The Restructuring of Sovereign Debt - Assessing the Benefits, Risks, and Feasibility of Aggregating Claims

The Restructuring of Sovereign Debt - Assessing the Benefits, Risks, and Feasibility of Aggregating Claims


The Restructuring of Sovereign Debt - Assessing the Benefits, Risks, and Feasibility of Aggregating Claims

Executive Summary

At its Spring 2003 meeting, the IMFC concluded that it was not now feasible to establish the Sovereign Debt Restructuring Mechanism (SDRM). It did, though, agree that work should continue on issues raised in discussion of the SDRM that relate more generally to the orderly resolution of financial crises, including aggregation.

There is now a growing recognition that a legal framework that aggregates different creditor claims for voting purposes could help the restructuring process. It could be of particular benefit in resolving collective action problems. The challenge is to devise a framework that maximizes the potential benefits while minimizing the risks.

There are two principal benefits. The first is the reduction in “holdout” risk that aggregation might bring. Any system that made it more difficult for a small minority of creditors to block a restructuring deal would be attractive. Collective action clauses, as currently used, are only able to do this to a limited extent. Aggregation could also improve creditor coordination significantly, not least by providing an incentive for creditors to organize at an earlier stage in the process. It may also be possible to design an aggregation framework that would make it easier to deal with inter-creditor equity issues, at least in the post-default context, by ensuring that all creditors be offered the same terms, or menu of terms.

But there are significant risks with aggregation and even where the perceived risk may be greater than the actual, any widely acceptable aggregation framework would have to overcome them. First, is the risk that debtors, in alliance with the majority of creditors, may be able to discriminate against a minority: this is most likely to occur in cases where some debt is secured, or, where the debt is restructured prior to a default, when different residual maturities are involved. Second, aggregation may create some rigidity in circumstances where the debtor may wish to provide different terms to creditors, reflecting their different economic interests. And third is the scope for potential abuse by the debtor who might try to manipulate the voting process by placing real or fictitious claims into the hands of creditors that are under the influence of the sovereign.

The SDRM aimed to balance these benefits and risks in a satisfactory way, but it does not currently have the support needed to pursue it further. Given this, efforts are under way in several quarters to devise an effective contractual framework that would capture the substantial benefits of aggregation while, at the same time, minimizing its risks. One option is to introduce some modification into the collective action clauses now being more widely adopted. In its most recent bond issues, Uruguay included provisions in the CACs for a limited form of aggregation. But in order to counter fears about the risks involved in aggregation, these provisions include important safeguards that limit some of the benefits of aggregation.

Other ways of designing contractual frameworks that would achieve some degree of aggregation are also being explored by the private sector. One is a “two-step” proposal by J. P. Morgan: it would allow creditors to exchange their existing instruments for an interim

debt claim that, in effect, would put all creditors on an equal footing; providing the framework for the second stage, the final restructuring. Another is the establishment of the Argentine Bond Restructuring Agency (ABRA). This works in a somewhat similar way to the “two-step” proposal, except that the first stage “swap” is reversible, albeit subject to a financial penalty.

Given the complexities involved and the fact that most approaches under consideration are still at an experimental stage, it is too early for the Fund to endorse any one approach. Much uncertainty also surrounds the market’s view of aggregation. But it does seem clear that aggregation offers important benefits, and, based on the efforts underway, that some progress —however incremental—can be achieved in the contractual context.

I. Introduction

1. At its Spring 2003 meeting, the IMFC, while recognizing that it is not feasible now to move forward to establish the Sovereign Debt Restructuring Mechanism (SDRM), agreed that work should continue on issues raised in its development that are of general relevance to the orderly resolution of financial crises, including aggregation. The Committee asked the IMF to report on progress to its next meeting. This paper is a step toward responding to this request.1, 2

2. Proposals for strengthening the arrangements for sovereign debt restructuring—whether contractual or statutory—share the same objective: to make the restructuring of unsustainable debt more orderly, predictable, and rapid.3 As has been recognized, the achievement of this objective requires addressing a number of weaknesses in the existing system, including collective action problems and creditor coordination issues. When designing a mechanism that will address these weaknesses, the question arises as to whether the “aggregation” of creditor claims for voting purposes would make an important contribution to the resolution of these issues. For example, the frameworks that have been established to resolve collective action problems are all based on the ability of a qualified majority of creditors to make decisions that are binding on all creditors. Such a framework, however, begs a broader question. For purposes of this decision, is the qualified majority that is needed to make the decision—and the overall creditor body that is bound by it—calculated on the basis of those creditors holding the same bond issuance or participating in the same bank syndicate? Or, alternatively, when calculating both the voting majority and the overall creditor body that is affected, should claims across different instruments be “aggregated”?

3. If one accepts the premise that collective action difficulties are exacerbated by the multiplicity of instruments and growing diversity of creditor interests, it is reasonable to conclude that there are benefits in achieving some degree of aggregation. While binding a minority within an issuance resolves the collective action problem for restructuring of that issuance, it does not necessarily resolve the collective action problems arising among different instruments. Thus, the potential benefits of aggregation across instruments.4

4. Although achieving aggregation has been a key attractive feature of the SDRM proposal, there is also a growing effort to address this problem under the contractual framework. For example, the successful debt exchange that was recently concluded by Uruguay included a limited form of aggregation provision in the new bonds (Box 1). J.P. Morgan’s proposal for a “two-step” procedure to achieve a rapid and orderly debt restructuring also assumes that there will be some form of aggregation of creditor claims. Finally, the recent establishment of the Argentine Bond Restructuring Agency (ABRA) is designed, in part, to obtain some of the benefits of an aggregated framework.

5. Notwithstanding these potential benefits, however, concerns have been expressed regarding the potential risks of aggregation. There is a fear that aggregation could give rise to inter-creditor equity concerns where, for example, a majority of creditors holding a certain type of claims impose an agreement on a minority of creditors that hold very different claims. While such a problem could arise where claims of different seniority (i.e., secured or unsecured claims) are aggregated, it may also occur where the claims being aggregated continue to have different maturities. As will be discussed in this paper, the latter issue is particularly problematic in the pre-default context, since all of the claims will not yet have become due and payable as a result of acceleration. Investors also have expressed the fear that the sovereign could use the aggregation technique to manipulate the voting process.

6. In light of the above considerations, this paper explores the potential benefits, risks, and feasibility of achieving aggregation of claims under a legal framework that resolves collective action problems. It also examines whether aggregation could contribute to making the process of debt restructuring more predictable by catalyzing creditor coordination, and facilitating the resolution of issues relating to inter-creditor equity and the dialogue between the debtor and its creditors. The first section provides a preliminary discussion of the potential benefits and risks of aggregation in the sovereign context. The second section examines the feasibility of designing a legal framework that seeks to capture these benefits while, at the same time, minimizing these risks.

Uruguay: Novel CAC Features in Recent Bond Issues

Bonds issued by Uruguay include novel features. Specifically, the innovative aggregate voting clause included in these bonds provides the option to amend payment terms on the basis of aggregate voting across affected bonds in cases where the amendment affects two or more series of bonds. In particular, the aggregate voting provision enhances the ability to restructure where there is a large number of holdout creditors concentrated in a particular series of bonds, but there is significant support for the restructuring among bondholders taken as a whole.

  • a. If the sovereign chooses to amend the bonds on an aggregated basis, the 75 percent majority needed for changing payment terms on each individual bond issue will be lowered to 66⅔ percent, provided that at least 85 percent of the aggregate outstanding principal of all issues that are proposed to be affected support the amendment.

  • b. The effectiveness of aggregate voting is, however, limited in two respects. First, while the required 66⅔ percent threshold for each individual series is easier to achieve than the otherwise applicable 75 percent, it still enables a creditor to obtain a blocking position with respect to a particular issuance, although it would be more costly to do so.1/ Second, aggregation applies only to new bonds governed by New York law that are issued under the same trust indenture.

At the same time, a number of features of Uruguay’s bonds provide greater investor protection than is available under sovereign bonds issued recently with CACs. In particular:

  • c. The bonds limit Uruguay’s ability to use coercive exit consents in future restructurings of such bonds through an undertaking that no future exchange offer will include amendments to make the current bonds less attractive than the bonds to be offered in the exchange.

  • d. The bonds include strengthened provisions (in addition to the disenfranchisement provisions) to ensure the integrity of the voting process—in particular, Uruguay is required to affirmatively certify in future restructuring the amount of bonds owned or controlled by the government or its public sector instrumentalities, and not to issue new bonds or reopen a series of bonds with the intention of placing bonds with investors that are expected to support a future restructuring.

  • e. The bonds contain a transparency provision which requires Uruguay to provide certain types of information to investors before any future modification of the bonds is sought.

1/ Although the failure to achieve the 66⅔ percent requirement for any series of bonds would preclude a restructuring from going forward with respect to that series, a restructuring could still be effected for other series so long as it is supported by bondholders of 85 percent of aggregate outstanding principal of all series that are proposed to be affected by the restructuring and of 66⅔ percent of outstanding principal in each series to be restructured.

