Chapter 5. The IMF’s Role in the Euro Area’s Crisis: What Are the Lessons from the IMF’s Participation in the Troika?

Moisés Schwartz, and Shinji Takagi
Published Date:
June 2017
  • ShareShare
Show Summary Details
G. Russell Kincaid


This chapter evaluates the ad hoc arrangement between the IMF, the European Commission (EC), and the European Central Bank (ECB)—known collectively as the troika1—that negotiated conditional economic policy programs and provided balance of payments financial assistance to four euro area countries during 2010–15: Greece (2010 and 2012), Ireland (2010), Portugal (2011), and Cyprus (2013). The terms of reference for the study exclude the joint programs with Greece (2012) and Cyprus (2013), which were still ongoing when the study was commissioned. Nor is Spain’s conditional financial assistance (2012) for bank recapitalization from the European Stability Mechanism (ESM) considered in any depth.2

The troika3 arrangement has raised several questions including, for example: Why was it created? What distinguishes it from other joint lending operations by the IMF? Did it operate efficiently? What governance issues does it pose for the IMF? Was the IMF a “junior partner” in the troika, as some have argued? This chapter attempts to provide preliminary answers to these questions.

The analysis focuses on the implications for the IMF of its participation in the troika. In each of the countries studied, this arrangement coordinated two conditional lending programs—one by the IMF and the other by the EC working in liaison with the ECB and on behalf of the Eurogroup. As discussed in the next section, the troika arrangement originated in coordinated conditional lending by the IMF and EU to Hungary, Latvia, and Romania (members of the EU though not the euro area) during 2008–09. Coordination is necessary because two simultaneous conditional lending programs need to be mutually consistent and coherent, given that the country authorities can only implement one set of economic policies. However, the two policy programs need not be, and were not, identical to each other in all aspects.

Coordination of conditional lending programs does not take place in a policy vacuum. Both the IMF and the EU institutions have mandates defined by their respective charters and elaborated by their respective policies and practices. The program country also has separate treaty obligations to the IMF and the EU institutions. A country’s treaty obligations to one institution prevail independently of any policy commitments it has made related to conditional borrowing from the other institution. Thus, for example, the Treaty on the Functioning of the European Union (TFEU) obliges EU members at all times to abide by the Stability and Growth Pact and Excessive Deficit Procedure as well as by provisions related to the single internal market (the most notable of which relate to state aid to the financial sector). Program conditionality that an individual EU member negotiates with the IMF must be consistent with its policy obligations to the EU, while conditionality that it negotiates with the EC must also be consistent with the TFEU and the decisions made under that treaty. Similarly, of course, all EU member states have treaty obligations under the Articles of Agreement of the IMF that they also should abide by when negotiating loan conditionality with the EC.

In these circumstances, a major challenge in studying the troika arrangement is to disentangle the implications for program (policy) design of the conditional loan coordination process from the underlying implications of programs countries’ membership in the euro area and EU. Because it may be impossible to adequately disentangle these two factors in the troika’s operations, it would be hazardous to apply lessons learned from the troika arrangement to other regional financing arrangements that are not currency unions. Further, if a troika-like arrangement were to be developed in other currency unions—the Central African Economic and Monetary Community (CEMAC), Eastern Caribbean Currency Union (ECCU), or West African Economic and Monetary Union (WAEMU)—it would not necessarily function in the same way as the troika has for the euro area.

Troika Origins

Preliminary Considerations in 1998

In September 1998, after the euro project was launched but before the currency union was created, the IMF Executive Board discussed modalities for conducting surveillance, and lending operations in euros, with euro area members based upon a trio of staff papers (see IMF, 1998a, 1998b, and 1998c). As regards IMF surveillance, euro area members were deemed subject to Article IV consultation for the economic and financial policies under their competency (e.g., national fiscal, financial, and structural policies), while discussions with the EC and ECB would need to take place pertaining to the economic and financial polices delegated to them (for example, respectively, area-wide fiscal and trade policies, and monetary and exchange rate polices). These EC and ECB discussions would constitute part of the Article IV consultations with individual euro area countries but, for practical reasons, they were considered best handled separately from the discussions with individual euro area countries. The Executive Board approved these surveillance modalities in December 1998.4

The Fund staff considered the use of Fund resources by a member of the euro area “extremely unlikely,” because this would signify that the union-wide financial system had become segmented (see IMF, 1998c). Such segmentation might arise if it were perceived that a currency union member might depart from the union, and thus an exchange risk could reappear. Even in the absence of exchange risk, lenders could possibly be deterred by country-specific risk, including macroeconomic risk (e.g., if a national recession endangered the financial viability of otherwise healthy companies); political risk; or risk arising from the insolvency of a national government. The staff thus considered movements in interest rate premiums or official accommodating transactions (such as ECB liquidity support—e.g., TARGET2)5 to be indicators of a euro area member’s need for balance of payments (BOP) support. Specifically, the staff expressed the view that “in the unlikely event that such risks assumed significant proportions, residents of a member of the European Economic and Monetary Union (EMU) could find themselves unable to borrow on suitable terms, as much as appropriate and necessary to avoid measures destructive of national or international prosperity.” In such circumstances, the Fund staff opined, a euro area member could request to use Fund resources just like any other Fund member.

While a euro member had the legal right to request the use of Fund resources, the IMF staff stated that, “it remains to be seen whether the EU would regard the use of Fund resources by EMU members as consistent with the ‘no bailout’ clause of the Maastricht Treaty.” In discussing the no bailout clause, Fund staff observed that “the EU, however, could provide exceptional financing under the terms of Article 103.a.2 of the Maastricht Treaty, which allows . . . financial assistance to a member state that ‘is in difficulties or is seriously threatened with severe difficulties caused by exceptional circumstances beyond its control.’” At that time, EU and euro members had no plans to operationalize this provision. Twelve years later, it would be called into use when a firewall was built in Europe to protect euro members from spillovers from Greece.

One staff paper (IMF, 1998c) dealt primarily with issues related to a balance of payments financing need by an EMU member and to BOP/reserve strength for inclusion of an EMU member into the Fund’s operational budget. Issues related to the use of Fund resources were seen to have arisen prior to 1998 with respect to other monetary unions (including the two CFA franc zones (14 countries), the ECCU (6 countries), and the Belgium–Luxembourg Economic Union). In particular, attention was drawn to whether the use of Fund resources by a member of a currency union financed a fiscal gap or a balance of payments gap. The staff concluded, based largely on experience with members of the CFA franc zones (IMF, 1995), that many “balance of payments needs have originated in fiscal needs,” especially in cases where the private sector’s access to international capital markets was limited or subject to sudden stops.

In the 1998 Board discussion on the implications of the EMU for the IMF, Executive Directors focused on surveillance modalities, ECB representation at the Board, and criteria for inclusion of an EMU member in the Fund’s operational budget. Issues related to the use of Fund resources and program design received scant attention, perhaps because Directors agreed with staff that that prospect was remote. But some of the contributions to the discussion are noteworthy. In particular, the U.K. Executive Director opined that “while a balance of payments need for an EMU member may seem an unlikely event . . . I agree that the Fund should be able to provide balance of payments assistance to EMU members in just the same way it provides financing to other members when they get into difficulties.” While echoing staff,6 the U.S. Executive Director noted that “in some circumstances, IMF conditionality associated with the use of Fund resources could involve measures that would conflict with the EMU objectives.” Continuing, the U.S. Executive Director wondered “whether it would be desirable to have an understanding with EMU participants whereby the ECB and/or other EMU members agree to provide euros to a member to enable it to fulfill its financial obligations to the Fund.” An Executive Director representing a group of developing countries felt that “the issue of conditionality for use of Fund resources in the case of euro members needs to be addressed.” In the concluding remarks, no mention was made of program design issues related to currency union membership, although Directors agreed to come back to the issue of identification of balance of payments need for members of a monetary union, notably the EMU. There was no direct or immediate follow up.7

An Embryonic Troika, 2008–09

The IMF had no experience in lending to euro or EU members between 1999, when the euro was adopted, and late 2008,8 when two EU, but not euro, members requested the use of Fund resources: Hungary for a Stand-By Arrangement (SBA) in November and Latvia for an SBA in December. The IMF had not lent to a EU member since the mid-1970s when it lent to Italy and the United Kingdom.9 As non-euro-area members of the EU, Hungary and Latvia were required under the Maastricht Treaty (Article 119 of the Treaty on the Functioning of the European Union (TFEU)) to consult with the EC and the Economic and Financial Committee (EFC) on their balance of payments needs before seeking conditional financial assistance from sources outside the EU. The Fund’s policy on exceptional access to Fund resources and its emergency financing procedures were called into play in both cases.

Hungary and Latvia also received medium-term balance of payments assistance from the EC based upon EU-determined policy conditionality. The EC had not lent to an EU member under its medium-term BOP assistance facility in 15 years (Greece in 1991 and Italy in 1993) and consequently it faced a very steep learning curve. Hungary was the first joint IMF-EC lending operation to an EU member; the ECB entered simultaneously into a repo facility in an amount of €5 billion with the Hungarian Central Bank. In Latvia, an ECB representative participated as an observer during the IMF and EC missions to that country, owing to Latvia’s membership in ERM2.

Romania’s request for financial assistance from EU and IMF in early 2009 prompted missions by staff from the EC and IMF, and the Fund provided assistance under its exceptional access policy. In the second half of 2009, the Polish authorities requested a precautionary flexible credit line (FCL), with exceptional Fund access. There are two apparent reasons why staff from the EC did not join IMF staff in Poland to discuss the use of EU financial assistance. One, the Polish authorities had not requested EU precautionary financial assistance, which did not exist at the time.10 Two, ambiguity existed about whether a request for a precautionary FCL, with its ex ante conditionality, would trigger the EU consultation clause. While the EU did not provide financial assistance, Fund staff reported that the Polish authorities had requested a swap facility with the ECB.

For the three countries with non-precautionary programs, the financing gaps (after any bank-exposure agreements) totaled almost €48 billion (Table 5.1). The IMF contributed 57 percent of this total, while the EU’s BOP assistance contributed 31 percent; the remainder came from the World Bank and the European Bank for Reconstruction and Development (EBRD). These averages mask substantial variation. The IMF covered 63–65 percent of the financing gaps for Hungary and Romania, and only 23 percent for Latvia, as examined below. No financing gap existed for the precautionary FCL with Poland.

Table 5.1.Financing Gaps and Official Contributions(In billions of euros; and in percent of totals)

World Bank.

Czech Republic, Denmark, Estonia, Finland, Norway, Poland, and Sweden.

World Bank and EBRD.

After bank exposure agreement reduced gross financing requirement by €7.5 billion.

After bank exposure agreement reduced gross financing requirement by €24 billion.

World Bank.

Czech Republic, Denmark, Estonia, Finland, Norway, Poland, and Sweden.

World Bank and EBRD.

After bank exposure agreement reduced gross financing requirement by €7.5 billion.

After bank exposure agreement reduced gross financing requirement by €24 billion.

Recognizing the unprecedented nature of IMF–EC cooperation, the IMF staff used a box in the 2008 staff report for Hungary’s SBA request to record five key cooperation principles: (i) early consultation and ongoing information exchanges during the program negotiations; (ii) contributions of both institutions to financing needs; (iii) joint announcement to underline broad support; (iv) consistency of program design and conditionality; and (v) consultation during the program monitoring process.

These principles borrowed implicitly from the Collaboration Concordat between the World Bank and the IMF (Box 5.1). At the Board meeting discussing Hungary’s SBA request, the ECB observer elaborated, saying that “given that the EU has its own policy and instrument framework, conditionality of the IMF has to be reflected or mapped onto our own requirements in terms of, for example, an update of the convergence program, in terms of the excessive deficit procedure and also in terms of the national reform program. This means that one would have two conditionalities running in parallel.” IMF–EC cooperation received scant attention in the interventions made by Executive Directors at this Board meeting.

The cooperation principles and the practices established with the experiences in the cases of Hungary, Latvia, and Romania laid the foundation for the troika arrangement with euro area members that followed in 2010.

Cooperation at the technical level between the staffs of the EC and IMF has been judged to be effective in ensuring consistency between the two programs’ conditionality, and having contributed to successful outcomes (see the ex post evaluations by IMF staff for these three countries and the detailed review contained in Annex 5.1 below). This cooperation drew extensively on the Fund’s cross-country experience and expertise in responding to financial crises, as well as on its ability to mobilize resources quickly in emergency situations. The EU’s assistance was embedded in the EU’s treaty-based policy framework, which provided a medium-term anchor to policies. But some challenges were identified. From the EU’s perspective, Fund staff did not sufficiently integrate the EU dimension, such as the EU’s surveillance framework—specifically the Stability and Growth Pact/Excessive Deficit Procedure (SGP/EDP) and competition policy/state-aid rules—into their analysis and operational procedures. As a consequence, frictions arose in all three countries over the operation of automatic fiscal stabilizers, which increased the overall fiscal deficit (above EDP targets) when real GDP turned out to be lower than projected.

Box 5.1.Bank-Fund Collaboration—Principles of the Concordat

When the World Bank provides quick-disbursing financial support in conjunction with IMF resources, their collaboration is guided by principles, agreed by the Bank President and IMF Managing Director, known as the Bank-Fund Concordat (Boughton, 2001). The Concordat was first articulated in 1989 and then was updated and reaffirmed in 1998. The Concordat attempted to identify areas of primary responsibility for each institution but increasing overlaps gave rise to legitimate difficulties/frictions and “made a strict delineation of responsibilities impractical and impossible to define properly” (Boughton, 2012).

Reflecting this delineation challenge, an updated Concordat sought to strengthen operational modalities and improve mechanisms to resolve disagreements. Procedures clarified modalities for exchange of information, including, inter alia, draft and final mission briefs, missions’ back-to-office reports, and technical assistance reports. Most major disagreements related to program design or its specific components were expected to be resolved at the staff level. When disagreements could not be so resolved, the issue was to be raised to more senior management, such as area department heads or country directors. At the overall institutional level, the focal point for collaboration was the SPR Director and the appropriate Bank counterpart. Regular, and as-needed, consultations were envisaged between the Managing Director and President as well as the Fund’s deputy managing directors and the Bank’s managing directors.

To avoid cross-conditionality—where a country’s failure to implement the lending conditions of one institution prevents the other institution from lending—each institution can proceed independently with its own financial assistance according to its own standards. In the latter circumstance, Bank/Fund management would present the case to an informal meeting of its Executive Board before proceeding.

The three central principles of Bank-Fund collaboration are:

Clarity for members. Countries in which both institutions are actively involved need to have a clear understanding of which institution has primary responsibility in any given area of policy advice and reform. When developing their policies and reform programs, countries should be able to draw upon the expertise of staff residing in both institutions under their respective mandates, and on other sources.

Full consultation between Bank and Fund. Before finalizing its position on key elements of a country’s policies and reform agenda, each institution will solicit the views of the other and share its evolving thinking at as early a stage as feasible. This should lead to better policy advice and program design benefiting from the perspectives of both institutions. When there are differences of view between the two institutions about policy and priorities in countries where both are involved, and the disagreement cannot be resolved at the staff level, the issue will be raised at the level of senior management for resolution. If agreement still cannot be reached, the views of the institution with primary responsibility will prevail in the final advice to, or negotiations with, a member country and such differences will be reflected in reports on the country to the Executive Boards of the two institutions.

Each institution retains separate accountability for its lending decisions. Programs supported by the Bank and Fund should be complementary and part of an overall reform agenda owned by the member country. The Executive Board of each institution will be made aware of the total package and of how the components covered by one institution complement the parts supported by the other. At the same time, each institution must proceed with its own financial assistance according to the standards laid down in its Articles of Agreement and the policies adopted by its Executive Board. Any difference of view between the two institutions will be reported to the Boards when approval to support a program is sought.”

Source: Report of the [IMF] Managing Director and the [World Bank] President on Bank-Fund Collaboration, SM/98/226, Revision 1, September 25, 1998.

In the Latvian case, the Fund’s relatively small financing contribution of 23 percent (nevertheless at 1,200 percent of Latvia’s IMF quota)—stemmed from the Latvian authorities’ strong desire to maintain a currency peg that the IMF staff considered significantly overvalued and did not feel comfortable supporting with the use of Fund resources. This disagreement over exchange rate policy had been simmering for some time. Indeed, owing to the inability to implement the 2007 bilateral surveillance decision in this case, the Fund’s Executive Board had not completed two annual Article IV consultations (2008 and 2007) with Latvia.11 The EU, by contrast, supported the Latvian authorities’ preference to maintain the euro peg under ERM2, and provided the requisite external and budget financing.12 The IMF and EC resumed their more usual financing pattern and policy roles in the case of Romania, which had a floating exchange rate and independent monetary policy.

The EU and IMF completed their program reviews and disbursements in tandem except in one instance. During a joint mission for the first review (May 2009) for the IMF and EU programs in Latvia, output estimates for 2009 foresaw a real contraction of 18 percent rather than the 5 percent that had been projected. Without new measures, the fiscal deficit was expected to widen to 16–17 percent of GDP, far exceeding the program target of 5 percent of GDP. A supplementary budget was adopted on June 16, 2009 with the full-year-equivalent of 13 percentage points of GDP and containing measures that included cuts in pensions and social benefits. The IMF mission—concerned about the adverse growth implications and the effect of these measures on vulnerable groups—returned to Washington to consult. The European Council, worried that the currency peg might unhinge without financial support, on June 19 “strongly supported the intention of the Commission to propose the swift disbursement of the Community’s balance-of-payments assistance….” In mid-July, an IMF team returned to Latvia to hold discussions to complete the first IMF program review. The subsequent staff report made clear that the IMF staff had serious reservations about the rapid fiscal adjustment endorsed by the EU Economic and Financial Affairs Council (ECOFIN). But, as it turned out, bold upfront fiscal adjustment coupled with strong country ownership helped produce an expectations-induced V-shaped recovery in 2010 and thereafter. The EC had seemingly made the right judgment call about the currency peg sustainability, feasible fiscal adjustment, and a small fiscal multiplier.

In September 2009, the Fund staff issued a Board paper (IMF, 2009d) that analyzed 18 crisis programs approved between September 2008 and July 2009, including the 4 IMF-supported programs with EU members. Three main conclusions were drawn. One, compared to previous capital account cases, the 18 programs involved less compression of domestic demand. Two, external adjustment in these programs was less wrenching than in past crises owing to more timely, greater, and more front-loaded financing and supportive macroeconomic policies. Three, banking crises were generally avoided, which was considered remarkable given the externally financed credit booms that had been taking place, especially in Central and Eastern Europe. In this context, the maintenance of private sector exposure through the Bank Coordination Initiative was cited as mitigating the potential effects of deleveraging. Financing support from the EC was found to enable risk sharing, but not to have produced less external adjustment. Indeed, based on the Fund staff’s statistical analysis, Latvia’s adjustment effort was significantly above the predicted level (based upon initial conditions), while no adjustment in the EU program countries was significantly smaller than predicted. No issues were raised concerning EC-IMF cooperation, suggesting that the staff had identified no problems worthy of Board discussion.13, 14 The Board paper did however refer to EU constraints pertaining to program design—noting that the SGP/EDP prevented Latvia from immediate adoption of the euro and limited Hungary’s scope for discretionary fiscal loosening. The paper did not examine the implications of EU/euro membership for program design and financing, leaving questions unanswered from the 1998 Board discussion.

In discussing this 2009 paper (see Minutes of Executive Board Seminar 09/6-1), IMF Executive Directors from EU countries expressed views on EC-IMF cooperation. In particular, the German Director commented that “staff tend to assess the EU policy framework as a constraint to defining an adequate policy response . . . it should be underlined that the governance framework of the European Monetary Union, which includes the Stability and Growth Pact and the exchange rate system ERMII, is an integral part of the institutional and legal framework in which each EU member state operates. Therefore, it should be fully respected in the design and monitoring of IMF program conditionality.” These remarks were endorsed by several other Directors from EU countries. No non-European Director spoke on this issue. The concluding remarks by the Chairman of this Board discussion mildly observed that, “Directors highlighted the importance of close cooperation in the design of financing packages with other bilateral and multilateral creditors, notably the European Union.”

The Troika Emerges with the Euro Crisis

In October 2009, the newly elected Greek government revealed that the fiscal deficit for 2008 had been misreported as 5 percent of GDP and was in fact 7¾ percent of GDP, while the projected fiscal deficit for 2009 was 12½ percent of GDP, rather than the 3¾ percent previously projected. Yields on Greek government bonds rose sharply as demand for them fell, limiting the government’s access to private market funds. The new Prime Minister telephoned the IMF Managing Director to seek help from the IMF.15 Euro members were not eligible for the EU’s Medium-Term Balance of Payments Facility that had been accessed by Hungary, Latvia, and Romania, and no mechanism specific to euro members existed for balance of payments assistance. This absence was intentional; it was consistent with the “no bailout” provision of the Maastricht Treaty (Article 125 of TFEU),16 which in turn was reinforced by the provision that no monetary financing of budgets would be provided by the ECB or national central banks (Article 123 of TFEU).17 Even if a request for Fund resources was admissible from an IMF standpoint, Greece—as a euro area member—still needed to consult with its European partners before making such an unprecedented request.

During the remainder of 2009, the Greek authorities worked with the EC to devise a stability program for 2010–12, which aimed to reduce the fiscal deficit to below 3 percent of GDP by 2012. The IMF staff provided technical assistance to the Greek government during this period. The Greek-EC stability program relied entirely on financing from private markets. The ECOFIN Council accepted the program in February 2010. Both the EC Commissioner and President voiced confidence that a European-only (no IMF financing) solution would be sufficient, and the ECB President publicly expressed opposition to IMF financial assistance for a euro member. In addition, during this period to February 2010, the German and French finance ministers made public statements excluding a financing role for the IMF in Greece (Bastasin, 2012).

