Chapter 2. Living with Rules: The IMF’s Exceptional Access Framework and the 2010 Stand-By Arrangement with Greece

Moisés Schwartz, and Shinji Takagi
Published Date:
June 2017
  • ShareShare
Show Summary Details
Susan Schadler


The handling of the IMF’s framework governing exceptional access to crisis countries will be one of the most important legacies of the euro debt crisis.1 The framework has as its objective to set procedures and substantive standards to guide IMF decisions on providing exceptionally large financing, typically in capital account crises. The initial framework was adopted in 2002. Its purpose was to distill best practices for resolving crises efficiently and to steer decisions away from mistakes made in previous crises. However, as the crisis in Greece loomed in early 2010, the substantive part of the framework—four criteria that had to be met for the IMF to provide exceptional access—was seen by some as preventing the optimal response to the crisis. Accordingly, the criteria were changed, and the amended version was applied to the IMF’s decisions first in Greece and later in Ireland and Portugal. This evaluation will focus on the amendment and its application during the 2010 Stand-By Arrangement (SBA) with Greece.

Evaluating the role of the framework for exceptional access is integral to understanding one of the most controversial aspects of the Greek program—the decision to proceed in 2010 and 2011 without restructuring privately held debt. Views remain deeply divided on whether that decision was a good one: some argue that it saved the euro area from intolerable stress, while others argue that, by delaying the resolution of a fundamental sustainability problem, it set Greece and possibly the euro area more broadly on a path of excessive uncertainty and escalating problems. The purpose of this chapter is not to judge that debate. Nor is the purpose to judge whether the IMF needs a constraining framework for exceptional access decisions.2 Rather the purpose is to consider whether the framework for exceptional access fulfilled its objective of facilitating a decision-making process in the IMF that is effective, clear, predictable, and protected from undue political influences. To this end the chapter examines two issues:

  • The thoroughness of the considerations that went into the amendment of the four criteria in the context of the approval of the 2010 SBA with Greece.

  • The thoroughness of the assessment of the four criteria during the 2010 SBA with Greece.

The remainder of this chapter is structured as follows. The second section briefly reviews the history of the framework for exceptional access. The third section examines the considerations behind the 2010 amendment to the framework and the process of gaining Executive Board approval. The fourth section considers the rigor of the application of the amended framework to decisions on the SBA for Greece. A final section states key findings and remaining issues.

The Framework for Exceptional Access—A Brief Summary of Its Inception and Evolution

The Fund’s policies on exceptional access evolved slowly and deliberately. With very few precedents of exceptional access, policies in the capital account crises of the 1990s (the 1994 Mexico crisis, Asian crises in the late 1990s, Russia and Brazil crises in 1998–99) drew on the “exceptional circumstances” clause formalized in 1983. That clause stated that access in excess of normal limits could be provided when a case was made that a member was experiencing “exceptional circumstances,” a term that “was left deliberately unspecified, reflecting what was seen as the inherent uncertainty of ‘exceptional circumstances’.”3

Eleven arrangements during 1995–2002 with countries experiencing “exceptional circumstances” focused attention on their terms and modalities. The Supplemental Reserve Facility (SRF) was put in place during the Korea crisis in 1997 as a channel for, in principle, unlimited exceptional access in capital account crises.4 Disbursements through the SRF were differentiated from those under pure SBAs or the Extended Fund Facility (EFF) by short repayment periods (reflecting the presumption that programs would restore investor confidence quickly and thereby catalyze private financing) and a surcharge on top of the General Resources Account (GRA) rate of charge. SRF resources were meant to support programs also backed by a SBA or EFF.5

The case for a formal framework (as opposed to a separate facility) for exceptional access emerged from a three-four-year period of reflection about the IMF’s role in the capital account crises of the late 1990s. Specifically, by 1999, concern had become strong about risks stemming from the absence of a clear and transparent framework for exceptional access. This concern surfaced in the IMF Annual Report for 1999 (IMF, 1999), a staff review (discussed in a Board seminar) of the experience during the Asian crisis (IMF, 2000a), and the Prague Framework for Private Sector Involvement endorsed by the International Monetary and Financial Committee at the 2000 Annual Meetings (IMF, 2000b). A dominant theme in these early documents was the need to delineate circumstances when it would be necessary to have private creditors bear some portion of the losses associated with the crisis—so-called private sector involvement (PSI). A chapter entitled “Strengthening the Architecture of the International Financial System” in the 1999 Annual Report captured the sense of this theme:

The effort to better involve the private sector in crisis prevention and resolution is seen as critical in bringing about a more orderly adjustment process, limiting moral hazard, strengthening market discipline, and helping emerging market borrowers protect themselves against volatility and contagion. . . (IMF, 1999, p. 47).

At the end of this multi-year period of reflection, the Executive Board discussed and then approved a framework to guide exceptional access decisions. The framework had both procedural and “substantive” components. On the former, several requirements for exceptional access were put in place: a process for early and regular consultation with the Executive Board on the progress in reaching agreement on a program; the presumption that staff reports for exceptional access arrangements would be published; an assessment of the risks to the Fund of high and/or concentrated exposure; and an ex post evaluation one year after the completion of exceptional access arrangements. These were all relatively uncontroversial. The substantive component—the four criteria for exceptional access—was the subject of several subsequent reviews, deliberations, and some minor revisions during 2002–09. Box 2.1 shows the Four Criteria as of end-2009.

Box 2.1.The Four Criteria for Exceptional Access as of End-20091

The following text is taken from IMF (2009b), pp. 30–31.

The Fund may approve access in excess of the limits set forth in this Decision in exceptional circumstances, provided the following four substantive criteria are met:

  • (a) The member is experiencing or has the potential to experience exceptional balance of payments pressures on the current account or the capital account, resulting in a need for Fund financing that cannot be met within the normal limits.

  • (b) A rigorous and systematic analysis indicates that there is a high probability that the member’s public debt is sustainable in the medium term. Debt sustainability for these purposes will be evaluated on a forward-looking basis and may take into account, inter alia, the intended restructuring of debt to restore sustainability. This criterion applies only to public (domestic and external) debt. However, the analysis of such public debt sustainability will incorporate any potential contingent liabilities of the government, including those potentially arising from private external indebtedness.

  • (c) The member has prospects of gaining or regaining access to private capital markets within the timeframe when Fund resources are outstanding.

  • (d) The policy program of the member provides a reasonably strong prospect of success, including not only the member’s adjustment plans but also its institutional and political capacity to deliver that adjustment.

1 This version of the four criteria reflects the cumulative changes to the criteria during 2003–09.

There were two recurring concerns in the establishment of the four criteria and in reviews during 2003–09. Though they took different directions, they both spun off the initial objective of putting in place a formal framework—to ensure that there is as clear an ex ante understanding as possible of when the private sector should assume a share of the burden of a country’s excessive accumulation of debt.

