This paper describes the evolution of ideas to apply bankruptcy reorganization principles to sovereign debt crises. Our focus is on policy proposals between the late 1970s and Anne Krueger's (2001) proposed "Sovereign Debt-Restructuring Mechanism," with brief reference to the economics literature on sovereign debt. We describe the perceived inefficiencies that motivate proposals, and how proposals seek to change debtor and creditor incentives. We find that there has been a moving concensus on what constitutes the underlying problem, but not on how to fix it. The range of proposed approaches remains broad and only recently shows some signs of narrowing.
This paper describes the evolution of ideas to apply bankruptcy reorganization principles to sovereign debt crises. Our focus is on policy proposals between the late 1970s and Anne Krueger's (2001) proposed "Sovereign Debt-Restructuring Mechanism," with brief reference to the economics literature on sovereign debt. We describe the perceived inefficiencies that motivate proposals, and how proposals seek to change debtor and creditor incentives. We find that there has been a moving concensus on what constitutes the underlying problem, but not on how to fix it. The range of proposed approaches remains broad and only recently shows some signs of narrowing.
IMF First Deputy Managing Director Anne Krueger’s recent proposal to create a “Sovereign Debt-Restructuring Mechanism” has triggered a new debate about procedures, based either on statutes, private contracts, or official policies, that would apply bankruptcy reorganization principles to the resolution of sovereign debt crises.2 As is widely known, this debate has some predecessors, particularly during 1995–96 and 2000–01, beginning with an influential speech by Jeffrey Sachs (1995) in the aftermath of the Mexican crisis. What is less known is that the post-Mexico debate was itself preceded by a steady stream of proposals beginning even prior to the 1980s debt crisis, which contained many of the ideas put forward since 1995.
This paper describes the evolution of this literature from the late 1970s until 2001, with particular emphasis on the early literature. Our starting point reflects the earliest postwar contributions on the subject: prior to the 1970s, with very little private lending to developing countries and no debt crises involving the private sector, the subject of sovereign bankruptcy received little attention.3 Since our objective is to survey the intellectual history of the current debate rather than this debate itself, which has been rapidly evolving in recent months, we end the survey just prior to the Krueger (2001) speech and the contributions that follow. However, we briefly return to the current debate in the conclusions.
We restrict the survey in two ways. First, our focus is on policy proposals. The formal economics literature on sovereign debt, which begins with contributions by Eaton and Gersovitz (1981) and D. Cohen and Sachs (1982), is discussed only to the extent that it sheds light either on the inefficiencies that the policy proposals are concerned with, or on the potential limitations of some proposals. Second, within the set of policy proposals, we concentrate on ideas to deal with sovereign insolvency through a structured arbitration or negotiation process between creditors and the debtor, often based on the analogy with domestic bankruptcy organization (e.g., Chapter 11 of the U.S. bankruptcy code). This implies, in particular, that two major policy initiatives to resolve the debt crisis of the 1980s are only referred to in passing, namely market-based mechanisms for debt reduction, and ideas to create an official “International Debt Facility” that would buy privately held debt at discounted prices and subsequently offer debtors some degree of relief. Both have been expertly surveyed elsewhere.4
We approach the subject by asking two questions. First, what are the assumed inefficiencies that the various proposals are trying to address? Second, how would the proposals change incentives in a way that addresses these inefficiencies, if at all? As we shall see, there is a clear historical progression with regard to the first question. From the late 1970s to the early 1990s, the emphasis was mostly on inefficient debt workouts due to free riding by individual creditors. Beginning in 1995, this is extended by two additional concerns: addressing self-fulfilling debt crises, and avoiding the moral hazard problem associated with large IMF-led bailouts.
On the second question, however—how to influence creditor and debtor incentives—the progression is less obvious. To be sure, from the late 1970s until about 1995, the range of approaches that were put forward to deal with the sovereign bankruptcy problem steadily expanded, moving from sketchy ideas on international cooperation to more complete and concrete statutory proposals, and finally the suggestion to focus on debtor-creditor contracts rather than new institutions. Until the end of 2001, however, there is not much of a trend—in the sense of either a movement toward consensus, or a “moving consensus”—to speak of. Only the most recent, 2002 round of the debate shows some signs of narrowing between the main proposals.
Several years before the 1980s debt crisis, rapidly growing developing country debt, crises in Zaïre and Peru, and finally the “specter of widespread defaults by LDCs” (Oechsli, 1981) prompted calls for better creditor coordination and the development of a formal restructuring process for international debt.5 The first policy initiative along these lines appears to be the proposal to create an “International Debt Commission” put forward by the Group of 77 developing countries during a meeting in Arusha in February 1979, in preparation for the fifth United Nations Conference on Trade and Development (UNCTAD) in Manila in June of the same year. The Debt Commission would consist of “eminent public figures with recognized knowledge and experience of debt problems and economic development. Any interested developing country which believes it has, or may have a debt problem could address itself to the Commission. The commission will: (i) Examine the debt and development problems of the requesting country; (ii) In the light of such examination … make recommendations on measures required to deal with the debt problem in the broader context of development including measures of debt reorganization and additional bilateral and multilateral finance; (iii) Convene a meeting of all parties concerned with a view to implementing the recommendations under (ii) above” (Group of 77 (1979), p. 157).
Although the “International Debt Commission” never materialized because of resistance from the creditor countries—and in any event would not have had powers other than making recommendations—the Arusha Program foreshadowed several aspects of subsequent proposals for an international bankruptcy mechanism. These include the “debt reorganization” objective, the desire to coordinate all parties, a role for a neutral arbiter or mediator, and the emphasis on new financing. This said, it appears that the primary objective of the G-77 was to make debt negotiations with official creditors more debtor friendly compared to the Paris Club framework,6 rather than addressing an inefficiency under the status quo. In particular, neither the Arusha Program itself nor a subsequent background report by the UNCTAD secretariat (UNCTAD, 1981) refer to problems caused by private creditors, or poor coordination among creditor groups.
The credit for the first proposal that invokes the Chapter 11 analogy and is explicitly motivated by problems of this kind goes to Christopher Oechsli (1981): “Many of the procedures set forth in Chapter 11 of the Bankruptcy Reform Act of 1978 for rehabilitating financially troubled businesses can be applied profitably to renegotiation of LDC debt” (p. 354). Of these procedures, Oechsli emphasizes three: a creditor committee, an independent “examiner, a monitoring party which does not displace the debtor from control of its business,” and a formal initiation procedure. Oechsli believes the monitor could be the IMF, but stresses the need for “inclusion of the debtor in the formulation process” of a restructuring plan. The initiation procedure should allow either creditors or debtors to take the initial step, although “creditors and the IMF need not accept the debtor LDC’s formal petition.”
According to Oechsli, sovereign debt restructuring under the status quo suffers from several problems. Negotiations take too long, and their outcome is too uncertain, harming the debtor and delaying the rehabilitation process. Moreover, they may be insufficiently focused on “an LDC’s basic development as the means to strengthen the country’s credit and debt service capacity” (p. 329). In Oechsli’s view, this is due to the “lack of an established procedure” and poor creditor coordination. By the latter, he seems to mean primarily a lack of coordination between classes of creditors negotiating separately—in particular, the private and official sectors—rather than across individual private creditors. In response, he proposes “an established procedural framework for debt renegotiation” that includes “all major official and commercial creditors in a comprehensive response. The procedure should avoid the long delays … which result from separate renegotiations by different types of creditors.” In particular, “the commercial creditors would not have to delay their reaction to the LDC debt problem until after the completion of the official creditor club negotiations” (p. 329).
Although he contemplates the creation of a “court-like entity” as a possibility, stating that “the IMF would seem the obvious choice for that role,” Oechsli concludes that this is unnecessary. “Alternatively, creditors could specify binding arbitration procedures in their loan contracts” (including an arbitration entity along the lines of the International Centre for the Settlement of Investment Disputes). “Neither, however, is necessary for Chapter 11 procedures to be applied successfully in the renegotiation context. Establishment of a renegotiation plan could continue to be by the agreement and consensus of the parties, and not by imposition from some international institution” (p. 333).
In short, in spite of invoking the Chapter 11 analogy, Oechsli seems to implicitly attribute the inefficiencies associated with debt renegotiation to shortcomings in the official sector’s approach, rather than collective action problems among private creditors. He does not discuss provisions for dealing with free riding within the private sector such as declaring a standstill, or imposing a majority-approved agreement on dissenting creditors. A Chapter 11-like procedure is important to him not as a solution of the free rider problem, but mainly because it provides a predictable timetable and clear communications channels.
This said, Oechsli’s view of debt restructuring as a three-way negotiation (debtor, private creditor, and official creditor), and his insistence that the uncertainty about the timing and nature of the official response can complicate negotiations between debtor countries and private creditors was farsighted, and arguably borne out during the debt crisis.7 Moreover, the three-way negotiation framework, formalized a few years later by Bulow and Rogoff (1988a), proved useful not only in understanding the ex post inefficiencies that Oechsli was concerned with, but also in analyzing the potential moral hazard problem ex ante facing creditor country governments and international financial institutions.8 As we shall see below, this developed into one of the principal arguments for a sovereign bankruptcy procedure in the debate that followed the 1995 Mexican crisis.