II. The Potential Benefits and Risks of Aggregation

A. The Resolution of Collective Action Problems


7. The incentives for individual investors to decide whether to participate in a restructuring, or to hold out in the hope of receiving more favorable terms, clearly depend on an evaluation of the extent to which a proposed deal protects their individual interests, and the likely payoffs of the alternative strategies in each case. In making these judgments, the following considerations would be of particular importance.

8. With regard to the protection of investors’ interests, a key question concerns the quality of assurance that the overall deal will achieve adequate inter-creditor equity. In some cases, a limited number of, and homogeneity among, instruments to be restructured, may provide sufficient comfort that forbearance will not be exploited to allow other creditors to receive more favorable terms. In other cases, however, it may be helpful to have in place a mechanism that provides more concrete assurances that the various groups of creditors will act in unison.

9. With regard to the evaluation of likely payoffs of alternative strategies, an investor will need to compare the likely market value of the claim on a post restructured basis, with: (i) the probability that the debtor would service the original claim, and the likely market value of such a claim that is continuing to be serviced; and (ii) the likely risk and return of seeking to obtain recoveries on distressed debt, in the event that the claim is not serviced.5, 6 In general, the holdout strategy may be more appealing in cases in which either the restructuring is conducted pre-default, and so it is likely that nonparticipating creditors will continue to be paid, or where the potential recoveries on distressed debt may be large in relation to the secondary market price (i.e., cases in which debt is trading at a steep discount). Conversely, collective action difficulties are likely to be less acute in cases in which there is a perception that non-participating claims will not be serviced, and in cases in which the potential gains of a holdout strategy are moderate—on account of some combination of the relatively high secondary market price, the potential costs and uncertainty of litigation, and unavailability of assets vulnerable to attachment. The viability of the recovery strategy also depends critically on individual investors being able to avoid being bound by collective action clauses or having the non-payment terms of their instruments modified by exit consents.

Multiple Bond Instruments

10. There are strong reasons for believing that the inclusion of collective action clauses in debt instruments could make a significant contribution to the restructuring process. Approximately one third of the existing stock of international sovereign bonds include collective action clauses. Until recently, bonds issued by emerging market sovereigns that are governed by New York law have not included majority restructuring provisions, though at the time of writing, most of the new issues of such bonds in New York since March 2003 have included these provisions.7 This development is welcome and, if sustained, would, over time, ensure that an increasing portion of outstanding emerging market sovereign bonds include CACs. Even when all sovereign bonds include CACs, however, it will be possible for a creditor or a group of creditors to obtain a “blocking position” in one or more issuances. In these circumstances, the question arises as to whether creditors holding a qualified majority in other issuances that are otherwise willing to reach an agreement with the debtor will be willing to pursue a restructuring without an assurance that investors holding “blocked” issuances would take similar action.8

11. The viability of such a “multiple issuance holdout strategy” would depend, in part, on the design of the collective action clauses. As has been discussed in earlier papers, collective action clauses are comprised of two different types of provisions.9 The first is the majority restructuring provision, which enables a qualified majority of creditors (typically 75 percent) to bind all holders of the same issuance to the terms of a restructuring agreement before or after a default. The second type of provision is the majority enforcement provision, which limits the ability of individual bondholders to disrupt the restructuring process by enforcing their claims after a default but prior to a restructuring agreement. Most bonds require a 25 percent of outstanding principal to accelerate the bond. Some bonds also allow bondholders of a simple or qualified majority of outstanding principal to reverse a prior acceleration of the issue if all events of default have either been cured or waived.

12. The following example illustrates the multiple issuance holdout problems. Suppose that the debtor has several outstanding bond issuances that it wishes to restructure and that all of these bonds include collective action clauses that are comprised of: (i) a majority restructuring provision that requires a vote of 75 percent of outstanding principal, and (ii) majority enforcement provisions that require a vote of 25 percent of outstanding principal to accelerate the bond and 50 percent of outstanding principal to decelerate. To block the use of the majority restructuring provision, the holdouts would need to acquire more than 25 percent of outstanding principal. Furthermore, if the holdouts were to litigate for recovery on their claims, they would need to activate the acceleration provision. To do so, they would have to acquire over 25 percent of outstanding principal of the bond. However, to ensure that the acceleration is not rescinded, and that the contractual provisions are not modified through the use of exit consents, the holdouts would need to acquire more than 50 percent of the bond issuance.

13. How difficult it will be to obtain a blocking position within a bond issuance will, of course, depend on the size of the issuance. In the context of normal market-based liability management operations (rather than a restructuring of the net present value of indebtedness), the issuer may offer creditors to exchange existing instruments for new bonds with longer duration. As not all of the original instruments will typically be tendered for such exchanges, these operations tend to leave behind relatively small residual amounts of the original issue. On account of their limited size, it may be relatively easy for an investor to acquire a blocking position in such instruments in anticipation of a possible restructuring exercise that may take place in the future.

14. While obtaining a blocking position in a single bond issuance enables potential holdouts to neutralize the potential benefits of the collective action clauses, it does not mean that other holders of the same issuance cannot participate in the restructuring. To the extent to which the restructuring is achieved through an exchange of instruments—the most commonly applied technique–holdouts could not prevent other holders of the same issuance from exchanging their instruments.10

15. An important implication of this analysis concerns the magnitude of the hold out problem. Specifically, if holdouts obtain a blocking position within individual bond issuances, and the debtor makes an exchange offer, the magnitude of the potential holdout problems is determined by the value of the bonds actually held by the holdouts—not by the value of the bond issuances over which the holdouts exercises control. Where the aggregate value of the bonds held by such holdouts is considerable, this may indeed make it more difficult for other bondholders to accept the terms of a restructuring. Moreover, it may also make it more difficult for the sovereign to achieve the necessary level of debt reduction. In these circumstances, even the most robust collective action clauses will be of no benefit to the restructuring process as long as they apply only within a single bond issue.

16. There is a question of whether this issue can be addressed through the use of trust deeds governed by English law. Such instruments provide one of the most effective types of majority enforcement provisions, where the rights of individual bondholders to initiate litigation is, with some limitations, effectively delegated to the trustee who is only required to initiate litigation if, among other things, it is requested to do so by the requisite percentage of bondholders (typically between 20 percent and 25 percent).11 Consistent with the trustee’s authority to initiate legal proceedings on behalf of all of the bondholders, any amounts recovered by the trustee through such proceedings are for the benefit of the bondholders as a group and, therefore, are distributed pro rata among all of the bondholders. Accordingly, even if a bondholder wishing to pursue litigation has managed to acquire a sufficient percentage of bonds to enable him to require the trustee to initiate litigation, the pro-rated distribution of any amounts received through litigation among all bondholders will reduce such bondholders’ incentive to do so.

17. Notwithstanding these benefits, however, the sharing provision will be of no benefit in the situation described above; i.e., where a single distressed debt purchaser has been able to neutralize other collective action clauses and the restructuring is achieved though an exchange of instruments. Where the exchange offer has “emptied out” the issuance of all holders other than the distressed debt purchaser, the requirement to share the proceeds of litigation with other bondholders is of limited deterrence: the litigating holdout would only be required to share the proceeds with itself.

18. In light of the above analysis, the inclusion of collective action clauses and the use of exchange offers will not eliminate the multiple issuance holdout strategy if the clauses only bind holders on an issue-by-issue basis. Of course, it is still too early to determine whether this “gap” will have significant impact on the speed and order of the debt restructuring process. However, to the extent that this gap is perceived as being problematic, aggregation of claims for voting purposes could effectively address it.12 Specifically, to the extent that the voting provisions could be structured so that voting is calculated across issuances, it would significantly reduce the viability of the multiple issuance holdout strategy: as long as the requisite majority of the entire community of bondholders supported the transaction, all bondholders would be bound, irrespective of the percentage they may hold in a particular issue. As will be discussed in the next section, however, such aggregation will create its own problems in circumstances where the terms of the differing bond issues vary.

Different Types of Debt

19. For purposes of resolving collective action problems, would there be a similar benefit in also aggregating the claims of other forms of debt, such as commercial bank debt? For purposes of analyzing this question, it is useful to distinguish between two different possibilities. The first would entail aggregation among commercial bank claims, while the second would go one step further and aggregate commercial banks claims with bonds.

20. Taking commercial bank claims as an example, a threshold question is whether the restructuring of such claims is actually hampered by collective action problems. To date, commercial banks have normally behaved in an orderly manner and the risk of litigation by commercial banks has been very limited. As has been discussed in other papers, this behavior is the result of, among other things, their ongoing relationship with the debtor, reputation concerns, and the moral suasion by regulators. Nevertheless, given the evolution of capital markets, the assumption that syndicated bank debt can be restructured on a more orderly fashion than bonds may be becoming increasingly tenuous. As result of securitization of banks’ claims, the instruments can be acquired by non-bank financial institutions, including vulture funds (as was demonstrated by the recent litigation against Peru). For this reason, there has been some discussion both within the official and private sectors as to the merits of including collective action clauses within syndicated bank loans.

