Chapter 5. IMF Governance
- International Monetary Fund. Independent Evaluation Office;International Monetary Fund. External Relations Dept.
- Published Date:
- September 2016
A. How Did the IMF Work with European Partners?
93. The term troika in the context of the euro area crisis appears to have been used in two alternative ways. First, in a strictly operational sense, the troika refers to an ad hoc coordinating device by which the IMF and the EC, in liaison with the ECB, economized on the process of negotiating with the authorities of crisis countries. Without such a device, the authorities would have had to negotiate with two or three different teams, increasing the costs of reaching agreement. Second, the term troika is sometimes taken to refer to a policy framework within which the IMF was expected to accept constraints imposed by the EMU membership. This distinction is important. The IMF could have been the sole lender to a euro area country, working within the legal, political, and other institutional constraints imposed by the country’s EMU membership. In this scenario, the IMF would have faced the same policy constraints due to EMU membership as it did in joint lending operations, though it might have enjoyed greater leverage.
How did the troika work?
94. There were several aspects to the way the IMF operated within the troika arrangement (Kincaid, 2016). First, the IMF does not allow the use of Fund resources to be directly subjected to the rules and decisions of other organizations. Thus, the IMF cannot disburse funds conditional on the judgment of another organization that the borrower has met its conditions—a situation known as cross-conditionality. Even so, judgements reached by other lenders could affect the IMF’s lending decision by virtue of what is known as the financing assurance policy, that is, the requirement that the program be fully financed. Thus, even if the IMF assesses that its own conditions have been met, it may not disburse if the IMF-supported program is not fully financed because the country has failed to meet conditions imposed by other lenders.
95. Second, there was no clear demarcation of responsibilities between the IMF and its European partners, and their areas of competence overlapped considerably (in contrast, in the case of IMF-World Bank collaboration, the IMF focuses on macroeconomic or macro-critical issues while leaving most structural and development-related issues to the World Bank).66 The overlap not only made coordination more complicated but also led to a duplication of staff assignments and an increase in the size of troika teams. Troika teams to Greece reportedly could total 30-40 persons.
96. Third, the IMF and its troika partners did not fully unify their analytical frameworks or approaches to structural conditionality. For example, fiscal performance criteria in IMF programs were based on cash nominal amounts for the primary deficit in Greece and Ireland and the cash nominal overall deficit in Portugal. The EC fiscal targets, in line with the Stability and Growth Pact and the Excessive Deficit Procedure, pertained to the overall general government deficit relative to GDP, using a form of accrual accounting. As noted in the IMF’s ex post evaluation of Ireland’s Extended Arrangement, “a unified approach would have helped communicate the program objectives more effectively and avoid possible uncertainties and mixed signals” (IMF, 2015a).67 In addition, structural conditionality was extensive and intrusive in EU programs, whereas it was focused on macro-critical issues in IMF programs, as noted earlier.
97. Fourth, collaboration required sharing confidential internal documents, including preliminary draft policy notes. The IEO was not shown any written documents to indicate that IMF management or the Board formally authorized confidential information sharing or defined confidentiality rules among the troika partners. Interviews with those involved suggest that practices appear to have varied considerably, from mission chief to mission chief, or from program to program.
98. Fifth, decision making at the IMF proceeded at a different level and pace from the process in the euro area. The IMF’s ex post evaluation of Greece’s SBA noted that decision making by the European authorities was more fragmented, spanning multiple institutions and varying levels; the Greek authorities felt that this exacerbated uncertainties and reduced the chance for early agreements (IMF, 2013c).68 In contrast, IMF mission chiefs had more delegated authority. As noted earlier, negotiations with the European partners to alter program assumptions or parameters, not to mention the negotiations to restructure sovereign debt, took a long time to complete, causing the IMF to lose its characteristic nimbleness and agility as a crisis manager.
99. Sixth, the troika was not the only vehicle that IMF staff used to interact with the authorities in Europe. There were multiple contacts between the IMF staff and senior officials of major euro area countries. While it is appropriate that IMF staff should consult with a wide range of stakeholders to seek their views, European members of the troika who were interviewed for this evaluation did not consider these contacts outside the troika arrangement favorably, seeing room for misunderstanding.
