IMF and the Private Sector—A Factsheet
Public Information Notice: IMF Board Holds Informal Seminar on Sovereign Debt Restructuring
Involving the Private Sector in the Resolution
of Financial Crises: A Staff Note on the Lerrick-Meltzer Proposal|
Prepared by the International Capital Markets, Policy Development and Review, and Research Departments in consultation with other Departments
April 10, 2002
1. At the informal briefing for Executive Directors on December 14, 2001 on A New Approach to Sovereign Debt Restructuring—Preliminary Considerations1 a question was raised concerning the recent proposal by Professors Lerrick and Meltzer for a mechanism to strengthen the ability of members that have unsustainable debt burdens to reach agreements on an orderly restructuring.2 This note, which is issued for information, provides a summary and a preliminary evaluation of the Lerrick and Meltzer proposal.
2. Professors Lerrick and Meltzer argue that members with unsustainable debt burdens need to reach agreement with their private creditors on restructurings that restore fiscal and balance of payments viability, and that the costs of any debt and debt-service reduction should be borne by the private sector.3 In practice, however, they consider that under the current institutional arrangements the costs of a sovereign restructuring are so great that sovereigns, and the international community more generally, will resort to large-scale bailout packages in order to avoid defaults. These allow for the possibility that private investors exit whole, and thereby generate moral hazard.
3. Lerrick and Meltzer argue that much of the cost of a default stems from a market failure. Specifically a default is likely to trigger panic, which would lead to an overshooting of the secondary market debt prices. This in turn leads to margin calls and demands for redemptions by retail investors in mutual funds, which forces fund managers into further rounds of panic selling, thereby transmitting the problem to other emerging markets. In conditions of financial panic, market liquidity would be low, and so relatively small sales could have large impacts on secondary market prices. This is illustrated by reference to the movements of secondary market prices following the defaults by Russia and Ecuador.
The Lerrick-Meltzer proposal
4. In order to lower the costs of default and to make it a more acceptable policy option for addressing extreme cases, Lerrick and Meltzer propose that the Fund (and the international community more generally) should use its resources to establish a floor in the secondary market price of debt of a country seeking a restructuring. The floor would be set at, say, 85 percent of the Fund's estimate of the long-run value of the debt. Lerrick and Meltzer argue that by providing bondholders a put option at zero cost to the purchaser, Fund resources would be used to correct a market failure associated with overshooting of secondary market prices. At the same time, they argue this would avoid creating a market failure stemming from the moral hazard associated with large-scale packages.
5. Under the proposal, a sovereign that had an unsustainable debt burden would simultaneously make three announcements: (i) a moratorium on foreign currency debt to the private sector; (ii) plans to offer a restructuring within a short period (say, three months) that would have a minimum specified restructured value; and (iii) the IMF had agreed to establish a line of credit that could be used to finance debt purchases during the period of the restructuring. Under the proposal the Fund would make its resources available in cases in which the member had elaborated and announced a fully credible and comprehensive adjustment program. In addition, the Fund would announce that it is committed to provide support to other emerging market countries affected by contagion.
6. The official cash support bid under the proposal would take the form of a collateralized loan from the Fund to the member. A trustee would receive irrevocable instructions from the Fund and from the member to purchase any bonds offered at the official cash support price and to draw automatically on the Fund loan facility for payment.4 The trustee would serve as collateral agent for the Fund and would hold the tendered bonds as security for the loan. If any bondholders accept the official bid and Fund resources are advanced, the member would agree to retire its debt to the Fund on a fixed repayment schedule and to dedicate one-half of the proceeds of all new borrowing to early redemption to ensure a rapid recycling of Fund resources. As the borrower repays the Fund loan, a proportionate share of the bonds held as security would be released and canceled.
7. Although the specific features of the proposal are open to debate, the novel feature—that warrants further consideration—is for the official community to try to assist countries in crisis by providing derivatives (in this case a put option) instead of conventional loans. In the following, this proposal is compared with conventional loans in terms of its effectiveness in mitigating the economic damage of a crisis and providing appropriate incentives, and its practical and legal feasibility.
Effectiveness and incentive issues
8. From the perspective of a country in crisis, the most obvious difference between receiving IMF assistance via a conventional crisis loan or via a credit line as envisaged by Lerrick and Meltzer's proposal is the much narrower use of the latter. Conventional Fund assistance, including the CCL, can be used in multiple ways: to support a new exchange rate policy after a devaluation, to smooth the path of fiscal adjustment, to back the domestic financial system, to buy back debt at a low price, etc. Lerrick and Meltzer's credit line would only serve to prevent debt prices from overshooting. How effective is this in reducing economic dislocation in a crisis? A crisis country may face a free fall of the exchange rate, a banking crisis, corporate distress, capital flight and a collapse in fiscal revenues. Stabilizing debt prices is not likely to ameliorate those problems a great deal.
