Money Has to Grow: How the Fund Was Financed

James Boughton
Published Date:
March 2012
  • ShareShare
Show Summary Details

As an official multilateral lending agency, the IMF has a daunting task to keep its finances in balance. Its mandate requires it to lend to countries in difficult financial straits after private sector lenders have pulled out and official bilateral lenders have become wary of the risks. In the 1990s, the rapid growth and increasing complexity and instability of international trade and capital flows raised the stakes immensely. The value of world trade more than doubled from the 1980s to the 1990s. The aggregate current account deficit of all deficit countries rose from $223 billion a year in the 1980s to $366 billion in the 1990s. The size and volatility of international financial flows grew much more rapidly. In response, IMF credit outstanding rose from $28 billion to $41 billion. Despite the size of new lending, the Fund was covering a smaller portion of its members’ financial needs. Raising the capital to keep this contribution from shrinking further provided a constant challenge throughout the decade.

The Financial Evolution of the IMF

The IMF began life in 1946 as a single fund: a pool of financial resources deposited by its member countries, mostly in the form of gold, U.S. dollars, and countries’ own currencies. Those nondollar currency balances were a major part of the IMF’s assets, even though they were—at the time—not traded or convertible internationally and therefore not usable for IMF lending. Over time, the number of “usable currencies” gradually grew. The financial assets of what eventually became the General Department of the IMF comprised gold, usable currencies, and other currencies. The First Amendment of the Articles of Agreement, adopted in 1969, enabled the IMF to allocate (i.e., issue) special drawing rights (SDRs) through a separate SDR Department, which then had its own balance sheet.1 Beginning in 1976, the IMF administered a series of trust funds to make loans to specified groups of countries, mostly low-income. The IMF sold a portion of its gold stock to finance those trust funds initially, and later borrowed from member countries, central banks, and other official bodies. By the 1990s, the IMF was no longer a single fund but a complex set of financial funds, each with its own balance sheet and operations.2

General Department

The IMF’s main balance sheet—that of its General Department, summarized in Table 15.1—has an unusual structure in that a portion of its assets consists of currencies that are not usable for operations. Each member country is required to deposit approximately a quarter of its assigned quota in the form of usable currencies or SDRs and the remainder in its own currency. Every three months, the IMF assesses the strength of the member’s balance of payments and of its reserves and determines on that basis whether to include the member’s currency in the Fund’s financial transactions plan, which determines the maximum amount of each currency that can be used in lending operations.3 Thus, the operational value of the Fund’s assets fluctuates with economic conditions.

Table 15.1.Balance Sheet of the General Department(Billions of SDRs)
December 31,

December 31,

Liquid cash assets42.696.253.6
Usable currencies40.992.451.6
SDR holdings0.92.51.5
SDA investments0.81.30.5
Loans and credits51.0114.663.6
GRA credits22.351.128.7
Other currency holdings27.163.035.9
SAF loans1.50.5(1.0)
Gold (book value)
Other assets0.00.30.3
Total assets98.7217.5118.8
Liquid liabilities25.554.829.3
Reserve tranche positions22.054.832.8
Borrowings (General Department)3.50.0(3.5)
Balance of quota subscriptions68.1155.587.3
SDA resources2.31.8(0.4)
Other liabilities0.40.50.2
Reserves, etc.
Ordinary reserves1.42.81.4
Deferred charges0.91.00.1
Total liabilities and net worth98.7217.5118.8
Gold at market price31.621.3(10.3)
Sources: IMF financial accounts and staff calculations.Note: SAF = Structural Adjustment Facility; SCA = Special Contingent Account; SDA = Special Disbursement Account. Details may not sum to totals, owing to rounding.
Sources: IMF financial accounts and staff calculations.Note: SAF = Structural Adjustment Facility; SCA = Special Contingent Account; SDA = Special Disbursement Account. Details may not sum to totals, owing to rounding.


At the beginning of the 1990s, the IMF held a total of $53.2 billion (SDR 40.9 billion) in 32 different currencies deemed usable for lending operations (Table 15.2), plus small amounts of 9 other currencies that were eligible to be received by the Fund from members in receipt of financial obligations.4 In volume, these currencies ranged from U.S. dollars ($14.4 billion) down to Botswana pula ($3.8 million). In addition, the General Department held $1.2 billion in SDRs that it could lend to members. Within the General Department, the Special Disbursement Account (SDA) held about $1.1 billion in short-term SDR-denominated securities, pending the approval and disbursement of loans to low-income member countries through the Structural Adjustment Facility (SAF). Altogether, these liquid cash assets—resources that could readily be lent out—amounted to $55.6 billion (Table 15.1).5

Table 15.2.Countries with Currencies in the Financial Transactions Plan, 1989 and 1999
December 31, 1989December 31, 1999
CountryHoldings (Millions of SDRs)Percentage of QuotaPercentage

of Total

CountryHoldings (Millions of SDRs)Percentage of QuotaPercentage

of Total

Brunei Darussalam115760.12
Korea, Rep. of285610.70
New Zealand422911.03New Zealand586660.63
Saudi Arabia2,739866.70
United Arab Emirates64320.16United Arab Emirates399650.43
United Kingdom4,9488012.10United Kingdom6,892647.46
United States11,0726227.08United States24,0566526.02
Total holdings40,88292,439
Number of countries3233
Sources: Annual Reports and International Financial Statistics.Note: Countries shown in italics joined the IMF during the 1990s. Details may not sum to totals, owing to rounding.
Sources: Annual Reports and International Financial Statistics.Note: Countries shown in italics joined the IMF during the 1990s. Details may not sum to totals, owing to rounding.

During the 1990s, the size of the Fund’s General Department more than doubled, as did the size of the resources in the financial transactions plan. By the end of the decade, liquid cash assets exceeded $132 billion. The number of usable currencies, however, was virtually unchanged. Although more than 40 countries qualified at some time during the decade, the list of eligible currencies never rose to more than 33. At a first approximation, therefore, the usable resources available to the Fund rose in parallel with the rise in Fund quotas, which remained the primary funding source for the department.

On a net basis, the Fund’s usable and uncommitted resources did not rise commensurately with gross resources until some of the major borrowers began repaying earlier loans in 1999 (Figure 15.1).6 Two causes stand out for the increased pressure on the balance sheet. First, the series of financial crises that engulfed much of the developing world from 1994 through the end of the decade led to sharp increases in Fund lending and a corresponding reduction in uncommitted funds. Second, the increase in IMF membership in the 1990s, to 182 countries from 152, applied further pressure, because most of those new members soon became borrowers. Brunei Darussalam, Slovenia, and Switzerland were creditor countries and contributed to the Fund’s usable resources, but almost all of the other new members were net debtors throughout much of the decade.

Figure 15.1.IMF Usable Resources, 1989–99

Billions of SDRs

Source: “Methodology Used in Reviews of the Fund’s Liquidity and Financing Needs,” EBS/97/60 (April 2, 1997), Table 2; “IMF’s Financial Resources and Liquidity Position 2000,” accessed at; and staff calculations.

Lending peaked in July 1998, right after the IMF made a large disbursement to the Russian Federation as part of an extended arrangement. The new members had an aggregate outstanding indebtedness of SDR 18.8 billion and a combined reserve tranche (creditor) position of just SDR 2.0 billion. At that point, the Fund’s net uncommitted usable resources totaled less than SDR 20 billion, well below the level at the start of the decade. The expansion of membership thus put substantial stress on the balance sheet. As discussed below, the IMF resorted to borrowing for the first time in several years. The 1999 quota increase and the first large repayments of earlier drawings then raised liquid resources sharply, to an all-time high at the end of 1999.


In addition to its currencies and SDRs, the IMF held more than 103 million ounces of gold in the 1990s—the third largest stock in the world after those of the United States and Germany. The Fund’s gold stock decreased 50 million ounces from its peak following the sale of 25 million ounces to member countries in the 1970s to finance the original Trust Fund and the restitution of another 25 million ounces to member countries at the same time. During the 1990s, as recounted in Chapter 13, much political discussion and internal debate took place over the possible sale of gold to help finance the Fund’s participation in the Heavily Indebted Poor Countries (HIPC) Initiative to help provide debt relief to poor countries. Surprisingly, in the end, the IMF did not sell any gold, and the physical stock remained constant from beginning to end (Figure 15.2). The Fund did, however, put some of this gold to use in the 1990s:

Figure 15.2.IMF Gold Stock, 1990–2000

Sources: IMF financial statements in Annual Reports; gold price from International Financial Statistics.

  • In 1992, the Fund “sold” a small amount of gold (21,396 ounces) to Cambodia and simultaneously accepted it back in partial settlement of the country’s overdue obligations to the General Department (Chapter 16).

  • In 1993, the IMF pledged to sell 3 million ounces of gold if needed to cover potential losses to creditor countries on loans from the Enhanced Structural Adjustment Facility (ESAF) Trust (also covered in Chapter 16). The need did not arise, and no gold was sold.

  • In 1996, the Fund agreed to sell up to 5 million ounces if needed to finance ESAF lending (Chapter 13). As before, the need did not arise, and no gold was sold.

  • The Fund “sold” 12.9 million ounces to Brazil and Mexico in a series of transactions from December 1999 through April 2000 and immediately accepted it back in repayment of outstanding obligations falling due. As explained in Chapter 13, these transactions—like the earlier one with Cambodia—raised the valuation of gold on the IMF’s balance sheet but did not affect the physical stock.7 In this case, the revaluation also restructured the Fund’s balance sheet to help finance the Poverty Reduction and Growth Facility-Heavily Indebted Poor Countries (PRGF-HIPC) Trust for the benefit of low-income countries.8

The market price of gold moved in a narrow range throughout most of the 1990s, usually between $330 and $400 per ounce from 1990 through 1997, and then fell to a low of $256 in 1999. That late decline meant the market value of the Fund’s holdings declined by about 25 percent from the beginning to the end of the decade. The book value in SDRs, however, remained stable until the increase at the end that resulted from the revaluations described above (see Figure 15.2).

Access Limits

The original text of the IMF Articles of Agreement specified that no member was to borrow more than its quota unless the Executive Board agreed to waive the limit in a particular case. Moreover, with the same exception, no member was to borrow more than 25 percent of its quota in any 12-month period.9 These access limits were intended to preserve the IMF’s resources to ensure that funds would be available to all countries that needed them and to ensure that no country would borrow more than it needed or could afford. The limits had an apparent numerical logic, but they were inherently arbitrary. No one knew at the time what the demands on Fund resources would be, nor how quotas would relate to each member’s financing requirements. By the 1960s, waivers were becoming routine. In the 1970s, the Fund began setting higher access limits based on experience, with the aim of offering adequate but sustainable amounts of financing to its member countries.10

At the outset of the 1990s, the IMF was applying a complex set of access limits that allowed countries to borrow up to a theoretical maximum of 590 percent of quota. Under the policy on “enlarged access to the Fund’s resources” (EAR) established in 1981 and last modified in 1985, the cumulative debt limit under the Fund’s tranche policies (stand-by and extended arrangements and outright purchases) was set at 440 percent of quota. In addition, the Fund had two “floating” facilities, which meant that drawings under those facilities did not count toward the general access limits. In principle, a country could borrow up to 105 percent of quota through the Compensatory and Contingency Financing Facility (CCFF) and another 45 percent through the Buffer Stock Financing Facility (BSFF) without reducing its access to other facilities.

The high ceilings fixed in the 1980s were intended to be temporary. They were set to recognize that quotas were unduly low and that countries had to borrow from official creditors to supplement quota-based resources.

The significance of the floating facilities was not just that use of the CCFF or the BSFF could raise potential access to Fund resources. A country borrowing through one or both of those windows could also borrow 25 percent of quota as a “first credit tranche” loan without having to submit to the Fund’s policy conditions. Because both of these floating facilities were designed to help countries cope with adverse circumstances beyond their control and thought to be temporary, this feature had a certain logical appeal. By the early 1990s, however, major creditors and Fund management firmly believed that borrowing countries needed strong policy conditionality to improve their prospects for economic growth and stability. Termination of floating was a long-suppressed but now openly espoused goal of several of the Fund’s main creditors.

When a general quota increase took effect in November 1992 (see below, under the Ninth Review), the Fund simultaneously reduced access limits in percentage terms. It eliminated the EAR and set new limits calculated so that the absolute amount a country could borrow would be the same as before the quota increase took effect; that is, the Fund replaced the temporary policy of lending borrowed money with a more sustainable regime of lending only its own resources. In percentage terms, the new limits became 300 percent of quota cumulatively and 68 percent in any 12-month period. At the same time, the Board modified the “floating” rule for the CCFF and the BSFF, so that those drawings would count the same as a first credit tranche loan for purposes of triggering upper-tranche conditionality. They remained separate, however, with respect to access limits.

The decision not to raise absolute access when the quota increase took effect may seem controversial and even odd. The principle, however, gained wide acceptance by Executive Directors, by borrowers as well as creditors. Everyone understood that the policy of enlarged access, financed by large-scale borrowing, could not be continued for long. The debate in the Fund in October 1992 was limited to the size of the cut in access limits and whether to discontinue the floating character of the special facilities.

At one extreme, G.A. Posthumus (the Netherlands) and Bernd Goos (Germany) proposed protecting the Fund’s liquidity by setting the annual access limit at 60 percent of quota and the cumulative limit at a level Posthumus called “the traditional long-term average” of 250 percent. That level would have reduced absolute access for many countries, and it therefore received little support from other Directors. Even among the major creditor countries, Muhammad Al-Jasser (Saudi Arabia) and Jean-Pierre Landau (France) called for limits of 70 and 300 percent to convey a stronger signal of support for countries with substantial financing needs. At the other extreme, a few Directors from developing countries argued for ceilings of 83 and 333 percent—levels calculated by the staff as ensuring that no member would suffer a decline in access.

The Board was badly split. A plurality, with 40 percent of the voting power, wanted limits in the higher range—at least 70 percent annually and 300 percent cumulatively. Those with about a third of the votes favored low limits close to those proposed by the Netherlands and Germany (60 percent and 250 percent). The rest were prepared to go along with the staff proposal for ceilings of 60 and 290 percent of quota. The Managing Director, Michel Camdessus, realized that reaching a consensus on a compromise was going to require bargaining with the other element under discussion, the floating character of the special facilities. He proposed a package in which floating would be eliminated (thus giving something to the hawks) and the access limits would be set at 68 percent annually and 300 percent cumulatively (thus coming close but not quite up to the ceilings preferred by the doves).

