment would be a bad idea.

In order to meet with any contingency that the IMF wants to guard against, the sensible defence would be the existence of a well-diversified portfolio of highly rated liquid assets, which can be liquidated at relatively short notice to generate resources. This portfolio could be built up from the income generated by the sale of gold proposed in this paper.

Finally, the IMF states that profits form the sale of any gold should be used, whenever feasible, to create an investment account from which only the income should be used.

We sympathise with this position and believe it to be prudent. However, we insist that the HIPC Initiative and the request for total debt cancellation is an exceptional circumstance and creates sufficient justification for a possible violation of this prudence principle especially given that the IMF would still have more than three quarters of its gold intact after having met its enhanced obligations under a program of total HIPC debt cancellation.

Alternatively, if the IMF sells a sufficient amount of gold, the investment income from the proceeds of the sale of gold would be enough to meet the funds' obligation for 100% HIPC debt cancellation. Under such a scenario, all the profits from the sale of gold could be invested in income generating assets.

The HIPC owe the World Bank a total of about $19.2 billion (NPV) of debt, of which the Bank has already cancelled or promised to cancel about $6.4 billion (NPV). Most of this debt is owed to the IDA.

The World Bank is using the HIPC trust fund operating under the trusteeship of the IDA as the vehicle for debt cancellation. The fund gets its resources through IBRD net income allocations and voluntary bilateral contributions. Neither the IBRD, nor the IDA have written off the HIPC loans from their books. Instead, as and when a HIPC debt falls due, equivalent funds are transferred from the HIPC trust onto their respective balance sheets.

To date, donor countries have committed about $2.5 billion and the IBRD about $1.5 billion to the trust. However, this is less than the amount needed even for the part of debt cancellation that the World Bank has already agreed to. In addition, about another $13 billion (NPV) must be mobilized to cancel all the debt owed to the World Bank by HIPCs.

Below we consider three possible sources of funding for World Bank debt cancellation.

Through transfers from its retained earnings

We suggest that the IBRD make an immediate transfer to the HIPC trust from its retained earnings.

We believe that the IBRD could transfer up to $10 billion from its retained earnings, which currently stand at $27 billion (total equity $37 billion). These include the past profits, which are yet unallocated or have already been allocated to the reserves. We have shown that the IBRD has the financial resources to make this transfer without jeopardizing or seriously impacting its operations in any way[24]. This transfer would take the IBRD's total reserves back to the 1997 level. The IBRD was active and successful at that point just as it is now.

1. Of this $10 billion, $5 billion can come from the unallocated portion of net income, which stood at $5.344 billion on the 30th of June 2003. This $5 billion would include the $240 million already allocated to the HIPC trust this year.
   
   
2. The rest of the $5 billion would come from the general reserve, which currently stands at more than $19 billion. Though this may raise some accounting issues, the transaction is similar in spirit and effect to the transfer from the unallocated part of the reserve. Hence, provided the proposal has political backing, the potentially problematic accounting issue can easily be circumvented.

Through transfers of its net income

We suggest that the IBRD publicly commit to an annual transfer of a fixed amount of its income to the HIPC trust for a fixed number of years.

We argue that the IBRD could transfer up to $800 million annually from its net income to the HIPC trust over the next 20 years. The IBRD's net income (profit) has been more than $1 billion annually for more than 15 years in a row. Moreover it has been increasing sharply over the last few years, up from just $1.2 billion in 1996 to $5.3 billion in 2003 We have shown that the IBRD has the financial resources to make this transfer without any serious impact on its financial soundness or normal operations[25].

Additionally, we recommend that the IBRD transfer a higher amount of income in years with good financial results.

Through transfers from provisions for loan losses

We suggest that the IBRD transfer excessive provisions out of its loan loss provision account to the HIPC Trust. The IBRD has built an accounting firewall so that the, accumulated monies in its loan loss provision accounts cannot be used to write off IDA shortfalls. However, this ignores the basic fungibility of funds.

We argue that because

• The Bank is a preferred creditor,
  
  
• The Bank has never written off any loan in its portfolio
  
  
• The Bank has excellent recovery rates for loans in non accrual status

it can afford to transfer at least $1 billion of resources from its accumulated provisions for loan losses, which currently stands at $4 billion. The IBRD's loan loss provisions are more than one and a half times its impaired loans.

We suggest that the best way for the IBRD to finance 100% HIPC debt cancellation is to use a judicious mix of the sources of funds discussed above.