III. Developments During the Debt Crisis
Oechsli’s proposal does not seem to have become widely known outside a narrow legal literature. Ironically, the 1980s debt crisis itself initially seemed to stifle rather than inspire similar ideas, perhaps because the debate about strategies for resolving the debt crisis—including market-based debt-reduction schemes, plans for a new “International Debt Facility” or “International Debt Discount Corporation” that would centralize private debt in public hands, and finally the Brady plan—crowded out loftier proposals on international bankruptcy reorganization.9 However, there were important exceptions as early as 1984, and by the time the debt crisis ended in the early 1990s, several proposals had been put forward that went significantly beyond Oechsli’s contribution, both in terms of motivation and in their statutory implications. In addition, some contributions explored the question to what extent Oechsli-style sovereign bankruptcy procedures could be supported by the existing institutional framework, in particular, the IMF’s Articles of Agreement. Finally, the 1980s were exceptionally fertile in improving the understanding of why sovereign debt might create problems that could not necessarily be dealt with by market forces, current institutions, or one-off policy measures. Beginning with the latter, we now survey these developments in turn.
A. A Better Understanding of the Problem
Collective Action Problems and the Allied Bank Case
A milestone in understanding what inefficiencies might justify a centralized resolution to sovereign debt crises was Jeffrey Sachs’s influential 1984 Princeton Study, “Theoretical Issues in International Borrowing.” Sachs’s paper, which extended earlier work with Daniel Cohen (D. Cohen and Sachs, 1982), presents a simple, formal statement of some of the collective action problems associated with international debt—both self-fulfilling debt crises, and free rider problems in the context of debt rescheduling or restructuring.10 Citing evidence from Cline (1981) on Peru, and data on a 1983 rescheduling of loans to Brazil, Sachs argues that “even in bank syndicates significant free rider problems remain” (Sachs (1984), p. 33).
A 1985 court decision, Allied Bank International v. Banco Credito Agricola de Cartago, was widely viewed as illustrating both the possibility that a single creditor might “hold out,” and the tenuous nature of protections against the claims of such a creditor. In 1981, Costa Rica had suspended debt payments to a 39-member bank syndicate. A restructuring agreement was subsequently reached with all creditors but one, Fidelity Union Trust of New Jersey, which sued through an agent, Allied Bank. In 1984, a U.S. Court of Appeals initially upheld a lower court ruling in favor of three Costa Rican banks that had acted on behalf of Costa Rica, arguing that “Costa Rica’s prohibition of payments of its external debt is analogous to the reorganization of a business pursuant to Chapter 11 of our Bankruptcy Code … Costa Rica’s prohibition of payment of debt was not a repudiation of the debt but rather was merely a deferral of payments while it attempted in good faith to renegotiate its obligations.” Upon rehearing the case in March 1985, however, the court reversed itself after the U.S. Department of Justice argued that contrary to the court’s initial assumptions, the U.S. government did not agree with “Costa Rica’s attempted unilateral restructuring,” but instead supported an IMF-guided renegotiation procedure, “grounded in the understanding that while parties may agree to renegotiate conditions of payment, the underlying obligations to pay nevertheless remain valid and enforceable.”11 In the end, the court ruled that “The Costa Rican government’s unilateral attempt to repudiate private, commercial obligations is inconsistent with the orderly resolution of international debt problems.”12
While the case was subsequently settled, the March 1985 ruling made it clear that in the United States, Chapter 11 style protections did not apply to sovereigns, as noted and lamented by Dell (1985) and many others writing at that time. For example, Sachs (1986a) tersely observes: “Individually, these creditors have an incentive to call in their claims against the overextended debtor countries, even if doing so injures the economic performance of the debtor so much that the creditors suffer collectively. Preventing such a destructive race to liquidate assets is one of the major purposes of a bankruptcy code, which restricts the ability of individual creditors to act against the group interest. Unfortunately, countries cannot file for Chapter 11 protection” (p. 418).13
This said, the Allied Bank case also illustrated the tension between addressing the collective action problem and maintaining creditor rights, which is characteristic of the debate today. While the case was initiated by a single holdout, it would be incorrect to interpret it as a straightforward conflict of interests between that holdout and the remaining creditors. Although they had settled with Costa Rica, the other creditors did not object to the litigation.
On the contrary; through the New York Clearing House Association, they filed briefs strongly supporting the claim (see Laryea, 2002). An explanation for this is that the whole creditor class—and indeed future creditors—benefited from the principle that “while parties may agree to renegotiate conditions of payment, the underlying obligations to pay nevertheless remain valid and enforceable.” The vindication of this principle strengthens the position of creditors in a voluntarily restructuring since they can negotiate from the position that they are legally entitled to 100 percent recovery in accordance with the contract terms. Thus, the lens of the collective action problem does not fully capture the basic creditor-debtor conflict central to the Allied Bank case (and, indeed, many other sovereign debt litigation cases).
Debt Overhang and “Market-Based” Debt Reduction
A second major development in the mid-1980s was the recognition that high levels of debt could lead to inefficiently low levels of growth because the need to repay creditors acted like a tax on investment (the “debt overhang” problem). Taken to the extreme, this implied that debt forgiveness might benefit not only debtors but also creditors if the write down of nominal claims was more than offset by an increased likelihood that the country might repay its remaining debt.14 The debt overhang argument was used to justify “market-based” debt-reduction schemes, in which the country itself took the initiative in reducing the debt stock by buying back debt at discounted prices, swapping bank loans for local currency that had to be invested in domestic equity (debt-equity swaps), or exchanging loans against discounted “exit bonds” with lower principal or interest.15
Market-based schemes enjoyed considerable popularity in the late 1980s and were tried in several countries (buy backs in Bolivia and Brazil, debt-equity swaps in Argentina, Brazil, Chile, and Mexico, and exit bonds in Mexico and Argentina).16 However, they soon came under criticism from several angles. Some authors argued that any efficiency gains from market-based solutions would mostly benefit creditors, not debtors (Bulow and Rogoff, 1988b; Dooley, 1988, 1989; and Bulow and Rogoff, 1991). In addition, it became clear that market-based schemes suffered from a very similar free rider problem as unilateral debt forgiveness or negotiations with uncoordinated creditors: participation in the scheme had the effect of increasing the repayment probability to the holdouts that chose not to participate (Sachs, 1988). Helpman (1989) showed that uncoordinated voluntary debt reduction will typically be suboptimally low. Indeed, the pure market-based approach mostly failed to achieve large-scale debt reduction and was eventually replaced by the Brady plan, which combined some elements of the market-based approach with coordinated negotiations and public sector funding.17
The Costs of Default and Incentives to Repay
Finally, an important intellectual development during the 1980s was a literature on what deters countries from repudiating, that is, how sovereign debt can exist in the first place. The link to proposals for a supranational bankruptcy regime is as follows. While the presence of some kind of collective action problem seems necessary to justify a Chapter 11 type of procedure at the international level, it is not necessarily sufficient. The reason is that the debtor may have the means to unilaterally protect himself from some of these collective action problems, such as the debt panics or creditor grab races described by Sachs (1984, 1986a), by declaring a unilateral payments moratorium.
Thus, to assess under what conditions Chapter 11 type of “protections” may nevertheless be necessary to achieve orderly debt restructuring, it is important to understand why countries are generally reluctant to default or suspend payments in the first place. The initial contribution in this area was Eaton and Gersovitz (1981), who portray a world in which sovereign lending can take place even if borrowers are truly immune to any direct actions by creditors in the event of nonrepayment. What motivates borrowers to make any repayments in these circumstances is a concern that they will lose their reputation in international credit markets, and thereby lose future access for an extended period. In contrast, D. Cohen and Sachs (1982) assume that if a borrower fails to make repayments, it will suffer a loss that is proportional to output. Though they do not explicitly motivate their analysis in terms of the legal rights of creditors, their framework is easiest to reconcile with the view that the international legal structure governing sovereign loans matters. In the reputational model, the international legal system is assumed to have little or no bite on debtors anyway. Thus, introducing an international Chapter 11 would not necessarily have an impact, except perhaps to the extent that it might mitigate the reputational loss of countries who “play by the rules” when restructuring their debts.