Camdessus’s proposal was met with strong resistance by the leading hawks, but it managed to pull a few Directors into a majority. After more discussion and a lunch break, the Secretary determined that the proposal carried with just 55 percent support. Consensus was obviously out of reach, so Camdessus closed the discussion and declared the package to be adopted.11

Experience with the new access limits was generally positive. The Fund was able to meet most countries’ financing demands through a combination of its own lending and the catalytic effect of mobilizing support from other official creditors. The main drawback was that many of the transition countries in Europe and Asia had such large financial needs associated with their efforts to establish market economies that the Fund’s self-imposed ceilings became a serious constraint. In April 1994, the Interim Committee noted this constraint with some alarm. “The transition to market economies by a large group of countries,” the communique warned, “is a task of historic proportions deserving full and concerted support of the international community.” The ministerial committee urged the IMF “to continue to play a central role in this process, including if needed through increased access to its own resources.”12

The Fund tried to respond to this call with new policies targeted specifically at the transition countries. For a time in 1994, it explored the idea of “cofinancing trust accounts,” which would link IMF support to trusts funded by donors and administered by the Fund. By late summer, though, management and the staff focused more on getting approval for a selective SDR allocation for countries that had joined the IMF following the last allocation, in 1981. In combination with an expansion of the Systemic Transformation Facility (STF) and perhaps a modest increase in access limits, selective SDR allocations would have gone a long way toward closing the anticipated financing gaps of most transition countries. As explained below, the selective allocation failed to garner enough support in the Interim Committee, and it was dropped for the time being. The STF was extended by several months, but without the proposed increase in access (see Chapter 5). As a fallback position, the committee recommended “a temporary increase in annual access limits from 68 percent to at least 85 percent of quota.”13

The Executive Board reacted with unusual alacrity to this request, acting on it just three weeks later. Even more remarkably, the hawk-dove split that had blocked consensus two years earlier was nowhere to be found. Early in the discussion, Stefan Schoenberg (Germany) stressed that “members should have confidence that the Fund would be in a position to respond on an appropriate scale in support of strong policies. We are prepared, therefore, to go substantially beyond 85 percent of quota in terms of the annual access limits; that is, we would consider appropriate an increase to 100 percent of quota.” The U.S. Director, Karin Lissakers, quickly supported that suggestion “on a temporary basis.” The Dutch Director, Posthumus, was still unhappy about the situation and noted correctly that the appeal to raise access limits “was a reflection more than anything else of the fact that donor countries were unprepared to provide substantial financing for the countries in transition.” Even he, however, was willing to go along with the emerging consensus. After a relatively brief discussion, the Board approved an increase to 100 percent, with only Saudi Arabia’s Al-Jasser abstaining.14

The access limits set in October 1994–100 percent of quota annually and 300 percent cumulatively—proved to be sustainable. Instead of treating them as temporary, the Fund kept those limits in effect throughout the rest of the decade, through the next quota increase, and until the doubling of the limits (to 200 percent annually and 600 percent cumulatively) in 2009.

Exceptional Circumstances

As frequently stressed at every review of access limits, these limits were just that: they were ceilings, not expectations of actual or average access. Throughout the 1990s, actual annual access in nonexceptional stand-by and extended arrangements averaged about 44 percent of quota, with little trend from beginning to end. The dramatic change in Fund actions after the October 1994 review was a shift toward more frequent invocation of “exceptional circumstances” to justify breaching the ceilings.

When the IMF established the Supplementary Financing Facility (SFF) in February 1979 as its first effort to supplement its own resources with borrowings for regular lending arrangements, it included a provision that became known as the “exceptional circumstances clause.” That clause allowed the Executive Board to approve arrangements and allow total cumulative indebtedness in excess of the established access limits. During the next 16 years, the Board invoked the exceptional circumstances clause only three times: in stand-by arrangements for Turkey and Zambia in 1978 and in an Extended Fund Facility (EFF) arrangement for Mexico in 1989.15 Then in February 1995, to enlarge the stand-by arrangement for Mexico after the U.S. authorities were unable to provide the anticipated amount of cofinancing (see Chapter 10), the Board again declared the circumstances exceptional. From that instance through December 1999, the Board coped with a wave of financial crises by invoking the clause 11 times for seven countries (Table 15.3). This extraordinary practice hit its peak in December 1997, with a stand-by arrangement for the Republic of Korea totaling almost 20 times the size of Korea’s quota. (On that occasion, the motivation was not only that Korea faced exceptionally dire circumstances, but also that its quota was unusually low for historical reasons.)

Table 15.3.Exceptional Access Under Fund Arrangements, 1990–99




of SDRs

of Quota
MexicoFebruary 1, 199518SBA12,0706888,758
Russian Fed.March 26, 199636EFF6,901160a
July 20, 1998b20EFF/CCFF15,3633568,344
ThailandAugust 20, 199734SBA2,9005052,500
IndonesiaNovember 5, 199736SBA7,338490
July 15, 1998c28SBA8,338557
August 25, 1998d26EFF4,669312
March 25, 1999c19EFF5,3832597,555e
Korea, Rep. ofDecember 4, 199736SBA/SRF15,5001,93814,413
BrazilDecember 2, 199836SBA/SRF13,02560010,474
TurkeyDecember 22, 199936SBA2,89230011,739f
Sources: “Review of Access Policy in the Credit Tranches and under the Extended Fund Facility—Background Paper,” EBS/01/134 (August 9, 2001), Table 13, accessed at; and IMF financial accounts.Note: CCFF = Compensatory and Contingency Financing Facility; EFF = Extended Fund Facility; SBA = Stand-by arrangement; SRF = Supplemental Reserve Facility.

Exceptional access was required only for the annual limit.

Augmentation and extension of the existing arrangement.


New arrangement to replace the existing one.

Cumulative amount drawn under the SBA and EFF.

The arrangement was augmented in December 2000 and again in May 2001, for a total approved amount of SDR 15,038 million. (A portion of the augmentation was under the SRF.) It was canceled in February 2002 and replaced by a new arrangement.

Sources: “Review of Access Policy in the Credit Tranches and under the Extended Fund Facility—Background Paper,” EBS/01/134 (August 9, 2001), Table 13, accessed at; and IMF financial accounts.Note: CCFF = Compensatory and Contingency Financing Facility; EFF = Extended Fund Facility; SBA = Stand-by arrangement; SRF = Supplemental Reserve Facility.

Exceptional access was required only for the annual limit.

Augmentation and extension of the existing arrangement.


New arrangement to replace the existing one.

Cumulative amount drawn under the SBA and EFF.

The arrangement was augmented in December 2000 and again in May 2001, for a total approved amount of SDR 15,038 million. (A portion of the augmentation was under the SRF.) It was canceled in February 2002 and replaced by a new arrangement.

The Source of Reserves: Income over Expenses

Despite the decade’s many challenges, the IMF managed to generate a steady flow of net income in the 1990s (Table 15.4). That income enabled the institution to build up its reserves as a cushion against the risks posed by its loan portfolio, which was heavily concentrated in large loans to a few countries with spotty economic and financial track records.

Table 15.4.Income Statement of the IMF General Department, Financial Years 1991–2000(Millions of SDRs, except as noted)
1991199219931994199519961997199819992000Average, 1991–2000a
SDRsMillions of

Operational Income
Interest and charges on GRA lendingb1,858.71,765.71,607.21,321.01,361.41,771.31,605.91,983.42,787.62,548.21,861.02,614
Interest on SDR holdings70.956.9217.4299.5216.440.357.637.469.5123.3118.9167
Interest on SAF loans8.
Income on SDA investments74.762.539.723.99.14.900030.124.534
SDA “profit” on gold transactions2,226.0
Net effect of change in accounting method for pension liabilities268.3
Gross Income2,012.51,893.41,872.81,653.11,595.01,824.01,669.62,025.32,860.35,198.02,025.52,845
Operational Expenses
Net remuneration to creditor countries1,140.5984.21,013.4838.5861.41,095.41,101.01,390.01,843.91,768.01,203.61,691
Interest paid on borrowings317.9286.6222.1147.2127.662.00078.859.0162.6228
Allocation to the SCA212.1229.7255.3242.6215.4263.5151.998.5106.7128.590.467
Administrative Expenses
Other administrative expenses32.136.142.849.345.347.454.070.478.186.154.276
Total Expenses1,871.91,748.01,770.51,565.51,515.21,745.91,600.41,857.02,421.52,403.91,850.02,598
Net Income140.6145.4102.387.679.778.069.2168.4438.82,794.2175.5247
Source: Annual Reports.Note:—= Not applicable.

Excludes gold transactions and accounting changes in financial year 2000.

Includes a small amount of miscellaneous income.

Source: Annual Reports.Note:—= Not applicable.

Excludes gold transactions and accounting changes in financial year 2000.

Includes a small amount of miscellaneous income.

The basic income model in place throughout the 1990s relied heavily but not exclusively on steady income from outstanding loans. For each financial year, the IMF aimed to generate net income by setting the interest rate it charged on loans at a level calculated to produce a target income, based on projected demand for loans. The Fund adjusted the rate of charge periodically during the year to stay on target. The target was set to keep the reserves of the General Department rising at a rate deemed sufficient to safeguard the Fund’s resources. Throughout the decade, the annual target income level was fixed at 5 percent of initial reserves. In addition, under the “burden sharing” mechanism the Fund devised to cover the risk associated with payments arrears (see Chapter 16), the rate of charge was set at a margin above the “basic rate” each year to fund a pair of Special Contingent Accounts (SCAs).

Through financial year 1993 (FY93),16 the Fund kept the basic rate of charge—the interest rate on most loans from the General Resources Account (GRA), before adjustments for burden sharing—slightly below the SDR interest rate (a weighted average of yields on top-quality short-term securities in the five countries with currencies in the SDR basket). In June 1993, as the Executive Board considered its policies for FY94, it confronted three forces that affected the Fund’s income adversely: a decline in the SDR interest rate, a troubling persistence in overdue financial obligations by several members, and a worsening risk profile associated with the prospect of large lending to a few countries. For the sake of prudence, the Board agreed to raise the basic rate of charge to about 111 percent of the SDR rate. The relationship between the two rates stayed near that level for the rest of the decade (Figure 15.3).

Figure 15.3.SDR Interest Rate and Rates of Charge and Remuneration, 1990–2000

Percent per year

Sources: Annual Reports and staff calculations.

Aside from small amounts of income from the SDR holdings of the GRA and investments in the SDA (pending the use of those funds for SAF loans), in most years the accrual of charges on GRA lending was the only major item on the income side of the ledger. The exception was FY2000, when the Fund had a large bookkeeping entry for the indirect transfer of gold from the GRA to the SDA, as described in Chapter 13, and an entry for a change in the method of accounting for future payout of pensions to retired staff.

The contribution of creditor countries to IMF net income came from the difference between (a) the interest that would have accrued at the SDR interest rate on the full amount of each creditor’s reserve position in the Fund and (b) the actual net remuneration paid by the Fund. That difference had two components.

First, a portion of each creditor’s reserve position was set aside as “unremunerated.” For each member country in the IMF in April 1978—when the Second Amendment took effect—the unremunerated position was equal to 25 percent of the quota in effect at that time. Because quotas were raised periodically after that date, this fixed amount correspondingly shrank as a percentage of current quotas.17 On average at the end of the 1990s, after the Eleventh General Review took effect, the unremunerated portion was just 3.8 percent of quota.18

Second, the adjusted rate of remuneration was set below the SDR interest rate. Throughout the 1990s, the basic remuneration rate was equal to the SDR interest rate. As with the rate of charge, however, the actual remuneration rate was adjusted (reduced) each year to fund the two SCAs as protection against potential losses (including both principal and interest charges) on nonperforming loans (see Figure 15.3).

For an illustration of these relationships, consider FY99, the penultimate year shown in Table 15.4. Borrowers from the various facilities of the General Department paid a total of SDR 2,791 million ($3.8 billion) in interest and fees. On the expense side, the Fund paid SDR 1,844 million ($2.5 billion) in remuneration to creditors. If the Fund had paid remuneration at the SDR interest rate on 100 percent of reserve positions, remuneration would have amounted to approximately SDR 2,156 million ($2.9 billion). The net implicit contribution from creditors therefore was about SDR 312 million ($0.4 billion), or just 11 percent of the direct payments from indebted countries.

The anomaly of this situation was readily apparent. Although every country benefited from the existence of an institution that helped to stabilize and provide analytical information about the international financial system, borrowing countries, most of which were less economically developed than the creditor countries, covered the great bulk of the expense.

The original conception of the IMF as a financial institution was that each member would contribute to the cost of running it in proportion to the member’s quota. This equitable distribution applied automatically at the outset because no remuneration was paid on the Fund’s holdings of currencies or gold, and every member was required to replenish its gold tranche contribution frequently (that is, repay any drawings on the gold tranche, which was the predecessor of the reserve tranche). By lending or investing these currencies, the Fund would generate income to cover its administrative expenses.19 Although the actual distribution of the burden differed from the distribution of quotas because the costs differed between debtors and creditors, the discrepancies were not usually large enough nor persistent enough to violate the underlying philosophy.

After the 1978 amendments, these principles no longer applied. Because countries were not required to repay reserve tranche drawings, a member could escape all costs by following Polonius’s dictum to “neither a borrower nor a lender be.” As long as the Fund held a currency in an amount exactly equal to the member’s quota, the member would incur no borrowing cost. With no reserve position, it also would have no unremunerated balance.20 Moreover, with remuneration paid on the lion’s share of reserve positions, creditor countries would pay a relatively small share of the cost. The IMF had become uncomfortably similar to a commercial bank, in that it depended on interest income from its loans to cover its costs.

When the Executive Board began considering alternative financing structures in 1994, it focused particularly on an idea awkwardly named the “uniform adjustable norm.”21 Under this proposal, every member country, including those with outstanding drawings, would be required to maintain a reserve tranche position at least equal to a specified percentage of quota, on which remuneration would not be paid. That percentage “norm” would be equal (“uniform”) for all members. With all members sharing this burden in proportion to quota, the role of interest charges in covering operating expenses could be reduced.

After a round of preliminary discussions in the Executive Board, the Interim Committee endorsed the general direction of the financial reform effort and asked the Fund to “accelerate its consideration” of the issue “with a view to ensuring a more effective and equitable mechanism” for financing its operations. The Executive Board held a second series of meetings, which resulted in a consensus view that the Fund’s operating expenses should be covered by the membership in proportion to quota. The “uniform adjustable norm” proposal also could have simplified the whole system of charges and remuneration by setting both interest rates permanently equal to the SDR interest rate. Unfortunately, doing so would require an amendment to the Articles of Agreement so that the Executive Board could specify an unremunerated percentage of each member’s quota instead of an unremunerated percentage of its reserve tranche position.22

This reform proposal was broadly accepted by Executive Directors as a principle for further discussion.23 Within a few weeks, however, the approval of an extraordinarily large stand-by arrangement for Mexico upset the consensus. Over the next several months, the focus of the Board’s deliberations shifted to the question of whether the Fund should impose a surcharge on large-scale borrowing. That idea, suggested by the U.S. authorities and supported by some other creditor countries, met with strong resistance from other members.24 In the process, the competing idea of shifting the burden more equitably to all members quietly disappeared. The lending-dependent income model stayed in place until a temporary but quite precipitous drop in demand for Fund loans in 2005–08 finally forced revisions in it.

Administrative expenses more than doubled in the course of the 1990s, in two stages. In the early part of the decade, the IMF absorbed a rapid influx of new members, many of which had large borrowing requirements and major complex structural issues to be analyzed. Responding to those demands required a sizeable increase in the staff and in the amount of staff travel. Then toward the end of the decade, the Fund was asked by its members to take on several new tasks: assisting in the development of poverty-reduction strategies for low-income countries, devising debt-reduction programs for heavily indebted poor countries (the HIPC Initiative), and undertaking new surveillance activities such as the Financial Sector Assessment Program and the preparation of reports on countries’ observance of internationally recognized standards and codes. These activities gave rise to a second large increase in administrative expenses in 1998–99.

