Adequacy of the Global Financial Safety Net— Review of the Flexible Credit Line and Precautionary and Liquidity Line, and Proposals for Toolkit Reform
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

protracted global uncertainty combined with frequent episodes of capital flow volatility have intensified demand for liquidity support.

Abstract

protracted global uncertainty combined with frequent episodes of capital flow volatility have intensified demand for liquidity support.

Introduction

1. Enhancing the Fund’s toolkit for crisis prevention is a central element in the ongoing efforts to strengthen the global financial safety net (GFSN). Since the global financial crisis, emerging markets (EMs) have been highly exposed to short-term moderate volatility in international capital markets.1 With shocks emerging suddenly, unexpectedly, and relatively frequently, demand for global liquidity backstops has intensified and seems unlikely to subside in the foreseeable future. Despite significant reforms of the GFSN since the crisis, gaps remain in the global architecture. The standing network of bilateral swap lines among central banks could protect well against such volatility, but it is unavailable to most countries. Many countries have instead resorted to building reserves, which is not only very costly for these countries, but also entails potential systemic costs and coordination problems.2 There is appetite, therefore, from a group of members for a new Fund facility that would provide liquidity support (akin to central bank swap lines). This liquidity backstop could act as a complement to other elements of the GFSN. The IMFC and G20 have called on the Fund to work on proposals for such an instrument.

2. At the Board discussion in November 2016, Executive Directors agreed to consider more detailed proposals for a new liquidity backstop, but raised several concerns, most notably on:

  • the coherence of the lending toolkit, particularly the risk of instrument proliferation; consequently, they requested that any reform proposals be informed by a review of the Flexible Credit Line (FCL) and Precautionary and Liquidity Line (PLL);

  • the impact on Fund resources of both a new backstop and, more generally, all precautionary arrangements;

  • the qualification criteria, with an emphasis on the need for a high qualification bar for any facility without ex post conditionality; and

  • the proposed qualification process, especially the risk of undermining the quality and candor of surveillance through alignment with Article IV cycles, and the adverse signaling effect of “disqualification.”

3. Some Directors have also continued to express concern about prolonged use of FCL arrangements at high access levels, while the use of the PLL has remained lackluster. The 2014 review of the FCL and PLL called for improved exit strategies from FCL arrangements and a reassessment in subsequent reviews of the usefulness of the PLL.3 Since then, requests for successor arrangements have been approved for all three members with FCL arrangements, including two arrangements at significantly higher access levels. This has contributed to continued concern from part of the membership about prolonged use of the FCL. Meanwhile, the single user of the PLL has signaled its commitment to a gradual exit from the instrument.4

4. This paper proposes three complementary reforms to improve the Fund’s toolkit for crisis prevention. First, a new liquidity backstop is proposed as a special facility to address a key identified gap in the GFSN.5 To help address Directors’ concerns regarding qualification standards and process, the new backstop will have the same qualification standards as the FCL. The backstop will be available year-round—eschewing the earlier proposal to align it with the member’s Article IV consultation cycle—and will follow a process largely similar to the FCL. The key features of the backstop—revolving access, no exit expectations, extension of a Fund “offer”, and option for sole central bank signatory—were designed through extensive consultations with the membership. Second, to further address existing concerns regarding the qualification framework, the transparency and predictability of the qualification criteria for the FCL (which will also apply to the new liquidity backstop) will be improved. Finally, the PLL will be eliminated to avoid instrument proliferation and ensure toolkit cohesion. The paper also discusses possible options to change the Fund’s commitment fee structure aimed at promoting a more balanced use of Fund resources, including by unlocking resources for the new liquidity backstop.

5. These reforms are part of a broader set of proposals to improve Fund facilities and instruments to address gaps in the GFSN. These proposals include establishing a new (non-financing) Policy Coordination Instrument to help signal commitment to reforms and catalyze financing, and strengthening the cooperation between the Fund and the Regional Financing Arrangements.

6. The remainder of this paper is organized as follows: section two reviews the experience with the FCL and the PLL since 2014; section three discusses reforms to the existing toolkit; section four sets out the proposal for a new liquidity backstop; and section five reviews the resource implications of the proposed reforms.

Review of the FCL and PLL

7. The 2014 review did not find evidence of unjustified use of the FCL and PLL. The review, nevertheless, proposed to develop an External Economic Stress Index (ESI) to strengthen the discussions of external risks and help guide access discussions. Given concerns related to a lack of transparency and predictability in qualification, the review also unified the qualification areas for the FCL and the PLL, and added quantitative indicators of institutions to help underpin the assessment of a member’s institutional strength.

8. There have been no new users of the FCL or PLL since the 2014 review. Only three members (Colombia, Mexico, and Poland) have used the FCL, with a total of 18 arrangements since its introduction in 2009. Only one member (Morocco) has used the PLL, with a total of three arrangements since its establishment in 2011.6 This section summarizes the main findings of the review of the FCL and the PLL—focusing on exit, access, and qualification—which inform the subsequent reform proposals. Additional findings and the analytical background work are provided in Annexes I-IV.

A. Exit and Access

9. In response to some Directors’ concerns about prolonged use of the FCL, staff reports for successive FCL arrangements have been more explicit about the expectation for exit. The revised 2015 FCL guidance note highlights specific expectations for exit discussions in staff reports, which subsequent staff reports have followed. For example, the recent staff report for Colombia’s FCL request highlighted that “the authorities consider access to the FCL to be temporary and dependent on external conditions.” Reports have also listed specific external risks and the expectation of exit as these risks recede, and the authorities’ own letters have included language on exit expectations (Box 1). The Polish authorities, in particular, highlighted their efforts to communicate to market participants and the broader public their intention to continue with a smooth and gradual exit strategy, and noted the muted market reaction to date.

10. Staff analysis shows that reductions in access do not appear to be associated with adverse market reaction. The six cases of access reductions to date—four under the FCL and two under the PLL—have not been, on average, associated with marked changes in yields, spreads, or yield volatility (Annex I). Furthermore, Poland’s case highlights the benefits of outreach to financial markets, as it has reduced access on three occasions—most recently to 159 percent of quota in January 2017 (Annex II)—without adverse market reaction.

Authorities’ Language on Exit Expectations

This box quotes specific language on exit from the authorities’ written communications requesting an FCL arrangement.

Colombian authorities’ 2016 written communication:

“We expect that as the growth transition in China progresses smoothly, investor concerns about stressed emerging market economies recede, and monetary policy normalizes in advanced countries, uncertainties will abate, and associated commodity price and global financial risks will recede. We also expect the outlook in regional trading partners to improve gradually as they undergo adjustments to restore internal and external balance. With substantial reduction of some of the global risks affecting Colombia, we would intend to reduce access to Fund resources in any subsequent FCL arrangements, and to phase out Colombia’s use of the facility. Successful adjustment to permanently lower oil prices and the ongoing productive transformation of the economy should also build resilience and reduce future access to Fund resources.”

Mexican authorities’ 2016 written communication:

“…we intend to treat the arrangement as precautionary. We do not intend to make permanent use of the FCL. Conditional on a reduction of some of the global risks affecting emerging markets during the proposed FCL, including a dissipation of the risks of increased protectionism in advanced economies, and a smooth continuation of the process of normalization of U.S. monetary policy, we intend to cut access to Fund resources in the future, with a view of phasing out Mexico’s use of the facility.”

Polish authorities’ 2017 written communication:

“…we believe that the FCL would continue to play an important role in mitigating external risks in case of tail events. Nevertheless, we find that these risks have to be reflected upon in the context of Poland’s improved economic fundamentals and enhanced policy buffers. On balance, we consider that a new two-year FCL arrangement, albeit at a substantially lower access, would provide additional insurance against adverse external shocks, while conveying a strong signal of Poland’s commitment to exit the facility as soon as external conditions permit. In this context, we have continued our efforts to communicate our intention to proceed with a smooth and gradual exit strategy to market participants and the broader public. Our outreach has met only muted market reaction.”

11. Still, no member has recently exited from FCL or PLL arrangements, and two members have recently increased access. Colombia and Mexico requested increased access levels in 2016, reflecting a rise in external risks. Access under Colombia’s FCL arrangement doubled to 400 percent of quota (about USD 11.5 billion) and access under Mexico’s FCL arrangement increased by more than 30 percent to 700 percent of quota (about USD 88 billion).7

12. In light of elevated external risks, there is no evidence that either prolonged use on a precautionary basis or high access levels have been unjustified. Access discussions in successor arrangements were generally consistent with developments in the ESI, and the ESI helped point to specific economic risks (including from global growth, commodity prices, and turbulence in global financial markets) that were elevated. Mexico’s and Colombia’s requests for higher access levels in 2016 were consistent with ESIs that pointed to heightened risks. In addition, the assumed shocks (used to establish the access level) were chosen further into the tail of the distributions (see the red areas in Figure 1). Poland’s 2016 request for a lower access level coincided with an ESI that pointed to reduced risks, though assumed shocks were broadly unchanged.

Figure 1.
Figure 1.

Evolution in the Severity of Shocks Relative to Previous Arrangement1/

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: IMF staff reports.1/ Colors denote the severity of the shock in the kernel distribution and level of access relative to the previous arrangement (or review in the case of Poland in 2017). Red: more severe; yellow: unchanged; green: less severe; gray: not available. Table considers only the shocks as they pertain to the kernel distributions. Other shocks, such as “other capital outflows” are not considered.2/ Mexico: red color in the 2012 request is relative to the 2010 request as the 2011 request assumed MEX to become a net oil importer in 2011–12.3/ Morocco: yellow color is relative to the shock in the 2012 request.

13. The implementation of the ESI has, nevertheless, been somewhat uneven and could be improved. While ESI adverse scenarios were constructed drawing on Fund flagship reports, such as the WEO and the GFSR, shocks were not always based on the most recent scenarios (see Annex III). Moreover, the size of assumed global shocks varied across shock scenarios in recent program requests. That said, staff reports have recently become more transparent regarding the assumptions underpinning the access scenario; both the Mexico and Colombia requests have listed all relevant assumptions for not only the current arrangement but also the previous two arrangements.

14. While the assumed use of reserves in adverse scenarios has become more prevalent, there may be scope for an even greater reliance on reserve drawdown in adverse scenarios, which would in turn support the case for lower access levels. While external risks have pushed in the direction of higher access levels, developments in fundamentals and the use of buffers have had an important role in reducing access levels (Annex IV). Improved economic fundamentals and buffers, for example, reduced financing needs in the 2017 Poland FCL request. Assumed use of reserves in adverse scenarios has also become more prevalent. Nonetheless, reserve levels in extreme stress scenarios have generally remained within the adequacy range (see text chart), in part reflecting concerns about adverse market reactions to a large and rapid reserve drawdown. The 2014 review and the 2015 FCL guidance note suggest considering 80 percent of the Fund’s ARA metric as a lower threshold in an extreme stress event (see Annex IV). This would imply that additional room for reserve drawdown may be available.

uA01fig01

Reserve Adequacy in Current Arrangements

(Percent of reserve adequcy metric; based on gross international reserves)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: IMF staff reports; and IMF staff calculations.1/ Augmented by buffer for commodity exports.2/ Net international reserves declined to 89 percent of the ARA metric in the adverse scenario.3/ Financing gap defined to maintain reserves at 90 percent of the standard ARA metric (before adjusting for capital controls). After adjusting for capital controls, baseline reserves amounted to 136 percent of the ARA metric and reserves in the adverse scenario remained adequate.

15. In addition, the discussion of exchange rate assumptions and the implied impact on financing needs in adverse scenarios could be strengthened. As noted in the FCL guidance note, “the adverse scenario should also take into account whether there is a case for an orderly exchange rate adjustment.” Furthermore, temporary exchange rate depreciation could impact the current account. However, while adverse scenarios have often implicitly built in exchange rate depreciation, which has led to import compression (e.g., Poland’s most recent FCL request), this has generally not been clearly communicated. Hence, the assumed impact of the exchange rate depreciation could be clarified in the description of the adverse scenario.

B. Qualification

16. The 2014 review modified the qualification criteria for the FCL and the PLL to improve the transparency and predictability of qualification and added new indicators to strengthen the assessment. The changes included (i) a clarification of the PLL qualification areas by introducing a direct mapping between the five PLL qualification areas and the nine FCL qualification criteria, strengthening comparability; and (ii) the introduction of indicators to inform the assessment of institutional strength. However, the review also pointed to ambiguity over what constitutes underperformance relative to the qualification requirements, especially with respect to the PLL.

17. The FCL criteria are currently assessed without clearly specified benchmarks. The examination of recent FCL and PLL staff reports highlights the lack of specific tools and benchmarks (i.e., specific indicators with thresholds) that can guide judgment in several of the nine qualification criteria (see Box 2). The absence of benchmarks against which the assessment is made reduces predictability in qualification. The assessment of a sustained track record of implementing very strong policies has also been based entirely on judgment, since neither the benchmarks nor the period over which they are assessed are defined. The next section puts forward proposals for refining the assessment of the nine qualification criteria and a member’s track record. No changes are proposed to the assessment of institutional quality, which warrants a holistic approach.

The Use of Quantitative and Qualitative Assessments in Qualification

This box examines the balance between quantitative and qualitative assessments used to measure performance against the nine qualification criteria.

Qualification area I: External position and market access

1. Sustainable external position. The assessment of the external position has benefitted from readily available tools, particularly on debt sustainability analysis and current account and real exchange rate assessment relative to fundamentals. As such, the assessment has been largely analytical. Nonetheless, specific tools to assess the level of external (as opposed to public) debt remain absent.

2. A capital account position dominated by private flows. While staff reports have referred to capital account flows and the extent of privately-held external debt, there is no specific threshold available. Hence, the assessment is entirely judgment-based.

3. A track record of steady sovereign access to international capital markets at favorable terms. Assessment for the current FCL/PLL users has generally benefitted from readily available information on access to international capital markets. However, the benchmark against which the assessment should be made is not specified.

4. A reserve position that is relatively comfortable when the arrangement is requested on a precautionary basis. Reserves have been assessed against standard metrics. As such, the assessment has been largely analytical.

Qualification area II: Fiscal policy

5. Sound public finance, including a sustainable public debt position. While judgment is applied in the assessment of fiscal policy and rules, existing tools to assess debt sustainability have provided an analytical framework to guide such judgment.

Qualification area III: Monetary policy

6. Low and stable inflation, in the context of a sound monetary and exchange rate policy framework. Deviations of inflation from the target (for inflation targeters) have generally been attributed to temporary, external shocks (see, for example, the 2016 FCL staff report for Colombia and the 2015 and 2017 FCL staff reports for Poland), highlighting that medium-term inflation expectations have remained anchored. Overall, however, the benchmark against which the assessment is made is not specified.

Qualification area IV: Financial sector soundness and supervision

7. Sound financial system and the absence of solvency problems that may threaten systemic stability. Staff reports have assessed banking sector capital against corresponding capital requirements, and used stress tests to assess resilience. However, the basis for assessing the overall soundness of the financial system could be made clearer.

8. Effective financial sector supervision. Staff reports have often referred to implementation of FSAP recommendations and other steps to strengthen supervision and stability.

Qualification area V: Data adequacy

9. Data transparency and integrity. Staff reports have benefitted from references to the latest data ROSCs, timely data provision, and SDDS observance, which have helped create benchmarks for data transparency and integrity. Overall, this has provided substantial guidance to inform judgment.

Proposed Changes to the Existing Toolkit

A. Qualification for the FCL

Making the qualification criteria more transparent would help ensure evenhanded treatment and give members a clearer sense, ex ante, as to whether they would meet (or continue to meet) the qualification criteria. As noted above, while judgment is essential in making the FCL qualification assessments, the role of indicators and thresholds in guiding judgment can be more clearly specified. Accordingly, the paper proposes to use a set of core indicators, based on existing analytical tools, to guide judgment in FCL qualification.

