Bail-Ins, Bail-outs, Burden Sharing and Private Sector ...



Bail-ins, Bailouts, Burden Sharing

and Private Sector Involvement in Crisis Resolution:

The G-7 Framework and Some Suggestions

on the Open Unresolved Issues

by

Nouriel Roubini

Stern School of Business

New York University

July 16, 2001

Introduction

PSI is the most controversial issue in architecture reform

The issue of bail-ins versus bailouts – or private sector involvement in crisis resolution - is the most controversial question in the debate on the reform of the international financial architecture. There is broad agreement on measures for crisis prevention; much less so about measures and approaches to crisis resolution. Even the definition of the problem has been debated with different terms used over time to characterize the problem: bail-in, burden sharing, private sector involvement in crisis resolution, constructive engagement of the private sector

Main Issue in PSI: what to do when there is a crisis and an external financing gap?

In principle there are three options:

1. A big “bail-out” in the form of an official support package filling the entire financing gap (where the term “bail-out” is loosely used to describe large official packages)

2. A full bail-in of private investors (debt rescheduling, restructuring, reduction)

3. A combination of official financing, “appropriate” PSI and policy adjustment by the crisis country

Note that the external financing gap is endogenous, thus there is a role for catalytic IMF financing in some cases.

Trade-offs in PSI

There is some trade-off between the amount of bail-in versus the amount of bailout given an external financing gap: more of one means less need of the other. Ideally, one would want to keep official support to the minimum necessary (to avoid moral hazard) but also to avoid more coercive forms of PSI (as they may negatively affect private flows of capital to emerging markets) For example, Kohler speaks of “constructive engagement” and wants small IMF packages and no coercive PSI

But there is some partial tension in this view: smaller IMF packages may mean more PSI and more PSI of a more coercive form. While less coercive PSI may mean the need for larger official packages. The new US administration faces a similar tension between the Meltzerite views of some (no more large bailouts and allowing for more restructurings and defaults) and the Wall Street and national security interest groups (who tend to prefer bail-outs to bail-ins). The catalytic approach, when successful, can partly obviate this tension as official financing together with policy adjustment will restore investors’ confidence and market access and thus avoid the need for more coercive forms of PSI.

PSI in the 1980s versus the 1990s

The 1980s developing countries’ debt crisis had its own PSI (suspension of payments on syndicated bank loans, concerted loans rollovers, new money) and eventually led to debt reduction (the Brady Plan)

So, what is new in the 1990s?

Instruments (bonds and short term interbank lines rather than syndicated medium-long term bank loans)

Creditors (bondholders in addition to banks)

Debtors (private debtors in addition to sovereign ones)

In the 1980s the challenge was to restructure medium and long term syndicated bank loans. In the 1990s the challenge was to restructure sovereign and private bonds and short-term interbank lines.

There are a number of flawed arguments on how easy was to do PSI in the 1980s versus the 1990s. It has been argued that in the 1980s it was easy to restructure loans of a small number of homogeneous regulated banks pliant to forbearance while in the 1990s it would be impossible to restructure bonds (without collective action clauses) held by thousands of creditors. Also, it would be hard to restructure interbank lines as investors would rush to the door before the concerted rollovers could be arranged. The 1990s reality instead had been that there has been lots of PSI, both bond restructurings and interbank rollover arrangements.

Moreover, in the 1980s PSI was not that easy to arrange as there were:

1. Collective action problems of coordinating many different creditors

2. Hundreds of banks with different interests

3. Holdout problems especially among smaller banks

4. Non-homogenous syndicated loans that had to be restructured in more homogenous instruments.

Conversely, the 1990s experience has been that:

sovereign bond restructurings are possible even without collective action clauses (CACs):; see the cases of Pakistan, Ukraine, Russia, Ecuador (and Argentina);

the bail-in of interbank lines is also possible: see Korea, Indonesia, Thailand, Russia, Brazil…and Turkey.

In general the 1990s crises were addressed with a combination of partial bailouts and bail-ins in spite of the superficial perception among some that international financial crises were mostly dealt with large bailouts.

Current G7 PSI doctrine and 1990s experiences with PSI

Rationale for PSI

The rationale for PSI is pretty straightforward:

1. If there is a crisis, it is likely that there will be an external financing gap even after policy adjustment by the country

2. Official support can help to fill the gap but not fully. Not enough official money under normal circumstances

3. Exceptional financing is not only not feasible but also not desirable apart from few special cases as large bail-outs may lead to creditor and debtor’s “moral hazard”

Given 1-3, there is a need for “appropriate” PSI that will help to fill in the external financing gap.