For the first few years in this decade, the IMF was incurring significant costs in the form of interest payments on earlier borrowing. Through FY92, the Fund offset much of that cost by charging higher interest rates on the high-access arrangements that were partially financed with borrowed money. After the quota increases under the Ninth General Review took effect, the Fund ended that practice and finished repaying the loans from the 1980s. For two financial years (FY97 and FY98), the Fund was debt free. The activation of the General Arrangements to Borrow for lending to Russia in July 1998 marked the temporary resumption of borrowing by the Fund.

SDR Department

Each participant [in theSDR Department] undertakes to collaboratewith the Fund and with other participants … with the objective of making the special drawing right the principal reserve asset in the international monetary system.

Article XXII of the IMF Articles of Agreement

Adopted April 1, 1978

By the end of the 1980s, it seemed that the SDR was an asset whose time had passed.25 Designed in the late 1960s when a global shortage of reserve assets was a major concern, the role of the special drawing right was no longer so obvious in a world of multiple reserve currencies, floating exchange rates, and cooperative swap arrangements among central banks. If the major countries could create liquidity at will, and if the dominant concern therefore was inflation rather than a shortage of liquid assets, then of what use was the SDR? Reflecting this view, the international community declined to create any additional SDRs after 1981. Over the next two decades, the share of SDRs in world nongold official reserves fell from 6.5 to 1.25 percent.26 The objective—enshrined in the Articles of Agreement in 1978—of making the SDR the principal reserve asset was effectively abandoned.

The Balance Sheet

The declining aggregate role of the SDR did not mean that it was of small importance. For many countries, the availability of SDRs as a supplement to foreign exchange served as a lifeline to international markets for goods and services. Receiving an SDR allocation gave them reserve balances that they could either hold as insurance against future deficits or draw upon to meet excess demand for imports or other payments needs. SDR holdings also helped countries pay quota increases or meet other obligations to the IMF. At one time or another during the 1990s, nearly three-quarters (132 out of 182) of the Fund’s member countries were net users (i.e., net borrowers) of the SDRs that had been allocated to them.

Table 15.5 summarizes the balance sheet of the Fund’s SDR Department around the beginning and end of the 1990s.27 No SDRs were allocated during this decade, so the total stock outstanding was constant at SDR 21,433.33 million. The main asset of the department was the stock of outstanding claims on net users of SDRs, which rose from SDR 6.5 billion at the end of FY90 to SDR 9.3 billion 10 years later. Most of that increase resulted from the payment of quota subscriptions in 1999 when the increases under the Eleventh General Review took effect. Consequently, the main counterpart to that increase on the liability side was an increase in SDR holdings by the GRA.28

Table 15.5.Balance Sheet of the SDR Department(Millions of SDRs)
April 30,

April 30,

Net usage of SDRs (allocations minus holdings, for net users)6,494.19,344.3
Charges receivable481.5221.2
Overdue assessments and charges43.1105.6
Total assets7,018.79,671.1
Holdings by participants in excess of allocations5,891.06,052.3
Holdings by the IMF General Resources Account628.52,723.9
Holdings by prescribed holders19.3673.2
Interest payable479.9221.8
Total liabilities7,018.79,671.1
Source: Annual Reports.Note: Details may not add to totals, owing to rounding.
Source: Annual Reports.Note: Details may not add to totals, owing to rounding.

Evolution of the SDR as a Financial Asset

Since 1974, the SDR has been defined as an asset equivalent in value to a basket of currencies. The initial basket comprised currencies of the 16 countries that accounted for at least 1 percent of world trade. When that proved unwieldy, the basket was simplified in 1981 to include only the five currencies in widest international use: the U.S. dollar, the deutsche mark, the French franc, the Japanese yen, and the pound sterling. From that point on, the Fund agreed to examine and, if desired, revise the composition of the SDR at five-year intervals. Effective January 1, 1986, 1991, and 1996, the selection of currencies was left alone, but the amounts of each currency in the basket were revised to keep the weights in line with each one’s importance in world trade and in monetary reserves (Table 15.6). The creation of the euro in 1999, replacing the mark and the franc and several other European currencies, occasioned a further revision in weights and a reduction to four currencies.29

Table 15.6.Composition of the SDR, 1969–2000(Local Currency Units)
July 1969–June 1974July 1974–June 1978July 1978–December 19801981–851986–901991–951996–981999–2000





















Gold (grams)0.888671
U.S. dollars0.3300.40000.3300.4000.420.5400.420.45200.400.57200.390.58200.390.5821
Deutsche marks0.1250.38000.1250.3200.190.4600.190.52700.210.45300.210.4460
Japanese yen0.07526.00000.07521.0000.1334.0000.1533.40000.1731.80000.1827.20000.1827.2000
French francs0.0750.44000.0750.4200.130.7400.121.02000.110.80000.110.8130
Pounds sterling0.0900.04500.0750.0500.130.0710.120.08930.110.08120.110.10500.110.1050
Canadian dollars0.0600.07100.0500.070
Italian lire0.06047.00000.05052.000
Netherlands guilders0.0450.14000.0500.140
Belgian francs0.0351.60000.0401.600
Swedish kronor0.0250.13000.0200.110
Australian dollars0.0150.01200.0150.017
Danish kroner0.0150.1100
Norwegian kroner0.0150.09900.0150.100
Spanish pesetas0.0151.10000.0151.500
Austrian shillings0.0100.22000.0150.280
South African rand0.0100.0082
Saudi Arabian riyals0.0300.130
Iranian rials0.0201.700
Source: International Financial Statistics.
Source: International Financial Statistics.

A little-understood characteristic of the SDR is that it is a variable-weight basket. That is, it is the sum of fixed amounts of each of its component currencies. As exchange rates change over time, the weights in the current valuation of the SDR also change. During the 1990s, for example, the weight of the Japanese yen ranged from a high of 24.1 percent in April 1995, when the yen was at its strongest, to a low of 14.2 percent in August 1998 (Figure 15.4). The SDR thus has a built-in “strong currency” bias because the weight of any appreciating currency automatically rises.

Figure 15.4.Weights in the Valuation of the SDR, 1981–99

(Valued at monthly average exchange rates)

Sources: Table 15.6 for currency composition of the SDR; International Financial Statistics for exchange rates; and author’s calculations.

aIn 1999, the deutsche mark and French franc were replaced by the euro.

Would There Be a Future for the SDR?

Although every proposal for allocating SDRs proved controversial, most countries found their allocations useful for one or more reasons. For low-income and other developing countries, SDRs were a relatively inexpensive source of official reserves and an inexpensive and reliable way to borrow foreign exchange (by drawing down their allocated balance). For the more advanced economies, having a stock of SDRs provided a buffer for the efficient management of reserves.30 By settling official obligations in SDRs, those countries were able to retain foreign exchange and manage the timing of its eventual disposition. By accepting SDRs in settlements from trading partners, large industrial countries were able to accumulate reserves in a stable and balanced form that was less subject to exchange rate fluctuations than any individual currency. Consequently, throughout the 1990s both the dominant holders and the main net users of SDRs included the major industrial countries (Table 15.7).31

Table 15.7.Leading Holders and Users of SDRs, 1990–99(Decade Averages)
Amount Held

(Millions of

Amount Held

(Percentage of

Percentage of

Total Stock

A.Leading holders (holding at least 1 percent of total stock)
1.United States7,261.9148.233.933.9
7.United Kingdom470.924.62.258.6
10.Saudi Arabia338.0172.91.663.9
B.Leading users (using at least 1 percent of total usage)
1.United Kingdom1,482.077.518.418.4
9.South Africa183.983.52.355.4
13.New Zealand140.599.41.762.9
22.Iran, Islamic Rep. of109.444.81.475.3
Source: IMF financial accounts.Note: Countries shown in italics in Part A were net users, on average.
Source: IMF financial accounts.Note: Countries shown in italics in Part A were net users, on average.

Controversy over proposals for new allocations arose for two reasons. First, if countries with ample foreign exchange reserves received an additional allocation, the temptation to borrow against the allocation or reduce other forms of reserves could add to aggregate demand and to inflationary pressures. That argument was the reason most often advanced by opponents. Advocates of allocations typically responded that the proposed stock of SDRs would be so small relative to total outstanding reserves that any such effects would be practically unnoticeable.32 The second issue, usually expressed less directly, was a concern about the unconditional extension of liquidity. Although the advanced economies no longer needed a supplementary reserve asset, for many developing and transition countries the acquisition of adequate reserves remained expensive. They could attempt to accumulate reserves by running external payments surpluses, but that strategy would run counter to their development needs. Alternatively, they could borrow from the IMF, conditional on the implementation of acceptable economic policies. An SDR allocation would open up a third avenue, without the conditionality and without the need to repay or periodically renegotiate the loan. The prevailing view among major creditor countries in the 1990s was that conditional lending by the IMF was preferred over the indiscriminate extension of credit lines to all members.33

Even though only a relatively small number of countries held these concerns, they held them strongly enough that no realistic possibility of a general allocation existed in this period. A general allocation could be undertaken only with the support of 85 percent of the voting power in the Board of Governors, but that did not deter Michel Camdessus from trying.

In April 1993, having been rebuffed repeatedly in previous efforts, the Managing Director made as strong a case as he could muster for a general allocation. He pointed out that more than 40 percent of IMF member countries (67 out of 162) held foreign exchange reserves in amounts below the customary notion of a minimum acceptable level (the equivalent of 10 weeks’ worth of imports). No SDRs had ever been allocated to the 37 countries that had joined the Fund after the last allocation in January 1981. The transition countries that had recently joined the Fund faced the particularly daunting task of trying to build reserves from scratch at the same time that they were trying to recover from devastating declines in output.

Camdessus calculated that an allocation of SDR 36 billion would be required just to restore the stock of SDRs to its average level in relation to global reserves over the preceding 20 years. “What meaningful interpretation,” he asked, “could be given to the obligation accepted by the Fund’s member countries, of making the SDR the principal reserve asset of the international monetary system, if we were not prepared to take a step in that direction in such pressing circumstances?”34

To strengthen the pertinence of his proposal to current circumstances, Camdessus revived a proposition, made on several occasions in the preceding decade, to redistribute part of the newly allocated SDRs from industrial to developing countries. In Camdessus’s formulation, the IMF would establish a new administered account. Countries with no need for additional SDRs could transfer them to this account, and the Fund could lend them as a supplement to stand-by or similar conditional arrangements. He suggested that the donors to this scheme should accept the risk associated with this lending: “the risk they encourage the Fund to take.”35

The Managing Director’s statement was an unusually direct challenge to the IMF’s major shareholders, several of whom he knew would be hostile to it. Indeed, almost at the outset of the Executive Board’s discussion of the proposal, Thomas C. Dawson II (United States) averred that he “remained unconvinced that there was a global need for additional liquidity at the present stage.” The United States had more than enough voting power to block any SDR allocation, essentially dooming the plan, at least for the moment. Within the Group of Seven (G7) major industrial countries, France, Italy, and Japan were open to the Managing Director’s proposal, but the others—Canada, Germany, and the United Kingdom, in addition to the United States—were not.36

With the G7 split, Camdessus pressed ahead. Shortly after the April 1993 Board meeting, the Interim Committee asked for further work on the rationale for, and implications of, an allocation of SDRs. Despite the reluctance of some to reanimate the SDR, the idea was not going to die. A pair of linked and disturbing facts—that Russia and other transition countries faced a severe shortage of foreign exchange reserves and that these countries had never received an allocation of SDRs, which could readily alleviate the problem—continued to haunt the corridors of the IMF.

Over the next year, the old redistribution scheme gave way to a new idea, that the Fund might allocate SDRs selectively, to help the recent member countries catch up. Dawson raised this possibility in April 1993 and asked the staff to try “to discover a mechanism compatible with the Articles [of Agreement] that would allow for a special purpose or separate allocation of SDRs to those countries that did not benefit from the allocation[s] earlier.” Unfortunately, the Articles did not allow for such a mechanism. Allocating SDRs selectively would require an amendment, and that would take both time and political capital.37

The resolution of this issue was supposed to be realized at the IMF/World Bank Annual Meetings in Madrid, Spain, in October 1994. These meetings were held somewhere other than Washington only once every three years, and the occasion usually raised expectations for the emergence of major agreements. The Madrid meetings were especially notable because they would be used to commemorate the fiftieth anniversary of the Bretton Woods conference at which the two institutions—and the postwar international financial system—were founded. Everyone wanted an agreement on allocating SDRs as the centerpiece of the celebration.

The staff, management, and the Executive Board labored for months to craft a proposal that would make a meaningful dent in the reserve shortages in the new member countries and preserve the role of the SDR as a stabilizing influence in the system. An acceptable compromise proved hard to devise. Developing countries and their allies insisted that a “global need to supplement existing reserve assets” existed and should be met by a general allocation of SDRs for all member countries. A blocking minority of major creditors insisted that any allocation should be targeted at the post-1981 members and should be based on an amendment to the Articles without any finding of global need. As the date of the Madrid meetings approached, the gulf between these two positions had not been bridged.

In the end, the report of the Executive Board to the Interim Committee presented four distinct proposals and asked the Governors on the committee to choose among them.38 The Managing Director proposed an allocation of SDR 36 billion, the same figure he had called for a year and a half earlier. As a compromise, Camdessus was asking for an initial allocation of SDR 20 billion based on a finding of global need, to be followed by an amendment allowing for selective allocation of SDR 16 billion. As a further concession to creditors, he proposed reintroducing the “reconstitution” requirement, which had been dropped in 1981. That provision required each country that drew on its SDR allocation to restore (“reconstitute”) its balance to a specified level within five years. The elimination (“abrogation”) of that requirement had enabled many countries to use all or most of their allocations as a permanent source of financing for international payments.39 Reintroducing the requirement would help restore the original concept of the SDR as a source of official reserve holdings.

Huw Evans (United Kingdom) and Karin Lissakers (United States) tabled a second proposal. Their proposal eschewed any general allocation on the grounds that “a consensus does not exist” for a declaration of global need. Instead, it called only for a “special one-time allocation” of no more than SDR 16 billion based on an amendment of the Articles. As a third proposal, two Directors—Marc-Antoine Autheman (France) and Jarle Bergo (Norway)—agreed with the Evans-Lissakers conception but conditioned their support on the allocation being at least SDR 22 billion. The fourth proposal—presented by the developing-country Directors who constituted the informal “Group of Eleven” on the Board—aimed to close the gap between the Managing Director and the leading shareholders. It called for a general allocation of SDR 14.5 billion, followed by a selective post-amendment allocation of 16 billion. The Group of Eleven also signaled their willingness to accept a reintroduction of the reconstitution requirement.

Hopes for a compromise dimmed in the run-up to the Madrid meetings. Officials from the G7, including Hans Tietmeyer (Germany’s central bank governor), Lloyd Bentsen (secretary of the U.S. Treasury), and Kenneth Clarke (chancellor of the exchequer in the United Kingdom), met with reporters to say they saw no need for a general allocation. Developing-country officials, mostly under the cover of anonymity, responded that they wanted a general allocation and predicted a “battle royal” in the Interim Committee. Meanwhile, Camdessus continued to go public with the case for a two-stage allocation of SDR 36 billion, effectively siding with developing countries in the prefight sparring.

On Saturday, October 1, 1994, the G7 finance ministers and central bank governors met, as was their custom on the day before the meeting of the full Interim Committee. Although they did not issue a formal communiqué afterward, they took a strong public stand against the proposals backed by Camdessus and the developing countries. Addressing reporters after the meeting, Bentsen stated flatly, “We will not accept a general allocation of SDRs.”40 Even so, the matter remained unsettled. The G7 had a keen interest in getting approval for an amendment that would enable the Fund to supply reserves to the transition countries. Without the Fund’s financial support, pressure would grow on G7 countries to fund the transition process themselves. Whether they would back away from the line they had drawn in the sand remained to be seen.