9. The Lerrick and Meltzer proposal would also crowd out other forms of financial assistance that may be more urgently needed. A large commitment of funds under the proposal would place serious limits on any other form of financial support from the international community, and would not generate much direct economic support for the domestic economy.5 The only direct benefit, perhaps, would come from limiting the losses incurred by domestic banks in their holdings of government debt in the form of international bonds. But this would not be a very well targeted method of supporting bank recapitalization.
10. The Lerrick and Meltzer proposal may help contain contagion in certain circumstances. However, in its current form at least, the Lerrick-Meltzer proposal is only activated in the context of a default, which may be too late in many cases. For example, if a default is widely anticipated it will not generate much contagion when it happens, although some more significant spillovers would take place well before the default.
11. It may be argued that the Lerrick and Meltzer proposal would help speed up debt restructuring while conventional crisis management does not. However, encouraging the renegotiation of debt when necessary, including in the context of a moratorium, is part of the conventional approach through the Fund's policy of lending into arrears. There remain important difficulties in the existing debt restructuring process, including protecting the debtor from litigation while negotiations are ongoing, and the holdout creditor problem. However, neither the Lerrick-Meltzer proposal nor the conventional approach directly addresses these problems. In particular, the incentive to hold out is possibly reduced but not eliminated under the proposal.6
12. To the extent that creditor moral hazard is a serious problem, the proposal would help to address it, since it guarantees that public funds will only be used in situations when creditors experience substantial haircuts. Indeed, one of the main motivations of the proposal is concern about the moral hazard associated with big bailouts. However, a similar effect could arguably be achieved by making traditional crisis lending conditional on negotiated debt restructuring, and lending into arrears.
13. The effects of the proposal on debtor incentives are more mixed. Once the support price has been set,7 if a debtor government engages in irresponsible policies or simply "scares" investors into selling their holdings to the IMF at the distressed price, it would end up benefiting from the ensuing reduction in the face value of its debts. Such actions would be shortsighted and ultimately hurt the debtor country, but they are possible under the rules. Of course, Fund programs can go off track owing to irresponsible policies as well, but conditionality and tranching limits the opportunity for reckless strategies. In the case of the Lerrick-Meltzer proposal, there is no conditionality after the Fund establishes its credit line, and the Fund would be unable to suspend the program under any circumstance.
14. There are several dimensions of the Lerrick-Meltzer proposal, which, taken together, undermine its feasibility as a useful instrument in crisis resolution when compared to conventional lending in the context of a Fund supported program.
15. Under the proposal, even in the nadir of crises, the Fund is assumed to be able to make reliable estimates of members' long-run capacity to service debt. These estimates would be used by the Fund to establish a support price. However, in such circumstances, it is very difficult to pin down members' medium-term capacity to generate resources for debt service, which depend on assumptions about the future path of such variables as GDP growth, interest rates, and fiscal adjustment.8 To be sure, conventional crisis lending also requires an analysis of the member's ability to repay the Fund, but this is less demanding than to estimate the market value of debt with sufficient accuracy to set an exact floor price. Moreover, the consequences of an incorrectly estimated price could be very serious, since the Fund may end up holding a substantial portion of the member's debt. Since the proposal does not envisage the usual safeguards such as conditionality during the implementation of the program or tranching of Fund resources, the only way the IMF could protect itself from this risk would be through a very conservative estimate of the debt recovery value. But in that case, the price support mechanism would not have much of an impact on secondary market prices.
15. Depending critically on market reaction to the announcement of a restructuring, the implementation of the proposal may require large-scale use IMF resources:
In summary, the announcement of an IMF-supported price floor may help to resolve the debt crisis if it is part of a convincing overall package. However, it will not, in general, remove the possibility of a panic in which investors rush to the exit. The latter would imply large-scale use of Fund resources unless the Fund established a very low floor price, in which case it would defeat the purpose of the operation.
17. The proposal does not address several operational and legal issues. There is no discussion of what happens if no agreement is reached during the restructuring period or of how long the price support would need to be maintained, and the credibility problems the Fund may face in this regard. There is also no discussion of how a situation in which policies deteriorate and the member is unable to repay the Fund would affect the Fund's financial position.
18. Finally, there is a question of consistency of the proposal with Article VI of the Articles of Agreement, which precludes members from using the Fund's resources to meet a large or sustained capital outflow. Financing under the proposal would be directly linked to capital account outflows and would be distinct from the current situation in which Fund financing in capital account crises is justified on the basis of restoring confidence and replenishing reserves.