Current G-7 PSI Doctrine

The current G-7 PSI doctrine, as evolved from the G-7 Kohl Summit architecture report through the April 2000 G-7 operational guidelines and the September 2000 IMFC statement, can be summarized as follows:

Case by case approach rather than strict, rigid rules

But discretion constrained by principles, considerations, tools, criteria and guidelines

I.e. “constrained discretion”

Rigid rules initially deemed not realistic because of complexity and novelty of the issues to be solved.

Idea of case studies leading eventually to “case law”

1. Some fuzziness and ambiguity on large systemic liquidity cases with preference for catalytic IMF financing when deemed likely to be successful.

2. Still open debate on when access should be high or low and when PSI should be concerted or catalytic.

3. Cooperative solution to be preferred to coercive ones when possible

The G-7 preference for “constrained discretion” rather than rigid rules in the 1999 Architecture Report derived from the fact that, in the absence of case law, there was a set of complex and difficult questions on how to do PSI that had no clear ex-ante answers in the absence of actual experience with cases studies:

When to do PSI and when not?

Which type of PSI (coercive, voluntary)?

Which claims to include?

Which creditors?

What do to in liquidity cases?

What to do in systemically important cases where contagion is possible?

How to distinguish insolvency from illiquidity?

How much adjustment versus financing?

How to divide the burden sharing between private and official creditors?

How to measure comparability?

Allow reverse comparability?

What is the appropriate sequencing btw Paris Club and private sector claims restructuring?

Which process to follow for debt restructuring? Unilateral exchange offers or a more formal process with use of CACs, formal creditor committees and negotiations between the debtor and creditors?

Thus, the case-by-case approach with principles and tools and criteria combined clear guidelines with necessary discretion and flexibility to address appropriately each case.

There were two other elements of the G7 PSI doctrine:

1. “Appropriate PSI”

2. Try “cooperative, voluntary solutions” rather than coercive ones as much as possible to avoid hurting long term capital flows to emerging markets.

“Appropriate” PSI partly depended on:

a. Where a country stands in the spectrum going from illiquidity to insolvency cases

b. Whether the country is large and systemically important or not, i.e. whether the crisis may lead to “international contagion”

Evolution of Cases Studies in the Last Decade

The Mexican peso crisis of 1994-95 was the closest to a full “bail-out” case as enough official money was provided to allow the exit of those investors who did not want to rollover their holdings of Tesobonos.

Asian Crisis in 1997-98: in Korea, Thailand and Indonesia concerted rollovers of interbank lines were arranged. No sovereign bonded debt restructurings as these were “de minimis” in Asia. But in Indonesia there was a suspension of payments by private corporates and financial institution as the collapse of the rupiah led to the near bankruptcy of a large section of the corporate sector

Paris Club comparability applied for the first time to bonds in Pakistan in 1999

Brazilian crisis, monitoring of interbank lines in 1999 and commitment to maintain interbank exposure after partial roll-off of such lines

Romania and Ukraine piecemeal restructurings of sovereign bonded and quasi-bonded liabilities in 1999

Russia default and restructuring of bank and bonded debt liabilities (1998-2000)

Ecuador Bradies default and restructuring in 1999-2000

Sovereign bond exchanges in Russia, Ukraine, Pakistan and Ecuador in 2000.

Turkey: monitoring of cross border interbank lines exposure in 2000-2001

Argentina’s mega swap in 2001: effectively a voluntary market solution.

Open Cases Studies and Issues

Cote d’Ivoire and Nigeria: Bradies restructurings in the near future?

1. What to do if Argentina and Turkey’s rescues do not work? More concerted PSI and/or more official finance?

2. How to deal with contagion if - as likely - Argentina and Turkey get in serious trouble again?

3. In case of contagion, ring-fence the “well-behaved” (CCL for Brazil for example?) with official finance or involve somehow the private sector?

Evolution of the G-7/G10 debate on PSI and G-7 doctrine[1]

The Rey Report. The G-7/G10 debate on PSI began when the G-10 organized a working group after the Mexican crisis bail-out. Some Europeans were concerned that the high levels of access and absence of a bond restructuring in Mexico would set a negative precedent and cause moral hazard. The Rey report recommended the adoption of bond clauses and called for the IMF to modify its policy of lending into arrears (LIA). But the report recommendations were not made into policy.