  • The first concern centered on how best to provide clarity, transparency, and predictability on the IMF’s exceptional access policy. This objective is a rather generic desideratum of most, if not all, IMF policies. But in the context of exceptional access, it is specifically seen as key to preventing the moral hazard that would exist if private creditors believe that they will in almost any circumstances be repaid and to reducing any tendency for the IMF to contribute to market uncertainty. More broadly, private creditors should be fully apprised that the IMF would not substitute official for private credit to heavily indebted countries if that involved imposing an excessive adjustment burden on the crisis country and worsening the chances for a successful recovery.

  • The second concern has been to protect the Fund’s decision-making process in exceptional access cases from undue political influence. Such protection is always seen as essential to ensuring uniformity of treatment among members. But a broader concern about political influence was also articulated. In the words of the 2002 staff paper proposing the four criteria, “the degree of discretion and flexibility in the [less formal pre-2002] framework may make the Fund more vulnerable to pressure to provide exceptional access even when prospects for success are quite poor and the debt burden of the sovereign is likely to be unsustainable.”6

Both of these objectives complement two other (and more generally applicable) features of IMF lending decisions: to safeguard IMF resources and to be aware of risks from excessive concentration of the IMF credits.

During some deliberations, questions arose about how risks of contagion during crises should influence exceptional access decisions. The 2002 staff paper proposing the four criteria directly addressed this question. It stated that “Regional and systemic implications [of severe debt crises] have often been cited as potential justification for exceptional access.” After a short discussion, the paper concluded that “it would be inappropriate to make the systemic criterion a necessary or a sufficient condition for providing exceptional access.”

Directors appear also to have taken a clear negative view on special consideration for contagion. The question about contagion during these deliberations was whether the risk of contagion should be a criterion in addition to the other four criteria. In other words, should exceptional access cases have to meet each of the four criteria and a fifth criterion that contagion was a significant risk (IMF, 2002)? The Public Information Notice (PIN) on the Board meetings establishing the four criteria states “a few Directors suggested further narrowing the definition of capital account crises that could warrant exceptional access by establishing a formal criterion relating to problems of contagion or the potential for systemic effects.”7 The PIN goes on to say:

Many other Directors, however, considered that such a criterion could create a bias toward higher access for larger members, which could not be reconciled with the principle of uniformity of treatment. Directors recognized that the Fund should be prepared to provide access above the normal limits in cases where the member’s problems have regional or systemic implications, when the other criteria are met [emphasis added]” (IMF, 2003b).

In short, the introduction of a specific consideration of contagion was rejected by both staff and Executive Directors.

Although the conclusion on contagion was well-reasoned, the term “contagion” was used rather loosely in the 2002 staff paper. No effort was made to define contagion, suggest ways to measure contagion, assess the implications of contagion or address whether there is a trade-off between actions that might be advocated to reduce contagion (including, for example, official bailing outs of the private sector) and actions that might most directly help the crisis country (for example, PSI).

In 2010, the Executive Board approved a major revision to the exceptional access framework. It introduced the option of granting an exemption to the second criterion (on debt sustainability) if it were judged that a debt restructuring needed to ensure a high probability of debt sustainability would have “adverse international spillover effects.” The following sentence was added to criterion 2:

However, in cases where there are significant uncertainties that make it difficult to state categorically that there is a high probability that the debt is sustainable over this period, exceptional access would be justified if there is a high risk of international systemic spillovers.8

Without formal discussion dedicated to this significant policy change, the Executive Board adopted this amendment as part of the approval of the SBA for Greece.

Despite concerns of some Executive Directors about the unusual process of changing this Fund policy, a long delay occurred before the Board revisited the issue. In 2013 in a Board seminar touched on the topic. A formal Board discussion that included issues related to the framework for exceptional access took place in 2014. In January 2016, the Executive Board approved an amendment to the four criteria.

Changing the Four Criteria: The Debate and the Process

The justification for seeking the 2010 amendment to the four criteria was that the criteria did not allow the optimal response to the circumstances in Greece and the euro area more broadly. This conclusion was not reached lightly. In contrast, the process of gaining Board approval of the change departed sharply from established practices.

The Debate Before Seeking Executive Board Approval

The amendment was made in the heat of the emergency of dealing with the Greek crisis. Many questions of substance about how to manage the crisis required rapid decisions simultaneously: how an IMF lending arrangement with the member of a currency union should be handled, how much fiscal adjustment could and should be undertaken, how to mesh the fiscal adjustment and structural reform components of the program, how to provide adequate financing, to name a few. With formal IMF involvement starting only in mid-April, 2010 and a deadline for decisions on the strategy of mid-May when a large debt-service payment was due, the time for debating any one of these issues was compressed.

Interviews of senior staff and management indicate that debate within those groups was fairly open and encompassed a reasonable spectrum of differences of view. It appears that positions ranged from the view that, in the absence of restructuring, debt was unsustainable (and therefore that the constraint imposed by the four criteria on proceeding without a restructuring was appropriate) to the view that debt would be sustainable if the right policies and sufficient financial support were put in place.

Those interviews also indicate that three considerations drove the decision on the way to characterize the sustainability of Greek debt.

  • The case for IMF involvement stemmed in large part from a conviction that the IMF was the best-equipped institution for the technical rigors of negotiating and monitoring the program. IMF participation was highly controversial among officials in Europe. By late March, however, the European debate had swung in favor, a position which, according to some interviewees, was supported by some large non-euro area members. IMF management was also eager to be involved.

  • Key European officials were resolutely opposed to debt restructuring by a euro area member. The commitment to this view was already deep by the time the IMF was formally invited into the inner circle of the crisis resolution process. While it is unclear whether an earlier IMF presence would have given the IMF a more influential voice on sustainability, it was clear that by April 2010, European opponents of restructuring had secured acceptance of their position. As one IMF staff member said on the issue of whether management could have insisted that Greece restructure, “the train had already left the station.”

  • The position within the IMF staff and management remained divided even after intense debate. The compromise position was that Greek debt would be deemed sustainable but not with a high probability. Interestingly, most interviewees were on either end of the sustainable/unsustainable spectrum. In other words, positions were rather polarized, and it is questionable whether many involved staff would have supported the formal compromise position that debt was sustainable but not with a high probability.

Given these three defining features, the only way to square the circle was to provide an exemption for Greece from at least the second criterion. In this context, the options were to make a special exception for Greece alone or to introduce a permanent exemption into the criteria. A one-off exception would have had the advantage of not changing, without due process, a Fund policy born of careful reflection on the IMF’s involvement in capital account crises since the mid-1990s. However, Legal Counsel was firmly of the view that the Executive Board did not have the authority to make a one-off exception to an approved Fund policy. Beyond that legal consideration, a permanent change was seen as having the advantage of even-handedness (vis-à-vis members that might one day seek exceptional access) and transparency.

