Public Information Notice: IMF Board Agrees on Changes to Fund Financial Facilities
September 18, 2000

Review of Fund Facilities—Follow Up

Prepared by the Policy Development and Review Department and the Treasurer’s Department in consultation with other Departments

August 31, 2000

Executive Summary

This paper follows up on two earlier Board discussions about the financial terms—charges and repurchase periods—and eligibility requirements of “core” Fund facilities, namely, stand-by arrangements, the Extended Fund Facility (EFF), the Supplemental Reserve Facility (SRF), and the Contingent Credit Lines (CCL). The paper also considers further certain features of post-program monitoring.

Terms of stand-by arrangements and the EFF

With the globalization of capital markets, concerns have been voiced that some members may be making excessive use of Fund resources—either unduly long or unduly large use. Although concerns about long and large use are interrelated, the following aspects of undue use could be distinguished:

• Prolonged use: The long use of Fund resources over a series of arrangements does not seem to be a very widespread problem, with only a few members making relatively intensive use of Fund resources over many years.

• Long use following the resolution of a balance of payments problem: There do appear to be many members whose balance of payments problems are resolved quickly following an arrangement, yet who make long use of Fund resources—that is, do not repay ahead of the repurchase periods of the resources provided, as envisaged in the Articles to enhance the revolving character of Fund resources.

• Large use: There is also a risk that members will make unduly large use of the Fund’s resources, even though decisions on access levels are within the Fund’s control, because there are situations where initial access and eventual need are not in step.

There are a number of instruments the Fund could use to try to counter undue use of its resources:

• The early repurchase policy (ERP) could, in principle, address the problem of long use following the resolution of a balance of payments problem. Under the ERP, early repurchases can be requested by the Fund at any time, based on an evaluation of the strength of the member’s balance of payments and reserve position. However, the ERP has not functioned well, and the staff’s assessment is that remedying the flaws in the ERP would create a system closely resembling the time-based repurchase expectations discussed below.

• The introduction of schedules of repurchase expectations (REs) could be another instrument to counter long use following the resolution of a balance of payments problem. Expectations of repurchases could be established uniformly at set dates ahead of the obligation dates, but members could request a waiver of the expectations if their situation remains weak. This approach would differ from the ERP in three main ways, and would address the three problems the staff believes have made the ERP less than fully effective:

• Instead of the initiative to request early repurchases being with the Fund, the normal course of events would be that the member repays early.

• The amounts to be repurchased would reflect an acceleration of repurchases due, with no attempt to fine-tune the amounts to the strength of a member’s external position (a difficult task, which the ERP attempts by linking these amounts to the level of and change in gross reserves).

• Discussions on REs would be separated from decisions on the financial transactions plan and would take place in the context of country-specific Board discussions.

• A uniform increase in the rate of charge could address both long and large use, but may not be the best way of ensuring that the cost of Fund lending provides the right incentives at the margin.

• Graduation of charges with the time elapsed within an arrangement could also address the problem of unduly long use following the resolution of a balance of payments problem. It would be a weak instrument with which to discourage prolonged use, since prolonged use involves a series of arrangements, each of which would start out at the basic rate of charge. Graduation within an arrangement could start from the date of each purchase, or from the date of the first purchase under an arrangement, although the latter would require major changes to the structure of facilities (namely, making stand-by arrangements the only means of accessing the credit tranches). Graduation with time elapsed within an arrangement would have both advantages and disadvantages:

• It would rely on a price incentive to encourage advance repurchases, and hence would give the member more flexibility than a system based on quantitative rules.

• It would apply to members with weak as well as strong external positions, which might be considered either a disadvantage or, to the extent that external weakness reflects inappropriate policies, a useful incentive for members to strengthen policies.

• It could be less successful than (effectively implemented) repurchase expectations in inducing early repurchases, as—absent a fairly steep schedule of surcharges—some members which the Fund would have found strong enough to repurchase early might choose not to voluntarily effect repurchases.

• It would provide an incentive to “refinance” older purchases, at the least by requesting new arrangements (and, in those variants of the system that could be implemented without major changes to the structure of facilities, even by making new purchases under existing arrangements).

• It would result in advance repurchases made at the member’s discretion, thereby triggering prepayment clauses in other creditors’ operations.

• Graduation of charges with time across arrangements could serve to address prolonged use, but could quickly become onerous and/or complex. Policies on conditionality and access could address prolonged use more directly.

• Graduation of charges with the level of credit outstanding could help discourage unduly large use of Fund resources—including prolonged use of large amounts, and long use of large amounts following the resolution of a balance of payments problem.

Any changes in the terms of Fund financing—introduction of REs, or graduation of charges—should apply only to new purchases. Credit outstanding at the time of adoption of the policy changes would not be affected.

The impact on Fund income and on members of the changes under consideration is extremely hard to predict, in large part because members’ response to the policy changes cannot be known in advance. Illustrative calculations suggest that the impact on Fund income could range from strongly negative to strongly positive, and that the effects on individual members would be modest.

The Contingent Credit Lines

There is a strong case for reducing the Fund’s discretion to require policy changes in the CCL activation review—in effect separating the “activation” review (which would be held expeditiously and would make available a predetermined, large amount of resources) from a “post-activation” review (which would be held immediately following in-depth discussions between the member and the staff regarding the possible need for policy adjustments). As regards whether a crisis is traceable to contagion, the staff believes that it would be useful to include, in a member’s request for a commitment of CCL resources, an illustrative list of situations in which the CCL might be activated. With regard to the other conditions for completion of the activation review—the existence of a crisis, and the member’s adherence to its program to date—the staff sees little opportunity to reduce the scope for judgment on the Fund’s part, but believes that these conditions introduce only limited uncertainty for the member.

There remain questions as to how a member might exit from the CCL—because of policy slippages or other reasons—without risking disturbing markets, and, correspondingly, whether the Fund might be reluctant to force a member to exit in the event of slippages. In this regard, precautionary arrangements are an attractive avenue for exit, but a member cannot be compelled to request an arrangement. A “graduated response” in the form, for instance, of a reduction in the amount available on completion of the activation review would bring with it the risk of a multi-tiered CCL, and of a loss of credibility. At the same time, exit may itself take place gradually rather than suddenly, as the prospects for completion of a mid-term review or approval of a new commitment under the CCL gradually dim with time.

There is broad agreement that the rate of charge on the CCL should be reduced, in particular so that it is below the rate on the SRF. It may be reasonable to halve the initial surcharge to 150 basis points (that is, 150 basis points below the initial SRF surcharge), but to keep its graduation with time unchanged, until the surcharge reaches a level of 350 basis points (again, 150 basis points below the ceiling on the SRF surcharge).

There is also broad agreement that the commitment fee should be reduced or eliminated. The mechanism for such a reduction would be a change in the Fund’s Rules and Regulations, that would make the commitment fee apply—for all arrangements—to the amounts scheduled for the next twelve months. Such a change would virtually eliminate the commitment fee under the CCL.

The Extended Fund Facility

There is widespread agreement that greater attention should be given to ensuring that access to the EFF is granted only in cases where the member’s balance of payments difficulties are expected to be relatively long-term. As a result, the EFF would probably not be used by members with well-established access to capital markets. There is also widespread agreement that use of the EFF on a precautionary basis should be exceptional, as there will rarely if ever be a high likelihood that a member without an immediate balance of payments problem would develop longer-term balance of payments difficulties.

Post-program monitoring

Stronger post-program monitoring (PPM) would provide for closer monitoring of the circumstances and policies of members whose arrangements have ended but that continue to have Fund resources outstanding. Such monitoring would be regular (at six-monthly intervals) and would involve staff discussions with the member, reported by means of a staff paper and discussed by the Board. Such PPM could be based on the existing consultation clauses in all arrangements.

It would be possible to adopt stronger procedures for PPM without adopting automatic criteria for it. PPM could be based on the recommendation of the Managing Director, and the agreement of the Board, on the occasion of a member’s Article IV consultation or during informal country matters sessions.

At the same time, there may be a case for automatic application of PPM, or for a presumption of application, if certain criteria are met (for instance, relating to the level of Fund credit outstanding in relation to quota), as ad hoc procedures tend to be little used.

I.   Introduction

1.   The Executive Board has been conducting a comprehensive review of the Fund’s nonconcessional financing facilities, with discussions in March and July 2000.¹

2.   Notwithstanding a rather wide consensus on the nature of the Fund’s financing role in a changing global environment, the July discussion brought to the surface considerable differences of opinion as to specific changes that may be required in the Fund’s core facilities.² While there was broad agreement on the general direction of the changes needed to the Contingent Credit Lines (CCL) and on the eligibility requirements under the Extended Fund Facility (EFF), there were much greater divergences of view as regards possible changes to repurchase periods and rates of charge (for stand-by arrangements and the EFF) and as regards post-program monitoring.

3.   The present paper seeks to follow up on these issues. In large part, it represents a stocktaking of the analysis presented in RoF I and RoF II, in the light of the views expressed by Directors at the last meeting. Its discussion should help clarify what the next steps in the process should be, including if possible a timeframe for moving to decisions.