Recommendation

We recommend that the IBRD should use

• $5 billion from its unallocated net income
  
  
• $2 billion from its general reserves
  
  
• $475 million of annual income pledged for 20 years with a NPV[26] of $5.9 billion

or another suitable combination of the sources of funds to write off 100% of the remaining $13 billion (NPV) of HIPC debt owed to it.

We further recommend that the commitments for the transfer of income that the IBRD makes for this purpose be additional to the transfers it already makes to the IDA and not in lieu of them.

The IBRD claims that its equity capital is leveraged at a ratio of about 5 through the issuance of AAA-rated debt. Therefore, its capacity to lend would be reduced by $5 for every $1 distributed to debt cancellation in respect of the concessional lenders' balance sheets.

We rebut this claim as inaccurate, irrelevant and misleading. The following points explain why:

• First, the Bank is actually leveraged at a much lower level than stated. Conservatively estimated, the rate at which the Bank's equity capital is leveraged, has declined steadily from 4.7 to 3.6 in the period 1998-2003. If the bank's usable callable capital[27] is included in the calculation, then the bank's leverage is actually even lesser than 1.

• Second, there is no prudential justification for capping the Bank's equity leverage[28] at 5. Historically, the Bank has had much higher equity leverage . In the late 1980's and the early 1990's the Bank's equity was leveraged almost 6 times. Given that the Bank was healthy and functioning in good shape through most of those times, the current leverage of 3.6 seems excessively prudent and inefficient.

In the private sector, for example, it is common for banks to have much higher levels of equity leverage. For example, Rabobank Nederland rated 'AAA' (just like the IBRD), has been leveraging its equity at a ratio between 10 to 12 times for many years. This ratio is at the lower end of the scale for large international banks, which typically operate at a even higher leverage.

• Third even if some capital is depleted to pay for debt cancellation, in order to keep the lending level the same, all the IBRD has to do is to increase the rate at which its equity is leveraged to retain its lending levels. For our recommended level of a transfer of $7 billion from the reserves, the equity leverage ratio would just rise to 4.7, which is well within the IBRD's stated level of 5.

• Fourth, it is misleading when the Bank says that its lending capacity would be reduced by $5 for every $1 appropriated towards debt cancellation. The Bank is currently lending only at 55% of its nearly $220 billion capacity. In fact, our debt cancellation proposal would just have the effect of reduce the Bank's total lending capacity only by 3.2% to about $213 billion. The Bank's lending capacity would actually be reduced only by $1 for every $1 appropriated to debt cancellation.

The Bank claims that it is likely that the write-off would result in a weaker equity capital position for the Bank and therefore an increased cost of lending to its borrowers. Debt cancellation, with substantially reduced borrowing, at higher cost, would have a serious impact on IBRD-eligible borrowers, which are home to 80% of the world's poorest people.

We disagree with the position of the Bank. We believe that the fundamental financial strength of the Bank derives from a number of factors of which capital reserves are only one. According to credit rating agencies such as S&P and Moody's and the Bank itself[29] its financial strength derives from:

• Shareholder support
  
• Consistent profitability
  
• Large liquidity
  
• Quality loan portfolio
  
• Prudent Financial Policy

To illustrate the point, we assume that the Bank decides to follow our recommendation and decides to write off $7 billion from its balance sheet. It then uses the amount for debt cancellation. This would mean that the equity of the IBRD would go back to the 2000-2001 level. The Bank was financially sound then and even at that time had one of the lowest borrowing costs in the financial markets.

We have shown, using the example of the African Development Bank, that changes to the Bank's financial situation induced by the recommended debt cancellation exercise, are not very significant to its market position as one of the lowest cost borrowers.[30] Some credit and bond market professionals who actually contribute the pricing decisions on the Bank's costs of borrowing have privately confirmed that the Banks borrowing costs would not rise by any significant level[31]. As we have already pointed out above, we think that the Bank is excessively prudent and financially inefficient.

If, for instance, the Bank were to double its current level of reserves, estimates show that its cost of borrowing would only go down by about three hundredths of a percent. Such an exercise would not be financially efficient as the opportunity cost of holding those reserves would be far higher than the small decrease in borrowing costs[32].

Any increase in interest costs would be minor (1-2% of current interest costs) and could easily be financed by a compensating reduction in the Bank's operating costs (through more efficient operations for instance). Even if this were not possible, the cost of increase would be spread across the Bank's borrowers and lie well within the range of frequent changes in market based lending rates.

The Bank has, for many years, been allocating a proportion of its net income to the IDA, a transaction identical in spirit and effect to what we have proposed[33].

Incidentally, some IBRD borrowers have been increasing Overseas Development aid (ODA) Budgets. India, for example, has recently said that it intends to double its current $700 million ODA budget and plans to add African nations its list of aid recipients.