During February–March 2010, it became increasingly clear that private financial markets would not provide the Greek government with the requisite funds on acceptable terms. Official financial resources would therefore be needed to prevent a default by Greece, which had large repayments coming due beginning in May 2010. The Europeans debated whether to have the IMF involved in Greece and, if so, how. According to Bastasin (2012) and senior euro area officials who were interviewed by the IEO, the German and some other euro area governments wanted the IMF to be directly involved in any European lending operation to Greece, desiring to benefit from the Fund’s technical expertise and experience in crisis management. Other euro area governments opposed IMF involvement, wishing to keep the resolution solely in European hands. The IMF’s possible financial contribution seems not to have played a significant role in these discussions. It was the Eurogroup that induced the ECB, whose independence and credibility was respected by governments and the European public, to be a troika partner. During this period according to interviews, the U.S. government in its contacts with European governments urged IMF involvement in Greece.

With Greek default looming, the heads of state and government of the euro area announced on March 25, 2010 that “As part of a package involving substantial International Monetary Fund financing and a majority of European financing, Euro area member states are ready to contribute coordinated bilateral loans. This mechanism, complementing IMF financing, has to be considered ultima ratio, meaning in particular that market financing is insufficient. Any disbursement on the bilateral loans would be decided by the euro area member states by unanimity subject to strong conditionality and based on an assessment by the European Commission and the European Central Bank.” At a related informal Board meeting on Greece, the nature of IMF engagement was not discussed; perhaps because Executive Directors were told that Greece did not expect to use this new mechanism and that staff had not been asked to discuss a program. The IMF did not issue a press release in response to this announcement either. Nonetheless, the envisaged IMF involvement seems to have been modeled on the IMF-EC lending to EU countries that had taken place in 2008–09.18 In early April 2010, the Executive Director representing the EU informed his colleagues that the same close collaboration that had been employed in Hungary, Latvia, and Romania was the best approach were it needed. Staff and management did not confirm or elaborate.

The Eurogroup announced on April 11, 2010 that its members had reached agreement on the “practical arrangements, notably financial, of the mechanism for financial support.” The meaning of “substantial IMF financing” was not given greater specificity by the Eurogroup nor defined by the IMF during this period. The same day (April 11), the Managing Director issued a press release stating, “an IMF team will hold discussions in Brussels on April 12 with the Greek authorities, the European Commission, and the ECB.” April 12, 2010 therefore saw the first meeting of the troika with Greek authorities, albeit it was not termed a negotiation. The Managing Director did tell Executive Directors that IMF staff were not going to share information with the European Commission without giving it to our member [Greece], too. It was only on April 15, 2010 that the Managing Director issued a statement stating that the Greek authorities had requested a Fund-supported program.

Staff prepared an initial concise note under the exceptional access policy, which was circulated to, and discussed by, Executive Directors on April 16, 2010. This note did not quantify the financing requirement, possible IMF access, or prospective European financial support. It did not discuss operational modalities for IMF engagement with the EC and ECB. All four criteria under the exceptional access policy were observed preliminarily. At the informal Board meeting, staff did not provide any additional quantification related to program financing. On April 29, 2010, the IMF mission chief for Greece told Executive Directors that his team was still looking at the external financing need and he couldn’t give numbers right now as it was too early. On May 2, Executive Directors were told that a staff-level agreement on a program had been reached and that a Eurogroup meeting was convening at the same time in Brussels. Later that same day, press releases from the EU and IMF formally announced the programs with Greece.

The IEO has found no evidence that Fund management and staff attempted to define the nature of the IMF’s possible involvement with Greece and the euro area, or to discuss the related issues with the Executive Board. In particular, such a discussion could have focused on the implications for program design and for financing of a request by Greece—a member of the euro currency union—to have a Fund-supported program.

Other options (than the troika with parallel conditional lending by the euro area) could have been considered for assisting Greece, though each might have had its own drawbacks. For example, the IMF could have been made solely responsible for program design and financing. This would not have altered Greece’s economic policy obligations with respect to the EC and ECB that stemmed from its currency union membership; these obligations potentially constrained the scope for Greek policy actions and therefore potentially affected the policy design of any IMF-supported program. Arguably, the EC might have had less influence on program design had it not provided financial assistance, although euro members carry considerable weight at the IMF and could have made their views known via their Executive Directors. To finance the entire Greek program, the IMF would have halved its ability to lend to other members or its forward-commitment capacity (FCC).19 The subsequent programs with Ireland and Portugal would have more than exhausted the FCC, requiring the IMF to borrow from official sources—such as in the euro area—the necessary resources. The IMF’s exposure to Greek credit risk would have been higher under this scenario than it actually was, while the credit exposure of euro area governments to Greece would have been correspondingly lower. Essentially, a risk transfer from the euro area to the entire IMF membership would have taken place.20

Alternatively, the IMF could have financed a significantly smaller share of Greece’s financing gap, even to the point of avoiding the need to trigger the exceptional access policy.21 The IMF would have still provided its program-design expertise and crisis-management experience (which were the chief reasons given for IMF involvement). It would still have needed to adhere to its policies and practices; as the Director of the Strategy, Policy, and Review (SPR) Department observed in the Latvia context, the IMF “cannot delegate responsibility for use of Fund resources. This applies whether we put in one cent or the entire financing of the program” (Blustein, 2015b). Nevertheless, the question naturally arises whether with “less, or no, skin in the game,”22 IMF staff would have had less influence over program design with the country authorities or with the EU institutions. In considering this possible money–influence tradeoff, it must be recognized that the IMF would still have put its reputation at risk.

Of course, other possible modalities for IMF involvement exist. The point here is not to be exhaustive or to judge what would have been the best option, but to show that a range of options was available in early 2010 that could have been considered by the Executive Board.

The Troika in Action

Follow the Money

Four countries in the euro area—Greece, Ireland, Portugal, and Cyprus—used Fund resources during 2010–15 (with only Cyprus not triggering the exceptional access policy), while simultaneously receiving financial support via one or more EU/euro financing mechanisms created for this purpose. In the 2010–15 programs the financing gaps typically included costs related to bank recapitalization, but, unlike in the earlier IMF-EU joint programs, foreign banks provided no maintenance-of-exposure agreements to reduce the program financing requirements. Burden-sharing contributions between the IMF and EU/euro area for these four country cases are set out in Table 5.2. In Greece (May 2010), the IMF covered 27 percent of the identified financing gap, or a somewhat larger share than in Latvia but a considerably smaller one than in Hungary and Romania.

Table 5.2.Financing Gaps and Official Funding for Euro Area Programs(In billions of euros, and in percent of total)
Memorandum items:
Greece II28.0143.6171.65

Includes €10 billion for a Financial Stability Fund.

Excludes Irish authorities’ contribution of €17.5 billion (or 20.6 percent of Ireland’s financing gap) from their cash reserves and liquid assets.

Includes €17.5 billion for bank recapitalization provided by Irish authorities per footnote 2.

Includes €12.0 billion for a Bank Solvency Support Facility.

This total includes €50 billion for bank recapitalization and about €50 billion to finance credit enhancements for the debt reduction with the private sector and to finance a debt buyback program. This total was reduced owing to a projected €50 billion in privatization receipts.

Includes €10 billion for a Financial Stability Fund.

Excludes Irish authorities’ contribution of €17.5 billion (or 20.6 percent of Ireland’s financing gap) from their cash reserves and liquid assets.

Includes €17.5 billion for bank recapitalization provided by Irish authorities per footnote 2.

Includes €12.0 billion for a Bank Solvency Support Facility.

This total includes €50 billion for bank recapitalization and about €50 billion to finance credit enhancements for the debt reduction with the private sector and to finance a debt buyback program. This total was reduced owing to a projected €50 billion in privatization receipts.

As the IMF Board approved the IMF-supported program with Greece (on Sunday, May 9, 2010), the EU Council was completing the design of a European firewall. The IMF Managing Director attended the EU Council discussions to encourage the creation of a firewall. On Monday morning, May 10, 2010, the EU Council announced new mechanisms23 totaling the equivalent of €750 billion, of which the IMF’s contribution was expected to be €250 billion, or one-third. The EU statement went on to say that “the IMF will participate in financing arrangements and is expected to provide at least half as much as the EU contribution through its usual facilities in line with recent European programmes.” (In fact, the average share of IMF financing in the three programs with EU members was considerably larger, or nearly double that of the EU financing (Table 5.1).)

The Managing Director welcomed immediately these European actions, noting that the IMF contribution would be made “on a country-by-country basis” and quietly walking away from the headline figure of €250 billion, while at the same time endorsing the contribution ratio.24 The First Deputy Managing Director clarified to the press in Washington on Monday, May 10, 2010 that the IMF had not “earmarked” any money for the euro area and that these announced figures were “illustrative” or “hypothetical.” In the first two subsequent programs (Ireland, December 2010 and Portugal, June 2011), the IMF covered one-third of the financing gap as expected, providing half the amount that was contributed by the euro area (Table 5.2).

In both the cases of Ireland and Portugal like that of Greece, the initial concise note under the exceptional access policy circulated to Executive Directors did not contain quantified estimates of the financing gap, IMF access, or European financing support, although the respective policy notes sent to management prior to the circulation of these concise notes contained such quantification (De Las Casas, 2017). At the informal Board meeting on Ireland (November 23, 2010), Fund staff was asked by Executive Directors the size of the financing package and of the total EFF access. Staff did not provide the requested quantification explaining that those numbers haven’t been finalized as yet. However, staff had already provided to IMF management preliminary quantification. Quantified estimates were provided to Executive Directors on November 28, 2010 just before the announcement later that day on a staff-level agreement. At the informal Board meeting on Portugal under the exceptional access policy (April 19, 2011), Fund staff told Executive Directors responding to questions on the size of the program, that it was too early to say. However, preliminary estimates had already been provided to IMF management. On May 2, 2011, staff informed Executive Directors that agreement on a program with the Portuguese authorities would probably be reached in the coming days and provided a quantified estimate of the still preliminary financing gap. The Portuguese Prime Minister announced agreement on a EU-IMF program the next day (May 3), although EU-IMF announcements took place only on May 5.

This two-to-one ratio did not last long as an “illustrative” benchmark; by 2012, it was gone. In particular, the IMF’s share in the total financing package for the second arrangement with Greece (March 2012) fell to 16 percent, and Spain received financial assistance for its banking sector from the European Financial Stability Fund (June 2012) without parallel use of IMF resources. In 2013 for Cyprus, the IMF’s share of the total financing package was only 10 percent. Cyprus’s access to Fund resources was 563 percent of quota, or below the 600-percent-of-quota threshold for obtaining exceptional access.

The European Stability Mechanism (ESM)—the permanent financing mechanism established by the euro area for its members—was created via an intergovernmental treaty among euro members. The Treaty came under immediate legal challenges within the EU but survived them (Box 5.2 and, for details, Schneider, 2013 and Van Malleghem, 2013). For our purposes, the most notable outcome of the Treaty-ratification process and subsequent legal decisions was the continuing role given to various national parliaments. In particular, the German Federal Constitutional Court ruled that ESM packages must be clearly defined and that the German Parliament must be given the opportunity to review the aid and stop it if needed. This parliamentary check was considered necessary to retain Germany’s sovereignty over its national budget—sovereignty that the Court saw as a “fundamental element” of the democratic process. Six other euro area parliaments have similar roles.

The ESM Treaty requires unanimity among its (19) members to enable the provision of ESM financial support and to empower the EC to negotiate the associated economic policy conditionality. Unanimity voting typically grants more influence to members with smaller voting weight/size, although these members may thus come under considerable peer pressure to join the majority. The 19 finance ministers of the euro area countries comprise the ESM Board of Governors. Their voting is constrained by national laws in some cases: in seven countries, notably Germany, the Finance Minister must obtain the consent of the national parliament before voting at the ESM. In considering how to vote, national parliaments may look to other actors, including the ECB and IMF, for assessments. National parliaments of ESM members are typically informed of the status of an IMF program. Many of the European officials who were interviewed by the IEO cited this prospective need for parliamentary action by some euro members, especially Germany, as a factor in giving the IMF a perceived “veto power” within the troika.

Box 5.2.EU/Euro Balance of Payments Financing Mechanisms for Euro Members

EU/euro countries financed adjustment programs for euro area members via four different modalities. The first three of these were announced in May 2010, consisting of the Greek Loan Facility (GLF), the European Financial Stabilization Mechanism (EFSM), and the European Financial Stability Facility (EFSF). The GLF had resources of €80 billion composed of bilateral loans from 14 euro members. The European Commission administered the GLF, disbursing funds based on decisions taken by the Eurogroup, which evaluated compliance under the EC’s MOU as assessed by the EC and ECB, and reviewed findings by the IMF. The GLF was intended as a temporary country-specific response.

The EFSM (€60 billion) and the EFSF (€440 billion) formed the European “firewall” of €500 billon that was expected by the Eurogroup to be supplemented by IMF financing arrangements equivalent to half of the EFSM/EFSF contributions. The EFSM was intended to safeguard EU financial stability “under current exceptional circumstances” (such as the problems of Greece), essentially replicating for all EU members the medium-term BOP financing facility available only to non-euro EU members. The EFSM operated within the Treaty on the Functioning of the European Union (TFEU), and its borrowing in international capital markets was backed by EU budget guarantees. The EFSM had its legal basis in Article 122 (TFEU), which allows the EU to provide a euro member with financial assistance “where a Member State is in difficulties or is seriously threatened with severe difficulties caused by natural disasters, or exceptional occurrences beyond its control….” This Article is an escape clause to the no bailout provisions of the Maastricht Treaty. Decisions to approve a loan, or to disburse tranches, are taken by a qualified majority of the European Council. The EFSM lent to Ireland and Portugal, totaling €46.8 billion at end June 2015, or 78 percent of the EFSM’s total lending capacity, and all of its lending as of that date.

In July 2015, the EFSM provided “bridge financing” (€7.2 billion) to Greece for a three-month period “in view of the severe economic and financial disturbances caused by exceptional circumstances beyond the control of the [Greek] Government” and “to avoid further default on its repayment obligations.” This short-term EFSM loan was repaid by a disbursement from a new loan from the European Stability Mechanism (ESM)—see below for details.

In June 2010 euro area governments agreed to establish the EFSF as a temporary crisis mechanism for euro members and as a private company under Luxembourg law. The EFSF disbursed €185.5 billion to Greece, Ireland, and Portugal as of June 2015. It borrows on international capital markets and euro area governments guarantee its debt. The EFSF has the same credit standing as any other sovereign claimant (that is, pari passu); it is not a preferred creditor. The EC was mandated to negotiate the policy conditionality, in consultation with the IMF and ECB.

Legal complaints were filed against Germany’s participation in these rescue efforts. While the German Federal Constitutional Court rejected these complaints in September 2011, the Court ruled that aid packages cannot be automatic and may not infringe on the decision-making rights of Parliament. Thus, the German Parliament must be given the opportunity to review the aid and stop it if needed.

In December 2010, euro area governments decided to establish the ESM as a permanent body to replace the EFSF with an effective lending capacity of €500 billion. To implement, the European Council amended in March 2011 the TFEU, adding: “The Member States whose currency is the euro may establish a stability mechanism to be activated if indispensable to safeguard the stability of the euro area as a whole. The granting of any required financial assistance under the mechanism will be made to strict conditionality.” The legality of this new TFEU provision—including its consistency with the no bailout provisions—was challenged. The European Court of Justice rejected this complaint in October 2012 (ECJ, 2012). The ESM Treaty was signed in February 2012. The ESM began operation in October 2012 and new requests for financial assistance by euro members have been directed to the ESM since July 2013. The ESM Treaty accepts that the IMF has preferred-creditor status over the ESM. The German Federal Constitutional Court and the European Court of Justice upheld in 2012 the consistency of the ESM Treaty with respectively, German Basic Law, and EU laws. The ESM has lent to Spain (€41.3 billion) and to Cyprus (€9.0 billion). On August 19, 2015, the ESM Board approved a new MOU with Greece and a new three-year loan (for up to €86.0 billion), following its endorsement by ESM members according to their national procedures. The IMF did not negotiate a corresponding Letter of Intent, or disburse resources.

The ESM’s financial assistance to Spain is noteworthy because a concurrent request for an IMF program was absent.25 The Eurogroup announced on June 9, 2012 that it would respond favorably to an expected formal request by Spain for financial assistance by the ESM to cover estimated capital requirements (plus a safety margin), for the Spanish banking system.26 The Eurogroup statement added that the EC in liaison with the ECB, European Banking Authority (EBA), and IMF would propose the necessary policy conditionality for the financial sector. The IMF Managing Director issued a statement on June 9, 2012 strongly welcoming the Eurogroup’s announcement and noted that “The IMF stands ready, at the invitation of the Eurogroup members, to support the implementation and monitoring of this financial assistance through regular reporting” (IMF Press Release No. 12/215).

As was made known subsequently, the absence of an IMF financial contribution (or IMF program) for Spain was explained to euro area national parliaments by the fact that the IMF did not have a facility to provide sectoral financial assistance. Thus, the “where possible” clause related to the IMF’s involvement in the ESM Treaty came into play. Nonetheless, while it is true that the IMF had no sectoral lending facility, IMF resources were used to help to fill financing gaps arising inter alia from requirements for bank recapitalizations in the cases of Greece, Ireland, Portugal, and Cyprus.

The Spanish government formally requested financial assistance from the EFSF/ESM on June 25, 2012. A joint mission that included staff from the EC, ECB, EBA, EFSF/ESM, and IMF visited Spain from June 27, 2012 to July 4, 2012, negotiating an EC memorandum of understanding for ESM financial assistance and terms of reference for IMF technical assistance (TA) in the context of ESM financial assistance. This joint mission (and subsequent ones) resembled a troika mission plus participants from the two additional European institutions (the EBA and ESM/EFSF). The memorandum of understanding (MOU), which defined financial sector conditionality, was agreed between the EC and the Spanish authorities on July 20, 2012, or the same day as the IMF’s terms of reference were finalized. The Spanish authorities also agreed to comply fully with the EU’s Excessive Deficit Procedure commitments and recommendations, which provided macroeconomic and nonfinancial structural policy elements to this financial assistance from the ESM (Véron, 2016). As regards the terms of reference, Fund staff preferred that the request for technical assistance be made by Spain plus other individual Eurogroup members. This was partly because TA requests from Fund members can be acted on without Board authorization, while TA requests from non-Fund members—such as the Eurogroup—do require Board authorization, and partly because in the staff’s view, the Fund’s “honest broker” role would be enhanced if the request were not made by the EC, ECB, or EBA. In any event, the staff felt that the Fund should be free to voice disagreement, including publicly, with policy recommendations (for example on the extent of deleveraging, bailout/state aid, or legacy asset management) made by EU institutions.

The IMF terms of reference (TOR) for staff monitoring of the EU program for Spain were sent to the IMF Executive Board, for information only (FO/DIS/12/135; 07/20/15), on July 20, 2012, the day they were agreed. The TOR were made public later that day. They specified that the Fund staff would not be party to the EC’s memorandum of understanding for financial assistance, nor would the Fund staff be responsible for the MOU’s conditionality; these were matters solely for the Spanish authorities and the EC. The monitoring to be conducted by the Fund staff was described as a form of technical assistance under Article V, Section 2(b). Thus, the staff would play a very different role than in the euro area cases that used Fund financial assistance. IMF staff monitoring was to be conducted “independently of the views of the authorities and EC.” This role represented a compromise between the wishes of those euro area governments (not least Germany’s), that wanted IMF involvement, and of the Spanish government, which insisted that the IMF play only an advisory role and not impose any conditionality (Véron, 2016). The TOR did not restrict the Fund staff from expressing its views regarding recommendations and policies formulated by the authorities and the EC. The TOR also made clear that the IMF staff reports would be provided to the IMF Executive Board for information only.

This was the first time (since the box on EC-IMF cooperation in the staff report for Hungary’s 2008 SBA request), that IMF staff described to the IMF Board the nature of IMF-EC collaboration. However, the Board was merely informed, rather than engaged in a decision making process. Indeed according to De Las Casas (2017), management and staff made a series of choices that effectively excluded the Board from a decision-making role on possible modalities for IMF engagement with Spain. On July 25, 2012 the Board concluded the 2012 Article IV consultation with Spain and “welcomed the European financial assistance for the recapitalization of Spanish financial institutions and the accompanying policies, as well as the envisaged role of the Fund in monitoring progress.”

Joint review missions related to Spain’s financial sector were conducted by the EC, in liaison with the ECB and EBA. These verified compliance with the MOU’s policy conditions, while the IMF staff supported implementation and monitoring with analysis, policy advice, and its own regular reporting (Véron, 2016). Progress in meeting the EDP commitments was regularly monitored by the EC. Thus, the EC’s surveillance procedures were reinforced by its conditionality on macroeconomic and structural policies via its lending operation.

Interviews conducted by the IEO with the Spanish authorities and relevant staff at the IMF, EC, ECB, ESM, and EBA portrayed the IMF staff as a co-equal partner with the EC, ECB, and EBA as regards providing insightful analysis of Spain’s financial sector and appropriately targeted policy recommendations to address identified financial sector problems. In Spain as in Ireland and Portugal, Fund staff debated with partners over the appropriate pace of bank deleveraging, arguing for a slower pace (Véron, 2016). Cooperation was deemed excellent by all parties and Spain successfully exited from this EC program with a stronger financial system. But some Europeans expressed doubts whether the “Spanish model”—with no parallel IMF program or conditionality—would be easily repeated in the future, and expressed a desire to have IMF “skin in the game.” Looking ahead, the ESM may be less involved with future bank recapitalization efforts because the Single Resolution Fund now provides an alternative instrument. Some European interviewees also noted that Spain’s macroeconomic policy performance under the ESM program could have been better; they noted in this context that the Excessive Deficit Procedure target for 2013 was not achieved, even though macroeconomic outcomes benefited from the sharply lower interest rates that were largely a response to the ECB President’s pledge in mid-2012 to “do whatever it takes” to save the euro.

What Is the Troika Arrangement?