100. Finally, an unusual feature of the troika arrangement was that the IMF partnered with European policy institutions at the staff level, potentially subjecting the IMF staff’s technical judgments to political pressure from an early stage. The European Commission, in the area of emergency crisis lending, acted as the agent of the Eurogroup, which in turn represented member states and decided whether to provide assistance.69 The intergovernmental nature of the euro area lending operations—given the EC’s consensus decision making—meant that negotiating positions had to be pre-agreed with individual creditor countries in order to ensure full European support (Pisani-Ferry, Sapir, and Wolff, 2013). IEO interviews and some internal documents suggest that political feasibility in creditor countries was an important consideration for EC staff and that IMF staff occasionally felt pressured to accept a less-than-ideal outcome. Because all members of the troika needed to agree on a unified position before jointly approaching the borrowing country for a program negotiation or review, this setup potentially exposed IMF staff to political decisions at an earlier stage than would normally be the case (see “Was the IMF a junior partner?” below).
101. From the point of view of the borrowing country, however, the troika arrangement may have been an efficient mechanism. Despite the potential governance and operational issues noted above, the arrangement allowed programs to be negotiated quickly and program reviews to be completed expeditiously in most cases. National authorities who were interviewed for this evaluation were generally satisfied with how the process worked. In addition, EC and ECB officials stated to the IEO that their working relationships with IMF staff were by and large effective and professional, although the modalities and practices of engagement evolved as they gained experience. The European Court of Auditors generally gave high marks to the operation of the troika (“the variety of expertise and experience among the three institutions’ staff made it possible to produce more thorough assessments”) and to the role of the IMF in particular (“working with the IMF helped the Commission in learning how to manage the programmes”) (ECA, 2015a). European and national officials interviewed by the IEO generally considered the IMF’s value-added and contribution to the troika process to be positive. A senior IMF official explained to the IEO that the arrangement had worked well as long as there were no major differences of view among the partners, a condition that may not have existed at all times.
102. There appears to be no formal procedure or guidance for staff on engaging on programs in the euro area. The IEO has not been given any documentation to show that IMF management consulted with the Board on the modalities for engagement with the euro area even after a considerable passage of time. At the Board discussion of the IMF’s 2011 Conditionality Review in September 2012, a number of Executive Directors requested an in-depth review of the troika, whereas those representing euro area countries expressed the view that troika cooperation “proved quite successful” and was “very effective” and “well-functioning” (IMF, 2012b). The summing up of this 2012 meeting noted: “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” (IMF, 2012c). The Board has not yet clarified its position on the modalities for engaging with the euro area, though at least three Board papers prepared by staff have addressed the need to establish guidelines for engaging with regional financing arrangements (IMF, 2011a, 2013a, 2015c).
Was the IMF a junior partner?
103. Critics have frequently characterized the role of the IMF in the troika as that of a junior partner. In terms of financing, the IMF clearly was a junior partner, providing at most one-third of the program financing for Greece, Ireland, and Portugal. In terms of policy inputs, however, the distinction between the two meanings of the word troika becomes critical. The IEO, based on all available evidence, concludes that the IMF was not a junior partner in the troika if the troika is defined as a device for inter-agency coordination. The IMF was not a senior partner, either—which contrasted with the IMF’s customary sole or lead role in its lending to emerging markets and developing countries. It is the consensus view of all interviewed for this evaluation that each member of the troika had a veto power. Thus, the troika arrangement was effectively viewed as consisting of coequal partners. The IMF’s veto power stemmed from recognition of its considerable expertise and crisis management experience, and its credibility with key euro area members and their parliaments whose consent was required in the context of EFSF/ESM lending decisions. Those interviewed for this evaluation did not share the view that the IMF’s relatively small financial contributions muted its voice in policy debates or its impact (Kincaid, 2016).
104. If the troika is defined more broadly as a policy framework in which the IMF was expected to operate, the question “was the IMF a junior partner?” is really that of whether the IMF too readily deferred to decisions that were made within the euro area. At the technical level, there is no evidence to suggest that the IMF staff too easily accepted the institutional constraints of EMU membership. A number of complaints the IEO has heard from European members of the troika—that IMF teams seemed not to understand or appreciate such constraints on policy options (e.g., EU regulations on state aid or competition in banking sector restructuring)—give support to the view that the staff often held its ground.
105. Serious conflicts arose at a higher, political, level but there was no agreed mechanism to address them. For example, in Ireland, the IMF staff’s judgment was that senior unsecured creditors of Irish banks should be bailed in, but this position was overruled by the European partners, who were concerned about the spillover to the euro area’s integrated banking market. In this case, as well as the case of preemptive debt restructuring in Greece, the ultimate decisions had the support of the IMF’s largest advanced economy shareholders. The larger question therefore is about the role of political constraints in the IMF’s decision-making process. In these cases, the IMF consulted with the United States and the other G7 countries. While consulting with major shareholders has been a standard practice in the IMF during times of crisis, it is legitimate to ask whether the Board—i.e., all shareholders—should have been consulted (see below).