18. The paper proposes to maintain the existing high FCL qualification standard. To approve an FCL arrangement, the Fund must assess that the member (i) has very strong economic fundamentals and institutional policy frameworks; (ii) is implementing—and has a sustained track record of implementing—very strong policies; and (iii) remains committed to maintain such policies in the future.8 In addition to a very positive assessment of the member’s policies by the Executive Board in the context of the most recent Article IV consultation, qualification is assessed against nine criteria.

19. The proposed changes to the qualification framework retain all existing qualification criteria but specify more clearly the role of indicators in guiding judgment. To provide a more transparent and predictable basis for assessing whether a member meets the qualification criteria, the framework would be anchored in a set of core indicators with thresholds that are based on established analytical frameworks (e.g., metrics such as the ARA and bottom-line assessments of a member’s external position).9 These core indicators are drawn from the indicators endorsed by the Executive Board in 2014.10

20. The bottom-line assessments on each criterion will remain a judgment, but will be clearly guided by the proposed set of core indicators. The set of core indicators and their thresholds will constitute a key element in determining whether the FCL criteria are met, but do not obviate the need to consult other sources of information—including the broader set of Board-approved indicators—deemed relevant for the bottom-line assessment of any given criterion in a specific country context. This approach would help address the issue of ambiguity over what constitutes underperformance relative to the qualification requirements identified in the 2014 review. Moreover, as under the current framework, very strong performance against all qualification criteria will not be necessary to secure qualification. However, significant shortcomings on one or more of these criteria—unless there are compensating factors underway—would generally signal that the member is not among the very strong performers for whom the FCL is intended. The specific proposal is set out in Box 3, and has been designed to ensure that the overall strength of the FCL qualification standard remains unchanged.

Qualification Criteria and Proposed Core Indicators

This box discusses the core indicators and thresholds proposed to underpin the overall assessment of the nine qualification criteria for the FCL. The bottom-line assessments on each criterion will be informed by other sources of information deemed relevant, some but not all of which are listed here. 2

1. A sustainable external position. Requires the member’s external position to have been assessed, in the most recent Board document (Article IV or ESR), as “broadly consistent”, “moderately stronger (weaker)”, “stronger”, or “substantially stronger” than implied by fundamentals and desirable policies.2 This assessment implies that members with “weaker” or “substantially weaker external positions” would not meet the criterion. The asymmetry in the assessment follows the reasoning that “weaker” or “substantially weaker” external positions (e.g., high current account deficit or net foreign liabilities, overvalued exchange rate, etc.) constitute early warning indicators for impending BOP crises.

2. A capital account position dominated by private flows. Requires public flows to account for less than half of a member’s direct, portfolio, and other asset and liability flows, on average in the past three years.3

3. A track record of steady sovereign access to capital markets at favorable terms. Requires public sector issuance or guaranteeing of external bonds or disbursements of public and publicly-guaranteed external commercial loans in international markets during at least three of the last five years for which data are available, in a cumulative amount over that period equivalent to at least 50 percent of the country’s Fund quota at the time of the assessment.4 The indicator also requires that the member did not, in staff’s assessment, lose market access at any point in the last 12 months. Following the Fund’s framework for loss of market access, deteriorating funding conditions and adverse changes in issuance patterns (volume, maturity, and frequency of issuance) that cannot be explained by funding needs would be indications that the member has indeed lost market access.5

4. When the arrangement is requested on a precautionary basis, a reserve position which— notwithstanding potential BOP pressures that justify Fund assistance—remains relatively comfortable. Requires reserves to have been greater than 100 percent of the ARA metric on average over three (the current and the two previous) years, and not below 80 percent in any of these three years. By including a lower—but not an upper—threshold for reserves, the assessment follows the reasoning that excess reserve holdings, while possibly undesirable from a systemic perspective, do not constitute a vulnerability for the member.6

5. Sound public finances, including a sustainable public debt position. Requires the member’s public debt to be assessed as sustainable with a high probability. The high probability assessment would explicitly take into account risks to public finances not immediately visible in current public debt projections.

6. Low and stable inflation, in the context of a sound monetary and exchange rate policy framework. Requires the member to have maintained single-digit inflation over the past five years. The bottom-line assessment would consider if the member’s performance is seen to reflect favorable external conditions and there are grounds to question the ability of its policy framework to maintain low inflation under normal circumstances. It would also consider persistent deviations from stated inflation targets, as well as sustained deflation, to the extent that it reflects deficiencies in the monetary policy framework.

7. Sound financial system and the absence of solvency problems that may threaten systemic stability. Requires the average capital adequacy ratio for the banking sector to be above regulatory thresholds, and that the most recent Article IV did not highlight significant solvency risks or recapitalization needs. The bottom-line assessment would consider other financial soundness indicators, as well as any relevant stress tests conducted by staff, to provide a forward-looking perspective. It would also reflect potential problems in large and systemic banks that may be masked by system-wide averages.

8. Effective financial sector supervision. Requires that the most recent FSAP or Article IV report did not raise substantial concerns regarding the supervisory framework. The bottom-line assessment would consider any significant changes in conditions since the latest FSAP.

9. Data transparency and integrity. Requires that the member is an SDDS subscriber or has made satisfactory progress toward meeting the SDDS requirements.

1 While judgment is important for all nine criteria, it is particularly important for criteria 6, 7, and 8. 2 The assessment of a member’s external position as per the mandatory external sector assessment in surveillance takes into account the following five key areas: current account (CA), real exchange rate, foreign exchange intervention and reserves, foreign assets and liabilities, and capital and financial account. The bottom line assessment of a member’s external position falls into one the following seven categories, guided by the corresponding indicative ranges for the staff-assessed CA gaps (in percent of GDP) with considerations of all other areas: (i) substantially stronger (CA gaps more than 4 percent); (ii) stronger (CA gaps between 2 and 4 percent); (iii) moderately stronger (CA gaps between 1 and 2 percent); (iv) broadly consistent (CA gaps between -1 and 1 percent); (v) moderately weaker (CA gaps between -2 and -1 percent); (vi) weaker (CA gaps between -4 and -2 percent); and (vii) substantially weaker (CA gaps less than -4 percent). 3 Public flows are flows to and from the domestic public sector, and are defined as the sum of the absolute values of reserve assets flows, and general government and central bank portfolio and other debt liability flows. In the absence of data for a large sample of countries, other official asset and liability flows of the public sector are assumed to be zero. 4 This indicator assessment broadly follows the market access criterion for (graduation from) PRGT eligibility. The bottom-line assessment will consider if there is convincing evidence that the sovereign could have tapped international markets on a durable and substantial basis, even though the scale or duration of actual public sector borrowing fell short of the specified thresholds. This would be a case-specific assessment, informed by factors such as the volume and terms of recent actual borrowing in international markets and the sovereign credit rating. 5 A methodology for making this assessment was articulated in “The Fund’s Lending Framework and Sovereign Debt—Further Considerations”, IMF Policy Paper, April 2015. 6 See Annex IV for an empirical justification of the 80 percent threshold. The overall assessment could consider other reserve metrics if the ARA metric is deemed inadequate for judging the member’s reserve position. This assessment should generally be reflected in recent Article IV reports.

21. The assessment of policy track record will be anchored in the same set of core indicators. Determination of whether a member meets the track record requirement would be guided by an assessment of the qualification criteria in each of the five most recent years.11 The bottom-line assessments on each criterion would follow the same approach as for the current year, including by drawing on additional information, e.g., the appropriateness of a member’s policy response to past bouts of capital flow volatility, and the extent to which individual indicator outcomes reflect favorable external conditions.12

B. Other Operational Considerations for the FCL

External economic stress index

22. The implementation of the ESI could be strengthened to enhance the comparability of external risks across countries. Particularly, staff plans to amend the guidance to ensure that downside scenarios are based on the most recent flagship reports (i.e., WEO, GFSR, or spillover reports). While the impact of a given shock would likely differ across countries, reflecting different exposures and hence different ESI weights, the underlying shock scenario would be expected to be similar across countries with arrangements falling close in time. However, should different assumptions (e.g., different shock size) be warranted, this could be clarified by specifying the context surrounding the given shock. Overall, this should help ensure that assumptions related to global shocks are similar across reports that fall close in time.

Use of international reserves

23. The potential impact on access levels of additional reserve drawdown could be considered to support lower access levels. In past cases, reserves have generally been assumed to remain adequate, even if FCL guidance provides scope to draw them below the relevant adequacy thresholds in adverse financing needs scenarios. Discussions of the access level during the internal review process could take into account alternative reserve drawdown scenarios (e.g., access levels under a lower bound of reserves, but with consideration to the pace of reserve drawdown to avoid adverse market reactions, see Annex IV). A decision to maintain significantly higher reserve levels in the adverse scenario should be clearly justified.

C. Commitment Fees

This section discusses considerations for raising commitment fees on high access commitments. Although the review has not found any evidence of undue use of the FCL, higher fees could nonetheless promote a more balanced use of Fund resources, including between the FCL and the proposed new liquidity instrument.

24. Over recent years, a number of Directors have called for strengthening exit incentives from prolonged precautionary use of Fund facilities at high access levels. When this issue was last discussed, in the context of the 2014 Review of the FCL and PLL, views were divided. Many Directors saw merit in further work on stronger incentives to encourage timely exit, such as time based commitment fees, with a few suggesting also a time limit on the use of the FCL and PLL. Many other Directors saw no compelling evidence of problems with exit, noting that exit should depend on the state of the external environment facing the member.13

25. While the current review reaffirms staff’s earlier view that the FCL has been used as intended, maintaining large commitments on a precautionary basis for extended periods entails costs to the membership. Specifically, prolonged use of large arrangements on a precautionary basis ties up a significant share of the Fund’s finite resources. It also potentially entails costs to creditor members, particularly to non-reserve-currency-issuing members in the Financial Transactions Plan (FTP) that maintain liquid assets to meet potential calls under such arrangements.14

26. Against this background, two commitment fee reform options could be considered to promote a more balanced use of Fund resources (see Annex V for details). Both are intended to strengthen price-based incentives to discourage large-scale use of precautionary arrangements. Such use typically arises in the context of the FCL, but as commitment fees apply to all GRA arrangements whether they are treated as precautionary or non-precautionary by the member, the fees would also apply to arrangements that are expected to be drawn. Ramifications for arrangements other than FCL would therefore need to be taken into account in the design of the specific policy.

27. A first option would be to steepen the current commitment fee schedule. The current upward-sloping fee structure was introduced as part of the broader reforms in 2009, including the creation of the FCL, to discourage unnecessarily high precautionary access and contain risks to the Fund’s liquidity. Building on the existing commitment fee structure, the current 60 bps rate applied at access above the highest fee threshold of 575 percent of quota could be increased by, say, 10–20 bps. This would be an operationally simple and transparent means to strengthen price-based incentives to discourage prolonged commitments under arrangements with very high access, which are typically precautionary FCLs, while not excessively penalizing high access where such use is warranted.15 A drawback would be that the structure would apply uniformly to all GRA arrangements irrespective of whether these are treated as precautionary, and could therefore also have implications for large drawing arrangements that go off-track.16

28. An alternative option would be to introduce a time-based commitment fee, which would be intended to discourage large-scale prolonged precautionary commitments. An arrangement could become subject to an additional fee once the level of undrawn credit has remained above a specified threshold for a defined period (“duration trigger”), say 4 years (see Annex V for details).17 For arrangements currently in effect, the clock toward meeting the duration trigger would thus start, at the earliest, in mid-2013 (see ¶33 on grandfathering). The arrangements would remain subject to the fee until the level of undrawn credit falls below this threshold, through purchases or upon expiration.18,19 The drawback of a time-based fee is that it would not be state-dependent, but rather come into effect on a pre-determined schedule, irrespective of underlying external conditions, and could become payable when risks remain elevated. Thus, it could penalize members that experience prolonged periods of adverse external circumstances, including those facing new risks. For this reason, it would be important that any time-based commitment fee be set at a modest level, e.g., 10–20 bps.

uA01fig02

Access under Successive FCL/PLL Arrangements

(Percent of post-14th review quota)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: IMF staff reports; and IMF staff calculations.

29. If adopted, the threshold for the time-based commitment fee should be set at a sufficiently high level to appropriately target large precautionary commitments and mitigate the risks of unintended consequences. The threshold could be set, for example, at the higher level-based commitment fee threshold of 575 percent of quota. Such a high threshold could address Directors’ concerns regarding particularly large and prolonged commitments, in practice applying only to prolonged use of high access FCL arrangements.20 A lower threshold would increase the risk of having unintended consequences for other high access arrangements that are expected to be drawn.

30. While consideration could be given to applying the proposed time-based commitment fee to all GRA arrangements, given the purpose of this fee to discourage large-scale prolonged precautionary commitments, it would be appropriate not to apply it to the extended arrangements and the new liquidity facility. If the fee were applied to all GRA arrangements, extended arrangements, which are generally not provided on a precautionary basis21 could be subject to the fee in instances that are not aligned with the fee’s purpose, such as when large arrangements go off-track on a protracted basis.22 Arrangements under the new liquidity facility are envisaged to have no exit expectations, and should therefore not be subject to the time-based commitment fee either.

31. Under either of the options above, members could still request new arrangements or renewals if warranted by external risks. Under the current commitment fee structure, costs to users of the Fund’s precautionary facilities are materially lower than staff’s estimates for the cost of accumulating reserves and the fees for comparable arrangements charged by other IFIs (Annex VI). Increasing the fees through either of the above options would bring the costs closer to these comparators (see Annex V for numerical examples).

32. However, in staff’s view, both options have significant drawbacks. As argued above, steepening the slope of the current fee structure would be administratively straightforward to implement, but would not be specifically targeted to prolonged precautionary use of Fund resources. A time-based fee would be more targeted on prolonged use but is a rather blunt tool, as it would be triggered irrespective of the state of external risks faced by members. It would also significantly increase the complexity of the Fund’s fee structure and be operationally burdensome, requiring careful calibration of several design elements (see Annex V).

33. Should the Board decide to implement one of the above options, staff would propose that current arrangements be grandfathered. Increased fees would not be payable under existing arrangements, but the new policy would apply to all new requests for use of Fund resources (e.g., augmentation or new arrangements) approved after the new fee structure comes into effect. In the case of a time-based commitment fee, past repeat use could be taken into account when determining whether a time-based fee is triggered.23 As such, any augmentation or successor arrangements to the current FCL arrangements could become subject to a time-based fee to the extent that the trigger condition for the fee has been met and the level of undrawn credit under the augmented or new arrangement continues to exceed the chosen threshold. Alternatively, the duration trigger could be applied from the date of effectiveness of the decision or the approval date of a new arrangement.

34. Alternative approaches such as charging an additional fee at times of lower external risks or imposing time limits on use of FCLs were also considered, but would be subject to significant legal and policy constraints. In the 2014 Review, some Directors proposed that an additional fee could be levied on the basis of the state of risks faced by a member or that a time-based commitment fee could be waived in certain circumstances. Such options were noted to be impermissible under the Articles of Agreement, as any difference in fees paid by members must be related to the use of Fund resources and cannot be based on members’ specific circumstances. A legally possible alternative would be to uniformly waive some portion of commitment fees, e.g., the time-based fee if introduced, for all members, on the basis of some global stress indicator. However, such an approach would make the current fee structure much more complex to administer and unpredictable for users, while still leaving the possibility that a member could face higher fees when it was subject to idiosyncratic risks. A few Directors have also raised the idea of imposing a time limit on members’ use of the FCL. Staff does not support such a limit, as it could come into play while external risks remain elevated, and would thereby limit the Fund’s ability to provide support to members that would otherwise be eligible.

D. Elimination of the PLL

Eliminating the PLL would help maintain a streamlined and coherent toolkit.

35. The PLL has had two users. The facility was introduced in 2011 as a replacement for the PCL. Despite offering more flexibility than its predecessor, including use for actual as well as potential BOP needs, and a new 6-month term “liquidity window” for short-term BOP needs, Morocco is the only member that has opted to use the PLL.24 This likely reflects the issue of tiering vis-à-vis the FCL, as well as the difficulty of identifying members that have strong fundamentals and remaining vulnerabilities, as per the 2014 review.