The importance of political factors in Multilateral Development Bank AAA rating; the case of the African Development Bank and its relevance for IBRD

In various statements made over the years, the World Bank has often implied that debt cancellation poses a large threat to its financial soundness. It has implied that in the event that it made more contributions to the cancellation effort, its 'AAA' credit rating would be under serious threat.

We have already shown, in other sections of the report, that these claims are completely unfounded. We have used rigorous financial analysis and comparisons with the private sector to make the point.

In this section, using the example of the African Development Bank in the 1990's we have highlighted the primary importance of the unique political and financial structure of multilateral development banks such as the IBRD in determining the credit worthiness of these institutions.

Serious problems at the African Development Bank

In the early 1990's the AfDB was close to collapse. "Morale is as low as I have ever known it to be," said one insider. "The bank has lost direction, it is demoralized, and the entire lending machinery has come to a grinding halt."

The bank had also become an international embarrassment. Political infighting had paralysed the bank's administration and it could not extricate itself from a sizeable equity investment in a collapsing commercial-banking empire.

Percy Mistry, a former chief of staff at the World Bank was commissioned by the AfDB's Nordic shareholders in 1993 to put together a report on the AfDB's financial condition. "The bank kept lending very large dollops of hard money to countries which simply could not afford its terms," he said. "I have still not got a satisfactory explanation for why this has happened."

Between 1989 and 1993, according to the bank's annual reports, it lent a total of $3.47 billion at near-market interest rates to countries, which, under World Bank classification, qualified only for concessionary lending. In 1994 Babicar Ndiaye, the bank's former president, announced that loan defaults and arrears to the AfDB had topped $700 million."

Worried by the refusal of some of its big-country shareholders to advance more money for lending on, the governors of the African Development Bank commissioned a panel chaired by former World Bank vice president, David Knox, to write a report on the state of affairs at the bank.

The report was dire in its findings and it warned that if the bank was not strengthened, "it may end up destroying itself". His specific charges read like a litany of the worst faults associated with multilateral institutions. The bank was described as a top-heavy bureaucracy, riddled with political intrigue.

The bank also had many other problems such as a lame-duck president at odds with his board who was unlikely to be re-elected for a further term of office; a portfolio of non-performing loans; shareholders refusing to put in any more capital.

Worse, the bank's monitoring of loans was found to be haphazard. Mr Knox could not find in the bank's head office in Abidjan any files on specific projects. No wonder, then, that the bank's loan losses and arrears stood at around $700m. "As a matter of urgency the bank must put in place a comprehensive reporting system to monitor projects and assess the status of the portfolio," Mr Knox concluded.

Vote of Confidence in the African Development Bank

If the bank had been privately owned, there are little doubts that such serious problems would have led to serious castigation by investors and credit agencies in the form of a massive downgrade. However, despite such acute troubles, the bank got a vote of confidence from both investors and credit rating agencies. This was primarily because of the callable capital in its capital structure and its' standing as a multilateral development bank.

On August 9 1994, US rating agencies Standard & Poor's and Moody's both reaffirmed the triple-A credit rating of the AfDB's senior debts. According to Standard & Poor's, the "outlook on the bank's ratings remains stable". It reported that the "ratings reflected the AfDB's traditional conservative statutory and policy controls on the bank's leverage and liquidity as well as stricter financial and lending policies". Moody's said that the ratings "reflected further strengthening of the support of the bank's members which should be rearmed during the bank's fifth general capital increase". Speaking just before the decision, AfDB's new treasurer James Ranaivoson said investors' confidence in the bank had been little affected by the problems. "Investors know that if the senior debt is less than the callable capital of the high-rated countries they have nothing to worry about".

Uwe Bott, an analyst with Moody's in New York, said that the bank remained fundamentally sound. He points out that, although the quality of the loan portfolio had deteriorated, net income in I993 was still positive, at just under $72 million. "For there to be a capital call, the bank's assets must deteriorate sufficiently to absorb all net income," said Bott. "Then the bank must draw on its reserves to pay its creditors. Only when reserves are depleted must the bank draw on its capital base, and it is backed by a large pool of triple-A-rated capital. This could only happen after a number of years of the bank making losses."

Mild Castigation of the African Development Bank

However, the political infighting intensified especially over the issue of the election of a new president. Finally, in October 1995 Standard & Poor's stripped the African Development Bank of its AAA credit rating, and demoted it one notch to AA+. The agency cited political influences in the running of the bank for the downgrading.

