Revolving fund for the purpose of implementing the WTO Marrakesh Decision relating to the Least-Developed and Net Food-Importing Developing Countries[81]

Executive Summary

Purpose of the Revolving Fund (RF) - The RF would provide a loan facility to address the short-term financing difficulties that least developed countries (LDCs) and net food-importing developing countries (NFIDCs) may experience in financing normal levels of commercial imports of basic foodstuffs. It is argued that the provision of financing would not require justification other than evidence that import bills were excessive.

Size of the RF - The key parameter that determines the size of the RF is the threshold level selected to define excess food import bills. Under the assumption of a 25 percent coverage, i.e. that food import bills in excess of 125 percent of normal levels are considered excessive, an initial capital of US$1 billion would suffice for “normal” years. In addition, extra resources on a stand-by basis would be needed for spike years and these would amount to another US$1 billion. However, these extra injections into the RF would be paid back to the contributors within two to three years from the time of the exceptional borrowing. Hence, the total net contribution to the RF is US$1 billion plus the obligation (in the form of “promissory notes” or in some other similar form) to provide stand-by resources on loan when spikes occur.

Capping drawings from the RF - To insure against the unlikely event that the total resources of the RF (including funds on call) are insufficient to cover loan applications in an exceptional spike year, loans would have to be capped. Thus, the maximum callable obligation of contributors would always remain fixed, i.e. would not exceed US$1 billion during a spike period (which could be of one or multiple years” duration). One capping option is to ration total lending - at the time of the spike - to all potential borrowers in a proportionate manner. Another option would be to introduce an element of special and differential treatment so that borrowing by NFIDCs is capped but not that by LDCs.

Institutional location of the RF - It is envisaged that the RF would be managed by an existing financial institution with the requisite logistics, manpower and experience in managing funds. It would also be desirable that it has experience in funding agricultural development programmes in the prospective beneficiary countries. The International Fund for Agricultural Development (IFAD) is one institution that meets these criteria.

1. Background

In a decision adopted by the WTO General Council on 15 December 2000, the WTO Committee on Agriculture (CoA) was instructed to examine possible means of improving the effectiveness of the implementation of the Marrakesh Ministerial Decision on Measures Concerning the Possible Negative Effects of the Reform Programme on Least-Developed and Net Food-Importing Developing Countries. In this context, a group of 16 developing country Members of WTO submitted on 25 April 2001 a proposal which calls for, inter alia, the establishment of an Inter-Agency Revolving Fund (RF) with two components[82]. These were outlined as follows:

“The first, and variable component of the Fund (to comprise existing and/or new financing facilities as appropriate), will be to ensure that adequate financing at concessional terms is made available to the NFIDCs and LDCs in times of high world market prices”. The second, fixed component of the RF, should “provide technical and financial assistance to LDCs and NFIDCs for specific projects linked to improving agricultural productivity and related infrastructure... over and above the regular bilateral and multilateral activities of donors in this area”.

2. Summary of the FAO informal expert consultation of September 2001

Discussions by the experts during the Rome Consultation in September 2001, and earlier consultations in Geneva in March 2001[84], identified several issues involved in the rationalization of the RF and its purpose, how it could operate in practice and how its resources could best be used. The views expressed are summarized here.

The purpose of the RF would be to provide a loan facility to address short-term financing difficulties eligible countries may experience in financing normal levels of commercial imports of basic foodstuffs. Loans from the Fund would be used to offset the extra costs of abnormal food import bills. Two approaches were considered in borrowing from the RF: ex post, i.e. borrowing takes place after the imports have been effected from a country’s own resources; and ex ante, i.e. borrowing takes place at the time the country imports the needed supplies. Both have implications for the attitude and ability of the country with respect to mobilizing resources to meet its food import needs as well as conditions that may be attached to drawings from the RF.

Where borrowing takes place ex post, in the year following the surge in the import bill, the condition for activating the drawing on the RF would have been met if the extra foreign exchange outlay had already been made to secure the extra food imports. This is the main condition for using the RF’s resources and one that is by definition met - i.e. the actual food import bill increased beyond a certain level. The very existence of the drawing right would be an assurance that governments could secure concessional finance for food imports beyond a certain level. Thus, countries facing either a spike in import prices and/or a surge in the volume of imports could confidently draw on their scarce reserves or borrow on commercial terms knowing that concessional finance for the extra imports would be forthcoming within a short time.