36. Staff proposes to eliminate the PLL.* Given the low use and general Board concerns about instrument proliferation, the paper proposes to eliminate the PLL. This might open a gap in the toolkit, but the gap should be small, reflecting the low use. Members not qualifying for the new liquidity facility/FCL could continue to have access to the precautionary SBA. Moreover, removing the PLL would eliminate the tiering vis-a-vis the FCL and avoid instrument proliferation. Staff assesses that, on balance, the benefits from eliminating the PLL would outweigh the costs. The existing PLL arrangement would be grandfathered (i.e., it will remain under pre-existing policies until it expires).

Design of the Short-Term Liquidity SWAP

The Short-term Liquidity Swap (SLS) will provide liquidity support within normal access limits for potential short-term moderate BOP needs, reflecting capital account pressures arising from external developments. It will have a revolving access feature to cover against repeated shocks, and there will be no exit expectation for as long as the member qualifies and has the special BOP need. Qualification for the SLS will be based on the same standards and criteria as the FCL (see Table 1 for a comparison of the SLS and FCL). This approach will facilitate the transition from the FCL to the SLS, if the special BOP need requirement is met (and vice-versa, if needed). The high qualification bar, 12-month duration of the arrangements, annual requalification, short repurchase period, and the facility-review clause will help to safeguard and conserve Fund resources.

Table 1.

Comparison of the Key Features of the SLS and FCL

article image

A. Objective and BOP Need

37. The 2008–09 review of the Fund’s lending toolkit marked a shift toward more flexible instruments that could address all types of BOP problems. The review recognized that, in practice, it is difficult to predict ex ante the nature, magnitude, or duration of a crisis, and to distinguish between different types of BOP problems, especially potential BOP needs, underscoring the importance of flexible access levels and the possibility of longer repurchase periods.25 Even in circumstances where the sources of a shock were exogenous, one could not assume that policy adjustment would not be necessary. It was therefore agreed to place greater reliance on the credit tranches as the vehicle for the delivery of GRA financing and to eliminate special facilities that were designed to address special, and narrower BOP needs, many of which had been seldom or never used.

38. Since that time, however, the global economy has changed. As noted in previous papers, the global economy is experiencing a period of protracted uncertainty, marked by frequent episodes of volatility.26 Global financial cycles have increased in amplitude and duration, capital flows have become more volatile, and nonbank finance channels have expanded. Demand for a liquidity backstop has intensified, particularly from EMs, which are increasingly exposed to liquidity events triggered by external events and channeled through the capital account (see Box 4).

39. Staff proposes to establish the SLS as a special facility to provide “swap-like” liquidity support for potential short-term BOP needs resulting from exposure to volatile capital flows.** The SLS will be anchored in a special BOP need, namely “potential short-term BOP difficulties reflected in pressure on the capital account and the member’s reserves resulting from volatility in international capital markets.” Such difficulties are expected to be limited in scale, and to require, at most, a fine-tuning of central bank policies (e.g., exchange rate adjustments, foreign exchange market intervention, and/or interest rate changes).

B. Qualification

40. SLS qualification will be based on the same standard and criteria as the FCL. The high qualification bar for the FCL aims to ensure that a member, when faced with an actual BOP problem of any kind, will undertake the necessary adjustment to correct the problem, and therefore, ex post conditionality is not needed, regardless of the size or nature of the shock. The high qualification standard for the SLS will provide similar confidence that the member will take the necessary policy actions to resolve its BOP problem under moderate shocks and repay the Fund, without the necessity for ex-post conditionality.27 28 It would also provide assurance that the member is inherently less susceptible to deeper or more protracted BOP problems. Discussions of subsequent SLS arrangements would also consider the member’s use of Fund resources under previous SLS arrangements in accordance with the special nature of the facility.

C. Terms

12-Month repurchase obligation

41. In light of the short-term nature of the special BOP need, staff proposes the establishment of a 12-month repurchase obligation.29 The proposed one-year repurchase period is consistent with the observed duration of the liquidity shocks that the SLS aims to cover. The acute phase of the recent liquidity events—including the 2013 Taper Talk, which was the most severe post-GFC liquidity event for the EMs—typically lasted a few months, while the impact of the shock dissipated within about a year (Box 4).30 31 Assuming a member draws on the SLS at the onset of the shock, a three-month repurchase period (in line with the central bank swap lines) would likely fall within, or very close to, the acute period of the crisis, putting undue additional strain on the BOP and reserves. A three to five-year repurchase period (in line with the instruments in the credit tranches), in contrast, appears unnecessarily long for a short-term liquidity shock.32 Establishing the liquidity backstop as a special facility with a repurchase period shorter than under the credit tranches will require an Executive Board decision adopted by an 85 percent majority of the total voting power.33

Recent Liquidity Episodes in Emerging Markets: Stylized Facts1

The introduction of the SLS is at least in part motivated by the major liquidity events that have affected emerging markets (EMs) since the Global Financial Crisis (GFC). This box summarizes stylized facts on the evolution of key variables and policy responses during these episodes, which inform the design of the facility.

The analysis is based on four major exogenous events that have generated financial stress in at least half of the 22 major EMs. These events are the GFC, the 2011 Euro Debt crisis, the 2013 Taper Talk, and the 2015 China market correction. The GFC is included primarily as the reference for a large global shock.

uA01fig03

Financial Pressures in Emerging Markets

(Percent of large EMs; VIX level)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Source: IMF staff calculationsNote: Daily data aggregated at weekly level. Sample includes 22 large EMs: Argentina, Brazil, Bulgaria, Chile, China, Colombia, Hungary, India, Indonesia, Kazakhstan, Malaysia, Mexico, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Turkey, Ukraine, and Venezuela

The liquidity events follow the same pattern: a short acute phase followed by a gradual normalization within about a year. The charts on the next page show the evolution of portfolio flows (as measured by Emerging Portfolio Fund Research, EPFR2), the exchange market pressure index (EMP, which combines exchange rate movements and use of reserves), and reserves. The first month of the crisis is the most severe one (as measured by the EMP), and the acute phase lasts around one quarter. During this period, even better-performing EMs (75th percentile) experience portfolio outflows. The outflows reverse within two to three quarters, and by the first anniversary of the crisis, portfolio investments either surpass their original level, or at least stabilize at a new equilibrium level (Taper Talk). Some major liquidity episodes are followed closely by smaller market-moving events (oil price collapse, Brexit, the US elections). Finally, while the GFC had a profound impact on growth outcomes, EM capital inflows rebounded as quickly in the GFC as in the more modest episodes that followed. Only the GFC had a profound impact on EM reserves, while in subsequent episodes reserves of possible SLS qualifiers showed little downward movement and remained at or above 100 percent of the ARA metric, suggesting a reliance on exchange rate flexibility.

The response is generally limited to central bank’s policies. The 2013 Taper Talk was the most severe post-GFC liquidity event for EMs, prompting a multi-prong policy response in several countries (see table below). But even in this extreme event, the response relied almost exclusively on FX interventions, liquidity measures, and monetary tightening, while fiscal and structural measures were deployed only in a few cases.

uA01fig04

Portfolio Flows and Exchange Market Pressure During Recent Liquidity Episodes

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: EPFR, IFS and IMF staff calculations.Note: Sample comprises Argentina, Brazil, Bulgaria, Chile, China, Colombia, Hungary, India, Indonesia, Kazakhstan, Malaysia, Mexico, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Turkey, and Ukraine. The solid blue line is the sample median and the shaded areas the sample interquartile range. The dotted line is the median for potential SLS qualifying EMs.

Policy Response Following the 2013 Taper Talk

article image
Sources: IMF desk economists; IMF FSAPs; and IMF staff reports. Adapted from Sahay et al., 2014, “Emerging Market Volatility: Lessons from The Taper Tantrum ,” IMF Staff Discussion Note. 3/ At least ten bank days of disorderly market conditions, as measured by the IMF DMC index.
1 The box focuses on EMs, as previous GFSN papers have identified them as the main beneficiaries of a new liquidity instrument given gaps in the current GFSN. 2 EPFR flows cover only a fraction of BOP-recorded flows. However, EPFR’s monthly data allows for a more granular analysis. BOP and EPFR trends look similar when compared on a quarterly basis.

Revolving access

42. Access would be defined as a limit on the stock of Fund credit outstanding under the SLS, up to a maximum of 145 percent of quota, as justified by the member’s potential BOP need. Access up to this level, together with the revolving feature (see below), should be sufficient to reinforce confidence in the member’s ability to manage relatively small-scale, short-term volatility.34 Actual access for a qualifying member would be based on an assessment of the member’s potential need for Fund resources considering other sources of financing, and maintaining a reasonable level of reserves and its capacity to repay, similar to the FCL.

43. The SLS would also have a revolving access feature, aligned with the special BOP need that the facility aims to address. The revolving feature enables repeated purchases and repurchases within an arrangement and across successive arrangements, so that members have liquidity support against shocks which are frequent and short-lived (see Box 5).35 By defining access as a limit on the stock of Fund credit outstanding, a member would have the right to make partial or full purchases at any time during the arrangement. Unlike other Fund facilities, a full drawing of access under the arrangement would not terminate the arrangement. Any repurchase would reconstitute the member’s right to purchase up to the maximum access approved under the arrangement in effect.36 If, at the time of the approval of a successor SLS arrangement, there is credit outstanding from an earlier SLS, initial rights to purchase under the new arrangement would be limited to the difference between the newly approved access and any credit outstanding under the facility. If the full 145 percent of quota were approved and drawn under an earlier arrangement, a member that continues to qualify would have the right to purchase under a successor arrangement only after credit has been repurchased.

44. Access under the SLS would not trigger the application of the exceptional access framework, similar to the FCL. In the case of the SLS, access is capped below exceptional access levels (i.e., 145 percent of quota). Moreover, the proposed rigorous qualification criteria include elements similar to the exceptional access criteria, including the requirement for sustainable debt with high probability. The SLS would thus, like the FCL, not need to be subject to the exceptional access framework.

Duration of arrangement

45. Individual arrangements will be approved for a period of 12 months. If qualification criteria are met, following Board approval, the arrangement will remain in place for 12 months unless unilaterally terminated by the member. There would be no restrictions on the Board’s approval of successor SLS arrangements for a member, provided that the member continues to meet the qualification criteria. There would be no expectation that members would exit the facility when their current arrangement expires. Repeat use will require that users continue to qualify by meeting the qualification criteria, and that they continue to wish to avail themselves of the facility. Members will be encouraged to initiate the process for re-qualification prior to the expiry of an existing arrangement if they wish to ensure uninterrupted access to Fund resources under the SLS.37 Several features of the SLS are designed to ensure that the temporary nature of the use of Fund resources is preserved: (i) each arrangement has duration limited to 12 months; (ii) approval of the initial and all successor arrangements under the SLS will require a decision by the Executive Board; and (iii) repurchases of amounts purchased under the SLS will need to be made within a shorter repurchase period (12 months) than current GRA instruments.

Operational Aspects of Revolving Drawing Rights Under the New Facility

This box describes the operational aspects of the revolving access feature for the SLS, where purchases followed by early repurchases re-establish purchase rights both within and across successive arrangements. It demonstrates how the Fund’s commitment would be for a maximum stock of credit, not a maximum flow of purchases.

Repurchases would have a “first out, first in” rule whereby, in the case of multiple purchases, any repurchase would be automatically applied to credit under the SLS that has been outstanding for the longest time (falling due soonest).1 This is in contrast to the credit tranches or extended arrangements where members making early repurchase(s) have the option to apply the early repurchase to any outstanding obligation. The proposal for the SLS would streamline the processing of Fund transactions and limit operational risk, while having only a limited impact on members.

Three stylized examples are presented below to illustrate how access under the revolving feature would work. The examples are based on a member that has 145 percent of access under the SLS, faces a balance of payments shock 3 months after its approval and purchases 100 percent of quota. The purchase would trigger a 12-month repurchase obligation and the member would continue to have the right to purchase up to 45 percent of quota during the remaining 9 months of the arrangement.

Example 1: Early repurchase and no further purchases. 9 months after the approval of the arrangement (i.e., 6 months after the original purchase) the member’s BOP position has improved, and the member makes an early repurchase for 100 percent of quota. This early repurchase (which is within the 12-month obligation term) restores, pro tanto, the rights to make purchases over the remaining life of the arrangement up to the stock limit of 145 percent of quota.

Example 2: Multiple purchases, across two successive arrangements. At the end of the first arrangement, the member has not made an early repurchase. The Fund considers that the member remains qualified to use the instrument, and the member opts for a new 12-month arrangement. At the time of approval, the member is entitled to a purchase up to 45 percent of quota (which would bring the outstanding stock of credit to 145 percent of quota). If the member makes a purchase of 45 percent of quota shortly after the approval of the successor arrangement and a repurchase of 60 percent of quota two months later, the repurchase would first be applied to retire part of the obligation associated with oldest purchase (i.e., the purchase of 100 percent of quota made under the first arrangement). If the member makes another repurchase of 50 percent of quota one month later, 40 percent of quota of the repurchase would be used to retire the remaining 40 percent of the first purchase and 10 percent of quota would be used to retire early part of the second purchase. The member’s right to make purchases during the rest of the successor arrangement period is restored pro tanto (to 110 percent of quota, representing the two repurchases).

Example 3: Different access levels across two successive arrangements. If at the start of the second arrangement access is reduced to 120 percent of quota, reflecting the member’s lower potential BOP need, the member is then at the beginning of the second arrangement eligible to make a purchase of 20 percent of quota given the outstanding credit of 100 percent of quota under the first arrangement.

1 Where a member has preexisting GRA credit outstanding from a purchase made outside of the SLS, as well as GRA credit outstanding under the SLS, and the member makes an early repurchase, the member will still have the election to apply the early repurchase to non-SLS GRA credit, or to SLS outstanding credit. To the extent that the repurchase is applied to outstanding purchases under the SLS, the “first out, first in” rule will apply.

Charges and fees

46. Staff proposes a commitment fee and service charge structure for the SLS that takes into account the short-term BOP need addressed by, and the revolving access under, the SLS policy. Specifically, staff proposes a non-refundable commitment fee of 8 bps and a service charge of 21 bps. While these fees differ from the commitment fee and service charge applicable to other Fund arrangements, the differentiation can be justified in view of the special, short-term nature of the BOP need to be addressed by the SLS and the proposed revolving feature.38 The proposed structure and level of the fees are intended to ensure that both the commitment fee and the service fee can fulfill their respective purposes in the context of the SLS and that the overall pricing for the SLS is comparable to other Fund credit.39

47. The commitment fee and the service charge serve different yet related purposes. Commitment fees are intended to cover the cost of establishing and monitoring arrangements and for setting aside resources, and to discourage unnecessarily high precautionary access and thereby help contain risks to the Fund’s liquidity. The purpose of the service charge is to generate income to cover administrative costs and discourage unnecessary purchases (Table 2). Taken together, these charges provide the incentives for a member to request only the amount of access it needs, without imposing prohibitive cost barriers that discourage members with potential BOP problems from requesting Fund resources and, if an actual BOP need arises, to purchase only the amount it needs. These incentives are reflected in the commitment fee being generally lower than the service charge, so that even after taking account of the commitment fee refunds, there is often a net cost to drawing under an arrangement.

Table 2.

Income from GRA Charges and Fees, FY2012–16

(Millions of SDR unless stated otherwise)

article image
Source: IMF Finance Department.

Excluding burden sharing adjustment for charges.

At end-April 2016, there was a further SDR 415 million of commitment fees collected (including SDR 396 million from FCL/PLL arrangements) which may be refunded (if drawn or cancelled before expiration) or recognized as income (if amounts remain undrawn upon cancellation or expiration).

Average USD/SDR exchange rates for FY2012–16 are from the annual income outturn papers. Average exchange rate for the 9 months in FY2017 is the daily average of USD/SDR exchange rate published by the IMF.