"The downgrading reflects the increasing politicisation of the bank's corporate governance and management in recent years," said an S&P statement, "a development which has weakened its financial flexibility and which sets it apart from other multilateral development institutions in Standard & Poor's highest rating category."

Such banks enjoy a AAA credit rating because they are ultimately backed by the governments that own them. So S&P's move, in spite of such a guarantee, amounted to a fierce indictment of the chronic mess that the AFDB was in. It was paralysed by poor management and by a bloated bureaucracy.

However, this move was criticized and Charlie Berman, head of capital markets at Salomon Brothers in London, was one of the many who disagreed with S&P's actions. "In our research, Salomon has always been on record stating that S&P's action was premature," he said. "There are issues that the bank has had to address, but we don't think that they justify a downgrade."

It was only earlier in 2003, that S&P reinstated the AfDB's AAA credit rating after having kept it on AA+ for nearly eight years. However, it is important to note that Moody's never downgraded the bank and rated it Aaa all the way through.

Political pillars of strength and support

Even after Standard & Poor's stripped the AfDB of its coveted triple-A credit rating, the bank retained a presence in the international bond markets. Over the next two years in the secondary market, the spread fluctuated between 25-40 hundredths of a percent and when the bank made another US$ benchmark issue in 1997, it was priced just 23 hundredths of a percent over treasury.

More recently, the bank's most recent $1 billion five year issue made in August 2003 has been priced at 28 hundredths of a percent over US treasuries.

The AfDB's cost of borrowing is only about 10 hundredths of a percent higher (for a five year maturity) than the IBRD. Even at the height of its troubles in the 1990's this cost of borrowing was not significantly higher than this.

Another important case that highlights the importance of political factors is that of the German Landesbanks. Many of these banks are heavily undercapitalised (with capital adequacy ratios far below those required by the Basel Accord). However, a number of them still enjoy a AAA credit rating mainly because of the implicit guarantee of the government[34].

Conclusion

It is clear from the above discussion of the AfDB, that even under conditions where a private sector bank would lose the confidence of the market, the strength of the political and financial backing behind the multilateral development bank's such as the AfDB, the ADB and the IBRD ensures that the market continues to perceive them as fundamentally default free.

This discussion also illustrates that while financial performance is important, it is the political structure of the multilateral development banks that makes them financially sound.

The Millennium Development Goals

1 Eradicate extreme poverty and hunger
Halve, between 1990 and 2015, the proportion of people whose income is less than one dollar a day.
Halve, between 1990 and 2015, the proportion of people who suffer from hunger.

2 Achieve universal primary education
Ensure that, by 2015, children everywhere, girls and boys alike, will be able to complete a full course of primary schooling

3 Promote gender equality and empower women
Eliminate the gender disparity in primary and secondary education preferably by 2005 and to all levels of education no later than 2015

4 Reduce child mortality
Reduce by two-thirds, between 1990 and 2015, the under-five mortality rate

5 Improve maternal health
Reduce by three quarters, between 1990 and 2015, the maternal mortality ratio

6 Combat HIV/AIDS, malaria and other diseases
Have halted, and begun to reverse, the spread of HIV/AIDS
Have halted by 2015, and begun to reverse, the incidence of malaria and other major diseases

7 Ensure environmental sustainability
Integrate the principles of sustainable development into country policies and programmes and reverse the loss of environmental resources
Halve by 2015, the proportion of people without sustainable access to safe drinking water
By 2020, to have achieved a significant improvement in the lives of at least 100 million slum dwellers

8 Develop a global partnership for development

About the IMF

1. The IMF is an international organization of 184 member countries. It was established to promote international monetary cooperation, exchange stability, and orderly exchange arrangements; to foster economic growth and high levels of employment; and to provide temporary financial assistance to countries to help ease balance of payments adjustment.
   
   
2. Financially, the IMF is a co-operative of its 184 member countries, lending reserve assets (foreign exchange) to some of its members from resources subscribed by all of its members.

   IMF Resources

3. The Fund is mainly a pool of currency and reserve assets built up from the member's fully paid up capital in the form of subscription quotas. After the last quota increase in 2000, the total quotas amounted to SDR 211 billion (about $290 billion at the current exchange rate) of which, a quarter is paid in and the rest is callable. The paid in subscription is denominated in reserve assets, in gold prior to the second amendment to the articles in 1978 and since then in SDRs or "usable" currencies of its members as determined by the fund. The callable portion of the member's subscription is payable in its own currency. Consequently, a portion of the fund's reserves consist of unusable currencies i.e. currencies of financially weak member leaving its lending capacity at about two thirds of total quotas.
   