The debate at the Interim Committee meeting on Sunday turned out to be as acrimonious as had been predicted and feared.41 Formidable figures led the two sides: Manmohan Singh, finance minister (and later prime minister) for India, the chief advocate for a general allocation; and Tietmeyer, joined by Bentsen and Clarke, speaking forcefully against and trying to get the committee to endorse the G7 position. Although everyone was prepared to accept a selective allocation following the requisite amendment of the Articles, neither side would agree to it except as part of a package that included acceptance of their position on general allocations. Those positions were irreconcilable. Moreover, neither Singh nor Camdessus would consent to allowing the G7 to dictate terms on an issue this crucial for developing countries and for the Fund. For their part, G7 officials refused to believe that the usually fractious and fragmented developing countries could subvert their collective will.

The usual press conference by the Managing Director and the chairman of the Interim Committee (Philippe Maystadt, the Belgian finance minister) was scheduled to take place in late afternoon, when the committee meeting should have ended. Instead, the committee continued to meet through the evening in an effort to break the deadlock. Only as the hour hand clicked toward midnight did Maystadt finally concede that no compromise was possible. The meeting adjourned, but it was already October 3 when the exhausted pair, Camdessus and Maystadt, dragged themselves onto the podium before an impatient press corps.

The questions at the midnight press conference ranged from incredulous to openly hostile. Because Camdessus had not embraced the G7 proposal, the first questioner wanted to know, “whom do you think you represent as Managing Director?” Another noted that some G7 officials were accusing Camdessus of “wildly exceeding [his] mandate as Fund Managing Director and engaging in a partisan effort on behalf of developing countries.” A third reporter wanted to know if Camdessus was going to resign as a result of the impasse. Camdessus gave the only answer that he sensibly could: “I am not here to support a given majority. I am here to serve the Fund.” And he vowed to keep fighting for a general allocation.42

The Fourth Amendment to the Articles of Agreement

Camdessus kept fighting in virtually a one-man crusade to restore the systemic role of the SDR. In March 1995, as part of a broad strategy to raise resources following the unprecedented size of the $17.8 billion stand-by arrangement with Mexico, he asked Executive Directors to reconsider the issue of a sizeable SDR allocation. Elaborating on that request a few weeks later, he noted that the Mexican peso crisis was resulting in a loss of access to international capital markets by many developing countries. A fresh allocation of SDRs would help ameliorate the effects.43

Although the Executive Directors’ positions endured essentially unchanged from a year earlier, the staff circulated a draft amendment to the Articles for consideration by the Board. Maystadt, however, decided to put the matter on the back burner after he determined that none of the active proposals had any chance of acceptance by the requisite 85 percent majority of the voting power. Preparing for the April 1995 meeting of the Interim Committee, he suggested that “we reconsider the SDR issue in the context of a wider review of its role in the international monetary system.”44 That suggestion led eventually to a high-level conference on “the future of the SDR,” held at IMF headquarters in March 1996 (Mussa, Boughton, and Isard, 1996). Although the conference brought together many of the world’s leading experts on reserve assets, and although it helped clarify a number of vital issues, it did little to resolve the profound differences of view on the wisdom of SDR allocations.

Despite the continuing political impasse, almost everyone wanted to find some way to solve the “equity issue” so that the post-1981 members of the IMF could enjoy the fruits of the SDR system. In June 1995, the G7 heads of state and government, meeting in Halifax, Nova Scotia (Canada), issued a summit communiqué endorsing “a one-time special allocation” for this purpose. In 1996, meeting in Lyon, France, they reiterated the call and noted that they “continue to hope for progress.” The persistent challenge was to sweeten the proposal for a selective allocation sufficiently to satisfy developing countries that their economic interests were not being threatened, without sweetening it so much that the anti-SDR industrial countries would withdraw their support.

Finally, in September 1996, a delicate compromise was devised to meet this challenge. It had three critical elements. First, the Articles would be amended to call for a special one-time allocation that would result in all member countries having equal total cumulative allocations in proportion to their quotas. Second, the special allocation would be large enough that all members would receive at least a small allocation. Third, the Interim Committee would issue a statement to the effect that the existing purposes and procedures for allocating SDRs, based on a finding of long-term global need, would remain in place. No general allocation would take place at this time, but the special allocation would not prejudice the future.45

Completing the process took another year, owing to technical disputes over the wording of the proposed amendment, an argument about whether countries in arrears to the Fund should get an allocation, and a lengthy battle over the size of the special allocation. An agreement that any special allocation to a country in arrears would be held in escrow until the arrears were settled resolved that question. Getting an agreement on the overall magnitude was harder. By September 1996, Camdessus had lowered to SDR 26.6 billion his 1994 proposal calling for SDR 36 billion. The developing-country caucus was insisting on 30.8 billion and some of the G7 countries, chiefly Germany, were holding to a ceiling of 16 billion. As late as April 1997, the main players were still separated by several billion SDRs.46

The winning formula, adopted by the Executive Board in September 1997, called for doubling the total stock of SDRs, from SDR 21.4 billion to 42.8 billion.47 An allocation of that magnitude, appropriately distributed to equalize cumulative shares, would leave each participant with a stock equal to 29.3 percent of quota. Once that figure was accepted by consensus in the Executive Board, it was endorsed by the Interim Committee on September 21 at its meeting in Hong Kong SAR. Just two days later, the full Board of Governors adopted Resolution 52–4, approving the Fourth Amendment of the Articles of Agreement.48 The remaining task was for national parliaments and legislatures to ratify the amendment.

Ratification—which required approval by 60 percent of member countries, holding 85 percent of the total voting power—turned out to be even more difficult than the struggle to produce the amendment. Almost two years after approval by the Board of Governors, only 60 countries, with 37 percent of the voting power, had ratified the amendment. It took another two years (to November 2001) to get to the required 110 countries, at which point the only remaining hurdle was ratification by the U.S. Congress. By that time, the special allocation was no longer a priority for the U.S. government, and it appeared that Camdessus’s long campaign had come to naught. It was revived, at long last, by the global financial crisis of 2008, which induced a fresh look at all avenues for raising the resources available to the IMF. In June 2009, the U.S. Congress approved the necessary legislation for U.S. acceptance. The Fourth Amendment took effect in August 2009, and the special allocation was made in September.

Members’ Quotas

Quota subscriptions from member countries are the main source of lendable resources for the IMF. The size of those subscriptions can be increased only with the consent of 85 percent of the Fund’s total voting power. As a result, keeping pace with the rapid growth of the world economy that began in the 1950s was a recurring struggle. In addition, quota size has a major bearing on each country’s voting power and the amount it can borrow from the Fund. Because countries grew at different rates, negotiations to adjust the distribution of quotas became more and more contentious.

The Size of the Fund

As every child is supposed to learn in school, a small difference in percentage growth rates eventually accumulates into a large difference in sums. From 1946, when the first quotas took effect for the 40 original members of the IMF, until 1999, total world trade (measured by total imports) grew at a remarkable average rate of 8.6 percent a year. IMF quotas, adjusted for the increase in the Fund’s membership, grew by 5.6 percent a year. Although the growth in quotas also seems impressive, the difference meant that quotas declined from the equivalent of 43 days of world imports at the outset to just 15 days at the end of the century (Table 15.8).

Table 15.8.General Reviews of Quotas, 1950–99



Number of

Stocka (Billions of



(Billions of


Average annual increase (percent)
Sources: IMF Treasurer’s Department (2001), Tables II.3 and II.4; International Financial Statistics; and author’s calculations.Note: This table updates Table 17.2 in Boughton (2001), p. 854.

Stocks are calculated as the sum of quotas for countries that were members when the quota review took effect. These figures differ in some cases from those in the text, which refer to the stock at a particular date.

Imports are at five-year intervals, except that the initial datum is for 1948.

Sources: IMF Treasurer’s Department (2001), Tables II.3 and II.4; International Financial Statistics; and author’s calculations.Note: This table updates Table 17.2 in Boughton (2001), p. 854.

Stocks are calculated as the sum of quotas for countries that were members when the quota review took effect. These figures differ in some cases from those in the text, which refer to the stock at a particular date.

Imports are at five-year intervals, except that the initial datum is for 1948.

Virtually the entire decline occurred in the late 1960s and the 1970s, when three increases in quotas fell far short of keeping pace with growth in international trade. From the equivalent of 41 days of imports in 1965, aggregate quotas settled at 14 days following the increase that took effect in 1980. After 1980, the relationship rose a bit and then fell again.

In principle, general reviews of the size of the Fund were to be held at least once every five years. The first two reviews, completed in 1950 and 1955, resulted in no agreement to raise quotas. The next six reviews occurred at roughly quinquennial intervals, and each one was concluded with an increase of at least 30 percent (an annual rate of 5.5 percent). Work on the Ninth Review, scheduled to be completed in 1988, was still under way as the 1990s began.

Ninth Review

Preparation of the Ninth Review began in 1987, four years after the Eighth took effect. As explained more fully in Boughton (2001, pp. 870–72), the staff initially calculated that growth in trade and other relevant variables since 1983 justified a quota increase on the order of 57 percent. Most Executive Directors favored a doubling of quotas to allow for the further growth that would occur before any quota increase was likely to take effect. Three Directors with a total of 29 percent of the voting power—representing the United States, the United Kingdom, and Saudi Arabia—strongly resisted any such move, and discussions dragged on inconclusively for more than two years.49

The fall of the Berlin Wall and the other revolutionary developments in Central and Eastern Europe shook up these entrenched positions by raising the prospect that the IMF would soon be called upon to undertake large-scale lending to new member countries. By the end of 1989, U.S. opposition to a quota increase was crumbling. In March 1990, Dawson (United States) informed his colleagues on the Board that his authorities were prepared to support a 50 percent increase but would not go higher. Dawson conditioned this support, however, on an agreement that the next quota review be extended by two years from its scheduled 1993 completion date.50 That proposal was resisted successfully via the tactic of throwing the matter to the Interim Committee to decide. When the committee met on May 7–8, 1990, it agreed that quotas should be increased by 50 percent under the Ninth Review and that the Tenth should be completed on schedule.51

Acceptance of the quota increase by the United States was also conditional on a strengthening of the IMF’s procedures for sanctioning countries in arrears on their payments to the Fund. As explained in Chapter 16, the U.S. authorities wanted the Fund to be able to strip noncooperating countries of their voting rights as an intermediate step leading to the unlikely eventuality of forcing the offender to withdraw from membership. That would require an amendment to the Articles, and the United States lacked the votes to get it approved on its own merits. In April 1990, as the date of the Interim Committee meeting approached, Dawson insisted that his authorities would not approve the quota increase unless the resolution to the Board of Governors included a provision linking the effectiveness of the increase to the approval of such an amendment. The Executive Board and then the Interim Committee reluctantly went along because they had no real choice in the matter.

The Executive Board approved the resolutions for a 50 percent quota increase and for the Third Amendment on May 21, 1990. The Board of Governors completed the approval on June 28. That step initiated what would turn out to be a lengthy consent process. Because no increases could take effect until adoption of the unpopular Third Amendment, many countries were in no hurry to give their formal consent to an increase in their quotas. Moreover, the resolutions implied that no increase could take effect before the end of 1991 unless the United States had given its consent, and that seemed unlikely owing to anticipated opposition within the U.S. Congress.52

The agreement to give the United States a second but temporary veto over quota increases in 1990–91 resulted from an Executive Board decision on the “participation requirement.” To ensure that most countries’ quota increases take place about the same time, the Fund traditionally had set a requirement that no increase would take effect until member governments with at least a fixed percentage of existing quotas had given their consent. In most quota reviews through the Eighth, the Executive Board set the floor in a range from two-thirds to three-quarters of total quotas. For the Ninth Review, the staff recommended setting the floor near the middle of that range, at 70 percent.53 At the request of the United States, a two-tier requirement was eventually adopted: 85 percent until the end of 1991 and 70 percent thereafter.54 Practically, however, this odd decision was of little consequence, because the linkage to adoption of the Third Amendment already implied that an 85 percent approval would be needed.

By the end of 1991, 18 months after the adoption of the quota resolution by the Board of Governors, 103 countries with 66.7 percent of existing quotas had consented to the proposed increases. By the same date, just 69 countries with 56.3 percent of the total voting power had accepted the proposed Third Amendment. The 70 percent participation requirement was satisfied in January 1992 with the addition of 13 more consenting countries, but ratification of the amendment was still making little headway. At this stage, all that really mattered was when the U.S. Congress would act. The United States finally ratified the amendment and consented to its quota increase in October 1992, and the increases took effect the following month. That raised total quotas—the “size of the Fund”—from SDR 96.2 billion to SDR 141.4 billion.55

Tenth Review

The requirement in the Articles of Agreement that the Fund must conduct a general review of quotas at least every five years gave rise to a timing anomaly that eventually squeezed out the Tenth Review. Because the Eighth Review was completed in 1983, the five-year period allowed for the Tenth began in 1988 even though the Ninth was still in progress. In 1990, as noted above, the Interim Committee refused to extend the completion date for the Ninth beyond 1993. As that date approached, however, an extension seemed inevitable because no further substantive work on the review had been completed. Accordingly, the Board of Governors approved a continuation of the review in April 1993 and asked the Executive Board to submit a report by the end of 1994.56

Demand for loans from the Fund remained moderate in 1994, and the Fund had adequate resources to meet that demand. Instead of focusing on the amount by which aggregate quotas should be increased, the work program for the Tenth Review in 1993–94 focused on the way the formulas for distributing quotas should be revised. As explained below, that work was inconclusive, and the final report on the Tenth Review suggested only that the next review aim to make further adjustments. In January 1995, the Board of Governors approved the completion of the Tenth Review without any changes in quotas.57

Eleventh Review

The extension of the Tenth Review beyond the scheduled five-year limit did not affect the timing of the Eleventh Review. It started officially in April 1993 and was scheduled to be completed by the end of March 1998. In a “business as usual” world, that would have meant starting work in the spring of 1997. But no sooner had the Fund disposed of the Tenth Review than the Mexican peso crisis applied new pressure to the Fund’s resources and forced an early start to the Eleventh.

