Chapter 4. Strengthening the Global Financial Safety Net
- International Monetary Fund. Independent Evaluation Office
- Published Date:
- January 2015
This chapter discusses the actions taken by the IMF to contribute to strengthening the global financial safety net in response to the crisis. First, it examines the IMF’s resource mobilization efforts, then the reforms of lending instruments, and finally the design and implementation of IMF-supported programs.
The chapter concludes that the IMF’s efforts in this area were largely successful. Although the IMF was not well positioned in advance to respond to a crisis of this magnitude, it responded quickly. It quadrupled its resources and lent almost $400 billion to 38 countries to help them deal with the crisis; it also raised additional concessional resources, facilitating an almost doubling of lending to LICs. It modified its lending instruments to make them better suited to the circumstances—speeding up negotiations, loosening access limits, increasing front-loading, and streamlining conditionality. It launched precautionary instruments, although their design still needs fine-tuning to address limited demand and concerns on exit. The current credit capacity at $1 trillion seems appropriate, but with an agreed increase in IMF quotas still pending, the size and modalities of the IMF’s financial resources remain an issue going forward.31
A. Resource Mobilization: Strategy and Results
In September 2008, IMF credit capacity stood at about $250 billion, of which $210 billion were in quotas and the rest in two standing arrangements, the General Arrangements to Borrow (GAB) and the New Arrangements to Borrow (NAB), through which the IMF could borrow from certain member countries in case of extraordinary needs.
The IMF and its members had twice assessed the adequacy of IMF quotas since they were last increased in 1998, but those discussions took place at the time of the “Great Moderation,” when country authorities and to a significant degree IMF Management and staff deemed it unlikely that substantial IMF lending would be needed.32 Further, in the years leading up to the crisis, IMF liquidity was high, because few members had sought financial support. As a result, there was no strong push from IMF Management or consensus within the membership for a general increase in quotas at either the 2003 or 2008 reviews.33
With the crisis escalating, policymakers turned their attention to increasing the IMF’s resources, as concern grew about their adequacy.34 In April 2009, the IMFC endorsed a multi-pronged strategy that had been articulated earlier that month by the G20 Leaders. This strategy consisted of borrowing from member countries (partly as a bridge to a quota increase), and accelerating the 14th General Review of Quotas for completion by January 2011. To boost global reserves the IMFC agreed on an issuance of new SDRs. The IMF also sought to double the concessional resources available for LICs.
The IMF thus dramatically increased its financial firepower to more than $1 trillion by end-2012. The resource mobilization effort allowed the IMF to respond to member country requests for financial support, and authorities interviewed for this evaluation were satisfied overall with the results of this effort.
However, the resource increase has, thus far, come solely from three waves of borrowing. First, a series of bilateral borrowing agreements with individual member countries almost doubled the IMF’s credit capacity to $460 billion by March 2010. Second, in March 2011, an expanded $580 billion NAB took effect, raising credit capacity to more than $725 billion. 35 Finally, in late 2013, a new round of bilateral borrowing provided potential additional resources of more than $400 billion. A doubling of IMF quotas and associated governance reforms was agreed in December 2010. These have not taken effect because the United States has not ratified all the necessary agreements.
The first wave of borrowing arrived only “just in time” to ensure that the IMF was not liquidity-constrained in responding to program requests. A number of the interviewed authorities pointed out that an important contribution of the IMF to global financial stability is the confidence it gives to financial markets that resources are available in advance to deal with crises. In the early months of the crisis, the IMF could not play this role of calming the markets.
Because the agreed quota increase has not yet taken effect, the IMF remains reliant on borrowing for 70 percent of its credit capacity, and access to more than half of the IMF’s credit capacity is controlled by a super-majority of creditors. Agreement on the resource mobilization strategy, and success in securing borrowed resources, hinged importantly on the understanding that borrowing would not substitute for a quota increase. This principle underlies the statement by the IMFC, in its initial endorsement of the strategy, that “while an expanded NAB is an important backstop for IMF resources, we recognize that it is not a substitute for a quota increase” (IMF, 2009e). Some of the authorities interviewed for this evaluation were also concerned about the risks involved in the need to renew and reactivate the NAB and to extend the bilateral borrowing agreements.
The prolonged reliance on borrowing undermines the IMF’s functioning as a universal cooperative that is governed by all members through a system of weighted voting. The quota increase, although it would bring only a small additional increase in credit capacity, would have important implications for IMF governance. 36 In addition to restoring the primary reliance on quotas, the 2010 quota reform would bring a shift in shares and chairs from advanced economies to faster growing emerging markets. Although the shift would still leave EMEs under-represented relative to their shares in the global economy, the 2010 reform has been seen as an important step to enhancing the legitimacy of IMF governance.