When two or more institutions engage in conditional lending to support a country’s adjustment program, consistency is necessary because a country can only adopt one set of policy measures (for example, only one target for the fiscal deficit) that must satisfy both institutions’ conditionality.27 Even if conditionality does not overlap directly, policy measures must not work at cross-purposes. The institutions also need clear rules of the game to handle situations where the conditionality set by one institution is not observed, preventing that institution from disbursing as scheduled. Coordination issues naturally arise. To address them, two broad options exist: (i) one institution “borrows” the conditionality set by the other, tying its disbursements to disbursements by the other institution—such arrangements can be termed “co-financing”; or (ii) the participating institutions provide “parallel” or “joint” financing and agree on modalities to assure consistent conditionality.

Before discussing how the troika parties handled the coordination of conditionality, this section looks more generally at how these coordination issues apply with respect to the IMF’s interactions with regional financing arrangements (RFAs) and currency unions.

Treatment of coordination in regional financing arrangements and currency unions

Both in surveillance and in the use of Fund resources, an IMF member’s membership of a currency union raises policy and procedural issues that do not apply for non-currency-union members. Policy and procedural constraints may also differ among currency unions,28 and the specifics of the financing mechanism and any associated conditionality of the RFA add further complications. Financing mechanisms for (or the policy rules of) currency unions may raise different issues from RFAs in regions without currency unions.

With regard to regional financing arrangements, the G20 finance ministers and central bank governors endorsed on October 15, 2011 six non-binding principles for cooperation between the IMF and RFAs (Box 5.3). Several of these principles are germane to the troika: (i) cooperation should respect the roles, independence, and decision-making processes of each institution, taking into account regional specificities in a flexible manner; (ii) cooperation should include open sharing of information and joint missions where necessary; (iii) consistency of lending conditions should be sought to the extent possible, to prevent arbitrage and facility shopping; and (iv) RFAs must respect the preferred-creditor status of the IMF.

Box 5.3.G20 Principles for Cooperation between the IMF and Regional Financing Arrangements, as endorsed by G20 Finance Ministers and Central Bank Governors, October 15, 2011

In November 2010, G20 leaders tasked G20 finance ministers and central bank governors to explore “ways to improve collaboration between RFAs and the IMF across all possible areas.” Based on contributions by the EU and by ASEAN+3 countries that are members of the G20, the following nonbinding principles for cooperation have been agreed. Also, collaboration with the IMF should be tailored to each RFA in a flexible manner in order to take account of region-specific circumstances and the characteristics of the RFAs.

  • (i) An enhanced cooperation between RFAs and the IMF would be a step forward towards better crisis prevention and more effective crisis resolution and would reduce moral hazard. Cooperation between RFAs and the IMF should foster rigorous and evenhanded surveillance and promote the common goals of regional and global stability.

  • (ii) Cooperation should respect the roles, independence, and decision-making processes of each institution, taking into account regional specificities in a flexible manner.

  • (iii) While cooperation between RFAs and the IMF may be triggered by a crisis, ongoing collaboration should be promoted as a way to build regional capacity for crisis prevention.

  • (iv) Cooperation should commence as early as possible and include open sharing of information and joint missions where necessary. It is clear that each institution has comparative advantages and would benefit from the expertise of the other. Specifically, RFAs have better understanding of regional circumstances and the IMF has a greater global surveillance capacity.

  • (v) Consistency of lending conditions should be sought to the extent possible, in order to prevent arbitrage and facility shopping, in particular as concerns policy conditions and facility pricing. However, some flexibility would be needed as regards adjustments to conditionality, if necessary, and on the timing of reviews. In addition, definitive decisions about financial assistance within a joint program should be taken by the respective institutions participating in the program.

  • (vi) RFAs must respect the preferred-creditor status of the IMF.

Though endorsed by the G20, these principles are not binding on the IMF or on any RFA. Importantly, the IMF Executive Board has not endorsed, nor even discussed, these G20 principles even though the International Monetary and Financial Committee (IMFC) in April 2011 “urged the Fund to work with regional financing arrangements to develop broad principles for cooperation with the IMF.” The IEO has not found any evidence that the Eurogroup, EC, or ECB have adopted the principles. The principles are too general to be used for meaningful assessment purposes.

The G20 held a seminar on an IMF staff paper entitled “Stocktaking the Fund’s Engagement with Regional Financing Arrangements” (IMF, 2013b) at the IMF on April 17, 2013. This staff paper had been circulated for information, but not discussion, to the IMF Executive Board on April 11, 2013, with a note that it provided background for the forthcoming seminar. According to published summary of seminar participants’ views (, they observed that while agreeing on the desirability of cooperation, the extent and form of such cooperation were “the most difficult question to answer.”

On May 10, 2013, IMF staff circulated to Executive Directors a note on issues for discussion (FO/DIS/13/64) related to the “Stocktaking” paper. Questions posed included: (i) whether Directors saw a need to review the nonbinding G20 principles; (ii) whether formal cooperation mechanisms should be put in place with individual RFAs; and (iii) whether financing mechanisms for currency unions raised different issues from RFAs in regions without currency unions. The IMF Executive Board had an informal discussion on May 13, 2013. No summing up or minutes were produced because the session was informal, but available records and interviews indicate that Executive Directors were not inclined to move towards a structured, formal arrangement with RFAs. In December 2015, in concluding their discussion of the IMF staff’s Crisis Program Review (IMF, 2015b), “many Executive Directors supported establishing operational guidelines that build upon the G20 principles for cooperation between the Fund and regional financing arrangements (RFAs)….”

The troika arrangement was uniquely developed by the Eurogroup to benefit from the IMF staff’s technical expertise and crisis-management experience, allowing coordination of the EU’s and IMF’s separate, but parallel, conditional lending operations. No other currency union has yet developed a financing mechanism such as the euro area’s ESM. In studying the troika arrangement, it is extremely difficult to distinguish its possible effect on loan conditionalities from the possible effect of the policy constraints that were imposed on the program countries by their membership of the euro area and EU. To the extent that lessons from the troika experience derive from the effects of euro/EU membership, any lessons would be less germane for RFAs without a currency union. For RFAs that are also a currency union, the lessons depend on the similarity of their policy and financing frameworks to those in the euro area and EU.

“Borrowed” conditionality

Against this background, the relevant policies established by the IMF Executive Board are examined. According to the Fund’s Guidelines on Conditionality (Decision No. 12864 (02/102), as amended), the IMF is prohibited from allowing the use of Fund resources to be directly subjected to the rules and decisions of other organizations. Thus, the Fund cannot “borrow” conditionality from another institution. The Conditionality Guidelines also state that “there will be no cross-conditionality, under which the use of the Fund’s resources would be directly subject to the rules and decisions of other organizations.” Fund staff reiterated this point in 2014 (IMF, 2014), saying that the Fund cannot delegate its responsibility, including to RFAs, in assessing whether the conditions for the use of its resources have been met. This is necessary in order for the Fund to ensure that “adequate safeguards” are in place to preserve the revolving character of Fund resources as required by its Articles of Agreement. If the Fund assesses that its conditions have not been met, it will not disburse, irrespective of the judgments reached by other lenders. Conversely, in cases where the Fund assesses that its conditions have been met but the conditions imposed by other lenders are not met, so that they do not disburse, the Fund may not be able to release its resources. The absence of financing assurances—a situation wherein the IMF-supported program is not fully financed—can block the IMF from disbursing, given the need to safeguard its resources.

Unlike the IMF, the Eurogroup could have decided to “borrow” IMF conditionality by choosing to trigger its financial assistance solely upon the program country’s observance of IMF conditionality, or by deciding effectively to cofinance the IMF program. It must be noted that the euro authorities did not consider delegating program conditionality to the IMF at the time they were debating the IMF’s involvement. Borrowed IMF conditionality has been used in debt restructurings by the Paris Club and London Club, and in official bilateral lending during the Asian crisis, Mexico (1995), and Brazil (1998). According to IMF staff, only one out of five regional financing arrangements requires an IMF program for use of RFA resources, and in that case the use of these resources must exceed a threshold amount (though the use of the RFA resources has never been activated).29,30

Even without separate European loan conditionality, the design of the IMF-supported program for a euro area member needs to contend with the country’s EU Treaty obligations as administered by the EC and ECB. The EC and ECB have an obligation to treat EU/euro members evenhandedly, while also considering the potential spillovers for the EU/euro area as a whole. Thus, the policy disagreements that arose between the IMF staff and EC/ECB staff within the troika (e.g., on the pace of fiscal adjustment, sovereign debt restructuring, bank recapitalization, or treatment of unsecured bank creditors) would likely still have emerged even without the troika arrangement.

Modalities for assuring consistent conditionality

What of the modalities to assure consistent conditionality by the EC (in liaison with the ECB) and the IMF? The IMF’s Conditionality Guidelines state that “the Fund’s policy advice, program design, and conditionality will, insofar as possible, be consistent and integrated with those of other international institutions within a coherent country-led framework.” In addition, Fund staff explained in the 2011 Review of Conditionality (IMF, 2012b) that the “[Conditionality] Guidelines do not provide explicitly for coordination with regional institutions [such as EU institutions]. However, they provide clear guidance regarding coordination with the World Bank that can be transposed to coordination with other institutions.” In 2014, the operational guidance to IMF staff on the 2002 Conditionality Guidelines was revised to add a paragraph on collaboration with regional financing arrangements; in particular, it was considered “useful for staff to understand the timing and phasing of RFA disbursements . . . and to reach mutual understandings on policy objectives and program design to remove or minimize any inconsistencies” (IMF, 2014).

Some but not all of the IMF’s experience in coordinating with the World Bank on joint conditional lending to developing countries is relevant to the EU program cases. Typically, the Bank-Fund collaboration process has involved exchanges of information and analysis, sharing of briefing papers, and joint or parallel staff visits. As regards policy substance, under Bank-Fund collaboration a division of labor applies that is consistent with their respective institutional mandates. This principle has resolved most (though certainly not all) Bank-Fund coordination issues.

Unlike that of the World Bank, the policy mandates of EU institutions with respect to euro members overlap extensively with that of the IMF. Hence the IMF and the EC developed a modus operandi for assuring “consistent and integrated” conditionality—per the conditionality guidelines—based upon their experience with joint lending programs for Hungary, Latvia, and Romania. One consequence of the overlapping responsibilities was duplication in staff assignments by troika partners. This increased the overall size of troika teams, which placed a burden on country authorities. Troika teams in Greece reportedly could total 30–40 persons, though teams in Ireland and Portugal were substantially smaller. In addition, country authorities, particularly in the case of Greece with the EU’s Task Force for Greece, needed to accommodate a great number of technical assistance missions in revenue administration, expenditure management, banking, and statistics.

The IMF, EC, and ECB all used similar internal procedures/practices for fielding their respective teams (ECA, 2015). Once tentative mission dates had been determined, each team would begin to prepare a policy brief/note that identified the main challenges facing the country, the principal policy recommendations of the respective team, and an assessment of financing requirements. Consultations would take place among troika partners (including via teleconferences and the sharing of preliminary notes), and the EC would consult with the EFSF/ESM on funding issues while, within the EC, the Directorate-General for Economic and Financial Affairs, which provided the EC team leader, would consult with other Directorate-Generals. Efforts would be made to converge on analysis, assessments, and policy prescriptions, while retaining needed flexibility given the uncertainties involved. Internal consultations would also take place within each institution. Senior officials within each—the responsible Deputy Managing Director in the case of the IMF—approved the final draft policy brief/note. The EC played a dual role: one, acting as agent for the euro area members (or EFSF/ESM), it sent the policy brief to the Economic and Financial Committee/Euro Working Group President; two, it represented the general interests of the EU community because formal ECOFIN Council decisions may be required (e.g., EDP), or EFSM disbursements, which are EU-wide matters.

On the IMF side, the policy on exceptional access (EA) mandates early informal consultation with the Executive Board once IMF management decides that new, or augmented, exceptional access to Fund resources may be appropriate. The EA policy requires that Executive Directors be provided a concise note that sets out “as fully as possible:” (i) a tentative diagnosis of the problem; (ii) outlines of the needed policy measures; (iii) the basis for a judgment that exceptional access may be necessary and appropriate, with a preliminary evaluation of four substantive criteria and including a preliminary analysis of the external and sovereign debt sustainability; and (iv) the likely timetable for discussions. Concise notes were circulated to Executive Directors in all three country cases. While the initial notes in each of the three cases signaled that exceptional access to Fund resources was anticipated, none of them provided quantitative estimates of the financing requirements, of expected European financing, or of possible access to Fund resources. This information was, however, contained in the respective policy notes that were sent to IMF management before the three initial notes were sent to the Executive Directors. According to interviews with various IMF Executive Directors/Alternates, or their staffs, who attended informal Board meetings under the exceptional access policy, quantitative estimates were not communicated orally either.

Such estimates were made available to the Eurogroup by EC staff to gain the Eurogroup’s authorization to negotiate loans and policy conditionality on its behalf. Consequently, an information asymmetry resulted among IMFC finance ministers, with finance ministers from the Eurogroup having more detailed information. Depending upon what information the Eurogroup shared with their IMF Executive Directors, this information asymmetry might have also extended to IMF Executive Directors. This information asymmetric is distinct from the usual information asymmetry enjoyed by the Executive Directors representing the country seeking an IMF-supported program.

In the field, troika teams met jointly with the country authorities whenever feasible. They also met regularly among themselves to share information, to revise the macroeconomic framework and estimated financing requirement, to discuss adjustments to proposed policy conditionality, and to give mutual feedback on evolving drafts of their MOUs/LOIs. Progress reports were provided to headquarters—in some cases daily—to seek additional guidance. Because the EC acts as agent, the Euro Working Group (EWG) President is kept informed of developments by EC staff. The EWG President may inform other Economic and Financial Committee/EWG members of important developments as appropriate. If substantial disagreements arise among troika partners or with country authorities, troika deputies are involved, working with their counterparts and their teams to devise solutions. As necessary, troika principals may discuss matters with the objective of allowing commonly agreed proposals to be presented to the Eurogroup.

The IMF’s policy on exceptional access provides that “additional [to the initial] consultations will normally be expected to occur between informal meetings and the Board’s consideration of the staff report. The briefings will aim to keep the Board abreast of program-financing parameters, including assumed rollover rates, economic developments, progress in negotiations, any substantial changes in understandings, and any changes to the initially envisaged timetable for Board consideration. . . . Management will consult with the Board specifically before concluding the discussions on a program and before any public statement on a proposed level of access.” Additional informal consultations with the Executive Board prior to the announcement of a staff-level agreement (see De La Casas, 2017): Greece (2); Ireland (1); and Portugal (2). As discussed earlier, Executive Directors were not provided quantified estimates of the financing gaps, possible IMF access, or European financial support in the initial concise notes for Greece, Ireland, and Portugal or during the respective informal Board meeting on exceptional access. Directors were only informed once staff-level agreement had been reached and was about to be announced publicly. Executive Directors were not in all likelihood as well informed as Eurogroup members.

Executive Directors were not consulted in advance under the EA policy on three key policy issues: (i) in Greece, related to the absence of “high probability” for sovereign debt sustainability; (ii) the need to amend the EA policy in the case of Greece, by introducing the systemic exemption clause; and (iii) whether to apply haircuts to senior unsecured bondholders in the case of Ireland. As to consulting “before concluding discussions on a program,” in the cases of Greece and Ireland the last informal briefings took place on the same day as the announcement of the staff-level program agreement, while in the case of Portugal the last informal Board briefing took place three days before the announcement of the staff-level agreement (De Las Casas, 2017). The same-day announcement of the staff-level agreements in the cases of Greece and Ireland raises a question whether the Board was consulted or merely informed.

The IMF staff needed to share confidential IMF information with EC/ECB/ESM staff (and vice versa) in order for the troika arrangement to function. The IMF’s code of conduct for its staff prohibits the communication of confidential information to outsiders (who include EC/ECB/ESM staff) without authorization. Such authorization could take the form of either direct instruction from management or general policies established by the management and the Executive Board. According to the Legal Department, management has the authority to consent to such sharing without the need for a policy approved by the Executive Board, and without the need to inform the Board of such sharing or to share the same information with the Board. Staff in EUR, LEG, and SPR were not able to provide the IEO with copies of written authorization by management to permit sharing of confidential information with troika partners. Nor were concurrent records (such as minutes or memorandums to files) provided documenting oral authorization by management. SPR and LEG maintain that the sharing of confidential IMF information with third parties is authorized by the Board in the context of obtaining financing assurances for the member’s program; creditors/lenders will not support the country’s program without knowing the Fund’s contribution and level/magnitude of policy adjustment. In any case, written staff guidance on sharing of confidential information under these circumstances was not provided to the IEO. The ECA in its 2015 audit report noted that no formal arrangement existed between the IMF and EC regarding the exchange of confidential information; the ECA also recommended formalizing the mechanism for information sharing and the handling of confidential information. It is also good practice to obtain assurances from a recipient party that it will treat shared confidential information confidentially. The IMF staff has not been able to provide written evidence of such assurances from the EC, ECB, or ESM.

Decisions by the European partners (the Eurogroup, European Council, and the EFSF/ESM/EFSM) related to euro/EU loans preceded the IMF Board meetings on use of Fund resources. This sequencing assured that the IMF-supported program was fully financed—satisfying the IMF’s financing assurances policy—by the time the IMF Board met. But this sequence also created the perception that the IMF Board was faced with a fait accompli, and that the IMF Board merely rubber-stamped decisions that had already been taken in Europe. Alternatively, the IMF Board could have held its meetings prior to the decisions by European partners, using a more cumbersome “in-principle” decision procedure. Under this procedure, the Board approves Fund action in principle, but that action only becomes effective once the European partners take the corresponding decision. Whether making this procedural change would alter these perceptions is open to debate.

How Operationally Efficient Was the Troika Arrangement?

Answers to the question about the operational efficiency of the troika arrangement may differ from one program country to another and even for a single country depending upon the period chosen. This section attempts to provide a high-level overview of troika operations in Greece, Ireland, and Portugal. At the outset, it must be noted that an assessment of operational efficiency of the troika arrangement is distinct from an assessment of the quality or suitability of its policy advice or program design.31

The operational efficiency of coordination within the troika arrangement was examined by the IMF staff in the 2011 Review of Conditionality and in the context of ex post evaluations (EPEs) for exceptional access arrangements for Greece and Ireland.32 Two of the three scheduled EPEs for programs with euro area countries have been completed; those for Greece (IMF, 2013e) and Ireland (IMF, 2015a) have been discussed by the Executive Board. The EPE for Portugal has not yet been completed and issued to the Board, though it should have been circulated to the Board in early July 2015 to adhere to the exceptional access policy.33

To place the analysis of the troika’s operational efficiency in context, the time between the request for financial assistance by the euro member and the announcement of the staff-level agreement of a program was calculated for Greece, Ireland, and Portugal. The shortest was two weeks and the longest was four weeks, indicating that the troika arrangement was able to quickly negotiate the initial programs for these euro countries. In addition, the quarterly program reviews were completed in a timely manner; any delays were due to substantive policy disagreements with the country authorities rather than to policy disagreements among the troika members.

According to the staff report for the 2011 Review of Conditionality (IMF, 2012b), troika coordination “functioned well operationally and improved over time, but nevertheless added an additional layer of complexity to conditionality design and decision-taking.” At times this added complexity produced extended periods of discussion on crucial issues such as the pace of fiscal consolidation, debt restructuring, or regaining competitiveness. Coordination developed “in the spirit”—as formal agreement was absent—of the Bank-Fund Concordat. The Fund was seen as focusing on short-term macro-critical policies, while the EC covered comprehensive medium-term structural reforms. Overlaps existed within this division of labor, notably on fiscal, competitiveness, and financial policies. (From the IMF’s perspective, fiscal and competitiveness policies are of particular importance in programs with currency union members, which cannot use exchange rate policy to achieve adjustments, yet in the euro area countries the EC also had responsibility for fiscal and competition policies.)

EC structural conditions have been observed to be more numerous and detailed than structural conditions in the Fund-supported programs.34 The large number of structural conditions identified in the EU program contrasted with the IMF’s principle of parsimony. Staff noted that, over time, the IMF and EC each ventured increasingly into areas of structural reform that were initially the province of the other institution. These overlaps increased the need for coordination, requiring “constant cross-checking and a gradual adaptation between the MOU and MEFP” (IMF, 2012b). Such coordination, plus the reliance on review-based conditionality, avoided situations where cross-conditionality might prevent IMF disbursements owing to an absence of financing assurances.

In light of these findings, the 2011 Review of Conditionality concluded that the conditionality guidelines remained broadly appropriate, while implementation needed to be strengthened by, inter alia, “improving partnerships with other institutions including in currency unions, where program success can be linked to union-level policies.” More specifically, the staff recommended “maintaining a standing dialogue with regional financial agencies on policies and procedures regarding program conditionality and design, including a discussion of approaches for dealing with recurrent problems.” But, arguing that to do so was premature, especially in the euro area context, the staff provided few details on how to improve these partnerships through policies and procedures pertaining to program conditionality and design for a member of a currency union.

At the Board meeting for this conditionality review (September 2012), Executive Directors acknowledged that experience with the troika arrangement was limited but many of them nonetheless wanted a more in-depth study of the troika. For example, the Director from Japan encouraged staff to conduct, if necessary, an ad hoc review of the conditionality guidelines in order to reflect lessons learned. The Director from Australia noted that “we would be interested in a more in-depth discussion of the role played by European institutions in program design. Lessons drawn from the more developed relationship with the World Bank may provide guidance on enhancing the operational aspects of cooperation.” Similarly, the Director from Canada believed “that the costs and benefits of the troika model merit additional consideration” and “ask[ed] the staff to look deeper into the troika partnership and report back to the Board with recommendations for this partnership.” The Director from the Netherlands suggested that “going forward, we would encourage some written framework of cooperation between the Fund and partner organizations.” On the other hand, many Directors representing euro area countries were of the view that troika cooperation “proved quite successful in the end;” was “very effective” and “well-functioning;” some said they would “insist more than staff on the positive aspects of this cooperation.” While the summing up endorsed the recommendation to have a standing dialogue with relevant regional organizations, it added that “many Directors encouraged staff to draw preliminary lessons from these [euro area] cases in a timely manner, including on coordination with troika partners and the modalities of designing programs and conditionality.” To date, the staff have not prepared a Board paper to present such lessons, although a short box on cooperation experience with the IMF and EU institutions appeared in a report (IMF, 2016) that was prepared for an informal discussion on strengthening the international monetary system.

