Exceptional Access Policy

2. The first area where the Argentine crisis had an important impact was on the decision to commit IMF resources. Even before the crisis, concerns over whether large-scale IMF financing was creating adverse systemic risks had been growing gradually. These concerns arose in large part from the unprecedented scale of IMF financing packages provided in the late-1990s. Under the existing IMF’s access policy at the time, financial assistance to members would normally not exceed 100% of quota on an annual basis and 300% of quota on a cumulative basis. While the policy authorized the IMF to exceed this limit in exceptional circumstances, prior to Mexico’s financial crisis in 1994, recourse to financing above the normal limits had been rare. However, the situation changed drastically in the late 1990s: the amount of IMF support provided to Mexico, Thailand, Indonesia, and Korea represented 688%, 600%, 490%, and 1,939% of quota, respectively.

3. These arrangements contributed to a perception that sovereigns were primarily using IMF to repay maturing debt obligations, thereby generating moral hazard, particularly in situations where debt sustainability was in doubt. Moreover, the degree of discretion and flexibility in the existing framework was seen as making the IMF more vulnerable to pressure to provide exceptional access even when prospects for success were quite poor and the debt burden of the sovereign was likely to be unsustainable. Experience had shown that, once debt was assessed unsustainable and a debt restructuring became unavoidable, delaying it was likely to raise the costs of the crisis and further complicate its resolution [40, 41]. Market participants also argued that the lack of clarity in the framework was creating uncertainty in financial markets when a country approached the IMF. More importantly, the provision of exceptional access under these circumstances called into question the adequacy of safeguards for IMF resources. These concerns set the stage for the reform of the Exceptional Access Policy (EAP), with the IMF aiming to move from a regime that allowed broad discretion in the use of IMF resources above normal access limits to a more systematic framework that would constrain that discretion.

4. The new EAP in capital account crises was adopted in 2002, and became fully operational in February 2003 [42, 44]. Building on the principles set forth in the “Prague Framework” endorsed by the International Monetary and Financial Committee (IMFC) in September 2000,⁶ the IMF agreed on a set of substantive criteria—related to balance of payment (BOP) pressure, debt sustainability, market access, etc.—that would need to be met to justify exceptional access for members facing a capital account crisis. The sustainability assessment was seen as a crucial input for the decision on whether exceptional access was justified or a debt restructuring was warranted instead. While the existing lending framework referred to the concept of debt sustainability, it did not present a clear methodology for this assessment. This motivated the IMF to discuss how the debt sustainability assessment could be strengthened and better organized in a common framework [43].

5. Further changes to the EAP policy were proposed in early 2004, but the amendments were not approved by the IMF Board [45]. The changes would have extended the application of the criteria for EAP to any request for exceptional access, even in noncapital account cases. However, it was decided instead that cases of members not experiencing a capital account crisis would take into account the EAP criteria but would not be conditioned on meeting all of them.

Addressing collective action problems—statutory approach

6. The second area in which the Argentine crisis highlighted a shortcoming related to the mechanisms to address collective action problems. As discussed in Chapter 2, restructuring unsustainable sovereign debts had become more difficult since the 1980s, with the shift from syndicated bank loans to bond issuances and recurrence of sovereign debt litigation. While the IMF and the international community had looked into alternatives for resolving collective action problems during the 1990s, little progress had been achieved. The complexity of Argentina’s sovereign debt structure, with 88 bond issuances outstanding in five legal jurisdictions, underscored the growing challenge to orderly debt restructurings. The emergence of creditors not participating in exchanges—commonly known as “holdouts”—and their enforcement strategies to exert leverage on sovereign debtors also posed additional difficulties for the debt restructuring process [46]. These developments provided a renewed impetus for the IMF to consider a different statutory or contractual approach for resolving collective action problems.