106. These and other conflicts arose in part because the IMF’s objectives were not fully aligned with those of the euro area. The overriding concern of the European authorities was to preserve stability, and especially to preserve the single currency project. In contrast, the IMF’s responsibility was also to the individual countries requesting financial assistance. The IMF’s Articles of Agreement stipulate that the use of its general resources is meant to “assist members to solve their balance of payments problems in a manner consistent with the provisions of this Agreement” (Article V, Section 3(a)). In most cases, conflicts do not arise because policies aimed to help solve a member’s balance of payments problem contribute to maintaining or restoring systemic stability. In the case of the euro area, debt restructuring was an issue where a conflict could arise between what was good for a country and what was good for the euro area as a whole.
How did ECB participation affect decision making?
107. Another controversial area was the role of the European Central Bank in the troika. Authorities from a number of countries outside the G20 (and Europe) found the arrangement in which the ECB sat on the same side of the table with the IMF inappropriate as “this implicitly took certain policy actions ‘off the table’ and constituted bad governance” (IEO, 2014c). Likewise, officials from other currency union central banks who were interviewed for this evaluation indicated that they found a conflict of interest in this arrangement—as did the European Parliament (Karas and Ngoc, 2014; Sapir and others, 2014).70
108. What kinds of policy actions were taken off the table by virtue of this arrangement? Could the ECB have been made subject to IMF program conditionality for a member country? Establishing the counterfactual would be difficult, especially because the ECB did take a number of measures to help preserve financial stability in the crisis countries, though without formal conditionality placed on it. Even in the context of financing arrangements with members of other currency unions (Box 7), there appears to have been no instance in which conditionality was explicitly placed on union-wide monetary and exchange rate policies. There have been instances of implicit conditionality, however. The most prominent case was the devaluation of the CFA franc in January 1994 (see IEO, 2007a, p. 33), which was applicable to all CFA franc zone members and implicitly considered by IMF staff as a prior action (Tan, 2016).
Box 7.The IMF and Currency Unions
Besides the euro area, there are three other currency unions in the world: the Eastern Caribbean Currency Union (ECCU), the Central African Economic and Monetary Community (CEMAC), and the West African Economic and Monetary Union (WAEMU); the CEMAC and WAEMU are collectively known as the CFA franc zone. The challenges they pose to the IMF are similar to those posed by the euro area: the IMF’s Articles of Agreement contain no provision for joint membership, creating complexities absent in its relationship with non-currency-union members in terms of surveillance and conditional lending.
The ECB’s troika role contrasted with the approach usually taken by the central banks of other currency unions when the IMF lends to one of their members. In negotiations involving IMF-supported programs for members of other currency unions, the regional central bank never sits on the IMF’s side of the table. As a general practice, officials from the regional central bank, if present, typically sit with or closer to the country authorities; letters of intent are signed only by the country’s finance or prime minister. In the euro area programs, the governor of the national central bank sat on the side of the national authorities and across from representatives of the ECB, EC, and the IMF, even though national central banks are part of the Eurosystem, and their governors are members of the ECB Governing Council.
To be sure, some features of the euro area distinguish it from the other currency unions. For example, the other currency unions do not have national central banks. In the euro area, these banks had their own national balance sheets and supervisory responsibilities over their own national banking systems, along with the ability to provide emergency liquidity assistance to national commercial banks. Financial markets are more highly developed and integrated in the euro area, and the euro is a reserve currency.Source: Tan (2016).
109. In other currency unions, instances exist of the IMF imposing country-specific conditionality on a regional central bank (Tan, 2016). Until the mid-1990s, IMF-supported programs in CFA franc zone members routinely specified as a quantitative performance criterion a ceiling on the net domestic assets of the program country’s national banking system. This practice ceased as it became increasingly infeasible to define a national monetary policy. Programs with euro area crisis countries did not contain conditionality on national monetary policy but, with financial fragmentation in the euro area and the development of new instruments (e.g., SMP, OMT, and macro-prudential tools), the scope for the ECB to shape national monetary conditions increased substantially. Likewise, with the transfer of supervisory powers from national authorities to the Single Supervisory Mechanism in November 2014, associated conditions could now be reassigned to the ECB, as is the case in the other currency unions.71
110. In the euro area crisis, the IMF could have been more proactive in advocating policy measures at the union level to help ensure the success of programs in individual countries (Truman, 2013; Dhar and Takagi, 2016). The IMF staff, in its recent review of IMF-supported programs following the global financial crisis, interprets the Articles of Agreement as “establish[ing] the Fund’s inherent ability to call for the adoption of union-level measures where such measures are necessary for the success of a member’s Fund-supported program” (IMF, 2015c). As the staff further notes, such measures need not take the form of formal conditionality; they could take the form of surveillance or policy commitments, which the Board seems to prefer.72 Such policy commitments could involve not only monetary and bank supervision policies but also other policies whose competency is held by union-level institutions. Providing advice on the general stance of fiscal policy at the union level, when some union members are experiencing a collapse of domestic demand, could be another such example.