How seriously should one take the reputational model of sovereign debt? Bulow and Rogoff (1989) dealt a blow to the pure reputational view, showing that pure reputation-based debt is not sustainable under a broad range of assumptions, unless the loss a country sustains by not repaying is much broader than just its image in credit markets. However, reputational theories of default costs have remained a hot topic of debate, and experienced something of a comeback in recent years (see Cole and Kehoe (1996), Wright (2001), and Tomz (2002)). The current consensus can be characterized as saying that reputation is important, but reputational losses are not the only costs of default. There remains a question as to what these costs exactly consist of, and how the introduction of a sovereign debt-restructuring procedure would make a difference in this regard. A potential answer given in the legal literature (see Debevoise, 1984; and Barnett, Galvis, and Gouraige, 1984, discussed below) is that even when there are few overseas assets that creditors can attach, unilateral creditor suits could hamper or destroy the prospect for orderly negotiations by triggering cross-default clauses (of course, this still requires an argument of why the inability to reach a settlement is in itself a bad thing). Another answer is that creditors might interfere with the ability of the debtor country to obtain trade credit.18
In short, although the economics literature during the 1980s does not provide any simple answers, it does suggest that the laissez-faire approach to international debt crises does not generally result in efficient outcomes. In these circumstances, an institutional framework governing sovereign debt workouts can improve matters, although the extent to which it does so will depend on the exact nature of default costs. In particular, the problem of helping defaulting countries later regain normal access to international capital markets is likely to present itself in even the most smoothly functioning debt-workout mechanism.
B. Initiatives Within the Existing Statutory Framework
Aside from proposals that would have required modifications to either national law or international treaties (see next section below), two policy initiatives during the 1980s sought to facilitate the orderly resolution of debt crisis without formal changes in statutes. The first was IMF lending into arrears, which first occurred in the context of a Stand-By Arrangement with Bolivia in June 1986, and was formally adopted as part of the IMF’s debt strategy in May 1989 (see Boughton (2001), pp. 485-98). The IMF’s prior policy had been to lend only if the projected balance of payments needs of a country were fully financed. “Accumulation of arrears did not count as financing.” This implied that “if bank creditors refused to reschedule the country’s debts, the Fund would normally suspend access to its own money” (Boughton, p. 477).
Under the new policy, arrears to commercial banks were generally tolerated. From the perspective of solving the collective action and incentive problems associated with sovereign debt restructuring, this had two consequences. First, in principle, debtors could now receive IMF support after a payments suspension while negotiations with creditors were in progress. This made the prospects of declaring a unilateral moratorium less daunting and weakened the bargaining position of private creditors, who were “no longer allowed to determine whether an [IMF] arrangement would be approved” (Boughton, p. 498). Second, it gave the IMF an instrument with which to exert leverage over a defaulting debtor. Cooperative debtor behavior during its negotiation with commercial bank creditors could be rewarded through lending into arrears.19 In 1998, the policy was extended to include arrears to bondholders.
A second notable initiative during this period was the Debevoise (1984) proposal to use Article VIII, Section 2 of the IMF Articles to extend legal protections to debtor countries declaring a unilateral payments moratorium. Article VIII, Section 2 reads as follows:
“Subject to the provisions of Article VII, Section 3(b) and Article XIV, Section 2, no member shall, without the approval of the Fund, impose restrictions on the making of payments and transfers for current international transactions.
“Exchange contracts which involve the currency of any member and which are contrary to the exchange control regulations of that member maintained or imposed consistently with this Agreement shall be unenforceable in the territories of any member.”
Debevoise argues that past attempts to use Article VIII Section 2(b) to stay creditor enforcement against debtor countries were unsuccessful either because courts did not regard the term “exchange contracts” as sufficiently broad to cover loan agreements, or because “the defendant had not met its burden of demonstrating that the currency regulations relied upon were ‘maintained or imposed consistently with the Fund Agreement.” Debevoise’s proposal is to “establish a deferral mechanism” that exploits the IMF’s power, under Article VIII, Section 2(a), to approve exchange restrictions. “A nation facing an unmanageable external debt profile would be eligible in principle. Application would be made to the Fund.” Following an “expedited appraisal of the applicant’s need for the deferral mechanism … the Fund would signify acceptance of an application by a formal decision of the Executive Board. This decision would include a statement that the applicant’s exchange restrictions, particularly those on the making of payments and transfers for certain current international transactions, were maintained or imposed consistently with the Fund Agreement” (pp. 63-64). To deal with the problem that courts might still not recognize controls approved by the IMF, Debevoise suggests three alternative avenues: domestic legislation “providing that in any case in which an Article VIII Section 2(b) issue is raised, a Fund decision involving the controls at issue will be determinative,” making the IMF’s deferral authority part of the debt contract, and finally, an authoritative, broad IMF interpretation of Article VIII Section 2(b).
The Debevoise proposal does not seem to have had practical consequences, even though in January 1988 the IMF Legal Department did in fact propose, in an internal report, “that consideration be given by the Executive Board to the adoption of an authoritative interpretation of Article VIII, Section 2(b)” The proposed interpretation included a broad definition of the term “exchange contract” to include any “contract providing either for a payment or transfer of foreign exchange, or for an international payment or transfer (that is, a payment between a resident and a nonresident, or a transfer of funds from one country to another).”20 However, the purpose of this broad interpretation was not to protect sovereigns during payments suspensions, but rather to “promote more uniformity in the interpretation of Article VIII, Section 2(b).”21 In any event, the Executive Board decided not pursue the issue. But the notion that the IMF might use its powers under Article VIII, Section 2 to “legalize” standstills proved popular, and periodically reappears in the literature.22
C. New Statutory Proposals
Although overshadowed by the policy debate on one-off initiatives to resolve the debt crisis, the 1980s and early 1990s also witnessed a number of proposals for an improved international bankruptcy regime that contain many ideas that were rediscovered after the 1994-95 Mexican crisis. In the following, we focus on four contributions by Barnett, Galvis, and Gouraige (1984), B. Cohen (1989a, 1989b), Raffer (1990) and Kaeser (1990), as well as some reactions to these authors.23
Barnett, Galvis, and Gouraige (1984) set out with an extensive legal analysis of the expected consequences of a creditor suit in U.S. courts following a unilateral debt payments moratorium. They conclude that “the legal limitations upon the Bank’s suit against the foreign sovereign would pale in comparison to the practical consequences of a unilateral suit.” In particular, a suit would “trigger cross default clauses in virtually all the debtor’s other loan agreements, possibly precipitating an avalanche of litigation and hampering coordinated attempts at recovery or renegotiation of the debt” (p. 113). In other words, a suit would trigger “a race to the courthouse” (p. 85). The authors then examine the feasibility of “a national level response” to this problem, in which a stay of payments and litigation would be imposed by presidential executive action, via the International Emergency Powers Act used by President Carter to freeze Iranian assets in the United States. Apart from constitutional concerns, they find that this would raise jurisdictional problems and not be suited to the international character of the problem. “Therefore, we must look to institutional reforms at the international level,” including “the creation of an adjunct to the IMF to handle debt problems on a unified, international basis.”
Barnett, Galvis, and Gouraige envisage a “supranational, multilateral body” that would be “independent from the IMF both in administration and decision-making.” Its powers would include the authority to (1) convene mandatory discussions between a debtor state and its commercial bank creditors; (2) order the commencement of and preside over debt renegotiation proceedings; (3) preempt unilateral creditor suits; (4) determine fair terms of debt renegotiation and establish a ceiling on those terms; (5) preclude the parties from undertaking other renegotiation efforts (6) permit creditor banks’ suits to proceed as a sanction against a debtor which refused to accept the renegotiated terms and (7) require the debtor to adopt internal adjustment measures as a condition to renegotiation (p. 135). The authors’ emphasis is thus on unifying creditor-debtor negotiations and preventing a “race to the courthouse.” The exclusive right to initiate adjunct proceedings would rest with the debtor state provided that specific criteria relating to debt sustainability were satisfied. Good-faith behavior by the debtor would be enforced by the implicit threat that the ban on litigation could be revoked.
While Barnett, Galvis, and Gouraige do not go into details on the legal basis of the proposed new institution, it is implicit that unlike Oechsli’s (1981) proposal, their plan would require a formal multilateral agreement. They concede that “the political obstacles to approval of the plan are formidable …” A more modest plan, however, may prove acceptable.” This would involve a legally nonbinding “central mechanism to handle debt problems,” as well as “the adoption of a multilateral agreement establishing guidelines for debt renegotiation.”
Much along the lines of Barnett, Galvis, and Gouraige’s proposal, which he was apparently unaware of, Benjamin Cohen (1989a, 1989b) calls for the creation, “by multilateral convention,” of a new “International Debt Restructuring Agency” (IDRA). “Ideally, it would be organized as a wholly new and independent entity in order to underscore its neutrality…. In practice, it might be more feasible to get IDRA started as a joint subsidiary of the two multilateral agencies most involved with the problem now, the IMF and the World Bank.” As in Barnett, Galvis, and Gouraige’s proposal, the IDRA’s primary role would be that of a facilitator, mediator, and monitor, but Cohen’s proposal allows for a more heavy-handed role if necessary: “IDRA could conceivably be authorized to compel agreement in the event of deadlock in order to suppress any remaining temptation among lenders to free ride. For example, dissenting creditors might be obliged to accept terms agreed by a qualified majority if IDRA declared the proposed settlement to be ‘fair and equitable’.” (B. Cohen (1989a), p. 33).