4.   The paper is structured as follows: Chapter II considers repurchase periods and charges; Chapter III, the CCL; Chapter IV, the eligibility requirements under the EFF; and Chapter V, post-program monitoring. Chapter VI suggests issues for discussion.

II.   Repurchase Periods and Charges

5.   The Board has had a lively discussion about whether the terms of the Fund’s core facilities—stand-by arrangements and the EFF in particular—remain appropriate in an environment of increasingly globalized capital markets. With many—though still by no means all—members able to access capital markets, questions arise as to whether repurchase periods and charges are well-tailored to encourage members to seek to use, restore, or establish such access to private financing, or whether they encourage members to rely excessively on Fund resources.

6.   An important theme of this discussion has been “whether there is a problem.” In particular, with how much concern should the Fund be viewing long or large use of its resources? And are members indeed making unduly long or large use of the Fund’s resources?

7.   In order to address these questions, it is useful to distinguish three aspects of the possibility that members may be making excessive use of Fund resources. Two relate to the possibility of unduly long use of Fund resources—where “use” can be interpreted either as the drawing of Fund resources or as the nonrepayment of Fund resources at the earliest opportunity. The third relates to the possibility of unduly large use of Fund resources. These three aspects are, of course, interrelated: the Fund is likely to be significantly more concerned about long use of a large amount of its resources than of a small amount. Nonetheless, the three aspects provide a useful broad taxonomy of the issue, and the following paragraphs consider each of them separately, examining in turn the appropriate degree of Fund concern and the likely prevalence of the phenomenon.

8.   The first aspect of unduly long use of Fund resources may arise when members draw Fund resources over a series of arrangements. This phenomenon is traditionally known in the Fund as “prolonged use:”³

• There is agreement that the Fund should be concerned about prolonged use, to the extent that it reflects a lack of progress toward external viability under successive arrangements aimed at broadly the same balance of payments problem (as opposed to repeated shocks). Such arrangements do not bring the benefits expected from Fund arrangements in general, either for the member or for the world at large.

• A priori, there would seem to be both incentives and disincentives for members to engage in prolonged use, but also scope for the Fund to control it. From the member’s point of view, on the one hand, the relatively low rate of charge represents an incentive to engage in prolonged use; on the other hand, because prolonged use involves a series of arrangements, members will also be taking into account other potential costs—including political and administrative—of having an arrangement from the Fund. As regards the Fund, it should, in principle, be able to limit prolonged use through appropriate policies on conditionality and access, which would ensure that members use Fund resources only in support of appropriate adjustment efforts.

• In practice, prolonged use does not appear to be a widespread problem in the GRA. RoF I examined the evidence in this regard, and found that there are only four members that (i) have made use of Fund resources in the GRA for more than 5 of the last 15 years, and (ii) continue to have substantial outstanding use of Fund resources (with credit outstanding above 100 percent of quota). There are three more members that would meet criterion (i), but whose credit outstanding is around 50 percent of quota or less—cases of a type the Fund has traditionally considered to be less cause for concern.^{4 5}

9.   The second aspect of unduly long use of Fund resources arises when a member has credit outstanding for an unduly long time following the resolution of a balance of payments problem. Questions about this phenomenon revolve around whether members might be able to repay sooner than is specified in the 5- or 10-year repurchase schedules of stand-by and extended arrangements, respectively:

• Long use is a source of concern when a member would be able to repay. In order to preserve the revolving nature of its resources, the Fund relies primarily on obligatory repurchase schedules to ensure timely repayment of its resources. However, the Articles also require a member to make repurchases as soon as it is able to do so.⁶

• A priori arguments suggest that unduly long use following resolution of a balance of payment problem addressed by an arrangement could be a problem. The member’s incentives in this case are clearer than in the case of prolonged use: no further arrangement—potentially with associated costs—is involved, and the relatively low rate of charge constitutes an incentive to retain the Fund’s resources to maturity (particularly if these resources substitute for higher-cost resources).

• Empirical evidence suggests that unduly long use following an arrangement may be widespread. Analysis of the duration of members’ balance of payments needs in RoF II, while subject to a number of caveats, yielded striking results: for almost half of members with nonprecautionary arrangements, the balance of payments need is eliminated within one year; and the number of members with shorter-than-projected balance of payments needs (60 percent) is far greater than the number with longer-than-projected needs (15 percent).⁷ At the same time, few members voluntarily make repurchases ahead of the scheduled dates (see Table 1 below).

10.   Finally, the possibility exists that members are making unduly large use of the Fund’s resources:

• This is the area where views differ most as to the appropriate degree of Fund concern. On the one hand, the Fund should encourage members to have arrangements, including at a relatively early stage of their difficulties, because arrangements bring substantial benefits in terms of discipline in policy-making and externalities for the rest of the world. On the other hand, the Fund has a collective interest in ensuring that its resources are used efficiently. The membership as a whole sets limits on the size of the Fund, and thus on the resources it can make available to borrowing members. Within this framework, the Fund should be focusing its resources on those members where, absent financing, “national and international prosperity” (Article I) is most at risk—and when a member has access to capital markets, as a growing number of members have, the alternative to Fund financing (worse borrowing terms for the member) is much less destructive of prosperity than when it does not.⁸ Moreover, large use may also accompany prolonged use, as use of Fund resources over a series of arrangements tends to raise the level of credit outstanding.⁹

• It is not clear that unduly large use is a widespread phenomenon, but it might still represent an important risk. Clearly, decisions on access levels are under the Fund’s control, based on the member’s balance of payments need and its adjustment effort. Access levels are generally well within the access limits, and represent only a relatively small share of members’ financing needs;¹⁰ and, as noted above, prolonged use of relatively large amounts of Fund resources is a fairly rare phenomenon. However, once access has been approved, the Fund has only relatively weak protection against purchases made in the absence of need,¹¹ and there are particular situations in which approved access and need are not in step. Two such situations stand out. First, approved access is based on an initial assessment of balance of payments need, and, as noted in RoF II and recalled above, need often turns out to be of shorter duration than expected—while it is relatively rare for members to turn their arrangements precautionary (unless this was expected from the outset).¹² Second, there is by definition little link between access and need when an arrangement is approved in the absence of a balance of payments need, and when the member intends to treat the arrangement as precautionary. Although members with arrangements that they have indicated they will treat as precautionary do not typically resort to purchases,¹³ the risk associated with purchases under precautionary arrangements will increase as the Fund experiments with larger precautionary arrangements than in the past.¹⁴

11.   If long and/or large use are problems to be addressed, then the question arises which instruments could best be used to address them. A number of instruments/options affecting the terms of Fund financing are candidates in this respect:

• The Fund’s early repurchase policy is an explicit attempt to address the problem of long use following the resolution of a balance of payments problem. The Guidelines for Early Repurchase clarify how members can meet the expectation provisions of Article V, Section 7(b) (see Box 1 below).¹⁵ Under this policy, early repurchase expectations can be formed by the Fund at any time while Fund resources are outstanding, based on an assessment of the strength of the member’s balance of payments and reserve position. The policy relies on judgments by the Executive Board, based on established criteria and guidelines, and tries to tailor the amount of repurchases to the strength of the member’s external position by linking them to the level and recent movements in the member’s gross reserves.

• The introduction of schedules of repurchase expectations (REs) to complement repurchase obligations would be another way of addressing the problem of long use following the resolution of a balance of payments problem. Under this proposal, which was discussed in RoF II and in which the Board expressed interest in the July meeting, time-based expectations of repurchase would be established ahead of the original obligation dates (as in the SRF and CCL). A member would be expected to adhere to the earlier schedule, but could request the Board to grant it a waiver, so that it could adhere instead to the obligations schedule. This approach would involve judgment by the Executive Board only when a member requested a waiver, and would not attempt to fine-tune the amount of repurchases, relying instead on a binary judgment as to whether the member should be able to meet a set schedule of expectations or not.¹⁶

• A uniform increase in the rate of charge could address both long and large use, but may not be the best way of ensuring that the cost of Fund lending provides the right incentives at the margin. A uniform increase would bear equally on all members at all times, when the Fund’s concern may be concentrated on members that have made long or large use of Fund resources.¹⁷ In the March and July meetings, there was little interest from Directors in a uniform increase in the rate of charge, and the present paper does not pursue this option further.

• Graduation of the rate of charge with the time elapsed within an arrangement could serve to address long use following the resolution of a balance of payments problem.¹⁸ It would ensure, in a flexible manner, that members had increasing incentives to repay as their external position improves over time. The Fund would thus be relying on incentives rather than rules to encourage members to repurchase. Graduation of charges with time under a single arrangement would, however, be a weak instrument with which to discourage prolonged use. Prolonged use stems from a series of arrangements, and under a surcharge based on time elapsed all new arrangements would start out at a lower rate of charge.

• Graduation of charges with time across arrangements could serve to address prolonged use, by strengthening members’ incentives to make progress toward external viability. As noted in RoF II, there are a number of practical problems with the design of graduation of charges across arrangements, which quickly becomes onerous and/or complex.¹⁹ Policies on conditionality and access could also be used to target (more directly) progress toward external viability. In the July meeting, only a few Directors expressed interest in graduation of charges across arrangements, but they were open to the possibility of graduation within an arrangement, and the present paper does not pursue the option of graduation across arrangements further.