In short, the decision to amend the four criteria was the result of debate within Fund staff and management. There is no evidence that any government officials played a direct role. In fact, many government officials interviewed were not fully aware of the framework for exceptional access, nor therefore, of the constraint it formally posed on IMF lending decisions. That said, the debate within staff and management about sustainability specifically was certainly influenced by the strong opposition especially of some European officials to restructuring. At this level, it is impossible to disentangle the possible roles in the formation of European views (and in turn pressure on staff and management) of hard analysis of sustainability and/or risks of contagion, pressure from private creditors, or personal self-interest.

The Role of the Executive Board

Executive Directors entered the meeting to approve the SBA for Greece with no preparation for addressing the proposal to amend the four criteria. As described above, the concept of introducing an exemption to the requirement of a high probability of debt sustainability had not been raised or discussed in any prior Board consideration of the exceptional access framework. Executive Directors interviewed had no recollection of any mention that Greek debt would not be considered sustainable with a high probability or that management would propose an amendment to the exceptional access framework. In short, decision to amend the four criteria was made without the usual consideration by the Executive Board of intended and unintended consequences that the vast majority of other IMF policies receive.

The staff paper for the Board meeting did not give any prominence to the proposal to amend the criteria. The paper stated staff’s final (compromise) position on the sustainability of Greek debt, but it did not provide any explanation of why a permanent change in Fund policy was effectively being put to the Board for approval. Rather, the proposal was embedded in the standard assessment of the exceptional access criteria. Specifically, the following three sentences appeared in the assessment of the criterion on debt sustainability.

On balance, staff considers debt to be sustainable over the medium term, but the significant uncertainties around this make it difficult to state categorically that this is the case with a high probability. Even so, Fund support at the proposed level is justified given the high risk of international systemic spillover effects. Going forward, such an approach to this aspect of the exceptional access policy would also be available in similar cases where systemic spillover risks are pronounced (IMF, 2010b).

Several Directors commented that staff’s position on the sustainability of debt was clear, but the necessity of a permanent policy change was not clear from the report. When a few Directors expressed a preference for making a one-off exception to the criteria for Greece, Legal Counsel clarified that a permanent amendment to the four criteria was necessary. Moreover, approval of the SBA carried with it approval of the amendment. In other words, votes on the two issues could not be separated. In the event, there were no votes against and no abstentions. In an interview, one Director stated that he felt “cornered” by the process.

Directors raised several questions at the Board meeting.

  • To proceed with the arrangement with Greece, was it necessary to make the systemic risk exemption a permanent change or could a once-off exception to the second criterion be made for Greece? Legal Counsel explained that the Executive Board does not have authority to make ad hoc exemptions to general Fund policy. The staff representative from the Strategy, Policy, and Review Department said that granting a once-off exemption would violate the IMF’s commitment to uniformity of treatment.

  • Were the four criteria meant to be simply guidance for Executive Board decisions, not an actual constraint? If so, the arrangement with Greece could be approved without a change in policy or specific exception made for Greece. Legal Counsel confirmed that the four criteria were an actual constraint.

  • If a permanent policy change were approved, one Director noted, and “spillover risks in a sense take priority over criterion 2, then we will in future cases have to have some assessment about the spillover risks.” He requested an update from the Director of the Monetary and Capital Markets Department (MCM) on “broader regional implications,” but no comment was provided.

Once it was understood that staff and management were proposing a permanent policy change, several Directors suggested that the Board return to the issue soon. Immediate follow-up would have moved the process closer to the normal standard of scrutiny (albeit almost without exception ex ante) that changes in Fund policy receive. However, no follow-up meeting was held. In fact, there was no mention of the issue in a staff paper on the closely related topic of the Fund’s mandate (IMF, 2010d) nor in the Board discussion of that paper the following month. Instead, appeal to the systemic exemption, without review, was again made in approving the arrangements with Ireland and Portugal and all subsequent reviews of the arrangements for Greece, Ireland, and Portugal.

Three years passed between the decision to amend the four criteria and the first discussion at the Executive Board of the amended policy. In mid-2013, a staff paper reviewing options for the framework was discussed in a Board seminar. A follow-up paper, sent to the Board and discussed in mid-2014, contained proposals for eliminating the systemic risk exemption and introducing other ways to deal with crisis countries where debt was assessed to be sustainable but not with a high probability. No Board decision was proposed. A third paper, sent to the Board in June 2015, built on the 2014 paper with farther-reaching proposals for addressing the risk of contagion when a country is assessed as having debt that is sustainable but not with a high probability. In January 2016, the Executive Board discussed the paper and approved the recommended elimination of the systemic spillover waiver. The resulting version of the four criteria is reproduced in an annex to this chapter.

Assessment of the Amended Four Criteria for Greece

Reviewing the analysis that supported approval of the SBA with Greece against the four criteria brings into focus many of the critical issues that were, and continue to be, debated with respect to the Fund’s role. Dominant among them are whether the framework for assessing debt sustainability was broad enough, whether the projections underpinning the Debt Sustainability Analysis (DSA) were rigorously constructed, whether prospects for market access were assessed against reasonable metrics, and whether prospects for contagion were adequately assessed.9This review identifies concerns about how each of these issues was handled.

Debt Sustainability, Market Access, and Program’s Prospects for Success

Ex post, Greek sovereign debt proved to be unsustainable. The baseline projection in the May 2010 staff report had gross public debt stabilizing at about 150 percent of GDP in 2012, beginning to fall in 2014, and reaching 120 percent of GDP in 2020. Market access was expected to resume in mid-2012. In fact, gross public debt reached 177 percent of GDP in 2014 (even after the unanticipated 2012 restructuring of privately held debt) and, in the October 2015 World Economic Outlook (WEO), was projected to rise above 200 percent of GDP in 2016 before starting to fall. Except for a brief period in 2014, Greece has not had market access. Forecasting errors (at times even startlingly large ones) are common in IMF-supported programs. They cannot be the standard for assessing a judgment at the inception of the program that public debt was sustainable. Rather, that standard should be the rigor of the analytics underlying the original judgment.

The assessment of public debt sustainability was based on a very narrow definition of sustainability in close to ideal circumstances. The central question in judging sustainability was whether in the medium term the baseline projection for the debt ratio would stabilize and then start on a downward trend. The baseline was constructed using assumptions of full implementation of the program policies (including large proceeds from privatization), realization of staff’s projections for the main endogenous variables—real GDP, the GDP deflator, government revenue and expenditure, and other influences on the primary government balance—and the opening of market access at “favorable terms” by mid-2012. The possibilities for worse-than-programmed policy implementation, progress toward market access, and developments in key macroeconomic variables were highlighted. But the staff report states that these risks prevented only the assessment of debt sustainability with a high probability, not the assessment that debt was sustainable.10 Questions of rigor, therefore, center on the definition of sustainability and the projections underlying the baseline DSA.

The definition of debt sustainability was narrow. Although a sine qua non of sustainability is the stabilization of the debt ratio, several other factors have important influences on debt sustainability: the level at which debt stabilizes and the known vulnerabilities associated with that level; the rollover rates in the country’s debt structure and, relatedly, the gross financing need; the analytical underpinnings of the assessment of when market access can plausibly be regained; and as criterion 4 notes, the prospects for success of the program “including not only the member’s adjustment plans but also its institutional and political capacity to deliver that adjustment.” The assessment with respect to market access and likelihood of success of the program were mere assertions without supporting argumentation.