The initial staff paper for the Eleventh Review, circulated in August 1995, calculated that aggregate quotas should be raised by at least 60 percent to restore the relationship between quotas and world output that had prevailed when the Eighth Review was completed in 1983.58 Allowing for the time that would elapse before completion of the review, a case clearly could be made for a doubling of quotas in 1998. The Board’s report to the Interim Committee raised that possibility and concluded that “a large increase in quotas is needed to support the activities of the Fund in the years ahead.”59

Work on the Eleventh Review dragged on for two more years. Differences in view on whether the Fund needed more quota-based resources were large but were overshadowed by disputes on how to adjust the distribution of quotas. In September 1996, after a year of such debates, the Board reported that “we have not reached firm positions on either of the issues.” As late as June 1997, Camdessus was pressing for an increase of 55–65 percent, a range that Lissakers (United States), considered too large. She responded by calling it “not credible in light of the Fund’s current and prospective financial position.”60

The outbreak of the financial crisis in Thailand in July 1997 finally focused all energy on reaching a consensus and concluding the review. The direct effect of the crisis on the Fund’s liquidity position and the prospect that several other countries might soon need financial help erased all doubt about the need for a sizeable increase. The main creditor countries, however, remained unconvinced that the increase should be as large as Camdessus was seeking. In August, Lissakers hinted that the United States was prepared to accept a 45 percent increase, but she insisted that “even looking at the effect of the Thailand program on liquidity, a number starting with 3 seems to us more defensible.” A month later, however, when the Executive Board convened in Hong Kong SAR just ahead of the IMF/World Bank Annual Meetings there, she proposed a 45 percent increase as a compromise. That figure was readily accepted, and the Interim Committee approved it the following day.61

The staff proposed that quota increases under the Eleventh Review should take effect as soon as countries with 70 percent of existing quotas had consented to their proposed increases. In general terms, that participation requirement would have been consistent with practice in most previous reviews, although the Ninth Review had employed a two-stage requirement in which 85 percent participation was required for the first 12 months. On the occasion of the Eleventh, Barry S. Newman (Alternate, United States), insisted on a straightforward 85 percent requirement, so that no increases could take effect without the participation of the United States.62

The situation in 1998 differed from that earlier circumstance in that the United States faced the possibility of losing its veto power over major decisions by the IMF. The Articles of Agreement specified certain types of decisions, including amendments to the Articles, quota increases, and allocations or cancellations of SDRs, that could be enacted only with the support of 85 percent of the total voting power. At the outset, in 1946, the United States held 36.2 percent of the vote; but by 1998, after several quota increases and a large increase in membership, the U.S. share was down to 17.8 percent. If every country except the United States consented to the quota increases proposed under the Eleventh Review, the U.S. voting weight would have declined to less than 13 percent. Newman, though, made two other arguments in favor of requiring U.S. participation. First, it would help pressure the U.S. Congress to act favorably on the proposal. Second, it would ensure that the quota increase would have the desired effect on Fund liquidity—otherwise it would not add many U.S. dollars to the Fund’s resources.

A slight majority of the Board, holding 60 percent of the vote, opposed Newman’s proposal and preferred to accept the staff recommendation for a 70 percent participation requirement.63 That raised the prospect that the Executive Board would adopt a resolution that would not be accepted by the 85 percent majority vote required in the Board of Governors. The First Deputy Managing Director, Stanley Fischer, who was chairing the meeting, decided to postpone a decision for the time being. At the next meeting, three weeks later, Camdessus reiterated his preference for 70 percent but asked the Board to accept the U.S. request in “a spirit of a consensus. … There were solid grounds,” he averred, “for giving the United States the benefit of the doubt, and for being confident that the United States would not be the last member to agree to its quota increase.”64

The Board of Governors approved the proposal for a 45 percent increase on January 30, 1998. After that vote, the only remaining hurdle of consequence was the consent of the U.S. Congress. Five days later, Camdessus made a rare trip to Capitol Hill to testify to an executive session of the Senate Budget Committee and urge passage of the necessary legislation. President Bill Clinton’s administration also made a strong appeal for approval, but it faced sharp opposition from a legislature controlled by the other major party. Curiously, this situation was the same as—but also the opposite of—that of 1992, when a Republican administration supported the quota increase and a Democratic legislature opposed it. In each case, the lack of a solid domestic constituency in favor of funding for international institutions was reinforced by the inevitable political battles resulting from two-party coexistence.65

Opposition crumbled in the autumn of 1998, in the middle of the Russian financial meltdown, the near collapse of the Long-Term Capital Management hedge fund, and growing concerns about financial stability in Argentina and Brazil. Republican leaders in the House of Representatives joined those in the Senate in supporting funding for the IMF, and both houses passed the enabling legislation for the quota increase in October. Other countries that had been waiting for the United States to act first began taking their own legislative actions. The 85 percent participation requirement was finally reached, and the quota increases began taking effect on January 21, 1999—al-most a full year after approval by the Board of Governors, but still much faster than the 28-month completion process for the Ninth Review in 1990–92.

With the effectiveness of the Eleventh Review, quotas totaled SDR 210.2 billion at the end of 1999, an increase of 45 percent and more than double the size of the Fund at the start of the decade.

The Distribution of Quotas

The starting point for any discussion of the appropriate distribution of quotas among member countries is a set of equations known as the “Bretton Woods formula” and its variants. The U.S. team at Bretton Woods in 1944 constructed the original equation to produce a list of quotas that would be broadly acceptable to the potential membership of the IMF. The equation’s arguments included estimates of each country’s national income, international trade, and official reserves; and its coefficients were determined heuristically (not by statistical techniques) to generate a desired set of results. Subsequent modifications to, and variations on, the Bretton Woods formula were negotiated by the Executive Board with the aim of updating the distribution through the periodic general reviews. In a typical round of discussions, the staff would compute a list of quotas based on updated data, Executive Directors would complain about the implications for their countries, and the staff would modify the equations to tweak the results until a consensus developed that the outcome was acceptable.66

Redistribution in the Ninth General Review

As work developed on the Eighth Review in the early 1980s, a political consensus emerged in the Interim Committee that the allocation of quotas should be brought up to date more aggressively than had been the case in previous reviews. The end result was that 60 percent of the increase in quotas was distributed according to the formula calculations and just 40 percent according to the existing distribution. As work proceeded on the Ninth Review in the late 1980s, this redistribution commitment continued, though with a smaller portion of the total (40 percent) dedicated to it.

Japan and the G7. One especially glaring misalignment was Japan’s low quota. Upon joining the IMF in 1952, Japan was assigned the ninth largest quota, just below the Netherlands and above Belgium. It gained a seat on the Executive Board by forming a constituency with a few other Asian countries—Burma, Ceylon, Thailand, and later Nepal. The Japanese economy’s growth then accelerated rapidly over an extended period, and the Fund responded by increasing Japan’s quota at above-average rates. Japan moved up to seventh place (tied with Italy) in 1959, surpassed Italy in 1966, and moved up again to fifth place (between France and Canada) in 1970.

With that leap in 1970, Japan replaced India as one of the five countries entitled (and required) to appoint its own Executive Director instead of participating in the biennial constituency elections. Japan’s quota (and voting share) remained in fifth place for the next two decades despite a sustained growth rate among the highest in the world and higher than that of any of the countries with larger quotas.

Table 15.9 shows the actual and calculated quota shares for the five leading countries as of December 1987. The only major country with an outsized shortfall was Japan, which by then had the second largest economy in the world. Also reflective of the inertia in the adjustment of quotas, the relatively slow-growing United Kingdom had the only sizeable overage.

Table 15.9.Actual and Calculated Quota Shares of the Five Largest Economies in 1987
CountryActual Quota ShareCalculated ShareGap
United States19.920.2-0.3
United Kingdom6.95.2+1.7
Source: “Ninth General Review of Quotas—Revised Quota Calculations,” EB/CQuota/87/5 (December 22, 1987), Table 1.
Source: “Ninth General Review of Quotas—Revised Quota Calculations,” EB/CQuota/87/5 (December 22, 1987), Table 1.

In the second half of the 1980s, the Japanese government decided it should play a larger and more active role in international affairs, including through multilateral institutions. The minister of finance, Kiichi Miyazawa, and his deputy, Makoto Ut-sumi, were particularly active in strengthening Japan’s role in the IMF. As a result, Japan became the leading bilateral lender to the IMF, a major contributor to the ESAF, and an important provider of funding for technical assistance and other support for low-income countries.

By 1987, when the Fund began work on the Ninth Review, Japan had a formidable case for being recognized through a substantial increase in its share in IMF governance. The Executive Director from Japan, Koji Yamazaki, made such a case repeatedly, based both on the calculated quota shortfall and Japan’s increased global role. As he stressed in March 1988,

the present situation of Japan’s quota is so incommensurate with economic reality that our authorities are put in quite an awkward position in their relationship with the Diet [the Japanese parliament] whenever the promotion of cooperation with the Fund, such as the contribution to the [ESAF], is called for. Proper rectification is by all means essential.67

Although no one disputed Japan’s claim, the proposed adjustment would be large enough to have a major negative impact on quota shares for many other countries. The French authorities specifically objected to the suggestion that their quota share should be reduced from fourth to fifth place. The British authorities were willing from the outset to see their rank decline from second to fourth, but they also objected to the idea of falling all the way to fifth. Although the French economy was the larger of the two, both the British role at Bretton Woods and centuries of tense history between the two neighboring countries made this issue more diplomatically stressful than would have made sense otherwise.

More generally, Executive Directors agreed that the shares of developing countries should not be reduced as a side effect of the proposed rebalancing at the top. Extensive discussions made it clear that any effort to rebalance the largest quotas within the constraints of the accepted formulas would lead to anomalous and unacceptable results for other countries. At that point, Frank Cassell (United Kingdom) and Miguel A. Fernández Ordóñez (Spain) independently suggested that the Executive Board should just leave the matter to the G7 to decide. That is, the Board would set an aggregate quota share for the G7 countries and let the G7 allocate that share among themselves. E.A. Evans (Australia) objected on principle, but most Directors seemed happy to be offered a solution that relieved them of having to slice such an awkwardly shaped pie.68

As the scene shifted to the G7, the solution seemed to slip out of sight once more. Finance officials argued for months about how to share the burden of the Japanese increase between Britain and France, both of which insisted on holding the fourth largest quota. (Germany qualified for third place by a clear margin and stayed out of the fight.) At the end of March 1990—the deadline that had been set for completing the review—the matter remained unresolved. The group’s finance ministers scheduled a meeting for April 7 in Paris to be hosted by the French minister, Pierre Bérégovoy, who was publicly and adamantly insistent on maintaining France’s rank in the hierarchy. When that meeting failed to produce an agreement, the next opportunity was the regularly scheduled G7 ministerial meeting in advance of the Interim Committee meeting, in Washington on May 6.

Two days after the G7 meeting in Paris, a larger group of 40 countries meeting in the same city concluded an agreement to establish the European Bank for Reconstruction and Development. That opened a new battleground on which several European countries would vie to have the new multilateral agency located in their capital city while France and the Netherlands would press to have one of their own officials named as the bank’s first president. For the next month, those contests played out in a series of negotiations while the international press enjoyed the spectacle. By late April, the choice of a headquarters location had narrowed to London or Paris, capital cities of the same two countries squabbling over the ranking of their IMF quotas. In the first week of May, the United Kingdom and France struck a bargain that resolved all of the issues in dispute. The European Bank for Reconstruction and Development would build its headquarters in London; Jacques Attali—special advisor to French President François Mitterrand, who had initially proposed establishing the bank—would become its president; and France and the United Kingdom would share fourth place in the IMF with identical quotas. Because Germany was about to expand via a merger of East and West Germany, the G7 ministers also decided that Germany (then in the third slot) should share the number two position with Japan. The Interim Committee ratified the quota agreement a few days later, generating a collective sigh of relief.

The final agreement provided for an increase in the G7 share from 48.9 percent of total quotas to 49.4 percent. Japan’s share rose from 4.7 to 6.1 percent; the British share fell from 6.9 to 5.5 percent; France’s share rose from 5.0 to 5.5 percent; and the other shifts were marginal.69

Other adjustments. The more general effort in the Ninth Review was to continue to adjust the distribution of quotas by giving fast-growing and more-open economies relatively larger percentage increases. For the great majority of member countries, the decision to devote 40 percent of the total increase according to calculated rather than actual quotas resulted in increases ranging from 32 to 70 percent. The only outlier on the low side was the United Kingdom, which received only a 19.7 percent increase for the reasons described previously.

Increases in excess of 70 percent were received by 11 countries (Table 15.10). Three (Bhutan, Maldives, and Seychelles) were very small countries with low quotas, for which the Fund agreed to round the results upward to ensure them a meaningful share in the governance structure. In addition to Japan, a few others—ranging from Singapore (287 percent) to Korea (72.8 percent)—received above-average increases in recognition of their rapid growth. The Korean authorities asked for a larger ad hoc adjustment to correct a ridiculously large gap between their actual and calculated quotas. They almost certainly would have succeeded, except that the Islamic Republic of Iran was asking for similar treatment. The Fund had not held a consultation with Iran for more than a decade, owing to tensions with the international community dating to the time of the country’s 1979 revolution and the year-long hostage taking and occupation of the U.S. embassy in Tehran. Agreeing to the Korean request while rejecting Iran’s would have made an awkward impression of political partiality. Both requests were dropped before the final report was prepared in 1990.

Table 15.10.Largest Percentage Quota Increases, 1992 and 1999
Ninth General Review (1992)Eleventh General Review (1999)
Seychelles100.0Korea, Rep. of104.3
United Arab Emirates93.5Congo, Democratic Rep. of83.2
Bhutan80.0San Marino70.0
Korea, Rep. of72.8
Memo: Average increase50.045.0
Source: International Financial Statistics.
Source: International Financial Statistics.

The Executive Board held extensive discussions on the treatment of countries with arrears to the Fund. The final decision was to award increases to them, with the proviso that no country could consent to its increase as long as it had outstanding arrears to the GRA.70 That prevented five countries—the Democratic Republic of the Congo, Liberia, Somalia, Sudan, and (until 1995) Zambia—from receiving any increase. Cambodia got a 160 percent increase in its quota in 1994 after it settled its arrears. That large increase was calculated to make up for Cambodia having had no increase in the preceding three reviews, when relations with the Fund were completely interrupted (see Chapter 16).

Quotas for New Members

Between the Executive Board’s approval of the Ninth Review in May 1990 and effectiveness of the increases in November 1992, 22 countries became members of the IMF. In a few cases, the Fund could set initial quotas in a straightforward manner by examining the data, comparing the economy with those of similar member countries, and slotting them in accordingly. For others, unique problems intervened.

The former Soviet Union. Fifteen of the new members emerged from the former Soviet Union. No reliable data existed on their national income, international trade, or reserves. As explained in Chapter 7, the best the Fund staff could do was collect aggregate data for the Soviet Union, convert it roughly from socialist to market concepts, and then develop some rules of thumb on the relative size of each of the 15 component economies. The depressed state of the Soviet economy in the late 1980s caused the staff’s calculations to produce results that Russian negotiators and officials considered unacceptably low. After further negotiation and some high-level political intervention, the Executive Board agreed to award Russia a quota share of 3 percent (just below Canada). Quotas for the other 14 newly independent countries were set in relation to Russia and ranged from 0.69 percent for Ukraine to 0.03 percent for Estonia and Turkmenistan. The aggregate quota share for Russia, the Baltic countries, and other countries of the former Soviet Union was set at 4.76 percent.71

Switzerland. Switzerland, which applied for membership in the IMF in May 1990, posed the other challenging case. C. Scott Clark (Canada) was selected to chair the membership committee (a subset of Executive Directors charged with recommending the terms on which membership would be offered to an applicant, including the quota). Jean-Pierre Landau (France) agreed to represent the interests of Switzerland on the committee. Normally, such committees completed the task fairly quickly by basing their deliberations on a staff background paper. The staff would make a series of calculations using available economic data for the country and typically would conclude with a recommended narrow range for the new member’s quota. In this case, that process got short-circuited.