21. To the extent to which one concludes that: (i) collective action problems can arise in the restructuring of syndicated bank debt; and (ii) the terms of these instruments would benefit from the inclusion of collective action clauses, one is confronted with the question as to whether a gap similar to the one identified for multiple bond issuances would also apply with respect to bank debt. Specifically, a restructuring could be undermined in circumstances in which there were a number of syndicated bank loans (all of which included collective action clauses) and one syndicate cannot participate on account of a holdout creditor having obtained a controlling interest in that syndicate

22. In one respect, the terms of syndicated bank loans already contain provisions that will make such a holdout strategy more difficult. Specifically, even though these agreements do not presently include collective action clauses, they do include sharing provisions, which require a bank that has recovered amounts from the sovereign as a result of litigation to share the proceeds of the recovery with other members of the syndicate in question. However, the effectiveness of this sharing provision to undermine a “multiple syndicate hold-out strategy” is also limited by the same problem that confronts trust deeds when addressing a multi-issuance holdout strategy; namely, to the extent that the restructuring is achieved through an exchange offer, the sharing provision will not constrain a lender participating in a syndicated loan if—as result of general participation by others in the offer—it is the only remaining creditor in the syndicate. Indeed, in the case of the litigation against Peru, although the terms of the agreement enforced by the distressed debtor provided for the sharing of any litigation proceeds, this did not undermine the distressed debt purchaser’s strategy because—as a result of the success of the Brady bond exchange—it was only required to share these proceeds with itself.

23. In light of the above, there may also be a benefit in aggregating claims for voting purposes among different syndicated bank loans. However, as with the aggregation among bond issuances, such an approach will need to resolve problems that could arise in circumstances where the terms of the various loan syndicates differ from each other (which is likely to be the case). With respect to aggregating claims among bond issuances and syndicated bank loans, the problem is compounded by the fact that bondholders and commercial bank creditors also have very different interests when they approach a restructuring with a sovereign. Whether the aggregation of other types of non-bonded debt (e.g., guaranteed trade debt) would be of benefit would need to take into consideration the particular features of that debt, including whether it is sold on the secondary market and the extent to which aggregation of these claims would be hampered by a diversity of economic interests. The latter set of issues is discussed further below.

B. Creditor Coordination Issues

24. To what extent would the aggregation of creditor claims facilitate creditor coordination and, thereby, enhance the speed and predictability of the debt restructuring process? There would appear to be two potential benefits. First, the aggregation of creditor claims for voting purposes may assist in catalyzing earlier creditor organization. Second, to the extent that aggregation assumes that, in a post-default environment, all creditors are treated the same, it may provide greater predictability regarding the resolution of inter-creditor equity issues. Each of these benefits is analyzed in turn.

Catalyzing Creditor Organization

25. Once the restructuring process is initiated, early organization of creditors could be helpful from the perspective of developing an understanding of the need for a restructuring, and economic developments and prospects, which will determine the financial parameters of a deal. It could also facilitate the design of a restructuring package.13 Clearly, however, it will not expedite agreement in circumstances in which with great uncertainty concerning economic prospects and the debtors’ payments capacity, the debtor and its creditors share a common interest in delaying an agreement.

26. In the absence of aggregation, the debtor will have to deal separately with representatives of bondholders whose instruments are subject to restructuring on a series-by-series basis. While this may be less of a problem where only few issues are outstanding, the creditor organization process may be protracted where there is a multiplicity of issues. The formulation of a representative creditors’ committee within one issue, even if formed quickly, would represent only the bondholders of that particular issue and would not facilitate overall creditor organization.

27. On the other hand, to the extent that the claims of creditors are aggregated for voting purposes across different instruments, creditors may have an additional incentive to organize at an early stage of the process. Under the existing framework, it is difficult for an organized group of creditors to “deliver” the creditor community. For example, where there are large numbers of bond issues outstanding, even a large, representative group of creditors will have difficulty in representing to the debtor that it can control the actions of those creditors that have obtained a blocking position in a number of issuances. However, if claims are aggregated across for voting purposes, it will be possible for a qualified majority of all external bondholders to represent to the debtor that: (i) they can make an agreement binding on all bondholders and (ii) they can prevent any litigation by bondholders prior to reaching an agreement. (The latter would be true only if the decision on whether to initiate litigation is subject to the aggregation feature.) The potential of such leverage will make it more likely that the debtor will wish to engage more rapidly in a dialogue with such creditors—a fact that, in turn, could further encourage early creditor coordination. Although this analysis is somewhat speculative, recent private sector proposals that provide for some form of aggregation (discussed in the following section) identify this benefit as being an important motivation for an aggregated legal framework.

Resolution of Inter-Creditor Equity Issues

28. One of the sources of uncertainty under the existing framework is how inter-creditor equity issues should be resolved. In circumstances where there is a diverse group of external creditors holding differentiated claims, to what extent will all of them be subject to the restructuring and, if so, what criteria will be used for purposes of determining whether the restructuring terms they receive should be the same as—or differ from—other creditors? While this problem has many different facets, including the relative treatment of domestic and external debt, one issue that has proven particularly difficult is the treatment of claims with different residual maturities. Specifically, for restructuring purposes, should one take into consideration the differing expectations of creditors under the original terms of the instruments? In a pre-emptive restructuring (i.e., prior to a default) it is likely that investors would consider that holders of a bond that was to mature one year after the date of the default should be treated differently from a creditor whose claim is maturing 20 years after a default. In contrast, investors may consider that in circumstances in which all of the debt is in default and has been accelerated (i.e., the claims are due and payable), they should be given the same treatment because all of their claims have the same maturity.

29. A number of commentators have suggested that prior to, and in the immediate aftermath of, default, debts of differing maturity should be restructured in a fashion that extends maturities on each instrument, but broadly preserves the original relative residual maturities. Thus, by way of example, individual bonds (or a group of bonds with similar maturities) could be restructured into new instruments with a maturity extension of, say, 5 years. It has been argued that this so called “bucket approach” can be seen as helping to preserve inter-creditor equity, while at the same time producing a post-restructuring debt stock that allows a market-based yield curve to be established rapidly. On the other hand, bondholders have also expressed concern that such differential treatment may result in discriminatory treatment, and that such a “bucket” approach may lead to manipulation, for example, by forcing investors holding the shorter maturities to accept a disproportionate share of the burden. It has also been pointed out that, post default, the secondary market processes of debt instruments do not vary on the basis of the original maturity or coupon, but rather trade according to market perceptions of recovery value.

30. In light of the above, aggregation may also be designed so as to provide some predictability with respect to the resolution of inter-creditor equity issues for the following reason. As will be discussed further in the next section, aggregation of claims for voting purposes is easier to achieve across different instruments if it is accompanied by a rule that requires all aggregated claims to receive the same terms, or the same menu of terms. Such a rule prevents one form of discrimination; namely, where all creditors have the same claims against a sovereign, the rule prevents a majority of creditors from agreeing to terms that give them preferential treatment vis-à-vis the minority of creditors holding the same claims. Indeed, the application of this rule is an important means of preventing discrimination under any decision framework that allows a qualified majority to make decisions binding on a minority—including, for example, traditional collective action clauses. While the imposition of an equality of treatment rule in a post-default environment will make it easier to achieve aggregation for voting purposes, the consequence of such a rule may also be to provide for some predictability as to how creditors with different instruments will be treated during the restructuring process; i.e., it will confirm that, in a post-default context, creditors holding instruments of different maturities are, nevertheless, to be given the same treatment under the restructuring agreement. The situation is more complicated in pre-default restructurings, however, as is discussed further, below.

31. A related issue arises with respect to differences between the coupons of various bonds. One approach would be to calculate the size of each creditor’s claim by taking into account not only outstanding principal but also any accrued interest. However, nonaccrued interest that would have been paid during the life of the loan had there been no default would not be a relevant factor for purposes of distinguishing between the value of the various claims of creditors. This is the approach that is reflected in the existing credit default market and which is followed under at least two of the legal frameworks discussed in the next section.14

C. The Potential Risks of Aggregation

32. As was suggested in the previous section, the potential benefits of aggregation need to be assessed in light of the risks that such a framework could create. For purposes of analysis, it is possible to identify at least three distinct types of potential problems that could be created by aggregation: (i) potential discrimination among creditors, (ii) potential manipulation by the sovereign, and (iii) an absence of flexibility.

Inter-Creditor Discrimination

33. One of the biggest potential risks of aggregation is that it may provide an opportunity for the debtor and a qualified majority of creditors to discriminate against a minority of creditors who, although they do not support the proposed agreement, will be bound by its terms. As noted above, in circumstances where all of the creditors have the same type of claims—because they hold bonds in the same issuance or they form part of the same syndicate—the problem of discrimination can be addressed by requiring that they all receive the same terms under the restructuring agreement—or the same menu of terms. However, when one aggregates different types of claims for voting purposes, the application of this rule may give rise to another form of discrimination. Namely, where a minority of creditors possess claims that are—in one way or another—preferential to those of the majority, allowing the majority of creditors to approve an agreement that treats all of creditors the same will be discriminatory; i.e., depending on the nature of the claims being restructured, inter-creditor equity may actually require differential rather than similar treatment.