Barnett, Galvis, and Gouraige, and Cohen differ slightly in their characterization of the underlying collective action problem. While the former worry about competing creditor litigation, Cohen is concerned with incentives to free ride on a settlement reached by a majority, as well as the underprovision of new financing. His proposed solution is to require less than unanimous creditor support for the acceptance of a restructuring proposal, and if necessary, giving the IDRA powers to impose a settlement. As far as debtor incentives are concerned, Cohen suggests that debt relief granted under the process could be, in part, made conditional on good debtor behavior: “Creditors would be permitted to withdraw all concessions on such matters as interest rates if IDRA determined that a debtor was not complying with its policy commitments.” He is less clear on how the IDRA would encourage debtors to negotiate in good faith and stick to the specified timetable. Unlike Barnett, Galvis, and Gouraige, Cohen does not discuss the possibility of having the agency authorize, and if necessary revoke, a stay of litigation while negotiations are ongoing.
Cohen’s proposal had broad resonance in the policy debate at the time. Miller (1991) dismisses the creation of an IDRA as unfeasible, and instead explores the possibility of amending the U.S. Bankruptcy Code (in particular, Chapter 9) to allow sovereigns to benefit from U.S. bankruptcy protections. However, he cautions that this may not protect debtors from claims by non-U. S. creditors, and could lead to jurisdictional problems. In contrast, Williamson (1992) embraces the IDRA idea and—echoing his 1985 contribution—suggests that it might operate largely on a private contractual basis: “An International Debt Restructuring Agency would base its legitimacy on clauses in future loan contracts specifying that the terms of the contract could be revised by the agency to take account of unforeseen contingencies, and that both creditors and debtors would be bound by its decision” (P. 95) 24
Roughly coinciding with Cohen is Raffer’s (1990) proposal for an international insolvency procedure modeled after Chapter 9 of the U.S. bankruptcy code, which applies bankruptcy reorganization principles to municipalities. Like Barnett, Galvis, and Gouraige, and Cohen, Raffer envisages a structured negotiation procedure overseen by a new international body. Raffer refers to it as a “neutral court of arbitration,” implying that it would have considerable powers, much like a domestic bankruptcy court. In his plan, creditors and debtor countries would nominate an equal number of arbitrators, who in turn would nominate a chairperson. Raffer argues that in all other respects, “an adaptation of Chapter 9 to the international setting would only require minor changes.”
What distinguishes Rafter’s proposal is the emphasis on Chapter 9 rather than Chapter 11 as the right domestic analogy for sovereign bankruptcy.25 He seems to prefer Chapter 9 for two reasons. First, it is not vulnerable to the objection that a Chapter 11 for countries would not work because of the impossibility of “liquidating” a state entity. Furthermore, as Raffer emphasizes, Chapter 9 limits court interference with the municipalities’ political or governmental powers, and gives certain groups that might be affected by the reorganization plan (such as unions and debtor employees) the right to be heard. In Raffer’s view, this provides an opportunity for balancing the interests of creditors with the welfare of domestic citizens as well as national sovereignty, which he thinks were disregarded in official attempts to resolve the debt crisis during the 1980s.
A final contribution in this group is a little-known proposal by Daniel Kaeser (1990), a Swiss treasury official. Building on suggestions by the “Languetin Group,” an independent commission charged by the Swiss government with proposing solutions to the debt crisis, the centerpiece of Kaeser’s proposal is the creation of a sovereign debt workout mechanism under the auspices of the IMF or some other international agency. However, Kaeser goes further in several respects. In essence, he wants to tackle three problems at once: first, create a mechanism for efficient debt reduction, second, discourage future overindebtedness, third, allow countries with sustainable debt levels to access private capital at relatively low cost. To reconcile the last two objectives, he proposes setting a hard, statutory, solvency criterion—for example, a debt-service threshold of 25 percent of export revenues. Actual indebtedness would be continuously monitored through a centralized registry. In the event that debt-service commitments were to rise above the threshold, a country could petition the international bankruptcy agency. Relief would be tranched and conditioned on adjustment policies, in the context of an IMF program. Countries below the threshold would not be eligible for debt relief.
In Kaeser’s view, this mechanism would serve as a disincentive to excessive indebtedness while still allowing countries with low debt to access the capital market. To further encourage such access, he additionally suggests an insurance fund that would partly guarantee interest payments by countries staying below the debt-service threshold. Furthermore, he proposes using the same criterion to differentiate the provisioning requirements of creditor banks. Since debt service would (by definition) fall below the threshold after a restructuring, this would also help address the underprovision of new financing emphasized by Sachs and many others.
Kaeser seems to be the first author to suggest a sovereign bankruptcy mechanism strictly geared to countries that are overindebted as defined by some “objective” criterion, as opposed to any country in payments difficulties. As we shall see, this idea fades in the 1990s, but eventually reappears (Krueger, 2001).
IV. After Mexico: The IMF as an International Bankruptcy Court?
The implementation of the Brady plan in 1991 and the subsequent resumption of capital flows to emerging markets brought a temporary lull to the literature on sovereign debt-restructuring mechanisms. This was quickly reversed in the wake of the Mexican crisis and the ensuing U.S./IMF crisis loan, beginning with an influential lecture by Jeffrey Sachs (1995), “Do We Need an International Lender of Last Resort?” (April 20, 1995).
The essence of Sachs’s argument is that the international financial system does indeed suffer from inefficiencies that could be used to justify a lender of last resort, but that in practice the IMF is so ineffective in exercising this function—partly by design,26 partly due to incompetence—that these inefficiencies would be addressed much more successfully if it were to give up its lending role and instead assume that of a bankruptcy court. “IMF practices should be reorganized such that the IMF plays a role far more like an international bankruptcy court and far less like the lender of last resort to member governments” (p. 14).
Sachs does not go into details on how his vision of the IMF as a bankruptcy court would be implemented. He argues that the IMF’s existing mandate to approve exchange rate restrictions under Article VIII Section 2(b) of its charter might give it legal cover to sanction payments moratoria, adding that “of course, more directly, the Articles themselves could be amended.” He points to the hold-out creditor problem at the beginning of his lecture, but does not state how hold-out creditors would be disciplined under his plan, and whether this would require changes in member countries’ domestic laws.
Sachs’s 1995 lecture did not break new ground either in the analysis of collective action problems and panics, which are amply covered in his 1984 study, or in terms of specific institutional suggestions, which are preceded by several proposals made during the 1980s, as shown in the previous section. Rather, Sachs’s innovation consisted in taking a much broader view of the problems that a bankruptcy mechanism might help resolve: not just the solvency crises experienced during the 1980s, but also self-fulfilling debt runs as (he argues) were experienced by Mexico in 1995, through a combination IMF-endorsed standstills and “administrative priority” of new private lending. This mirrors the idea that standstills and last resort lending are alternative ways of dealing with self-fulfilling runs, as argued in the academic literature by Diamond and Dybvig (1983) and particularly Wallace (1988).27 Admittedly, contemporary theorists no longer believe it is appropriate to draw a sharp line between pure runs and solvency crises (Morris and Shin, 1998), and the distinction has always been difficult to make in practice. In Sachs’s proposed bankruptcy mechanism, however, the IMF would not have to distinguish between these types of crises.
In effect, Sachs proposed a radical form of what much later became known as “private sector involvement” in crisis resolution. Remarkably, he did this without invoking the argument that official crisis lending leads to moral hazard. His argument was merely that private sector involvement of this kind would do the job of official crisis lending much more effectively, to the point that it should replace official crisis lending altogether. Furthermore, Sachs argued that this could be achieved without fundamentally new treaties or institutions, by merely refocusing an existing institution, possibly even without changing its statutes.
No wonder, then, that the Sachs lecture had immediate and widespread resonance and a considerable impact.28 It provided the main impulse for a discussion of the topic at the G-7 Halifax summit in June of 1995, which in turn led to the G-10 paper on the resolution of sovereign liquidity crises (see next section).29 While earlier proposals did not elicit much response, virtually all the literature published after mid-1995 discusses Sachs. Although it was never formally published, Sachs’s lecture probably did more to popularize the idea of international bankruptcy than all of the literature that preceded him.
It is not widely known that at the time when Sachs was delivering his lecture, the IMF Legal Department (1995) was preparing an extensive paper on how a international bankruptcy mechanism under IMF auspices might be designed and practically implemented (“Note on an International Debt Adjustment Facility,” May 26, 1995).30 In some ways, this paper provides the counterpoint to Sachs’s proposal. It is short on motivation: while free rider problems are occasionally referred to in the text, there is no explicit analysis of the inefficiencies embodied in the status quo, except to say that it is presently difficult to negotiate an orderly restructuring for lack of a legal process. However, it is more specific than any of the previous studies on how such a process could be legally established, structured, and implemented. It is also the first to contain a detailed discussion of the desirable scope for a sovereign bankruptcy mechanism, arguing that it ought to cover domestic as well as all external debt (except for debt owed to lFIs).