Under the ex ante approach some additional guarantee may be needed to ensure that foodstuffs are actually imported in amounts above what otherwise would have been possible to maintain normal levels of consumption. This may present some difficulties. First, databases for assessing the level of food consumption in the current year are rarely sufficiently up-to-date or adequate. Only cereal consumption levels for all countries are monitored by FAO’s Global Information and Early Warning System (GIEWS), with regular up-datings. For other foodstuffs, data only become available after a lapse of a year or so. Moreover, estimates of food consumption depend not only on data for production and trade, which are reasonably reliable, but also on stock movements and utilization levels of feed, seed and other inputs, which are harder to measure. In other words, much of the data on food consumption in the current year would consist of estimates of often doubtful value for determining whether the level of consumption was maintained by increasing food imports, which in turn led to excessive food import bills.

Second, it would be necessary to establish criteria on how the extra imports would be utilized. Even if it is assumed that consumption for the population as a whole is maintained, there may be doubt as to whether consumption of the poor would actually be protected. It may quite possibly fall while that of the better-off members of society was maintained, but no such telling data are available for all countries in the current year.

In the absence of reliable up-to-date data to monitor current consumption, especially of the weaker segments of the population, the best possible indicator of the efforts made by governments is whether food safety nets for the poor are put in place and steps taken to moderate domestic price increases by such means as drawing on reserve stocks and lowering import duties.

There is a variety of publicly funded social safety net programmes already in place in a large number of developing countries (see Box 1). During periods of price spikes, one or more of these programmes could provide the mechanism through which the government could intervene to protect the consumption of the poor.

Box 1. A typology of government-funded social safety net interventions in developing countries that support food security at the individual or household level
Targeting	Food	Income	Nutrition
Targeted	Emergency food aid	Cash transfer “safety nets’	Weaning foods for low-income families
	Food stamps for the poor	Unemployment benefits
	School feeding for vulnerable groups	Disability grants	Vitamins or mineral supplements for specified groups
	Supplementary feeding for vulnerable groups	Pension for the elderly
Untargeted	General food price subsidy Overvalued exchange rate for imported food	Universal child benefit	Water fluoridization
Self-targeted	Food-for-work projects Price subsidy on “poor foods’	Cash-work projects	Iodization of salt

Besides demonstrating ex post that food import bills increased excessively during the preceding year (or as soon as data become available), countries receiving credit from the RF could be expected to undertake measures to protect the food consumption of the poor and food insecure as an objective of the loan agreement. Recipient countries could be expected to report back to the RF as to how the loan was used - e.g. to replenish food reserve stocks or repay loans used to finance food safety net programmes. The evidence accumulated through such reporting would not only contribute to establishing the credibility of individual countries vis-à-vis the fund and facilitate future borrowing, but would also provide needed feedback to contributors supporting the RF.

Another issue addressed at the September Consultation was possible administrative arrangements for the RF. The meeting agreed that it would not be appropriate to create a separate, new, institution to house the fund. Rather, there was consensus that it should be operated by an existing agency that has experience with managing funds and has the necessary logistic infrastructure and manpower. That would not only ensure that the resources were well managed but also that administrative costs would be kept to a minimum.

The IMF was considered as one possibility, largely because it already operates a Compensatory Financing Facility (CFF) that is available for excess import bills on cereals. However, the International Fund for Agricultural Development (IFAD), which is a specialized international financial institution for lending for agricultural development to low-income countries, was considered as a more promising possibility. Having the RF located in IFAD could be expected to strengthen the complementarity between a short-term financing instrument (the RF) and longer-term financing for food and agriculture development (IFAD). These twin-financing needs are not only stressed in the proposal by the 16 WTO Members, but also emphasized in the Marrakesh Decision. Also, establishing the RF in IFAD would facilitate coordination and complementarity with other Rome-based organizations of the United Nations system that deal with food and agricultural issues.

3. The requisite size of the Revolving Fund and its modalities

Trends in food imports

Figure 1 shows total food import bills of LDCs and NFIDCs over the past 35 years (1965-99), along with the five-year moving average. Two aspects of the trend should be taken into account in considering the features of the RF.

Figure 1. Annual and moving average of food import bills of LDCs and NFIDCs, 1965-99

First, there has been a steady increase in food import bills throughout the period, averaging just over 7 percent per annum. The increase was particularly sharp during 1972-1984 (over 11 percent p.a.) and slowed down thereafter (to about 3 percent p.a.). Since the RF concerns the future, e.g. 2002-2012, some upward adjustment would be required to take into account this trend inasmuch as the absolute level of excess imports would be increasing. Thus, for example, a percentage deviation from the moving average level in, say, 2005 would give a much higher excess import bill than the same percentage deviation in 1990.