Conflicting views on the relative size and importance of Switzerland’s economy confronted the exploratory staff mission, led by Harilaos Vittas (Assistant Director, European Department), sent to collect data and exchange information with the Swiss authorities. Total output, as measured by GDP, was smaller than that of any G10 country and was closer in magnitude to the larger developing countries (smaller than Saudi Arabia but larger than Nigeria). Switzerland, however, was a major player in international finance. Its official reserves of gold and foreign exchange, for example, were estimated to be the eighth largest in the world.72 The fact that Switzerland was potentially among the largest creditors to the IMF could scarcely be ignored.

Using standard data sets, equations, and procedures, the staff calculated several possible quotas for Switzerland, most of which were in a range between SDR 1.35 billion and SDR 1.45 billion. The staff report acknowledged that if financial variables were taken into account more directly, a higher figure would result. One outlier in the computations was a comparison with the United Kingdom, which also was a major financial center and which (as discussed above) still enjoyed an unusually large quota as a legacy from its role in the origins of the IMF. To be comparable to the British quota, Switzerland’s would have to be as high as SDR 1.79 billion. Anticipating pushback, the staff declined to suggest a specific range. Instead, the report merely suggested that these figures “may be considered as useful indicators by the Committee in making its recommendation of an appropriate quota for Switzerland.”73

The Swiss authorities reacted negatively to these estimates.74 The government asked to be represented directly at the initial meeting of the membership committee by the Swiss ambassador to the United States, Edouard Brunner. The committee agreed, and Brunner delivered a statement at the meeting rejecting the staff estimates out of hand. “My authorities expect,” he averred, “a quota of SDR 2.1 billion, which is commensurate with Switzerland’s financial, monetary, and economic importance in the world.”75 That approach did not go down well with some on the committee, who insisted that the Fund had to apply its own rules and logic, not those of the applicant, and who thought SDR 2.1 billion grossly unrealistic. A few committee members, including Thomas Dawson (United States), were willing to consider numbers between those of the staff and the high figure requested by the Swiss but supported only by Landau. A “very strong majority,” however, wanted to set the quota no higher than SDR 1.5 billion, according to the committee’s chairman.76

Without a staff recommendation as a clear guidepost, committee members were left to haggle among themselves. In a series of five meetings over six months, they debated detailed and obscure technical issues such as the treatment and measurement of income from international banking activities and the role and valuation of nonmonetary gold transactions. The real issues, though, were political and were serious matters for the IMF—Switzerland’s ranking in the hierarchy, especially relative to other European countries, and the potential effect on the composition of the Executive Board.

The challenge in placing Switzerland into the ranking of quotas is best illustrated by reference to Spain, a European country with a GDP more than twice that of Switzerland. Spain’s relatively low quota dated from its entry into the IMF in 1958, when still under the dictatorship of Francisco Franco. In 1990, Spain’s quota (SDR 1.29 billion) was ranked eighteenth among existing members, between Venezuela and Mexico, and the Spanish authorities were waging a largely unsuccessful campaign to move up the ladder. Swiss membership with any of the quota numbers being discussed would reduce Spain’s position to nineteenth, but a much larger quota would add insult to injury.77

Speaking for the Spanish authorities, Miguel A. Fernández Ordóñez (Alternate, Spain) came close to threatening a withdrawal of his country’s support for the institution if Switzerland got a larger quota than was justified by the staff’s calculations. Spain would not “withdraw” the financial contributions it had already made, he promised. “Nevertheless, it would be very naïive to think that a decision to worsen the situation of a current member could have no consequences at all. Certainly a decision to alter the ranking that is obvious in the real world, and consistent with the calculations made by the staff, would certainly affect relations with the institution and in the long run could be costly for other members.”78

As for the second underlying issue, the membership committee did not have a mandate to discuss the effect of the quota decision on the composition of the Executive Board. That issue would be taken up by the full board at the time of the next election of Executive Directors, as discussed in Chapter 17. Nonetheless, granting a large quota to Switzerland had the potential to squeeze out the constituency with the smallest aggregate voting power. That constituency comprised the group of 24 francophone countries in sub-Saharan Africa. Everyone understood that preserving a seat for that group was essential, but the Executive Director for the group did not want to be placed in the awkward position of having to negotiate a solution. When it became clear that most developing countries were going to resist Switzerland’s request for a larger quota, the Swiss finance minister, Otto Stich, wrote to a number of Fund Governors to reassure them. His letter to S.M.H. Adeli, governor of the central bank of Iran and chairman of the Group of Twenty-Four (G24) developing countries, assured him that Switzerland “does not intend to claim a seat to the detriment of the developing countries.”79 Vague as that promise was, it helped to minimize the rebellion before it became too entrenched.

After four committee meetings failed to produce a consensus, Clark met bilaterally with each member to try to narrow the range of differences. At the outset of the fifth meeting, on March 4, 1991, he suggested that the range under consideration be narrowed to SDR 1.55 to 1.75 billion. In a spirit of weary compromise, 9 of the 12 committee members acquiesced to a quota of SDR 1.7 billion. Three others—Grant H. Spencer (Alternate, New Zealand), G.A. Posthumus (the Netherlands), and Coren-tino V. Santos (Cape Verde, speaking for francophone Africa)—dissented and asked for a quota no higher than 1.55 billion.80 The full Executive Board approved the compromise later that month, and the Board of Governors approved it as part of the offer of membership to Switzerland on April 24. After another long delay, Switzerland finally entered the Fund on May 29, 1992, with a quota of SDR 1.7 billion: the fourteenth l argest in the Fund, between Belgium and Australia and four places above Spain.81

Redistribution in the Eleventh General Review

In the Eleventh Review, the Board continued the policy of making modest adjustments by dedicating a portion of the overall increase to redistribution in favor of countries that had experienced relatively rapid economic growth. On this occasion, the distribution component was further reduced, to one-fourth of the total 45 percent increase. All but eight countries got increases ranging from 34 to 63 percent. The outliers all got larger increases, led again by Singapore (see Table 15.10). Two countries—the Democratic Republic of the Congo and Iraq—got catch-up increases for having missed out in the previous round. The Congo had been prevented from consenting to its earlier increase because it had arrears to the Fund. Iraq had not consented to its increase because it was subject to international sanctions after the 1991 war.

The lack of major realignment during the Eleventh Review reflected what Camdessus called “the tradition of the Fund to allow any realignment of the rankings of countries to occur only with the utmost gentleness and care.”82 That tradition injected a high degree of inertia into the distribution of quotas, and the resulting inflexibility was becoming a growing problem by the late 1990s. Although no one wanted to scrap this tradition and attempt a zero-based, all-new quota calculation in the context of this review, a consensus was emerging that sooner or later the IMF would have to do just that. Accordingly, in April 1997 the Interim Committee asked the Executive Board “to review the quota formulae promptly after the completion of” the Eleventh Review.83

In June 1999, Camdessus appointed a committee of external experts, chaired by Harvard economics professor Richard N. Cooper. The terms of reference for the Cooper Committee called for it to review the adequacy of the quota formulas “to help determine members’ quotas in the Fund in a manner that reasonably reflects members’ relative need for and contributions to the Fund’s financial resources, taking into account changes in the functioning of the world economy and the international financial system and in light of the increasing globalization of markets.”84 The committee responded with a report heavily in favor of a formulation that was “simple and transparent” and that would “have a sound economic basis and reflect the relevant changes in the world economy.” Although the committee’s terms of reference asked it to consider including population as one variable determining quotas, a majority of the committee’s members decided against it. Instead, they recommended adopting a simple linear function of a country’s output (GDP) and “external vulnerability.”85

The results pleased almost no one, even though it would have modernized the quota system. Although some countries would have benefited by getting well-deserved increases, the shifts overall would have reinforced the initial perceptions of dominance by the few. The United States and most other major industrial countries would have received increased quota shares, as would rapidly growing emerging-market countries such as Brazil, Korea, and Mexico. Shares for China, India, Saudi Arabia, and South Africa would have declined, along with those of slowly growing European countries, including France, the Netherlands, and the United Kingdom. The aggregate share of the least-developed countries would have declined from 3.6 to 2.1 percent.86 No action was taken on the report, and substantive reform of quota distribution was again put on hold.

Borrowing by the IMF

The IMF has borrowed episodically over much of its history, whenever its quota-based resources have become insufficient to meet its members’ financing demands.87 In 1962, the Fund entered into an agreement with a group of major industrial countries to establish the General Arrangements to Borrow (GAB). That agreement created a standing line of credit to the Fund that could be activated to meet large, systemically important demands for Fund financing from the GAB creditors themselves. On four occasions before 1990, the Fund also borrowed bilaterally from national monetary authorities or governments: from Italy in 1966, from the Swiss National Bank in 1977, from Saudi Arabia in 1981–82, and from Japan in 1987–91. In addition, on three occasions, the Fund borrowed from ad hoc groups of countries and monetary authorities: in 1974–75 to finance the oil facilities, in 1979–84 to finance the Supplementary Financing Facility (SFF), and in 1981–86 to finance the policy on enlarged access to the Fund’s resources (EAR). By the end of the 1980s, only the GAB and the bilateral arrangement with Japan were still active, although some of the SFF and EAR obligations were also outstanding. Total outstanding obligations in April 1990 amounted to $4.6 billion, or 4 percent of total Fund quotas, down from peaks of $16.7 billion in April 1986 and 26 percent of quotas in April 1977.

The 1992 increase in quotas provided the resources the IMF needed for the next several years. The Fund undertook no new borrowing until 1998, when it activated both the GAB and, for the first time, the New Arrangements to Borrow (NAB), as discussed below.

General Arrangements to Borrow

Unlike the ad hoc borrowing arrangements the IMF undertook from time to time, the GAB and associated agreements provided a standing commitment by a group of 12 countries and monetary authorities to lend up to SDR 18.5 billion to the Fund on an as-needed basis.88 Its purposes, however, were limited. Originally it could be activated only for loans to a member country of the Group of Ten (G10), which was composed of the GAB creditors which—at the time—were considered to be the only systemically important countries. That restriction was dropped in 1983 so that the Fund could activate the GAB if necessary to manage the Latin American debt crisis. It was not activated then, but it remained in reserve in case “supplementary resources are needed to forestall or cope with an impairment of the international monetary system.”89

The Fund considered borrowing from the GAB on a few occasions in the 1990s, but it only activated the arrangements in 1998. In 1992, the G7 countries were looking for ways to provide financial support to the newly independent Russian Federation. Camdessus proposed activating the GAB to finance a $6 billion fund to stabilize the exchange value of the ruble. The G7 liked the idea, but the currency stabilization fund was never established (Chapter 7). Two years later, when the Mexican peso crisis erupted, it became clear that the IMF would be called upon to lend a large sum of money to Mexico. In the view of Fund and U.S. officials, the crisis was a threat to the international monetary system. As recounted in Chapter 10, Camdessus considered asking the G10 to activate the GAB for this purpose, but the staff talked him out of it because some European members of the G10 were unlikely to support the request without a battle. During the East Asian crises in 1997, the Fund did not need the GAB to finance its lending to Thailand, Indonesia, or Korea, although the surge in Fund lending did spark discussions on the need for new borrowing arrangements, as described below. In 1998, the impending collapse of the Russian banking system necessitated emergency loans that were large enough and systemically important enough that activating the GAB was both necessary and widely supported.

In July 1998, the Fund determined it should augment its existing Extended Fund Facility (EFF) arrangement with Russia by the equivalent of $8.4 billion (SDR 6.3 billion). By this time the Fund’s liquidity had already been squeezed by its lending in East Asia, so Camdessus decided the augmentation should be financed entirely by borrowing from the GAB. After preliminary discussions with GAB participants, the Acting Managing Director, Alassane D. Ouattara, wrote to each participant on July 14 to seek approval of the request. The Fund received acceptances from all 11 participants over the next few days, in time for the Executive Board to approve activation as part of its augmentation of the commitment to Russia on July 20.90 The first tranche of the borrowing took place immediately, for $1.9 billion (SDR 1.4 billion). Two future calls were anticipated to take place in September and December 1998, but Russia’s debt default in August put those plans in abeyance. The September tranche thus was the only call on the GAB in this decade.

At the insistence of the Executive Director for Germany, Bernd Esdar, who was supported by Directors from several other GAB participants, the Fund91 In March 1999, the Fund repaid the loans in full.92

New Arrangements to Borrow

After the Mexican crisis in 1995, the limitations on the use of the GAB and its narrow base of creditors led to consideration of new borrowing arrangements. The idea of creating a parallel system to attract lenders from among the smaller advanced economies and the emerging-market economies in the developing world originated within the G7 as officials prepared for the group’s annual summit meeting in 1995.93 As expressed in the summit report, the proposal was to create a new “emergency financing mechanism” within the IMF and to finance it by doubling the size of the GAB with resources to be lent by “the G10 and other countries with the capacity to support the system.”94 The IMF embraced both suggestions but demurred at the link between them. If the Fund needed to supplement its own quota-based resources, it should do so independently of the need for a new lending facility. Moreover, the staff pushed for a broader approach that would create a new facility open to all creditors, not just those approved by the existing participants in the GAB.95

To carry the expansion idea forward, the G7 appointed a working group, chaired by the assistant deputy minister of finance of Canada, Thomas A. Bernes (later Canada’s Executive Director in the Fund). That group prepared a report that was discussed first by the finance deputies of the G7 at a meeting in Rome in September 1995 in which the Fund’s Treasurer, David Williams, and its Economic Counsellor, Michael Mussa, participated. The deputies decided against trying to expand the GAB, which could have effectively spelled the end of the G10 by inviting a host of other countries to share in one of the group’s flagship functions. Instead, they devised a scheme more in keeping with the Fund’s preferences. Under this plan, the availability of borrowed resources to the IMF would be doubled by creating a new arrangement open both to the original participants and to others, while retaining the GAB in its existing form.96

This two-arrangements scheme was fleshed out progressively over the next year. The discussions gradually expanded to include all of the countries and institutions qualified and willing to participate in the New Arrangements to Borrow (NAB). The participants agreed on the text of a draft instrument at a meeting in Brussels in July 1996, and the IMF Executive Board formally approved the NAB on January 27, 1997.97

The essence of the NAB agreement was to nearly double the size of the pool of funds the IMF could borrow under a standing arrangement, to increase the number of potential lenders from 12 to 25, and to inject more flexibility into the procedures of the borrowing arrangements. Whereas activation of the GAB required agreement by all GAB participants, the NAB could be activated with an 80 percent majority.98

Most of the NAB creditors–21 out of 25—were member countries of the IMF (Table 15.11). Three others were the central banks of IMF members Germany, Sweden, and Switzerland. Inclusion of the remaining participant, the Hong Kong Monetary Authority (HKMA), posed special technical problems for two reasons. First, it was not the central bank of a Fund member or even of a country as that term was normally understood. While the negotiations for the NAB were occurring, Hong Kong was a territory of the United Kingdom, but preparations were under way for sovereignty to be transferred to China on June 30, 1997. The HKMA operated independently and would continue to do so after the hand over, but for purposes of the NAB it was considered to be an official institution of the member whose territory included Hong Kong. The NAB agreement, therefore, specified that the consent of the United Kingdom or China (as appropriate) would be required before the HKMA could participate in a call on the arrangements. Second, the Hong Kong dollar was not a usable currency for IMF operations. In all other cases, each participant was expected to lend to the IMF in its own currency. Loans from the HKMA to the IMF would have to be made in another currency, normally the U.S. dollar. In all cases, however, the loans would be denominated in SDRs.