34. In the sovereign context, this type of problem can arise in at least two different circumstances.

35. The first case is where the claims of certain creditors are senior by virtue of the fact that they benefit from some enforcement privilege, e.g., a secured claim. In the event of a default, the value of the secured creditor’s claim is preserved—at least relative to that of the unsecured creditor—by virtue of its ability to foreclose upon the collateral (which may, for example, be made up of irrevocable export marketing agreements and arrangements for the associated receivables to be channeled through an offshore account). If, during the restructuring process, the claims of these creditors were to be aggregated with those of unsecured creditors (including bondholders), there is a risk—indeed, a strong likelihood—that, in the event that unsecured creditors constitute a qualified majority, they would approve an agreement that would strip this collateral from the secured creditors. For this reason, it is unlikely that investors would accept contractual provisions for aggregation in the context of secured credits.

36. The second case is where the claims of the creditors—while all unsecured—have different residual maturities. As noted in the previous sections, this problem is mitigated in the post-default context by virtue of the fact that, at that stage, their original terms have been effectively superceded by an acceleration that makes the face value of all claims immediately due and payable. To the extent that a debtor seeks a restructuring prior to a default, however, creditors with differing residual maturities have differing economic interests. If all creditors holding nondefaulted claims were forced to accept the same long-term instrument in a restructuring, creditor holding claims with relatively short residual maturities would bear a disproportionate burden as compared to those holding claims with relatively long residual maturities.

37. As has been discussed in previous papers, there are numerous advantages to restructuring debt prior to a default. This may allow the level and profile of debt to be brought to a sustainable level, with a minimum of uncertainty that is inimical to maintaining confidence in policies and domestic financial markets. Moreover, an interruption in a sovereign’s contractual obligations can undermine the credit culture that underpins the efficient operation of banks, and the domestic financial system more generally. At the same time, securing agreement on a debt reorganization without default may help to minimize the damage to a sovereign’s reputation in capital markets, and may help to facilitate an early return to spontaneous market access. Accordingly, to the extent that the complete aggregation of claims for voting purposes in a pre-default context runs the risk of discrimination, this would constitute a significant issue, which would alter the incentives facing investors and could, thereby, potentially influence the scale and composition of capital flows to emerging markets.

Limiting Flexibility

38. Even putting discrimination issues aside, there may be cases where the complete aggregation of claims for voting purposes may actually limit the flexibility needed for the debtor to secure a critical mass of support for an agreement. To the extent to which all creditors whose claims are being aggregated are required to receive the same terms—or menu of terms—the application of this rule may fail to take into consideration the fact that different creditors may have very different interests—notwithstanding the fact that their claims are the same. Thus, for example, although a domestic bank and a foreign investor may hold foreign currency bonds that have become due and payable, they may be willing to consider very different types of terms in the restructuring process. While the foreign investor will be seeking a restructured claim instrument that is liquid, a domestic bank may be willing to accept a claim that is inferior in exchange for some regulatory forbearance. Being able to take advantage of these different preferences will be critical for a sovereign that is seeking to secure a restructuring on terms that provide it with maximum economic latitude. However, this will be constrained to the extent that, as a consequence of aggregation, it is required to provide all creditors the same terms or menu of terms.

Potential Abuse by Debtor

39. Could aggregation provide sovereigns with the opportunity to distort the voting process? At one level, this risk already exists in a framework that allows a qualified majority to bind all creditors within the same bond issuance. Because of its unique powers, there is a risk that, in circumstances where a majority of the bonds are held by residents over which the sovereign issuer exercises some regulatory influence, this majority will be coerced into accepting terms that are particularly favorable to the sovereign, at which point the agreement will also become binding on the minority of foreign investors. On one level, aggregation could actually serve to limit this risk, particularly if the only claims being aggregated are external claims; i.e., claims that are governed by foreign law or subject to exclusive jurisdiction of foreign courts. While it may be possible for the sovereign to exert enough influence over domestic banks and other domestic entities to affect the voting of a single issuance, it will normally be more difficult to do so when all foreign law instruments are aggregated.

40. Nevertheless, there are at least two ways in which aggregation, even if limited to external claims, could provide a sovereign with the opportunity to manipulate the voting process. The most extreme device would be the creation of fictitious claims. To the extent that a sovereign issues securities or makes a private placement to an entity it owns or controls and does not receive value from that entity in return, it can distort the restructuring process in two ways. First, those claims could be voted in a manner that would support onerous debt restructuring terms for holders of valid claims. Second, since these claims would be recognized under any restructuring agreement, they would reduce the amount to be received by valid creditors.

41. A less extreme device would be for the sovereign, in anticipation of a restructuring, to issue new external debt to its domestic banks in exchange for outstanding domestic debt. If the amount of the new external debt is significant, aggregating these claims with pre-existing external claims would provide an opportunity for the sovereign to shape the voting process through its influence over the domestic banks.

III. Designing an Aggregation Framework

42. Given the above analysis, the key challenge when designing a legal framework that can achieve some degree of aggregation is to find a way to capture the considerable benefits of aggregation while, at the same time, minimizing its potential risks. Over the past two years, considerable effort has been expanded—both in the official and private sectors—to arrive at a framework that can achieve such a balance. This section provides a comparative analysis of these frameworks. It begins with a discussion of the feasibility of aggregating claims under a contractual framework. It then analyses how aggregation would be achieved under a new proposed statutory framework, drawing on earlier work that has been done with respect to the design of the SDRM.

A. Contractual Framework

Collective Action Clauses

43. Given the considerable benefits of aggregation, the question arises as to the extent to which such a feature can be successfully introduced into collective action clauses in a manner that effectively addresses the risks identified above. The most recent bonds issued by Uruguay include a collective action clause that, to a limited extent, provides for aggregation. For this reason, the analysis set forth below regarding the feasibility of aggregating claims through a contractual framework benefits from this helpful development.

Establishing the Contractual Basis

44. By definition, the establishment of a contractual basis for aggregating claims across different debt instruments requires that the terms of these instruments actually provide for such aggregation. This is in contrast to a statutory framework (discussed below), which imposes an aggregation mechanism irrespective of the terms of the contract. Thus, for example, to the extent that the collective action clauses of a particular bond issuance provide that, for voting purposes, the bonds may be aggregated with future bond issuances, this will only occur if the future bond issuance also provides for aggregation with that bond.

45. In light of this requirement, there would appear to be two ways in which aggregation provisions may be integrated into collective action clauses. The first would involve the inclusion of a provision in each particular bond issuance or syndicated bank loan agreement that would identify the types of instruments with which aggregation could be achieved. For example, the bond issuance could provide that, for voting purposes, aggregation could only occur with respect to future sovereign bond issuances that also contain a collective action clause that allows for aggregation with that bond. The establishment of aggregation by such “reciprocity” would most likely require that the collective action clauses contained in the respective bond issuances be identical. It would also require that all such instruments include uniform voting rules and procedures.

46. A second approach, which was used by Uruguay, would involve establishing a single collective action clause in a master agreement (e.g., the trust indenture) that could then govern all (or some) future bond issuances. In the case of Uruguay bonds issued in April 2003, all bond instruments are issued under a single trust indenture. The indenture contains a majority restructuring provision that aggregates, for voting purposes, all of the bonds issued under the indenture under certain circumstances. It also provides for voting rules and procedures with respect to aggregation that applies to all bonds. Uruguay may issue an unlimited additional amount of bonds in different series under the same indenture and all these bonds will be subject to the same aggregation provision in the indenture. Thus, all future bond issuances, when issued under this indenture, will be aggregated with the outstanding bonds for voting purposes under the terms of the indenture. As noted above, the aggregation provision in the Uruguay bonds applies only to modification of certain key terms of the bonds and does not apply to majority enforcement provisions (e.g., acceleration, de-acceleration, or initiation of legal proceedings). However, one could also envisage using a master agreement as the basis for aggregating for these other decisions.

47. The advantage with this approach is its simplicity and predictability—since a single collective action provision governs all of the bonds that are issued under the trust indenture, there is no need to ensure that future bonds contain an aggregation provision that is compatible with those contained in earlier bonds. The limitation, however, is that the bonds must be issued under the same indenture in order for the aggregation provision to apply. Market participants have raised a question as to whether the introduction of a new indenture which provided for aggregation with a large stock of bonds that do not include such provisions would be accepted by investors.

48. Under both approaches, the contract could be formulated so as to allow the debtor to decide on the coverage of a restructuring. For the purposes of decision taking, the aggregation would apply only to instruments within the scope of the restructuring; investors holding excluded claims would not have a vote.

Aggregating across jurisdictions

49. One of the potential limitations that applies to aggregating claims pursuant to contract is the difficulty of aggregating claims of instruments that are governed by different laws or subject to different jurisdictions: in the event that a dispute arose regarding the application or interpretation of the aggregation provision, there would be a risk that holders of different bond issues would find themselves in different courts—which could provide different interpretations of the provision. How serious a problem this is will, of course, depend on whether a sovereign issuer chooses to issue in different jurisdictions. Moreover, even if individual bonds are governed by different laws and subject to different jurisdictions, the question arises as to whether it would be feasible to include the aggregation provision in a master agreement that is governed by a single law and subject to a single jurisdiction.