Private sector resistance to the Rey Report. The private sector, namely the Institute for International Finance, strongly objected to both the bond clause recommendation and the LIA recommendation.

The G22 working group on financial crises. The next official initiative was the G22 working group on financial crises. This group tried to bring emerging market debtors into the discussion and, unlike the Rey report that had stressed the problems deriving from bonded debt driven crises, it also addressed the question of interbank rollover crises such as those observed in Asia. In addition to the question of collective action clauses in bonds it covered new issues and recommendations: contingent credit lines from private banks to provide emergency liquidity, a call for better insolvency regimes, and a suggestion that in some cases, after receiving indications about the amount of official financing that would be available, debtor countries would approach their creditors for a restructuring. Such restructuring would ideally allow a country to avoid default and a suspension of payments, but such a positive outcome could not be guaranteed. But note that the G22 did not call for restrictions on the availability of official financing or indicate that some form of debt restructuring would always accompany the provision of official finance.

Lending into arrears policy change by the IMF. The G22 also recommended that the IMF needed to make it clear that if country was working with its creditors in good faith and taking appropriate reforms, it would lend into arrears. The IMF subsequently amended its policy on lending into arrears to allow lending into arrears on bonded debt.

Contagion and creation of the CCL. In the aftermath of the Russian crisis and its contagion to Brazil and other emerging markets, the G-7 leaned towards developing improved mechanisms to ward off contagion, such as the IMF’s CCL.

The Kohln Architecture report: PSI principles, considerations and tools (“constrained discretion”). The G-7 directly took up the PSI issue in the debate on architecture in spring and summer of 1999. This was a framework of considerations, principles and tools that basically introduced a set of principles and considerations, presented a list of ways the private sector could be involved, and said that the right approach would depend on the circumstances of a given case. The official sector also made clear that it did not consider bonds to be inherently privileged relative to banks. This framework can be described as a “case by case” approach constrained by principles, considerations and tools, i.e. a “constrained discretion” approach to PSI.

The G-7 framework for PSI was partly a response to several concerns:

1. concern that large bailout packages in Mexico and Asia were leading to moral hazard;

2. concern that new official forms of financing such as the SRF, the CCL and the renewal of IMF quotas should be accompanied with rules restricting excessive official financing;

3. concern that the application of PSI in specific case studies in Asia and afterwards did not follow clear rules and was thus a source of uncertainty and confusion for market participants.

Paris Club comparability principle applied to sovereign bonds. In early 1999 for the first time ever, the Paris Club applied the comparability approach to sovereign bond by asking Pakistan to restructure a set of bonds that were coming due at the end of 1999 and early 2000. This case and that of Ecuador opened the issue of whether the Paris Club, the IMF and the G-7 were forcing the countries to default and restructure their bonds.

The April 2000 operational guidelines. The G-7 attempted to give operational substance to their Kohln report principles - also taking stock of the lessons from experience of bond restructurings- in their April 2000 “operational guidelines” for PSI. The G-7 statement considered three types of cases – cases where the emphasis should be placed on catalytic official financing (Mexico), cases where official action could help overcome coordination problems (Korea, Brazil) and “restructuring cases” (Ukraine, Ecuador). It then laid out a set of guidelines for official action in such restructuring cases. Intentionally, the statement did not address the controversial issue of the right level of official financing, i.e. when exceptional financing is needed or not. Thus, the hard questions of what to do in large, systemic semi-liquidity cases were largely finessed again.

G-7’s Central Banks support for standstills and coercive PSI. The G10 central banks have been generally more sympathetic to the idea of standstill and the necessity to limit exceptionally large official packages to limit moral hazard. The Bank of England and Bank of Canada argued in a paper that limited official financing and payments standstills might be the right approach, even in systemic liquidity cases. Support for standstills and limited official finance in crises has also been expressed by the Bank of Italy and Bundesbank in documents and official statements. Quite interestingly, central banks have expressed greater sympathy for standstills and limited official money than finance ministries and treasury departments. There may be a principal agent problem here: the closer is an agent to decisional power (as central banks in monetary policy and bank regulation) the greater its sympathy for qualified benevolent discretion and aversion to rules (thus the finance ministries preference for discretion in PSI issues). The farther is an agent away from direct decisional power (as central banks on PSI issues) the greater their sympathy for rules and aversion to discretion.