Table 15.11.Borrowing Arrangements in Effect, 1998–99(Millions of SDRs, except as noted)

(= NAB)


Deutsche Bundesbank (central bank of Germany)3,5572,380.0202.1327.9530.0
Hong Kong Monetary Authority34031.331.3
Korea, Rep. of3400
Saudi Arabia1,7801,500.0a0
Sveriges Riksbank859382.532.579.2111.7
(central bank of Sweden)
Swiss National Bank (central1,5571,020.086.6143.5230.1
bank of Switzerland)
United Kingdom2,5771,700.0144.3237.6381.9
United States6,7124,250.0360.9618.8979.6
Total in U.S. dollarsc46,12225,0961,9583,9025,860
Source: Annual Reports.Note: Details may not sum to totals, owing to rounding.

Associated Agreement.

Includes the Associated Agreement with Saudi Arabia.

At average 1998 exchange rate.

Source: Annual Reports.Note: Details may not sum to totals, owing to rounding.

Associated Agreement.

Includes the Associated Agreement with Saudi Arabia.

At average 1998 exchange rate.

An innovative feature of the NAB agreement was that it included a cap on total borrowing by the IMF from the two arrangements plus the associated agreement with Saudi Arabia. That cap was set at SDR 34 billion, which was also the size of the NAB.99 The new total was twice the amount available under the GAB, but slightly less than that ratio when the associated agreement with Saudi Arabia was taken into account. That is, the maximum amount the Fund could borrow increased to SDR 34 billion from SDR 18.5 billion.100 The intention was that the NAB would be the “facility of first and principal recourse” relative to the GAB for Fund borrowing. The only exceptions would arise if a call on the arrangements was for the purpose of financing a loan to a GAB participant or if NAB participants refused to accept a proposed call.

Ratification of the NAB by participants took close to two years. Like the quota increase under the Eleventh Review, U.S. participation in the NAB had to be approved by the U.S. Congress. That body got entangled in domestic political disputes for some months, but finally adopted the required legislation on November 17, 1998. In Germany, the Deutsche Bundesbank approved its participation on the same day, and the NAB was in business.

The delay in ratification meant that the NAB was not available to help finance the Fund’s management of the major financial crises that afflicted East Asia and Russia in 1997 and 1998. The agreement came into effect just in time for the next shock: the collapse of financial stability in Brazil, which struck in December 1998. The IMF borrowed SDR 2.9 billion ($3.9 billion) to finance the initial disbursement under a new stand-by arrangement with Brazil. As with the activation of the GAB for lending to Russia earlier that year, the Fund repaid the loans in March 1999 when it obtained an augmentation of its own resources through the general quota increase.

These various borrowing arrangements and their activations in 1998 are summarized in Table 15.11. The activation of the GAB for the Russian EFF arrangement was supported by calls on all 11 GAB participants in proportion to their standing commitments. A larger group of 21 NAB participants, again in proportion to their commitments, shared in the lending in support of Brazil. On that occasion, four countries—Korea, Malaysia, Thailand, and Saudi Arabia—opted not to participate. The first three were still recovering from their own financial crises, and Saudi Arabia had a weakened external payments position as a result of the low level of world oil prices. Once the GAB and NAB loans were repaid in March 1999, the Fund’s General Department had no outstanding debts for the rest of the year.

Appendix: A Primer on the SDR

In the second half of the 1960s, the U.S. dollar was the only reserve currency in widespread usage. With almost all currencies fixed in value against the dollar, and the dollar in turn pegged to gold, the continual growth in world trade and finance required a commensurate increase in dollars to be added to reserves. The unsustainability of this relationship (known as the “Triffin dilemma”) led to discussions about the creation of an additional reserve asset. In 1969, the IMF’s member countries adopted the First Amendment to the Articles of Agreement, establishing the special drawing right (SDR) for this purpose. As the unit of account for the IMF, the SDR had a value fixed at the gold content of the U.S. dollar at that time (0.888671 grams, that being the equivalent of the official gold price of $35 an ounce). As a medium of exchange, the SDR could be held or traded only by the IMF, those member countries that elected to participate in the newly created SDR Department of the Fund, and multilateral institutions such as development banks that would be accepted by the IMF as “prescribed holders.” As a financial asset, SDRs would be created—”allocated” was the official terminology—upon approval by the Board of Governors, with an 85 percent majority vote being required.

In three annual installments, 1970–72, the IMF allocated SDR 9.3 billion (equal to $9.3 billion) to the 105 participating member countries, in proportion to each country’s quota in the Fund. Coincidentally, by the end of that sequence, the dollar was no longer convertible into gold, and exchange rates were floating between the dollar and other key currencies. In response, the valuation of the SDR was changed in 1974 to a basket of 16 currencies, with fixed amounts of each currency in the basket. To simplify the calculation, the Fund reduced the basket to five currencies in 1981. Meanwhile, the Fund conducted a second series of annual allocations in 1979–81, bringing the total stock to SDR 21.4 billion (then worth about $27 billion, reflecting the dollar’s depreciation during the 1970s).101

The primary financial role of the SDR is as a de facto line of credit between authorized holders. When a participant in the SDR Department (which since 1980 has included all IMF member countries) receives an allocation, it treats its SDR holdings as part of its official reserves, offset by a corresponding long-term liability (the allocation). It receives interest on its holdings and pays interest on its allocation. For both purposes, the SDR interest rate is a weighted average of top-quality short-term security rates in each of the five component countries. An allocation thus is costless. The country has the option of using some or all of its allocation to settle obligations with another participant, the Fund, or a prescribed holder. As its holdings fall below its allocation, the country pays net interest to the SDR Department. As other countries receive those SDRs, they earn net interest. Until 1981, net users were required to “reconstitute” their holdings periodically by repaying these loans. The reconstitution requirement was then abrogated so that participants could become permanent net users, in effect borrowing up to the amount of their allocation from other participants at the short-term (floating) SDR interest rate.


    BoughtonJames M.2001Silent Revolution: The International Monetary Fund 1979–1989 (Washington: International Monetary Fund).

    ClarkPeter B. and Jacques J.Polak2004“International Liquidity and the Role of the SDR in the International Monetary System,”Staff Papers Vol. 51No. 1 pp. 4771.

    deVriesMargaretGarritsen1976The International Monetary Fund, 1966–1971: The System Under Stress Vol. I: Narrative Vol. II: Documents (Washington: International Monetary Fund).

    deVriesMargaretGarritsen1985The International Monetary Fund, 1972–1978: Cooperation on Trial Vol. I and II: Narrative and Analysis Vol. III: Documents (Washington: International Monetary Fund).

    HorsefieldJ. Keithed.1969The International Monetary Fund 1945–1965: Twenty Years of International Monetary Cooperation. Vol. III: Documents (Washington: International Monetary Fund).

    International Monetary Fund Treasurer’s Department2001Financial Organization and Operations of the IMF,IMF Pamphlet Series No. 456th edition (Washington: International Monetary Fund). Accessed at

    KaeserDaniel2004La Longue Marche vers Bretton Woods: Chronique des relations de la Suisse avec le Fond Monetaire et la Banque Mondiale [The Long Road to Bretton Woods: A Chronicle of Swiss Relations with the IMF and the World Bank] (Chêne-Bourg and Geneva: Georg Éditeur).

    LissakersKarin2006“Is the SDR a Monetary Dodo? This Bird May Still Fly,”in Reforming the IMF for the 21st Centuryed. byEdwin M.Truman (Washington: Institute for International Economics) pp. 48392.

    LockeMary2000“Funding the IMF: The Debate in the U.S. Congress,”Finance and Development Vol. 37No. 3 pp. 5659.

    MomaniBessma2009“Throwing Swiss Financial Weight Around to Push for a Seat” (unpublished; Waterloo, Ontario: University of Waterloo).

    MussaMichaelJames M.Boughton and PeterIsardeds.1996The Future of the SDR in Light of Changes in the International Financial System (Washington: International Monetary Fund).

    WijnholdsJ. Onno de Beaufort and ArendKapteyn2002“International Financial Crises: The Role of Reserves and SDR Allocations” in Financial Policies in Emerging Marketsed. byMario I.Blejer and MarkoŠkreb (Cambridge, Massachusetts: MIT Press) pp. 95162.

The division of the IMF’s accounts into separate “departments” (General and SDR) should not be confused with the organizational division into area and functional departments such as the African Department and the Fiscal Affairs Department. The two concepts are completely independent.

The operations of the trust funds, or administered accounts—chiefly the Enhanced Structural Adjustment Facility, the Poverty Reduction and Growth Facility, and the Heavily Indebted Poor Countries Trusts—are covered in Chapter 13.

The term in use through the 1990s was “operational budget.” That ill-fitting term was replaced by “financial transactions plan” in 2000.

Eligibility for receipts-only usage meant only that the country had a sizeable reserve tranche position in the IMF. That is, if a country had a sufficiently large reserve position in the Fund but a weak balance of payments, its currency could be used in receipts but not in lending operations; see “Principles for Calculating Amounts of Currencies under the Fund’s Operational Budgets,” EBS/89/201 (October 17, 1989). At the end of 1989, most of those countries were so small that the Fund decided to exclude them from the operational budget and not use their currencies for receipts. Only Bahrain and Paraguay were included as receipts-only countries. Mauritius and Thailand were excluded from the currency list for lending operations because they had outstanding obligations to the Fund; see “Operational Budget for December 1989-February 1990,” EBS/89/227 (November 29, 1989); and Suppl. 2 (December 15, 1989).

The data cited here are in U.S. dollars. Those shown in Table 15.1 are in the IMF’s unit of account, the SDR. The official data have been converted to dollars at the average monthly exchange rate ($1.302 per SDR in December 1989; $1.373 in December 1999).

The data series “net uncommitted usable resources” was constructed by the Treasurer’s Department by subtracting from usable resources an estimate of commitments under existing arrangements that were likely to be used and a portion of members’ reserve tranche positions that was set as an estimate of minimum working balances. This data series was replaced in 2002 by a slightly different concept, the Fund’s “one-year forward commitment capacity.”

The IMF values the gold on its balance sheet at the price at which the gold was acquired (historical cost). Most of the gold stock was acquired before the Second Amendment of the Articles of Agreement took effect in 1978 and was valued at the then-official price of SDR 35 an ounce. Also see Chapter 13, p. 663.

This transfer mechanism is explained further in Chapter 13, pp. 669–71.

Article V, Section 3(a)(iii), in the original Articles adopted at Bretton Woods in 1944 (Horsefield, 1969, Vol. III, p. 191). In the 1978 amendment to the Articles, this provision was renumbered Section 3(b)(iii), and the annual limit (25 percent of quota) was dropped. In the language of the Articles, indebtedness of 100 percent of quota is described as the Fund holding 200 percent of quota of the member’s currency.

For a brief history of access policies through 1989, see Boughton (2001), pp. 875–84.

See minutes of EBM/92/128 and EBM/92/129, October 28, 1992. At the next meeting, on October 30, Posthumus and Goos tried to amend the proposed decision by inserting language to the effect that the Executive Board expected to lower the access limits in future annual reviews. When that last-ditch effort failed, the decisions on access limits and termination of floating were adopted without further discussion, effective November 4.

Interim Committee communiqué (April 25, 1994), paragraph 7 (Annual Report 1994, p. 201).

Interim Committee communiqué (October 2, 1994), paragraph 3 (Annual Report 1995, p. 207).

Minutes of EBM/94/95 (October 25, 1994), pp. 14 (Schoenberg), 15–16 (Lissakers), 25 (Posthumus), and 30 (Al-Jasser). The decision is on p. 31.

See Boughton (2001), pp. 878–79.

The IMF’s financial year runs from May 1 through April 30. FY93 refers to the financial year ending on April 30, 1993.

For countries that joined the Fund after April 1, 1978, the “norm” for remuneration (the maximum percentage of the reserve tranche on which remuneration would be paid) was set equal to the average norm for existing members on the date that the new member joined.

“‘Norm’ for Remuneration,” EBD/99/76 (June 22, 1999). Before the 1999 quota increase, the unremunerated portion averaged 5.5 percent of quota.

The IMF began investing in interest-bearing securities in 1957 to supplement income from lending and build up a cushion of reserves. It ended that practice in 1972 because reserves were then adequate for the foreseeable future (Boughton, 2001, p. 899n131).

In the 1990s, an average of 37 countries—mostly, but not exclusively, low-income or very small countries—had financial positions in the GRA within 1 percent of being balanced. In most cases, that meant that the country had withdrawn all of its reserve tranche and had no outstanding debts.

In the initial discussions, this concept went by an even less clear name, the “variable uniform norm.” The concept was introduced in “The Cost of Financing the Fund and its Distribution—Review of Burden Sharing,” EBS/94/28 (February 18, 1994), pp. 23–24. It was fleshed out in “Reforming the Financial Structure of the Fund—The Role of the Variable Uniform Norm,” EBS/94/139 (July 1, 1994).

Article V, Section 9(a), specified that the “Fund shall pay remuneration on the amount by which the percentage of quota prescribed … below [as the unremunerated portion] exceeds the Fund’s average daily balances of a member’s currency held in the [GRA],” subject to certain exceptions. For a discussion of the required amendment, see “Reforming the Financing of the Fund’s Operations,” EBS/94/235 (December 6, 1994), p. 12.

See minutes of EBM/95/1 (January 6, 1995) and EBM/95/2 (January 9, 1995).

See “Charges on Large-Scale Use of the Fund’s Resources,” EBS/96/57 (April 2, 1996), and references therein. The proposal was later embodied in the terms of the Supplemental Reserve Facility, adopted in 1997.

For a primer on the nature of the SDR, see the Appendix to this chapter.

The peak for this ratio was 8.4 percent, at the end of the first round of allocations in 1972.

Because SDR obligations are settled quarterly on a financial-year basis, these accounts are presented in Table 15.5 at the end of the financial years 1990 and 2000.

By construction, the net income of the SDR Department equals zero each year. Interest received on net uses and interest paid on net holdings are both assessed at the SDR interest rate and are exactly offsetting. Separately, the Fund levies an assessment on each member country at the end of each financial year, in proportion to the country’s SDR allocation. Those assessments are charged at a rate that is calculated exactly to offset the administrative expense of running the department. Arrears on interest or assessments are offset on the liability side by a temporary increase in GRA holdings.

Technically, the basket continued to comprise five currencies, including the euro as the currency of both France and Germany, until January 1, 2001. See “EMU and the Fund—Valuation of the SDR and the SDR Interest Rate,” SM/98/221 (September 1, 1998). Also see Decisions 11801-(98/101) G/S and 11802-(98/102) G/S, both dated September 21, 1998; in Annual Report 1999, p. 172. The 0.3519 euro component in the 1999 basket was the sum of 0.228 euro (Germany) and 0.1239 euro (France); see “IMF Incorporates the Euro into the SDR Valuation and Interest Rate Baskets,” PR/98/67 (December 31, 1998). The principal effect of this five-currency treatment was to link the basket to the five separate national securities used to compute the SDR interest rate.

Major countries also occasionally drew on their allocations to supplement their use of foreign exchange reserves in settling payments deficits. Notably, in November 1978 the United States used SDR 1.1 billion (48 percent of its allocation) as one element in a package of measures to support the value of the U.S. dollar in exchange markets.