Like Raffer (1990) and basically for the same reasons (namely, the public and political nature of the debtors), the paper mainly utilizes the analogy with Chapter 9 rather than Chapter 11, although bankruptcy codes from other countries are referred to as well. As appropriate for an IMF staff report, the paper lays out options rather than pushing a specific “plan.” In particular, it distinguishes between a “consensual approach” justified by Article V, Section 2(b) of the IMF Articles, which authorizes the IMF to perform financial and technical services, and a “mandatory approach” requiring “the creation of international treaty obligations, which would, either directly or through specific enactments, become part of the domestic law of participating states.” However, the paper quickly dismisses the “consensual approach” as inadequate from the perspective of dealing with free riders and establishing orderly procedures for negotiations, and concentrates on the design of a “Facility” (meaning mechanism or body, not lending facility or “debt facility”) established through “an amendment of the Fund’s articles or a new treaty.”
The proposed “Facility” combines features of Barnett, Galvis, and Gouraige’s “IMF adjunct” and Cohen’s International Debt Restructuring Agency. Like both Barnett, Galvis, and Gouraige, and Cohen, the IMF Legal Department (1995) favors an arbitration body that would be independent from the IMF. This could be created without a new treaty, through an amendment of the IMF’s Articles of Agreement, either as “a separate organ within the Fund” (following the example of the IMF’s Administrative Tribunal), or by establishing an independent “affiliate institution” following precedents within the World Bank Group.The restructuring process overseen by the facility would involve initiation by the debtor, formation of a creditor committee and/or creditor registry, debtor preparation of a restructuring plan (possibly with IMF participation), preferential status for new financing, approval by a qualified majority of creditors, enforcement “against all creditors, including creditors who voted against the plan” (p. 20), and “monitoring of the debtor’s performance of the terms of the debt adjustment plan” (possibly with IMF support). The paper also suggests endowing the facility with “cramdown” powers along the lines of U.S. bankruptcy legislation: if “one or more classes of creditors fail to approve the plan, the Facility may deem the plan approved if the plan does not discriminate unfairly and a majority of creditors, representing at least two-thirds of the total amount of debt owed to all creditors, have approved the plan.”
Like Barnett, Galvis, and Gouraige, the paper envisages a stay of litigation that would come into effect automatically after initiation. This “would have to become part of the law the signatory states, and domestic courts would thus have to enforce this provision.” The document also resembles Barnett, Galvis, and Gouraige, with respect to the initiation procedures: while the debtor—and in exceptional cases, the creditors—should be entitled to petition the facility for debt adjustment, “the final determination as to whether initiation of the debt adjustment process meets the required conditions will be made by the decision-making organ of the Facility” (that is, not by the creditor committee). The “required condition” is described as “its actual or prospective inability to service external debts as they become due, i.e., illiquidity” (p. 15).
It is worth noting that a “Facility” designed along these lines would have been sufficiently broad to play the role that Sachs wanted the IMF to play, in particular, by (1) not attempting to distinguish between solvency and liquidity crises on initiation; (2) legalizing a standstill via an automatic stay; and (3) giving preferential treatment to new creditors.
While the Legal Department’s document served as the basis for an IMF Executive Board discussion in June of 1995, it did not result in any public initiative, and never registered in public. Instead, the role of giving legal and practical content to Sachs’s version of an international bankruptcy court was assumed by John Chun (1996).31 Even more than Sachs—whose perspective is broader, marrying the experience of the 1994-95 crisis with that of the 1980s—Chun is motivated by the Mexican crisis. The emphasis is squarely on international bankruptcy as an alternative to last resort lending, and the market failure which the author seems to have in mind is primarily a self-fulfilling debt crisis, not an externality arising during debt workouts (although the creditor free rider problem is mentioned to justify a cramdown provision).
Chun’s article was written after the June 1995 Halifax summit, at which both the possibility of an International Bankruptcy Agency (IBA) and a large-scale “Emergency Financing Mechanism” (EMF) were discussed, and takes the form of comparing these two proposals. He comes down strongly on the side of the IBA. His ideas on how this agency would function rely heavily on the Chapter 9 analogy, and include an automatic stay, preference for new financing, monitoring powers for the IBA, and a cramdown provision. Like Barnett, Galvis, and Gouraige, and the IMF Legal Department, Chun argues that the IBA should be established as a separate and independent affiliate of the IMF.
The main novelty of Chun’s article is the way in which the desirability of an independent IBA is argued. The primary comparison is not with a situation of uncoordinated default, but rather with crisis management involving large-scale IMF lending. Chun makes Sachs’s point that a bankruptcy procedure would improve over the IMF’s current international lender of last resort function because of the inherently hesitant nature of IMF lending, which requires a reform program, conditionality, and possibly tranching. In addition, however, Chun argues that an IBA is better than an EMF on the grounds that it does not create moral hazard, a point not made by Sachs. Together with Macmillan (1995b) and the G-10 Working Group (see below), Chun is one of the first authors to explicitly invoke the moral hazard argument to argue for a sovereign debt-restructuring mechanism.32
V. Orderly Workouts Without a Bankruptcy Court
At about the same time at which an international bankruptcy mechanism via the IMF or an independent IMF subsidiary was being explored by Sachs, IMF staff, and others, several authors were working on alternative proposals for “orderly workouts” that take a less sweeping approach. These proposals reject a centralized international bankruptcy mechanism based on a new convention or an amendment of existing treaties, arguing that this either is unfeasible, undesirable, or a combination of both. Instead, they seek to improve crisis resolution via legal changes in some creditor countries, IMF and debtor country policies, and changes in bond contracts between sovereign debtors and private creditors.
The first paper in this category is a short Euromoney article by James Hurlock (1995). Hurlock is one of the first to emphasize that the classic problems of international debt restructuring—uncoordinated litigation, underprovision of new financing, and the holdout creditor problem—are made worse by the shift to bond financing since the early 1990s. He rejects an IMF-based bankruptcy court as a solution to these problems, however, on the grounds that “the Fund is ill-suited to the role of neutral arbiter of sovereign debt disputes” because of its “political nature and voting structure.” Instead, he proposes working through the U.S. and U.K. legal systems to impose a stay and deal with rogue creditors. “The essential predicate would be for certain key nations, such as the U.S., to close their courts on a limited basis to creditors seeking to undermine a legitimate and fair restructuring process that had been endorsed by an overwhelming majority of similarly situated creditors.” This could be achieved by an amendment of the U.S. Foreign Sovereign Immunities Act “that would render a foreign state immune from suit, or its property immune from attachment, if, in the context of a sovereign debt workout, the litigating creditor were attempting to bring suit notwithstanding a restructuring plan that was being negotiated in good faith by, or had been accepted by, a supermajority of similarly situated creditors.”
An extensive study by Eichengreen and Portes (1995) takes a similar view of the underlying problems, but surveys a much broader set of potential solutions. After reviewing the rationale for bankruptcy procedures and the history of institutions for renegotiating sovereign debts, the authors consider the potential role for “an international bankruptcy court or tribunal,” but ultimately reject the idea, primarily for feasibility reasons. Closing the courts to rogue creditors, as suggested by Hurlock, is also dismissed, on the grounds that a change in statute in a single country would not solve the problem, and enacting a treaty seems unlikely given “the trend in recent years away from sovereign immunity,” as well as international human rights law guaranteeing court access.
The authors then come down in favor of a set of pragmatic institutional reform proposals, including the creation of an international “Bondholder Council” to complement the Paris and London Clubs, a redefined role for the IMF that could include sanctioning standstills as a signaling device (“a definitive reinterpretation of Article VIII(2)(b) would support the IMF in this role even if it did not have legal effect in national courts”), a bigger emphasis on information dissemination, and large scale financing in a narrow set of circumstances, including contagion and self-fulfilling runs. The most influential idea among Eichengreen and Portes’s proposals, however, was to use majority clauses in debt contracts as the main device for overcoming creditor collective action problems in the aftermath of a debt crisis. While such clauses had long been included in bonds issued under U.K. law, this was not true for most other jurisdictions, including New York. By pushing for the universal adoption of such clauses, Eichengreen and Portes became the fathers of what is now referred to as the “contractual approach” to orderly crisis resolution.
A third contribution in this group is a comprehensive legal article by Rory Macmillan (1995b). Like Hurlock and Eichengreen and Portes, Macmillan is particularly concerned with coordination problems among bondholders, which he describes in some detail.33 In the literature surveyed here, he is the one of the first to reject large-scale crisis lending on the grounds that it creates moral hazard.34 He shares Hurlock’s distaste for an international bankruptcy agency, particularly in the form of a reborn IMF, but disagrees with Hurlock’s solution on the grounds that it is too heavy handed and insufficiently protective of creditor rights.