48. When the SLS is used as envisaged, i.e., on a short-term and revolving basis, the Fund’s current standard commitment fee and service charge structure could undermine the key features of the new facility. If the current rules were applied to the SLS, a member would pay a refundable commitment fee upfront for the approved amount, receive a refund if it makes purchases, and pay a pro-rated commitment fee if it reconstitutes access by repurchasing within the arrangement period. Requiring members to pay (pro-rated) commitment fees each time they reconstitute access under the SLS, however, would provide financial disincentives to early repurchases and could undermine the SLS’s unique revolving feature.40 Similarly, if the current 50 bps service charge is levied on each purchase under the SLS, it could be prohibitively expensive for members to make repeated purchases/repurchases within the arrangement as their short-term BOP needs subside and reemerge, which would also discourage the use of the revolving feature. For example, if a member with an SLS were to fully draw, repay, and draw again under an SLS arrangement with access of 145 percent of quota during a 12-month period, total fees from the combination of the current refundable 18 bps average commitment fee and the 50 bps service charge would amount to 100 bps. In the view of staff, this pricing is out of line with the pricing of other Fund facilities, and would not provide adequate incentives for members to make use of the SLS.

49. Against this background, and to support the use of revolving credit, staff proposes applying a nonrefundable commitment fee for the SLS. As discussed above, requiring members to pay pro-rated commitment fees when they reconstitute access could discourage early repurchases. Therefore, staff proposes that the commitment fee for the SLS be paid on a nonrefundable basis.41 This would also apply in the context of successor arrangements where drawing under a previous SLS remains outstanding at the time a successor SLS is approved.42 If purchases are made under the SLS, members would not receive refunds, and if they make early repurchases and thereby reconstitute their right to make further purchases, they would not incur further commitment fees. Consideration could be given to billing the commitment fee at the end of the arrangement rather than upfront to simplify operations.

50. To be comparable with the refundable commitment fee under the credit tranches, staff propose that the nonrefundable fee be 8 bps. The expected commitment fee paid by member and received as income by the Fund for use of other GRA facilities of the same size and duration, i.e., 145 percent of quota for 12-months on a precautionary basis, would be between 0–18 bps.43 A nonrefundable commitment fee of 8 bps for the SLS would make the expected costs and related income to the Fund broadly comparable with credit tranche facilities under a combination of drawing and non-drawing scenarios.

51. Staff also proposes to establish a lower service charge of 21 bps for purchases under the SLS. This would support the SLS’s revolving feature and, together with the proposed 8 bps non-refundable commitment fee, would allow for the expected cost of the facility to be broadly comparable to that of other GRA facilities as its revolving feature is used. The service charge should be established at a level where it provides an additional net cost to the user of Fund credit above the commitment fee, but lower than the standard service charge of 50 bps, which would add significantly to the overall cost of using the SLS’s revolving feature and thus discourage the use of the revolving feature. Applying a non-refundable commitment fee of 8 bps while setting the service charge for all purchases under an SLS arrangement at 21 bps would result in a cost of the SLS that is comparable to other Fund credit if the facility is drawn on a revolving basis.44

52. It is expected that the special commitment fee and service charge structure for the SLS would be reviewed within two years in the context of a broader review of the SLS. The review will take into consideration whether the experience with the use of the facility is consistent with the assumptions that underpin the proposed pricing structure.

D. Process

53. The proposed process for the approval and use of the SLS largely follows the process for the FCL There are three notable differences: the extension of an “offer” by the Fund to those members that qualify, the absence of a prior informal Board meeting, and an option for sole central bank signatory. The main steps are:

  • Initial confidential consultation. Staff discusses with the authorities their interest in pursuing an SLS arrangement and the likelihood of meeting the qualification criteria. Discussions are confidential and could happen at any point during the year, similar to the FCL process.

  • Assessment of qualification. Staff begins the assessment process once the authorities have indicated their wish to be assessed.45 Staff carries out an assessment of qualification and the level of access that might be warranted based on the latest information. If further consultation with the authorities is required, it could take place via a staff visit or video teleconferencing. Once the assessment is complete, management will assess if access to Fund resources is appropriate.46 47

  • Preparation of the Board paper. If a member confirms its intention to proceed, staff prepares the staff report setting out the basis for the Board to support an arrangement. The report includes: (i) the assessment of qualification; and (ii) the assessment of the potential BOP need, appropriateness of the proposed level of access, and repayment capacity.48

  • Formal Board meeting to consider the case for an offer and conditional approval of an arrangement. In contrast to other Fund instruments where the Executive Board may approve a request for the use of Fund resources, the Board instead approves the “extension of an offer” to the member. The offer is contingent on the authorities’ acceptance within a specified period (see below) and their commitment to maintaining very strong economic policies and responding appropriately to shocks that may arise. If, after the Board meeting, there are concerns about market-sensitive leaks or misinformation regarding the offer, the member could request that the Fund issue a press release indicating the Board’s conditional approval of an arrangement. The press release would take special care not to prejudge the authorities’ decision on whether to avail themselves of the arrangement49

  • Extension of an “offer.” The Board’s conditional approval of the arrangement, with a specified access level, is communicated to the authorities within 24 hours of the Board meeting. The authorities are required to respond within two weeks from the date of the conditional approval.

  • Effectiveness of the arrangement. The arrangement becomes effective upon the Fund confirming receipt of the signed written communication from the member, including the acceptance of the “offer” and policy commitments (see Annex VII for an illustrative letter). Staff will subsequently issue the communication to the Board for information.50

  • Press release and publication. A press release is published on the date the arrangement enters into effect, while the staff report, including the authorities’ written communication, is published shortly afterwards.

  • Purchases. A member could make one or multiple purchases during the period of the SLS arrangement subject to the approved access level. In the event of a purchase, staff would inform the Board promptly.51

  • Transition between the FCL and the SLS. A member could cancel an existing arrangement (FCL or SLS) and request a new arrangement (SLS or FCL, respectively). As the proposed qualification criteria for the SLS and the FCL are fully aligned, and the member is already qualified for the existing arrangement, it is expected that the member will also qualify for the new arrangement (in the case of the SLS, if the member has the potential BOP need which the SLS is aimed to address), and the process for the new arrangement will be expeditious.

54. The central bank could be the sole signatory of the written communication, if the member so chooses, subject to certain requirements being met. First, the central bank must have the domestic legal authority to commit the member to the financial obligations to the Fund.52 Second, it must be the exclusive responsible authority for any necessary adjustment. As the SLS is designed to address a special BOP need that by nature requires only adjustment of a monetary and exchange rate nature (e.g., interest rate changes, exchange rate policies), any needed adjustment would likely be confined to measures implemented by the central bank.53, 54 Lastly, the written communication would need to convey, on behalf of the member’s authorities, the member’s commitment to maintain its current very strong policies that form the basis for the qualification assessment. As these policies will necessarily include fiscal policies, staff will confirm with the Ministry of Finance that they are indeed committed to maintaining their current fiscal stance during the course of the arrangement.

55. The process for the SLS is strictly confidential and designed to both minimize the risk and mitigate the impact of potential leaks. The information would be handled the same way as for the FCL, with great care taken to maintain the confidentiality of the process. The qualification assessment is made public only for countries that meet the criteria and ultimately enter into an arrangement. The risk of a damaging leak is therefore mitigated by the fact that reports are only produced for countries that have passed staff’s qualification assessment. The negative signal that might be sent if a previous user does not renew its arrangement is thus muted, since it would be unclear whether this was the member’s choice or the member no longer qualified.55

56. A group of qualifying countries could simultaneously opt in. Synchronized extension of offers to, and opt-in, by qualifying members can help kick-start the use of the SLS by minimizing first-mover problems, and strengthen the effectiveness of the response to a common shock. The process is similar to that for the synchronized approval of FCLs, as per FCL guidelines. Confidentiality is preserved in discussing with members the potential interest and qualification for the backstop; staff only reveal that other members have expressed interest in the backstop if those members have agreed that such information can be shared. While each member’s qualification is assessed individually, and is subject to all the requirements for the backstop, the Executive Board considers the extension of offers to multiple countries at a single meeting. External communication could be coordinated, e.g., in the form of a single Fund press release announcing the multiple new arrangements, subject to agreement by the members in question.

E. Relationship with the FCL

57. The SLS and the FCL are designed to address different types of BOP needs, but in practice may be complementary. The SLS aims to provide support against medium-sized BOP shocks of a special nature, while the FCL is for any type and size of BOP needs. For those members looking to access a Fund arrangement on a precautionary basis, staff envisages that the SLS would provide support against medium-sized liquidity shocks, while the FCL could be the exceptional-access counterpart, for use as a temporary backstop against extreme shocks/tail events that may require comprehensive policy adjustments, with an exit expectation (see Figure 2). The maximum access level under the SLS should provide sufficient coverage for the type of volatility that the SLS is envisaged to address. If, however, the financing need of a member with an active SLS ends up exceeding 145 percent of quota (or the remaining credit available under the SLS), the member would need to transition to a different instrument, e.g., the FCL, which would provide access to more resources and a longer repayment period. The transition from the FCL to the SLS (and vice-versa) would be facilitated by the identical qualification criteria.

Figure 2.
Figure 2.

Envisaged Configuration of the Fund Precautionary Toolkit

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Source:IMF staff estimates.1/ Showing historical precautionary arrangements in percent of 2016 quota.

F. Other Modalities

58. Review clause. To ensure that the Fund has sufficient resources for crisis lending, staff propose that the SLS be reviewed two years after its creation, or earlier if total combined commitments under the SLS and the FCL exceed SDR 150 billion (see resource section below). The Board could also consider incorporating a sunset clause providing for the expiration of the facility in eight years after its establishment. A decision to extend the SLS beyond the initial sunset date would, like the original decision establishing the facility, require a Board decision adopted by an 85 percent majority of the total voting power.

59. No safeguards assessment. The SLS would be subject to the same requirements as the FCL regarding safeguards assessments. FCL arrangements are not subject to the Fund’s policy on safeguards assessments and a modified safeguards approach applies.56 Under the FCL, the very strong institutional quality criterion justifies a streamlined approach and since the SLS has the same qualification criteria, the same limited safeguards procedures are proposed for the SLS.

60. No concurrent use. Members that have other GRA arrangements (e.g., FCL or SBA) would not use the liquidity backstop to “double up.” As a general matter, members do not access financing under concurrent GRA arrangements. If a member has an existing precautionary arrangement under another facility, this would provide support in the event of short-term liquidity shocks, even if that were not the main objective of the arrangement.

61. Post-program monitoring (PPM). Like the FCL, it is proposed that credit outstanding under SLS arrangements will be taken into account for the purpose of the PPM policy.

62. Consequential policy changes. Establishing the SLS with its proposed features will require changes to a number of Fund policies. These include changes to the decision on overall access to the Fund’s resources in the GRA, which will need to be amended to reflect the specific access provision of the SLS. Changes to the transparency policy will also be made to clarify that the Board’s approval of an SLS arrangement will be conditional upon the member’s commitment to publication in its written communication accepting the offer, and to link the publication timelines for the SLS to the activation of the arrangement (instead of the Board approval). Further changes will need to be made to the decision on Article IV consultation cycles to provide that, upon approval of an SLS arrangement, a member will remain on a 12-month Article IV consultation cycle.

Resource Implications

A. Commitments and Second-Round Effects

63. Staff’s preliminary estimates indicate that potential commitments under the SLS would likely not exceed SDR 90 billion, and would probably be much lower. Staff has established a tentative list of potential qualifiers, based on a preliminary assessment of the qualification framework described in Box 3.57 The resulting list was used to construct different scenarios for possible resource implications (Table 3). All scenarios assume that members that opt in receive the maximum access under the facility of 145 percent of quota. • Scenario A assumes that all potentially eligible members, except those with current active FCL arrangements, decide to avail themselves of SLS arrangements. This could be considered an upper-bound estimate under current circumstances and would result in commitments of about SDR 88 billion.

  • Scenario B assumes that only the potentially eligible members that have so far expressed a clear interest opt in. Under this more plausible scenario, commitments would be about SDR 32 billion.

  • Scenario C builds on Scenario B, further assuming that, once their current arrangements expire, the three current FCL users would qualify and opt in for an SLS arrangement. In this case, commitments under the SLS would rise to around SDR 54 billion, but aggregate commitments under the SLS and FCL arrangements would fall considerably relative to Scenario B, by about SDR 55 billion, raising the Fund’s forward commitment capacity (FCC) by SDR 23 billion.

Table 3.

Potential Resource Implications

(Billions of SDR)

article image

Potential SLS commitments assuming maximum commitment of 145 percent.

Current FCL commitments (Mexico, Poland, Colombia). They are already reflected in the current FCC.

Assumes all members with an SLS or an FCL draw simultaneously.

64. This analysis suggests that potential commitments under the SLS could be accommodated comfortably within the Fund’s current resource envelope. Under all scenarios, commitments under the SLS would be well below the current FCC of SDR 210 billion, which as of April 11, 2017, comprises only quota resources.

65. The liquidity impact of SLS commitments could become temporarily higher if some members were to draw on the arrangements. Members meeting the SLS qualification criteria would generally be expected to also have sufficiently strong external positions to be included in the Financial Transactions Plan (FTP).58 Members included the plan would be removed temporarily from the FTP if they draw on their arrangements, since they would not be in a position to provide resources to other members when they have BOP needs themselves. This, in turn, would reduce the FCC, giving rise to a second-round effect on the Fund’s liquidity.59 Such second-round effects of SLS commitments could be significant, albeit temporary given the anticipated short-term nature of the special BOP need.60 For instance, in Scenario B, the maximum temporary second -round effects (for as long as Fund credit remained outstanding) would reach SDR 12 billion in the extreme event where all members with SLS arrangements in that scenario made purchases.

66. Staff proposes to review the SLS if total combined commitments under the SLS and the FCL exceed SDR 150 billion.61 62 This is the same threshold as the current review trigger of SDR 150 billion of combined commitments under the FCL and the PLL. Maintaining a trigger based on combined commitments under the FCL and SLS is also consistent with the proposed replacement of the PLL with the SLS and positioning the SLS as complementary to the FCL. If the Fund has concerns in the future regarding the burden the SLS places on Fund resources, it could decide to modify or terminate the facility, with transitional arrangements for those arrangements in effect.

B. Scoring of Precautionary Commitments in the FCC

67. The Fund’s forward commitment capacity currently counts (“scores”) all Fund commitments at their full value. The one year FCC, which was adopted in 2002, is the primary measure of the Fund’s liquidity position, providing a transparent measure (derived from published data) of the resources available for making new financial commitments over the next 12 months.63 Its calculation treats all commitments as equivalent, including under arrangements treated as precautionary (“precautionary arrangements”) by country authorities that are historically less likely to be drawn.64

68. Directors last reviewed the above approach in the context of the 2014 FCL/PLL/RFI Review. The staff paper noted several drawbacks with partial scoring, and proposed that the current approach be maintained. Directors generally supported this conclusion, though a few were open to consider some flexibility in their treatment.65 Some Directors have since asked staff to revisit the practice again in the context of the current review.

69. Partial scoring would not increase the Fund’s overall resource envelope. Nor would it change the size of total commitments the Fund is obliged to meet on demand. Rather, a shift to partial scoring would signal that the Fund is willing to commit more resources than it has available when part of these commitments are precautionary in nature. To avoid raising doubts about the Fund’s ability to honor its commitments, staff believes it is important that the Fund retain a low tolerance for liquidity risk.

70. Against this background, two conditions can be identified that could justify partial scoring: (i) that the Fund’s exposures under precautionary arrangements are sufficiently well diversified to assure, with a high degree of confidence, that only a fraction of its total commitments will be drawn upon within a short period of time; and/or (ii) that the Fund has a credible mechanism in place to mobilize the necessary resources quickly should large-scale drawings materialize.