• Graduation of charges with the amount of Fund credit outstanding could serve to address large use—including prolonged use of large amounts and long use of large amounts following the resolution of a balance of payments problem. While access policy would remain the Fund’s main instrument for limiting large use, graduation of charges with amounts would provide disincentives to draw large amounts under existing arrangements, as well as incentives to repay large amounts early. Moreover, graduation of charges with amounts would provide some disincentives for prolonged use, as it would make the substitution of further Fund financing for adjustment more costly in cases where large amounts of Fund resources have already been drawn under previous arrangements.

12.   The next two sections discuss, respectively, the instruments available to address long use and large use. The sections also evaluate the impact of each of these instruments on individual members and on the Fund’s income. As noted above, under “long use” the focus is on long use following resolution of a balance of payments problem, rather than prolonged use over a series of arrangements (which is instead captured, to some extent, under large use).

13.   Any changes to the facilities would be forward looking, and members should not bear an additional financial burden for decisions made to date.²⁰ In this light, it would seem reasonable to apply possible changes in charges and repurchase expectations only to purchases that take place after the effective date of the decision. All such purchases, including those pursuant to arrangements approved prior to that date, would be subject to those changes. In any event, the membership as a whole would not face the full effect of the changes for a number of years, and the full impact on the Fund’s net income would not be apparent for some time.

14.   Like the previous papers, the present paper treats stand-by arrangements as one of the Fund’s core facilities (see footnote 2), even though SBAs are not strictly speaking a facility. SBAs are a means for members to access the credit tranches, and can also be used to deliver other types of resources (e.g., SRF or CCL resources, which are provided under stand-by arrangements in combination with credit tranche resources). The paper uses the term SBA, for presentational simplicity, to denote a stand-by arrangement in the upper credit tranches. Any changes that the Fund wishes to make to SBAs (e.g., introduction of repurchase expectations or graduation of charges) would actually need to be made to the credit tranches.

15.   The non-core facilities in the credit tranches would be automatically affected by the changes discussed below—but, if it were desired, it would be possible in many cases to prevent such consequential changes. Any changes introduced with a view mainly to affecting stand-by arrangements would affect other policies in the credit tranches—namely, the first credit tranche and emergency assistance for natural disasters and post-conflict cases. In some cases, this would not make a difference in practice; for instance, graduation of charges with the level of credit outstanding would not affect the first credit tranche, as long as graduation did not start below 25 percent of quota.²¹ In the case of the two forms of emergency assistance, any changes could be prevented by turning these into special policies (which would require a 70-85 percent majority of the total voting power, depending on the features of the special policies).²² The Compensatory Financing Facility (CFF) is not part of the credit tranches, and would not be automatically affected by most of the changes discussed in this paper. However, should the Fund make changes to the terms of the credit tranches and the EFF, it might want to consider making similar changes to the CFF, so as not to bias members toward use of the latter facility.

A.   Addressing Long Use

16.   As noted above, there are three instruments that can address the problem of long use: the early repurchase policy, time-based repurchase expectations, and graduation of charges with time. The present section will discuss first the early repurchase policy and then, together, time-based repurchase expectations and graduation of charges with time.

(i)   The early repurchase policy

17.   Under existing policies, the Fund’s primary instrument to address long use of Fund resources following the resolution of a balance of payments problem is the early repurchase policy (ERP), and it is natural to ask why, if there is a problem of long use, this policy does not achieve its objectives. The answer this section will give is that the ERP is flawed in ways that, if remedied, would make it closely resemble time-based repurchase expectations. The functioning of the ERP is described in Box 1 (which reproduces Annex III of RoF I).

18.   Experience shows that the ERP has not worked well, especially recently. Any measure of the overall effectiveness of the ERP must take at least some account of voluntary advance repurchases, since members may be encouraged to make advance repurchases by the prospect that early repurchase expectations could be imposed (Table 1).²³ However, even taken together, the number of members making early or (voluntary) advance repurchases is consistently small relative to the number of members with Fund credit outstanding, and is certainly not commensurate with the substantial proportion of members found to experience a rapid turnaround in their balance of payments following an arrangement (as examined in RoF II). Moreover, the number of members making early and advance repurchases has fallen since the 1980s (as has the number of early and advance repurchases), with the decline in early repurchases especially marked; no early repurchase expectation has been imposed under the ERP since 1994.²⁴

Total number of transactions

Total amount (in millions of SDRs)

Number of members per year in percent of number of members with credit outstanding at the end of previous year (period average)

Amount per year in percent of credit outstanding at the end of the previous year (period average)

1/ Excluding SRF resources. Historical data on early repurchases are maintained on an annual basis (June-May) that differs slightly from the Fund’s financial year (May-April).
2/ Excluding 11 transactions in which an early repurchase expectation was met fully by maturing obligations and advance repurchases.
3/ Advance repurchases include about SDR 2.6 billion by Mexico in 1997.

19.   The main problems with the early repurchase policy are the following:

• The initiative to request early repurchases of a member is with the Fund. Under the ERP, the “null hypothesis” or presumption is that the member is not able to repay earlier than scheduled, and thus, by construction, the Fund must have a high degree of confidence that the opposite is the case in order to request an early repurchase. In cases where views might differ on the member’s ability to repay early, the construct gives the benefit of the doubt to the member—whereas if, in such a situation, the member were required to make a case to the Board that it cannot repay early, it might well in some cases decide not to make such a case. Indeed, and symptomatic of the high standard required under the ERP, the staff typically does not bring proposals for early repurchases under the ERP to the Board unless the member agrees. Overall, the high degree of confidence that is required for the Fund to request early repurchases under this construct is difficult to secure on an issue that is ultimately a matter of judgment, and a complex one.²⁵

• ERP decisions have been linked to decisions on the inclusion of members in the financial transactions plan (FTP). The use of a single framework to identify both potential creditor members and debtor members that are in a position to repay the Fund early has had the unintended effect of making early repurchases less likely. While the two balance of payments tests are essentially identical, the implications for the member are quite different: early repurchases result in a reduction in gross reserves, whereas participation in the FTP results only in an exchange of reserve assets.²⁶ In practice, the reduction in gross reserves arising from inclusion of debtor members in the FTP has made it more difficult to agree on their inclusion for early repurchases.

• The amount to be repurchased under the ERP is currently determined by a formula based entirely on the level of and changes in gross reserves. While reserves remain an important indicator of capacity to repay the Fund, they are not an adequate indicator of overall balance of payments strength, not least because they are under the member’s own control, especially in the presence of access to international capital markets.

20.   The staff believes that remedying the flaws in the ERP would essentially replicate the time-based repurchase expectations (REs) proposed in RoF II:

• There is little prospect of a well-functioning ERP as long as the initiative to request early repurchases remains with the Fund. As noted above, the ERP is biased toward acceptance of the presumption that members are not able to repay ahead of schedule. It is not clear that this presumption is consistent with a world in which a large share of the membership may be able to repay ahead of obligatory maturity dates. By contrast, if the Fund had to make a judgment as to whether a member was justified in not repaying early, members would probably more readily accept the notion of early repurchases. The conclusion that the Fund should be making judgments about members that cannot repay early, rather than about members that can, is a key component of the proposal for REs.

• Fine-tuning of the amounts to be repurchased—as attempted in the ERP—has become increasingly problematic, and there is a case instead for simply accelerating repurchases falling due, for all outstanding amounts. In principle, there is a case for fine-tuning (as it tailors amounts repurchased to the strength of a member’s external position), but there is also a case for linking the amount of repurchase to the amount of Fund credit outstanding (as the purpose of seeking early repurchases is to ensure the revolving character of Fund resources). Moreover, especially in an environment of increasingly important private capital markets, fine-tuning is difficult. Gross reserves are not, or at least are no longer, a reliable indicator of external strength, and no small (or even large) set of indicators reliably captures countries’ external strength and vulnerability.²⁷ In these circumstances, the staff believes that, to be effective, the instrument should not try to fine-tune the amount of repurchases when a member’s external position is strong, but instead simply accelerate repurchases relative to the original schedule—as proposed for REs.

• Early repurchases would be separated from the FTP. This would allow discussions and decisions on whether members should meet REs to take place more naturally in the context of country-specific Board discussions.

21.   The staff thus believes that, among instruments relying on rules rather than incentives, time-based repurchase expectations stand the best chance of ensuring that members repay as their external position improves. Such repurchase expectations would be included in all arrangements and would complement the existing ERP, which could continue to play a useful role for the exceptional situations where countries’ positions improved so rapidly that they were able to repay before the dates specified in the expectation schedules. The staff could return at a later date to the issue of strengthening the early repurchase policy (which, as noted above, is likely to involve reconfiguring the policy to resemble more closely, and to the extent possible, time-based repurchase expectations).

(ii)   Graduation of charges with time and repurchase expectations

22.   Graduation of charges with time and time-based repurchase expectations can be seen as two alternative ways of addressing the problem of long use, with the emphasis placed, respectively, on quantitative rules and on incentives.