In addition to dramatically increasing resources for its general lending, the IMF nearly doubled its concessional lending capacity by raising additional loan and subsidy resources. Also, in 2012, the IMF put in place a strategy for a self-sustaining framework for concessional lending with an annual capacity of about SDR 1.25 billion going forward.
Another important contribution to global liquidity was the increase in global SDR holdings by the equivalent of $250 billion in August 2009.37 The new allocation expanded global SDR holdings tenfold, with nearly $100 billion going to EMEs and developing countries. This represented a significant increase in their reserves; and more broadly, it boosted global liquidity and arguably contributed to market confidence.
B. Updating the Lending Toolkit
When the crisis struck, the IMF was already in the midst of reconsidering its lending facilities. With virtually no demand for nonconcessional lending, many felt that the existing IMF lending instruments did not match member needs. The potential for an IMF crisis-prevention instrument had been much discussed, but no consensus had been reached about design and terms. Further, countries considered that approaching the IMF for support entailed stigma and were accumulating large precautionary reserves as self-insurance or were pursuing alternatives such as reserve pooling arrangements.
The crisis intensified the discussion of the IMF’s lending toolkit, resulting in decisions in March 2009 to recast the terms of existing lending instruments and introduce new instruments for precautionary lending. The reforms included:
A doubling of the limits on the level of resources normally available under nonconcessional programs; greater front-loading of resources at the start of a program; and a rationalized structure for charges, maturities, and fees.
Streamlined conditionality, including by eliminating structural performance criteria, and recommitting to greater parsimony and criticality in conditionality.
Two new precautionary instruments to make resources rapidly available with high or no access limits: the FCL and the Precautionary and Liquidity Line (PLL).38 These instruments require pre-qualification based on strong policies; the FCL has a higher qualification bar and no ex post conditionality for drawing the resources.
In July 2009, the IMF established the Poverty Reduction and Growth Trust (PRGT) which has three concessional lending windows, to better address the needs of LICs. It also doubled access limits and temporarily set a zero percent interest on concessional credits, which has been extended and continues through end-2014.
Three FCL arrangements were approved shortly after the creation of this new instrument—for Mexico (SDR 31.5 billion), Poland (SDR 13.7 billion), and Colombia (SDR 7 billion). Three successor arrangements have been approved for each country, with their terms extended to two years.39 As intended at the time they were approved, countries have not drawn on these FCLs.
Authorities in countries with FCL arrangements believe that the FCL played an important role in calming markets and continues to be a useful tool in maintaining confidence in a time of uncertainty in the global economy. They praised the FCL as having served as a signal of support for their macroeconomic policies and a “seal of approval” that has helped promote market confidence.
However, no additional FCL arrangements have been approved, even in the face of waves of global market stress in the five years since its creation. Some authorities interviewed for this evaluation argued that the FCL’s strict qualification criteria may preclude many countries from accessing it, but surveys conducted by the IMF indicate that a preference for self-insurance, access to alternative financing, and stigma were key factors inhibiting FCL use.40
None of the FCL users has yet exited the instrument. Authorities in these countries believed that the FCLs should remain in place until the IMF unequivocally communicates that global risks have subsided. A number of other authorities indicated a concern that continued use of the FCL ties up IMF resources for an extended period. In any case, there is widespread understanding that pushing users to exit could create signaling problems and undermine the confidence-building objective of the instrument. These issues suggest a need for further experimentation and innovation in precautionary lending instruments.
Overall, the reforms of lending instruments addressed many of the concerns of member countries about the lending toolkit, and helped make IMF lending more helpful in coping with the crisis.
C. Extending Financial Support to Member Countries41
Member countries hit by the crisis began turning to the IMF for financing support immediately in September 2008. The primary tool of support was the SBA: the IMF approved 17 SBAs for more than SDR 50 billion in the first year of the crisis and an additional 20 SBAs for SDR 50 billion between September 2009 and the end of 2013 (eight countries were supported by more than one SBA). The IMF has also deployed the EFF several times since 2008, increasing the use of this instrument over time. In addition, it moved quickly to launch the FCL, rapidly approving the first three arrangements under this new instrument, which together initially totaled SDR 52 billion, and rose to SDR 76 billion.
The IMF also provided substantial concessional support to LICs. LICs were not at the epicenter of the crisis that struck in September 2008, but they faced volatility in food and fuel prices and subsequently confronted declines in trade and tourism, as well as fluctuations in aid flows. The IMF expanded its average annual concessional lending commitments to SDR 1.6 billion on average in 2009–12, up from SDR 0.7 billion on average annually in 2000–08. A recent IEO review identified progress in addressing issues in the IMF’s engagement with LICs during the last decade, while also pointing to a need for continued attention to supporting broad-based growth, poverty reduction and the safeguarding of social and other priority spending (IEO, 2014).