71. However, in staff’s view, neither of these conditions is currently in place:

  • It is difficult to conclude, with high degree of confidence, that only a fraction of the Fund’s precautionary commitments will be drawn within a short period of time. Given the high degree of inter-connectedness in today’s global economy, the number of Fund’s individual precautionary exposures is unlikely to be large enough to meaningfully diversify liquidity risks, even assuming greater demand for the new liquidity instrument that has emerged to date for the FCL/PLL. Moreover, the Fund’s precautionary arrangements, including the proposed new liquidity instrument, are designed to address external shocks that are likely to impact a number of countries simultaneously. Even if the probability of such shocks were to remain relatively low, the potential concurrent drawings in such a low probability event could be very high.66

  • As currently designed, the Fund’s primary liquidity backstop, the NAB, could not be used to meet commitments made when the NAB was not activated. While the NAB can be activated at short notice with a Board decision and 85 percent support from the NAB participants, NAB resources can only be used to finance commitments made after its activation.67 This implies that such commitments could only be financed with quota resources and the Fund would need to ensure that it has sufficient quota resources available at all times to meet the full amount of potential drawings under existing arrangements. Relatedly, under recent legislation, the U.S. can only support an activation proposal during the next NAB renewal period if the FCC is expected to fall below SDR 100 billion, which effectively sets a quantitative threshold to meet the required 85 percent majority for activation.68 Therefore, under partial scoring, the Fund could be in a position where available quota resources are fully committed but the FCC remains above SDR 100 billion, such that it would be unable to activate the NAB to finance new commitments (see Annex VIII for illustrative calculations).

72. Under partial scoring, the FCC would also become an increasingly misleading indicator of the Fund’s ability to enter into new commitments as the level of precautionary commitments expands. It should also be noted that full scoring of precautionary commitments is embedded in the Fund’s Borrowing Guidelines as well as in each individual bilateral borrowing agreement, which would need to be amended if partial scoring were adopted.69

73. Based on these considerations, staff believes that full scoring of precautionary arrangements in the FCC remains appropriate. The main reasons are (i) the importance of avoiding any doubts over the Fund’s ability to meet its financial commitments under precautionary arrangements and thus ensuring their effectiveness as part of the GFSN; (ii) the potential for drawings to be highly correlated in response to external shocks that affect several members at the same time; and (iii) the absence under the current framework of an alternative source of liquidity that could be tapped in the event of such correlated drawings. If such an alternative source of liquidity were identified in the future, this treatment could be revisited.

74. That said, it would seem reasonable for the Board to take account of the level of precautionary commitments when assessing the adequacy of the Fund’s liquidity. In particular, the same level of the FCC could be considered more or less comfortable depending on the relative size of drawing versus precautionary commitments. In circumstances where precautionary commitments are a relatively large share of the total, the Fund could be willing to tolerate a lower FCC before activating the NAB, and the Managing Director could take such considerations into account in making a proposal on activation. To facilitate such judgments on when a need exists to supplement quota resources, staff proposes that the FCC calculation in future provides a breakdown of total commitments between arrangements expected to be drawn and those treated as precautionary by the authorities.

Issues for Discussion

75. Directors may wish to consider the following issues for discussion:

  • Do Directors agree with the use of a core set of indicators with thresholds to guide judgment in FCL qualification, to improve the predictability and transparency of the assessments?

  • Do Directors see scope for increasing price-based incentives to encourage exit from the FCL? What are Directors’ views on the options, i.e., steepening the current commitment fee schedule and introducing a time-based commitment fee?

  • Do Directors concur with the proposed elimination of the PLL?

  • Do Directors agree that the qualification standards and frameworks for the SLS should be fully aligned with the standards and framework for the FCL?

  • Do Directors support the new design features of the SLS, including revolving access, no exit expectation, the option of sole central bank signatory, and the Board’s conditional approval of an arrangement?

  • Do Directors agree with the proposed special fee structure for the SLS?

  • Do Directors agree that precautionary commitments should continue to be counted at their full value in the calculation of the FCC?

76. A forthcoming supplement will further elaborate on the specific proposed changes to Fund policies, and will include proposed decisions.

Annex I. Access Reduction and Market Impact

Potential negative market reaction upon exiting from FCL/PLL arrangements could deter exit. Hence, this annex examines market reactions to past reductions in access. The analysis does not find evidence of negative market impact from reducing access under FCL or PLL arrangements.

1. Since the inception of the FCL and the PLL, six cases of access reduction have taken place. Morocco reduced access at the time of its second (July 2014) and third (July 2016) arrangements (Figure AI.1). Poland reduced access during its fifth arrangement, both at the time of the approval (January 2015) and the review (January 2016), and at the time of the sixth arrangement (January 2017). Colombia reduced access under the FCL at the time of the second arrangement (May 2010), but has since reversed this.

Figure AI.1.
Figure AI.1.

Access in FCL and PLL Arrangements

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

2. Lower access, which signals gradual exit, was discussed positively in staff reports. With Colombia’s first arrangement approved at the height of the global financial crisis, reduced access under the 2010 arrangement reflected the fact that global risks had subsided. In the case of Morocco, lower access reflected stronger fundamentals and an improved balance of external risks relative to previous arrangements. The reductions in access under Poland’s FCL arrangements reflected improved fundamentals and increased policy buffers, and was communicated as a clear signal of the authorities’ intention to fully exit from the FCL once external risks recede. At the same time, the authorities highlighted their commitment to maintaining very strong policies and fundamentals. The authorities were keen on communicating their plans of gradual exit through outreach to investors and the general public, including by top Ministry of Finance officials. In turn, the market reaction at the time of access reduction was muted.

3. High-frequency data point to generally muted market reaction to access reductions, with no negative impact on spreads and exchange rates. Daily data confirm the absence of sudden changes to EMBI spreads and nominal bilateral exchange rates on the day immediately following the announcement. In the case of Poland, financial variables appeared to improve at the time of the January 2015 announcement, with observed exchange rate appreciation and reduction in spreads. Similar patterns were observed when looking at yields and the volatility of spreads and exchange rates (Figure AI.2).

Figure AI.2.
Figure AI.2.

Access Reduction and Market Reaction

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Source: Bloomberg.1/ The January 13, 2016, access reduction occurred shortly before an announcement of a potential costly foreign -currency mortgage convertion scheme.

4. Regression analysis supports the finding of no adverse market reaction. Regressions of EMBI spreads adjusted for global factors on an access reduction dummy show no evidence of negative market reactions to access reductions.1 The coefficient on the access reduction dummy is consistently negative and statistically significant in some specifications. This suggests a potentially favorable market reaction to access reduction, which, in turn, underscores the importance of effective communication surrounding exit decisions.

Annex II. Recent FCL and PLL Arrangements

Access levels under recent FCL and PLL arrangements have diverged. This annex provides an overview of the recent FCL and PLL arrangements.

1. Among FCL users, Colombia and Mexico have recently increased access, while Poland has reduced access under the arrangement.

  • Colombia has recently increased access. A sixth arrangement under the FCL in an amount equivalent to SDR 8.18 billion (400 percent of quota) was approved in June 2016. This corresponded to more than a doubling in access from the previous arrangement of SDR 3.870 billion, though it was only moderately above the level of access under the first arrangement in May 2009 in light of a significant access reduction at the time of the second review.

  • Mexico has recently increased access. A sixth arrangement under the FCL in an amount equivalent to SDR 62.389 billion (700 percent of quota) was approved in May 2016, thereby allowing for a more than 30 percent increase in access from the previous arrangement. This is also considerably larger than the SDR 31.5 billion approved under the first arrangement in April 2009.

  • Poland has continued a gradual exit from the FCL. After an initial arrangement in May 2009 and increased access in terms of SDR in the context of the third and fourth arrangements, Poland requested a close-to-30 percent reduction in access at the time of the fifth request in January 2015, an additional 16 percent reduction at the time of the review of the fifth arrangement one year later, and a further 50 percent reduction in access at the request of the sixth arrangement. This resulted in a reduction in access to 159 percent of quota from a peak of 537 percent of current quota.

2. Access has gradually declined for Morocco, the only current PLL user.

  • Morocco has continuously lowered access. Morocco is currently the only user of the Precautionary and Liquidity Line (PLL). After the approval of the initial arrangement in August 2012 in an amount equivalent to SDR 4.1 billion, Morocco has demonstrated its gradual exit strategy by requesting lower access in both the second and third arrangements, reflecting the evolution of external risks and the strengthened resilience of the economy. In turn, access under the third (2016) PLL arrangement is close to 40 percent lower than that under the first arrangement.

Annex III. Evaluation of the External Economic Stress Index

The External Economic Stress Index (ESI) was developed to guide access discussions under FCL and PLL arrangements. This annex reviews the performance of the ESI since its implementation. Overall, the analysis suggests that ESI downside scenarios have developed in line with access levels, but that some strengthening of guidance pertaining to the use of the ESI would help support cross-country comparability.

1. The 2014 FCL/PLL review introduced an External Economic Stress Index (ESI) to guide access discussions and help inform exit strategies in program documents. Subsequently, the 2015 FCL guidance note fleshed out modalities of the ESI and asked country teams to develop and maintain an ESI that indicated the evolution of the external economic environment for the given country. The guidance note proposed to compute the ESI as a weighted sum of standardized deviations from means of external variables. When the index points to more elevated risks, the adverse scenarios that help determine access levels in FCL and PLL arrangements would normally be expected to include assumptions that are more extreme, and vice versa. As such, the ESI provides information about the external environment, whether a successor arrangement is warranted, and in which direction potential access is expected to change.

2. To support consistency over time, guidance on the chosen ESI and the adverse scenario was provided. Specifically, once risks, variables, and weights in the ESI were decided, these were expected to remain unchanged during an arrangement, unless a compelling economic reason for changes could be provided. Furthermore, downside scenarios should draw, as appropriate, on corresponding World Economic Outlook (WEO) downside scenarios, the Global Financial Stability Report (GFSR), or the Global Risk Assessment Matrix (G-RAM).

3. Staff implemented ESIs with data-based weights. The first implementations of ESIs among FCL and PLL users were in 2014 (Mexico) and 2015 (Colombia, Morocco, and Poland). While weights related to the various external factors in the ESI could be either data- or model-based, all four country teams opted for data-based weights. Under this method, weights are determined by the economic size of the respective trade and financial exposures relative to the overall size of the economy, thus differing across countries. None of the teams implemented model-based weights, where econometric methods are used to estimate the importance of each of the external risks for observed balance of payment pressures.

4. The projected downside ESI scenarios have helped justify access levels in successor arrangements. Overall, downside ESI scenarios have generally been closely aligned with potential financing needs in adverse scenarios. This consistency between ESIs and financing needs tables was evident also across time. Notably, in Colombia’s and Mexico’s 2016 requests for higher access under the FCL arrangements, projected downside ESI values worsened relative to their respective 2015 and 2014 arrangements (Figure AIII.1). In 2016, Morocco’s (PLL request) and Poland’s (FCL review) requests for lower access were associated with marginal improvements in the adverse scenarios of the ESIs relative to earlier program documents. That said, despite deteriorated external conditions at the time of Poland’s 2017 request, Poland was able to reduce access on the back of improved fundamentals and buffers.

Figure AIII.1.
Figure AIII.1.

Evolution of ESI

(Negative: adverse external economic conditions)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: IMF staff reports; and IMF staff calculations.1/ For Poland, the adverse scenario documents denote the 2015 and 2016 reports.

5. However, guidance could be strengthened to improve comparability of the ESI adverse scenarios across countries, including related to the size of underlying shocks.

  • Comparability of adverse scenarios across countries. While downside ESI scenarios have generally been computed based on values of external variables in adverse scenarios in the WEO, GFSR, and spillover reports, not all reports have based their downside scenarios on the most recently available scenarios. Hence, relevant guidance could be strengthened by specifying that ESI downside scenarios should be based on scenarios developed in Fund documents no older than six months, to the extent that relevant scenarios are available. Alternatively, specific common scenarios as they pertain to the G-RAM would help ensure consistency.

  • Comparability of the size of shocks across countries. Adverse scenarios in program requests that fall close in time should be comparable in terms of the size of assumed underlying shocks, given similar supporting information available. However, some shock scenarios in recent program requests have not been fully comparable. For example, while the May 2016 staff report for Mexico incorporated a U.S. growth shock of 1.5 percentage points (referencing the April 2016 GFSR), the June 2016 staff report for Colombia involved a smaller U.S. growth shock. Hence, while downside scenarios for the two reports were similar, the quantification of the shocks differed. That said, even if underlying shocks would be similar, the impact of the shock would likely differ across countries given their different exposures to external risks, as reflected in the ESI weights.

6. In sum, the ESI has been helpful in supporting discussions related to the need for successor arrangements, but could be improved. As external risks have remained elevated as assessed by the ESI, there is no evidence of unjustified successor FCL and PLL arrangements. Adverse scenarios, computed using the ESI, have generally developed in line with the direction of access levels. Nonetheless, comparability across countries could be improved. Hence, to ensure consistency going forward, it will be important that Fund country teams rely on the most recently published flagship reports (i.e. WEO, GFSR, spillover reports) or specific scenarios related to the G-RAM, as appropriate. In a similar vein, global shocks should be treated equally across reports that fall close in time. To the extent that differences are warranted, these could be clarified by specifying the context surrounding the given shock.

Annex IV. Use of Reserves in Adverse Scenarios

Reserve use has been increasingly prevalent in recent FCL and PLL arrangement access scenarios. However, to the extent that reserves have been assumed to remain adequate in adverse shock scenarios, additional reserve drawdown could be warranted. This annex examines to what extent there is room for additional reserve drawdown in adverse scenarios.

1. The FCL and PLL allow for an enlarged backstop against downside risks without the need to accumulate excess international reserves. As such, they complement adequate reserve levels. However, as central banks accumulate adequate reserves for use in the event that downside risks materialize, some reserve drawdown should be expected in a tail risk event. Therefore, reserve drawdown should be assumed to occur in adverse scenarios when determining the appropriate access level under FCL and PLL arrangements. To the extent that reserve drawdowns in past access scenarios have been modest, more severe shock assumptions could be partially met through a larger drawdown of reserves, rather than an increase in access levels.

2. Fund staff’s guidance has pointed to the need for increased use of reserves in downside scenarios. The 2015 FCL guidance note highlighted that “for countries for which reserve levels are plentiful, and well above adequate, the adverse scenario should include the use of international reserves to cover part of the financing gap, implying that not all the potential financing need is met through Fund resources. In such an adverse scenario, reserves could go below the relevant adequacy thresholds (since this is an extreme stress event).” In this respect, the 2014 Review of the FCL, the PLL, and the RFI pointed to a lower threshold of 80 percent of the Fund’s ARA metric. Using a large sample of EMs over a 22-year period, the reserve threshold was derived so as to minimize Exchange-Market-Pressure crisis prediction errors.1 Hence, with an 80 percent threshold separating crisis from non-crisis signals, countries may be able to draw down reserves to that limit in severe stress events without igniting crisis risks.

3. In turn, the use of reserves in adverse scenarios has become more prevalent. Requests for FCL arrangements in 2012 (Mexico) and 2013 (Colombia and Poland) were characterized by the absence of a drawdown of reserves in the adverse scenarios, with Poland assuming reserve accumulation corresponding to half of that projected under the baseline scenario. However, since then, all adverse scenarios in FCL arrangements have been accompanied by reserve drawdown, with the drawdown playing an increasing role in reducing the financing gap across most arrangements (Figures AIV.1 and AIV.2).

Figure AIV.1.
Figure AIV.1.

Reserve Accumulation in Access Scenarios

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Figure AIV.2.
Figure AIV.2.