23.   At the same time, it is not clear that time-based graduation of charges and repurchase expectations would be suitable complements. If both instruments were used at the same time, weaker members, which had been found not able to adhere to the repurchase expectation schedule, would have to pay a surcharge.²⁸ Whether this was deemed appropriate would depend on the degree to which the weakness of such members’ external positions was considered to reflect a weakness in policies, rather than exogenous constraints.

24.   The present section will discuss in turn options for graduation of charges with time and for time-based repurchase expectations, and will then provide tentative estimates of the possible impact of these policies on members and on the Fund, with the aim of eliciting Directors’ views on which of these changes, if any, they might support.

Graduation of charges with time

25.   A price incentive to encourage advance repurchases could be introduced as a surcharge graduated with the length of time that credit has been outstanding.²⁹ Such a surcharge could begin at a low rate, starting, for example, at the time when repurchases begin to fall due—after 3¼ years for SBAs and after 4½ years for extended arrangements. In recognition of the longer maturities of the EFF, the surcharge under an EA could rise more gradually than under an SBA (see Figure 1 for two possible options).

26.   A surcharge that graduates with the time credit has been outstanding could begin, in principle, either from the date of each purchase, or from the date of the first purchase under an arrangement. The former would be feasible within the present structure of facilities, but the latter would require important changes in that structure. As noted in paragraph 15, at present SBAs are but one means of accessing the credit tranches: other means comprise outright purchases under emergency assistance, as well as other outright purchases in the credit tranches (although there have been none of the latter since 1968). Because the concept of “first purchase under the arrangement” is meaningless for outright purchases, and because the principle of uniformity of treatment dictates that purchases under arrangements in the credit tranches must be treated the same way as outright purchases in the credit tranches, graduation with time in the present structure would have to begin separately for each purchase from the date of that purchase. To make feasible graduation with time from the date of the first purchase for all amounts drawn under an arrangement, stand-by arrangements would have to be made the only means of accessing the credit tranches—implying that (i) the two forms of emergency assistance would have to be turned into special policies, (ii) the possibility of other outright purchases in the credit tranches would have to be terminated, and (iii) the Board would have to agree that credit tranche purchases through SBAs would constitute the general policy on access to the Fund’s general resources (i.e., available for all types of balance of payments problems) that the Fund is required to have according to Article V, Section 3(a). These changes would require a 70-85 percent majority of the total voting power, depending on the features of the special policies thus created on emergency assistance (see footnote 22).

27.   A time-based surcharge would have both advantages and disadvantages:

• It would rely on a price incentive in order to encourage advance repurchases, and hence would give the member more flexibility than a system based on quantitative rules. This incentive might be particularly effective in the case of members with alternative sources of finance, and of members that no longer have a balance of payments need but may be reluctant to make advance repurchases because they believe they are able to receive a rate of return on their reserves that is higher than the rate of charge.

• It would apply to all members regardless of the strength of their external position, and could be said to bring both advantages and disadvantages on this count. The surcharge could have adverse consequences for those members with a weaker balance of payments position, particularly those without access to alternative sources of external finance. At the same time, to the extent that the weakness of some members’ external positions reflects inappropriate policies, a time-based surcharge would constitute an incentive for members to strengthen their policies, so that they would be in a position to make repurchases sooner.

• It could be less successful than repurchase expectations in inducing early repurchases, although this would depend on effective implementation of REs, as proposed below. Absent a fairly steep schedule of surcharges, some members which the Fund would have found strong enough to repurchase early might nevertheless choose not to voluntarily effect advance repurchases.

• A surcharge graduated over time within the context of a single arrangement would provide an incentive to “refinance” Fund credit. Under either variant of such a surcharge (described in paragraph 26), there would be an incentive for members to seek new arrangements, under which purchases would be at lower rates of charge than outstanding credit; and while the Fund could curtail access to its resources when members appear to be seeking refinancing, it might have difficulties doing so in practice, and, in order to do so, it might also need to be more restrictive in granting precautionary arrangements. If graduation started from the date of each purchase, there would even be an incentive to refinance older purchases with new purchases under the same arrangement.³⁰ The possibility of refinancing is also the reason why graduation with time under a single arrangement—under either variant—would be a weak instrument to discourage prolonged use, since prolonged use involves recourse to Fund resources over multiple purchases and arrangements (see paragraph 11).

• Since advance repurchases made under a surcharge incentive would be made at the member’s discretion, they would trigger prepayment clauses in other creditors’ operations (see Annex). This would reduce the member’s incentive to effect advance repurchases.

Time-based repurchase expectations

28.   Introducing time-based repurchase expectations could give the Fund a flexible instrument to address directly unduly long use of Fund resources. Repurchase expectations would consist of a repurchase schedule with earlier repayment dates than the dates in the obligations schedule.

29.   A member would be expected to meet the earlier schedule of repurchases, but a member whose external position does not allow it to meet repurchase expectations could request the Fund to grant it a waiver of the expectation. The member would request such a waiver in advance of the repurchase expectation date. If a waiver were not granted and the expectation were not met, remedies would apply (see below). The member could request a waiver at any time—including, for instance, after it has met several repurchase expectations, should its external position deteriorate. The Fund could grant a waiver for several or all repurchase expectations at once.

30.   Remedies would apply if the member did not adhere to the repurchase expectation schedule and a waiver had not been granted. In particular, the member’s right to make further purchases, including under ongoing arrangements, would be suspended.³¹ This remedy would be stronger than under the ERP, where no remedies are currently provided for—yet, even in the absence of remedies, no member has failed to meet any of the 38 early repurchase expectations imposed under the ERP in the 22 years since the Guidelines were established. In addition, consideration could be given to establishing a policy that would allow unmet REs to be converted into obligations (as under the SRF and CCL); failure to meet this obligation would then leave the member in arrears. Allowing the conversion of repurchase expectations into obligations would raise the majority required for the introduction of REs from a simple majority of the votes cast to an 85 percent majority of the total voting power.³²

31.   The Fund could establish criteria to guide its decisions as to whether to grant waivers of the expectations. Possible criteria are presented in Box 2 for preliminary consideration by Directors.³³

32.   It is important to note that the proposed time-based REs could constitute a means to encourage early repurchases when a member’s external position is strong, but would not be an instrument with which to induce early repurchases when the member’s policies are weak. As noted in RoF II (paragraph 50), under this approach, the Fund would not refuse a member’s request for a waiver of the expectation if its external position was weak as a consequence of its own policies. In this respect, a system with REs would resemble the current system, where members are permitted to continue to use Fund resources (in the sense of not repurchasing them) following the expiration of an arrangement, regardless of developments in their policies. The Fund could instead seek to influence the conduct of policies following an arrangement by means of post-program monitoring, which is discussed in Chapter V.

33.   Consideration could be given to introducing REs into SBAs that bring the maturity structure forward by one year, as with the SRF and CCL. Repurchase periods would thus be advanced to 2¼ - 4 years from 3¼ - 5 years (Box 3).³⁴ As noted in RoF II (paragraph 53), such a change would not necessarily be inconsistent with the maximum duration of SBAs of 3 years, since long SBAs are normally expected to turn precautionary in their later years, and, were this not the case, the member could request the Fund to grant a waiver of the expectation.

34.   Consideration could be given to introducing REs into the EFF that accelerate repayments after the grace period, by doubling repurchases until all credit is repaid. Purchases under the EFF carry repurchase periods of up to 10 years, including the grace period of 4½ years; REs could shorten the repurchase period from 10 to 7 years (Box 3). As noted in RoF II (paragraph 52), it may not be appropriate to shorten the grace period under the EFF in view of the need for members with EAs to undertake deep-seated reforms for several years.

35.   Both time-based repurchase expectations and time-based surcharges would apply only to purchases made after the changes came into effect, including purchases under arrangements already in existence at that time. If a surcharge schedule for time-based surcharges were to key off the first purchase under an arrangement, the Board would need to decide whether the first purchase under an existing arrangement made after the changes would be treated as “the first purchase under an arrangement.”

Impact on Fund income and on members

36.   The policy options described above would clearly have impacts on both individual members and on the Fund’s net income, but their effects cannot be predicted with confidence. Any quantitative exercise assessing the impacts of the changes on individual members and on Fund income must, by necessity, be based on certain assumptions, about both members’ responses to surcharges and the impact of repurchase expectations. In effect, any quantitative assessment will have to anticipate behavioral reactions about which very little is known.

37.   A particular difficulty lies in estimating the changes in total Fund credit outstanding that might result from the policy changes, and any analysis has to make simplifying assumptions in this respect. Both surcharges and the establishment of REs would serve to decrease outstanding Fund credit from what it would otherwise be. The impact on the Fund’s income (and on total charges paid by members) will therefore reflect both the increase in charges and the decrease in credit outstanding. The net effect, however, cannot easily be estimated ex ante or ascertained ex post. Indeed, it is entirely possible that the Fund’s net income would decline. Moreover, because surcharges and repurchase expectations would apply only to new purchases, the impact on the Fund’s net income will be felt only some time in the future. Nevertheless, to make some headway, the staff has made simplifying assumptions (described below) to gauge the impact of the changes: assuming there are no behavioral reactions, and assuming an acceleration of repurchases in response to the introduction of surcharges. The following results should be viewed in light of the caveats described above.