Second, the data show that there were three distinct occurrences of “spikes” over the entire 35-year period. The first, during the food crisis of the 1970s, lasted for three years (1973-75), when actual food imports were 42 percent above the MA level. The second was in 1980 and 1981 (31 percent above the MA level) and the third was in 1995 and 1996 (a 20 percent excess). This “stochastic” pattern of food imports has implications for the design of the RF: in the sense that the distribution of “excess food import bills” is not normal but positively skewed, i.e. there are some extreme values (spikes) but the chance of their occurring is small.

Also important for the operation of the RF is the number of “consecutive” spikes. During 1965-99, only once did food import bills remain high for three consecutive years (1973-75), whereas the spike lasted for only two consecutive years in the other two periods.

Definition of excess food import bills under various threshold levels

While assumptions can be made about future food import bills and their pattern over the first 10 years over which the RF may be expected to operate, much of the analysis necessarily has to be based on past experience. In view of the marked shift in the trend of food imports in the early 1980s (Figure 1), the analysis below is based on the data for 1985-99. As noted above, some adjustment to the results would have to be made eventually to reflect the future period.

The size of the RF depends on the mean level of excess food imports, their frequency distribution and their sequence of occurrence. Thus, the starting point of the analysis is to define excess import bills.[85]

For any year t, the excess import bill for a given country is the actual level of imports minus (1+a) * MAt, where:

MAt is the moving average value for year (t) calculated as the mean of actual import bills for that year and the previous four years (e.g. it is 1991-95 for the year 1995);

a is a threshold level and the term (1+a) * MAt measures the food import bill that is not considered excessive and so is covered by the country’s own resources. For example, if a is 0.2, only imports more than 20 percent above the MAt level are considered excessive.

Figure 2. Excess food import bills by threshold levels (annual average 1985 -99)

The 1985-99 annual average level of the excess food import bill ranges from US$730 million for a threshold level of 10 percent to US$160 million for one of 35 percent. Thus, the threshold level is the key parameter determining the size of excess food import bills and consequently the size of the RF.

Table 1. Excess food import bills for various threshold levels assumed
(value in US$ million)

Threshold level (excess over normal imports - %)	1985-99 annual average	Spike years		1985-1999 average (excluding the two spike years)
		1995[86]	1996
10	729	2 338	1 780	524
15	558	1 903	1 328	395
20	416	1 482	916	296
25	301	1 119	553	219
30	217	835	354[87]	161
35	159	658	270[88]	122

Figure 3 shows the excess food import bills on the basis of the 25 percent threshold level. They were lower than US$400 million during the first 10 years (except in 1992 when they were slightly higher) followed by a spike in 1995 that also lingered into 1996, and were again below US$400 million in the last three years.

Figure 3. Excess food import bills in 1985-1989 on the basis of a 25 percent threshold level (US$ million)

The size of the Revolving Fund

Borrowing and repayment assumptions

The basic characteristic of the RF, which is essential to ensure that the fund remains viable over a sufficiently long period, is that it should be self-financing. Consequently, borrowing countries must assume the obligation to repay their loans.

Under normal conditions, the amount borrowed is assumed to cover 100 percent of a country’s excess import bill. However, in periods of inadequate resources in the RF, the level of borrowing would be capped.

Interest charged on the loan is assumed to be 0.75 percent per annum. Alternatively, different concessional interest rates could be applicable to different eligible countries, depending, for example, on their GNP per capita, or based on the World Bank’s IDA classification. Annexes 2 and 3 show the various lending terms of IMF and IFAD.

The repayment period is assumed to be two years. Thus, an amount borrowed in 2005 would be repaid in full in 2007, together with the interest accrued.

At any year (t), the fund balance, i.e. the balance after disbursements (lending) and receipts (repayment) would be equal to:

Balance (t) = Balance (t-1) - disbursements (t) + receipts (t)

Estimates based on experience of excess import bills in 1985-99

The size and the sequence of the excess food import bills depicted in Figure 3 provide the main basis for estimating the size of the RF. Given the stochastic nature of their distribution, it is inevitable that the excess import bills, and hence the funding needs, would be characterized by several years of “normal” or low levels of borrowing and some years of spikes.