The EPE for Greece painted a similar picture to that in the 2011 Conditionality Review, while the EPE for Ireland did not specifically refer to troika-coordination problems. In the EPE for Greece (IMF, 2013e), Fund staff concluded that the troika coordination was good despite differences in its members’ internal procedures, documentation requirements, and confidentiality rules. The EPE pointed out that the IMF and European institutions had different perspectives: the IMF was more accustomed to analyzing issues from the vantage point of the specific country, while European institutions emphasized possible spillovers within the euro area. The Fund staff observed that a clear division of responsibilities within the troika was difficult to achieve, given the overlapping responsibilities of the three institutions. Synergies were seen to arise from cooperation in areas with shared expertise, while European institutions had a comparative advantage in structural areas that were outside the Fund’s core areas of expertise. Thus, work in areas that were not macro-critical could have been assigned more efficiently, while scope existed to streamline procedures and documents to reduce the burden on country authorities.

According to the same EPE, the Greek authorities felt the troika arrangement suffered from coordination problems. They noted that the troika took time “to gel” as a unit to formulate, for example, a common macroeconomic view, but that dealing with the troika was fairly smooth. Detailed conditionality posed coordination challenges, while a lack of continuity in the troika teams added to the burden on the country authorities. Moreover, while the IMF made decisions in a structured fashion, decision making by the euro area was more fragmented, spanning multiple institutions and varying levels, including heads of states. All in all, the Greek authorities found the process to have exacerbated uncertainties and reduced the possibility of early agreements. They also endorsed the Fund staff’s recommendations to streamline troika procedures and documents.

In a joint statement to the Board, responding to the 2013 EPE for Greece, Executive Directors representing euro area countries “beg[ged] to differ on the assessment of the relative areas of expertise within the troika” and believed that “the functioning of the troika in Greece was overall much better than described in the paper.” Moreover, although internal troika discussions were acknowledged as protracted at times, those discussions “improved the quality of the policy advice.” The Board summing up concluded that “mindful of the need to ensure equal treatment across the Fund’s membership, Directors generally saw scope for tailoring the Fund’s lending policies to the particular circumstances of monetary unions, including appropriate modalities for collaboration with the union-level institutions.”

The EPE for Ireland (IMF, 2015a) did not discuss troika coordination itself, but noted that close and effective interaction between the IMF and relevant union-level authorities was required for program success. Interviews with troika teams for Ireland revealed that troika coordination was smooth, notwithstanding internal policy disagreements that were significant at times. To some extent, according to those interviewed, this smooth process may have reflected early lessons from the Greek experience. That said, the EPE for Ireland identified similar issues with the troika process as in Greece: initial teething issues as the teams learned to establish what would become a “very effective” working relationship based in part upon complementary expertise; and the difference in perspective between the IMF, with its country focus, and the European institutions, with their euro area focus.

The Irish authorities did not comment on the troika process for the EPE report. However, at the relevant Board meeting, the Alternate Executive Director for Ireland stated that “from a practical point of view, where the IMF is involved in a multi-institution program, it is much better for the program country if there is some form of coordination body, such as was in place with the troika. The troika worked reasonably well in Ireland and it certainly would have been more difficult to run a multi-institution program without such a coordinating entity.” Executive Directors representing euro area countries in their joint statement expressed the view that “the success of the Irish program also illustrates the effectiveness of cooperation with the troika.” The Board summing up did not mention issues related to troika collaboration.

The IEO conducted not-for-attribution interviews in June 2015 with staff of the EC, ECB, and ESM/EFSF who had worked as team members in the cases of Greece, Ireland, and Portugal.35 Overall, they saw troika teamwork as a continuation and deepening of a rather successful EU-IMF coordination experience in Hungary, Latvia, and Romania. They pointed out that while the troika arrangement was broadly similar in each case, differences emerged that reflected the individual countries’ economic circumstances and political situations, as well as the personalities and working styles of troika team leaders and other members. Some differences in working style stemmed from differences in institutional procedures; in particular, IMF mission chiefs had more delegated authority than EC/ECB heads, although as they gained experience the EC/ECB teams felt they were given more room for maneuver. Effective cooperation was seen to require trust and direct, open communication among all troika partners. Typically, troika teams took time to build the requisite trust in each other and personnel changes could necessitate a partial reset. Trustful and constructive personal relationships were viewed as vital for successful cooperation. Experience also showed that some policy disagreements among troika teams could only be resolved at a political level. That said, some European interlocutors expressed the view that direct contacts by IMF staff (on topics such as debt restructuring) with major euro members (such as Germany) outside the troika arrangement could create possible misunderstandings, to the detriment of troika cooperation.

European interviewees expressed some annoyance and surprise at the fact that IMF teams seemed not to understand or appreciate the constraints placed on national policy options by countries’ membership in the European Union and the euro area currency union. In their view, IMF teams appeared to have an individual country focus and to pay only limited attention to the implications for—or spillovers to—other EU/euro countries. They contrasted this focus with the EC/ECB emphasis on preserving the single market and currency union; on minimizing spillovers (such as could have affected other euro area members from the proposed “haircut” for Irish senior bondholders); on avoiding tilting the competitive playing field via state aid (particularly in the financial sector); and on adopting common practices for all EU countries. Nonetheless in their view, collaboration with the IMF resulted in valuable creative tensions, forcing EC/ECB teams to question their implicit operating assumptions and to encourage changes to various EU/euro rules/ policies.

The smooth coordination of troika conditionality could have suffered, but in fact did not, from two stumbling blocks. One, fiscal conditionality by both the IMF and EC was set consistently ex ante, but could have been inconsistent ex post. In the IMF arrangements the fiscal performance criteria were based upon cash nominal euro amounts for the primary deficit in the cases of Greece and Ireland and the cash nominal overall deficit in the case of Portugal. The EC’s fiscal target—in line with the Stability and Growth Pact and the Excessive Deficit Procedure—was set on the overall deficit relative to GDP, using European System of Accounts (ESA) accrual accounting. The European Court of Auditors criticized in its 2015 audit the EC’s monitoring of fiscal targets based upon ESA accrual, owing to problems with timely measurement, and recommended instead the use of quarterly cash balances with arrears limits. Automatic fiscal adjustors were features of the fiscal performance criteria for Greece and Ireland but not of the EC’s corresponding fiscal targets. These definitional differences could have meant that the fiscal targets set by the IMF or EC could have been missed while the other institution’s fiscal targets were met. As recommended in the EPE for Ireland, “a unified approach would have helped communicate the program objectives more effectively and avoid possible uncertainties and mixed signals.”

The second possible stumbling block stemmed from the extensive and detailed structural measures that the EC included in its MOU—estimated by the ECA at nearly 400 in the cases of Ireland and Portugal—compared with the Fund’s more parsimonious approach to structural benchmarks.36 With respect to structural conditionality, both the EC and IMF followed a review-based approach to determining whether to disburse their respective tranches. A review-based approach allows considerable flexibility to determine whether specified structural measures have been adequately implemented and whether to disburse. Thus as a result, the more numerous and detailed structural measures imposed by the EC did not produce inconsistent outcomes from the IMF’s more parsimonious approach to structural conditionality. However, such consistency is not assured and greater procedural clarity would be desirable. As the numerous and detailed structural measures contained in the EC’s MOU have been criticized (for example by Pisani-Ferry, Sapir, and Wolff, 2011b), some movement toward the IMF practice might contribute to lessening this potential problem. In addition, the IMF staff concluded (IMF, 2015b) that extensive structural conditionality may have resulted in reform fatigue in some cases; as the number of structural measures increased, the percentage that were promptly implemented declined. The staff also observed that the combination of IMF and EC structural conditionality may have strained the authorities’ implementation capacity. The European Court of Auditors (ECA, 2015) recommended that “[structural] conditions should be used sparingly, and should clearly relate to reforms that are essential to crisis resolution or the repayment of assistance. Programme teams should be obliged to justify the need for each and every condition.”

On Which Side of the Negotiating Table Should the ECB Sit?

The IEO’s evaluation of “The IMF Response to the Financial and Economic Crisis” (IEO, 2014) reported that in the context of the euro crisis some G20 authorities thought it was “inappropriate, from a governance perspective, for the IMF to be seated at the negotiating table alongside the monetary authority of a member country. In their view, this implicitly took certain policy actions ‘off the table’ and constituted bad governance.”

In their statement on April 11, 2010, the euro area heads of state defined the European Central Bank’s role in the troika arrangement as to work “in liaison” with the EC, which was tasked with negotiating conditionality—the MOUs—for the European financial assistance program for Greece. The ECB was also to provide assessments of economic developments under the program to the Eurogroup as an input for its disbursement decisions.37 Subsequently, these liaison and assessment functions of the ECB were enshrined in the ESM Treaty, which was signed on February 2, 2012.

Some IMF Executive Directors, country authorities, and commenters have expressed the view that it was inappropriate for the monetary authority—the ECB—to be seated on the same side of the negotiating table as the IMF (Bernes, 2014). Pisani-Ferry, Sapir, and Wolff (2011b) have cited several potential conflicts of interest between the ECB’s euro-wide policy responsibilities and its role in the troika. One, that the ECB’s focus on price stability might bias its recommendations toward fiscal consolidation in program countries. Two, that the ECB’s responsibility to provide liquidity assistance to banks could conflict with its responsibility to protect its balance sheet by bailing in the private and official sectors. Three, that Securities Markets Program/Outright Monetary Transactions operations in a program country can cause the ECB to become a significant sovereign creditor, possibly causing it to take a tougher line on fiscal adjustment and debt restructuring.

These same governance issues and conflicts of interest, albeit over different policies, could be seen to arise with the EC’s role in the troika, too. For example, bank restructuring and competition policies at the EU level had significant implications for program design and implementation for Greece, Ireland, and Portugal. Moreover, conflicts have been perceived between the EC’s role as “guardian of the Treaty” and its tasks as Eurogroup agent in the troika (European Parliament, 2014a). These various concerns plus others led the European Parliament in February 2014 to adopt a resolution that, inter alia, called for the creation over the medium term of a European Monetary Fund by combining the financing role of the European Stability Mechanism with the EC’s conditionality functions. The European Parliament also requested that the “ECB be given the status of a silent observer with a transparent and clearly defined advisory role, while not allowing it to be a full negotiation partner.”

In written testimony to the EU Parliament, the ECB described its troika role as follows: “ECB staff provides advice and expertise on a broad range of issues which are relevant for ensuring a proper functioning of the transmission mechanism of monetary policy (including debt sustainability), contributing to financial stability, and ultimately supporting the general economic policies of the Union. The decision to grant financial assistance, including the conditionality attached, lies with ECOFIN and the ESM Board of Governors” (ECB, 2014). As described, ECB staff sat next to the EC staff at the negotiating table and across from the national authorities of the euro member, listening and advising, while conditionality associated with euro area financial assistance was set by the ECOFIN/ESM and not the ECB. While this arrangement was formally and legally valid, the ECB nonetheless played a significant role in program design, owing to its views on several threshold issues (such as debt restructuring; fiscal adjustment; bank “deleveraging”) and actions (such as providing bank liquidity; lowering sovereign interest rates via the Securities Markets Program). Meanwhile, the authorities of the national central bank (which like the ECB and other national EU central banks is part of the European System of Central Banks), sat alongside their national authorities, voicing their bank’s views on topics related to bank supervision, bank restructuring, and national emergency-liquidity assistance.

As noted earlier, when the IMF conducts surveillance of a currency union, it does so at two levels—the national and the supranational, or union, level—based on where the policy competency is located. Thus, when dealing with the euro area, the IMF conducts its surveillance over the EC and ECB as well as individual euro members. The supervision of banks within the euro area was a national policy competency until November 2014, when the ECB became the single supervisor. Provision of bank liquidity, or effectively in some circumstances the lender-of-last-resort function for banks within the euro area, is split between the ECB and the emergency liquidity assistance (ELA) provided by the national central bank. This same two-level arrangement applies to IMF surveillance of the CEMAC, WAEMU, and ECCU, but in these three currency unions bank supervision and the provision of emergency liquidity to banks is a union-level responsibility, according to the IMF staff (IMF, 2012b). The supranational and national authorities of these four currency unions can be viewed as sitting on the same side of the table and across from the IMF during Article IV consultations.

Currency unions, or other regional financing arrangements, are not IMF members and therefore they cannot request to use Fund resources.38 However, for a country within a currency union the split in policy competencies that affects the conduct of IMF surveillance also affects the design of programs and the implementation of policy. In particular, the IMF’s conditionality guidelines state that “conditions will be established only on the basis of those variables or measures that are reasonably within the member’s direct or indirect control. . . .” Policy competencies that have been transferred to a supranational institution can reasonably be assumed to be outside the control of the national authorities. While the Fund’s surveillance policy and associated operational guidelines explain how surveillance for a currency union member should be conducted, the Fund’s conditionality guidelines do not offer similar explicit instructions to IMF staff, national authorities of currency union members, or supranational institutions of the currency union.

IMF-supported programs are customarily negotiated with the country’s fiscal and monetary authorities. Thus typically the finance minister and governor of the central bank sign the IMF’s letter of intent. When the IMF member is also a member of a currency union, the program negotiations take place with the national authorities, usually led by the finance minister. In the case of the CEMAC, WAEMU, and ECCB, the respective regional central bank often sends a representative (from the local/national office) to follow developments and to clarify issues pertaining to monetary policy. Unlike the European Central Bank, these regional central banks have not participated in joint missions with the IMF staff to design program conditionality. In these three currency unions, letters of intent are customarily signed only by the finance minister of the country using IMF resources. In the cases of Greece, Portugal, Ireland, and Cyprus, the letters of intent were signed by both the finance minister and the governor of the (national) central bank. The latter signed because the national central bank has a separate legal identity from the ECB and possesses germane policy competencies (bank supervision, emergency liquidity assistance).

According to the IMF Legal Department, under Article V, Section 3(a) of the IMF’s Articles of Agreement the IMF can impose program conditionality on union-level institutions such as the central bank (or can more generally require union-level measures) under certain circumstances (IMF, 2015b).39 Measures at the union level must be macro-critical and needed for the success of the Fund-supported program with a member of the currency union. In several instances, program conditionality has been effectively directed at the regional central banks of the CEMAC and WAEMU as well as the ECCU. In one instance, the CFA franc was devalued against the French franc to establish the necessary conditions for IMF-supported programs with members of the CEMAC and WAEMU. Following the devaluation, eleven Fund-supported programs were in place by end-March 1994 (IMF, 1995). In another instance, prior to the monetary reforms of 1993–94 in the CEMAC and WAEMU, the national fragmentation of financial markets within these currency unions led the IMF to impose national limits—quantitative performance criteria—on the net domestic credit of the national agency of the regional central bank. In the CEMAC, ECCU, and WAEMU, limits that were set on net credit to the government from the banking system typically took the form of performance criteria for the program countries. Limits on net credit to the government from the banking system were also used in IMF-supported programs for euro area countries. In a third instance, a special audit of the regional central bank (BEAC) for CEMAC revealed a significant risk that unauthorized outflows from BEAC’s reserves could occur due to poor oversight and inadequate internal controls. Under the IMF’s safeguard assessment program, remedial measures were identified and implemented in accordance with a time-bound action plan: IMF program reviews and new IMF programs for CEMAC countries would only proceed as long as BEAC made satisfactory progress. Board consideration of program reviews scheduled for the Central African Republic and Republic of Congo was postponed because of BEAC’s delays in implementing some of the actions (IMF, 2010c).

On two different occasions, structural conditionality was linked to actions under the authority of the Eastern Caribbean Central Bank (ECCB). In the first case, the letter of intent from the Finance Minister of Antigua and Barbuda, dated May 2010—the same month/year that the IMF program with Greece was approved—attached a letter from the Governor of the ECCB that “took note” that structural benchmarks—on the recapitalization of the Bank of Antigua and onsite inspection of domestic commercial banks—required direct actions by the ECCB. The Governor welcomed the inclusion of these benchmarks and gave assurances that the ECCB would take the necessary steps to observe both within the specified time frame. In the second case, in December 2011, the ECCB Governor sent a similar letter in the case of St. Kitts and Nevis promising that bank stress tests would be conducted and that the results would be shared with IMF staff as specified in the relevant structural benchmark.

In the euro area, the European Central Bank announced in August 2012 the creation of a new instrument—Outright Monetary Transactions (OMT) in the secondary sovereign debt market—that is intended to safeguard an appropriate monetary policy transmission and the unitary nature of euro area monetary policy. According to the ECB, no ex ante limit would be set on the size of OMT. To qualify for OMT, a member country must conform to strict and effective conditionality attached to an appropriate ESM program; the ECB will terminate OMT operations wherever there is non-compliance with the macroeconomic adjustment program. The European Court of Justice (2015b) ruled that the ECB’s OMT program as constructed was consistent with EU treaties. Thus, the ECB could use OMT to reduce or eliminate excessive risk premiums in sovereign yields, but it should not go further than necessary. The Court also ruled that the ECB has the authority to purchase government bonds in the secondary market—but only so long as such purchases would not have an effect equivalent to direct purchase of government bonds, and thereby undermine the effectiveness of the prohibition in Article 123 of the Treaty on the Functioning of the European Union. Interestingly, the Court did not mention the recommendation made by the Advocate General in January 2015 that the “ECB refrain from any direct involvement in the financial assistance programmes to which the OMT programme is linked.” Consequently, the ECB can still legally participate in the troika.

Using various central bank instruments (especially where currency-union financial markets are fragmented), a regional central bank could help individually small national economies to adjust their monetary conditions to their cyclical situations without adversely affecting monetary conditions in the currency union as a whole (see Kincaid and Watson, 2015). To tailor the design of fiscal policy to the prospective monetary situation in a euro-program country, troika teams need insight into that situation. Such insight could come from the ECB being more forthcoming with the IMF/EC teams about its policy intentions with respect to program countries within the currency union. Another example where the ECB could be more forthcoming concerns its ex ante commitments to as-needed bank liquidity support; Ajai Chopra, former IMF mission chief to Ireland, in testimony (September 2015) before the Irish Parliament, criticized this lack of ex ante commitment, noting that it hurt confidence in the banking system and likely increased the required amount of Euro-system funding. Moreover, in the summing up for the EPE on Ireland, Executive Directors “noted that securing strong commitments upfront from monetary union authorities would be important when those are critical for program success.” Ex ante commitments need to be followed through. In the case of Greece, the Fund staff noted (IMF, 2011c) that “contrary to program expectations,” the ECB Governing Council had not made a decision on accepting eligible collateral from the proposed tranche of government-guaranteed bank bonds. In the judgment of Fund staff, this was “itself a negative factor for system stability and is almost certainly contributing to tight credit conditions.”

Beginning in November 2014, the Single Supervisory Mechanism for the euro area banking system came into force with the ECB as the central prudential supervisor. The ECB directly supervises the largest banks while the national supervisors continue to monitor the remaining banks following instructions given by ECB. Thus, the national competencies for bank supervision have been transferred to the ECB. The Single Bank Resolution Mechanism for the euro area, with its own board (Single Resolution Board), became fully operational at the start of 2016. This means that the structure of euro area bank supervision is now more similar to that in the CEMAC, WAEMU, and ECCU. Thus, rather than the relevant national central bank or supervisory agency, the ECB and the Single Resolution Board would now seem to be the proper interlocutors for the IMF/EC on bank supervision and bank restructuring for an individual national system within the euro area. In addition, the ECB has policy instruments that can be directed toward monetary conditions in individual national economies without compromising its area-wide responsibilities. For example, macro-and micro-prudential tools could be used to affect bank lending and deposit rates only in program countries. Given that ELA provision and some macro-prudential tools remain in the hands of national euro area authorities, the relevant national authorities would seemingly also merit a seat at the table.

Was the IMF a Junior Partner in the Troika Arrangement?

The nature of the IMF’s role in the troika arrangement was questioned from the very beginning with the IMF being termed a “junior partner” in the troika arrangement. Two aspects have received attention: financial contributions and policy substance. As regards financial contributions, the IMF clearly was a junior partner, committing at most one-third of the program financing for Greece, Ireland, and Portugal, and substantially less in later joint programs for euro-crisis countries (Table 5.2 above). Only Ireland received the full amounts that the IMF and EC committed. Portugal decided not to request its last disbursement from either the IMF or the EC, following an adverse Constitutional Court ruling on expenditures, requiring more time to formulate a comprehensive response. In Greece, two-thirds of the amounts that were committed by the IMF and under the Greek Loan Facility were disbursed before these programs were replaced. Interestingly, this pari passu approach to disbursements did not continue with the Greece II or III programs, to which the IMF has not committed financial resources. With the Greek program off track, the last IMF purchase occurred in June 2014, lifting IMF credit outstanding to Greece to SDR 24.7 billion; subsequently, however, IMF credit outstanding to Greece declined to SDR 12.5 billion at end-January 2016, shortly after the Greek authorities cancelled the EFF. During this same period (June 2014 to end January 2016), the European Stability Mechanism (ESM) made net disbursements of €12.5 billion to Greece.

European interviewees observed that IMF loans are legally senior to loans from the European Financial Stability Fund and the European Stability Mechanism. The ESM treaty formally recognizes that IMF loans are senior to ESM loans. The original pari passu clause in the EFSF/ESM loans was waived for Ireland and Portugal to allow early repayment to the IMF. Moreover, when Greece failed to make scheduled repayments to the IMF in mid-2015, creating overdue obligations to the IMF, European partners accorded Greece enough European financial assistance, with appropriate conditionality, to allow it to extinguish its overdue IMF obligations and to help prevent a recurrence of arrears to the IMF.