38.   REs by themselves would likely entail a loss of net income for the Fund, as they would lead to a reduction in the average amount of Fund credit outstanding compared to what it would have been without such expectations. Assuming that about half of members with arrangements would have been in a position to meet repurchase expectations, Fund credit would have been reduced by about 10 percent, based on illustrative repurchase expectations for SBAs and EAs (Box 3). Under this scenario, Fund credit would have been SDR 4.3 billion lower at end-June 2000 than it actually was. Such a reduction would have entailed a reduction in net Fund income on an annual basis of about SDR 30 million.³⁵

39.   By contrast, time-based surcharges could possibly generate additional net Fund income, particularly—an extreme assumption—should members not advance repurchases in response to the surcharges. With illustrative surcharges, based on the date of the first purchase under an arrangement, beginning to apply after the end of the grace period, a maximum amount of net income on the order of SDR 120-235 million could be generated on the basis of Fund credit outstanding at end-June 2000 (Tables 2 and 3).³⁶ This assumes that members would not make any advance repurchases in response to the surcharge (i.e., the elasticity of repurchases with respect to surcharges is assumed to be zero).

40.   When members respond to the surcharges by advancing repurchases, the net income gain to the Fund would be much smaller, because fewer surcharges are paid and the Fund loses net income at the regular rate of charge on the repurchases made in advance of when they were due. To illustrate the sensitivity of the calculated income gains to the repurchase behavior of members faced with a time-based surcharge, an alternative calculation (Table 3) assumes that members would reduce their average use of Fund credit by 10 percent as a result of the incentive effect of the surcharge. Members are assumed to advance the repurchases on their purchases subject to the highest surcharges. The illustrative gain in net income in this case would be about SDR 35-105 million, much smaller than that in the “zero elasticity” scenario. A stronger advance repurchase response by members than the assumed reduction in use of Fund credit by 10 percent could actually lead to net income losses for the Fund.

41.   The impact on individual members faced with a time-based surcharge would in general be relatively modest. The impact can be illustrated by the additional charges members would pay relative to regular charges, or as an equivalent annual average surcharge (Table 4). The two illustrative options for a time-based surcharge would raise total charges paid by members under “typical” SBAs by 5 - 10 percent, and 8 - 16 percent under “typical” EAs, assuming that members did not make advance repurchases in response to the surcharges (the “zero elasticity” scenario). Were members to reduce their use of Fund credit by 10 percent over the course of the arrangement by advancing their repurchases subject to the highest surcharges, their surcharge payments would be reduced by about a third compared to the zero elasticity scenario.³⁷

1/ Excluding overdue credit.
2/ Includes credit extended under same payment terms as stand-by arrangements (such as CFF).
3/ Includes credit extended under same payment terms as EFF (such as STF).

1/ Based on data in Table 2, and assuming time-based surcharge on the basis of individual arrangements had always been in effect. Specifically, the calculations assume time-based surcharges had been applied to all outstanding GRA credit not in arrears.
2/ Advance repurchases of SDR 4.3 billion applied to outstanding credit under SBA and EFF terms in proportion to the respective amounts of Fund credit outstanding beyond 3 years for SBA credit and beyond 4 years for EFF credit. Advance repurchases assumed to be made by members subject to highest surcharges.
3/ Number of members that would pay surcharge as of June 30, 2000 if scheme had been introduced then and no advance repurchases were made.
4/ See Figure 1 for definition.
5/ Includes all credit outstanding under the GRA with similar terms.
6/ Includes 15 members that would pay surcharges on both stand-by and EFF credit.

1/ See Box 3 for presentation of the terms of “typical” arrangements.
2/ Advance repurchases applied when surcharges are highest.
3/ See Figure 1 for definition.
4/ Sum of surcharges as percent of sum of charges in typical arrangement (without discounting). Assumed rate of charge is 5.49 percent.
5/ Surcharge over and above the rate of charge in typical arrangement.

B.   Addressing Large Use

42.   As noted above, unduly large use of Fund resources can be seen as an issue separate from, though related to, the issue of unduly long use. While access policy would remain the Fund’s main instrument for curtailing large use, there is a case on this score for graduation of charges with the amount of credit outstanding. Such graduation would also help address the problem of prolonged use, to the extent that the Fund is concerned mostly about prolonged use of relatively large amounts of its resources.

43.   The present section will describe various options for graduation of charges with amount, including some tentative estimates of the possible impact of these options on members and on the Fund, with a view to eliciting Directors’ views on which of these options, if any, they might support.

Graduation of charges with the amount of credit outstanding

44.   A surcharge graduated with the level of credit outstanding would provide members with an incentive to curtail large use. The SRF, which is designed to be used in circumstances where large use of Fund resources may be expected (above the annual access limit of 100 percent of quota or the cumulative access limit of 300 percent of quota), imposes a surcharge over the rate of charge beginning at 300 basis points.

45.   A surcharge under SBAs and the EFF that peaks at 300 basis points would go some way toward harmonizing surcharges on the large use of GRA resources. Such a surcharge could be graduated, for example in three steps of 100 basis points, with the first step at outstanding credit of 100 percent of quota and additional steps as credit rises by each additional 100 percent of quota. Surcharges would apply to the corresponding marginal amounts of credit, not to the total amount of outstanding credit. Other options may be to exclude either the first or the first two steps (see Figure 2). However, a surcharge imposed only on amounts above 300 percent of credit outstanding would likely apply to very few members.³⁸ A surcharge on amounts above 100 percent of quota would apply to rather more members, including “prolonged users” of Fund resources as these have usually been defined (see paragraph 8).

46.   Credit outstanding under the SRF and CCL would not be subject to the level-based surcharges being considered here. SRF resources would continue to carry their present surcharge over the basic rate of charge, of 300 basis points for the first year after the first purchase, and rising by 50 basis points every six months thereafter, up to a ceiling of 500 basis points. The surcharge on CCL resources—currently identical to that on SRF resources—is reconsidered below.

47.   Level-based surcharges would apply to purchases made after the policy was introduced. Credit existing prior to the introduction of the policy would not bear a surcharge. However, such credit could be taken into account to determine which rate of surcharge applied to new drawings from the Fund, in which case a member with a large amount of Fund credit outstanding when the policy was introduced would pay a higher rate of charge on purchases under a new arrangement than would a member with little or no credit outstanding. Alternatively, credit outstanding when the policy was introduced could be excluded altogether from the determination of the surcharge, in which case it is likely to take some time before members’ new purchases cumulate to the levels where surcharges would apply.

Impact on Fund income and on members

48.   As noted above in the case of time-based surcharges, the impact of the introduction of a level-based surcharge on Fund income and on members is difficult to gauge. The impact depends on the amount of credit outstanding, the level of credit at which the surcharge begins to apply, the rate of surcharge, and how members would respond to the surcharge. For the purpose of illustrating the possible impacts on Fund income and on members, a level-based surcharge is assumed to be combined with REs (Table 5). While it would be possible to combine level-based and time-based surcharges, the combination would suffer from a number of drawbacks (see RoF II, paragraph 39). The illustrative REs discussed above were assumed to have reduced Fund credit at end-June 2000 by SDR 4.3 billion.

49.   The impact of a level-based surcharge on net Fund income could range from strongly positive to strongly negative, depending on which scheme is assumed to be adopted and on members’ responses to it. Estimates of the impact of a level-based surcharge on net annual Fund income are based on the hypothetical distribution of Fund credit that is obtained after taking into account REs. If it is assumed that members did not advance their repurchases in response to the surcharge (“zero elasticity of repurchases with respect to surcharges”), the Fund might see a gain in its annual net income ranging up to SDR 180 million if graduation begins at credit of 100 percent of quota (Table 6).³⁹ If the surcharge applies only at credit levels over 300 percent of quota, however, the net income gain would be small. If the opposite extreme assumption is made that members would advance their repurchases or forego purchases so as to avoid paying any surcharges (“infinite elasticity”), the Fund would stand to lose substantial net income due to the loss of regular net income on the advance repurchases or foregone purchases (which is added to the loss of income due to REs). How close to either extreme members’ behavioral response would actually be is not known.

1/ Excluding overdue credit.
2/ Outstanding credit by member allocated to successive categories. For example, of a member’s total credit of 150 percent of quota, 100 percent would fall in the “up to 100 percent of quota” category and the remainder in the “100 to 200 percent of quota” category.
3/ Number of members with total Fund credit outstanding in that category.
4/ This entails a reduction in Fund credit of SDR 4.3 billion, which is distributed in proportion to the outstanding amounts in each category to obtain the hypothetical amounts.