A broader and more rigorous view of sustainability would have probed the overall characteristics of Greek debt, Greece’s financing needs, and specific risks to the outlook. Within the narrow definition of sustainability actually used, the DSA drove the assessment. The DSA conformed to the technical template provided in the then-prevailing staff guidance. It covered 2010–20 and included a baseline projection for the debt ratio, six shock scenarios, projections of gross financing needs, projections for the debt ratio with all variables set at their “historical averages” (no time frame is provided), and projections for the debt ratio in a no-policy change scenario (a primary surplus of 0.9 percent of GDP, although it is not clear how this surplus would have come about with no policy change). A debt stabilizing primary surplus (1.9 percent of GDP) was calculated for the period after 2020. At least two parts of this template—the projections based on historical averages and those based on no policy change—were meaningless for a country in a major debt crisis. They therefore at best cluttered, and at worst undermined, the credibility of the exercise.

The staff report repeatedly qualified the baseline scenario as subject to serious downside risks. As such, the baseline was by construction not a central scenario; it rather characterized an outcome if program execution was complete, and a V-shaped recovery took place. As early as the 2002 introduction of a DSA template, the likelihood that DSAs would have such a bias was recognized. As long as baseline scenarios assumed full implementation of agreed policies—rather than assigning some probability of full policy implementation—baseline scenarios would be biased. A further bias would result if projections for other macroeconomic variables that are influenced by any shortfalls in program implementation—for example, GDP growth, inflation, terms of market access—erred on the optimistic side.

Guidance notes for the DSA therefore saw the sensitivity analyses as a critical tool for evaluating the debt paths under less optimistic assumptions. Informally, this could be seen as establishing an upper path for the debt ratio which would capture a plausible range of worse outcomes, though not extreme tail events. Sensitivity tests for Greece fell far short of this standard: the shocks were relatively mild;11 no explanation was provided for how the shocks were chosen or therefore why they were considered to capture most adverse outcomes; and there was no effort to consider interactions or feedback loops among the shocks. The debt ratio stabilized or fell in all scenarios (albeit peaking anywhere from 155–180 percent of GDP) except in the combined adverse shocks scenario. There was no country-tailored sensitivity analysis even though earlier guidance papers had encouraged staff teams to devise them.

The conclusion that debt was sustainable had immediate credibility problems. Initially, skepticism took three broad tacks.

  • The first centered on doubts about the feasibility of the fiscal adjustment program. Such doubts implicitly concerned whether Greece met the fourth criterion. The program entailed an improvement in the primary fiscal balance by some 9½ percentage points of GDP over three years through a combination of revenue and expenditure measures.12 In addition, privatization revenues, though not large, were important to reduce reliance on debt-creating financing. To a large extent, the division within the staff between those who saw debt as sustainable and those that did not appears to have hinged on the credibility/sustainability of the fiscal program. Some outside bloggers who challenged the debt sustainability assessment also pointed to the credibility of the fiscal program.13 Several Executive Directors explicitly questioned whether the fiscal adjustment proposed was realistic from a social point of view.

  • The second concerned the assumption that Greece would regain market access at favorable interest rates by mid-2012. Staff consistently emphasized that the market access assumption was critical not only to financing the program (including repaying the Fund) but also to the resumption of growth. Yet at least three bloggers (Kirkegaard, 2010; Mussa, 2010; and Schadler, 2010) questioned its plausibility when even in the baseline scenario debt would have just peaked but not yet started to fall and would still be exceptionally high. Staff’s conclusion that market access would return in 2012 was based solely on the projection that the debt ratio would peak in that year.

    Relatedly, some bankers interviewed asserted that they had had severe doubts that Greece could meet its medium-term amortization schedule. These concerns were not discussed or recorded in public. However, that they were serious is confirmed by reports that at least two banks reached out to major European governments and senior management of the IMF with proposals for maturity extension and/or coupon write-downs. These efforts were undoubtedly self-serving, but they indicate skepticism that the program or its financing would be adequate to restore stability without a debt operation. Some bankers saw these early proposals as direct precursors to the aborted July 2011 debt-reprofiling agreement.

  • The third broad concern was whether Greece could avoid a severe and prolonged output contraction, especially without devaluation. Lachman (2010), for example, argued that without a devaluation, the path to stronger competitiveness and therefore a resumption of growth would be extremely difficult. He expected that Greece would eventually leave the euro area in order to devalue. This would be highly disruptive to the Greek economy in the short term and in that context threaten sustainability. Several Executive Directors strongly questioned the growth projections.

In the event, divergences from projections of activity and prices have been far larger than divergences from the fiscal projections. In terms of the pure mechanics of the path of the debt ratio, the continuing drop in GDP is by far the most important factor behind the massive overshoot relative to the initial DSA. A falling GDP deflator—reflecting an internal devaluation accomplished more through wage and price cuts than productivity increases—is a distant second factor. The fiscal adjustment has been only slightly less than originally agreed, although privatization revenues have disappointed and costs associated with a weaker than expected banking sector have been higher than anticipated. These facts suggest (with the benefit of hindsight) that the rigor of the DSA (and the conclusion that debt was sustainable) should be considered against the rigor of projections for GDP growth and the design of the related sensitivity analysis.

What can be said about the growth projections from the perspective of the data and information available at the time they were made? Two observations stand out.

First, though favorable precedents for the inputs to the projections existed, Greece combined the most difficult conditions of any recent capital account crisis countries. For example, the timing of the projected V-shaped recovery in real GDP starting three years after the previous peak (2008), was not out of line with other large capital account crises since 1995. The examples of Turkey in 2001 (which had a very large primary balance adjustment but devalued substantially) and Latvia in 2007–08 (which had maintained its fixed exchange rate but had relatively little fiscal adjustment) were held up by proponents of the projections. However, Greece faced an exceptionally large fiscal adjustment without a devaluation and in a weak external environment, so comparisons with these relatively favorable previous crisis outcomes were a false comfort. The credibility of the projections suffered from the absence of explicit accounting for the deep differences between the circumstances for Greece and those for other crisis countries.

Second, the case in the staff report for a sharp but short drop in real GDP is not well developed. The scant explanation of the basis for the GDP projection is not out of line with common practice in staff reports. Most reports refer only qualitatively to some influences on GDP growth and provide little analytical or quantitative detail. But the absence of analytical underpinnings proved a particularly serious problem for Greece: it weakened the quality of the projections at the time and also left staff’s projections open to serious criticism, as outside commentators focused increasingly on the massive forecasting gap.

Without transparency about the analytical framework in which GDP projections in particular were made, the Fund invited questions from many angles.

  • What was the assumed fiscal multiplier?14

  • Was account taken of a possible credit crunch?

  • Was a low Okun’s law assumption (relating employment growth to GDP growth) used?