   
4. The Fund's pool of resources can also be augmented through additions to its precautionary balances. These balances comprise the Fund's reserves as well as resources that were set-aside in the special contingent account (SCA). These resources are not segregated from other resources of the Fund and can therefore finance the extension of Fund credit.
   
   
5. Reserves represent mainly net income of the Fund retained over the years, including from investments. Reserves are intended to meet losses of a capital character or administrative deficits (negative net income), and are categorized as "general" or "special." The main difference between the two reserves is that the Fund's Executive Board may distribute the general reserve but not the special reserve. Any distribution of the general reserve must be made to all members in proportion to their quotas.
   
   
6. Resources in the SCA are balances collected from debtor and creditor members through the operation of the burden-sharing mechanism and constitute one of the Fund's instruments for dealing with the persistence of overdue financial obligations to the Fund, and are refundable to the contributing members.
   
   
7. The Fund may also borrow from official and private sources, but has done so only from the former, reflecting mainly its nature as a cooperative intergovernmental institution. The Fund has in place the General Arrangements to Borrow (GAB) and the New Arrangements to Borrow (NAB). These arrangements supplement the Fund's quota resources when needed to forestall or cope with a threat to the stability of the international monetary system.
   
   
8. Under the GAB, which was originally established in 1962, the Fund may borrow up to SDR 18.5 billion from 11 industrial countries or their central banks, and from Saudi Arabia under an associated agreement. Under the NAB, which became operational in 1998, the Fund may borrow up to SDR 34 billion from 25 official lenders. The NAB is the facility of first recourse should the Fund need to borrow, and the limit of SDR 34 billion under the NAB also applies to the combined amounts borrowed under the GAB and NAB.

   Relevant IMF Accounting Procedures

9. The Fund's accounts are formally organized under three separate headings: the General Department, the Special Drawing Rights Department, and the administered accounts, where the last includes the PRGF and PRGF-HIPC Trusts. The General Department contains the General Resources Account (GRA), used for the Fund's main lending operations as well as it's borrowing, and the Special Disbursement Account (SDA). The Fund's various accounts are operated, recorded, and accounted for separately. Assets in one department or administered account may not be used to discharge the liabilities or to meet losses incurred in the administration of other accounts or departments, except as permitted by the Articles.
   
   
10. The administered accounts are the vehicle for the Fund's concessional lending and grants to low-income countries under the PRGF and PRGF-HIPC Trusts. The SDA is the vehicle for receiving and investing the proceeds from the sale of the Fund's gold in excess of SDR 35 per fine ounce, and the use of those resources for special purposes. In addition, the General Department contains an inactive account, the Investment Account, which may hold assets arising from gold profits, a transfer of currencies held in the GRA, or the investment income of the account.
    
    
11. The Fund's finances are similar to those of other financial institutions. A typical financial institution holds liquid assets and loan claims and securities among its assets, financed by its deposit (monetary) liabilities and capital resources. The Fund holds reserve assets (usable currencies, SDRs, gold) and credit outstanding to its members, and issues monetary liabilities (referred to as reserve tranche positions), and its capital includes the usable component of quota subscriptions. The Fund receives SDRs as called for under various provisions of the Articles. Though held as an asset, gold is not normally used by the Fund in its operations and transactions.
    

    IMF Income

12. The Fund earns income from charges on members' outstanding use of Fund. The rate of charge is determined as a proportion of the SDR interest rate that would achieve a target amount of net income for the incoming financial year. In other words, the rate of charge is set so as to generate revenues that more than cover the Fund's cost of funds, permitting the Fund to add to its precautionary balances and linking the rate of charge to market interest rates. The Fund also receives income from debtor members in the form of service charges, credit commitment fees, and special charges, though the amounts involved are relatively small.
    
    
13. The Fund has generally aimed at a modest annual increase in its reserves-3 percent in FY 1981-84, 5 percent in 1985-99, 3.9 percent in 2000, and 1.7 percent in 2001. Additions to SCA have been made under the burden sharing mechanism at 5 percent of reserves at the beginning of the financial year.
    
    
14. In deciding on the amount of income to be added to reserves, the Fund is guided by two principles: precautionary balances should cover outstanding credit to members in protracted arrears to the Fund, and such balances should also include a margin for the risk related to credit outstanding to members in good standing. Precautionary balances have, since FY 1993, covered the stock of outstanding Fund credit to members in protracted arrears. In recent Board discussions, it was agreed that the excess, called "free reserves," should be within a range of 3-5 percent of outstanding Fund credit in good standing. Free reserves rose to 7.0 perce