With respect to policy substance, the perception expressed by outside commenters has been that program design decisions were taken by the EC and ECB, backed by the Eurogroup. For example, “if a regional grouping can set IMF conditionality, what is the point of the Fund anyway? This could create a very dangerous precedent” (Goldstein, as reported in the Financial Times, April 2010). The Fund’s “credibility is being squandered by the IMF serving as the junior partner…” (Chowla, 2011). The Fund “is the junior partner in a ‘troika’ of institutions…” whom “the pro-austerity ECB and EC has outmuscled” and “the Fund’s views count for less than its partners” (Coggan, 2012). The IMF has been “used as a cover for the continent’s policy makers and its independence lost” (Mandeng, 2013); in a similar vein, “the IMF has toed the European/German line on the crisis, possibly to the disservice of Europe and the world” (Subramanian, 2012). The occasional contrary view appeared in the press: “the Fund could be a junior partner in terms of financing but a senior partner in terms of negotiations” (Prasad, quoted in Beattie, 2011b). Reflecting this debate, the IEO (2014) observed that the troika arrangement “raises questions as to whether it afforded greater traction of IMF’s policy advice, or whether it increased the pressure on the IMF to compromise its positions.”

To obtain a view from inside the troika, not-for-attribution interviews were conducted by the IEO with staff from the EC, ECB, ESM, and IMF who had participated in troika activities for Greece, Ireland, and Portugal. Their clear and common opinion was that the IMF was not a junior partner with respect to the policy substance of these programs. But neither was it a senior partner. This contrasted with the IMF’s customary sole, or lead, role in its lending to emerging market and developing countries. From a European perspective, the IMF needed to get accustomed to not being alone in the driver’s seat and to learn to act in tandem with EU institutions. The three troika partners were frequently described as each having a veto power on actions, which forced them to find collectively an approach that each of them could accept. The European agencies’ veto power derived from their financial contribution but also from the need for the program country to have its policy actions endorsed by the EC, ECB, or European Council, given its EU treaty obligations. The IMF’s veto power stemmed from the recognition of its considerable expertise and crisis-management experience, and its credibility with key euro members and their parliaments whose consent was required in the context of EFSF/ESM lending decisions. Thus, the troika arrangement was effectively viewed as comprised of co-equal partners.

These interviewees also disputed the notion that the IMF’s relatively small financial contributions muted either its voice in policy debates or its impact. To support this contention, they observed that it was the IMF’s expertise and experience, and not its financial resources, that prompted its invitation from the Eurogroup to participate. In addition, they pointed out that in the second Greek program (the EFF) and the Cyprus program, the Fund’s influence over program design remained unchanged even though its financing share had fallen (to 16 percent and 10 percent, respectively, from roughly one-third). They noted the IMF staff’s key roles in assessing (in 2011/12) the suitability of the Greek sovereign debt restructuring and the bank restructuring in Cyprus.

Clearly various outside commentators viewed the situation differently from the troika participants. To draw one’s own conclusions, it is necessary to identify specific situations in which the IMF and EU institutions apparently had, at least initially, a difference of view on the preferred policy approach, and then to discern whether the troika arrangement as a coordinating device was the responsible driving force or whether membership in the euro area currency union was the dominant force. Specific situations examined below are: (i) the disagreement about sovereign debt sustainability in Greece, along with the introduction of the systemic exemption clause to the Fund’s exceptional access policy; and (ii) the disagreement about how to treat senior bank bondholders in Ireland. This section draws heavily on other chapters of this volume.40

The IMF has been criticized (and has criticized itself in the EPE for Greece (IMF, 2013e) for not restructuring Greece’s sovereign debt in early 2010 and for introducing the systemic exemption clause to the exceptional access policy in May 2010. In 2010, both the EC and ECB were opposed to sovereign debt restructuring, as were the Greek authorities. The IMF staff was divided on this issue. At that time, Fund management decided not to press for debt restructuring, owing to worries about possible contagion within the euro area (which lacked an adequate firewall) and about spillovers to a fragile world economy struggling to recover from the global financial crisis. These concerns were shared by at least some major IMF shareholders, notably the United States. If the troika—a coordinating arrangement—had not existed, would anything have changed? Since Fund management (and major non-euro IMF shareholders) considered debt restructuring by Greece to be too risky for the euro area and the global economy in early 2010, the Fund would probably not have proposed debt restructuring even had it been alone in the driver’s seat.

What about the decision to introduce the systemic exemption clause to the second criterion into the Fund’s exceptional access policy? This decision, taken at the IMF Board meeting to approve the SBA request by Greece, was purely an internal IMF matter, and IEO interviews suggest the euro area partners were as surprised by this change as other IMF shareholders. Indeed, euro area partners tended to regard Greece’s sovereign debt as sustainable if supported by the requisite fiscal consolidation, resulting in the programmed primary surplus of 6 percent of GDP. But some senior IMF staff, in particular those in the Legal; Research; and Strategy, Policy, and Review departments, had serious reservations about debt sustainability; they did not think that “a rigorous and systematic analysis indicate[d] that there is a high probability that debt will remain sustainable” as required under the exceptional access policy. At the same time, senior staff in the European; Monetary and Capital Markets; and Fiscal Affairs (FAD) departments argued that restructuring Greece’s debt would be too risky for the rest of the euro area if not the world.41 Moreover, some Fund senior staff argued that there was insufficient time to organize an orderly debt restructuring before large debt repayments fell due in mid-May.

Faced with serious doubts about debt sustainability, IMF management searched in April 2010 with senior staff for ways forward. IMF management was concerned that changing the debt sustainability criterion under the exceptional access policy might send an adverse signal to financial markets about the strength of the program, undermining its chances for success. Some IMF senior staff advocated that the IMF should approach European partners to obtain assurances that European lending over the medium term would be sufficiently concessional to help achieve debt sustainability. Other senior staff (SPR/LEG) noted that any change to the exceptional access policy could be “done quietly” in the SBA staff report; the Board discussion would enable further oral clarifications. In the end as observed by Schadler (2017), the decision to introduce the systemic exemption clause was made at the last minute and staff did not call attention to this policy change which was [“quietly”] embedded in the assessment of the second exceptional access criterion. Staff only offered oral clarifications after one Director questioned during the Board meaning of this passage in the staff report. Interviews with government officials from major non-euro IMF shareholders indicated that they supported introduction of the systemic exemption clause on the grounds that it was deemed necessary to allow the IMF to lend, albeit considering this change to merely be a “housekeeping” matter at the time.

The above analysis should not be construed as validating the decisions made, or as endorsing the IMF’s decision-making process, particularly regarding the introduction of the systemic exemption clause. Indeed, the IMF eliminated the systemic exemption clause in early 2016. Findings and conclusions related to IMF decision-making are outside the scope of this study, but are examined by De Las Casas (2017).

As economic developments in Greece turned out to be worse than projected and the euro area made policy changes such as the creation of the European Financial Stability Facility and the ECB’s Securities Markets Program, IMF management and staff became convinced that sovereign debt restructuring was necessary and feasible. They made their arguments for debt restructuring within the troika, starting in late 2010 and extending into 2011. With the passage of time, more European partners recognized the changed fundamentals. In July 2011 euro governments announced that the EFSF loan terms would be softened by extending their maturity and lowering their interest rate—effectively constituting official sector involvement—and that private sector involvement on a voluntary basis would take place as “an exceptional and unique solution” (DG-ECFIN, 2011c). Euro leaders accepted an initial proposal by the Institute of International Finance (Henning, 2011) for sovereign debt restructuring. However, IMF staff analysis concluded that this proposal was overly generous to private creditors and would not achieve debt sustainability (IMF, 2011d). As a consequence, the Institute’s proposal was revised to give private creditors a bigger haircut and a new agreement with euro leaders was reached in mid-October 2011. The deal closed in March 2012.42 These developments demonstrate that the IMF played an influential role but at the same time the key decisions were taken in Brussels and not in Washington. Greece’s membership in the euro currency union was the underlying reason for the decision-making locus to be in Brussels with the Eurogroup.

In the case of Ireland, IMF staff and management supported a “bail-in” of senior unsecured bank bondholders, or private sector involvement (IMF, 2015a; Donovan, 2017). While this bail-in would have benefitted Ireland albeit its size was modest, concerns about adverse spillover effects to the rest of the euro area caused the ECB and EC to oppose it. The Irish authorities were caught in the middle. IMF management took its case to the G7 finance ministers, and in a teleconference in November 2010 the U.S. Treasury Secretary vetoed any haircut for senior unsecured bank bondholders because it might spread via contagion to the European banking system and then back to the U.S. banking system. This episode shows that the IMF management had a “court of appeals” for policy disagreements with its euro partners. In this instance, the “court” supported the euro area’s policy view and not the IMF’s. This scenario could have played out in the same way without the troika arrangement—witness the involvement of the G7 during the Asian and Latin American crises. Also of interest in this case is that the G7, and not the G20 or the IMF Executive Board, was the forum selected to resolve this dispute.

The pace of fiscal consolidation and its implications for real growth were recurrent tension points among the troika partners in all three programs with euro area countries, as they had been in the three earlier programs with non-euro EU members (Annex 5.1). As noted by Kopits (2017), these three programs were constrained by currency union membership; in particular, the pace of fiscal consolidation was influenced by the EU institutions’ desire for program countries to reach the Maastricht reference level for the fiscal deficit (3 percent of GDP) by the end of the program period. The program fiscal targets proposed to the country authorities thus represented compromises reached within the troika. In addition, as the program countries’ real GDP turned out to be lower than projected, debates ensued among the troika teams pertaining to the operation of automatic fiscal stabilizers. They burst into the open with the publication of the October 2012 World Economic Outlook. That WEO reported empirical results showing that the short-term fiscal multipliers used for the three euro area programs had been systemically underestimated, implying a larger than projected fiscal drag on real activity. Both the EC and ECB published rebuttals (Box 1.5 in the EC’s European Economy No. 7/202, and Box 6 in the ECB’s Monthly Bulletin, 12/2012), adding other relevant macroeconomic variables to the equations and producing fiscal multipliers close to the value (0.5) employed in these programs.

As regards the three program countries, interviews conducted by the IEO indicated that their authorities tended to side with the EU institutions because they wanted to be considered “good euro area citizens.” Moreover, since the three countries had lost access to private creditors, following a more gradual fiscal adjustment path would have required greater official financing, which was not forthcoming from the EFSF/ESM or from the IMF. A more gradual fiscal adjustment would have also exacerbated already-existing concerns about the sustainability of these countries’ public debt, and adding relatively more senior debt from the IMF could be seen as adding to the possible haircut for private creditors as well as augmenting the risks for the IMF. These conundrums were not a product of the troika arrangement but were a product of the underlying economic situations and currency-union constraints.

Turning to financial sector policies, IMF staff were considered to be better prepared than the other troika partners, particularly in the initial years. According to Véron (2016), IMF staff had the needed experience with the politically difficult sequence of bank triage, recapitalization, and bank restructuring, which was most evident in Ireland and Spain and least in Portugal. On several occasions, the IMF’s heft, leadership, and problem-solving built consensus with the troika, and the Fund also showed itself more adept at interacting and learning from financial market participants. IMF staff members working on financial sector issues also earned the respect of key national policymakers for their competence, impartiality, and discipline. Tensions nonetheless arose among the EC, ECB, and IMF, in large part owing to institutional differences of perspective and interests. Véron (2016) reports that the need to reach a troika consensus often resulted in better policy assessments and choices than if the IMF had been acting alone. He also observes that the IMF staff on several occasions appeared unwilling to acknowledge the EU’s institutional realities, especially as regards state aid to the financial sector, to the detriment of the IMF’s own effectiveness.

In summary, the IMF was clearly a junior financing partner compared with the euro area institutions. According to troika participants, the junior financing status did not lessen the IMF staff’s impact in policy debates within the troika. This judgment (based upon interviews) is most strongly supported by the basically unchanged policy influence of the IMF staff in the design of the later programs for Greece (EFF) and Cyprus, notwithstanding the sharp drop in the IMF’s relative financial contribution.

Of course, the IMF’s influence might still diminish as the IMF’s financing share approaches zero. Here the experience in Spain is instructive. Spain’s program financed by conditional lending from the European Stability Mechanism was not accompanied by access to Fund resources, and the role, scope, and impact on policy conditionality of IMF staff lessened—by design—dramatically from those in the euro area country cases that used Fund resources.

In conclusion, the IMF was not a junior—policy—partner in the troika arrangement but neither did it play its customary role as the senior, or lead, policy partner. Co-equal partnership seems the appropriate characterization.

Key Findings and Conclusions

The Eurogroup devised the troika arrangement in 2010 to meet their requirements and capabilities at that time. IMF management and the IMF Executive Board implicitly accepted this arrangement, as the modus operandi for joint efforts to lend to euro countries in crisis. Because IMF management and the Executive Board did not approve explicitly the IMF’s participation, define its role in the troika arrangement, or produce written operational modalities, it is not possible to assess outcomes relative to expectations. Nor has the Executive Board yet endorsed the 2011 nonbinding G20 Principles for Cooperation between the IMF and regional financing arrangements. In any event, those principles are crafted with too high a degree of generality for evaluation purposes.

The IMF has a long history of parallel, conditional lending operations with the World Bank (and regional development banks). Indeed, IMF staff have asserted (IMF, 2012b) in the context of the troika that the Bank-Fund Concordat “can be transposed to coordination with other [regional] institutions.” These principles thus can provide a frame of reference to evaluate the troika arrangement from the IMF’s perspective. The Concordat was agreed mutually at the highest level—by the World Bank President and IMF Managing Director—and then circulated to the respective Executive Boards.

No mutually agreed principles exist for the troika arrangement. Agreed written principles on joint lending operations between the heads of the EC, ECB, and ESM and the IMF Managing Director would provide clarity for all parties plus member countries.43 Establishing clear principles that are mutually agreed for parallel conditional lending operations would promote more efficient interactions among troika partners and between the troika and the borrowing country. These mutually agreed principles endorsed by the appropriate governing bodies would also enhance the legitimacy of the troika arrangement and the accountability of the troika institutions.

Similar agreed cooperation principles adapted to the circumstances of each regional financing arrangement might also prove useful for the Fund’s possible program involvement with regional financing arrangements (RFAs). The International Monetary and Financial Committee called in 2011 for the IMF to work with RFAs to develop broad cooperation principles. Meanwhile the G20 established in late 2011 nonbinding principles, but the Executive Board has had no formal—decision-making—discussion of those principles; it has only discussed them informally—non-decision-making. The Executive Board in both 2014 and 2015 generally supported the development of cooperation guidelines with RFAs. It is time for the Board to discuss formally the G20 principles for cooperation between the IMF and RFAs and for the IMF staff to develop individualized principles for cooperation with each RFA.

The troika arrangement was uniquely developed by the Eurogroup to benefit from the IMF staff’s technical expertise and crisis-management experience, allowing coordination of their separate but parallel conditional lending operations. In these circumstances, a major challenge in studying the troika arrangement is to disentangle the implications for program (policy) design of the conditional loan coordination process from the underlying implications of membership in the euro area and EU. Because it may be impossible to adequately disentangle these two factors in the troika’s operations, it would be hazardous to apply lessons learned from the troika arrangement to other regional financing arrangements that are not currency unions. As for RFAs with currency unions, the lessons would depend on their similarity with the policy and financing frameworks developed by the euro area and the EU. Consequently, cooperation frameworks between the IMF and any RFA would need to be individually tailored, although based upon broad principles in order to ensure consistent treatment across RFAs.

The troika arrangement proved to be operationally efficient, although areas for improvement were also identified. Conditional lending programs were negotiated quickly by the troika with the country authorities and program reviews were by and large completed expeditiously; program delays could not be attributed to the troika process itself. This assessment is based on IMF staff reviews, EPEs for Greece and Portugal, and IEO interviews with troika participants and relevant country authorities. Areas for improvement in the troika arrangement include: (i) agreed procedures among the troika institutions that are transparently shared with their memberships and the public; (ii) enhancing the information flow to, and role of, the IMF Executive Board; and (iii) efforts to reduce burdens placed on country authorities by large missions, staff turnover, duplicate documentation, and extensive conditionality.

While the principle of lead institution based upon areas of primary responsibility that is used in the Bank-Fund Concordat cannot be easily applied to the troika arrangement, the Concordat does recognize that “there is a broad range of matters which are of interest to both institutions” and that therefore enhanced collaboration is needed between the institutions. In particular, close contacts between the two staffs including “sharing of information and views at the earliest possible stages” is expected to produce “improved and consistent policy advice.” Thus, sharing of confidential information by IMF staff with Bank staff and vice versa is formally authorized; this is consistent with a G20 cooperation principle. No written principles for sharing of confidential information amongst troika institutions has been made available to the IEO. Written guidance to IMF staff on the sharing of IMF confidential with troika partners, which is consistent with the staff code of conduct, is also absent.

The IMF and EU institutions should regularize their mutual sharing of confidential information. Such an agreement has been proposed by the European Court of Auditors. An internal guidance note on the sharing of confidential IMF information would avoid possible inadvertent violations of the staff code of conduct. A cooperation document could also clarify issues related to the various IMF expectations described in the ESM Treaty. Procedures could be established to resolve differences of view at the mission level by involving their respective superiors. If such matters remain unresolved, interactions could take place between the appropriate IMF Deputy Managing Director and EU counterparts, and if necessary the IMF Managing Director and the corresponding head of institution can resolve matters. Such procedures would also provide that (in the words of the Bank-Fund Concordat) “in those cases, which are expected to be rare, the managements will wish to consult their respective Executive Boards before proceeding.” This latter provision may need to be modified to fit differences in the memberships, voting powers, and governance structures of various currency unions and RFAs.

Articulating three other topics from the Bank-Fund Concordat could improve the IMF’s relations with the EC, ECB, and ESM (and perhaps RFAs more generally):

  • The IMF needs to avoid inconsistent conditionality with these institutions, especially in overlapping policy areas such as fiscal policy, financial sector restructuring, and structural reforms, while also avoiding cross-conditionality. Each institution should be allowed to proceed with its own financial assistance according to the standards required by its legal charter and governing bodies; this also is a G20 cooperation principle. However, in the event one institution were to consider proceeding without the others, the conditions for such action should be understood in advance—including the scope for prior communication.

  • Cooperation principles could spell out efforts to reduce the burdens placed on country authorities by large mission teams, frequent changes in team staffing, and needless duplication of documents.

  • The implications of overdue obligations, or arrears, to one institution by a borrowing country for the actions of the other institutions could be usefully clarified in cooperation principles, recognizing of course the different nature of the various institutions (for example that debt obligations to the IMF are senior to those to the ESM); a G20 cooperation principle specifies that the IMF’s preferred-creditor status must be respected.

Such an institutional agreement could also clarify the conditions for requests to the IMF to provide technical assistance when EU institutions are lending to a euro member, such as took place in the case of Spain, and the modalities to be used by the IMF in such a case. Thus, the rules of the game for such TA provision would be jointly and transparently established. In this connection, it would be useful if the IMF staff prepared an ex post evaluation for Board consideration of its technical assistance activities with Spain and the EU institutions during 2012–13. This staff evaluation could identify lessons learned that could inform possible future such TA operations.

Cooperation principles—such as embodied in the Concordat—are more about process than about the substance of program design and conditionality. The IMF has long recognized (see IMF, 1995) that program design and conditionality for countries that are members of currency unions need to differ from that for countries that have a flexible exchange rate and an independent monetary policy (in particular, fiscal policy and structural reforms must play a larger role in programs with countries that are members of currency unions). Moreover, policy competencies in a currency union are split between national- and union-level authorities. The implications of this split for the conduct of Article IV consultations are explicitly considered in the various IMF surveillance decisions and in the corresponding guidance notes to staff. But neither the 2002 Conditionality Guidelines nor the Revised Operational Guidance (IMF, 2014) inform the IMF staff or country authorities as to its implications for program design or conditionality. Where should the EC and ECB sit at the negotiating table and under what circumstances could program conditionality be appropriately assigned to them? A country’s ownership of its policy program is crucial for successful implementation and is a central tenet of IMF conditionality guidelines. But the authorities in the euro area program countries also own their euro membership, wanting to be considered “good euro area citizens” in the eyes of the Eurogroup, EC, and ECB. This dual ownership can give rise to policy tensions, given the more constrained policy options associated with a currency union. Moreover, the economic governance structures built to support the euro area are stronger and more extensive than those in other currency unions, reinforcing member countries’ tendency to follow the policy advice given by EU institutions. Amending the Fund’s Conditionality Guidelines by introducing an explicit treatment of issues germane to countries in a currency union would bring these guidelines into conformity with surveillance policy and practices, and promote evenhanded treatment of IMF members in different currency unions.

The IMF was clearly a junior financing partner compared with the euro area institutions. However, this junior status did not appear to lessen the IMF staff’s influence in policy debates within the troika. Nor did the Fund see its influence decline when its relative financing contribution declined in the EFF for Greece or the EFF for Cyprus, whose access to Fund resources was below the threshold to trigger the IMF’s exceptional access policy. Of course, the IMF might see its influence diminish as its financing share approaches zero. At some point, the IMF might find its program (conditionality) involvement switched to an advisory (technical assistance) role with less influence, as was the case in Spain.

This junior financing role had advantages for the IMF, diminishing its exposure to credit risk and its need to borrow resources. Moreover, European partners were able to assist in the clearing of arrears by Greece to the IMF in mid-July 2015. Finally, the IMF’s smaller financial contribution made it possible for the IMF to reduce its exposure to these three program countries even as the ESM maintained or increased its exposure.

Though the IMF was a junior financing partner in the troika, the evidence and analysis marshaled in this paper indicates that it was not a junior policy partner. But neither did it play its customary role as senior, or lead, policy partner. A co-equal partnership seems to be the appropriate characterization of the troika arrangement in the three cases examined. On occasions, tensions emerged within the troika regarding proper policy recommendations, but their resolutions were typically constructive and represented differences in judgment and institutional realities. The consistency of conditionality that was achieved by the troika partners enhanced effectiveness and reduced the burden on country authorities. But the policy product of troika cooperation should also be consistent with the IMF’s mandate, policies, and practices, and recognize its independent decision making. (The same applies to the EU institutions.) In addition, the IMF should feel free to air publicly major policy differences with the EU institutions/Eurogroup and national authorities in order to preserve its credibility and independence—balancing of course the need to maintain its trusted advisor role. The disagreement between the IMF and EC over debt sustainability and fiscal sustainability in Greece, which became public in mid-2015, is a case in point.