  • Was account taken of likely impacts of an unusually severe crisis on conventional rules of thumb—such as fiscal multipliers and Okun’s law?

  • Was account taken of the lagged effects of a decade of falling competitiveness?

  • What was the basis for assumptions about the flexibility of domestic prices and (especially wage) costs?

  • Through what channels were the structural reforms expected to support growth and how quickly could they be expected to have an impact?

  • What were the projections for foreign demand?

None of these issues was addressed with any specificity or rigor in the staff report. A short paragraph in the May 2010 staff report mentioned some of them qualitatively. Specifically, it stated that the needed internal devaluation was likely to be a “long and painful process,” in a “relatively closed economy, the fiscal multipliers are bound to be large,” and that the “external environment is expected to remain weak.” But there was no indicative quantification provided and other factors were not mentioned. Growth, following a V-shaped pattern, was expected to return by 2012 on the basis of “confidence effects, regained market access, and comprehensive structural reforms.”

A second forecasting error with significant effects on the DSA was that for the GDP deflator. Fundamentally this came down to the fact that the internal devaluation that was achieved—though most data suggest it was less than originally planned—came about not through productivity increases but through falling wages, which in turn had a more-depressing-than-expected effect on the GDP deflator. The analytics behind projections of the GDP deflator are also scant.

In sum, Fund-wide standards for assessing debt sustainability and, accordingly, the actual assessment in May 2010 for Greece had serious shortcomings. The lack of specificity on the analytical underpinnings of staff projections for developments ranging from market access to GDP and prices adversely affected both the IMF’s strategy in Greece and its plausibility. Together these left a great deal of room for low contemporaneous credibility and ex post criticism.

Contagion: Was There Sufficient Analytical Evidence for Invoking the Exemption?

The 2010 amendment introduced contagion (or systemic spillover effects) into decisions on exceptional access.15 With no clear definition of contagion and no existing template or precedent for assessing it, staff analyses were in uncharted territory. Evaluating how thorough and convincing the assessments of contagion were must rely on the clarity of the framework created in real time and the transparency of the analysis.

As of May 2010, the analysis of contagion in the Greek crisis was developing, though it was to a large extent backward-looking.16 Several internal communications from staff to management (mostly from MCM, but one from an interdepartmental group) appear to have constituted the argumentation of the group within IMF staff and management that favored the European—and especially the European Central Bank—view that contagion was serious enough to trump concerns about insufficient confidence in debt sustainability in Greece. For the most part these characterized the possible channels of contagion and examined recent developments in Credit Default Swap (CDS) and Sovereign Swap Rates. Spillovers from sovereign spreads to banking sector spreads were prominently reported. The identification of channels of contagion analyses were thorough in the sense of calling attention to apparent channels for contagion, but they were mainly qualitative without indicating with much, if any, quantification which channels were likely to be important. Detail was greatest on the channels from Greek sovereign risk to the banking sectors in Greece and in other European countries. The main variables for which quantification was provided were various banks’ (domestic and foreign) exposure to Greek sovereign bonds and recent changes in banks’ CDS spreads, deposit outflows, and bank funding costs.

A more sophisticated analysis examined recent data and searched for changes in the degree of distress dependence. One internal memorandum in December 2009 reports computations of “conditional probabilities of distress” of European and Greek banks and a number of sovereigns in the event of distress (defined as a CDS event or an event that triggers activation of CDS) of the Greek government or major Greek banks.17 Staff describes the CDS event considered (a Greek government default on 80 percent of its foreign liabilities with a loss given default at 100 percent) as “very severe.” This is a forward-looking exercise in the sense that it is based on prices of financial assets that reflect market expectations of future developments.

The staff report for approval of the SBA for Greece has only a generic comment on contagion. This leaves unclear the extent to which the above-mentioned analyses fed into the design of the program and particularly to what degree there was coordination within the troika on a strategy to mitigate contagion. In a one-paragraph feature, the staff report states that “a worsening of the economic crisis in Greece could precipitate powerful spillovers to other countries.” It lists three channels of contagion: to sovereign debt and financial markets of other euro zone countries with relatively weak fiscal finances; to foreign financial institutions with substantial exposures to Greek paper; and to southeastern European economies (SEEs). There is neither quantification nor analysis to establish the potential importance of these channels. It is not even clear whether the passage relates to contagion from the adverse developments in Greece generally or from some form of possible default, restructuring, or exit from the euro area.

To the extent that analyses of contagion were forward-looking the event on which they focused differed. Most of the MCM analyses focused on a CDS or credit event that took the form, as noted above, of a severe default. A memo from the Research Department considered the implications, including for contagion, of a Greek exit from the euro area with a default. It is not apparent that any analysis considered an orderly restructuring of the type advocated by restructuring experts outside or inside the Fund. This apparent concentration on the extremes without recognition of more controlled modalities of PSI meant that the discussion of those extreme events might have crowded out discussion of orderly restructuring scenarios.

There is no written record to indicate that contagion counterfactuals were examined. The work described above presented evidence of correlations of spreads and conditional distress probabilities in the event of a credit event (variously defined). There was, however, no analysis of contagion that might stem from markets viewing large-scale official support as simply delaying restructuring (and raising the burden of a future restructuring on private creditors with long maturities). A question from a Board member during the May 9, 2010 Executive Board meeting crystallized the problem.

There is concern that default/restructuring is inevitable—even with the announcement of the program, bond spreads have risen. It is argued that trying to avoid default with the program simply increases the debt load and actually increases the probability of the default. On the other hand, it is argued that Greece is the sovereign version of Lehman Brothers and, therefore, it is advisable to put off restructuring for some time. We look forward to staff comments.18

In short, the staff report had not addressed the implicit question critical to the use of the newly-approved systemic risk exemption: even if one accepted that the risk of contagion in the event of a restructuring of Greek debt was substantial, was the counterfactual—proceeding aggressively (for example with an early restructuring of Greek debt)—likely to result in a better or worse outcome than a full bailout of creditors? The staff report did not address this issue, and the question raised at the Executive Board meeting was not answered.

Outside the IMF, some practitioners and academics proposed ways to restructure Greek debt adapting methods used in previous emerging market countries. The most widely circulated of these was Buchheit and Gulati (2010). While recognizing that restructuring in a currency union presented special challenges, they proposed concrete procedures based on their understanding of the legal structure of debt outstanding. They argued that the operation could be done within five-six months (“less if necessary”) with high creditor participation.

There appears to have been little if any dialogue within senior levels of the troika about options for restructuring and how disruptive they would be. Several interviewees expressed doubt about how well senior officials outside the IMF understood the possibilities for and technicalities of a debt restructuring in an emerging market or advanced country. In effect, the discussion of restructuring—of adhering to the need for a high probability of debt sustainability and not invoking the systemic risk waiver—was closed down in the lead-up to the May 2010 approval of the Greek program.