Adverse Scenario: Contributions to First-Year Financing Gap

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

4. However, room may remain for additional reserve drawdown in a tail risk scenario. While reserve drawdown has been increasingly prevalent, the level of reserves has generally remained above 100 percent of the ARA metric in the adverse scenario. For example, while the Colombia 2016 FCL arrangement request incorporated a reserve drawdown of more than USD 6 billion, reserves remained well within the adequacy range, in part reflecting concerns about adverse market reactions to a large and rapid reserve drawdown. For Morocco, while the financing gap was defined to maintain reserves at 90 percent of the standard ARA metric, reserves remained adequate in the adverse scenario after adjusting for capital controls. Hence, with reserve drawdown a key element in determining access levels, limited reserve drawdown should be clearly justified in light of the guidance provided by the 2014 review. Specifically, it would generally be expected that assumed levels of gross reserves are allowed to fall below 100 percent of the ARA in the adverse scenario, while reserve adequacy ratios well into the adequacy range in the adverse scenario should be avoided. The pace of the reserve drawdown should also be considered to avoid adverse market reactions.

uA01fig09

Reserve Adequacy in Current Arrangements

(Percent of reserve adequcy metric; based on gross international reserves)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: IMF staff reports; and IMF staff calculations.1/ Augmented by buffer for commodity exports.2/ Net international reserves declined to 89 percent of the ARA metric in the adverse scenario.3/ Financing gap defined to maintain reserves at 90 percent of the standard ARA metric (before adjusting for capital controls). After adjusting for capital controls, baseline reserves amounted to 136 percent of the ARA metric and reserves in the adverse scenario remained adequate.

Annex V. Commitment Fee Reform Options

The Annex provides a summary of the purpose and application of the current commitment fee policy and a more detailed discussion on the two reform options discussed under “Proposed Changes to the Existing Toolkit:” (i) steepening the current commitment fee schedule by increasing the rate for access above the highest fee threshold and (ii) introducing a time-based commitment fee. The Annex also presents numerical illustrations for possible calibrations of the two options, including implications to current FCL/PLL users.

Purpose and application

1. The Fund’s commitment fees serve two purposes: (i) they cover the cost of establishing and monitoring arrangements and of setting aside resources to be used if a purchase were to be made; and (ii) discourage unnecessarily high precautionary access and thereby help contain risks to the Fund’s liquidity. The relative importance attached to these purposes has shifted with the evolving nature of Fund lending (see Box AV1).

2. Commitment fees are charged uniformly for all GRA arrangements.1 The fee is due and payable on the date an arrangement enters into effect, and is billed annually for a 12-month period (or for the period left under the arrangement, if shorter). The level of the charge depends on the level of access during the relevant period, as specified by the arrangement’s phasing, following the current tiered fee structure. The fee is refunded when purchases are made, prorated by the size of the purchase within each tier. At the expiration of an arrangement, the un-refunded portion of the commitment fee, if any, is recognized as income by the Fund.

Option 1: Steepening the Current Commitment Fee Schedule

3. The 2009 calibration of the current upward sloping fee structure sought to strike a balance between two considerations. On one hand, the upward-sloping structure would increase incentives against unnecessarily high precautionary access and provide Fund income to help offset the cost of setting aside substantial financial resources. On the other, commitment fees would not be set so high as to discourage members from seeking precautionary arrangements.

4. The SDR value of the fee thresholds was since increased in 2016 in the context of the Review of Access Limits and Surcharge Policies in line with evolved economic metrics.2 While this effective lowering of fees was recognized to run counter to the views expressed by some Directors that such fees should rather be increased, it was noted that commitment fees, including their level, could be reviewed with more experience in the context of future FCL reviews.

5. Against this backdrop, one possible option is a modest upward adjustment to the commitment fee level at the highest tier (above 575 percent of quota). Such a change would steepen the current commitment fee schedule while preserving the quota-based thresholds set at the time of the 2016 Review. Adjusting the rate only at the highest tier would target only the largest Fund commitments, recognizing that the commitment fee schedule applies to all arrangements, irrespective whether these are treated as precautionary. The rate increase should nonetheless be modest so as not to excessively penalize high access arrangements when they are warranted.

History of Commitment Fees (1952–2009)

Commitment fees were originally put in place to help manage incentives and compensate the Fund for cases in which commitments were not drawn. They were first introduced in conjunction with the establishment of the stand-by arrangement in 1952. Directors emphasize that while the charge should not discourage countries with need, it would serve as a deterrent to those who had no real reason to request Fund assistance. The initial fee level was set at 25 bps per year and, if a member drew under the SBA, the fee would be credited against the service charge on a pro rata basis.

In the context of the review of Fund facilities in 2000, a two-tier commitment fee schedule was adopted. The initial fee remained at 25 bps per year for commitments up to 100 percent of quota, while a lower 10 bps fee was levied on amounts above this threshold. The lower fee for higher amounts was adopted mainly to encourage the use of the Contingent Credit Line (CCL) and declining fee schedule was motivated by the lower probability of drawing under the CCL, which made refunds less likely. This was consistent with the prevailing view at the time that the basic rationale for charging commitment fees for contingent credits was to cover the cost of establishing and monitoring Fund arrangements.

The current upward sloping fee structure stems from the 2009 GRA lending toolkit reform and places more prominence on management of the Fund’s liquidity risks. With the introduction of the FCL and expected greater use of high-access precautionary SBAs (HAPAs), it was recognized that large commitments have costs associated with the finite availability of Fund resources and such costs are likely to increase at the margin as resources available for other lending decline. Accordingly, the fee schedule was revised to increase progressively with the level of access: (i) an initial fee of 15 bps for access up to 200 percent of quota; (ii) 30 bps for access between 200 and 1000 percent of quota; and (iii) 60 bps for access above 1000 percent of quota. The higher threshold was initially set to support the implicit 1000 percent of quota cap on FCL access that was subsequently abolished in 2010, while the lower threshold coincided with the 200 percent of quota annual access limit. In the context of the 2016 review of access limits and surcharge policies following effectiveness of the 14th Review of Quotas, the SDR value of the two thresholds was increased by 15 percent to the current levels of 115 and 575 percent of quota, respectively.

Option 2: Introducing a Time-based Commitment Fee

6. A time-based commitment fee (TBCF) would operate similarly to the current policy on time-based surcharges and be charged on top of existing level-based commitment fees. It would apply once the level of undrawn credit for each relevant (12-month) period under successive arrangements has remained above a specified threshold for a defined duration of time (“duration trigger”).3 The fee would be charged only for the period when undrawn credit remained above the threshold, analogous to Fund policy on time-based surcharges. The count toward meeting the duration trigger would start as the threshold is breached and would pause whenever the level of undrawn credit falls below the threshold level or the arrangement expires or is cancelled. The count would reset only once a “cooling-off” period is exceeded. Once triggered, the TBCF would be levied ex-post at each anniversary date of the relevant period (or at the end of the arrangement if sooner) for the time during which undrawn balances remain above the threshold.

7. Specific design features of the TBCF would need to carefully balance its objectives and risks. On the one hand, the TBCF should provide a meaningful incentive to reduce the size of commitments at high-access levels. On the other hand, its design should acknowledge that prolonged high access may be justified in circumstances where a member experiences an extended period of elevated external risks. The design features should also limit the risk of unintended ramifications from the TBCF’s application to arrangements not meeting its intended purpose.4 Beyond the choice of threshold levels and TBCF rates discussed in the main text, the following design aspects are pertinent:

  • Duration trigger. The length of the trigger period should be short enough to effectively support the revolving nature of Fund resources, while recognizing the possibility of extended periods of stress justifying prolonged commitments. One option is to link the duration trigger to the typical length of a severe shock, i.e., around 3 or 4 years.5 In view of the maximum two-year duration under the current FCL arrangements, a 4-year duration trigger could be considered, effectively being applied at the beginning of the third FCL arrangement, provided access would remain above the threshold. Such long duration trigger would also significantly limit the risk of unintended ramifications across arrangements, as the level of undrawn balances is less likely to remain above the threshold for a period of four years unless the arrangement is treated as precautionary.

  • Cooling-off period. Short intervals between successive arrangements, or purchases reducing the level of undrawn credit to below the threshold for some time, could “pause” the count toward meeting the duration trigger. The count would continue once the level of undrawn credit increases above the threshold again, for example, because of a new 12-month period starting, a successor arrangement being approved, or access being augmented. However, the count would be reset once a certain “cooling off” period is exceeded. Such a period should be long enough to effectively promote a durable exit from prolonged arrangements and ensure that delays in successor arrangements do not interfere with the policy. On the other hand, too long a cooling-off period could increase the likelihood of penalizing members that would request a successor arrangement with a delay on the basis of new external risks that were not foreseen when the previous arrangement was allowed to expire. On balance, a cooling off period when undrawn credit remains below the threshold continuously for one-year would seem appropriate.

  • Ex-post billing and non-refundability. The fee could be billed and paid ex-post, at each anniversary date and upon expiration/cancellation for the period in an arrangement where commitments exceed the threshold. In the event purchases brought the level of undrawn credit below the threshold, a pro-rated TBCF would be levied at the end of the relevant period, covering only the time for which the threshold was exceeded. The same policy would be applied to cancellations and changes in the level of access under the arrangement. This implies that, in contrast to current policy on level-based commitment fees, the time-based fee, once triggered, would not be refundable.

Numerical illustrations

8. For illustrative purposes, staff presents alternative numerical examples for each of the above options. Figure AV1 illustrates the per annum marginal and effective fee schedules for each option, while Table AV1 illustrates hypothetical implications for current FCL/PLL arrangements if the changes were to apply to them once the policy were adopted. Since the thresholds and fee levels discussed are the same, the only difference is that the TBCF would apply only once the level of undrawn credit has remained above the 575 percent of quota threshold for the duration trigger period.

  • Under the option to steepen the current schedule, the level-based commitment fee rate for the highest tier (access above 575 percent of quota) is increased by 10 and 20 bps for the first and second alternatives respectively. For an arrangement with 700 percent of access, both alternatives would moderately increase the effective commitment fee rate from the 33 bps under the current fee structure to 35–36 bps (The illustrative scenarios assume no grandfathering; however, in practice, arrangements for which the fee would be higher would be grandfathered—see ¶33).

  • Similar increases in the highest marginal fee rate under the TBCF option would apply only after being triggered. With the clock starting at the earliest in mid-2013 (see main text), the only arrangement that would be affected would be Mexico’s FCL if it were renewed at its current access level beyond May 2020.

Figure AV.1.
Figure AV.1.

Illustrative Commitment Fee Structures 1/

(Basis points)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Source: IMF staff calculations.1/ The effective rate is the total commitment fee payable relative to total access available for purchase over a 12-month period.
Table AV1.

Implications of Illustrative Structures on Current FCL/PLL Arrangements 1/

article image
Source: IMF staff calculations.

Effective rates and nominal fees are based on amounts available for purchase over a 12-month period.

A time-based fee for access above 575 percent would not yet be triggered under successor arrangements to the current FCL arrangements.

Rounded to closest percentage point.

Computed on basis of approved access.

Annex VI. Cost Comparison of Commitment Fees

The Annex compares costs of the Fund’s current commitment fees to staff estimates for members’ costs of accumulating reserves as well as costs of contingent credit from other multinational and regional institutions.

Cost of Accumulating Reserves

1. Because reserves are a form of self-insurance, and, in a sense, an alternative to contingent credit, the cost of accumulating reserves may be seen as a useful “upper” benchmark for guiding the level of Fund commitment fees. Estimates of the cost of reserves generally comprise two main components: (i) the foregone return on an alternative investment; less (ii) the foreign currency return on reserves. A common estimate for the marginal cost in literature is the international foreign exchange denominated funding cost, typically proxied by the EMBI spread, adjusted for the yield on reserves held in shorter-term instruments.1 This measure reflects the opportunity cost of reserve holdings and is here estimated as: EMBI 5-year spread in USD + (5-year US bond yield – Average of 2-year US bond yield and 3-month US Treasury bill).2

2. As members’ funding costs change over time, the benchmark would need to be based on a long-term average. For gauging a useful benchmark, staff examines USD denominated EMBI spreads for 40 emerging market economies for which data is available over the period since 2000. Noting that Fund’s precautionary arrangements are intended to supplement reserves in periods of heightened external risks, which may be reflected in increased funding costs, peak observations associated with unfavorable market conditions are excluded from the calculation of the long-term average. Furthermore, as the motivation for potentially increasing the fees is to discourage excessively high access more typically witnessed under the FCLs, the calculation is based on a limited set of membership at lower deciles of spreads, intended to reflect the strong qualification criteria for accessing arrangements with no ex-post conditionality.3

3. On the basis of such benchmark, the Fund’s current commitment fees stand substantially below members’ estimated cost of accumulating reserves. As illustrated in Figure A1, for most of the period since 2000, the adjusted EMBI spreads for lowest 10th and 25th percentiles of countries have remained above the Fund’s highest marginal commitment fee rate of 60 bps and their average for the period ranges from about 210 to 260 bps.

Figure AVI.1.
Figure AVI.1.

Cost of Accumulating Reserves for Select Members

(Basis points)

Citation: Policy Papers 2017, 069; 10.5089/9781498346139.007.A001

Sources: Bloomberg; and IMF staff calculations.1/ The adjusted EMBI comprises the 5 year US dollar EMBI spread adjusted by adding the spread between 5 year Treasury bonds and the average of 2 year Treasury bonds and 3 month Treasury bills to account for the lower maturity of reserve portfolios. The 25th and 10th percentile are used to gauge the credit risk profile of potential FCL qualifiers.2/ Based on averages for the entire period since 2000, excluding the highest 20 percent of observations to reflect the period in which conditions for reserve accumulation are favorable.3/ For Poland the 5 year euro EMBI spread is used.

4. Such comparison should nonetheless be interpreted with caution. First, there is generally a premium to the cost of contingent credit relative to having the resources directly available on the balance sheet. Second, there can also be additional (immeasurable) stigma-related costs for engaging in Fund precautionary arrangements. Third, commitment fee costs at the Fund are partially offset by the refundability of Fund commitment fees against purchases, in contrast to the permanent opportunity costs associated with building reserves. Fourth, the cost of market borrowing may be volatile and changes may have both cyclical and structural components, implying that the difference between this cost and the level of commitment fees is unlikely to remain constant.

Cost of contingent credit from other institutions

5. Comparison of lending costs between institutions is not straight-forward due to differences in financing models and risk management frameworks. In comparison to the IBRD, for instance, the Fund does not borrow from private markets or leverage its capital base and employs policy and price incentives to promote timely repurchases once members’ balance of payments positions improve, whereas the Bank levies penalties for early repayments of loans and uses country exposure limits to manage its risks (an option not available for the Fund).

6. Bearing this caveat in mind, the Fund’s commitment fees are generally low on the basis of comparisons to corresponding fees charged on credit lines by other multinational or regional institutions. For instance:

  • The IBRD’s DPL-DDO (Development Policy Loan with Deferred Drawdown Option) is subject to a 25 bps front-end fee, due within 60 days of the effectiveness date, as well as an annual stand-by fee of 50 bps on the undisbursed balance. The fees are not refundable in event of cancellation after the effectiveness date, nor linked to the size of the loans, which are subject to country exposure limits and a country cap of $16.5 billion ($17.5 billion for India).

  • The ESM charges an annual commitment fee for all its financial support facilities, including the Precautionary Conditioned Credit Line (PCCL) and Enhanced Conditioned Credit Line (ECCL). The purpose of the commitment fees is to cover the possible negative carry of the ESM’s liquidity buffer and issuance cost of instruments as part of the prefunding activity. Commitment fees are applied ex-post at the end of the calendar year on the basis of the negative carry actually incurred and allocated to all ESM members benefiting from the ESM facilities on basis of the share of each such member’s Programme Amounts. In case of draw-downs under the precautionary credit lines, the commitment fee applies to the remaining maximum agreed amount under the facility plus the amount outstanding under the credit line. The fees charged are not refundable.

Annex VII. Illustrative Written Communication from the Central Bank

[member capital city], [date]

Ms. Christine Lagarde

Managing Director, International Monetary Fund

700 19th Street NW

Washington, DC 20431

Dear Ms. Lagarde,

I would like to thank you for your conditional approval of a 12-month Short-term Liquidity Swap arrangement with access in the amount equivalent to SDR [X].

I hereby confirm the intention of [member] to avail itself of the arrangement, and can confirm that the Central Bank of [member] will take any actions necessary to respond appropriately to shocks that may arise and is committed to maintaining its very strong economic policies during the course of the arrangement.