His own proposal comprises two main elements. First, the creation of an international bondholder council along the lines proposed by Eichengreen and Portes, or several national bondholder councils in the major financial centers, each representing the holders of bonds issued there. Second, addressing the free rider problem among bondholders, through one of two alternative solutions. The first is a much milder variant of the Hurlock idea of a stay via sovereign immunity, which would apply only to emergency situations, and to bondholders individually but not collectively. “Rather than granting complete immunity to the debtor, legislation might remove bondholders’ rights to sue but vest those rights collectively in the bondholder council.” His second, preferred, approach is to “engineer solidarity” among bondholders by changing some of the rules under which creditors could sue. Specifically, he proposes: (1) “a sharing obligation imposed by simple legislation” that would force bondholders to share payments received from a court judgment with other bondholders; (2) legislation requiring that “all legal actions over the bonds be consolidated into a single legal action”; and (3) a legal minimum threshold of bondholders to bring a suit. This threshold would need to be sufficiently high to solve the free rider problem. Macmillan reckons that 10 percent would be sufficient. Macmillan argues that this approach would also solve the problem of encouraging new financing: “if the incentives of existing creditors can be aligned so that prioritized new money is in their collective interest because there is no alternative, then they will agree to a restricted subordination of their own debt.”
The final contribution in this set is the report of the G-10 Deputies Working Group (1996), written in response to the Halifax summit’s request for “further review by G-10 Ministers and Governors of other procedures that might also be usefully be considered for [the] orderly resolution [of debt crises].”35 The report is close to the pragmatic approach of Eichengreen and Portes (1995)—whose book was commissioned as a background study to the report—but much more cautious. Like Macmillan (1995b), the G-10 Working Group is very concerned with “minimizing moral hazard for both creditors and debtors,” It shares Eichengreen and Portes’s concerns about the feasibility of an international bankruptcy procedure, but in addition is skeptical on the applicability of the bankruptcy analogy as such, on the grounds that the management of economic policies in a sovereign state cannot be taken over by a trustee, and that litigation “has not in the past been a serious problem for sovereign debtors. Such debtors have few assets to seize and some of these benefit from sovereign immunities” (p. 10).
The paper recognizes the potential usefulness of temporary standstills, but unlike Eichengreen and Portes, does “not consider that it would be feasible to operate any formal mechanism for signaling the official community’s approval of a suspension of payments by the debtor.”36 The proposed approach is to encourage standstills “in exceptional circumstances” via IMF lending into arrears. Most of the emphasis is on “contractual or statutory provisions governing debt contracts” that would improve communication between debtor and creditors and discipline rogue creditors. “Such provisions are those that (a) provide for the collective representation of debt holders in the event of the crisis, (b) allow for qualified majority voting to alter the terms and conditions of debt contracts, and (c) require the sharing among creditors of assets received from the debtor.” Developments in this direction should be “market led” but “should receive official support as appropriate.” No specifics are offered on what form this official support could take.
VI. Proposals During 2000–2001
After 1996, there was a brief lull in the literature on sovereign debt workouts, as world attention focused on the Asian financial crisis, which revolved mainly around private debt. However, this soon gave way to a second wave of crises in which sovereign debt played a significant role (including in Russia, Ukraine, Brazil, Ecuador, Pakistan, Turkey, and Argentina). By 2000, proposals to improve the handling of sovereign debt crises were back on center stage. For the most part, they relate closely to the preceding discussion round during 1995-96, but a few go substantially beyond. In what follows, we survey the main contributions prior to Anne Krueger’s November 2001 speech proposing a Sovereign Debt-Restructuring Mechanism, focusing on contributions by Haldane and M. Kruger (2001), Eichengreen (2000), Lerrick and Meltzer (2001), and Schwarcz (2000). As we shall see, these four papers occupy the full spectrum from proposals that would require little or no statutory changes to ambitious statutory initivatives.
Haldane and M. Kruger (2001) propose IMF-endorsed payments moratoria as the centerpiece of a structured crisis resolution mechanism, in which large bailouts are avoided except as a last resort in crises that threaten the stability of the international financial system.37 Their approach has some remarkable parallels with the original Oechsli (1981) proposal. First, it is a nonstatutory approach (no laws or treaties would need to be changed to adopt it, only policies). In addition, it follows the bankruptcy reorganization analogy quite closely, as the authors envisage (1) a payments moratorium legitimized by an independent authority such as the IMF (although this authority would not have particular legal powers); (2) seniority to new financing; and (3) bargaining between debtors and creditors during the duration of the standstill. What is lacking—in both Haldane and M. Kruger, and Oechsli (1981)—is protection from litigation, and a provision to impose agreement on a dissenting minority.
What makes Haldane and M. Kruger a “modern” proposal and differentiates it from Oechsli is, first, the motivation. They think of IMF-endorsed unilateral standstills as a mechanism to deal with both liquidity crises and debt crises. Moreover, like most papers written on this topic since 1995, they worry about the moral hazard implications of large bailouts. Second, unlike Oechsli, they are explicit about how to create incentives for debtor good behavior—a critical issue in their proposal, since there is no independent authority with statutory powers—namely, by offering the “carrot” of IMF lending into arrears. Debtor good behavior is effectively defined as the kind of behavior that would result under a full-fledged, statutory Chapter 11 process. It would be required by IMF conditionality attached to lending into arrears, and would include good faith bargaining with creditors during the standstill (“good faith” being defined as not seeking debt reduction beyond what is necessary to establish medium-term debt sustainability), equal treatment of creditors, giving seniority to new money, transparency (i.e., information provision to creditors), and—perhaps less plausibly—a time limit to the standstill. In other words, the debtor would be expected to lift the standstill and re-expose itself to the creditor grab race if agreement were not reached within a specified time frame (unless the IMF can be convinced that this was because creditors refused a reasonable offer).
Haldane and M. Kruger recognize that the holdout creditor problem is not addressed in their proposal, but argue that this problem is overstated. They also suggest that if the debtor played by IMF-recognized rules of good behavior, this might strengthen its hand in domestic courts. They do not refer to the possibility of using Article VIII Section 2(b) as a basis for “legalizing” standstills.
Barry Eichengreen’s (2000) main thesis is that attempts to limit “the moral hazard caused by IMF bail-outs” are not credible, and will not be effective, so long as the international community does not find alternative ways to resolve sovereign debt crises. Like Haldane and M. Kruger, he argues in favor of a nonstatutory approach, along the lines of his 1995 report with Richard Portes. He concentrates on two of the proposals advanced in that report: IMF-endorsed standstills and collective action clauses in bond contracts. Standstills would deal with liquidity problems and self-fulfilling runs, while collective action clauses would be the main instrument for addressing the “restructuring problem.” Thus, standstills serve a more limited purpose in Eichengreen’s proposal than in Haldane and M. Kruger (2001).
Eichengreen’s view that standstills are good only as panic breakers but insufficient in the context of debt restructuring seems to rest on two arguments. First, standstills do not address the collective action problems associated with debt restructuring negotiations—in particular, the holdout creditor problem—and do not by themselves protect the debtor from litigation. In the case of a panic, this is less of an issue, since the debtor pays in full after the panic is over. Second, unless backed up by Article VIII or changes in national laws, the leverage of the IMF over the debtor in a standstill is relatively weak. The IMF has some effect on the debtor’s reputation and the potential “carrot” of lending into arrears (emphasized in Haldane and Kruger’s proposal), but it cannot influence the debtor by, say, threatening to remove the standstill. This does not matter in a pure panic, which, by definition, can be resolved without corrective actions on the side of the debtor, but it may matter in a debt restructuring, when policy adjustments and good faith bargaining by the debtor are required.
Eichengreen argues that collective action clauses are the right instruments to solve these additional problems, and goes a bit further than Eichengreen and Portes (1995) and the G-10 Working Group in discussing what the official community could do to encourage their adoption. Because there is an externality involved in not adopting such clauses, there is, in principle, a rationale for official intervention. He proposes that “the official community should lead by example, as the British and Canadian governments have done, and that it should subsidize the issue of bonds featuring CACs [collective action clauses], perhaps by having the IMF express a readiness to provide emergency assistance on more attractive terms of countries prepared to adopt the measure” (p. 39).38
Lerrick and Meltzer (2001) propose an IMF-supported debt workout mechanism motivated along very similar lines as Eichengreen (2000). Repeated large-scale crisis lending “creates moral hazard and subverts incentives”, consequently, some other way must be found to address the market failures that generate pressures for bailouts in the first place. The main elements proposed are somewhat analogous to the Haldane and M. Kruger approach: first, an IMF-endorsed moratorium on debt payments; second, debtor-creditor negotiation during the moratorium; third, IMF financial support. However, the form that this support would take is radically different. Rather than lending funds conditionally for general balance of payments use, the IMF would lend unconditionally, for a limited time period, to place a floor below secondary market debt prices at 80–85 percent of the fraction of debt that is deemed sustainable. Thus, the sole purpose of this form of IMF financing would be to prevent debt prices falling below “reasonable” levels during the restructuring period. While the Lerrick-Meltzer approach does not require a new treaty or changes in national laws, this constitutes a large departure from the IMF’s traditional role that could require an amendement of the Articles of Agreement.39
In contrast with Haldane and M. Kruger’s proposal, the role of this form of IMF support is not to create good debtor incentives while the country is restructuring (Lerrick and Meltzer do not believe this to be a significant issue), but to prevent creditor panic and contagion by maintaining the liquidity of the sovereign debt market during the restructuring period. Thus, Lerrick and Meltzer would disagree with Eichengreen’s (2000) view that “officially sanctioned standstills,” by themselves, constitute a “solution to the panic problem.” In part, this might be because Eichengreen has a broader version of standstills in mind that would restrict most cross-border flows, not just sovereign debt payments. What mainly drives Lerrick and Meltzer’s proposal, however, is the view that it is primarily the collapse of debt prices itself that fuels panic and contagion. With debt traded on international markets, this could occur even with perfect capital controls, and thus would require an extra instrument to address it.