50.   The amount that an individual member would pay under a level-based surcharge would depend on its particular circumstances, but would, in general, not be very large. A member with an arrangement of 125 percent of quota might pay very little or nothing under any of the three surcharge options, even if no advance repurchases were made (Table 7). A member with an arrangement of 250 percent of quota might pay surcharges equivalent to about 10 percent of regular charges (if graduation began at 100 percent of quota) if no advance repurchases were made. Under the option involving the highest starting point for surcharges, only members with outstanding credit above 300 percent of quota would pay surcharges (not shown in Table 7). Of course, should members advance repurchases or forego purchases to avoid the surcharge, no surcharge payments would be made (also not shown in Table 7).

III.   The Contingent Credit Lines

51.   The Board had a preliminary discussion in March, and a more detailed discussion in July, of possible changes to the CCL that could make this as yet unused facility more useful to members, and thus would better encourage their efforts at crisis prevention. The present chapter follows up on these discussions, with the intention of laying the groundwork for completion of the review of the CCL (now scheduled for end-September 2000) and for the implementation of a number of experimental changes, which might help establish, over the coming months, whether the CCL can become a useful instrument.

A.   The Activation Review

52.   There was broad agreement, in the July Board discussion, that the scope for judgments by the Fund in the CCL activation review should be reduced. In the activation review, the Board currently verifies that “(i) the member is experiencing exceptional balance of payments difficulties due to a large short-term financing need resulting from a sudden and disruptive loss of market confidence reflected in pressure on the capital account and the member’s reserves, (ii) these difficulties are judged to be largely beyond the member’s control and to be primarily from adverse developments in international capital markets consequent upon developments in one or several other countries, (iii) up to the time of the crisis, the member has successfully implemented the economic program that it had presented to the Board as a basis for its access to CCL resources, and (iv) the member is committed to adjusting policies to deal with any real economic impact that may follow from contagion.”⁴⁰

1/ Based on hypothetical amount of Fund credit outstanding at end-June 2000 (see Table 5), and assuming repurchase expectations and level-based surcharges had always been in effect. Includes hypothetical income loss of SDR 30 million due to lower outstanding credit resulting from early repurchase expectations.
2/ See Figure 2 for definition.

1/ See Box 3 for presentation of the assumed arrangement terms (apart from access levels). Elasticity of repurchases with respect to surcharges assumed to be zero.
2/ See Box 3 for assumptions regarding repurchase expectations.
3/ See Figure 2 for definition.
4/ Sum of surcharges as percent of sum of charges in arrangement without repurchase expectations (without discounting). Assumed rate of charge is 5.49 percent.
5/ Surcharge (in arrangement without repurchase expectations) over and above the rate of charge.

53.   Three separate issues may be distinguished here—the Fund’s discretion to require policy changes, the scope for judgments by the Fund on the circumstances the member faces, and its judgment as to whether the member has adhered to its program. These relate respectively to condition (iv), conditions (i) and (ii), and condition (iii) above. The staff had proposed, in RoF II, reducing the Fund’s discretion to ask for policy changes in the activation review (condition (iv)). However, a number of Directors were also concerned about the extent to which the other conditions for the activation review reduce the degree of assurance given to the member and to markets regarding the availability of CCL resources. The remainder of this section will consider the three broad aspects of the CCL activation review in turn.

(i)   Policy changes

54.   The staff had proposed, in RoF II, that the Fund’s discretion to require policy changes in the activation review be reduced. The Fund would verify conditions (i)-(iii) above in the review, but the member would be given the strong benefit of the doubt as to condition (iv). The “strong benefit of the doubt” here should be taken to mean that the Fund would proceed expeditiously, without comprehensively examining the member’s policy intentions, unless it had clear evidence that the member planned not to adjust policies to deal with any real economic impact from contagion. The amount available on completion of the activation review would be specified at the time of approval of the commitment under the CCL, and would—either as a rule or as a normal practice—be large (e.g., a set amount of 100 percent of quota, or a set fraction, perhaps a third, of the arrangement).

55.   The activation review would thus fulfill two functions, and indeed in this construct two separate but possibly simultaneous reviews would be conducted (which are subsumed under the heading of the activation review in the present structure of the CCL): an “activation” review, and a “phasing” or “post-activation” review. The Fund would complete the activation review if it determines that conditions (i)-(iii) above are satisfied, and unless it had clear evidence that the member planned not to adjust policies to deal with any real economic impact of contagion; the prespecified amount would be made available at that time. However, for the purpose of specifying phasing and conditionality for the remainder of the amounts committed, the Fund and the member would have to reach understandings on the policies to be pursued from that point onward—understandings which would be sought in the “post-activation” review. The “activation” and “post-activation” reviews could be completed simultaneously, in which case the amount immediately released could be larger than that prespecified, on the grounds that the Fund and the member have already reached understandings on any policy adjustments that may be required. In most cases, however, it would seem likely that the activation review would be completed first, and expeditiously, and that it would be immediately followed by in-depth discussions between the member and the staff regarding the possible need for policy adjustments.

56.   In the July discussion, there appeared to be close to a consensus on the staff proposal, with a few Directors suggesting more far-reaching changes. In particular, it was suggested that the activation review—or rather, in the terminology introduced above, the “post-activation” review—should focus exclusively on the stance of macroeconomic policies, since a temporary capital account shock would not be expected to signal a need for adjustments in structural policies. While it would indeed generally be expected that any “post-activation” policy changes would relate to macroeconomic policies, the staff would not recommend that the Fund irrevocably tie its hands in this regard, not least because it is conceivable that a crisis will reveal structural issues that had earlier gone unnoticed.

(ii)   The existence of a contagion-driven crisis

57.   In the July Board discussion, a number of Directors expressed concern that the scope for judgments by the Fund as to whether a contagion-driven crisis was really underway could reduce the confidence of the member and of markets in the availability of resources. It was emphasized that the external circumstances in which the credit line might be drawn had been specified in the private contingent credit lines that at least one member had negotiated.

58.   The staff believes that it is important to retain conditions (i) and (ii) for the activation review:

• It is important to limit availability of CCL resources to cases where a crisis is underway, and where that crisis is triggered by external developments.⁴¹ Thus, it would not seem desirable, as was suggested by some Directors in the July Board discussion, simply to eliminate conditions (i) and (ii) for the activation review.⁴²

• It is incumbent on the Board, on the staff and management’s recommendation, to make a determination whether these conditions are met, because of the link to the decision to release resources. Thus, the suggestion made at the July discussion, to create a presumption that conditions (i) and (ii) were satisfied unless the staff had strong doubts about this, would put too much power in the staff’s hands.

59.   Alternatively, one could consider whether prespecified tests could be introduced, on which the judgment whether conditions (i) and (ii) were satisfied would be based. Such tests would specify “threshold conditions” which, if met, would constitute an assurance that conditions (i) and (ii) had been satisfied. In principle, such tests could be designed based on general principles, taking into account country-specific characteristics, and their specifics could be discussed between the authorities and the Fund, as part of the arrangement with CCL resources:

• With regard to condition (i), the staff doubts whether sufficiently robust “threshold conditions” could be prespecified that would indicate that a capital account-driven crisis was underway. The test would have to relate, essentially, to the sudden reversal of private capital flows. Even abstracting from issues of timeliness of data, a serious problem with this approach is that the size of the reversal in capital flows would depend on the country’s exchange rate and interest rate policies, so that it would not be an objective measure of the ex ante pressure on the member’s capital account. Some aggregate measure of net capital flows (perhaps approximated in the short-term based on changes in reserves), interest rates, and the exchange rate could be designed to address this problem. Standard early warning models, which try to predict external crises, define the left-hand side variable—the crisis index—in a similar way, although they usually exclude the interest rate component owing to data availability problems.⁴³ These models, however, do not find cross-country measures to perform well, and thus define occurrence of a “crisis” as a shock that is large relative to a country’s own experience in the past. As a result, these indices more readily indicate that a crisis is occurring in a country that has enjoyed a long period of stability than in one that has undergone episodes of instability. Use of such an index in the CCL would not seem appropriate. Moreover, in general it would be difficult to estimate an entirely robust relationship between the various variables involved that could be expected to survive in a crisis environment, especially if such an estimation is based on a relatively calm sample period, and so a threshold condition designed to indicate whether capital outflows had reached “crisis levels” would probably give the wrong signal in some cases.

• With regard to condition (ii), specification of precise tests would be even more difficult, but the staff sees potential for reducing uncertainty for the member by means of an illustrative (but in no way restrictive) list of external shocks that might trigger a capital account-driven crisis. Given the infinite number of possible shocks in the global economy, and the uncertainty involved in projecting the reaction of international capital markets to any such shock, it would not be possible to provide a complete listing of possible shocks, much less of the magnitude such shocks would have to reach in order to be considered possible triggers for contagion. It should, however, be possible, for the Fund and the member in consultation, to identify the shocks considered most likely to trigger a crisis for a particular member, and to include such a listing in the member’s request for a commitment of CCL resources. Such a listing would not in any way be binding: shocks that were on the list could be determined not to have caused a crisis, and shocks not on the list could be determined to have done so. Nevertheless, the listing would give some assurance to the member that the Fund and the member see eye to eye on the type of shocks that—if they occasion a capital account-driven crisis—would be considered to qualify for activation of an arrangement with CCL resources.