50. How does one address the problem of discrimination among creditors? Assuming for the moment that the claims being aggregated for voting purposes are different bond issuances, what safeguards are in place to ensure that holders of instruments with certain terms do not approve a restructuring that places an undue burden on holders of claims with different terms? As discussed earlier, if all of the bonds being aggregated were in default at the time the restructuring takes place, the approach could be taken that discrimination can be avoided by simply requiring that all holders receive the same terms or menu of terms: since, in this scenario, the maturity of claims is the same (all claims are accelerated), discrimination will be addressed through equal treatment.

51. The issue is complicated, however, by the fact that collective action clauses may also be used outside the context of a default. Accordingly, the concern regarding potential discrimination will arise in circumstances where the sovereign’s debt burden is sustainable but it faces acute liquidity constraints. In these circumstances, it may wish to use collective action clauses to secure creditors’ agreement to extend the maturities of certain claims falling due but leave the claims of those holding longer-term instruments untouched. In these circumstances, the holders of the long-term instruments will readily agree to such a restructuring since such a step will reduce the possibility of the need for a more comprehensive restructuring at a later stage.

52. One approach to addressing this potential risk, utilized by Uruguay, is to include a two-tier voting threshold in the collective action provision: while a certain percentage of support from all holders of all of the bond issues that are proposed to be affected would be required for the overall restructuring to go forward (in Uruguay, the aggregate voting threshold was 85 percent), no particular bond issuance could be restructured in these circumstances unless there was a minimum amount of support from holders of that bond issuance (in Uruguay, the threshold was 66⅔). To the extent that the minimum amount of support from a particular bond issue is not received, however, this would not prevent all other bonds being restructured to the extent that the above thresholds had been met.

53. While the above approach effectively addresses the risk of discrimination, the safeguard that it introduces does limit the benefits of aggregations—at least from the collective action perspective. As discussed in the earlier section of the paper, one of the objectives of aggregation is to limit the ability of creditors to disrupt the restructuring by obtaining a controlling percentage of the issuance. To the extent that the threshold for the restructuring of each bond issuance is 75 percent (the traditional collective action clauses), a creditor only has to obtain more than 25 percent of the bond issuance. Under Uruguay’s aggregation provision, it would still be able to obtain a blocking position by acquiring more than 33⅔.

54. The extent to which the dual threshold will be of benefit from a collective action perspective will depend on the circumstances of the case. Specifically, such an aggregation provision would facilitate a restructuring in cases where there is a large number of holdout creditors concentrated in a particular series of bonds, but there is significant support for the restructuring among bondholders taken as a whole. As an interesting illustration of this point, legal counsel for Uruguay calculated the extent to which the recent debt exchange would have received an even higher participation rate if the bonds being surrendered in the exchange had contained the type of aggregation provisions that were included in the new bonds. As can be seen from the results set forth in Box 2, the percentage of bonds that would have been restructured increases as the design of the aggregation provision becomes more ambitious.

Uruguay: An Aggregation Hypothetical

This chart takes the actual results of the Uruguay exchange and shows, on a pro forma basis, the potential effects of both a single series majority restructuring provision (with a 75% voting threshold) and a majority restructuring provision with an aggregated voting mechanism. 1/

Of the 18 series of Uruguay’s existing international bonds (excluding the Samurai), 12 series had tenders in excess of 75% (and thus would have been amended by single series voting with a 75% threshold had such a majority restructuring provision been in place for those old bonds); 3 series attracted tenders of less than 66% (and thus would not have been affected even by Uruguay’s new aggregated voting threshold); but another 3 had tenders of between 66% and 75% (and these would have been restructured in full through the use of the aggregated voting mechanism in Uruguay’s new clause).

article image
1/ Prepared by New York legal counsel to Uruguay.

55. As an alternative—or perhaps as a supplement—to the above approach, one could consider designing a collective action clause that provides for aggregation only in circumstances where all bond issuances have been accelerated. In these circumstances, the need for a minimum threshold of support from each bond issuance would no longer be needed: since the maturities of all instruments would be the same, the potential for discrimination would be addressed by applying the general rule that all creditors receive the same terms (or menu of terms) in the restructuring.

Potential Abuse

56. To address creditor concerns about manipulation of the voting process by a sovereign issuer through its effective control of certain bondholders, Uruguay introduced provisions in its recent bonds that require Uruguay to affirmatively certify whether any bonds are owned or controlled by Uruguay or its public sector instrumentalities prior to any vote on a modification affecting any of those bonds. In addition, Uruguay undertakes not to issue new bonds or reopen a series of bonds with the intention of placing bonds with investors that are expected to support a future restructuring. This certification requirement was welcomed by the investor community as a helpful safeguard against debtor abuse. It would not address the situation, however, where the sovereign influences the voting process through its regulatory influence rather than control.

Establishing a Contractual Framework in the Context of a Crisis

57. The collective action clauses relied upon to facilitate the resolution of financial crises, as discussed above, are included in the original bond documentation and, therefore, the relevant legal framework is already in place when the crisis actually arises. In order to facilitate the restructuring of the existing stock of bonds that do not contain collective action clauses, the private sector has made proposals that envisage bondholders adopting a contractual framework in the context of a crisis that provides for some degree of aggregation. As will be discussed, one of the difficult questions is whether bondholders would be willing to adopt such a framework in the context of a crisis, particularly in circumstances in which the authorities may not have fully elaborated a package of macroeconomic and structural policies to resolve the crisis, and there is substantial uncertainty regarding the debtor’s capacity to generate resources for debt service.

The J.P. Morgan “Two-Step” Proposal

58. J.P. Morgan has suggested a contractual approach that seeks to achieve the benefits of aggregation with respect to a stock of debt that does not include collective action clauses.15 The approach involves a two-stage process. In the first stage, investors holding different instruments would exchange their existing instruments for an “Interim Debt Claim” (the “IDC”). The terms of the IDC would provide a legal framework that would facilitate an agreement on the second stage, which would be the final restructuring agreement. The terms of the IDC would be designed to achieve a number of the benefits of aggregation identified in the earlier sections of the paper. As noted in the proposal, the IDC “aligns investors’ interests and provides for a legal mechanism for collective action sharing and representation.” This legal mechanism would facilitate the second stage, which would be an exchange of the IDC for the restructured instruments.

59. In terms of the various benefits of aggregation that have been identified in the first section of this paper, the J.P. Morgan proposal is noteworthy in several respects.

60. As a means of providing predictability on the resolution of inter-creditor equity issues, the proposal assumes a certain degree of uniformity of treatment. First, although the instruments being exchanged for an IDC would have different maturities under their original terms, holders of the IDC would all hold instruments of the same maturity. With respect to the IDC coupon, all holders would receive the same interest rate or the same menu of rates. However, the interest rate that had accrued under the original bond prior to the exchange for the IDC would be reflected in the par value received under the IDC. A distinction would not be made, however, on the basis of future, non-accrued interest. All holders of the IDC would receive the same restructuring terms or menu of terms at the time the IDC is exchanged at the second stage for the final restructuring instrument. Since the first step would occur “as soon as possible after a default,” the proposal does not address how such uniformity would be accepted in the pre-default context; i.e., before all claims become due and payable.

61. With respect to the resolution of collective action problems, the IDC will place all creditors within a single contractual framework that will contain a collective action clause. Specifically, “any restructuring terms approved by a defined super majority (e.g., 85 percent) of IDC holders would have to be accepted by all holders.” This framework would be designed to address the multiple bond holdout strategy. To the extent that bond issuances from different jurisdictions were exchanged into a IDC (which would be governed by the law of a single jurisdiction), this framework would allow for aggregation across different jurisdictions.

62. Regarding the facilitation of creditor coordination and negotiations with the debtor, the fact that the claims of all creditors are subject to the same legal framework will “give the sovereign a single credible party with whom to negotiate.” To provide further predictability in this area, the terms of the IDC would provide that a creditors committee appointed by the IDC holders would be empowered to negotiate on their behalf with the sovereign.

63. The key challenge in implementing the two-stage restructuring framework is persuading investors to participate in the first stage of the exchange. Shortly after a default, and in circumstances where there may be substantial doubts about the future course of economic policies and the terms of an eventual restructuring proposal, investors may be unwilling to surrender their individual contractual rights and enter into a collective process. In a period of considerable uncertainty regarding the conduct of a debtor, creditors may be unwilling to surrender their individual rights, Indeed, there is a question of whether doing so would be consistent with portfolio managers’ fiduciary responsibilities.

64. In recognition of these problems, the J.P. Morgan proposal includes incentives for creditors to exchange their claims for an IDC.16 One incentive would be an upfront cash payment for participation. Another would be the aggressive use of exit consents, which—by eroding the contractual rights of investors who elect not to participate in the first stage—would reduce the attractiveness of the holdout strategy.

Argentine Bond Restructuring Agency (ABRA)

65. To facilitate organization of small holders of Argentine bonds for the future restructuring negotiations, a special purpose company, the Argentine Bond Restructuring Agency (ABRA) has been established under the laws of Ireland. Bondholders who wish to participate in the initiative may surrender their bonds to ABRA in exchange for certificates to be issued by ABRA. As the legal owner of the bonds, ABRA will represent the participating bondholders in the restructuring negotiations. The negotiations will be conducted, on behalf of ABRA, by a special negotiation team retained by ABRA on a contractual basis.