The overall picture of the contagion debate suggests a serious anomaly. As some officials persisted in the view that contagion from restructuring could be catastrophic, the market and especially the largest holders of Greek government bonds (GGBs) were actively discussing and developing proposals for a restructuring. These proposals almost definitely would not have been adequate to render debt sustainable (even with full program implementation), but they had two potential attributes: they would have been a basis for starting discussion on restructuring at an earlier point than actually occurred, and they might have prevented banks from reducing their positions in GGBs.

In sum, staff work on contagion made available to the IEO team was rather thin and did not address the counterfactual issues essential to assessing contagion. Moreover, accepting that restructuring would be excessively risky took the Fund out of potentially useful dialogue with restructuring experts and market participants.

2010–11 Reviews of the SBA: How Quickly Did the Fund Analysis Evolve?

The decision on whether to extend exceptional access to Greece in 2010 without a restructuring was by any standard extremely difficult. A major debt crisis (albeit in a small country) was roiling a relatively new and globally important currency area when the Lehman crisis was fresh in the memory of officials and markets. Time to assess the relative risks of differing strategies for dealing with the crisis was virtually nonexistent. Thus, at least as important as the rigor of the assessments behind the decision to invoke the systemic risk exemption in May 2010 was the rigor of continuing reviews. This section considers the reviews of the SBA (and by extension the ongoing assessment of debt sustainability) and the continuing reliance on the systemic risk exemption.

Developments in real GDP were the largest and most important divergence between the original program projections and actual outcomes. Notwithstanding considerable skepticism about the GDP projections among Executive Directors at the May 2010 Board meeting and, subsequently, growing doubts of outside commentators about growth prospects, the projected trajectory for GDP was revised in a substantive way only in the fifth review (December 2011), over 18 months into the arrangement. In interviews, staff involved noted that undertaking substantive revisions to the projections proved very difficult in the absence of a clear event or other decisive piece of news that would have necessitated a significant revision. Instead, disappointing news as well as historical data revisions dribbled in, and changes to the GDP projections were small.

There is no publicly available detailed reexamination during the actual course of the SBA of the underlying framework that informed the initial projections.19 For example, staff might have undertaken and reported on a deeper analysis of the fiscal multiplier, Okun’s law relationships, and speed of adjustment of prices, productivity, and wages. Such a reexamination would have addressed questions raised at the May 2010 Board meeting and growing public skepticism about the basis of the projections for GDP and the debt ratio. The absence of such an effort was reflected in the almost unchanged description of the GDP projections in the staff reports for the second–fourth reviews and the fact that cumulative real GDP growth between 2009 and 2020 was revised only by 0.7 percentage points between May 2010 and July 2011. In the fifth review (December 2011) the projection for cumulative growth between 2009 and 2020 was revised down by over 10 percentage points (Table 2.1). In other words, large revisions to the projections took place only after the decision within the troika to restructure debt. GDP deflator projections were also small but in the upward direction. This suggests that the deflationary effect of the program was still not internalized in the projections.

Table 2.1.Selected Macroeconomic Projections
2009 data and cumulative changes from 2009 to-date indicated at top, projection as of date in left column (in percent)
Real GDP
May 2010−2.046.5−5.5
December 2010−2.6−4.2−7.1−6.1
July 2011−2.0−4.4−8−7.5
December 2011−3.3−3.5−9−12
October 2015 WEO−4.4−5.4−13.9−17.8
GDP deflator
May 20100.
December 20101.534.55
July 20111.32.544.8
December 20112.
October 2015 WEO2.
Level projections as of date at left of sheet for date listed at top of sheet
Primary balance/GDP (in percent)
May 2010−8.6−2.4−0.91
December 2010−10.1−3.3−0.8
July 2011−10.3−4.9−0.81.5
December 2011−10.4−5−2.30.2
October 2015 WEO−10.3−5.2−3.0−1.4
Primary balance (euro billions)
May 2010−20.4−5.6−2.12.4
December 2010−23.7−7.6−1.8
July 2011−24.1−11.4−1.93.3
December 2011−24.1−11.4−4.90.4
October 2015 WEO−24.4−11.9−6.2−2.7
Notes:(i) 2009 column shows 2009 values.(ii) Euro stat fiscal and GDP data revisions completed November 2010.(iii) December 2010 SR says program primary balance/GDP for 2010 was −2.2.(iv) December 2010 SR says program primary balance for 2010 was –5.3.
Notes:(i) 2009 column shows 2009 values.(ii) Euro stat fiscal and GDP data revisions completed November 2010.(iii) December 2010 SR says program primary balance/GDP for 2010 was −2.2.(iv) December 2010 SR says program primary balance for 2010 was –5.3.

The October 2010 WEO (IMF, 2010e) included a special topic chapter on fiscal multipliers. This chapter was not focused on Greece, but the analysis was clearly relevant to the projections for Greece. Indeed, the multipliers assumed for Greece would eventually generate vigorous public controversy about the IMF’s role in Greece. Broadly, the WEO analysis can be summarized as follows: fiscal multipliers for advanced countries have historically averaged about 0.5; expansionary effects of a fiscal contraction occur mainly in the long term (and in the short term in only very specific circumstances); and fiscal multipliers are likely to be at least twice the historical average when interest rates cannot be lowered or interest rate cuts cannot be taken simultaneously by several countries (effectively preventing a nominal depreciation). Though a link to the program projections for GDP in Greece is not drawn, it would seem likely that had it been, the WEO analysis would have seriously challenged them.

Analysis of the risks of contagion was not carried out in any detail until the fourth review (July 2011). That review included a box entitled “Greece: Spillover and Contagion Risks,” which concluded: “The direct spillovers of a Greek debt operation can remain manageable, provided that necessary safeguards (liquidity and capital backstops) and an effective communication strategy are put in place.” The conclusion was supported by data on banking and trade links between Greece and SEE countries and on foreign bank holdings of GGBs. The box speculated that “risks could escalate dramatically under a poorly implemented debt operation without adequate safeguards or under a disorderly default scenario. These instances could threaten stability in the euro area with substantial spillovers to the global financial system.”

Analyses related indirectly, or to a lesser extent directly, to the questions of fiscal and debt sustainability and contagion took place outside of the program reviews. The Global Financial Stability Reports (GFSRs) in October 2010 (IMF, 2010f) and April 2011 (IMF, 2011a) had several special mentions of the four euro area countries with the highest fiscal and/or banking vulnerabilities (Greece, Ireland, Portugal, and Spain). Though these were mostly embedded in more general analyses of vulnerabilities in advanced countries, they introduced several recurring concerns directly or indirectly pertaining to Greece. The most prominent such themes were the scope for transmission of stress from sovereigns to banks (the strongest channel being from the Greek sovereign to domestic banks rather than foreign banks) with feedback loops to the fiscal accounts, rising bank holdings of sovereign debt, and rising interest bills relative to fiscal revenue. Between the April and October 2010 GFSR’s concern about contagion seems to have diminished slightly in large part because the escalation of sovereign swap spreads in the middle of 2010 had ceased at least partly as a result of European policy initiatives and national fiscal adjustments. However, the concern rose again in the April 2011 GFSR as many volatility and spread measures had worsened during the preceding six months. The section includes an explicit admonition for the European crisis facilities to lend on “sufficient scale and [with] flexibility, and should lend at interest rates low enough to support debt affordability, subject to strict conditionality.”20

In mid-2011, the IMF published, in its debut Euro Area Spillover Report (IMF, 2011b), its most trenchant examination of the risks and nature of contagion from a “credit event” in Greece, Ireland, or Portugal.21 The report has quantitative and qualitative assessments of the main channels and potential sizes of transmission of shocks from the three crisis countries to other euro area countries and to other economies. Three quantitative exercises examining financial sector spillovers (using financial market prices from 2007–11) dominate the analysis.