The policy framework for exceptional access to Fund resources sets out stronger procedures for Board decision making on management’s proposals for exceptional access than exist for regular-access proposals. These stronger procedures were intended to provide additional safeguards for the use of Fund resources and to enhance accountability and include, inter alia, an early, informal Board session on needed policy measures and program-financing parameters whenever management decides that exceptional access is appropriate. Additional informal meetings are to take place as needed to keep Executive Directors “abreast” of progress in policy negotiations and program financing. Such Board meetings are intended to “provide the basis for consultations with capitals, and the issues that emerge would be addressed in a further informal session.” Two interrelated questions can be posed in this connection. One, were these strengthened Board decision-making procedures followed in the spirit envisaged under the exceptional access decisions? Two, did these procedures keep the IMF Executive Board and their capitals as well informed as the euro area finance ministers who attended the Eurogroup meetings?

Based on written documents circulated to Executive Directors, transcripts of informal Board meetings, and interviews with Executive Directors, Alternates, and their staff, serious shortcomings existed in the information provided, and issues presented, to Executive Directors, notably about estimated financing gaps; preliminary figures on European and IMF financing; doubts about debt sustainability in Greece; need to introduce a systemic exemption clause into the exceptional access policy; and the IMF recommendation, plus EC/ECB opposition, to apply a haircut to senior unsecured bondholders in Ireland. In most cases, the Eurogroup had the relevant information or was aware of the issues. (The Eurogroup was not aware of the extent of the doubts by IMF staff about Greek debt sustainability in 2010 or the need to modify the exceptional access policy by introducing the systemic exemption clause.) The information asymmetry was not caused by the troika arrangement itself, but stemmed from internal IMF practices/decisions. Both SPR and LEG contend that Executive Directors were provided all necessary information to make decisions under Fund policies; in particular all the requirements of exceptional access policy were observed. In light of these findings, the Executive Board might consider commissioning an independent review of experience with the implementation of the exceptional access policy, especially the extent of information provided and the policy issues that were presented during informal sessions.

Finally, the troika arrangement has been cited as one facet of a broader IMF “Europe is different” mindset, producing more favorable treatment of the EU and euro area than of other IMF members. The above recommendations—to define program design and conditionality for currency unions; to develop mutually agreed cooperation principles with the euro area (and other regional financing arrangements), especially procedures to settle policy disputes; and to enhance internal IMF decision-making procedures—would collectively go some way toward remedying such actual or perceived uneven treatment. Evenhanded treatment is reinforced by clearly defined rules of the game, mutually agreed implementation procedures, and transparent, informed, and broadly based decision making.

Annex 5.1. EC-IMF Cooperation in Lending Programs with EU Countries, 2008–09

Prior to the four IMF-supported programs with euro area countries (Greece, Ireland, Portugal, and Cyprus), the IMF and EC cooperated in joint lending operations to three non-euro EU members (Hungary, Latvia, and Romania) during 2008–09. The cooperation principles developed during these IMF-EU programs became the model upon which the troika arrangement was later built. In all three cases, IMF staff and EC staff butted heads over the pace of fiscal adjustment and the application of the SGP/EDP rules. Latvia stands out because it had a currency peg to the euro that IMF staff considered unsustainable, but that the country authorities with the financial support of the EU sought to maintain through the use of substantial fiscal consolidation and structural reforms. By 2010, when the joint programs with Greece, Ireland, and Portugal were designed, the lesson learned from Latvia’s case seemed to be that the EC got right the economic forecast and policy judgments (about the currency peg sustainability, feasible fiscal adjustment, and small fiscal multipliers) and the IMF apparently got it wrong. Subsequently, however, the applicably of this lesson to other economies has been called into question (Blanchard, Griffiths, and Gruss, 2013).


The Hungarian authorities contacted the Fund staff on October 9, 2008 to request possible use of Fund resources, owing to stresses in Hungarian financial markets, particularly the government debt market. However, as a non-euro-area member of the EU, Hungary was required under EU treaties to consult with the EC and the Economic and Financial Committee (EFC) on its balance of payments needs before seeking assistance from sources outside the EU that are subject to conditionality. Fund staff consulted immediately with EC staff and, in view of the severity and urgency of the situation, the EC agreed that parallel consultations could take place.

A Fund mission arrived in Hungary on October 13, 2008 and was subsequently joined by an EC team. During the negotiations, both teams operated and coordinated efforts to proceed at the same pace. As required under the Fund’s exceptional access policy, an informal Executive Board meeting was held on October 28. Once staff-level understandings had been reached, coordinated IMF-EC announcements were made before financial markets opened on October 29. Both the EC and IMF teams attended a press conference in Budapest, organized by the authorities. On November 4, 2008, the staff report for the SBA request under the exceptional access policy was issued to the Fund’s Executive Board; this request was approved on November 6, 2008 under the Emergency Financing Mechanism. Thus, the elapsed time from the authorities’ call to IMF disbursement was less than one month—which is very fast.

The BOP financing gap identified for the program period (17 months) was €20 billion. This gap was filled by commitments from the EU (€6.5 billion or 32.5 percent of the gap), World Bank (€1.0 billion, or 5 percent of the gap), and the IMF (€12.5 billion, or 62.5 percent of the gap). The IMF program was heavily front-loaded; the initial purchase was €4.2 billion or almost one-third of the IMF total. This financing pattern reflected in large part EU constraints. For example, when the IMF and EC teams were in the field, the size of the EU’s BOP financial assistance facility was only €12 billion. It was recognized that the facility was too small. Therefore, on November 4, the European Council authorized an increase in its size to €25 billion, while also granting a loan to Hungary of €6.5 billion. This Council decision stated that the first EU disbursement (€2 billon) to Hungary would be released once a memorandum of understanding (MOU), which would lay out the EU’s policy conditions, was signed by the EU and Hungary. The MOU was finalized on November 19, 2008 and the first EU disbursement took place on December 9, 2008, or only about a month after the first purchase from the IMF. Under the circumstances, this pace was fast.1

While the IMF and EC teams were in Hungary, the ECB was engaged in separate discussions with the Hungarian Central Bank (MNB) on a repo line. The IMF team discussed with their Central Bank counterparts the workings of this repo line, including its collateral requirements. As recorded in the LOI (and MOU), the MNB established on October 16, 2008 a foreign exchange swap facility, which would be supported by a repo facility with the ECB amounting to €5 billion, to improve liquidity in domestic financial markets. This euro provision promised by the ECB was not counted toward filling the financing gap, because it was viewed as largely a domestic monetary operation and because of uncertainties considering its drawdown. Later, a foreign exchange swap line replaced the repo line.

The 2008 staff report for Hungary’s SBA request (IMF, 2008) recognized (in its Box 1) the precedents that were being established for cooperation between the IMF and EC in a joint lending context. Specifically, the box observed that, “prior to the recent events in Hungary, no operating procedures had been developed for such interaction between the EU and the IMF. The process as developed in the case of Hungary could, however, become a reference on how to proceed should further cases of a similar nature arise. . . .” The box recorded five key principles: (i) early consultation and ongoing information exchanges during the program negotiations; (ii) contributions of both institutions to financing needs; (iii) joint announcement to underline broad support; (iv) consistency of program design and conditionality; and (v) consultation during the program monitoring process.

For our purposes, this box made two important elaborations. One, it was expected that EC conditionality (to be included in the MOU (yet to be written)) would be consistent with IMF conditionality and that, in particular, EC surveillance mechanisms such as the Excessive Deficit Procedure would incorporate the policy commitments made by the Hungarian authorities. This sequencing implied that the Fund-supported program effectively determined fiscal targets, but this did not take place later in the program period (see below). Two, if program deviations occurred, the authorities were to inform in parallel the EU and IMF and both institutions would coordinate closely during the related discussions. This second principle would be first tested in Latvia. The two principles were supported by the ECB observer at the Fund Board meeting, saying that “given that the EU has its own policy and instrument framework, conditionality of the IMF has to be reflected or mapped onto our own requirements in terms of, for example, an update of the convergence program, in terms of the excessive deficit procedure and also in terms of the national reform program.”

At the Board meeting on Hungary’s SBA request, Executive Directors generally did not comment on IMF-EC cooperation. Two Executive Directors, however, welcomed this cooperation and no Director expressed any reservations. Two Directors did question the lack of specificity on the conditionality to be imposed by the EC (and World Bank).

EC-IMF cooperation in the case of Hungary has been deemed successful by all parties (see, for example, IMF, 2011b). All program reviews were completed together. Both programs moved to a precautionary mode at the same time. The final two reviews of the IMF and EU programs were not completed owing to policy disagreements with the authorities. Cooperation was facilitated by several modalities. Frequent communications took place via telephone and email by the country teams for the EC and IMF. In addition, the IMF staff shared their draft policy note, formerly the briefing paper, with EC (and World Bank) staff at the same time as it was circulated to Fund departments. EC staff provided written comments electronically, while Bank staff could attend the policy consultation meeting. Meetings could take place in Brussels prior to reaching Budapest to clarify any outstanding points. The IMF, EC, and ECB staff attended meetings together with the country authorities. IMF staff shared their spreadsheets, programming expertise, and cross-country crisis experience, while EC staff shared their greater knowledge of EU policies and practices, particularly as it applied to Hungary. The two teams worked together in Budapest on their respective policy-intentions documents—the LOI and MOU. Starting with the first program review (March 2009), these two documents were signed by the authorities on the same date. This practice helped ensure consistent conditionality (even though EC conditionality was more detailed than IMF conditionality especially in the fiscal and structural areas) and more rapid EU disbursement.

Areas of difference or light friction emerged on occasion, particularly related to fiscal monitoring and the fiscal stance. As regards fiscal monitoring, the IMF’s performance criteria were set on the primary cash balance of the central government, while the EU fiscal benchmark was set on the overall accrual balance of the general government—using the EDP, or ESA-95, definitions. Thus three definitional differences existed—primary vs. overall balance, cash vs. accrual, and central vs. general government. The IMF definition ensured timelier reporting, while the EC definition was more comprehensive and consistent with EU obligations. The IMF and EC teams took steps to lessen associated communication and signaling risks; IMF fiscal targets as reported in the staff reports were consistent with the EC’s concept of the overall deficit of the general government, while EC monitoring took into account progress in achieving the IMF’s cash-flow deficit target for the central government.

Friction developed over the fiscal stance during 2009 because Hungary’s real GDP contracted by more than projected (by 6.7 percent compared with the projected 1.0 percent), placing pressure on the fiscal deficit to exceed the SGP limit of 3 percent of GDP. IMF staff advocated allowing automatic stabilizers to operate in order to cushion aggregate demand. The second program review was completed on the basis of additional (pro-cyclical) fiscal measures that offset about half of the automatic stabilizers and a two-year extension in the time to reach the EDP target. The IMF staff report recommended that if economic activity contracted by more than was currently envisaged, automatic stabilizers should accommodate fully. The EC MOU did not contain similar language and, as reported in the EPE (IMF, 2011b), the authorities had a strong commitment to adhere to their EU convergence program and the EDP.


To fully appreciate the program design issues that arose in Latvia, it is necessary to understand the policy debate that raged starting in 2007 over the exchange rate peg. As documented by the IEO (Wagner, 2010), Article IV consultation staff reports for Latvia were increasingly “alarmist,” starting in 2004 and continuing in 2006, sending clear messages about overheating, massive imbalances, and financial sector vulnerabilities. But the 2007 Surveillance Decision with its emphasis on external instability (and the associated “labeling”—fundamental exchange rate misalignment) created a rift on its application and led Fund management not to issue to the Executive Board the draft 2007 Article IV staff report on Latvia. Nevertheless, the Board did receive a Financial Sector Assessment Program update on Latvia that warned of serious threats to systemic financial stability and called for a strengthening of Latvia’s contingency framework, and a selected issues paper that concluded that Latvia’s real effective exchange rate appeared to be significantly overvalued—by some 20 percent.

In early September 2008, the Executive Board was issued an Article IV consultation report on Latvia that observed a much-needed slowing in domestic demand but noted that significant concerns still existed regarding external stability. In addition, an FSAP update supplement judged that the downside risks had risen, owing to the domestic economic slowdown and fragile global liquidity conditions. The revised staff estimates for exchange rate overvaluation ranged from 16 percent to 37 percent, averaging 27 percent. The staff was of the view that adjustment under the prevailing exchange rate peg remained the preferred option but entailed risks. Because the staff could not yet make a determination regarding fundamental misalignment, it recommended an ad hoc consultation with another Board discussion in about six months. The Board meeting on this staff report, which was scheduled for September 22, 2008, never took place. Relatedly, a Board meeting for the 2008 Article IV consultation with China was scheduled for September 26, 2008; in its staff report, China’s real exchange rate was considered to be substantially undervalued, but staff was “not yet making a determination regarding specific findings under the 2007 Surveillance Decision” and likewise recommended an ad hoc consultation in about six months. The Executive Board never discussed this staff report on China. These two Article IV consultations were delayed owing to “ongoing internal discussions on the implementation of the 2007 Surveillance Decision” (IEO, 2011; Blustein, 2013).

From a European perspective, IMF surveillance posed a complication because Latvia was an ERM2 member. According to the ERM2 operating procedures (March 2006), the ECB “shall closely monitor, on a permanent basis, the sustainability of bilateral exchange rate relations between each participating non-euro area currency and the euro” and it “shall have the right to initiate a confidential procedure aimed at reconsidering central rates.” At the same time, the EC was responsible for monitoring and assessing Latvia’s progress toward euro adoption, employing the Maastricht convergence criteria. Moreover, the Latvian authorities disputed the IMF staff’s analysis and assessments and were supported by their IMF Executive Director. There therefore was a risk that IMF surveillance might call into question aspects of the peer review conducted by EU institutions.

On November 15, 2008, a deposit run on Parex Bank turned into a speculative attack on the currency peg. The Latvian authorities requested financial assistance from the EC and the IMF in mid-November. An informal Board session to activate the emergency procedures was held on November 17, 2008, while preliminary talks (with the IMF/EC/ECB) took place in Latvia during November 17–23. The ECB joined this mission as an observer, owing to Latvia’s ERM2 membership. The IMF staff reaffirmed its estimates of real exchange rate misalignment (about 30 percent), while the EC’s and ECB’s estimates were at, or below, 10 percent. Any change in the peg was strongly opposed by the Latvian authorities, the EC, the ECB, and Sweden. The latter three parties worried about possible contagion to other EU members and to Swedish banks. Immediate euro adoption after a parity change was ruled out by the EC and ECB. The Latvian authorities also strongly desired to maintain their euro peg unchanged. According to Fund staff, because of the EU’s key role in overall policy design, including on the exchange rate strategy, the EC’s financing contribution was expected to substantially exceed that of the IMF. Thus, the EU’s key role in program design—supported by the Latvian authorities—led to the EU’s predominant share in the financing package rather than to the EU’s financing contribution giving it sway in the program’s design.2

The staff report requesting a three-year SBA for Latvia was issued to the Board on December 19, 2008 and the Board meeting was held three days later on December 23.3 With unusual candor, the staff made clear that a change in parity and immediate adoption of the euro had been discussed, but that this “technically more attractive” option had been “firmly ruled out” by the EU authorities and strongly opposed by the Latvian authorities. The envisaged fiscal adjustment was substantial and front-loaded (7 percentage points of GDP), and was expected to produce a moderate contraction of real GDP, owing to positive but limited fiscal multipliers; the inability of households to borrow to smooth consumption spending; and the lack of a monetary policy to offset fiscal tightening. The SGP limit (3 percent of GDP) was targeted for 2011, or the last year of the three-year program.

Latvia’s gross financing requirement was estimated at €15 billion. However, the program contained commitments from Swedish (and another Nordic country) parent banks to maintain their exposure, while the program assumed rollover rates of 40 percent for other foreign creditors. Consequently, the net financing gap was lowered to €7.5 billion, or about 50 percent of GDP. This financing gap was to be filled by the EC (€3.1 billion, or 41 percent of the total gap), four Nordic countries (€1.8 billion, or 24 percent), the IMF (€1.7 billion, or 23 percent), three European emerging market countries (€0.5 billion, or 7 percent), and the IBRD/EBRD (€0.4 billion, or 5 percent). Thus, the ratio of other financing to IMF resources was 3 to 1; the IMF share was only 24 percent for Latvia compared with 65 percent for Hungary. Nonetheless access relative to Fund quota was somewhat higher in Latvia (1,200 percent) than in Hungary (1,015 percent).

As in Hungary, the initial IMF purchase was heavily front-loaded—constituting nearly one-third of the entire arrangement—to allow Latvia to take out a bridge loan granted by Nordic central banks. The EC loan was similarly front-loaded (e.g., about one-third of the total in the first disbursement). The EC MOU was signed on January 28, 2010 and the first disbursement occurred on February 25, 2010. The EC MOU was signed about one month after the IMF LOI was signed (a time difference similar to that for the first EC MOU with Hungary).

As documented in the EPE for Latvia, the IMF worked closely with the various financing partners who also contributed technical expertise. The EC was heavily involved in the fiscal and financial sectors, as reflected in the MOU conditionality. The World Bank provided inputs on social safety nets and the legal framework for the financial sector, while the EBRD tackled the resolution of Parex Bank; both provided sectoral financing in these areas. In addition, interviews demonstrated that frequent communications took place via telephone and email between the country teams for the EC and IMF. The IMF, EC, and ECB staff (as did other partners) attended meetings together with the country authorities.

At the time of the first program review mission (February 2009), there were clear signs that output was contracting much deeper than envisaged and that the program was off track. Mass demonstrations against the program took place. As the governing coalition was blamed for Latvia’s economic problems, the Prime Minister resigned in late February. The review mission left an aide-mémoire to help the incoming government to identify measures to bring the program back on track. During this pause, changes were made in the team leaders for both the IMF and EC. These new team leaders visited the newly elected government in April. It was agreed that a supplementary budget would be sent to Parliament after municipal/EU elections in early June. Deposit outflows and loss of international reserves placed pressures on the currency peg, triggering under the IMF program the need for a discussion on contingency plans with the Latvian authorities. Technical assistance on revenue administration and public expenditure management from the IMF continued in April and May to assist the Latvian authorities in preparing a supplementary budget.

Efforts to complete the first program review resumed in late May 2009 between the Latvian authorities and the IMF, EC, and ECB along with representatives from the World Bank, EBRD, and Nordic countries. Output was now expected to fall by 18 percent in 2009 compared to the program projection of a 5 percent decline. This far deeper recession increased the projected budget deficit to 16–17 percent of GDP—far exceeding the program target of 5 percent. In addition, Latvia was seen to be at risk of running out of money, as its international reserves had declined by 25 percent since end-2008. On June 16, Parliament approved a supplementary budget of the full-year equivalent of 13 percentage points of GDP and containing measures that included cuts in pensions and social benefits. The IMF staff was concerned about the adverse growth implications and the effect of the measures on vulnerable groups. The EC (and the Nordics) were worried that the currency peg would not be able to withstand further delays in disbursements, seeing the IMF as too willing to risk a currency crisis to obtain improvements in fiscal policy.

In the event, the European Council met on June 19, 2009 and concluded that it supported “the adoption of the new budgetary measures in Latvia aiming at sizeable fiscal consolidation this and next year. It stresses that rigorous implementation of the measures adopted together with credible medium-term strategy will deliver a successful outcome of the current adjustment programme. The European Council strongly supports the intention of the Commission to propose the swift disbursement of the next installment of the Community balance-of-payments assistance in the framework of the adjustment programme.” On June 24, the EC sent a note to the ECOFIN Council proposing disbursement of the second EU installment, and stating that “the Commission services carried out a review mission from 27 May to 17 June in close cooperation with the IMF staff (which undertook its first full review under their SBA in parallel). . . . Based on the findings of the above-mentioned review mission and additional available information, the economic policy criteria for the second installment as laid down in the MoU can be considered to be fulfilled.” In its review, the EC acknowledged that the supplementary budget contained poor-quality measures that would have negative distributional consequences and doubtful sustainability, but the EC felt that these deficiencies could be corrected in the 2010 budget. The formal decision was taken on July 2 and the second installment was disbursed later in July. The Fund mission was surprised that the EC would disburse without the IMF.

After high-level discussions between the IMF and EC, a joint IMF/EC/ECB mission along with the Nordic representative returned to Latvia on July 12, 2009 to complete the first IMF program review. The mission completed its work by end-July and the IMF staff report was circulated to the Board on August 7, 2009. In this report, the staff made clear that it preferred “a slightly higher budget deficit in 2009 to protect basic services and to rebalance the burden of adjustment, while preparing for structurally sound adjustment in 2010.” The staff also took issue with the rapid fiscal adjustment pace for 2010–12 that was proposed by the Latvian authorities and endorsed by ECOFIN. Unusually, an IMF “program scenario” was also presented with less rapid fiscal adjustment—reaching the Stability and Growth Pact target of 3 percent of GDP in 2014 rather than 2012—that was considered to be more credible and was projected to yield somewhat faster output growth. The IMF Executive Board completed the program review on August 27, 2009. Eight Executive Directors from EU countries supported the rapid fiscal adjustment strategy, while most other Directors supported the less rapid one.

The Latvian authorities’ bold upfront fiscal adjustment sparked a revival in market confidence (Giudice, 2012), easing liquidity pressures, and real growth resumed unexpectedly in the fourth quarter of 2009. Their strong program ownership continued to be exhibited in their 2010 budget formulation and implementation. A V-shaped recovery was increasingly in evidence during 2010, while the current account was in surplus. By early 2011, the stronger Latvian economy allowed the authorities to stop drawing on the amounts under both the IMF and EC programs.