B. How Well Did the Executive Board Perform Oversight?
111. The Executive Board played only a perfunctory role in key decisions related to the IMF’s engagement in the euro area crisis (De Las Casas, 2016). This is not surprising. The weakness of the Board in exercising its oversight responsibilities has been a recurring issue in the governance of the IMF (IEO, 2008; IEO, 2014b). In the case of the euro area crisis, the Board’s absence was particularly problematic in two respects. First, the IMF’s exceptional access policy explicitly highlights the role of the Board. While the letter of the framework was complied with, the spirit was not fully respected. Second, earlier and more active involvement by the Board, even in situations where it would not have changed the decisions, could have lessened the public perception that the IMF was giving unfairly favorable treatment to the euro area.
112. In particular, the IMF’s exceptional access policy was followed only in a perfunctory manner. The policy requires that management consult the Board early on an informal basis, in order to provide a basis for consultation with national authorities and to reinforce “careful and systematic decision-making” (IMF, 2003a). However, the information provided to the Executive Board prior to the announcement of the programs did not contain important details, including on quantified estimates of the financing requirements, expected European financing, or possible access to Fund resources. The IEO has found no evidence to show that such information was even orally communicated to the Board, though that information was available to IMF management and staff as well as to the Eurogroup (Kincaid, 2016; De Las Casas, 2016). Nor was the Board consulted on three key policy issues: (i) in Greece, the absence of a high probability for sovereign debt sustainability; (ii) introduction of the systemic exemption clause; and (iii) in Ireland, whether to apply haircuts to senior unsecured bondholders (management’s consultation took place with G7 finance ministers, not with the Board).
113. The informal briefings, during which staff informed the Board of the details of the program, took place on the same day as the announcement of the stafflevel program agreement in the cases of Greece and Ireland and three days ahead in the case of Portugal (De Las Casas, 2016). By not providing the Board with key information early on a timely basis, management and staff undermined the ability of Executive Directors to consult effectively with their authorities and to perform their duties. From late 2009 through the first months of 2010, the Board was not fully kept informed of internal discussions that took place regarding possible IMF participation in a joint lending operation for Greece. An early, open consultation with the Board might have pushed management and staff to consider all available options, as noted previously.
114. The lack of Board involvement was an ongoing feature throughout the euro crisis period. Most members of the Board were not kept informed of ongoing developments that concerned the IMF’s role in the euro area; some who were interviewed for this evaluation complained to the IEO that they had learned more from the press than from informal Board meetings. There was no Board discussion of the role of the IMF in the 2012 Treaty Establishing the European Stability Mechanism. No formal discussion took place in the Board on cooperation principles between the IMF and regional financing arrangements (RFAs), even though the IMFC called for such a discussion in April 2011. By not consulting more fully with the Board, management effectively excluded at least a part of the IMF membership from key decisions. As a result, the legitimacy of what the IMF did in Europe has been challenged, with a lingering perception that the Fund treated Europe differently (Park, 2016).
115. In terms of IMF governance, the experience with the euro area crisis posed two fundamental problems (De Las Casas, 2016; Kincaid, 2016). First, at least some Executive Directors representing euro area countries may have had information that was not available to other Directors, creating an information asymmetry within the Board. Information asymmetry is always present whenever the Board discusses issues related to the use of Fund resources, as Directors representing the borrowing countries inevitably possess an informational advantage. In the case of the euro area crisis, this informational advantage extended to all Directors representing euro area countries (including creditor countries) and not just the Director representing the borrowing country. Given the conditional lending by euro area governments, their information requirements may have been warranted. But this situation should have placed upon IMF management and staff a greater responsibility to share information more frequently and completely with the entire Board than was customary. Second, in all instances, decisions by the European partners preceded IMF Executive Board meetings on the use of Fund resources. This sequencing ensured that the IMF-supported program was fully financed by the time the Board met, but it caused some to view the Board as faced with a fait accompli and created the perception that the IMF Board merely rubberstamped decisions already taken in Europe.