66. Participation in ABRA is voluntary and investors may withdraw, subject to an early redemption fee, before an agreement is reached; i.e., they may reacquire their original claims. Upon the conclusion of a restructuring agreement, ABRA will collect payments made and/or new securities issued under the agreement and distribute them among the certificate holders, subject to payment of certain fees.

67. As with the J.P. Morgan proposal, ABRA establishes a framework in the context of a crisis that seeks to achieve some of the benefits of aggregation. The chief among them is that of creditor coordination: one of the major objectives of the framework is to give small investors more leverage by uniting them under a single legal framework that delegates negotiations to a single party. It is less clear, however, whether ABRA will provide greater predictability on the resolution of inter-creditor equity issues. While the ABRA offering circular provides that ABRA may decide on a uniform write-down of all bonds regardless of the underlying maturities and interest rate coupons, it does not rule out alternative approaches. Interestingly, no interest accrues on ABRA certificates, irrespective of the coupon of the original instrument.

68. The extent to which ABRA addresses collective action problems is not entirely straightforward. One the one hand, the fact that ABRA is the sole legal owner of all of the claims that have been surrendered means that the problem that gives rise to collective action problems—the multiplicity of creditors with diverse interests—is addressed. On the other hand, the fact that bondholders can reverse the exchange (i.e., receive back their original claims) at any time prior to finalization of a restructuring agreement means that problems of collective action may continue to persist. Because it provides the right to reverse the exchange, ABRA is a voluntary rather than a mandatory framework. In this respect it differs considerably from the J.P. Morgan proposal: the reason why creditors may be unwilling to enter into the first stage of the J.P. Morgan proposal is because they are entering into a legal framework from which they cannot escape. Under ABRA, obtaining participation in the first stage is made easier by the fact that creditors have the right to withdraw and revert to their original claims if at any time prior to reaching an agreement, they are unhappy with the process. Of course, the financial penalty that is imposed on those wishing to opt out of ABRA prior to a restructuring may mitigate the potential collective action problem that may arise at that stage.17

B. Establishing a Statutory Framework

69. Consideration was given to establishing a statutory framework for aggregation in the discussions regarding the proposal for an SDRM. Indeed, the ability to achieve aggregation was one of the SDRM’s motivating features. Its statutory basis enables it to be more ambitious than the frameworks described above in a number of different respects. First, since its does not rely on contractual provisions to achieve aggregation, it can apply to the outstanding stock of debt. Second, since it envisages the establishment of a single dispute resolution process, it can achieve aggregation of claims that are issued in a number of different jurisdictions. Finally, the SDRM’s statutory basis would enable it to maximize the extent of aggregation while, at the same time, safeguarding against the most important risks. As was discussed during its development, however, the SDRM would not be able to address all of the challenges of aggregation. It was recognized, for example, that certain claims -because of their nature – could not be aggregated and would need to be restructured outside the SDRM framework.

70. Of course, the establishment of the SDRM would require substantial reform. As noted above, at its Spring 2003 meeting, the IMFC recognized that it was not feasible to move forward to establish the SDRM at this time. In the absence of the level of support needed for an amendment of the Fund’s Articles of Agreement, efforts are presently concentrated on improving the restructuring process through alternative frameworks. Nevertheless, the features of a statutory framework continue to provide a useful reference for understanding the range of issues to be addressed and the scope for reform.

71. It should be noted that, while the primary statutory debt restructuring model that has been considered over the past two years has been the SDRM, there have also been proposals for different types of statutory frameworks. One of the more innovative proposals would rely on existing legislation in the United States to restructure sovereign debt in the context of a class action procedure. Even the architects of this proposal acknowledge, however, that some legislative changes may be needed to ensure predictability in its application. A brief summary of the proposal is set forth in Box 3.

72. Under the SDRM proposal, particular attention was devoted to limiting the risks associated with aggregation that have been identified in Section I of this paper. Drawing on earlier Board papers, set forth below is a summary of how these particular risks were addressed. A number of these issues also arise when restructuring nonsovereign debt and, for this reason, the development of the proposal for a sovereign statutory framework benefited considerably from the experience that had accumulated regarding both the design and implementation of domestic insolvency laws.(An analytical overview of those aspects of domestic insolvency laws that are of greatest relevance is set forth in the Annex.)


73. As noted in the previous section, aggregation can lead to discrimination in a number of different ways. With respect to the potential discrimination of privileged claims, (i.e., those that benefit from some enforcement privilege, such as collateral), the SDRM would not aggregate these claims with other claims. Indeed, as a means of both achieving simplicity and limiting interference with contractual claims, it was decided that such claims would not be subject to the SDRM’s legal coverage; i.e., to the extent that a claim benefited from a privilege it could only be restructured with the consent of the creditor in question.18 While it was recognized that this might create incentives for an increase in the use of secured credits, it was recognized that this impulse could be contained by other means, including the reliance on negative pledge provisions.19

74. How would the SDRM proposal address the potential discrimination among unsecured creditors that possess claims that have different terms (maturity and interest rate) from each other? Could holders of long-terms bonds that have a qualified majority vote on a restructuring agreement that would provide, inter alia, for a significant lengthening of maturities of all short-term instruments? As noted earlier, this problem is resolved in the post-default context by virtue of the fact that all creditors can be considered to have claims of the same maturity; i.e., they are all due and payable. However how would the SDRM address this issue in restructurings that occur prior to a default, where the claims of creditors holding different instruments are not the same?

75. As was noted in the relevant staff papers, this issue also arises in the nonsovereign context, where a corporate debtor and its creditors agree upon a restructuring proposal before a commencement of insolvency proceedings and in circumstances where the debtor continues to service its obligations. Under these pre-negotiated or “pre-packaged” arrangements, proceedings are commenced solely for the purpose of making the agreement binding on the entire creditor body. When these plans are negotiated, unsecured creditors with different maturities are generally willing to have their claims aggregated for voting purposes and to receive the same restructuring terms as others because they recognize that failure to agree to a viable restructuring plan will result in general default and insolvency, where they will fare less well.

Limited Aggregation Through Class Action Procedures

An approach suggested by two US lawyers to overcome the multi-issuance holdout problem is to use existing statutory rules of civil procedure. 1/ These authors have suggested that, in particular, the class action procedures under US law can be used to bind all creditors to a collective settlement. 2/ In essence, the idea is that a group or committee of representative creditors, with the acquiescence of the sovereign debtor, would commence in US court a class action on behalf of all creditors. Under this approach, there would be a judicial mechanism for resolving disputes among creditors and between a debtor and its creditors. A proposed restructuring—in the form of a settlement of the class action—that enjoyed widespread creditor support could then be made binding on all affected creditors through a court judgment.

The proposal only involves “aggregation” in the sense there would be a common pooling of creditors across debt instruments with a settlement across all affected claims. However, it does not involve aggregation in terms of a vote by creditors across debt instruments: the terms of the settlement would be struck by the representative creditors and the sovereign debtor, subject to approval by the court.

Although innovative, even the authors have highlighted legal uncertainty in a number of areas. For example, would a single US domestic court be willing to exercise mandatory jurisdiction over the sovereign debtor and its creditors? Even if the court had such legal authority, what criteria would be used for determining the acceptability of the terms of a restructuring for a creditor body that has diverse and often indeterminate interests? 3/ Finally, where the debtor had issued in various jurisdictions, would the U.S. court have jurisdiction with respect to claims that are governed by a foreign law or subject to the jurisdiction of a foreign court? For these reasons, the authors have indicated that such legal uncertainties would need to be resolved—possibly through new legislation—in order to provide the necessary predictability to make the proposed approach workable.

1/ See Lee C. Buchheit and G. Mitu Gulati, Sovereign Bonds and the Collective Will, 51 Emory L.J. 1317 (2002).2/ Class action procedures under U.S. federal law are found in Rule 23 of the Federal Rules of Civil Procedure. “Representative actions” are available in English law, but are limited to proceedings by “parties with the same interest in a claim” (notably far less expansive than the U.S. federal law class action).3/ It is instructive that in the recent decision of New York district court denying class certification to two cases filed against Argentina, the court held that the class actions would have been un-manageable. See Lightwater Corp. Ltd v. Argentina & Macrotecnic Int’l Corp. v. Argentina, April 14, 2003 opinion of Judge Griesa of the United States District Court of the Southern District of New York. The class actions against Argentina had been framed in such a way that they would have allowed creditors within the defined class to opt out of the proceedings, in contrast to the Buchheit/Gulati proposal that would involve a mandatory class action without opt-out.

76. It was recommended that the SDRM should adopt a similar approach. Specifically, the claims of unsecured creditors holding different instruments could be aggregated for voting purposes (and receive the same terms under a restructuring agreement) because the activation of the SDRM would only be used where the debt was unsustainable; i.e., where, in the absence of a restructuring, creditors could expect there would be a general default anyway. In these circumstances, creditors would be willing to be treated as if their claims had, in fact, already been accelerated. Of course, this served to emphasize the importance of ensuring that the SDRM only be triggered in circumstances where the member’s debt was unsustainable.