Figure 4 shows the hypothetical sequence of the fund balance for the period 1985-99 (computed as explained above), on the assumption that the RF had been established in 1984, with an initial endowment of US$600 million. On this basis, the fund would have covered the needs until 1992, after which it would have exhausted its resources and, in consequence, would have stopped operating in 1993[89]. Obviously, fresh injections would have been required for the RF to continue operating.

Figure 4. Hypothetical fund balance in 1985-99 with an initial capital of $600 million and 25 percent threshold

Figure 5 shows what would have happened if such fresh injections had been made in order to avoid funds running out. This case assumes also that positive balances of the RF, as they accumulated, were drawn on to pay back the contributors of such extra injections. Under this “stand-by arrangement”, three fresh injections would have been necessary: $100 million in 1993, $600 million in 1995 and $400 million in 1996. Contributors would have been paid back this total of $1,100 million in two instalments - $800 million in 1997 and $300 million in 1998.

Figure 5. Fund balance in 1985-99 with fresh injections in spike years[90]

It should be reiterated that under the “stand-by arrangement”, the initial capital of US$600 million is the only contribution expected by contributors. In addition, it is assumed that contributors would have been prepared to make available extra resources on a stand-by basis for spike years. However, these extra injections would have had to be paid back. Hence under this modality of operation, the total net contribution to the RF would have been US$600 million plus the obligation (in the form of “promissory notes” or in some similar form) to provide stand-by resources on loan when spikes occur.

Alternative estimates, assuming other sequencing

Spikes in the very early years of the RF. What would have been the consequences if the spikes such as occurred in 1995 and 1996 had occurred earlier on, say in 1985 and 1986? Such a change in the sequence of shortfalls would not have altered the mean level of the RF, but would have had implications for the timing of fresh injections (and pay-backs). On this hypothesis, fresh injections of US$600 million would have been required in 1985 and US$500 million in 1986. As the borrowings in 1985 would have been repaid in 1987, the RF would quickly have accumulated a large balance, making it possible to repay US$800 million in 1987 and US$200 million in 1988. The remaining US$100 million could have been re-paid only in 1991 in order to keep the RF operating. Thus, basically, the total size of the RF (US$600 million initial capital and US$1 100 million stand-by) would not be affected nor would the nature of the RF change.

Spikes occurring in the middle of the sequence or other years in a random fashion. Neither the total resources of the RF nor the nature of its operation would need to change if the spikes were to occur in the middle of the sequence. This conclusion also holds when the sequences of excess food import bills are generated through Monte Carlo random sampling.

Spikes that last for three consecutive years (as in 1973-75). The food crisis of 1973-75 was the worst for the past four decades. As seen above (Figure 1), the spike was not only sharp but also of unprecedented duration. What would be the consequences for the RF if such a situation were to recur? Simulations show that there would be a need not only for a somewhat larger initial RF (because of an increase in the average annual excess food import bill) but also for bigger injections (over the three consecutive years), but that would not alter the nature of the RF. Moreover, the injection implicit in the third year is small as borrowings during the first year are repaid.

Adjusting the size of the RF for operations in the future

Estimates of the size of the RF discussed above were based on experience in 1985-99, but food imports of the LDCs and NFIDCs have been trending upwards, as was noted above. There is an upward trend in excess food import bills since these are measured relative to a moving average - which rises as actual bills increase.[91] In consequence, some upward adjustment would be required for the size for the estimated initial capital (calculated for the period 1985-99) so as to adequately cover future excess food import bills. Projected trends of food import bills, covering a 10-year future period (e.g. starting in 2002), imply that the initial capital requirement for the 1985-99 period would have to be scaled up by a factor of about 1.6. Thus, if the RF were to be set up with a time horizon of 10 years, the necessary initial capital endowment would amount to about US$1 billion.

As far as the stand-by resources are concerned, it is difficult to speculate on the nature of spikes in the future. What has been established from past data is that there is no trend in the severity of production shortfalls leading to ever-increasing price spikes[92]. There could be a repeat of the 1973-75 situation or things could get worse, although the probability of that happening should be small. In any event, would it be justified to provide sufficient funds to cover such an eventuality? Common sense would suggest that extreme situations require extreme responses - a standard instrument is unlikely to be effective for such situations. Moreover, as noted in Section II, an open-ended commitment is unlikely to be acceptable to contributors and therefore the stand-by component of the RF would need to be capped; an amount of US$1 billion was suggested. Thus, the maximum callable obligation of contributors should remain fixed, i.e. would not exceed US$1 billion during a spike period (which could be of one or multiple years” duration). One capping option is to ration total lending to all potential borrowers in a proportionate manner. Another option would be to introduce an element of special and differential treatment so that borrowing by NFIDCs is capped but not that by LDCs.