Broadly these exercises led to similar conclusions: if any increase in stress, including a “credit event” that triggered CDS contracts, were confined to the periphery, it would most likely have “modest spillover effects.” To the extent that stress were initiated in the core euro area (or presumably if periphery stress were to spread to the core euro area to a greater extent than the exercises indicated) the likelihood of contagion to non-euro area economies and financial systems would be far larger.

Despite the apparent fluctuation in concerns about contagion risks, staff assessments of the four criteria (when they were explicitly reexamined) were unchanged22: (i) public debt was assessed to be sustainable but not with a high probability; (ii) the systemic risk exemption was invoked; (iii) it was expected that Greece would regain market access within the period IMF resources were outstanding; and (iv) the program was expected to be successful. Perhaps the most puzzling part of the ongoing application of the four criteria to Greece is that the systemic risk waiver continued to be invoked in the fourth and fifth review of the SBA (July 2011 and December 2011, respectively) even though agreement had been reached that a restructuring of debt was needed. A possible explanation is that the exact terms of the rescheduling agreement and the extent of creditor participation were still uncertain, so invoking the exemption could have been seen as a precaution. This explanation, however, would not cover a larger puzzle: the systemic risk exemption was again invoked for the approval of the Extended Arrangement in March 2012 after the agreement on and high commitment of participation to a large rescheduling of private debt and two changes in the terms of the EU financial support that eased terms substantially.

In sum, refinement of the analysis and review of the strategy adopted in the heat of the outbreak of the crisis were slow and piecemeal after May 2010. In the first 14 months after approval of the Greek SBA (during which time four reviews took place), the analysis underlying staff projections was little deepened or adjusted. The assessment of contagion was elaborated somewhat, but it was not put in a counterfactual context. The staff’s assessment against the amended criteria accordingly remained unchanged.

Key Findings and Recommendations

The decision to amend the four criteria in order to extend exceptional access to Greece in 2010 was made in extremely difficult conditions; a large, imminent amortization payment threatened to lead Greece to default. The dominant European officials were adamant that an orderly restructuring could not take place in the context of euro area institutions at the time, while a decision had been made at the political level to involve the IMF in managing the Greek crisis. In these circumstances, the motivation that drove the decision to amend the criteria is clear.

The amendment, however, was not a small change and did not receive appropriate ex ante or ex post Board consideration. Rather it was a significant and substantive change to a policy framework that had resulted from careful deliberation and debate lasting for over a decade. An important strength of the IMF is that decisions of such import receive careful review so that intended and unintended consequences as well as implications for the future work of the IMF are clearly understood. Even if this process of deliberation could not be observed before the amendment decision was taken, it should have been undertaken as soon as possible afterwards (as some Directors requested at the May 2010 Board meeting). Instead, the first staff paper on the issue was circulated and discussed at an Executive Board seminar in mid-2013. In a revamping of the four criteria in 2016, the Executive Board eliminated the systemic risk exemption, but still retained an option for discussion in situations where debt is considered sustainable but not with a high probability.

The assessment of the (amended) four criteria in the approval of the 2010 SBA with Greece was not convincing. A large body of external commentary on the assessment questioned the conclusion that debt was sustainable and that market access could be restored within the period that IMF resources were outstanding. Fewer, but still a significant number, doubted that with reliance only on fiscal adjustment and structural reform, the program provided reasonably strong prospects of success that the growth projections were realistic. The low level of credibility of the projections harmed both the Fund’s reputation and any possible catalytic role that the Fund might hope to play. Whether this problem was the result of political influence (staff responding to the requirement to meet the amended criteria so that the Fund could participate in the lending arrangement with Greece) or true differences between the views of staff and those of outsiders can be debated, but probably not resolved. In any event, shortcomings were apparent in four areas:

  • The debt sustainability analysis. In the 2010 assessment, debt sustainability was equated with the stabilization of the ratio of debt to GDP, notwithstanding the fact that that ratio stabilized at a high 150 percent of GDP and rollover needs after the disbursement of IMF and European funds would remain exceptionally large. The baseline scenario—which many would expect to be a central scenario—was biased to the optimist side because any risk of incomplete policy implementation was precluded. The sensitivity analysis was not grounded in the types of uncertainties and shocks that a country in the midst of a major debt crisis was likely to face.

    A process of overhauling the DSA template started with a staff paper in mid-2011 and a more thorough staff paper in 2013. Many of the problems noted here have been addressed. Nevertheless, the DSA template should remain under close review, efforts to align the baseline with a central scenario should be deepened, and tailored shock scenarios should be actively encouraged.

  • Assessments of prospects for market access. The assessment in 2010 was essentially a statement of faith; it was assumed that as soon as the debt ratio reached the level at which staff projected it would stabilize, markets would reopen to Greece, even though debt would be some 150 percent of GDP. This assessment was possible because at the time there was virtually no agreed framework to provide guidance on how to assess prospects for market access.

    Staff guidance on indicators for assessing prospects for market access were issued in 2013. These have substantially improved the analytical basis for assessments. The approach to these assessments, however, is not based in firm theory. Refining the assessment of prospects for market access should be a continuing priority.

  • Contagion. According to all evidence provided to the IEO, both internal staff evaluations of contagion and staff work made available to the public during 2010 and the early months of 2011 were paltry. Much effort was devoted to describing the channels of contagion and reporting relationships between recent developments in prices of financial assets across countries. The more sophisticated of these examined cross probabilities of default with rather extreme assumptions on shocks. There is no written evidence that the relative risks, costs, and course of contagion in a restructuring scenario were compared to those in the no restructuring/ full bail-out scenario. In short, as mentioned in point 3, there was no clear analytical framework for judgments.

    The elimination of contagion as a consideration in the four criteria makes better measurements of contagion less urgent for the exceptional access framework per se. But the Fund would benefit more generally from work on measures of contagion in counterfactual conditions.

  • Macroeconomic projections. Staff reports for the initial program request and for subsequent reviews had scant elaboration of the underpinning of the central macroeconomic projections. This problem had the widest ramifications for the credibility of the IMF and the DSA when it came to projections for real GDP and the GDP deflator, both of which play a determining role in the DSA. Staff reports presented at best brief verbal comments on the projections and did not address the controversies that grew as time went by.