77. While the SDRM proposal would give the sovereign the right to aggregate the claims of all unsecured creditors for voting purposes, it was recognized that, in some circumstances, this would create inflexibility. As noted in the previous section, there may be circumstances where the sovereign may wish to avoid aggregation so that it could provide different types of creditors with different treatment. To provide for this flexibility—and drawing upon the experience in the nonsovereign context—the SDRM proposal contemplated the possibility of the creation of “optional classes” that would be created to accommodate the particular preferences of different creditors. Under such an approach, support by a qualified majority of creditors in each class would be required to approve restructuring terms offered to all classes. While votes would be aggregated across instruments within the same class—thereby greatly reducing the leverage of holdouts—there would be no aggregation across different classes. Finally, while all creditors within the same class would need to receive the same restructuring terms (or menu of terms) treatment of creditors across groups could be different.

78. By way of example, a sovereign may need to restructure claims that were originally inter-bank claims or trade credit but have become claims of the sovereign because a guarantee had been called. In these cases, the sovereign may need to provide these creditors with terms that are preferable to those offered to bondholders, because of the need to resume normal inter-bank and trade financing after a the crisis subsides. Since these preferential term could only be offered to these trade and bank creditors, they would be placed in a separate class from bondholders (If the claims of bondholders and bank creditors placed in a single class, a menu of options would need to be offered to all creditors within the class on a uniform basis). The restructuring would only become effective if a qualified majority in each class supported the restructuring. For their part, bondholders may be willing to accept being treated less favorably because of the recognized need to resume trade and inter-bank credit. Such a classification system would become an optional tool for the sovereign (it could not be imposed by creditors) and would be used in circumstances where complete aggregation into a single class would not provide the most effective means of achieving the most viable restructuring agreement.


79. The SDRM proposal recognizes that any framework that allows for broad aggregation of claims will need to include safeguards against abuse. While it includes several features that are designed to address this risk, by far the most important is the Dispute Resolution Forum (DRF). To address the possibility that the sovereign would established fictitious claims, the DRF would be charged with overseeing a claims verification process that would be designed to ensure that only creditors with valid claims against the sovereign would be eligible to participate in the voting process and receive a claim upon the effectiveness of the restructuring agreement. Similarly, the DRF would be charged with resolving disputes that are likely to arise as to whether the claims, although valid, are being effectively voted by the sovereign. Specifically, claims would be excluded from the voting process if there was evidence that: (i) the creditor in question is under the control of the sovereign or (ii) undue influence has been placed by a sovereign on a creditor to vote in a particular manner.

80. Notwithstanding these safeguards, there was a recognition that the SDRM could not eliminate all risks regarding abuse. For example, even if the verification process required that the end-investors register their claims, this would not preclude the sovereign from establishing a special purpose vehicle as the end-investor. Moreover, it was also recognized that making a determination as to when the influence exerted by the sovereign upon, for example, a commercial bank would be “undue” for purposes of the SDRM would also be a complicated process. Would, for example, the provision of regulatory forbearance for domestic banks as compensation for their acceptance of onerous restructuring terms be considered inappropriate when such terms would also become binding on foreign investors?

IV. Conclusion

81. This paper has reviewed the potential contribution that “aggregation” can make to the resolution of collective action problems and problems associated with creditors coordination. At the same time, it has identified a number of risks associated with such a mechanism. Accordingly, when considering ways in which the restructuring of sovereign debt can be made more orderly, predictable and rapid, the critical question is whether an aggregation framework can be designed that manages to capture the benefits of aggregation while minimizing its risks. Since there is insufficient support for the establishment of the SDRM at this time, there is considerable merit in exploring the potential scope for aggregation under a contractual framework. As has been discussed, the private sector has already begun to make progress in this area, in the context of both the design of collective action clauses and the design of contractual frameworks that are established in the context of a crisis. Notwithstanding this progress, however, designing and implementing such a contractual framework faces a number of challenges and it is still too early to determine the extent to which aggregation will be accepted by the market, or the precise form of the contractual provisions. Against this background, staff is of the view that it would not be appropriate for the Fund to endorse a particular set of aggregation provisions at this time. However, staff will continue to monitor the use and evolution of aggregation provisions and will report to the Board on any significant developments.

Annex Aggregation in Corporate Restructuring Context

1. Corporate rehabilitation laws normally include robust features that are designed to address collective action and creditor coordination issues that arise from the existence of multiplicity of claims. At the same time, however, these provisions contain safeguards to ensure that aggregation does not give rise to discrimination among creditors with very different claims. Of course, one of the important factors that distinguishes the legal framework that restructures corporate debt from that which restructures sovereign debt is that corporate rehabilitation laws operate in the “shadow” of the corporate liquidation law. Specifically, if a rehabilitation plan fails to receive support from the specified majority of creditors, the corporation will normally be liquidated and assets will be distributed in accordance with the priority provisions set forth in the law. With respect to aggregation, these priority provisions play an important role in determining how different classes of creditors may be treated under a rehabilitation plan and, accordingly, provide important guidance on creditor classification issues.1

Secured and Priority Claims

2. In many cases, secured claims will represent a significant portion of the value of the debt owed by the debtor. To the extent that the law ensures--as is the case with traditional “composition” plans--that an approved plan will in no way preclude secured creditors from exercising their rights, there is generally no need to give secured creditors the right to vote, since their interests will not be impaired by the plan. Priority creditors are treated similarly under such a composition plan. These creditors (including, for example, post-petition creditors and--depending on the law--employees) do not vote on the plan but, upon the plan’s approval, they are entitled to receive full payment on their claims; the plan cannot impair the value of their claims.

3. The limitation of the above “composition” approach is that it effectively reduces the chances for a successful rehabilitation. For example, in the case of a secured creditor, the assets securing the claim may be vital to the success of the rehabilitation plan. Accordingly, unless the secured creditor is bound by the plan or the plan provides for full satisfaction of the secured creditor’s claims, the exercise of the creditor’s rights may render the plan’s implementation infeasible. Similarly, in certain circumstances, the only way in which a rehabilitation plan may succeed is if priority creditors receive less than the full value of their claims immediately upon the plan’s approval.

4. One approach that has been adopted by some countries to address this problem is to allow for secured creditors and priority creditors to vote as separate classes on a plan that would otherwise impair the value of their claims. The creation of separate classes is considered necessary since the nature of their rights under liquidation differ from those of unsecured creditors, and they accordingly also have different interests. To the extent that majority support is obtained from each of these classes, all secured creditors and priority creditors would be bound to the terms of the plan. In these circumstances, the law requires that any dissenting creditors be entitled to receive at least as much as they would have received under liquidation. A majority of secured creditors may be willing to accept an impairment in the value of their claims in circumstances where they have a long-term interest in the continuation of the enterprise (e.g., financial institutions), and such continuation requires the adoption of a plan that provides them with less than immediate cash payment on their collateral. Similarly, employees that are priority creditors may very well be willing to receive less than full payment on back wages if this is necessary to ensure the survival of the enterprise.

General Unsecured Creditors

5. Even if the law does not provide for voting by secured or priority creditors, it must provide an effective means by which general unsecured creditors can vote on a plan. A number of mechanisms may be used to increase the chance that a rehabilitation plan will be approved by these creditors.


6. Irrespective of whether the law provides for voting of classes of creditors, all insolvency laws must set forth rules that identify the minimum threshold of support of general unsecured creditors required to bind such creditors, and the voting procedures that are to be used to determine this support. Various majorities can be envisaged (two-thirds or three-fourths of the total value of unsecured claims), with the chances of approval increasing as the minimum percentage of required support goes down. One issue that needs to be addressed in this regard is whether calculation of votes should be based exclusively on the percentage of the value of the debt that supports the plan or whether it should also take into consideration the number of creditors that are supportive. For example, the law may require that the plan must be supported by both (i) two-thirds of the value of the debt, and (ii) one-half of the creditors in number. While such a two-tiered voting requirement will effectively raise the hurdle for approval, it may be justified on the basis of the principle that the insolvency law is designed to be a collective proceeding: if a single creditor holds a majority of the value of the debt, this rule prevents that creditor from imposing its support of the plan against the will of all the other creditors.

7. Regarding voting procedure, many countries have found it preferable to calculate the percentage of support on the basis of the percentage of those creditors that actually participate in the voting. Absentees are considered to have little interest in the proceedings, normally because their claims are small. To the extent such an approach is relied upon, it is critical that there be adequate notice provisions and that these notice requirements be effectively implemented. This is of particular importance when many of the creditors are nonresidents.


8. Some countries that have established classes for secured creditors and priority creditors also provide for the division of unsecured creditors into different classes.2 The creation of such classes is designed to enhance the prospects of rehabilitation in at least two respects. First, as in the case of secured and priority creditors, such classes are a useful way to identify the varying economic interests of unsecured creditors and, therefore, provide an appropriate framework for structuring the terms of the plan. Experience demonstrates that, as with secured and priority creditors, some unsecured creditors have different interests in terms of what they feel they need to receive under the plan. For example, while certain unsecured creditors may only be interested in immediate cash payments (e.g., discontinued vendors), other creditors that have a long-term interest in maintaining a relationship with the enterprise (e.g., ongoing trade creditors) may be willing to accept deferred payment or equity. The creation of classes on the basis of these interests and the structuring of the plan to accommodate them provide a greater chance that a rehabilitation plan will receive adequate support.