Adjustment would also be required if a threshold level other than the 25 percent assumed here were chosen. Figure 2 provides a rough guide to the additional adjustments that would be necessary to the initial capital endowment.

With respect to the stand-by component of the RF to cover spikes, it was pointed out above that there is no way of determining the magnitude of such spikes in the future and that in any event there are practical reasons why a cap would need to be set. Consequently, it would also be necessary to cap borrowings from the RF so that its long-term viability is assured.

Annex 1 - Revolving Fund for the Purpose of Implementing the WTO Marrakesh Decision: Informal Expert Consultation held in Rome on 17 September 2001

List of Participants

IMF

Grant B. Taplin
Special Representative to
the WTO
and Assistant Director,
Office in Geneva.

IFAD

Kwabena Offei Dei
Assistant Controller,
Accounting

Robert Cassani
Coordinator, Resource
Mobilization

Carol Upham
Programme Officer
Supplementary Funds

WFP

Robin Jackson
Senior Policy Advisor

European Commission

André Libens
Principal Administrator
(DG DEV)

FAO

Harmon Thomas
Chief, Commodity
Policy and Projections
Service

Ramesh Sharma
Senior Economist,
Commodity Policy and
Projections Service

Panos Konandreas,
Senior Liaison Officer,
FAO Liaison Office in Geneva

Other Experts

Jim Greenfield
Commodities and Trade Expert,
Rome, Italy

Alberto Valdes
Consultant, Santiago,
Chile

Annex 2 - General Terms of IMF Financial Assistance

Facility or policy	Charges	Repurchase terms
		Obligation schedule (years)	Expectation schedule (years)	Instalments
Stand-by Arrangement	Basic rate plus surcharge[93]	3¼-5	2¼-4	Quarterly
Extended Fund Facility	Basic rate plus surcharge[94]	4½-10	4½-7	Semiannual
Compensatory Financing Facility	Basic rate	3¼-5	2¼-4	Quarterly
Emergency Assistance	Basic rate	3¼-5	Not applicable	Quarterly
Supplemental Reserve Facility	Basic rate plus 300-500 basis points	2-2½	1-1½	Semiannual
Contingent Credit Lines	Basic rate plus 150-350 basis points	2-2 ½	1-1½	Semiannual
Poverty Reduction and Growth Facility	0.5 percent per annum	5½-10	Not applicable	Semiannual
Memorandum items:
Service charge	0.5 percent
Commitment charge	25 basis points on committed amounts of up to 100 percent of quota, 10 basis points thereafter

Note: IMF’s market-related interest rate, known as the "rate of charge", is based on the SDR interest rate which is revised weekly to take account of changes in short-term interest rates in the major international money markets. The rate was about 4 percent in September 2001.

Source: IMF (available on the Internet at
http://www.imf.org/external/np/tre/lend/terms.htm)

Annex 3 - Lending Terms and Conditions of IFAD

Highly concessional terms (applicable to countries with GNP per capita of US$1 063 or less) Free of interest Service charge of three-fourths of one percent (0.75%) per annum Maturity period of 40 years, including grace period of 10 years. Intermediate terms (applicable to countries with GNP per capita of between US$1 064 and US$1 724) Interest per annum equivalent to 50% of variable reference interest rate (as determined annually by the IFAD Executive Board) Maturity period of 20 years, including a grace period of five years. Ordinary terms (applicable to countries with GNP per capita of US$1 725 or above) Interest rate per annum equivalent to 100% of the variable reference interest rate (as determined annually by IFAD Executive Board) Maturity period of 15 to 18 years, including a grace period of three years.

Note: For the purposes of implementing the HIPC Debt Initiative, the IFAD Executive Board may amend the terms upon which an approved loan is provided to a country. In determining the grace period, the maturity date and the amount of each instalment for the repayment of loans, the Executive Board shall take into account an assessment of a country’s debt sustainability produced under the HIPC Debt Initiative.

Source: IFAD Lending Policies and Criteria (available on the Internet at
http://www.ifad.org/pub/basic/lending/e/02polcri.pdf)