These events arguably showed the IMF staff’s judgment to be in error regarding the sustainability of the currency peg, the feasible fiscal adjustment, and the implications of fiscal consolidation for real growth. On each issue, the EC staff supported the country authorities, which had a strong political commitment to joining the euro at the earliest possible date. Essentially, the EC seemingly got it right and the IMF got it wrong. The EC by deciding to go it alone without the IMF displayed its independence and confidence. While in retrospect the IMF seems to have made the right, albeit risky, decision of completing the first program review, this decision may have contributed to a perception of the IMF as a junior partner in this international rescue package. In addition, EC staff believed Latvia provided lessons—particularly as regards the confidence-enhancing role of fiscal consolidation and structural reforms—that could be applied to euro-crisis countries (Giudice, 2012; Deroose and others, 2010), while the IMF’s Economic Counsellor (Blanchard) blogged in 2012, “The sad truth is that many of these conditions [that led to Latvia’s V-shaped recovery] are not satisfied elsewhere. So the lessons are not easily exported.”


Severe pressures on Romania’s balance of payments became evident in late 2008, as the domestic economy overheated and access to foreign liquidity dried up with the global financial crisis. Recognizing these pressures, an IMF staff visit took place in Bucharest in late January/early February 2009 to assess the situation and provide policy advice. As the situation continued to worsen, preliminary discussions on a possible IMF-supported program were held in Washington in early March with the Romanian authorities. These discussions were quickly followed by a visit to Bucharest from an IMF/EC negotiating team. The IMF mission chief stopped in Brussels to coordinate with the EC before proceeding to Bucharest. Teams from the EBRD, World Bank, International Finance Corporation (IFC), and European Investment Bank (EIB) also flew to Bucharest. An informal IMF Board meeting under the exceptional access policy was held on March 13, 2009.

A staff-level agreement on a program for Romania was concluded on March 25, 2009. Cooperation among all parties (IMF, EC, WB, EBRD, and EIB) was close throughout the negotiations although complicated by their numbers. In addition, from an IMF perspective the EC’s occasional internally conflicting objectives and cumbersome decision-making procedures were viewed as hurdles. On a more positive note, the EBRD, EC, IMF, WB, and EIB organized the first country meeting under the newly created European Bank Coordination Initiative on March 26, 2009. This meeting was attended by the nine largest foreign-owned banks incorporated in Romania; their parent banks; the National Bank of Romania; representatives of the home country authorities (Austria, France, Greece, and Italy); and an observer from the ECB. These banks committed to maintain their exposure to Romania and to recapitalize their subsidiaries as needed following stress tests.

The staff report to request a 24-month SBA (IMF, 2009c) was circulated to the IMF Board on April 24, 2009 and the Board meeting was held on May 4. Owing largely to the bank exposure agreement, the gross financing requirement (€44 billion) was reduced to a financing gap of €20 billion. This gap was filled by the IMF (€13 billion, or 65 percent of the total gap), the EC (€5 billion, or 25 percent), the World Bank (€1 billion, or 5 percent) and the EBRD, EIB, and IFC (collectively €1 billion, or 5 percent). These contribution shares were closer to those observed for Hungary than for Latvia. The IMF contribution was equivalent to 1,111 percent of quota—similar in quota size to Hungary and Latvia—and required use of the exceptional access policy. As in Hungary and Latvia, the SBA for Romania was heavily front-loaded with the initial purchase (€5 billion) representing 38 percent of the total. The EC’s MOU with Romania was signed on June 23, 2009 and the first installment of the EC loan (€1.5 billion or 20 percent) was released on July 27.

As regards the design of program policies, the IMF took center stage in framing the macroeconomic stance, identifying the financing requirement, and coordinating the institutional players. A key feature of the macroeconomic policy design was the front-loaded fiscal tightening in 2009 (with measures equivalent to 3 percentage points of GDP) to tackle overheating and to establish a credible path toward fiscal viability, seeking to reduce the fiscal deficit below the Maastricht target in 2011.4 Passage by Parliament of the 2011 budget was a structural benchmark under both the IMF SBA and the EC MOU. As in Hungary, monetary policy in Romania was conducted under an inflation-targeting regime with a floating exchange rate. The IMF team negotiated this aspect of the program; monetary policy did not feature in any of the EC’s MOUs. Addressing vulnerabilities in the financial sector was another major program element; here the EC was heavily involved, ensuring consistency with EU directives.

At the time of the first program review mission (August 2009), output contraction was projected to be more severe than initially envisaged (8 percent in 2009 compared with 4 percent originally) with a sharper 2010 recovery anticipated (1¾ percent compared with zero originally). The IMF team proposed to accommodate the bulk (80 percent) of the projected cyclical deterioration in the fiscal deficit in 2009; the additional fiscal adjustments for 2009 and 2010 were 0.6 percentage points of GDP and 0.4 percentage points, respectively. The 2010 fiscal effort came on top of already programmed measures equivalent to 1.4 percentage points of GDP. In the view of Fund staff, Romania’s ability to achieve the Maastricht target in 2011 would depend on the strength of the recovery. The EC advocated a larger fiscal adjustment effort in order to maintain the scheduled date (2011) for achieving the Maastricht target that had been endorsed by the ECOFIN Council. (The public debt ratio was projected to rise to only 34 percent of GDP in 2011, or well below the relevant Maastricht value.)

After discussions with the Romanian authorities by the IMF and EC teams, the output contraction in 2009 deepened to 8½ percent, while the projected 2010 recovery was reduced to ½ percent. To counter automatic fiscal stabilizers, the Romanian authorities adopted fiscal measures equivalent to 0.8 percentage points of GDP (or about 2 percentage points annualized) in 2009 and 2.0 percentage points in 2010. These consolidation efforts were somewhat larger than specified in the IMF’s policy note. The IMF’s LOI and staff report for the first program review contained no mention of the fiscal target for 2011, although a staff report table contained a figure of –4.2 percent of GDP, exceeding the corresponding Maastricht value. As no EU disbursement was scheduled to correspond with the IMF review/purchase, the EC team did not complete a formal review under the EU’s financial assistance program.

Owing to political tensions within the governing coalition, related to presidential elections, the government fell in October 2009. As the “caretaker” government could not adopt a 2010 budget, program reviews were delayed until after the presidential elections held on December 23, 2009. In January 2010, IMF and EC teams returned Bucharest to complete the relevant program reviews. As high-frequency data indicated stronger-than-expected real growth starting in the fourth quarter of 2009, the estimated 2009 output contraction was lessened to 7 percent and an output expansion of 1¼ percent was projected for 2010 (compared to zero at the time of the first review). Nonetheless, pro-cyclical fiscal consolidation measures of 2½ percentage points of GDP, or ½ percentage point more than envisaged at the first review, were agreed in order to maintain unchanged the deficit target of 5.9 percent of GDP for 2010. The Romanian authorities expressed their intention in the IMF LOI (dated February 5, 2010) to reduce the fiscal deficit below the Maastricht limit of 3 percent of GDP by 2012. The Fund staff estimated that to achieve this target, the authorities would need to implement additional fiscal measures equivalent to 1¼ percentage points of GDP in both 2011 and 2012. On this basis, the IMF Board completed both the second and third reviews under the SBA on February 19, 2012.

The EC MOU was signed on February 22, 2010. The ECOFIN Council on February 16, 2010 (or three days before the IMF Board met) endorsed the revised gradual fiscal adjustment strategy agreed by the EC team, including the extension by one year—to 2012—for dipping below the Maastricht limit of 3 percent of GDP for the fiscal deficit. The MOU did not contain an explicit fiscal deficit target for 2011, but it did cite the need to be consistent with achieving a deficit below 3 percent of GDP in 2012. Parliament’s passage of a draft fiscal responsibility law, prepared with input from the EC, IMF, and World Bank, was a structural benchmark for both the Stand-By Arrangement and the MOU. Once again, monetary policy was not discussed in the EC MOU. EU-consistent reforms in the financial sector, prepared with technical assistance from the EC and IMF, were also benchmarks under both programs. The EC disbursed its second installment (€1 billion) on March 11, 2010.

At the time of the fourth program review in mid-2010, Romania’s real growth for 2010 was revised downwards to a contraction of 0.5 percent (from +0.8 percent). The fiscal deficit was projected, based upon unchanged policies, to reach 9.1 percent of GDP compared to the programmed 5.9 percent of GDP. Consequently, fiscal measures—primarily on the spending side—equivalent to 4½ percentage points of GDP annualized, were implemented, while the fiscal target for 2010 was lifted to 6.8 percent of GDP. The IMF projected that the fiscal deficit would slip below the Maastricht reference value in 2014. With lower core inflation and fiscal tightening, the central bank reduced interest rates and the IMF staff saw room for further reductions going forward. The IMF Board completed the IMF review on July 2, 2010. The corresponding supplemental MOU was signed on August 2, 2010 and the disbursement took place on September 22, 2010. This MOU revised upwards the fiscal deficit target for 2010 in a manner consistent with the IMF target, but the supplemental MOU did not change, or even mention, the fiscal deficit targets for 2011 or 2012. As a consequence, an implicit, though perhaps not meaningful, difference in fiscal targets emerged.

Though an EC team accompanied the IMF team to Bucharest as part of the SBA review mission, no corresponding EU program review took place. The IMF team revised further downward its growth estimates for 2010 to –1.9 percent, but projected a very sharp recovery in 2011 and 2012 (of 1.5 percent and 4.4 percent, respectively). Notwithstanding the deeper contraction in 2010, the IMF staff and authorities considered the fiscal target (–6.8 percent of GDP) to be within reach. The IMF Board paper did not contain a table with 2012 fiscal projections, but the text stated that “robust economic recovery and continued expenditure restraint could make the achievement of the Maastricht fiscal target feasible by 2012 without further major adjustment measures.” Thus, the differences in the fiscal targets for 2011–12 that had emerged at the time of the fourth IMF SBA review had disappeared by the time of the fifth review, avoiding any possible complication for the EU disbursement on September 22. The fifth IMF SBA review was completed by the IMF Board on September 24, 2010.

From a troika standpoint, the remaining six months of Romania’s program hold little interest. The reviews were completed as scheduled by the IMF and EC. The Fund Board completed the final program review in March 2011; the Romanian authorities decided not to draw upon the last purchase made available under the IMF SBA, treating it as precautionary. At the same Board meeting (March 2011), the authorities requested cancellation of the existing SBA and the Board approved a new 24-month SBA, which the authorities intended to treat as precautionary. The final two installments of the EU loan were disbursed on March 24, 2011 and June 22, 2011. The EU also entered into a new BOP assistance program with Romania on a precautionary basis. Although the new EU loan was announced in March 2011, the MOU was signed on June 29, 2011, after the final disbursement of the previous loan.

Overall, cooperation between the EC and IMF was successful and effective, as noted in the IMF staff’s EPE for Romania (IMF, 2012a). According to the IMF staff, the cooperation details were similar to those reported in the EPE for Hungary, which were deemed effective by all parties. As in Hungary and Latvia, the IMF team tended to prefer a more gradual path of fiscal adjustment than was spelled out by the Excessive Deficit Procedure and was more willing than EC staff to adjust the fiscal deficit targets upward in response to unanticipated lower economic activity, allowing the automatic fiscal stabilizers to operate. The two teams sorted out these differences among themselves. Both teams were well aware of the different definitions used for their respective fiscal targets and the associated compliance/signaling risks. These risks did not materialize. EU structural funds were not absorbed as programmed, owing to bureaucratic barriers in both Romania and the EU—creating tensions, wasting time, and complicating the conduct of demand management. On the other hand, the IMF, EC, and ECB participated effectively and smoothly in the European Bank Coordination Initiative, or Vienna I/II.


    AslundAnders and ValdisDombrovskis2011How Latvia Came Through the Financial CrisisMay (Washington: Peterson Institute for International Economics).

    BastasinCarlo2012Saving Europe: How National Politics Nearly Destroyed the Euro (Washington: Brookings Institution Press).

    BeattieAlan2011aIMF Poised to Take a Backseat Role in Greece RescueFinancial TimesJuly22.

    BeattieAlan2011bSpain and Italy Pose Conundrum for IMFFinancial TimesAugust4.

    BernesThomas2014IMF Leadership and Coordination Roles in the Response to the Global Financial and Economic CrisisIEO Background Paper No. BP/14/06 (Washington: International Monetary Fund).

    BlanchardOlivier2012Lessons from LatviaIMF blog iMFdirectJune11. Available at:

    BlanchardOlivierMarkGriffiths and BertrandGruss2013Boom, Bust, Recovery: Forensics of the Latvia CrisisBrookings Papers on Economic Activity Fall (Washington: Brookings Institution).

    BlusteinPaul2013Off Balance: The Travails of Institutions that Govern the Global Financial SystemOctober (Waterloo, Ontario: Centre for International Governance Innovation).

    BlusteinPaul2015aOver Their Heads—the IMF and the Prelude to the Euro-Zone CrisisCIGI Paper No. 60March (Waterloo, Ontario: Centre for International Governance Innovation).

    BlusteinPaul2015bLaid Low—The IMF, The Euro Zone, and the First Rescue of GreeceCIGI Paper No. 61April (Waterloo, Ontario: Centre for International Governance Innovation).

    BoughtonJames2001The Silent Revolution: The International Monetary Fund 1970–1989 (Washington: International Monetary Fund).

    BoughtonJames2012Tearing Down the Walls: The International Monetary Fund 1990–1999 (Washington: International Monetary Fund).

    BovaElvaNathalieCarcenac and MartineGuerguil2014Fiscal Rules and Procylicality of Fiscal Policy in the Developing WorldIMF Working Paper No. 14/122July (Washington: International Monetary Fund).

    Chopra Ajai2015Witness Statement before the Joint Committee of Inquiry into the Banking Crisis. Available at:

    ChowlaPeter2011Christine Lagarde as IMF Chief? This is a Gift to the Fund’s CriticsThe GuardianJune29.

    CogganPhilip2012Less Cash, More Impact: The IMF Needs to Change Tack if Spain Asks for a BailoutThe EconomistOctober6.

    CottarelliCarloLorenzoForniJanGottschalk and PaoloMauro2010Default in Today’s Advanced Economies: Unnecessary, Undesirable, and UnlikelyIMF Staff Position Note No. 10/12September (Washington)

    De las CasasMiguel2017The IMF Executive Board and the Euro Area Crisis: Accountability, Legitimacy, and Governance(Chapter 1 in this volume).

    DerooseServaasElenaFloresGabrieleGiudice and AlessandroTurrini2010The Tale of the Baltics: Experiences, Challenges Ahead, and Main LessonsECFIN Economic Brief No. 10July (Brussels: Directorate-General for Economic and Financial Affairs, European Commission).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)1990European Economy: One Market, One Money: An Evaluation of the Potential Benefits and Costs of Forming an Economic and Monetary UnionNo. 44October (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2006The EU Economy: 2006 Review—Adjustment Dynamics in the Euro Area—Experience and ChallengesNo. 6 (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2008Memorandum of Understanding between the European Commission and HungaryDecember (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2009aMemorandum of Understanding between the European Commission and LatviaJanuary (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2009bMemorandum of Understanding between the European Commission and RomaniaMarch (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2010aRecommendation for a Council Opinion on the Updated Stability Programme for Greece, 2010–13February (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2010bAction Taken by Greece in Response to the Council Decision of 16 February, 2010 and to Implement Its Stability ProgrammeMarch (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)_2010cStatement by the Heads of Government and State of the Euro AreaMarch (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2010dThe Economic Adjustment Programme for GreeceOccasional Papers No. 61May (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2011aThe Economic Adjustment Programme for IrelandOccasional Papers No. 76February (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2011bThe Economic Adjustment Programme for PortugalOccasional Papers No. 79June (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2011cStatement of the Heads of Government and State of the Euro AreaJuly (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2012The Financial Sector Adjustment Programme for SpainOccasional Papers No. 118October (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2013The Economic Adjustment Programme for CyprusOccasional Papers No. 149May (Brussels: DG-ECFIN).

    Directorate-General for Economic and Financial Affairs European Commission (DG-ECFIN)2015Ex Post Evaluation of the Economic Adjustment Programme—Ireland 2010–2013Institutional Paper 004, July (Brussels: DG-ECFIN).

    DonovanDonal2017The IMF’s Role in Ireland(Chapter 8 in this volume).

    EichenbaumMartinSergioRebelo and Carlos deResende2017The Portuguese Crisis and the IMF(Chapter 9 in this volume).

    EichengreenBarry and NgaireWoods2016The IMF’s Unmet ChallengesJournal of Economic Perspectives30 (1).

    European Central Bank (ECB)2014ECB’s Replies to the Questionnaire of the European Parliament Supporting the Own Initiative Report Evaluating the Structure, the Role and Operations of the ‘Troika’ (Commission, ECB, and the IMF) Actions in Euro Area Programme Countries” (Frankfurt).

    European Central Bank (ECB)2014Monthly ReportDecember (Frankfurt).

    European Court of Auditors (ECA)2015Financial Assistance Provided to Countries in DifficultiesSpecial Report No. 18 (Luxembourg: European Union).

    European Court of Justice (ECJ)2012Case C-370/12: Thomas Pringle v. Government of IrelandNovember (Luxembourg: European Court of Justice).

    European Court of Justice (ECJ)2015aThe OMT Programme Announced by the ECB in September 2012 is Compatible with EU LawECJ Press Release 70/15June15 (Brussels).

    European Court of Justice (ECJ)2015bCase C-62/14: “Decisions of the Governing Council of the European Central Bank (ECB) on a Number of Technical Features Regarding the Eurosystem’s Outright Monetary Transactions in Secondary Sovereign Bond MarketsJune (Luxembourg: European Court of Justice).

    EuropeanParliamentCommittee on Economic and MonetaryAffairs2014aReport on the Enquiry on the Role and Operations of the Troika (ECB, Commission, and IMF) with Regard to the Euro Area Programme CountriesA7-0149/2014February (Brussels: European Union).

    European Parliament Committee on Economic and Monetary Affairs2014bCommittee Study on the Troika and Financial Assistance in the Euro Area: Successes and Failures” prepared by A.SapirG. Sousa and A.TerziFebruary (Brussels: European Union).

    European Stability Mechanism (ESM)2012Treaty Establishing the European Stability MechanismFebruary (Luxembourg).

    Financial Times2010Greece Rescue Fills in Bailout BlanksApril13.

    GiudiceGabriele2012Against the Odds: Lessons from the Recovery in the Balticsspeech given at conference held in Riga Latvia on June 5 2012. Available at:

    Group of Twenty (G20)2011G20 Principles on the Cooperation between the IMF and Regional Financing Arrangements as Endorsed by the G20 Finance Ministers and Central Bank GovernorsOctober.

    HenningC. Randall2011Coordinating Regional and Multilateral Financial InstitutionsWorking Paper 11-9March (Washington: Peterson Institute for International Economics).

    HitajErmal and Yasin KursatOnder2013Fiscal Discipline in WAEMU: Rules, Institutions, and MarketsIMF Working Paper No. 12/216October (Washington).

    Ireland Government of2010aThe Irish Banking Crisis: Regulatory and Financial Stability Policy 2003–2008Report to the Minister for Finance by the Governor of the Central BankMay.

    Ireland Government of2010bA Preliminary Report on The Sources of Ireland’s Banking Crisis” prepared by K.Regling and M.WatsonMay.

    Ireland Government of2011Misjudging Risk: Causes of the Systemic Banking Crisis in IrelandReport of the Commission of Investigation into the Banking Sector in IrelandMarch.

    Independent Evaluation Office of the International Monetary Fund (IEO)2011IMF Performance in the Run Up to the Financial and Economic Crisis: IMF Surveillance in 2004–07 (Washington: International Monetary Fund).

    Independent Evaluation Office of the International Monetary Fund (IEO)2014The IMF Response to the Financial and Economic Crisis: An IEO Assessment (Washington: International Monetary Fund).

    International Monetary Fund (IMF)1995CFA Franc Countries: Recent Adjustment Experience and Policy IssuesIMF Policy PaperOctober (Washington).

    International Monetary Fund (IMF)1998aThe European and Monetary Union and the International Monetary Fund—Main Legal Issues Relating to Rights and Obligations of EMU Members in the FundSM/98/131June (Washington).

    International Monetary Fund (IMF)1998bEMU and Fund SurveillanceSM/98/215August (Washington).

    International Monetary Fund (IMF)1998cEuropean Economic and Monetary Union—Surveillance; Use of Fund Resources and Use of Euros in Operational Budget—Preliminary ConsiderationsEBS/98/132December (Washington).

    International Monetary Fund (IMF)2005Fund Surveillance over Members of Currency UnionsIMF Policy PaperDecember (Washington).

    International Monetary Fund (IMF)2006Board Decisions on Surveillance of CEMAC and ECCUJanuary (Washington).

    International Monetary Fund (IMF)2008Hungary: Request for a Stand-By ArrangementIMF Country Report No. 08/361November (Washington).

    International Monetary Fund (IMF)2009aLatvia: Request for a Stand-By ArrangementIMF Country Report No. 09/3January (Washington).

    International Monetary Fund (IMF)2009bRepublic of Poland: Arrangement Under the Flexible Credit Line—Staff Report; Staff Supplement, Press Release, and Statement by the Executive Director for the Republic of PolandIMF Country Report No. 09/138May (Washington).

    International Monetary Fund (IMF)2009cRomania: Request for a Stand-By ArrangementIMF Country Report No. 09/103June (Washington).

    International Monetary Fund (IMF)2009dReview of Crisis ProgramsIMF Staff Policy Paper (SM/09/246; 9/14/09)September (Washington).

    International Monetary Fund (IMF)2010aGreece: Request for Stand-By ArrangementIMF Country Report No. 10/110May (Washington).

    International Monetary Fund (IMF)2010bAntigua and Barbuda: Request for a Stand-By ArrangementJune (Washington).

    International Monetary Fund (IMF)2010cSafeguards Assessments: Review of ExperienceIMF Policy Paper 070110July (Washington).

    International Monetary Fund (IMF)2010dIreland: Request for an Extended ArrangementIMF Country Report No. 10/366December (Washington).

    International Monetary Fund (IMF)2010eEx Post Evaluations of Exceptional Access Arrangements—Revised Guidance NoteIMF Policy PaperFebruary (Washington).

    International Monetary Fund (IMF)2011aPortugal: Request for a Three-Year Arrangement under the Extended Fund FacilityIMF Country Report No. 11/127June (Washington).