60.   All in all, the staff sees only rather little scope to reduce the room for judgment by the Fund, in the activation review, as to whether a crisis is underway and whether it was triggered by external developments. At the same time, the degree of uncertainty introduced by this judgment may be relatively small. The experience of the last few years suggests that most observers agree when a crisis is underway, and even agree (given an assessment of domestic policies, see below) on whether external developments played a major part in its genesis.

(iii)   The member’s adherence to its program

61.   It would also be difficult to eliminate the element of judgment in the final condition for the activation review, that the member should have adhered to its program. The member’s adherence to its program signals that it continues to meet the CCL eligibility criteria, which are complex and cannot be monitored simply by means of a limited number of conditions that could all be set out in the arrangement (along the lines of performance criteria).⁴⁴ At the same time, and despite this complexity, the staff believes that the scope for judgment by the Fund on this score should not create excessive uncertainty. The member’s program is specified at the time of commitment of CCL resources, and reviewed at the time of the mid-term review, and there are thus clear benchmarks against which to measure performance in many areas.

B.   Exit

62.   At the July Board meeting, many Directors were concerned about how members might exit from the CCL, especially in a case where it becomes clear that the CCL eligibility criteria are no longer met. RoF II (paragraphs 88-89) discussed the difficulty of exiting for any reason—as markets may suspect there have been policy slippages even if there have not—and noted that this difficulty may represent an obstacle to members’ entering into the CCL. Directors seemed especially concerned, however, that the risk of triggering a crisis by making a member exit from the CCL in the event of a weakening of policies could make the Fund reluctant to enforce the CCL criteria as strictly as it should.⁴⁵

63.   Precautionary arrangements were mentioned as one possible route for exit in the event of a weakening of policies, which would reassure markets that policies remain reasonably sound. The staff believes that it would be useful to encourage a member that needs to exit from the CCL on this account to request an arrangement it would treat as precautionary (or indeed, if the need arose, one under which it would make purchases).⁴⁶ On the other hand, it is not possible to compel a member to request an arrangement.

64.   Many Directors were interested in a “graduated response” to policy slippages under an arrangement with CCL resources. It was suggested, for instance, that the amount that would be available upon completion of the activation review could be reduced in stages, and finally eliminated, in the event of a weakening of policies.

65.   The notion of a graduated response, however, brings with it the idea of a “graduated CCL”—a CCL under which some members would be more highly “rated” than others. The graduated response would in itself require very finely-tuned judgments by the Fund, which would be even more difficult than the already difficult binary judgment that the Fund currently needs to make about CCL eligibility. Moreover, while it would be possible, in principle, to maintain a single standard for approval of a commitment of CCL resources and to introduce “graduation” only over the life of the commitment, the existence of different CCL “ratings” later in the commitment period would almost surely contaminate the signal the Fund seeks to give upon approval of access to the CCL. In addition, there is a substantial risk that any step in a graduated response—e.g., a reduction in the amount to be made available upon completion of the activation review—would have similar adverse effects on market perceptions as an outright exit.

66.   The staff thus sees little scope for new procedures to resolve the issue of exit—although, at the same time, it would emphasize that exit from the CCL as currently designed is likely to take place in a way that in itself leads to a “graduated” reaction by markets. The reason for this is the same as in other arrangements. Inevitably, a typical Fund arrangement does not go off track suddenly, but gradually. It is not possible for the Fund to decide and declare that a member’s program has gone off track beyond hope of recovery; the possibility remains open that the program will come back on track, although the likelihood of this generally diminishes as time goes on. Thus, in practice, when a program goes off track a program review (or the approval of a new arrangement) becomes subject to increasing delays, until finally it may become clear that it will most probably not be completed. (And it is worth noting in this connection that delays in completing reviews under arrangements with emerging market economies—likely to be closely watched by markets—have rarely led to sudden or significant shifts in market sentiment.) Should a member that has been granted a commitment under the CCL no longer satisfy the eligibility criteria, similar mechanisms would come into play: non-completion of the mid-term review (or the absence of approval of a new arrangement with CCL resources, should the member have requested it) would probably be regarded by markets with increasing seriousness as time went on.

C.   The Rate of Charge

67.   There is considerable sympathy in the Board for the idea of reducing the rate of charge on the CCL, so as to better position the facility in relation to the SRF. The CCL rate of charge is currently identical to the SRF rate of charge—incorporating a surcharge of 300 basis points over the GRA rate of charge for one year from the date of the first purchase under the facility, rising by 50 basis points every six months up to a maximum of 500 basis points.

68.   In the July Board discussion, Directors considered whether, as suggested in RoF II, the initial rate of charge on an arrangement with CCL resources should be aligned with the rate on a large stand-by arrangement entirely in the credit tranches.⁴⁷ A number of Directors expressed a preference for a rate of charge above the rate on a regular stand-by arrangement. At the same time, other Directors considered that the rate should be the same as that on a regular stand-by arrangement.

69.   Overall, unless the Board comes to a decision on charges quickly and finds a consensus on how the initial CCL rate of charge should relate to the rate on a regular stand-by arrangement, there may be a case for simply halving the starting surcharge on CCL resources. The initial surcharge would thus be 150 basis points—the lower end of the range discussed in RoF II. This surcharge would be only a little above the surcharge (120 basis points) that would be required to equalize the initial rate of charge on an arrangement with CCL resources and a stand-by arrangement in the credit tranches of 500 percent of quota (the upper end of the expected range of access under the CCL), if the Fund imposed a surcharge of 300 basis points on all use of Fund credit above the cumulative access limit. Reducing the starting surcharge to 150 basis points should be sufficient to introduce a significant difference in cost relative to the SRF.

70.   At the same time, and as suggested in RoF II, the surcharge on CCL resources could rise on the same schedule as the surcharge under the SRF, in recognition of the likely short duration of the member’s difficulties. The surcharge would thus rise by 50 basis points one year after the first purchase under the facility, and by a further 50 basis points every six months thereafter. The ceiling on the surcharge could be reduced, relative to the ceiling in the SRF, to 350 basis points, maintaining the 150 basis point differential between SRF and CCL rates at all times.

D.   The Commitment Fee

71.   There is considerable support in the Board for the proposal to reduce the commitment fee applying to arrangements with CCL resources in order to encourage members to utilize the CCL. This is intended to address the concern that the commitment fee that applies to arrangements would apply, on approval, also to CCL resources, even though there is little prospect that these resources would be used. In order to address this concern while ensuring that the commitment fee applies uniformly to all arrangements (with or without CCL resources), as required by the Articles of Agreement, an effective reduction could be implemented through a change in Rule I-8 of the Fund’s Rules and Regulations.⁴⁸ In particular, the basis for the payment of the commitment fee could be changed from “the total amount of the arrangement that could be purchased during each twelve-month period,” as currently provided, to “the total amount of scheduled purchases during each twelve-month period.” With this change, an arrangement with CCL resources would, on approval, trigger a commitment fee only on the purchases scheduled on approval (typically, only the first purchase of 5-25 percent of quota).⁴⁹ No commitment fee would apply on the CCL resources until the activation review (or the post-activation review, as proposed), at which time the Executive Board would determine the schedule of availability of later purchases. However, there would be no effective commitment fee on the amount of the purchase upon activation since this portion of any commitment fee would be refunded. Portions of the commitment fee pertaining to subsequent scheduled purchases would be refunded only if and when those purchases were actually made.⁵⁰ The necessary change in Rule I-8(a) could be made by a decision adopted by a 70 percent majority of the total voting power.

IV.   The Extended Fund Facility

72.   Directors generally agreed at the July Board meeting on RoF II that greater attention should be given to ensuring that arrangements under the EFF are granted only to cases fully consistent with the spirit and terms of the 1974 EFF Decision.These would be cases where a member’s balance of payments difficulties are expected to be relatively long term. Programs underpinning requests for EAs would be expected to continue to include appropriately strong structural reform programs, focused on achieving external viability in the longer-term as well as enhancing growth prospects. However, in contrast to some recent practice, the strength of the structural reform effort per se should not be considered sufficient reason for the use of EFF resources. In the absence of longer-term balance of payments difficulties, such structural programs could be implemented with the support of resources under a stand-by arrangement.⁵¹ Further, the need for a longer-term planning horizon, i.e., reforms that span a three-year period rather than the more typical one- to two-year period of the majority of stand-by arrangements, would also not be sufficient reason for granting resources under an EA. Such a planning horizon could be accommodated by a three-year SBA as well.

73.   A priori, members that would be expected to be eligible for financing under the EFF would be primarily countries just above the threshold for PRGF eligibility, including PRGF graduates; PRGF-eligible countries (when EFF and PRGF resources are blended); and some early transition countries. EFF financing would continue to be available to other countries as well—including to countries opening their capital markets in a well-defined, structured process—as long as such members have reasonable grounds to expect their balance of payments difficulties to persist over the longer-term. The EFF would probably not be used by members with well-established access to captial markets, which are not likely to suffer from the problems described in the EFF Decision—severe payments imbalances, or slow growth and an inherently weak balance of payments—and whose balance of payments difficulties are likely to be of a relatively short-term nature.