The remainder of this section focuses on 25 SBAs approved during 2008–11. The focus of the analysis is on the main shared characteristics of the programs, and not on the details of the negotiations, design, or implementation of each of them. Some of these details are referenced in this report to illustrate the “big picture.”
Learning from experience
As a whole, experience with these programs suggests that the IMF took into account lessons from past crises. Programs were designed and negotiated faster than in the past. They accommodated a larger share of countries’ needs, understanding that current account adjustment was more difficult in a global crisis and that capital markets were more volatile and undergoing a flight to quality. These larger programs also allowed for more front-loading, and conditionality was streamlined (see Box 3). Overall, these SBAs helped the member countries cope with the crisis, but clearly there was room for further learning and improvement in each of them. Some of these country-specific issues are presented in ex post evaluations prepared by IMF staff, but they are not generally discussed in this report.
Emergency mechanisms to speed program design and approval were activated for five new programs and two augmentations. The first 14 SBAs took an average of 6.2 weeks from the start of program discussions to Board approval, with the Ukraine program approved in only 4 weeks—among the fastest processing times in IMF history. The time taken to approve programs lengthened as the crisis progressed, averaging 12.2 weeks overall for SBAs approved in September 2008 through end-2011.
All but one of the SBAs approved in the first years went beyond the normal access limits. Four early SBAs—for Hungary, Iceland, Latvia, and Romania—exceeded 1,000 percent of quota. The IMF noted in its initial review of crisis programs that programs in the initial wave amounted to nearly 6 percent of GDP, which was higher than the average of about 4 percent of GDP in past capital account crises, although similar in terms of the share of gross financing needs (IMF, 2009g). IEO analysis (Takagi and others, 2014) confirms this conclusion: access was on average almost 4 percentage points of GDP larger than that in SBAs approved in 1997–99. In particular, these four European programs were three to five times larger in relation to GDP than were the programs approved in 1997 for Thailand, Indonesia, and Korea. In this regard, authorities—mainly but not only in EMEs—indicated that they hoped that the larger access that characterized IMF financing in the post-2008 European programs would serve as a precedent for future crisis lending.
Box 3.IMF Learning from Crises in Asia and Latin America
Considerable learning has taken place at the IMF since the emerging market crises of the late 1990s and early 2000s. As a result, IMF-supported programs in response to the current crisis have included:
Large, front-loaded access, in collaboration with other partners.
Proactive involvement with the private sector, for example, the Vienna Initiative.
Streamlined structural conditionality that is more focused on the IMF’s core areas of competence.
Greater awareness of balance sheet effects in designing exchange rate policy and forecasting the impact of the crisis on growth.
Explicit recognition of risks and contingency planning in case assumptions fail to hold.
Flexibility in targets and approaches (including direct budget support and judicious use of capital and exchange controls).
A public communications strategy to build investor confidence and public support through enhanced transparency and by explaining the logic of the programs.
IEO analysis found an average front-loading factor of about 25 percent in 2008–13, almost 10 percentage points higher than in pre-2008 programs. Front-loading was greater in the earlier programs, and decreased as the crisis subsided. It was greater in countries that faced both current account and banking crises, and where the fiscal deficit was larger in relation to GDP; it was smaller in successor arrangements, in programs that were larger in relation to quotas, and where countries had larger reserves relative to GDP.
Departing from normal practice, IMF-supported programs allowed direct budget support, including in Ukraine, Hungary, Latvia, Romania, and Bosnia and Herzegovina.42 Although this was not without precedent, it reflected the IMF’s adaptation to the reality of governments losing access to funding; this approach was subsequently incorporated in guidelines for staff issued in March 2010. Several senior officials in Europe interviewed for this evaluation indicated that the prospect of direct budgetary support raised the attractiveness, and helped overcome the stigma, of IMF financing at a critical time.
SBA-supported programs generally targeted a gradual reduction in the fiscal deficit, seeking to manage the trade-off between supporting the economy during a downturn and achieving medium-term fiscal sustainability. Unlike the 1997 Asian crisis programs, no post-2008 crisis program sought to achieve a surplus in the short run. In Iceland and Latvia, the programs targeted a large initial increase in the fiscal deficit, in light of the expected costs of bank restructuring. In the event, fiscal deficit outturns were larger than programmed, because the IMF relaxed targets when the crisis proved to be more severe than forecast. Even so, in most instances IMF financing did not appear to have accommodated the full extent of the fiscal shortfall, with automatic stabilizers partially offset by fiscal measures.