As in the Haldane and M. Kruger approach, there is no legal instrument in Lerrick and Meltzer’s proposal that would protect the debtor from litigation and impose a majority-backed agreement on holdouts. Lerrick and Meltzer argue that this is not necessary, both because the international assets of a sovereign debtor are hard to attach, and because the existence of a debt price floor would reduce the profitability of holding out. “Vulture funds” could no longer buy debt at extremely low prices, and the margin between the price floor and the face value of the debt might not be worth the costs of a court battle.
Steven Schwarcz (2000) presents perhaps the most comprehensive legal treatment so far on how the provisions of Chapter 11 (or Chapter 9, the distinction is dismissed as secondary) could be applied at the international level. According to Schwarcz, the status quo embodies three inefficiencies: “the collective action problem of reaching agreement among creditors,” “moral hazard” created by IMF bailouts (both vis-à-vis countries and creditors), and the underprovision of new private financing, leading to an excessive reliance on public money via the IMF. To address these, he proposes “a supranational legal framework for sovereign debt restructuring,” a draft of which is attached to his paper. He argues that “recent proposals to contractually solve the collective action problem in bonds by introducing supermajority voting clauses in new bond issues are unlikely to be successful” because “only consenting bondholders would be bound. As a result, a state cannot rely on a contractual approach to bind holders of the large stock of existing long term bonds, much less future creditors that choose not to consent” (p. 166).
However, Schwarcz takes a much more minimalistic approach than previous statutory approaches. His intellectual strategy is to set out desirable principles for an international Chapter 11—basically, fostering economic rehabilitation of the debtor while “minimally affecting non-bankruptcy incentives”—and ask how these could be attained with as little “adjudicatory discretion” as possible. The result is remarkable for what it does not contain: (1) no automatic “stay” of creditor claims following inititation, on the grounds that the sovereign can declare a payments moratorium, and creditors have no significant recourse against this since there are “relatively few assets located in other jurisdictions”; (2) no officially organized creditor committees, since his convention would provide sufficient incentives for creditors to organize themselves on an ad hoc basis; (3) no cramdown rule (over and above the basic supermajority rule) to impose a restructuring agreement on a dissenting creditor class, on the grounds that implementing cramdown requires valuing the debtor as a going concern, which is hard to do for sovereigns;40 and (4) no international bankruptcy court or greatly extended role for the IMF, on the grounds that his proposed convention would be largely self-enforcing.
What survives from Chapter 11 in Schwarcz’s convention are three simple proposals regarding initiation, new private financing, and supermajority agreement to a restructuring plan: “(1) only a State itself, and not its creditors, may commence the restructuring case, and must do so in good faith; (2) financiers of the debtor-State’s debt restructuring must have priority over claims of other creditors …; and (3) all creditors be bound to a plan or reorganization that is agreed to by super-majority voting by classes of claims, and, upon such agreement, debts not provided for in the plan be discharged. The Convention would also require each ratifying State to enact the Conventions’ rules into national law” (pp. 158–9).
Schwarcz argues that while a full-fledged international bankruptcy court would be superfluous under his convention, “a tribunal would be required to settle interpretive disputes in very limited circumstances” (p. 179). For this limited role he proposes creating an arbitration panel along the lines of the World Bank Group’s International Centre for the Settlement of Investment Disputes. He also sees a role for the IMF as “the location where States file their debt-restructuring cases,” to help the arbitration panel decide whether a filing occurred in good faith and if excessive new financing undermined the rights of existing creditors, and as a source or intermediary for interim financing.
VII. Comparing Ideas, 1976–2001: A Synthesis
This study set out to describe the development of ideas on international bankruptcy on the basis of two questions First, what is the economic inefficiency that the proposals seek to address? Second, how do they propose to change incentives in a way that would address this inefficiency? We now summarize the answers to these questions.
A. The Basic Conundrum: Collective Action Problems and Moral Hazard
As far as the perceived inefficiencies go that motivate the proposals for institutional and legal improvements in sovereign debt restructuring, there is a clear progression. From the late 1970s to the mid-1990s, the emphasis was on inefficient debt workouts caused by various incentive problems on the creditor side. For the early contributors, particularly Oechsli (1981), the main problem was a coordination failure between the public and private sectors. Most subsequent authors characterized the problem as free riding among private creditors in the context of debt workouts, but emphasize different aspects. The legal literature, particularly Barnett, Galvis, and Gouraige (1984) and IMF Legal Department (1995), viewed a potential “race to the courthouse” as a significant threat impeding orderly negotiations. The economic literature, including Sachs (1986a, b) and B. Cohen (1989 a, b) was more concerned with the underprovision of new financing and free riding that impedes or delays the successful conclusion of debt restructuring negotiations. Since a debt writedown by a portion of creditors will increase the capacity of the debtor to repay the remaining creditors, such free riding could be an impediment to restructuring even when litigation is not an issue (assuming there are other reasons that keep the debtor from simply repudiating).
Beginning with Sachs (1995), the literature—motivated by the Mexican crisis—emphasizes an additional inefficiency: debt panics, that is, self-fulfilling runs. In technical terms, the difference between this and inefficient debt workouts is that the former emphasizes multiple equilibria (one of which, the “run” equilibrium, is inefficient) while the latter involves a unique, but inefficient, equilibrium. In more practical terms, the difference is between a situation where a creditor collective action problem can cause a crisis, and one where it is an obstacle to its efficient resolution. While the distinction had been made in the academic literature since the early 1980s, it was only after the Mexican crisis that bankruptcy mechanisms of various kinds were proposed as a solution to both problems (Sachs, 1995; Eichengreen and Portes, 1995; Chun, 1996). Sovereign bankruptcy mechanisms or agencies were now being suggested not just as a complement of existing institutions such as the IMF, whose traditional responsibilities include lending to members with liquidity problems, but as alternatives to traditional crisis lending. This view was reinforced by a growing consensus that large-scale crisis lending à la Mexico created “moral hazard,” particularly on the creditor side (Macmillan, 1995b; G-10 Working Group, 1996; and Chun, 1996).
The desire to avoid the moral hazard attributed to large-scale crisis lending is common to virtually all proposals related to orderly debt workouts that have been made in recent years (Haldane and M. Kruger (2001), Eichengreen (2000), Lerrick and Meltzer (2001), Schwarcz (2000), Krueger (2001, 2002) and Taylor (2002)). In these papers, collective action problems are still viewed as the primary inefficiency that could justify public intervention in the form of IMF bailouts, but the focus is now on the new distortions introduced by the bailout itself. An international bankruptcy mechanism of some form is viewed as an alternative, less distortionary, way of dealing with the underlying market failures, and perhaps a necessary element in committing the international community not to undertake large bailouts.
B. Dealing With Creditor Incentives: Competing Approaches
Given the shared concern with a creditor collective action problem in some form, all papers in this literature contain proposals to change creditor incentives almost by definition. However, they vary widely on how to approach the issue. One end of the spectrum is occupied by what is now referred to as “statutory” proposals, including Barnett, Galvis, and Gouraige (1984), B. Cohen (1989), Kaeser (1990), Raffer (1990), Reinisch (1994), Greenwood and Mercer (1995), Hurlock (1995), IMF Legal Department (1995), Sachs (1995), Chun (1996), and Schwarcz (2000), among others. These papers propose changes in either national or international law to create rules or institutions under which (super) majority-backed agreements could be imposed on holdouts, new financing could be given seniority, and—in some proposals—sovereigns would be shielded from litigation during payments moratoria or while negotiating. The main alternative to the statutory proposals is the “contractual approach,” which focuses on the way contracts between creditors and debtors are written. This is associated with Eichengreen and Portes (1995), the G-10 Deputies Working Group (1996), Eichengreen (2000), and to a lesser extent Macmillan (1995b).41 These authors propose addressing the holdout problem—at least for a given debt instrument—through the incorporation of clauses in bond contracts that would impose a majority-backed restructuring on dissenters, and perhaps require the sharing of proceeds from litigation. There remains a potential for free riding across debt instruments—referred to as the “aggregation problem” in the more recent literature—although this could perhaps be contractually addressed as well.
Broadly speaking, the contractual and statutory approaches differ in two respects. First, the pure contractual approach focuses on resolving the basic holdout problem and—in some proposals—on discouraging litigation. In contrast, statutory proposals usually take aim at a broader range of problems, including providing seniority to new money by administrative fiat Second, by definition, the statutory approach requires changes in laws, while the contractual approach does not. However, the incorporation of collective action and sharing clauses in debt contracts is of course an endogenous variable, raising the question of how creditors and debtors can be persuaded to adopt them. The simplest answer, given by Macmillan (1995b), is: through legal changes in the major creditor countries, which would take one back to a version of the statutory approach. However, there may also be nonstatutory incentives as suggested by Eichengreen (2000): publicly subsidizing changes in private contracts, or IMF conditionality.