The accumulation of SDR holdings above any country’s allocation was entirely voluntary in the 1990s. Article XIX of the Articles of Agreement permits the IMF to designate a country to provide a usable currency in exchange for SDRs, but this designation mechanism was not used after 1987. Each financial year in the 1990s, the Fund prepared a quarterly series of “designation plans” that listed the countries eligible to be designated in this way, but the lists were purely precautionary.

See, for example, “The Rationale for SDR Allocation Under the Present Articles of Agreement of the International Monetary Fund” by Michael Mussa (the IMF’s Economic Counsellor and Director of Research) in Mussa, Boughton, and Isard (1996), pp. 57–87.

For a cogent argument against the use of SDRs to extend unconditional liquidity, see Wijnholds and Kapteyn (2002), pp. 120–24. In contrast, Clark and Polak (2004) argued that the savings in resource costs from an SDR allocation would bring efficiency gains for the world economy. Lissakers (2006) made a case for the SDR as a stabilizing force in reserve management, especially to protect against a disorderly decline in official demand for U.S. dollar reserves.

“Statement by the Managing Director on the Need for and Modalities of an SDR Allocation,” BUFF/93/16 (April 12, 1993), p. 8.

“Statement by the Managing Director on the Need for and Modalities of an SDR Allocation,” BUFF/93/16 (April 12, 1993), p. 7. The initial redistribution proposal in the Executive Board was made by Jacques de Groote (Belgium) in 1983. Similar suggestions, with varying details, were made by Bruno de Maulde (France) in 1984 and by Arjun K. Sengupta (India) in 1987. For a review of those three overtures, see “Proposals for Post-Allocation Adjustment in the Distribution of SDRs,” SM/86/154 (June 27, 1986). The Japanese authorities proposed a similar scheme at the April 1991 meeting of the Interim Committee. That formulation was similar to Camdessus’s 1993 suggestion, except that industrial countries would have made voluntary transfers to the GRA instead of to a special administered account. The GRA then would have absorbed the risk associated with additional lending. For a summary, see statement by Hiroo Fukui (Japan) at EBM/93/58 (April 19, 1993), p. 14.

Minutes of EBM/93/58 (April 19, 1993).

Minutes of EBM/93/58 (April 19, 1993), pp. 21–22.

The draft report discussed by Executive Directors (see the next footnote) stated that the principles underlying at least a selective allocation were supported “by all Directors, except one,” referring obliquely to strong opposition by Stefan Schoenberg (Germany). The final text stated more diplomatically that there was “nearly unanimous support.”

On the abrogation of the reconstitution requirement, see Boughton (2001), p. 933. On the four proposals submitted to the Interim Committee in 1994, see “Report to the Interim Committee on Access to Fund Resources and an Allocation of Special Drawing Rights,” ICMS/Doc/43/94/10 (September 26, 1994). The evolution of the debate may be traced through the six versions of the draft report, SM/94/205, from August 25 through September 23, 1994; and the minutes of Executive Board meetings from EBM/94/83 on September 12 through EBM/94/90 on September 23.

“G7 Countries to Call for ‘Special’ Allocation of 16 Billion SDRs,” Agence France-Presse, October 1, 1994. Instead of a communiqué, the G7 issued a set of notes under the heading of “media guidance.” That document endorsed the U.S.-U.K. proposal as the basis for agreement on a selective allocation; see “Texts of G7 Briefing Notes,” Reuters News, October 1, 1994. Both documents accessed at

This account is based on interviews with participants.

Transcript of the press conference, UNDOC/94/223 (October 6, 1994).

Minutes of EBM/95/28 and EBM/95/29 (March 27, 1995) and EBM/95/39 (April 12, 1995).

“Interim Committee Meeting—Message from the Chairman,” EBD/95/46 (March 24, 1995).

“The Committee emphasized that such an amendment of the Articles would not in any way affect the Fund’s existing power to allocate SDRs on the basis of a finding of long-term global need to supplement reserves as and when that need arises”; Interim Committee communiqué, PR/96/49 (September 29, 1996). Also see minutes of EBM/96/86 (September 12, 1996).

Minutes of EBM/96/86 (September 12, 1996) and EBM/97/44 (April 23, 1997).

This compromise was accepted by Executive Directors at EBM/97/91 (September 3, 1997), after numerous informal discussions with the Managing Director.

Resolution 52–4 of the Board of Governors may be accessed at

To meet the legal requirement of the five-year deadline, the Board of Governors adopted a resolution in April 1988 asking the Board to continue its work.

Minutes of EB/CQuota/90/13 and EB/CQuota/90/14 (March 19, 1990). Earlier, at EBM/89/154 (November 28, 1989), Dawson had indicated that the U.S. authorities “believe that a quota increase of 35 percent would be sufficient” (p. 3). Before that meeting, the U.S. Director had declined to commit to a position on whether an increase was warranted. On December 20, 1989, Dawson issued a further statement (BUFF/89/236) indicating a willingness to go to 40 percent or “slightly higher if the right conditions are met.”

Interim Committee communiqué, paragraph 7; Annual Report 1990, p. 113.

At the time, the U.S. president—George H.W. Bush—was a Republican, while the Democratic Party controlled both houses of Congress.

“Ninth General Review of Quotas—Consent and Participation Requirement,” EB/CQuota/89/10 (November 8, 1989).

“Ninth General Review of Quotas—Report and Proposed Resolution,” EBD/90/91, Revision 5 (May 23, 1990).

For a complete list of quotas by country before and after the Ninth Review, see the quota table at

The text of the resolution is in “Tenth General Review of Quotas—Report to Board of Governors and Proposed Resolution,” EBD/93/35 (February 26, 1993). It was approved by the Executive Board on March 12 and by Governors on April 14.

For the report of the Executive Board, see “Tenth General Review of Quotas—Report to Board of Governors and Proposed Resolution,” EBD/94/186, Suppl. 1 (December 14, 1994).

“Eleventh General Review of Quotas—Preliminary Quota Calculations,” EB/CQuota/95/1 (August 10, 1995), p. 39.

“Fund Policies—Fund Financial Resources and Assistance—Eleventh General Review of Quotas,” ICMS/DOC/45/95/12 (September 21, 1995), p. 7.

“Eleventh General Review of Quotas—Further Considerations,” EB/CQuota/97/2 (June 10, 1997), p. 2; and minutes of EB/CQuota/MTG/97/1, p. 17.

Minutes of EB/CQuota/MTG/97/2 (August 22, 1997), p. 8; minutes of EBM/97/97 (September 20, 1997), pp. 3, 12, and 13; Interim Committee communiqué (September 21, 1997), paragraph 8 (Annual Report 1998, p. 159). In December 1997, following the spread of the crisis to Indonesia and Korea, Camdessus proposed that the size of the increase be raised to at least 70 percent. That suggestion was strongly opposed by the United States and other major creditors, and Camdessus withdrew it even before it could be discussed by the full Executive Board; see “Statement by the Managing Director,” BUFF/97/118 (December 10, 1997) and minutes of EB/CQuota/MTG/97/5 (December 18, 1997).

Minutes of EB/CQuota/MTG/97/4 (December 5, 1997), p. 11.

Two compromises were also considered. Thomas A. Bernes (Canada) suggested that the participation requirement be set at 85 percent only for a limited time, as had been done in the Ninth Review. Jon Shields (Alternate, United Kingdom) suggested that the requirement could be defined as 70 percent and participation by the largest shareholder. Those proposals gained some support but were eventually discarded; minutes of EB/CQuota/MTG/97/4 (December 5, 1997), pp. 32–34.

Minutes of EB/CQuota/MTG/97/5 (December 18, 1997), pp. 19–20.

For an analysis of the congressional debates, see Locke (2000).

Developments through 1989 are discussed in Boughton (2001), pp. 859–64 and 915–17.

Minutes of EB/CQuota/MTG/88/3 (March 14, 1988), pp. 9–10.

Minutes of EB/CQuota/MTG/90/4 (January 10, 1990). Agreement on this approach was reached at the next meeting; see “Concluding Remarks by the Chairman,” BUFF/90/14 (January 12, 1990).

“Ninth General Review of Quotas—Calculation of Proposed Quotas,” EBD/90/91, Revision 4, Suppl. 1 (May 17, 1990). When the quota increases finally went into effect in December 1992, these shares were all lower, owing to the accretion of new members in the interim. Incidentally, the G7 did not hold the seven largest quotas, either before or after the Ninth Review. Saudi Arabia was in sixth place, with a share of 3.6 percent before the increases and 3.8 percent afterward.

The Fund did not have the legal power to prevent countries with arrears to other parts of the IMF (the SDR Department or administered accounts such as the ESAF Trust) from consenting to quota increases; see statement by Francois Gianviti (General Counsel) at EB/CQuota/MTG/97/5 (December 18, 1997), pp. 3–7.

For the Eleventh Review, the staff used data from each individual country to make its calculations. Because of the limited scope for realignments of quotas in that review, the distribution of quotas within the region changed relatively little. Russia’s and Ukraine’s shares rose the most, and those of Kazakhstan and Turkmenistan fell correspondingly.

The staff’s initial calculations and analysis were presented in “Switzerland—Calculation of Quota,” EB/CM/Switzerland/90/1 (August 10, 1990). The staff acknowledged that the calculations understated the relative size of Switzerland’s reserves because they valued gold at the official price used in the Fund rather than the much higher market prices. Switzerland’s gold reserves were a larger portion of the total than in most other major countries.

“Switzerland—Calculation of Quota,” EB/CM/Switzerland/90/1 (August 10, 1990), p. 40. For specific comparisons with comparator countries, see Table VI.3, p. 45 in that document. In general, the comparisons were derived by computing “calculated quotas” from the standard formulas for each country and then applying the ratio of actual to calculated quota for a country or group of countries to obtain an actual quota for Switzerland.

For a detailed analysis from inside the Swiss delegation, see Kaeser (2004). For an external academic analysis, see Momani (2009).

Minutes of EB/CM/Switzerland/MTG/90/1 (September 11, 1990), p. 2.

At EB/CM/Switzerland/MTG/90/2 (October 4, 1990), Dawson concluded that the quota “should be significantly larger than the SDR 1.45 billion implicit in the initial staff calculations. At the same time, however, a quota as large as SDR 2.1 billion appears to us excessive”; minutes, p. 6. Clark’s “very strong majority” conclusion for the smaller number is on p. 23.

For an account of this contentious exercise from the Swiss perspective, see Kaeser (2004), pp. 88–94.

Minutes of EB/CM/Switzerland/MTG/90/2 (October 4, 1990), p. 10.

Letter dated November 26, 1990, attached as an Annex to minutes of EB/CM/Switzerland/MTG/91/2 (March 4, 1991), p. 8.

Minutes of EB/CM/Switzerland/MTG/91/2 (March 4, 1991).

At the decisive March 1991 committee meeting, Angel Torres, who in the meantime had succeeded Fernández Ordóñez and become Executive Director for Spain, expressed a preference for SDR 1.55 billion but agreed to go along with the consensus. Kaeser (2004) p. 94, suggests a link between this acquiescence and a promise by France, Italy, and Switzerland to support Spain’s campaign to join the G10. That did not bear fruit, but the Spanish case for enlargement of its own IMF quota eventually was more positively received. Spain got above-average increases in both the Ninth and the Eleventh Reviews. At the end of the 1990s, Switzerland and Spain were ranked fourteenth and sixteenth, respectively.

Minutes of EBM/97/97 (September 20, 1997), p. 3.

Interim Committee communique (April 28, 1997), paragraph 8; Annual Report 1997, p. 211.

“Review of Quota Formulas—Establishment of Committee and Terms of Reference,” EBAP/99/63 (May 26, 1999), p. 3.

“Report to the IMF Executive Board of the Quota Formula Review Group,” EBAP/00/52 (May 1, 2000). Accessed at The reason given for rejecting population, which would have increased the shares of developing countries, was that “population does not bear directly on international monetary issues, and of course a country’s relative size (including population) is already captured by GDP” (p. 61).

Statistical Appendix to the committee report, Part B, Table III.1; accessed at

This review covers borrowing by the Fund’s General Department. In addition, as described in Chapter 13, the Fund has administered a number of trust funds, such as the ESAF Trust, financed in part by borrowing. For a comprehensive review of borrowing through the mid-1990s, see “Borrowing by the Fund—A Chronological Review,” EBS/95/122 (July 25, 1995).

Formally, the GAB creditors comprised eight governments—Belgium, Canada, France, Italy, Japan, the Netherlands, the United Kingdom, and the United States—and the central banks of Germany, Sweden, and Switzerland. Initially, the participation of the Swiss National Bank was through an associated agreement. That arrangement was replaced in 1983 by an amendment to the GAB that formally incorporated the bank (the central bank of a country that was still not a member of the IMF) into the GAB as a creditor. The GAB agreement by itself provided a credit line of a little less than SDR 7 billion up to 1983 and SDR 17 billion thereafter. A separate but parallel agreement was signed with the government of Saudi Arabia in 1983 for the Fund to borrow up to SDR 1.5 billion “for the same purposes and in the same circumstances as are prescribed in the GAB.” For the text of the Saudi agreement, see Annual Report 1983, pp. 154–57. For more on the 1983 amendments to the GAB, see Boughton (2001), pp. 894–99.

Preamble of the GAB. For the full original text of the GAB, see Horsefield (1969), Vol. 3, pp. 246–52. For the 1983 revision, see Annual Report 1983, pp. 146–53. Also see “The General Arrangements to Borrow—A Review,” EBS/95/117 (July 14, 1995).

“General Arrangements to Borrow—Proposal for Future Calls for Exchange Transactions Under an Augmentation of the Current Extended Arrangement for the Russian Federation,” EBS/98/123 (July 17, 1998) and Suppl. 1 (July 20).

Minutes of EBM/98/79 (July 20, 1998), pp. 74–75.

Russia completed the repayment of the EFF drawing in January 2005.

The Interim Committee communique of April 1995 (paragraph 5) gave impetus to this move by noting the “need to examine the issues related to borrowing by the Fund from members and, in particular, the role of the [GAB]”; Annual Report 1995, p. 210.

“The Halifax Summit Review of the International Financial Institutions: Background Document,” released by the G7 summit, June 16, 1995; accessed at

See minutes of EBM/95/72 (July 27, 1995), and “Establishment of Supplementary Lines of Credit for the Fund,” EBS/95/129 (August 2, 1995).

See Williams’s report to the Executive Board; minutes of EBM/95/92 (September 25, 1995), pp. 3-7.

See report by Camdessus to the Executive Board in minutes of EBM/96/72 (July 26, 1996), pp. 3–4; “New Arrangements to Borrow—Proposed Decisions,” SM/96/307 (December 31, 1996); and minutes of EBM/97/6 (January 27, 1997).

For the text of the NAB agreement, see

Under the new arrangements, each GAB participant committed to lend up to a specified sum through the NAB. It also continued to commit to lend up to a smaller amount through the GAB, but its total commitment was equal to its NAB commitment. When the Fund activated the GAB, as it did for Russia, calls were proportional to and limited by the GAB commitments. When the Fund activated the NAB, as for Brazil, then calls were limited by the NAB commitments. The two commitments were not additive.

The GAB-associated agreement with Saudi Arabia remained in force, and the cap could be changed without an amendment to the NAB agreement. These features are set out in paragraph 21 of the NAB agreement.

For more-detailed histories and explanations of the SDR, see de Vries (1976, 1985); chapter 18 of Boughton (2001); and the two Appendixes in Mussa, Boughton, and Isard (1996).

    Other Resources Citing This Publication