    International Monetary Fund (IMF)2011bHungary: Ex Post Evaluation of Exceptional Access under the 2008 Stand-By ArrangementIMF Country Report No. 11/145June (Washington).

    International Monetary Fund (IMF)2011cGreece: Fourth Review under the Stand-By Arrangement and Request for Modification and Waiver of Applicability of Performance CriteriaIMF Country Report No. 11/175July (Washington).

    International Monetary Fund (IMF)2011d Greece: Fifth Review Under the Stand-By Arrangement Rephasing and Request for Waivers of Nonobservance of Performance Criteria” IMF Country Report No. 11/351December (Washington).

    International Monetary Fund (IMF)2012aRomania: Ex Post Evaluation of Exceptional Access under the 2009 Stand-By ArrangementIMF Country Report No. 12/64March (Washington).

    International Monetary Fund (IMF)2012b2011 Review of Conditionality—Overview PaperIMF Policy Paper (SM/12/148; 6/18/12) plus four background papersJune (Washington).

    International Monetary Fund (IMF)2012cCrisis Program ReviewIMF Staff Presentation to Executive Directors circulated June 26 2012 as FO/DIS/12/106 (Washington).

    International Monetary Fund (IMF)2012dWorld Economic Outlook—Coping with High Debt and Sluggish GrowthOctober (Washington).

    International Monetary Fund (IMF)2012eSpain—Financial System Stability AssessmentIMF Country Report No. 12/138June (Washington). See also Minutes of Executive Board Meeting 12/55-1.

    International Monetary Fund (IMF)2012fIMF Says Spain’s Core Financial System is Resilient, but Important Vulnerabilities RemainIMF Press Release No. 12/212June (Washington).

    International Monetary Fund (IMF)2012g2012 Article IV Consultation with Spain: Concluding Statement of IMF MissionJune14 (Madrid) (Washington).

    International Monetary Fund (IMF)2012hIMF Managing Director Christine Lagarde Welcomes the Eurogroup’s Decision to Grant Financial Assistance to SpainIMF Press Release No. 12/266July (Washington).

    International Monetary Fund (IMF)2012iSpain—Terms of Reference for Fund Staff Monitoring in the Context of European Financial Assistance for Bank RecapitalizationFO/DIS/12/135July (Washington).

    International Monetary Fund (IMF)2012jGuidance Note for Surveillance under Article IV ConsultationsIMF Policy NoteOctober (Washington).

    International Monetary Fund (IMF)2013aLatvia: Ex Post Evaluation of Exceptional Access under the 2008 Stand-By ArrangementIMF Country Report No. 13/30January (Washington).

    International Monetary Fund (IMF)2013bStocktaking the Fund’s Engagement with Regional Financing ArrangementsIMF Policy PaperApril (Washington).

    International Monetary Fund (IMF)2013cSovereign Debt Restructuring—Recent Developments and Implications for the Fund’s Legal and Policy WorkIMF Policy PaperApril (Washington).

    International Monetary Fund (IMF)2013dCyprus: Request for an Arrangement under the Extended Fund FacilityIMF Country Report No. 13/125May (Washington).

    International Monetary Fund (IMF)2013eGreece: Ex Post Evaluation of Exceptional Access under the 2010 Stand-By ArrangementIMF Country Report No. 13/156June (Washington).

    International Monetary Fund (IMF)2013fSafeguard Assessments 2013 UpdateOctober (Washington).

    International Monetary Fund (IMF)2014Revised Operational Guidance to IMF Staff on the 2002 Conditionality GuidelinesIMF Policy NoteJuly (Washington).

    International Monetary Fund (IMF)2015aIreland: Ex Post Evaluation of Exceptional Access under the Extended ArrangementIMF Country Report No. 15/20 January (Washington).

    International Monetary Fund (IMF)2015bCrisis Program Review,” SM/15/277, 11/10/15; and “Summing UpBUFF/15/11 12/15/111 (Washington).

    International Monetary Fund (IMF)2016Strengthening the International Monetary System: A StocktakingIMF Policy PaperMarch (Washington).

    IrwinNeil2013The Alchemists: Three Central Bankers and a World on Fire (New York: Penguin Press).

    Kincaid G. Russell and Max Watson 2015”Avoiding Another Crisis in the Euro Area: Public and Private Imbalances and National Policy Responses” Journal of European IntegrationSeptember (Maastricht, Netherlands: Maastricht University).

    KopitsGeorge2017The IMF and the Euro Area Crisis: The Fiscal Dimension” (Chapter 6 in this volume).

    MandengOusmene2013The IMF Must Quit the Troika to SurviveFinancial TimesApril17.

    Pisani-FerryJeanAndréSapir and GuntramWolff2011aTRS External Study—An Evaluation of IMF Surveillance of the Euro AreaIMF Policy PaperJuly (Washington).

    Pisani-FerryJeanAndréSapir and GuntramWolff2011bEU-IMF Assistance to Euro Area Countries: An Early AssessmentBruegel Blueprint 19 (Brussels: Bruegel).

    SchadlerSusan2017Living with Rules: The IMF’s Exceptional Access Framework and the 2010 Stand-By Arrangement with Greece(Chapter 2 in this volume).

    SchaechterAndreaTidianeKindaNinaBudina and AnkeWeber2012Fiscal Rules in Response to the Crisis—Toward the ‘Next-Generation’ Rules: A New DatasetIMF Working Paper No. 12/187July (Washington).

    SchneiderKarsten2013Yes, But…One More Thing: Karlsruhe Ruling on the European Stability MechanismSpecial Section—The ESM before the Courts—of German Law Journal14 (11).

    SubramanianArvind2012The IMF Should Heed This ResignationFinancial TimesJuly26.

    TanLing Hui2017IMF Engagement with the Euro Area versus Other Currency Unions(Chapter 3 in this volume).

    Van MalleghemPieter-Augustijn2013Pringle: A Paradigm Shift in the European Union’s Monetary ConstitutionSpecial Section—The ESM before the Courts—of German Law Journal14 (11).

    VéronNicolas2016The IMF’s Role in the Euro Area Crisis: Financial Sector AspectsIEO Background Paper No. BP/16-02/10 (Washington: International Monetary Fund).

    WagnerNancy2010IMF Performance in the Run-Up to the Financial and Economic Crisis: Bilateral Surveillance in Selected IMF Member CountriesIEO Background Paper No. BP/10/03 (Washington: International Monetary Fund).

    WyploszCharles and SilviaSgherri2017The IMF’s Role in Greece in the Context of the 2010 Stand-By Arrangement(Chapter 7 in this volume).

    ZettelmeyerJerominChristophTrebesch and G. MituGulati2013The Greek Debt Restructuring: An AutopsyWorking Paper No. 2103-13-8August (Washington: Petersen Institute for International Economics).

The earliest press reference (in English) to the EC, ECB, and IMF as the troika appears to have been in June 2010—when the Greek media so dubbed these institutions. A troika is a vehicle drawn by three horses.

The troika did not negotiate Spain’s financial assistance because there was no parallel IMF lending program and thus no IMF conditionality. But because the IMF staff provided concurrent technical assistance to Spain and the EC, which resulted in joint troika-like missions, the case of Spain receives some attention.

Over time, staff from the European Financial Stability Facility (EFSF)/European Stability Mechanism (ESM) became more integrated with the IMF/EC/ECB missions, contributing financial analysis to debt-stability assessments prepared by both the IMF and EC staffs; however, the EFSF/ESM staff did not offer policy advice. Collectively, these four teams have been called in the press “the Quadriga,” which is Latin for a four-horse chariot. To simplify the presentation, the term troika is used throughout this chapter even when it refers to these four institutions.

Decision No. 11846 (98/125). Subsequently, these euro area surveillance modalities were extended to other currency unions (CEMAC, ECCU, and WAEMU). Surveillance modalities for currency unions were generalized as part of the 2007 Bilateral Surveillance Decision and then incorporated in the Integrated Surveillance Decision (2012).

The staff postulated that “if country-specific risks triggered a liquidity squeeze and thus pressures on interest rates in an individual union member, the union central bank or the national authorities (within the confines of their limited authority) might be prompted to take official action, if they perceive a risk to the prosperity of the individual member and/or of the union as a whole.” The union central bank could “intervene in the money or credit markets of the member, supplying liquidity or credit to residents (in the form of open market operations or other lending). In the case of EMU, prospective participants have been explicit in ruling out any such intentional intervention.” In the staff’s view, the central bank’s efforts to alleviate a liquidity squeeze in an individual country could be seen as an accommodating BOP flow induced by official action.

In particular, “Main Legal Issues Relating to Rights and Obligations of EMU Members in the Fund” (SM/98/131; 6/8/98) observed that an EMU member still had the right to request the use of Fund resources notwithstanding various provisions in the Maastricht Treaty. This legal paper also asserted that the Fund had the right to request a euro area member to impose capital controls in accordance with IMF Article VI, Section 1(a).

Issues related to balance of payments need, albeit not necessarily specific to currency unions, were examined a decade later (see “Review of Fund Facilities—Analytical Basis for Fund Lending and Reform Options” (2009), “The Fund’s Mandate—The Legal Framework” (2010), and “Staff Guidance on the Use of Fund Resources for Budget Support.”

The IMF did have considerable experience in conditional lending to 16 IMF members that were also members of the two CFA franc zones (e.g., CEMAC and WAEMU), or the ECCU. Notwithstanding this experience, the IMF’s Guidelines on Conditionality did not specifically address the implications for program design of membership in a currency union. Tan (2017) provides a more complete history and comparison of the IMF’s engagements with currency unions.

Technically, the EU only came into existence in 1993 (under the Maastricht Treaty). The IMF lent to Portugal in 1983, but at that time Portugal was not a member of the European Community, joining in 1986.

Subsequently, the EU created an instrument to grant precautionary BOP assistance; Romania was the first EU case in 2011. The IMF also approved a precautionary Stand-By Arrangement for Romania in 2011. It was also later clarified that prior consultation with the EC is required by EU members for a precautionary SBA with the IMF because an SBA involves ex post conditionality that could run counter to EU policy recommendations.

In retrospect, it was highly unfortunate that the Executive Board did not discuss in September 2008 the Article IV staff report on Latvia. The Executive Board could have provided its views on the interrelated issues of overvaluation of the exchange rate and sustainability of the currency peg. The Board’s views would have been particularly timely coming only months before program negotiations. This sequence of events demonstrates that regular discussions by the Executive Board of Article IV staff reports should not be deferred because of tricky or difficult economic situations.

This perspective—IMF staff’s unwillingness to support with Fund resources an exchange rate regime considered unsustainable—contrasts with the one presented by Blustein (2015a). He argues that the EU’s relative large financing contribution (77 percent)—a “reverse Hungary”—put the EC in the driver’s seat, making the EC the senior partner in designing the program, and relegating the IMF to being a junior partner. Thus in Blustein’s view, a reversal of financing roles caused a reversal in the policy roles.

Latvia’s preference to maintain its exchange rate peg was described as the IMF respecting the authorities’ choice of exchange rate regime, while ensuring the peg’s consistency with macroeconomic policies.

About the same time, the EC finalized a note on practical implementation issues related to joint EU-IMF programs. See Appendix I of the EPE for Hungary (IMF, 2011b). The EU’s Economic and Financial Committee and ECOFIN discussed this note, which spelled out the work sequence for EC BOP assistance missions including coordination with the IMF, which entailed the systematic mutual sharing of draft briefs/policy notes in order to ensure consistent policy conditionality.

Interestingly, the Fund’s General Counsel, in a speech given in Frankfurt in January 2009, entitled “Ten Years of the Euro: An IMF Perspective,” had observed that the IMF could provide financial assistance to an IMF member in the euro area with a BOP need, even if the EU could not, although this case was considered “somewhat theoretical.”

Specifically, this provision states that the union, or member states, shall not be liable or assume the commitments of central governments, regional, local, or other public bodies.

This provision prohibits the ECB and national central banks from extending overdrafts or any type of credit facility to any level of government. It also prohibits them from directly purchasing national debt. However, the ECB and national central banks can provide liquidity support to solvent commercial banks.

See the Managing Director’s speech, “Strengthening European Integration and Cooperation,” delivered on March 29, 2010 to the Warsaw School of Economics: “The IMF has also partnered very effectively with the European Union during the crisis—jointly providing balance of payments support to countries in the region. We see this as both a reflection of our common interests and as a template for better cooperation with regional financing mechanisms in the future.”

At end-April 2010 (before Greece), the IMF’s one-year FCC was equivalent to about €220 billion; roughly half of this stemmed from borrowing arrangements with EU central banks. The financing need estimated for the Greek program was €110 billion, or half the FCC.

Credit-risk transfer (to the IMF) could have been resolved by maintaining access under the IMF program at its actual level while having loan disbursements from the euro area governments be triggered solely by IMF disbursements (that is, with no separate conditionality imposed by euro area governments via EC/ECB). Of course, euro area governments would have still needed to obtain approval by their national parliaments for their respective loan (budget) contributions. Would national parliaments have entrusted their taxpayers’ money solely to the IMF without separate euro area conditionality and a role for EU institutions? Any answer is purely speculative.

The exceptional access policy was triggered in 17 of 23 (74 percent) of the General Resources Account arrangements outstanding at end-2009. Clearly, IMF-supported programs providing exceptional access were not unusual at that time. Nonetheless, the IMF would avoid triggering its exceptional access policy in the case of Cyprus (2013).

The IMF could have arranged a no-money program with upper credit tranche conditionality by extending eligibility for its Policy Support Instrument to the entire IMF membership. Policy Support Instrument programs are classified as a form of IMF technical assistance (see Decision No. 3561-(05/85).

To restore the monetary transmission mechanism in certain market segments, the ECB also announced on May 10, 2010 that it would begin to intervene in dysfunctional euro-area public-debt markets (the Securities Markets Program) and would adopt longer-term refinancing operations for banks.

Not only did such a commitment raise legal issues (for example, no Board decision, and the inability of the IMF to lend to euro area institutions as opposed to euro area countries); it was also made in a context where, as seen above, the IMF’s forward-commitment capacity at end-April 2010 was only about €220 billion. After Greece, the FCC was below €200 billion.

Nor did the third European program with Greece (8/19/2015) have a concurrent request for a use of Fund resources (UFR) program with the IMF, but such a request was expected subsequently. The EC MOU that was signed by the Greek authorities stated that the MOU was prepared in liaison with the ECB and with input from the IMF. Separately, the EC indicated that the IMF would take part in the regular review missions and was expected to participate financially later. From the EC’s perspective, the IMF was a “partner in the ESM programme as envisaged under the specific arrangements of the ESM Treaty.” The ESM stated that both the MOU and the ESM loan agreement were approved by ESM members according to their national procedures, which included parliamentary approval in several countries. The IMF mission chief to Greece confirmed in August 2015 (IMF Press Release 15/377) that the IMF would “make an assessment of its participation once the steps on the authorities’ program and debt relief have been taken, expected at the time of the first review of the ESM program.”

On June 8, the day before the Eurogroup’s announcement, the IMF Board considered a Financial System Stability Assessment (FSSA) report on Spain that identified a need to increase capital buffers by €40 billion. The FSSA report (IMF, 2012e) and the Board minutes (Minutes of Executive Board Meeting 12/55-1), particularly interventions by Executive Directors from euro countries, do not reveal any foreshadowing of the Eurogroup’s announcement, although press reports about a potential EU loan were noted by one non-euro-area Executive Director. Five Executive Directors questioned the departure from usual practice in considering a stand-alone report (without the usual Article IV staff report) of the FSSA and the shortened circulation period for the FSSA report. Indeed, one Director opined that “once again the Board is being led to diverge from recommended procedures to suit the situation and preferences of a euro area country.” Some unease was also expressed about the IMF—the Board—taking an official view, especially in a press release, on the strength/resilience of Spain’s financial system without the backing of an accompanying Article IV analysis. The 2012 Article IV consultation mission was in Madrid at that time, and completed its work on June 14 when it issued its concluding statement. Thus, a Board discussion of the FSSA and Article IV staff report could have taken place without a lengthy delay with some effort.

As the ECB’s Outright Monetary Transactions (OMT) program has not been used, only the EC and IMF conditional programs are reviewed in this section. The OMT program is discussed in the section “On Which Side of the Negotiating Table Should the ECB Sit?” below on the ECB’s troika role.

Notably, the economic governance architecture of the CEMAC, ECCU, and WAEMU differs from that of the euro area; for example, the fiscal rules are generally less restrictive than EMU rules and their enforcement mechanisms are also weaker (Schaechter and others, 2012; Hitaj and Onder, 2013; and Bova, Carcenac, and Guerguil, 2014). There are also major economic differences among these currency unions (Tan, 2017). The euro area has systemic importance and the euro has a role as a reserve currency. Relatedly, the EC and ECB have more staff and broader responsibilities than their equivalents in other currency unions. In addition, and more controversially, nationals from these three currency unions do not hold as prominent senior positions within the IMF, including management positions, as do nationals of the euro area and EU more broadly. This greater prominence could result in a tilt—even if unknowingly—toward “European exceptionalism.” Finally, the voting power and voice—number of Executive Directors or Alternates— for the euro area is considerably larger and louder than for other currency unions (Eichengreen and Woods, 2016).

According to IMF (2013b), three RFAs include no explicit role for the IMF. The Chiang Mai Initiative Multilateralization (CMIM) requires the existence of a Fund-supported program for disbursements above 30 percent of its member’s maximum quota. Below that threshold, the CMIM may set its own conditionality. Neither provision has yet been used. The North American Framework Agreement does not require an IMF-supported program; however, a letter from the IMF Managing Director, stating confidence in the borrower’s policies, is needed by the U.S. Treasury Secretary to authorize use of the Exchange Stabilization Fund.

Would “borrowed” IMF conditionality have been politically feasible for European governments? Given that their large loans frequently required authorizations by their respective national parliaments, would these national parliaments have accepted less involvement by the EC and ECB? Would the various national courts and the European Court of Justice have viewed differently the legal challenges to the euro area’s rescue mechanism? Answers to these counterfactual questions are left to the reader.

Even though troika partners worked well together and with the country authorities, they still could produce agreed policy advice judged to be less than appropriate. For an understanding of the appropriateness of policy design, see Donovan (2017); Eichenbaum, Rebelo, and de Resende (2017); Kopits (2017); Schadler (2017); Véron (2016); and Wyplosz and Sgherri (2017).

See the four background papers for that review (IMF, 2012b).

The EPE guidelines (IMF, 2010d) state that the ex post evaluation should be completed within one year of the end of the arrangement, where “completion” means approval by management for circulation to the Board. The EFF with Portugal expired on June 30, 2014. Thus, this EPE should have been approved by management for circulation to the Executive Board by June 30, 2015.

IMF staff noted that the exact numbers of EC measures were difficult to establish because they were typically broken down into sub-measures. Extensive EC structural conditionality has been noted, for example by Pisani-Ferry, Sapir, and Wolfe (2011a, 2011b) and by the ECA (2015). The latter estimated EC structural conditions at nearly 400 in the cases of Ireland and Portugal. The ECA did not review the European program with Greece.

The ECB in written testimony to the EU Parliament described troika cooperation as conducted in “a very good and fruitful manner. The different perspectives and experiences that the three institutions bring to the table provide for a more complete assessment and minimise possible errors or omissions.”

Albeit less parsimonious in these three euro program cases than in Fund-supported programs with countries outside the euro area (IMF, 2015b).

Pisani-Ferry, Sapir, and Wolff (2011b) offer three reasons for ECB involvement in the troika: (i) the ECB had significant exposure in these countries, particularly Greece and Ireland; (ii) euro area leaders trusted the ECB’s judgment; and (iii) the ECB could directly counter any IMF recommendation that might challenge ECB policies.

Only individual members of a currency union can request Fund financial support. Fund conditionality is applied in order to safeguard the Fund resources used by the requesting member.

Union-level measures may be difficult to implement in practice, because policy changes that may be desirable from the point of view of a particular member may not be so for others in the union, particularly if spillovers from the member to others are not considered to be systemic. More generally, union-wide policies can be hard to change quickly as they can involve complex decision-making procedures and multiple countries.

Country case studies by Donovan (2017); Eichenbaum, Rebelo, and de Resende (2017); and Wyplosz and Sgherri (2017); and studies on fiscal policy by Kopits (2017). For financial sector policies see Véron (2016).

For an FAD perspective on default in an advanced economy like Greece see Cottarelli and others (2010).

Even this revised debt restructuring has been deemed too expensive for Greece (Zettelmeyer, Trebesch, and Gulati, 2013), and as continuing a pattern of “too little, too late” (IMF, 2013c).

Any agreed collaboration principles would usefully be supplemented by operational guidelines for IMF staff. Such guidelines would help ensure that the IMF staff understands how to implement the agreed collaboration principles; prompt their consistent application; assist in the training of new staff members; and provide a means to disseminate experience within the IMF.

The conditionality for the World Bank loan was also not agreed by the time of the IMF Board meeting in November 2008. The World Bank loan was approved only in September 2009, but was never signed or disbursed.

Blustein (2015a and 2015b) argues, citing various published and unpublished sources, that the IMF’s smaller funding role—a “reverse Hungary”—in Latvia made the IMF a junior partner to Brussels, reversing the roles in Hungary. On the other hand, Aslund and Dombrovskis (2011) see program design as determined by strong country ownership, swaying an IMF staff that was of two minds on the currency peg.

Informal Board meetings under the exceptional access policy were held on December 5, 2009 and December 10, 2009 to outline the major features of program design and the likely financing requirements.

While both the IMF and EC focused on the overall deficit of the general government, the IMF definition was on a cash basis to permit timely quarterly monitoring, while the EC used an accrual, ESA-95 definition that was EDP-consistent. The IMF definition typically produced a deficit figure that was approximately ½ percentage point of GDP smaller than the EC figure. This difference could be larger if government payment arrears increased, as did occur; such increases, however, were not allowed under the IMF program. The EC MOU reported both sets of fiscal targets, while the IMF staff reports did not.

    Other Resources Citing This Publication