Separately, I have been requested to convey on behalf of the government of [member] its commitment to maintain during the arrangement its current very strong policies.

We consent to the IMF’s publication of this letter and the related staff report.

Sincerely yours,

Governor of Central Bank of [member]

Annex VIII. Scoring of Precautionary Arrangements in the FCC

This annex briefly reviews the background to the Fund’s current practice of counting (or “scoring”) precautionary arrangements at full value for the purpose of measuring Fund liquidity, elaborates on the relevant key considerations, and presents numerical scenarios illustrating the constraints to partial scoring.

Background

1. The primary measure of the Fund’s liquidity position is the Forward Commitment Capacity (FCC). Introduced in 2002, the FCC provides a transparent measure of the resources available for making new financial commitments over the next 12 months that can be derived directly from readily available data on the Fund’s finances.1

2. The current approach of fully scoring precautionary commitments when calculating the FCC was adopted when the new measure was introduced. Staff noted at the time that the previous approach of making adjustments to precautionary commitments when calculating the Fund’s liquidity ratio was not transparent, could lead to dramatic changes in the liquidity ratio as the status of arrangements changed from precautionary to non-precautionary, and could result in the Fund implicitly committing resources that had already been committed. In supporting full scoring, many Directors noted that the potential reduction in the FCC resulting from this treatment would need to be taken into account in future judgments about the adequacy of the FCC.

3. The Board reviewed this approach in 2014. The staff paper noted several drawbacks with partial scoring, including the highly concentrated nature of the Fund’s exposures and constraints on its ability to diversify liquidity risks, its limited ability to mobilize new resources at short notice, and the adverse impact on transparency and potentially undesirable signaling effects.2 Directors generally agreed that full scoring of precautionary arrangements remained appropriate, though a few others were open to considering some flexibility in their treatment, given the low probability of drawing under these arrangements.3

Key considerations

4. Partial scoring of precautionary arrangements would not create any additional resources for Fund financing; nor would it change the Fund’s total precautionary commitments, which it is obliged to meet on demand if a member so requests.4 Rather, a shift to partial scoring would signal that the Fund is willing to commit more resources than it has available in circumstances where part of these commitments are precautionary in nature. The FCC, which measures how much the Fund can commit in the next 12 months, would rise immediately without any corresponding change in the Fund’s lending resources or its total loan commitments.

5. The Fund’s facilities available to provide support for potential BOP needs, including the potential new liquidity instrument, can only be effective elements of the GFSN if users of these facilities have full confidence that the resources will be available in case of need. If there were any doubts about the Fund’s ability to honor its precautionary loan commitments, the incentives for members to rely on such facilities instead of self-insurance would be substantially reduced. Such doubts could also create perverse incentives for members to draw down their precautionary arrangements in a crisis to ensure early access to the Fund’s limited pool of resources.

6. These considerations suggest the Fund’s tolerance for liquidity risks should remain low. The question is whether partial scoring is consistent with such a low risk tolerance.

7. Two circumstances can be envisaged where partial scoring of precautionary arrangements could be justified without implying a significantly higher tolerance for liquidity risk.

  • a. One is where the Fund’s precautionary exposures are sufficiently well diversified to assure, with high degree of confidence, that only a fraction of its total commitments will be drawn upon within a short time period. However, it would seem hard to draw such a strong conclusion for the Fund, particularly for facilities designed to handle global shocks in a highly inter-connected world. For one thing, the number of individual exposures is unlikely to be large enough to meaningfully diversify risks, given the inter-connections across countries and even assuming greater demand for the new liquidity instrument than has emerged to date for the FCL/PLL.5 For another, while members could be subject to idiosyncratic shocks, the main motivation for an insurance-type instrument is to help members facing external shocks that impact a number of countries at the same time. Even if the probability of such shocks remains relatively low, the potential drawings in such a low probability event could be very high.6

  • b. Second, partial scoring could also be considered if the Fund has a credible mechanism in place to mobilize the necessary resources quickly should large-scale drawings materialize. The most logical vehicle would be the NAB, the Fund’s standing borrowing facility that can be activated at short notice with a Board decision and 85 percent support from the NAB participants. Bilateral borrowing agreements would also be potentially available, but only as a third line of defense after quotas and the NAB have been largely exhausted.

8. Two constraints limit the use of the NAB as a liquidity backstop to justify partial scoring:

  • a. NAB resources can only be used to finance commitments made when the NAB is activated. Thus, NAB resources would not be available to finance drawings under precautionary arrangements approved when the NAB was not activated. This implies that such commitments could only be financed with quota resources, and the Fund would need to ensure that it has sufficient quota resources available at all times to meet the full amount of potential drawings under existing arrangements. This constraint also implies that the FCC could become a misleading indicator of the Fund’s ability to commit new resources under partial scoring (see next section).

  • b. Required majority support effectively sets a threshold for NAB activation. The current NAB decision allows the MD to make a proposal for activation when there is a need for supplemental resources to forestall or cope with an impairment of the IMS, and does not specify any threshold for the FCC that must be met before an activation can be proposed.7 However, under recent legislation, the U.S. can only support an activation proposal during the next NAB renewal period if the FCC is expected to fall below SDR 100 billion. This requirement effectively sets a new quantitative threshold to meet the required 85 percent majority in support of an activation proposal. As a result, under partial scoring, the Fund could find itself in a position where its available quota resources are fully committed but the FCC remains above SDR 100 billion, such that it would be unable to activate the NAB to finance new commitments.

9. A similar situation could not currently arise with the 2016 Borrowing Agreements, which provide a further liquidity backstop after quotas and the NAB. While these agreements include an explicit threshold for activation (the FCC must be below SDR 100 billion), full scoring of precautionary commitments is embedded in the Borrowing Guidelines as well as in each individual bilateral borrowing agreement.8 This means that the Guidelines for Borrowing and each individual bilateral borrowing agreement would need to be amended if partial scoring were adopted. As with the NAB, however, the bilateral agreements cannot be used to finance drawings under commitments made prior to their activation, and would therefore not be available under partial scoring to finance drawings under existing commitments.

10. Based on these considerations, staff believes that full scoring of precautionary arrangements under the FCC remains appropriate. The main reasons are (i) the importance of avoiding any doubts over the Fund’s ability to meet its financial commitments under precautionary arrangements and thus ensure their effectiveness as part of the GSFN; (ii) the potential for drawings to be highly correlated in response to external shocks that affect several members at the same time; and (iii) the absence under the current framework of an alternative source of liquidity that could be tapped in the event of such correlated drawings.9

11. This said, it would also seem reasonable for the Board to take account of the level of precautionary commitments when assessing the adequacy of the Fund’s liquidity. In particular, the same level of the FCC could be considered more or less comfortable depending on the relative size of precautionary vs drawing commitments. In circumstances where precautionary commitments are a relatively large share of the total, the Fund could be willing to tolerate a lower FCC level (e.g., below SDR 100 billion) before activating the NAB, and the Managing Director could take such considerations into account in making a proposal on activation.10 To facilitate judgments by the Board and NAB participants on when a need exists to supplement quota resources, staff would propose in future to provide a breakdown of total commitments between drawing and precautionary arrangements in the FCC calculation.

Illustrative scenarios of partial scoring

12. The following scenarios illustrate how constraints to partial scoring of the Fund’s commitments would play out in practice. These scenarios, summarized in Table AVI.1, are based on the FCC as of end-December 2016 (for simplicity, it is assumed that two elements—projected inflows from scheduled repurchases and repayments of borrowing due in the next 12 months—net out and are therefore not shown). In addition to the FCC, the table shows a new concept of available liquidity, which measures the Fund’s resources available for new financing after fully accounting for existing commitments.

  • Under full scoring, shown in the first column, these two concepts are equal. The FCC indicates that the Fund could commit about SDR 208 billion in any combination of drawing and precautionary arrangements based on its existing (quota) resources, i.e., without requiring activation of the NAB. In this example, the Fund would need to commit at least SDR 108 billion before NAB activation would be possible, based on an assumed SDR 100 billion threshold.

  • The second column illustrates the potential impact of partial scoring with the same level of commitments. As discussed in the 2014 FCL review, it would be problematic in terms of complexity and market signaling to apply changing actual probabilities of drawing over time or across members. A simpler approach, illustrated here, would be to apply a constant level of partial scoring, say 50 percent, to all precautionary arrangements. On this basis, the FCC would rise to SDR 249 billion, while the Fund’s actual available liquidity to finance new commitments would remain unchanged at SDR 208 billion.

  • The gap between the two measures would widen further if there is sizable demand for the new liquidity instrument. For example, as shown in the third column, if this demand reached SDR 100 billion, the FCC under partial scoring would fall to SDR 199 billion, only slightly below its current level. However, the Fund’s actual available liquidity would be close to half this level (SDR 108 billion).

  • The fourth column illustrates the potential for the Fund to over-commit its quota resources under partial scoring in the event of very strong demand for the new liquidity instrument. For example, if total such demand reached SDR 250 billion, the FCC under partial scoring would drop to SDR 124 billion, still significantly above the threshold for NAB activation, but actual available liquidity would have turned negative. In this example, the Fund would have over-committed its quota resources available to meet potential drawings under its existing precautionary commitments, while remaining unable to activate the NAB to finance new commitments under either drawing or precautionary arrangements.11

Table AVII.1.

Fund Liquidity under Illustrative Scoring of Precautionary Commitments1/

(Billions of SDR)

article image
Source: IMF staff calculations.

Data as of December 30, 2016.

Undrawn balances under existing precautionary GRA arrangements, excluding NAB-financed portion.

Excluding projected inflows from scheduled repurchases and repayments of borrowing due in the next 12 months.

13. These examples show that, as precautionary commitments expand, the FCC would become an increasingly misleading indicator of the Fund’s ability to make new commitments under partial scoring. The FCC would continue to indicate that the Fund has a sizable ability to commit new resources when in fact its actual resources available to meet drawings under those commitments are substantially if not more than fully exhausted. Thus, a key strength of the current measure in terms of transparency would be undermined.12

1

Adequacy of the Global Financial Safety Net—Considerations for Fund Toolkit Reform,” IMF Policy Paper, December 2017.

2

Adequacy of the Global Financial Safety Net,” IMF Policy Paper, March 2016.

3

Review of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument,” IMF Policy Paper, January 2014.

4

Contingent on a decline in external risks, the Moroccan “authorities remain committed to continue with their gradual exit to allow full exit from PLL support at an early stage.” IMF Country Report No. 16/265, Statement by Mr. Daïri. The three FCL users have also signaled their intention to reduce access should risks subside, with a view to exiting the instrument.

5

Adequacy of the Global Financial Safety Net—Considerations for Fund Toolkit Reform,” December 2017, and “The Chairman’s Summing Up––Adequacy of the Global Financial Safety Net––Considerations for Fund Toolkit Reform”.

6

The Precautionary Credit Line (PCL), the predecessor to the PLL, also had only one user (the FYR Macedonia), with only one arrangement.

7

The FYR Macedonia fully exited from the PCL/PLL after one arrangement. However, the country experienced an actual BOP need only two months after the PCL approval. Also, it did not exit from a position of economic strength, and, as such, it is excluded from the analysis in the paper.

8

Decision No. 14283-(09/29), adopted March 24, 2009, as amended.

9

The proposed set of core indicators will be included in a revised draft of the FCL guidance note.

10

See “Reform of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument— Specific Proposals–Revised Annex I,” IMF Policy Paper, May 2014, and “Flexible Credit Line—Operational Guidance Note,” IMF Policy Paper, June 2015.

11

The only exception is the “Sound public finances, including a sustainable public debt position” criterion as the underlying assessment of debt sustainability with high probability is forward-looking in nature.

12

As in the case of the bottom-line assessment on the nine qualification criteria, very strong performance against all qualification criteria is not required.

13

See Press Release No. 14/84, March 5, 2014.

14

While high-access drawing arrangements which go off-track for extended periods would similarly end-up tying up Fund resources and potentially entailing costs to creditor members, in staff’s view, it would be more appropriate and effective to rely on policy discussions to facilitate progress with policy implementation, rather than imposing additional fees that would add to members’ external imbalances.

15

As this option builds on the existing fee structure, it is envisaged to apply to all GRA arrangements. The impact on drawing arrangements would be mitigated by the fact that commitment fees are refunded upon purchases.

16

While this drawback could be mitigated by not steepening the current commitment fee schedule for extended arrangements, which are not generally provided on a precautionary basis, it would add complexity to the Fund’s overall fee structure. See footnote 14 above and footnote 21 below.

17

For the purposes of commitment fees, the level of undrawn credit means the amount that could be purchased during the relevant period (12 months or the period left under the arrangement, if shorter) as per Rule I-8.

18

This is analogous to the application of time-based surcharges, which are levied for the period that members’ credit outstanding remains above a threshold, currently set at 187.5 percent of quota.

19

The clock toward meeting the TBCF trigger would pause once undrawn balances fall below the threshold, and would reset once undrawn balances remain below the threshold for a defined continuous period of time (“cooling off” period). See Annex V for details.

20

To date, based on quotas in effect after completion of the 14th General Review of Quotas, only one FCL arrangement has had access above 575 percent of quota (Mexico’s FCL arrangement approved May 2016; see text figure).

21

In the context of the 2000 Review of Fund Facilities, “the Board confirmed that it saw a need to ensure that arrangements under the EFF be granted only in cases that met fully the terms and spirit of the EFF Decision. These would be cases where there is a reasonable expectation that the member’s balance of payments difficulties will be relatively long-term, including because it has limited access to private capital, and where there is an appropriately strong structure reform program to deal with the embedded institutional or economic weaknesses. The Board agreed that extended arrangements should generally not be formulated on a precautionary basis, as circumstances where potential balance of payments difficulties were likely to turn out to be longer-term are probably very rare.” See “Summing Up by the Acting Chairman at Review of Fund Facilities––Proposed Decisions and Implementation Guidelines.”.

22

See footnote 14 above.

23

Past commitments would be remeasured in percent of 14th Review of Quotas and compared against the time-based fee threshold.

24

The FYR Macedonia’s January 2011 PCL was converted into a PLL arrangement by the time of the second review discussions. Morocco’s first PLL arrangement was in 2012.

*

This staff proposal was revised in the subsequent paper “Adequacy of the Global Financial Safety Net—Review of the Flexible Credit Line and Precautionary and Liquidity Line, and Proposals for Toolkit Reform—Revised Proposals,” IMF Policy Paper, December 2017.

25

See “Review of Fund Facilities—Analytical Basis for Fund Lending and Reform Options,” 02/06/09 and “GRA Lending Toolkit and Conditionality: Reform Proposals,” 03/13/09.

26

See footnote 1.

**

For updated information with respect to this proposal, see the subsequent paper “Adequacy of the Global Financial Safety Net—Review of the Flexible Credit Line and Precautionary and Liquidity Line, and Proposals for Toolkit Reform— Revised Proposals,” IMF Policy Paper, December 2017.

27

There will be no ex-post conditionality, including no standard continuous Performance Criteria (PCs), and no reviews during an arrangement.

28

Qualification criterion 4 will be amended slightly from that in the FCL arrangement to eliminate the language “When the arrangement is requested on a precautionary basis” to reflect that an SLS arrangement can only be requested when the member faces a potential but not an actual BOP need.

29

Pursuant to Article V, Section 7(d), the Fund, by an eighty-five percent majority of the total voting power, may adopt repurchase periods other than that which applies within the credit tranches.

30

Liquidity episodes in advanced economies also tend to be short-lived, as revealed by episodes of FX market dysfunction, when the covered interest parity breaks. The bulk (about 80 percent) of these episodes lasted for no more than two weeks, with the GFC being the sole exception. For details, see Box 1 in “Assessing Reserve Adequacy—Further Considerations” IMF Policy Paper, 2013.

31

Purchases would be subject to a one-year repurchase obligation without installments. As such, each purchase would have to be repurchased within a year of the purchase.

32

Even in the case of the GFC, the BOP pressures only lasted for about one year.