7. The proposal for a Sovereign Debt Restructuring Mechanism (SDRM) was launched on November 2001 by the then-First Deputy Managing Director of the IMF.⁷ The SDRM envisaged the resolution of collective action problems through a treaty-based framework that would enable a debtor and a qualified majority of its creditors to make critical decisions that would be binding on the minority, including decisions regarding the acceptance of final restructuring terms. A key aspect of the SDRM was that, for voting purposes, claims would be aggregated across different instruments, even in the absence of a contractual framework linking these different instruments [47, 48]. Over the subsequent 18 months, the design of the SDRM evolved based on discussions within the IMF and with the international community.

8. SDRM discussions focused on the role of the IMF and other key design issues [49, 50]. Considerations included whether the activation of SDRM should be made conditional upon the IMF’s determination that the member’s debt was, in fact, unsustainable; the scope of the debt to be covered by the mechanism; the appointment and role of creditor committees; and how to address inter-creditor equity issues when aggregating claims. Notwithstanding these complex design decisions, in April 2003 most of the IMF’s Executive Directors endorsed a proposal setting forth the core features of the SDRM [51].

9. However, support for the SDRM proposal was not strong enough. The SDRM would require an amendment of the IMF’s Articles of Agreement, which required support of three-fifths of its members holding 85 percent of the voting power. Support was insufficient; some Executive Directors at the Board, including those from emerging market economies, were not convinced of the need for the SDRM. In the 2003 Spring Meetings, the IMFC recognized that the establishment of the SDRM was not feasible and instead welcomed the enhancement of the contractual approach—namely the inclusion of CACs in international sovereign bonds.⁸

Addressing collective action problems—contractual approach

10. In June 2002, the IMF took stock of the CACs experience, focusing on the design and effectiveness of CACs and how to encourage their greater use in sovereign bond contracts [52, 55]. The stocktaking confirmed that majority restructuring and majority enforcement provisions provided a legal framework that could help facilitate an orderly restructuring process, and that these provisions could be incorporated into all international sovereign bonds governed by New York and English laws.

11. The IMF’s stocktaking was followed by a debate in the official community and with the private sector on the design of CACs. A working group formed by the G-10 issued a report in September 2002 with recommendations for the design of CACs, which included draft model clauses.⁹ Shortly thereafter, the Institute of International Finance and six other financial industry trade associations put forward for discussion a draft set of model CACs developed for bonds governed by New York and English laws.¹⁰ While the two proposals recognized the importance of the majority restructuring and majority enforcement provisions for orderly debt restructurings, they differed with respect to design features, particularly voting thresholds.

12. In April 2003, the IMF discussed developments in the design of CACs and their incorporation into international sovereign bonds [53, 54]. Executive Directors agreed that the most effective strategy was to promote the more widespread use of those types of provisions that already existed in many international sovereign bond contracts, particularly those governed by English law. While recognizing that any decision as to the design of CACs would ultimately be made by the issuer and its creditors, the IMF encouraged the use of CACs in the New York market that were broadly in line with the provisions of the G-10 Report.

13. In March 2003, Mexico issued bonds governed by New York law that included both majority restructuring provisions and majority enforcement provisions. The Mexican bond generally followed the design features recommended by the G-10 Report for majority restructuring and majority enforcement provisions. The issuance was considered a success and helped overcome the “first mover” problem, and, over the course of the following year, several countries followed suit. The IMF continued monitoring the adoption of CACs in sovereign bonds through progress reports. As of March 2005, CACs had become the market standard in new issues of New York law-governed sovereign bonds [57].

14. In September 2003, IMF staff looked into the benefits, risks, and feasibility of aggregating claims. There was growing recognition that the establishment of a legal framework that aggregated different creditor claims for voting purposes could be of considerable benefit to the restructuring process, particularly with respect to the resolution of collective action problems. At the same time, there were also concerns that aggregation could create its own risks, including fears that it could give rise to inter-creditor discrimination and allow for manipulation by the sovereign debtor. Against this background, staff was of the view that it would not be appropriate at that time for the IMF to endorse any particular set of aggregation provisions [56].