9. The second way in which the creation of classes of unsecured creditors enhances the chances of rehabilitation is that it provides a means for the court to utilize the requisite majority support of one class to make the plan binding on other classes, which do not support the plan. This is discussed more generally below.

“Cram-down” Authority

10. A few countries that provide for voting by secured and priority creditors and for the creation of different classes of unsecured creditors also include a mechanism that will enable the support of one class to make the plan binding on other classes (including classes of secured creditors and priority creditors) without their consent. Such a mechanism, which is designed to further enhance the chances of rehabilitation, is often referred to as a “cram-down” provision. If such a mechanism is relied upon (the merits are discussed below), it is important that protection be provided to the dissenting class to ensure that the priority rules that are established in liquidation procedures are respected. In particular, in addition to providing for the minimum level of protection for each dissenting creditor (which ensures that a creditor receives at least as much as it would have received under liquidation), discussed earlier, laws that seek to protect the relative rights of “crammed-down” classes of creditors also apply what is known as the “absolute priority rule.” Under this rule, a dissenting class of creditors may not be forced to receive less than the full value of its claims if creditors of a junior class receive any value.3

11. The creation of classes and the application of “cram-down” rules complicate both the law and its application by the court and the administrator. Where the institutional infrastructure is relatively well developed, such complexity may be merited, especially given the limitations of the traditional “composition” model, which can severely limit the opportunities for rehabilitation. However, where the capacity of the institutional infrastructure is limited, the inclusion of such rules will require careful consideration given that, among other things, their application may require the exercise of considerable discretion on economic issues. For example, the creation of separate classes of unsecured creditors will often require a categorization of such creditors by the court on the basis of their economic interests. Where creditors do not have confidence in the ability of the institutional infrastructure to exercise this discretion in an informed, independent, and predictable manner, such rules may actually undermine creditor confidence.


12. Some laws provide for the approval of plans by shareholders of the debtor enterprise, at least where the corporate form, the capital structure or the membership will be affected by the plan. In addition, when the debtor’s management proposes a plan, the terms of the plan may already have been approved by the shareholders. Depending on the type of enterprise in question (publicly traded or privately held), this may be required under the terms of the constitutive instrument of the enterprise. This is particularly the case when the plan involves debt-for-equity conversions, either through the transfer of existing shares or through the issuance of new shares. However, in circumstances where the law permits creditors or an administrator to propose a plan, and such a plan contemplates a debt-for-equity conversion, some countries allow this plan to be approved over the objection of the shareholders, irrespective of the terms of the constitutive instrument of the enterprise. As a result, such plans can result in existing shareholders being entirely displaced in the new enterprise without their consent.

Court Approval

13. As has been discussed above, many countries enable the courts to play an active role in “binding in” creditors by making the plan enforceable upon a class of creditors in circumstances where they have not approved the plan. Conversely, in cases where the plan has been approved by the requisite majority of creditors, the court will normally have the authority to reject the plan on the grounds that the interests of dissenting creditors have not been adequately protected (because, for example, they have not received as much as they would have received in liquidation) or if there is evidence of fraud in the approval process. In addition, some laws may give the court the authority to reject a plan on the grounds that it is not feasible. This may be justified, for example, where secured creditors are not bound by the plan but the plan does not provide for full satisfaction of the secured claims of these creditors. In these cases, the court may reject the plan if it considers that secured creditors will exercise their rights against the collateral and that such an event will render the plan nonviable. The risk of this occurring is quite limited, however, if a qualified and independent administrator has been involved in the plan’s preparation and approval. In such circumstances, the court would normally be expected to approve a plan that has been approved by the requisite majority of creditors.

14. After the approval of the plan by the court, several countries permit the court to authorize continued supervision of the affairs of the debtor, to varying degrees, by a supervisor or administrator after the confirmation of the plan.


Communiqué of the International Monetary and Financial Committee of the Board of Governors of the International Monetary Fund, Washington, D.C. April 12, 2003 (available at:


In the preparation of this paper, staff benefited from extensive discussions with market participants, workout professionals, and academic economists.


The discussion of the potential use of aggregation in this paper is limited to the restructuring of debt; it does not include a discussion of voluntary debt swaps in the context of liability management operations for which the use of a collective framework would not be appropriate.


For purposes of this paper, aggregation of claims for voting purposes applies to both majority restructuring provision and majority enforcement provisions.


Factors that will have a bearing on the first point include the economic circumstances of the member; the extent to which nonparticipating claims may be small, rated as “selective default” by credit rating agencies (notwithstanding the fact that they continue to be paid); and the likely liquidity of the instrument.


Factors that will have a bearing on the second point include the appetite for litigation (including willingness to bear the financial costs and possible reputational damage), and the availability of assets or payment streams vulnerable to attachment.


In addition, a number of emerging market sovereigns have reopened old bond issues, and the Philippines has launched a new issue, all governed by New York Law, none of which include CACs.


A recent paper by Barry Eichengreen and Ashoka Mody (“Is Aggregation a Problem for Sovereign Debt Restructuring,” AEA Papers and Proceedings, May 2003, pp. 80–84), shows that—controlling for fundamentals—launch spreads are higher for debtor countries that have many bonds outstanding. This suggests that market prices may reflect the problems that could be encountered by a debtor seeking to restructure multiple bond issues.


For a detailed discussion of collective action clauses, see The Design and Effectiveness of Collective Action Clauses, SM/02/173 (06/07/02) and Collective Action Clauses—Recent Developments and Issues, SM/03/102 (3/25/03).


It is worth noting that Ukraine used a debt exchange in combination with the use of collective action clauses in order to ensure that the instruments issued in the restructuring would be fully fungible.


While the trustee may also initiate proceedings at its own discretion, it will not normally do so because of the risks and costs involved.


The Uruguay bonds issued in a recent exchange offer are governed by New York law and are issued under a trust indenture. The bonds require a 25 percent majority for acceleration and a 66⅔ percent majority for de-acceleration. Under the terms of the trust indenture, litigation can be initiated by the trustee either at its discretion, or if, among other things, the holders of not less than 25 percent of outstanding principal of a particular series shall have made a written request to the trustee to do so.


As discussed earlier, for bonds issued under a trust deed, the required majority of bondholders can also effectively block litigation after a default has occurred.


It is not, however, the approach advocated by a minority of market participants who are of the view that future interest should be taken into account so that the net present value of the original claim can be calculated.


See: Bartholomew, Ed. “Two Step Debt Restructuring” 2002, J.P. Morgan, New York. (available at


The ability of investors to hold out under their original instruments, as opposed to being bound into a collective framework in an IDC, can be analyzed as an option. For creditworthy debtors, such an option would be out of the money, and would thus trade at a low price. As credit quality deteriorates, or in periods of extreme asset price volatility, the option would be in the money and have economic value. This is consistent with the recent experience with the inclusion of CACs in sovereign bonds issued in the New York market, which suggests that there is no premium for the inclusion of such clauses, and the expectation that in the context of an imminent restructuring, investors would require a financial incentive to exchange their claims for an IDC. It is also consistent with empirical work which has found a relationship between the market pricing of collective action clauses and the debtor’s credit rating. See: Eichengreen, B. and Moody, A., 2000, “Would Collective Action Clauses Raise Borrowing Costs?,” NBER Working Paper No. 7458 (available at


If one assumes that Argentina will treat all similarly situated retail investors equally, one could argue that ABRA creates its own collective action problem: such investors may wish to stay out of ABRA to avoid paying the fee but will be able to benefit from the benefits of its negotiations with the sovereign debtor.


There was an early recognition that aggregating the claims of official bilateral creditors with those of private creditors would not be feasible, particularly if it meant that these two groups of creditors would be required to receive the same terms or menu of terms under a restructuring. Two different options have been considered. The first would exclude official claims, but would envisage the development of procedures to ensure that these claims were restructured in parallel with privates claims in a manner that addressed inter-creditor equity. The second approach would involve establishing a separate mandatory class for private and official creditors—a qualified majority in each class would be necessary for the overall restructuring agreement to become effective. By placing creditors in different classes, it would enable each group of creditors to receive different term from each other.


Multilateral development banks normally control the amount and type of security that is granted in collateralized borrowings through negative pledge clauses contained in loans extended by them. These clauses typically preclude borrowers from pledging or encumbering their present or future assets to secure other creditors, without equally and ratably securing those creditors whose contracts contain the negative pledge clause. For a detailed discussion of negative pledge clauses, see Assessing Public Sector Borrowing Collateralized on Future Flow Receivables, SM/03/210, 6/16/03.


The analysis set forth in this Attachment is drawn from Orderly and Effective Insolvency Procedures: Key Issues (IMF Legal Department, 1999).


Some countries also allow for the creation of different classes of secured creditors on the basis that, depending on the nature of their claims, they may have different economic interests from each other.


For this purpose, seniority is based on the ranking applicable in liquidation (secured creditors, priority creditors, general unsecured creditors, subordinated creditors).