    Staff reports in general, but especially for requests for Fund assistance or reviews thereof, should provide a rigorous elaboration of the analytical underpinnings of the projections for key variables especially real GDP and GDP deflators. This is essential to ensure the analytical rigor of the projections and in turn their credibility.

The Fund was slow to revise its analysis and approach in subsequent reviews. Many of the shortcomings in the Fund’s initial assessment of sustainability and its components are understandable in view of the emergency conditions in which the program and projections were prepared. But lack of time cannot be used to explain why a reexamination of the macroeconomic projections did not start immediately after approval of the arrangement. Understandably, it is hard to change high profile projections which would require revision of a program that was just agreed. Moreover, news in the first few months of the arrangement was not unduly negative. However, the strong controversy surrounding the approval of the program—starting with several Executive Directors expressing doubts about the basis of the GDP projections, but extending to outside critics where commentary only became more negative—constituted a strong reason to review the projections in a fundamental way immediately.

Although early revision to program projections is difficult and costly the, IMF must have the resolve to undertake early and thorough revisions especially when the initial projections are made in haste, possibly under political pressure. It may therefore be desirable in exceptional access cases to require a more detailed and explicitly analytical review of exceptional access programs within six months of approval. This would include an exceptional review akin in terms of documentation, analysis, and projections to program approval with no presumption that changes must be small.

The disconnect between the perspective of several large banks pursuing reprofiling proposals on the one hand and the staff’s assessment that Greek debt was sustainable on the other is puzzling. First, it is difficult to understand how the Fund was taking the position that contagion from a restructuring was a major risk when banks were increasingly anticipating it. Second, communication between staff and management on the one hand and banks on the other appears to have been scanty. It may be that banks were highly focused on their communications with large euro area governments to the exclusion of the IMF. Yet, in view of the fact that significant numbers of Fund management and staff had come or were coming to the view that debt was not sustainable, it would seem that more intensive communication would have helped inform the Fund’s perspective on options for earlier rescheduling.

Annex 2.1. Four Criteria for Exceptional Access (Revised January 2016)1
  • (a) The member is experiencing or has the potential to experience exceptional balance of payments pressures on the current account or the capital account, resulting in a need for Fund financing that cannot be met within the normal limits;

  • (b) A rigorous and systematic analysis indicates that there is a high probability that the member’s public debt is sustainable in the medium term. Where the member’s debt is assessed to be unsustainable ex ante, exceptional access will only be made available where the financing being provided from sources other than the Fund restores debt sustainability with a high probability. Where the member’s debt is considered sustainable but not with a high probability, exceptional access would be justified if financing provided from sources other than the Fund, although it may not restore sustainability with high probability, improves debt sustainability and sufficiently enhances the safeguards for Fund resources. For purposes of this criterion, financing provided from sources other than the Fund may include, inter alia, financing obtained through any intended debt restructuring. This criterion applies only to public (domestic and external) debt. However, the analysis of such public debt sustainability will incorporate any potential relevant contingent liabilities of the government, including those potentially arising from private external indebtedness.

  • (c) The member has prospects of gaining or regaining access to private capital markets within a timeframe and on a scale that would enable the member to meet its obligations falling due to the Fund.

  • (d) The policy program of the member provides a reasonably strong prospect of success, including not only the member’s adjustment plans but also its institutional and political capacity to deliver that adjustment.


Access is termed “exceptional” when it exceeds the Fund’s normal access limits. Although exceptional access can occur in any circumstances—i.e., in traditional current account crises or in capital account crises—it is more typically necessary in capital account crises, the very definition of which is the loss of market access (usually when a government or banks have been borrowing heavily) and accordingly inability to roll over debt (which is usually large) coming due.

See Schadler (2013) for an in-depth review and assessment of the range of views on this topic.

IMF (2001), p. 54.

Although the notion of setting exceptional terms for exceptional access in capital account crises had been debated for several years, it is fair to say that the Asia crisis (and the Korea crisis in particular) accelerated consideration. The Executive Board approved the establishment of the SRF on December 17, 1997. The following day it was activated for the Korea SBA, which had been approved on December 4, 1997.

For a concise description of the SRF, see IMF (2003a). The SRF was modified on several occasions as experience was gained. It was eliminated in 2009 when the Fund’s “toolkit” was overhauled (IMF, 2009b).

Outside the IMF understanding of the four criteria, or even knowledge that the four criteria exist, is mostly confined to country officials that follow IMF developments very closely. Most market participants interviewed were unfamiliar with them. Nevertheless, most interviewees, whether inside or outside the IMF, were focused on the issues underlying the four criteria even if they were not explicitly knowledgeable about the framework.

The May 2010 staff report states, “on balance, staff considers debt to be sustainable over the medium term but the significant uncertainties … make it difficult to state categorically that this is the case with a high probability.”

The shocks comprised one positive shock (GDP growth higher than the baseline by 1 percentage point each year) and five negative shocks (GDP growth below the baseline by 1 percentage point each year, inflation below the baseline by 3 percentage points cumulatively during 2010–12, a permanent 200 basis points higher spread over German bund rates on new market debt, a once-off shortfall in the primary balance of 1 percent of GDP relative to the baseline, and realization of implicit or contingent claims of about 10 percent of GDP in 2010. A scenario showing combined adverse effects was also included.

A further increasing the primary deficit relative to GDP of 5 percentage points was projected for 2012–15.

For example, Wyplosz (2010) and Lachman (2010). Eichengreen (2010) also questioned whether Greek residents would tolerate severe fiscal adjustment when at least part of it was dictated by the need to repay foreign banks.

In interviews, staff stated that a multiplier of 0.5 was used in the baseline. This was reportedly the Organisation for Economic Co-operation and Development’s central estimate for the multiplier in its member countries. This information was not provided in the staff report.

The remainder of this chapter will use the terms “adverse systemic spillover effects” and “contagion” interchangeably.

Assessing this material was made difficult by the fact that too often terms are not defined, tables and charts are not fully labeled, and little effort is made to make the material accessible to non-technicians. It was also clear during interviews that some of the analyses were not well-understood by staff members who were not the ones actually carrying out the analysis.

Interdepartmental memo to the Managing Director and Deputy Managing Directors, December 18, 2009.

As recorded in the minutes of the meeting (IMF, 2010c).

The exceptional access framework mandated that an ex post evaluation be carried out within a year of the end of arrangements with exceptional access. A comprehensive evaluation was published in June 2013 for Greece.

Spillover reports, introduced in 2011, had their origins in the concerns about global imbalances earlier in the 2000s. They are issued once a year for each of the five systemically important countries/currency areas (China, euro area, Japan, the United Kingdom, and the United States).

The four criteria are supposed to be assessed in all staff reports initiating exceptional access arrangements and in all reviews. However, for Greece the criteria were not assessed in the second and third reviews. The first time they were assessed after May 2010 was in July 2011.

Reproduced from IMF (2016).

    Other Resources Citing This Publication