74.   The majority of Directors agreed that use of the EFF on a precautionary basis should be exceptional. Several Directors saw merit in precautionary EAs, because they provide confidence to members making difficult policy decisions, by giving assurance that resources will be available over a reasonably extended period if a balance of payments need should develop during a period of comprehensive economic adjustment. At the same time, it was recognized that the circumstances in which a member without an immediate need for Fund resources would develop longer-term balance of payments difficulties would be rare. It is not unlikely that a member without an immediate balance of payments problem may, in the course of implementing a medium-term structural adjustment program, develop a balance of payments need. This could be the result of adverse market developments not related specifically to the reform program or directly stemming from the program itself, for example, as a result of extensive resource reallocation or the relaxation of capital controls. However, there will rarely if ever be a high likelihood that this need would develop into a longer-term balance of payments problem justifying longer-term financing. The presumption is that if a member has appropriate policies and no need for Fund resources to begin with, its balance of payments position would likely recover within the repayment terms provided under a stand-by arrangement. Should this not be the case, the member could always request a follow-up arrangement under the EFF.

V.   Post-Program Monitoring

75.   The objective of stronger post-program monitoring (PPM) is to provide for closer monitoring of the circumstances and policies of members whose arrangement has expired but that continue to have Fund credit outstanding. Stronger PPM could provide an early warning of policies which could call into question a member’s continued progress toward external viability, and thus could eventually imperil Fund resources, or at the least indicate that such resources were not being used (in the sense of continuing to be used) for their intended purpose.

76.   In RoF II, the staff proposed PPM procedures that would aim to provide additional safeguards to Fund resources. Under current practices, the staff monitors developments in members with outstanding Fund resources relatively intensively, but this does not usually formally involve the Executive Board. RoF II proposed procedures that entailed regular (say, semi-annual) staff discussions with a member, with a particular focus on macroeconomic and those structural policies that have a bearing on external viability, which would be reported to the Board in the form of a brief Board paper. These procedures would be grounded in the use of the existing consultation clauses in all arrangements. The staff suggested that this monitoring could be carried out for members with relatively high Fund credit outstanding and/or for members where the risk of encountering difficulties in repaying the Fund was seen as high.

77.   During the July Board discussion, a number of Directors saw no need to formalize practices on PPM. In particular, they expressed a strong preference for PPM on a case-by-case basis rather than on the basis of the automatic application of formal criteria.

78.   An alternative proposal—noted by the Chairman in his concluding remarks—would involve more formal procedures, but applied on a case-by-case basis. PPM could be based on the recommendation of the Managing Director, and the agreement of the Board, on the occasion of a member’s Article IV Consultation discussion or during informal country matters sessions.⁵² It would involve discussions with the member and the preparation by the staff of a Board paper focusing on recent developments and policies, and how they affect the member’s medium-term prospects and capacity to repay the Fund. The Board paper would recommend, and the Executive Board would decide, whether there should be further PPM. The Managing Director would be guided in his or her recommendations by developments in a member’s policies and external position that he or she would see as posing risks to the member’s progress toward external viability.⁵³

79.   At the same time, it is important to note that this proposal could represent only a minor strengthening relative to current practices. The ad hoc application of the procedure could well result in very infrequent implementation, just as, for instance, the existing ad hoc procedures for supplemental/special consultations (which are instruments of surveillance) have been used only twice since the relevant decision was adopted in 1979. The fear of adverse market reaction and concerns about uniformity of treatment could contribute to such an outcome. A rule, or alternatively a presumption, that members meeting certain criteria (e.g., a level of Fund credit outstanding above a certain threshold) would engage in PPM might help ensure a more frequent application, which in itself should alleviate the concern that members “singled out” for PPM might suffer an adverse market reaction.

80.   Consistent with the Fund’s policy on transparency, the staff would recommend that, following a PPM discussion, publication of a Public Information Notice (PIN), based on a summing up of the Board discussion, be encouraged. While dissemination to the public of discussions on use of Fund resources⁵⁴ is through a Chairman’s Statement, a PIN would be a more natural vehicle for a summary of a PPM discussion that might involve substantial concerns about the course of a member’s policies. As with other PINs, the member’s consent would be required for publication. The use of PINs following a discussion about use of Fund resources would require a change in the Decision governing PINs.

81.   The Board might wish to review the implementation of, and procedures for, PPM in about 18 months.

VI.   Issues for Discussion

82.   Directors may wish to indicate whether they believe that there is a problem, or a significant risk, of excessive use of Fund resources, and, if so, whether they believe this problem to be concentrated in the area of prolonged use, long use following the resolution of a balance of payments problem, or large use (including long use of large amounts of Fund resources).

83.   Directors may wish to comment on the following approaches to address potential issues relating to unduly long or large use of Fund resources:

(i)   with a view to addressing long use of Fund resources:

• introducing time-based repurchase expectations (REs) into stand-by arrangements and the EFF, and if so whether the profiles presented in Box 3 would be appropriate; or

• graduating the rate of charge with the time resources have been outstanding, either from the date of each purchase or from the date of the first purchase under an arrangement (with the latter requiring major changes to the structure of facilities), and if so how they would view the options presented in Figure 1 (or others they might find preferable); or

• introducing both REs and graduation of charges with time; or

• whether Directors believe there is potential to strengthen the early repurchase policy in a way that does not make it closely resemble REs, and if so how this might be achieved;

(ii)   with a view to addressing large use of Fund resources (including long use of large amounts): graduating the rate of charge with the amount of resources outstanding, and whether the options presented in Figure 2 (or others) would be desirable or acceptable.

84.   Directors may wish to comment on whether the following changes should be introduced into the Contingent Credit Lines (CCL):

• giving the member the strong benefit of the doubt, in the activation review, as to policy adjustment in the event of crisis; and specifying, upon approval of a commitment of CCL resources, the amount to be made available upon completion of the activation review—an amount which would be large (perhaps 100 percent of quota, or one third of the arrangement) either as a rule or as a normal practice;

• encouraging the member and the Fund to discuss, at the time of approval of a commitment of CCL resources, an illustrative list of circumstances in which the arrangement might be activated;

• reducing the initial surcharge on CCL resources to 150 basis points, and maintaining the same graduation with time as in the SRF, with a ceiling at 350 basis points; and

• amending rule I-8 of the Fund’s Rules and Regulations to base the commitment fee for all arrangements on scheduled purchases during each twelve-month period, so that the commitment fee on arrangements with CCL resources would be virtually eliminated.

Directors may also wish to comment on whether they see possibilities further to reduce the scope for judgment by the Fund in other areas of the activation review, and whether they see scope for new procedures that would make it easier for a member to exit from the CCL in case of policy slippages.

85.   On the EFF, Directors may wish to confirm the sense of the Board in the July discussion—that access to the EFF should be granted only in cases where a member’s balance of payments difficulties are expected to be relatively long-term, and that use of the EFF on a precautionary basis would be expected to be exceptional.

86.   Directors may wish to comment on possible procedures for stronger post-program monitoring—in particular:

• whether they see a case for more formal monitoring (with the involvement of the Board) of certain members after their arrangements have ended,

• if so, whether post-program monitoring should take place automatically for some members based on established criteria (e.g., the level of Fund credit outstanding), whether it should be based on ad hoc recommendations of the Managing Director, or whether there should be a presumption that members should engage in post-program monitoring when they meet certain criteria;

• in the event of formal post-program monitoring, whether they believe that publication of a Public Information Notice should be encouraged.

87.   The introduction of some of these policies would require Board decisions. The voting majorities required for these decisions are set forth in Table 8.

1/  In percent of total voting power, unless otherwise specified.
2/  See footnote 22.

Prepayment Clauses in Other Creditors’ Operations

In this paper, prepayment clauses in creditors’ financing operations are understood as clauses in official or commercial credit agreements with sovereign borrowers whereby mandatory repayments ahead of an agreed repayment schedule are linked to early repayments to other creditors, including to the Fund. These clauses are potentially relevant when designing or implementing policies targeting early repayments to the Fund, because of their possible negative financial consequences for member countries making repayments.

Prepayment clauses are common in commercial syndicated loans and Brady bonds,⁵⁵ but not widespread—if present at all—in other types of bonds, and particularly in the growing Eurobond market. Given that the composition of private capital flows to emerging markets has shifted away from syndicated loans in recent years, prepayment clauses to commercial creditors have become progressively less important in the 1990s, though they could become more relevant should syndicated lending again grow in importance in the future.

Implicit or explicit prepayment clauses are present in official creditors’ concerted lending operations, but usually not in their individual operations.

The Fund has tried to minimize the financial consequences to members of early repayments to itself by discouraging members from signing clauses in commercial credit agreements that link prepayments to early repurchase to the Fund, or by encouraging them to seek modifications of these clauses in anticipation of the possibility of an early repurchase to the Fund.⁵⁶ The Fund’s arguments in this respect are strengthened by the fact that early repurchases are not at the member’s discretion, but rather part of the repurchase structure of the Fund.

By contrast, this approach would not help to contain the indirect financial costs to individual members of an entirely incentive-driven policy to address long use, since incentives would result in advance repurchases that would be clearly voluntary, and thus could not help but trigger prepayment clauses.