Simulation of the proposal
19. In November 2001, Professors Lerrick and Meltzer arranged for a simulation of their proposal using technology originally developed for war games. It attracted a high quality of participants from the private sector.9 In the simulation, a restructuring by the fictional debtor country of "Mañana" had been widely anticipated, giving room for some capital gain in an eventual restructuring. Mañana was able to reach agreement on a restructuring on terms consistent with the views of the debtor, the IMF, and market participants regarding medium-term sustainability. Minimal use was made of the Fund's price support mechanism. Mañana was able to impose collective action through the aggressive use of exit consents to limit incentives for dissident investors to hold out.10
20. In a number of important respects the simulations assumed away many of the problems that are likely to come to the fore in the real world. For example, the simulation assumed that (a) at the time when the government announced its intention to restructure, it had a fully-elaborated and credible program, (b) there were no uncertainties surrounding the stability of the banking system and the sustainability of the exchange rate, (c) the restructuring had been widely anticipated and the debt had been traded down to a very low level, thereby giving room for some capital gains in an eventual restructuring, and (d) once the sovereign achieved a 65 percent of principal participation rate, exit consents could be used without legal challenge to strip away the enforcement provisions of the remaining bonds, thereby creating strong incentives for holdouts to tender their instruments for the exchange.
21. The proposal of Professors Lerrick and Meltzer is aimed at reducing contagion and the damage caused by investor panic when a country needs to restructure its debt. By so doing, it aims to encourage countries to restructure rather than avoiding this by drawing on large Fund financing packages, and to encourage investors to weigh risks more carefully. The staff believes that these are important objectives, but does not believe that the proposal provides a practical solution.
22. The proposal would not do much to ameliorate the economic situation of the debtor country. It would help to address potential creditor moral hazard problems, but after the floor price for debt has been set, the framework could also reward irresponsible policies by the debtor country. The proposal would commit large financial resources by the Fund and the international community to support debt prices, effectively crowding out other forms of international support for crisis management. It would only be activated in the context of a default, limiting its usefulness in preventing contagion in cases in which overshooting of debt prices has already occurred. While the proposal may help a faster resolution of debt crises, it does not fully address issues of collective action and holdout creditors, and it does not protect the debtor from litigation.
23. There are also a number of problems regarding the feasibility of the proposal. To be in a position to implement the proposal, the Fund would have to have access to a much larger pool of resources than is now the case. Also problematic is the ability of the Fund to determine the floor price with the accuracy that the proposal requires, as the cost of overestimating a member's repayment capacity would be high and an underestimation would make the price support irrelevant. It is also not clear how the Fund could ensure that negotiations succeed within the envisaged period, and whether the operations envisaged by the proposal are consistent with the Articles of Agreement.
24. In summary, the staff believes that the proposal would not be a valuable addition to the existing instruments of the Fund and its members.
1A New Approach to Sovereign Debt Restructuring—Preliminary Considerations, FO/DIS/01/151 (11/30/01), (pdf file, 725 kb).
2See "Blueprint for an International Lender of Last Resort" unpublished manuscript, October 2001. See also, "Beyond IMF Bailouts: Default without Disruption," Quarterly International Economics Report, May 2001.
3Lerrick and Meltzer argue that in order to achieve a restructuring in which private creditors make genuine concessions, the sovereign must first default. They appear to draw no distinction between restructuring conducted against the background of credible threats of default and actual defaults.
4In fact, under the Articles, the Fund can only disburse to a member, not to a trustee.
5Lerrick and Meltzer give the example of a country with $90 billion in foreign debt and a cash support price of 50 percent of face value, which would imply a commitment of $45 billion by the IMF and the international community.
6Creditors with a penchant to sue may not be dissuaded by the availability of a cash support price, and the buyback trustee arrangement could well create room for creative attachments and other litigation maneuvers. Moreover, the proposal could increase the chances of (or at least skew the incentives related to) the collective action problem: e.g., a larger number of creditors may become holdouts, anticipating that the buyback arrangement will reduce debt and thus increase the returns of those that wait. This is the typical approach pursued by vulture creditors.
7Prior to the announcement of the support price, the proposal would not raise debtor moral hazard issues because the level of the support floor has not been set yet.
8In addition, defaults may occur before policy understandings on the basis of a Fund-supported program have been reached, which would make these projections even more difficult. For example, Ecuador defaulted several months before the approval of a Stand-By Arrangement. In the case of Ukraine, the bond restructuring was announced when the Fund-supported program was off track, and the next drawing from the Fund occurred almost eleven months after the completion of the bond exchange.
9A Fund staff member attended the simulation as an observer.
10In the simulation all bonds were assumed to include contractual provisions that allowed the use of exit consents. Moreover, in the simulation Mañana's use of exit consents was not subject to legal challenge.
IMF EXTERNAL RELATIONS DEPARTMENT