About half the programs called for greater exchange rate flexibility, although the IMF was alert to the possibility that a large depreciation could have adverse balance-sheet effects.43 While early SBAs saw a similar size of depreciation to that in SBAs before the crisis, their currencies were stabilized once the program was in place; later programs saw little or no currency depreciation. Coupled with large, front-loaded financing, judicious use of capital and exchange controls may have contributed to this outcome, as was evidenced in Iceland, for instance.
Structural conditionality was more streamlined and more focused on the IMF’s core areas of competence. Structural conditionality, as measured by the sum of performance criteria, structural benchmarks, and prior actions, was considerably lighter than it was in programs of the period 1997–2000 (see Figure 3).44 As the crisis evolved, however, the number of structural conditions increased somewhat, with the average number rising from 8.8 per year in the SBAs approved in 2008 to more than 10 in those approved in 2010 and 2011 (compared to 15.3 per year in programs approved in 1997).45 Similarly, the focus of structural conditions increasingly extended beyond the IMF’s core areas of competence over time. In the sample of 25 SBAs studied by the IEO, the share of structural conditions that fell in core areas declined from 87 percent in those programs approved in 2008 to 68 percent in those approved in 2011.
Figure 3.Structural Conditionality in IMF Stand-By Arrangements, 1997–2000 Versus 2008–11
Sources: IEO (2007) for 1997–2000; IEO staff estimates for 2008–11.
Program risks were covered in Board documents for all programs but were presented in a pro forma manner. An IEO review of internal staff documents indicated that IMF staff had done serious due diligence in contingency planning during program design and negotiation. However, staff had difficulties finding ways to convey these contingent plans without risking undermining program implementation.
The IMF devoted significant efforts to explaining that the programs were credible—a critical element, given that an important aim of the programs was to restore investor confidence. There were frequent outreach activities to explain the programs to politicians, business and labor leaders, journalists, and academics, in order to build national ownership of the IMF-supported programs.
The IMF collaborated with other multilateral and bilateral donors in a transparent manner, especially in the early European programs. At least 17 of the 25 SBAs studied by IEO, including 7 of the 8 European programs, involved some degree of collaboration with other agencies. Organizations collaborating with the IMF characterized their working relationships as effective, although there was a learning curve due to the lack of established protocols for such collaboration with some of the organizations. 46 The program for Hungary represented the first case of IMF-EU collaboration, and set a precedent for future requests for financial support by EU members. It was followed by Latvia and then others. In all cases, IMF staff enumerated these additional sources of financing in a transparent way, which helped to enhance the credibility of the financing packages.
Private sector involvement
The IMF proactively sought private sector involvement, particularly in those European countries where foreign-owned banks had a large presence. In Hungary, the IMF organized a meeting of public entities and the foreign strategic owners of six large banks immediately after starting to negotiate the program. Subsequently, the IMF actively participated in the Vienna Initiative that aimed at keeping the private sector involved, as discussed in Chapter 2.
Overall, IMF programs in the crisis aftermath helped contain the economic and financial fallout from the crisis. Unlike in previous emerging-market crises, a widely-feared financial meltdown was avoided (except in Iceland, where the collapse of the banking sector was a fait accompli by the time the country approached the IMF). A limited number of bank failures occurred (e.g., in Ukraine and Latvia), but even there the fiscal costs were contained.47 Latvia could not avoid a deep recession, but nonetheless succeeded in its primary objective of defending its currency peg, allowing it to adopt the euro on January 1, 2014. Large programs also contributed to restore investor confidence.48
Despite these overall successes, IMF engagement also encountered policy reversals and program interruptions. In a number of countries, especially in high-access cases, structural reforms and fiscal consolidation efforts did not progress much or were reversed after the program engagement ended. For example, in Belarus, fiscal policy was relaxed as soon as the program ended, and quasi-fiscal activities expanded (their containment had been the program’s key objective). In Hungary, although substantial fiscal consolidation had been accomplished, some of the achievements were reversed after the program relationship ended. These are not isolated instances, and they highlight the perennial issue of whether structural and long-term fiscal issues can be effectively tackled by conditionality during a crisis—or more practically of how to design reforms that will be sustained beyond the program relationship with the IMF.
Staff often found it difficult to build consensus on reforms during the short duration of a program and, coupled with their legitimate concerns about downside risks, did not always press their case vigorously. The authorities’ interest in continuing with the program engagement was not sustained once the acute phase of the crisis was over. Only 62 percent of the committed resources were drawn; and 11 of the 25 SBAs were ended ahead of their original expiration date, sometimes without completing several of the programmed reviews.49 This is not a new phenomenon; experience shows that countries with IMF programs frequently revert to their own policy framework and timetable once they no longer need IMF support. Nonetheless, beyond mitigating the immediate acute crisis, it leaves a question about whether recent crisis programs contributed to medium- or long-term sustainability.