33

In addition, as the SLS, like other special facilities, is proposed to “float” against the reserve tranche (i.e., members would be able to maintain a reserve tranche position in the Fund despite making purchases under the facility), an 85 percent majority of the total voting power would be required to establish this feature (Article XXX(c)(iii)).

34

See “Adequacy of the Global Financial Safety Net—Considerations for Fund Toolkit Reform,” IMF Policy Paper, December 2017.

35

Revolving access within an arrangement is not a new design element in the Fund’s toolkit. For example, prior to the effectiveness of the Second Amendment in 1978, members were allowed to reconstitute drawing rights through early repurchases before the expiration of a stand-by or extended arrangement. Reconstitution of access was also a feature of the Currency Stabilization Fund (CSF), which was a window established in 1995 to provide Fund financial support in the context of Fund arrangements. Purchases were subject to the member’s compliance with the arrangement, including supplementary conditions set out at the establishment of the CSF. For details see “The Acting Chairman’s Summing Up at the conclusion of the Discussion on Fund Support for Currency Stabilization Funds,” Executive Board Meeting 95/86, September 13, 1995; Box 8 in “Review of Fund Facilities––Preliminary Considerations,” Executive Board Specials 00/37, March 2, 2000.

36

Accordingly, in line with normal practice, SLS arrangements would be scored against the Fund’s Forward Commitment Capacity (FCC). However, repurchase obligations resulting from purchases under an SLS arrangement would not be added to the FCC, unless the SLS arrangement expires or is cancelled and no approved successor SLS arrangement is in place.

37

The process of re-qualification will be the same as if there had been no prior arrangement (see subsection D). For members with credit outstanding when the successor arrangement is approved, rights to make purchases would build pro tanto as outstanding credit is repurchased (see subsection C).

38

As a legal matter, charges and fees must be uniform (Article V, Section 8 (d)). Uniformity, however, does not mean that all Fund charges and fees need to be the same across Fund arrangements; rather differentiation would be permitted based on criteria that are relevant to the use of Fund resources. Specifically, the differentiation could be made in light of the nature of the Fund arrangements and the purpose of the fees and charges. In the Fund’s practice, “[d]ifferentiation of charges has been limited to relevant differences in members’ use of the Fund’s resources (e.g., having a different balance of payments need as addressed by a special facility)”, and such differentiation needs to be consistent with the purposes of the fees and charges. See, e.g., “The Fund’s Mandate—Future Financing Role,” IMF Policy Paper, March 26, 2010, at p. 24.

39

Once credit is drawn under the SLS, the rate of charge would apply as it does for any credit outstanding under the GRA.

40

A similar idea was discussed, but not adopted, in 1995 in the context of the establishment of the Currency Stabilization Funds (CSF), which was also intended to have revolving nature. Specifically, staff had considered charging the commitment fee on the “net amount of resources” committed by the Fund under the CSF during a relevant period, so that reconstituted access from repurchases would not increase that net amount. However, staff noted that this approach may not be of much relevance because the commitment fee is refundable at the time purchase.

41

Eliminating the need to pay commitment fees when access is reconstituted, but keeping the commitment fee refundable upon purchases could leave the Fund with outstanding commitments to members for which no (net) commitment fee had been paid.

42

For example, if at the time of approval of a successor SLS of 145 percent of quota, a member has drawings outstanding under its previous SLS of 100 percent of quota, the proposed 8 bps non-refundable commitment fee would be applied to the entire access of 145 percent of quotas, notwithstanding that at the time of approval access of only 45 percent of quota would be available for drawings.

43

Under the credit tranches, if no actual BOP needs arise during a 12-month precautionary SBA or FCL, the commitment fee paid upfront, 18 bps for 145 percent of quota, is retained by the Fund as income at the end of the arrangement. If the member experiences an actual BOP need and purchases the commitment in full, it would pay 50 bps in service charge, and receive a full refund of its commitment fee. If the member makes a partial drawing, it pays 50 bps for the drawn amount and receives a commitment fee refund pro-rated by the drawn amount. The remaining commitment fee is retained as income at the end of the arrangement.

44

Under the credit tranches, there is no revolving feature. The cost of a single purchase, with refundable commitments fees, is the 50 bps service charge. The cost of using the SLS’s revolving feature once, i.e., making two drawings, is also 50 bps under the proposed pricing structure of a 8 bps nonrefundable commitment fee and a 21 bps service charge for each of the drawings.

45

The expression of interest in being assessed is informal (e.g., could be verbal or written) and serves the purpose of eliminating the members that do not want to be assessed.

46

Staff will make clear to the member that the Executive Board takes the final decisions on qualification and access.

47

Given the SLS’s low access limit and the fact that the member accepts an SLS arrangement after the Executive Board conditionally approves the arrangement, it is envisaged that an informal Board meeting will not be required. In contrast, an informal meeting prior to Board approval is a requirement for the FCL and PLL, given the potential for much higher access levels.

48

The normal periods would apply to document circulation, but circulation could be expedited under special circumstances, similar to the FCL.

49

Staff does not see the need to discourage members from making public disclosures of the offer; since countries need to initiate the qualification process, as under the FCL, the likelihood of a country qualifying but not opting for the instrument is considerably reduced.

50

A further Board meeting would not be required as the Board had already approved the extension of an offer. Restricting the time between the extension of an offer to a member and the opt-in by the member to two weeks protects the Fund against a significant change in members’ conditions.

51

As is standard under Fund facilities, while the Fund would not challenge a representation of need by a member for a purchase requested under the SLS, the member’s drawings would have to be commensurate with its actual BOP need at the time of the purchase, notwithstanding the available amount of approved access.

52

Staff will generally rely on the member’s representation regarding its domestic legal authority.

53

Moreover, the central bank would need to be the designated fiscal agent for the Fund to be able to communicate the member’s request for purchase under the facility.

54

Irrespective of who the signatory of the written communication is, the counterpart obliged to make repurchases and pay any applicable charges vis-à-vis the Fund is the member country, not the specific agency or authority signing the written communication.

55

The Fund has extensive experience in dealing with avoiding a negative signaling impact. Moreover, the SLS largely follows the FCL and PLL in terms of process, and as such, any signaling effects created by the SLS would already exist under the FCL/PLL. While all members have continued to qualify and no member has yet exited from the FCL or PLL arrangements, the six cases of access reductions to date have not been associated with marked changes in yields, spreads, or yield volatility (see Annex I).

56

See paragraph 6(b) of Decision No. 14283-(09/29), adopted March 24, 2009, as amended.

57

Qualification is assessed against the nine criteria as well as the strength of the member’s policy track record, its institutional policy framework, and the overall assessment of policies in the most recent Article IV staff report. The assessment was done for the full membership, except reserve currency issuers (or countries with access to a reserve currency-issuing central bank, in the case of the Euro area). Any final assessment would require further in-depth consultation with area departments and would also be subject to Board approval.

58

Under the policies and procedures adopted by the Executive Board, the assessment of the strength of members’ external positions for the purpose of selecting their currencies for transfers during an FTP period is conducted in consultation with members and, while anchored in a number of indicators, is ultimately a matter of judgment.

59

The second-round effects on Fund liquidity of purchases made by members that are included in the FTP comprise two elements: (i) the reduction in the Fund’s holdings of currencies that are available to finance purchases and (ii) conversely, the reduction in the Fund’s need to set aside quota resources as a prudential balance. The second-round effects depend on the level of members’ Reserve Tranche Positions and Fund holdings of currencies and fluctuate over time. On average, second-round effects are estimated to be about 70 percent of a member’s quota.

60

If a member repurchases the credit outstanding under the SLS in full and continues to have an SLS arrangement, consideration can be given to automatically add the member back to the FTP so that transfers of its currency may be made for the remainder of the plan period, subject to the strength of the member’s external position and in consultation with the member. These procedures would be laid out in a future paper for the Executive Board’s consideration.

61

In line with past practices, these thresholds are based on committed resources and therefore exclude any potential second-round effects.

62

Ceteris paribus, i.e., assuming no change in the current commitment under the FCL (SDR 77 billion), the threshold would be reached with SLS commitments of SDR 73 billion. This would reduce the FCC to SDR 136 billion.

63

Specifically, the FCC is defined as the IMF’s stock of usable GRA resources minus undrawn balances under existing commitments, plus scheduled repurchases less repayments of borrowing, both measured over the coming 12 months, minus a prudential balance.

64

As noted at the time of the FCC’s introduction, the prior approach of making adjustments to precautionary commitments under the Fund’s liquidity ratio was not transparent, could lead to dramatic changes as the precautionary status of arrangements changed and could result in the Fund implicitly recommitting already committed resources.

65

See “Review of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument,” IMF Policy Paper, January 28, 2014, and the Press Release No. 14/84, March 5, 2014.

66

There have been no drawings to date under the FCL. However, staffs analysis for assessing potential drawings shows that these would likely take place in a correlated manner in event of a severe downturn scenario (see Annex V in “Review of the Adequacy of the Fund’s Precautionary Balances.” Also, the proposed new liquidity instrument is designed to handle shocks of more frequent nature, suggesting increased probability of potential drawings.

67

Bilateral borrowing agreements would also be potentially available, but only as a third line of defense after quotas and the NAB have been largely exhausted. As with the NAB, these cannot be used to finance drawings under commitments made prior to their activation, and would therefore not be available under partial scoring to finance drawings under existing commitments.

68

Beyond this practical constraint, the current NAB decision allows the Managing Director to propose activation when there is a need for supplemental resources to forestall or cope with an impairment of the international monetary system. The decision specifies that such proposals would only be made if the Fund’s capacity to make commitments from quota-based resources has been or is expected shortly to be substantially depleted, but does not specify any threshold for the FCC that must be met before an activation can be proposed.

69

All these documents reference the FCC as defined in Decision No. 14906 (11/38), adopted April 20, 2011.

1

Regressions control for global factors by using as the left-hand-side variable the residuals from a regression of EMBI spreads on the VIX. The “access dummy” takes the value of 1 on the day when lower access was approved by the Fund’s Executive Board, and 0 otherwise.

1

The “optimal” threshold minimizes the sum of type I and type II crisis prediction/misclassification errors to avoid missing crises (threshold too high) and predicting false crises (threshold too low).

1

PRGT borrowers under the SCF pay 15 bps in availability fees on undrawn balances at the end of each 6-month period.

2

The marginal increase was not assessed to have a meaningful impact on liquidity risks facing the Fund. For details, see “Review of Access Limits and Surcharge Policies.”

3

The level of undrawn credit means the amount that could be purchased during the relevant period as provided by Rule I-8.

4

Due to the legal requirement of uniformity of charges, application of the TBCF cannot be ring-fenced to specific arrangements within the credit tranches. Furthermore, the application cannot legally be restricted only to arrangements that are treated as precautionary.

5

See, for example, Dias, D. A. and Richmond, C. “Duration of Capital Market Exclusion: An Empirical Investigation,” July 2009; Mecagni, M., Atoyan, R, Hofman, D., and Tzanninis D., “The Duration of Capital Account Crises—An Empirical Analysis,” IMF Working Paper 07/258, November 2007; and Hatchondo, J.C. and Martinez, L, “Sudden stops, time inconsistency, and the duration of sovereign debt,” IMF Working Paper 13/174, July 2013. The duration could in principle also be based on a typical length of heightened global stress, but this would require more subjective assessment.

1

The Fund’s reserve adequacy literature also identifies alternative metrics that may be relevant in specific circumstances, including cost of borrowing or sterilizing the intervention for acquiring reserves when their level is low. However, the opportunity cost of reserves is a more suitable measure for comparisons to Fund commitment fees, as the costs of the latter are more relevant under precautionary arrangements where the member does not face an immediate balance of payments need and has broadly adequate reserve holdings.

2

Conceptually, such measure should be further adjusted for the fact that higher reserves may reduce the local yield of reserves as they reduce risks. However, as noted in “Assessing Reserve Adequacy––Specific Proposals,” IMF Policy Paper, December 19, 2014, the feedback from higher reserves to lower yields appears to diminish with the level of reserve holdings. The impact is not found significant for advanced economies and while relevant for EMs in the past, the rise in reserve holdings in recent years have essentially eliminated this effect.

3

In principle, high access may also apply to SBAs treated as precautionary, but there have been no instances of such arrangements in recent history. Also, the more gradual accumulation of access rights under SBAs imply that these are less likely to lead to commitments subject to higher commitment fees for extended time periods.

1

Specifically, the FCC is defined as the IMF’s stock of usable GRA resources minus undrawn balances under existing commitments, plus scheduled repurchases less repayments of borrowing, both measured over the coming 12 months, minus a prudential balance.

2

Review of the Flexible Credit Line, the Precautionary and Liquidity Line, and the Rapid Financing Instrument,” IMF Policy Paper, January 28, 2014.

3

Press Release No. 14/84, March 5, 2014.

4

For the purposes of this annex “precautionary arrangements” or “precautionary commitments” refers to arrangements or to commitments under arrangements which the authorities expressed an intention to treat as precautionary and under which no purchases have been made. “Drawing arrangements” in this annex refer to arrangements with no expressed intent of precautionary use by the authorities.

5

Only 3 countries have so far made use of the FCL, with the largest individual commitment currently representing three-quarters of the total (one member has made use of the PLL).

6

There have been no drawings to date under the FCL. However, the new liquidity instrument is designed to handle shocks of a more frequent nature, suggesting that the probability of drawings could be expected to increase.

7

The decision specifies that such proposals would only be made if the Fund’s capacity to make commitments from quota-based resources has been or is expected shortly to be substantially depleted.

8

All these documents reference the FCC as defined in Decision No. 14906 (11/38), adopted April 20, 2011.

9

The last constraint could be addressed if the current NAB Decision were amended to allow it to be used for existing commitments. However, it is not clear whether such a change would be feasible, given that the NAB decision was only just renewed. An alternative would be to explore new credit lines with central banks that could only be drawn upon in specific circumstances involving a widespread drawdown of precautionary arrangements.

10

Such judgment should also be informed by the precautionary commitments’ potential second round liquidity implications as discussed in section III.

11

The Fund could enter into new precautionary commitments of up to SDR 208 billion without over-committing its available quota resources. In this event, the remaining FCC would be about SDR 143 billion whereas the Fund’s actual ability to commit new resources would be fully exhausted.

12

While the Fund could potentially seek to reduce the risk of over-committing its existing resources by capping commitments under the new liquidity instrument (e.g., at SDR 100 billion), it would still be reporting a FCC (e.g., SDR 199 billion in the third column) significantly larger than its actual ability to make new commitments over the next 12 months, which is what the FCC explicitly purports to measure.

  • Collapse
  • Expand
Adequacy of the Global Financial Safety Net—Review of the Flexible Credit Line and Precautionary and Liquidity Line, and Proposals for Toolkit Reform
Author:
International Monetary Fund
  • Figure 1.

    Evolution in the Severity of Shocks Relative to Previous Arrangement1/

  • Reserve Adequacy in Current Arrangements

    (Percent of reserve adequcy metric; based on gross international reserves)

  • Access under Successive FCL/PLL Arrangements

    (Percent of post-14th review quota)

  • Financial Pressures in Emerging Markets

    (Percent of large EMs; VIX level)

  • Portfolio Flows and Exchange Market Pressure During Recent Liquidity Episodes

  • Figure 2.

    Envisaged Configuration of the Fund Precautionary Toolkit

  • Figure AI.1.

    Access in FCL and PLL Arrangements

  • Figure AI.2.

    Access Reduction and Market Reaction

  • Figure AIII.1.

    Evolution of ESI

    (Negative: adverse external economic conditions)

  • Figure AIV.1.

    Reserve Accumulation in Access Scenarios

  • Figure AIV.2.

    Adverse Scenario: Contributions to First-Year Financing Gap

  • Reserve Adequacy in Current Arrangements

    (Percent of reserve adequcy metric; based on gross international reserves)

  • Figure AV.1.

    Illustrative Commitment Fee Structures 1/

    (Basis points)

  • Figure AVI.1.

    Cost of Accumulating Reserves for Select Members

    (Basis points)