Finally, some proposals—notably Oechsli (1981), Haldane and M. Kruger (2001), and Lerrick and Meltzer (2001)—focus on official policies and do not propose any legal changes, whether contractual or statutory. These proposals tend to be less concerned with free riding among private creditors and more with other particular inefficiences: Oechsli with coordination failures between private and official creditors; Haldane and M. Kruger with sudden stops and creditor panics as well as debtor incentives to negotiate; and Lerrick and Meltzer with creditor panics and contagion. The last two papers recognize that their proposals will not directly deal with litigation and holdouts but play down the magnitude of this problem.
C. What About Debtor Incentives?
Most papers discussed in this survey ignore bad debtor incentives as a potential reason why debt workouts can be protracted and inefficient. Since negotiation delays and perhaps failures could in principle arise from debtor actions as much as creditor actions—either as a consequence of strategic behavior, or because the debtor side exhibits collective action or political economy problems of its own—this is a rather striking omission. However, the majority of authors surveyed here do recognize implicitly or explicitly that debtor incentives can become an issue as a byproduct of the measures that are put in place to address the creditor collective action problem. A stay of litigation, protection from holdout creditors, or an IMF-endorsed standstill, imply a measure of debtor protection from the usual market discipline. What is to guarantee that this protection is not abused?
One answer is that having debtor countries sign an international convention that subjects them to certain rules might by itself induce good behavior, just like countries that have signed the IMF Articles of Agreement tend to abide by its rules (Schwarcz, 2000, footnote 210). But this argument applies only to the statutory proposals, and even there raises questions of remedy in the event that rules are broken. An alternative answer is that good incentives for the debtor can be created by the threat of reverting to the status quo ante in the event that the debtor does not negotiate in good faith. For example, in Haldane and M. Kruger (2001), the “carrot” of IMF endorsement (including lending into arrears) would be conditional on country good behavior. A similar argument can be made for proposals that envisage a stay of litigation during and after negotiations, as in Barnett, Galvis, and Gouraige (1984) or IMF Legal Department (1995).
An interesting implication of this argument is that transposing bankruptcy procedures to the international level need not require an arbiter with statutory powers vis-à-vis the debtor. All that is required to create incentives for debtor good behavior is the presence of an initial inefficiency that hurts debtors, and which the arbiter (or a majority of creditors acting collectively) would have the power to remove and reinstate if necessary. Whether this would in fact be the case under the specific proposals that have been put forward is of course not obvious. It depends on why defaults are costly in the first place, and whether the proposals contain anything that would affect that cost. In particular, how important are the potential legal costs of defaults relative to other mechanisms, such as reputation? And how successful would an orderly bankruptcy procedure be in mitigating the reputational losses arising from a debt restructuring? While the economics literature has attempted to address some of these issues since the 1980s, there are no hard answers to these questions.
A related problem is whether smoother debt workout procedures could undermine debtor incentives ex ante, that is, create debtor moral hazard—for example, with respect to fiscal policy and debt management, or simply because debtors may try less hard to avoid defaults. A possible consequence would be a higher cost of borrowing and lower capital flows. Only a minority of the papers we surveyed—particularly Eichengreen and Portes (1995), Macmillan (1995b), Eichengreen (2000), and Haldane and M. Kruger (2001)—discuss this potential problem in some detail. Predictably, most of these authors conclude that debtor moral hazard is not an issue we should be overly worried about. A mix of the following arguments is invoked: (1) even from the perspective of getting debtor incentives right ex ante, the costs of defaults may be too high at present (for example, they may encourage gambles for redemption—anything to avoid defaults); (2) assuming defaults do indeed become more frequent as consequence of lower default costs, this may well be efficient—it may be better to have occasional low-cost debt restructuring than more infrequent but disastrous crises; (3) if that is the case, it is not obvious that the cost of capital would in fact go up, since the likelihood of debt restructurings would rise but creditor losses conditional on debt restructurings would fall;42 and (4) even if the cost of capital increases and the average volume of capital flows declines, this may be welfare improving if it goes along with a higher stability of capital flows.
The general challenge to this line of argument is as follows. Sovereign lending lacks collateral and/or judicial contract enforcement of the kind that we see at the domestic level. Consequently, could it not be that the painful consequences of defaults are the market’s answer to the lack of collateral, the device that makes international debt markets possible? And if so, will official or institutional tinkering with these default costs not make things worse?43 The answer, in equally general terms, is “yes” only if debt-restructuring costs are an efficient market response (in particular, through the way in which contracts are written) to the lack of collateral or enforcement. However, if it is impossible to write complete debt contracts (for example, because the debtor cannot commit to refrain from further borrowing), that may not be the case (Bolton and Jeanne, 2002). Inefficient capital structures may arise, for example, excessive short-term borrowing. In that case, new institutions, such as an international bankruptcy mechanism, can bring debt outcomes closer into line with the first best. Whether the existing proposals would achieve that is a separate question, but in principle, the argument can be made.
We have characterized the development of ideas on international bankruptcy over the last 25 years. Turning to the present, how do the most recent proposals fit in? Krueger’s (2001) proposed “Sovereign Debt-Restructuring Mechanism” (SDRM) is a statutory proposal in the tradition of Barnett, Galvis, and Gouraige (1984) and IMF Legal Department (1995). It defines an important role for the IMF, particularly in endorsing a debtor country’s request for activation of a stay, and in revoking the stay (or refusing to extend it) if the country does not implement corrective policies and negotiate in good faith. In contrast, the statutory approach in Krueger (2002) these roles would be played by a majority of creditors acting collectively, rather than an intervening international agency. In addition, Krueger envisages an independent court-like arbitration panel to resolve disputes, but with narrow functions, along the lines of Schwarcz’s (2000) proposed tribunal for settling interpretational disputes. Finally, Taylor (2002) is a contractual proposal in the tradition of Eichengreen and Portes (1995), but goes further in two respects. First, by suggesting clauses stipulating how the debt-restructuring process would be organized in addition to the standard majority action clause; and second, by being more explicit on the incentives that might encourage the adoption of such clauses. Taylor suggests that having such clauses might be a precondition for receiving IMF financial support and/or that countries might qualify for lower IMF interest rates.44
Figure 1 presents a classification of most proposals covered in this paper in terms of the main inefficiency that motivates them (vertical axis) as well as their basic approach: changes in official policies, changes in debt contracts, or changes in laws and treaties at either the national or international level (horizontal axis). These classifications do not do full justice to many papers. For example, Eichengreen and Portes (1995) and Eichengreen (2000) are not just about the contractual approach and changes in official policies, but include some statutory ideas at the international level as well, and Kaeser (1990) is not just concerned with the efficiency of workouts but also with preventing overindebtedness. Moreover, some proposals are missing from the figure because they have ambiguous classifications in terms of Figure 1. Williamson’s (1985, 1992) call for an international debt-restructuring agency would seem to put him in the right column, but he also suggests that an IDRA could be established mainly on a private contractual basis. Lerrick and Meltzer (2001) think of themselves as an approach based only on changes in official policies, but the IMF’s Legal Department believes that it would require statutory changes (to the IMF’s Articles). In spite of these shortcomings, the figure may be useful as a rough summary.
At first blush, the message of the figure is simple: while it is easy to order the various proposals so that there is a steady chronological progression along the vertical axis, that is, in terms of motivation, the same cannot be said of the horizontal axis. In particular, the proposals put forward during 2000–2001 cover the entire horizontal spectrum, from an approach that is mainly based on changes in IMF policies (Haldane and M. Kruger, 2001) to one that calls for far-reaching statutory changes (Krueger, 2001).
However, the figure also suggests a trend of a different kind. Not only is the history of ideas on sovereign bankruptcy procedures far longer and far more elaborate than is commonly known, but the call for sovereign bankruptcy reform, which began 25 years ago as a fringe phenomenon pushed by some debtor countries and a few academic sympathizers, has steadily moved into the academic and policymaking mainstream. Although there continue to be substantial disagreements on which approach is the best, the basic idea of applying domestic bankruptcy reorganization to sovereigns is now supported by key representatives of both the official creditor community and private creditors.45 The reason for this new consensus appears to be a deep discontent with conventional tools for crisis resolution—in particular, large-scale crisis lending—which is unlikely to go away. In addition, very recently, there seems to have been a narrowing of positions on how to reform sovereign bankruptcy. The main poles of the debate are now the contractual approach proposed by Taylor (2002), who addresses the problem of implementation incentives more explicitly than his predecessors, and Krueger (2002), who dispenses with the notion that a statutory approach to sovereign bankruptcy reorganization requires concentrating broad decision-making authority in the hands of multilateral institutions. Judging from these developments, it seems likely that the call for orderly workout procedures will have permanence, and result in actual reforms in the foreseeable future.