The IMF’s attempt to frame operationally PSI guidelines. The IMF did not find the G-7 “guidelines” neither very operational nor helpful—as they clearly finessed the most difficult PSI issues, i.e. the right amount of official financing in crises and whether or not private creditors should be involved in liquidity cases. In response to the guidelines, the IMF produced two papers that provided a more precise and clear, if controversial, set of operational guidelines. The first presented an IMF’s operational “framework” which effectively reduced the possible contingencies to a two by two matrix: official finance could either be exceptional or not, and “PSI” could either by “catalytic” or “concerted.” The second paper discussed the pros and cons of standstills. The G-7 were somewhat critical of the IMF framework. The Europeans, being more wary than the US of the use of large packages, were skeptical of the IMF’s framework matrix box that allowed “exceptional access” to be combined with “catalytic” PSI in some cases. While the U.S. objected to narrowing down to a binary distinction between catalytic and concerted the wide range of possible actions that the private sector could take to contribute to crisis resolution.

IMFC September 2000 statement. The G-7 debate next led to the fall 2000 IMFC statement, which represented a compromise between the IMF’s framework and the G-7’s April 2000 guidelines. The IMF statement was adopted by the IMF staff as its operational framework for PSI. The IMF’s operational framework, however, finessed again the hard questions of large systemic semi-liquidity cases: when to have low levels of official financing and when to have high levels of official financing; when should high levels of official financing be associated with concerted PSI and when should they be associated with catalytic PSI. Thus, a lot of room was left to discretionary policy in specific cases.

Coercive PSI? One of the dimensions of the G-7 debate remained the one of how coercive PSI should be. Kohler found the term private sector involvement too much associated with coercive solutions and prefer to talk of constructive engagement of the private sector. But he also expressed, at least in theory, sympathy for more limited official financing, leaving open the question of how less official money could be consistent with less coercive PSI. While catalytic solutions, if successful, may allow for less coercive PSI, they usually tend to be large to be successful. One is still looking for a case where low access had successful catalytic effects. Similarly, the new US administration position is against coercive and concerted forms of PSI, excessive official sector involvement in the restructuring process and against sanctioning or inducing sovereign defaults. But its support in principle of more limited official finance needs to be squared with its preference for less coercive PSI solutions.

The 2001 G8 Genoa summit report transmitted by the Finance Ministers. In preparation for the Genoa G-8 summit, the G-7 finance ministries tried again to find a compromise on the difficult PSI issues. The rise of a new US administration that was, at least verbally, more skeptical of large IMF packages led to a formal greater convergence between the European views that large access should be limited and the US views on this, with Japan still favoring large access, and linking to the low quotas in Asia. However, as the Turkey (and partly Argentina) case showed, there has continued to be a gap between official rhetoric against large bailout packages and actual policy; in Turkey the program was significantly augmented after the devaluation amounting to an effective bailout of the country; Argentina was allowed to let the fiscal targets slip and keep previous large official finance flowing. The US remains wary of standstills, forced defaults, coercive PSI and automatic linking of official financing to concerted PSI. While the Europeans remain somewhat more sympathetic towards concerted forms of PSI, use of standstills in some cases and linking large financing to meaningful PSI (see the April 2001 EFC document). Since it was recognized that previous G-7 agreements that the IMF be more explicit about financing assumptions in programs and whether PSI was needed or not had not been systematically implemented, the G-7 compromise decoupled IMF support from “automatic” PSI but requested that the IMF be more explicit and informative to its shareholders about these issues and provide a better monitoring of PSI in its programs. Again, the real tough issues of high/low access and catalytic/concerted PSI were finessed to leave room for discretionary decisions in specific cases. Note that there is almost always PSI in IMF programs if PSI is defined as the residual that is not filled by official finance, given scale of possible outflows. In this sense the G-7 current view is consistent with the view that there always need to be some private financing but no requirement that country obtain that financing in coercive/concerted ways rather than relying on catalytic approach. The trouble is that in many cases the catalytic approach may not succeed and thus one needs more clarity on when and how to have more concerted forms of PSI.

Open issues in PSI doctrine and practice and the problems of large systemically important “illiquid” countries

State of play and the difficult questions that the G-7/IMF need to address

Small non-systemic countries are easier to deal with as recent cases have shown that bonded debt restructurings are possible (Pakistan, Ecuador, Russia, Ukraine) with unilateral exchange offers.

Even in these cases there is an ongoing G-7 debate on whether defaults/debt suspensions should be effectively sanctioned and how pro-active the official sector should be in the debt restructuring process (i.e. whether PSI should be concerted and deals monitored by the official sector for medium term consistency or whether, once the official sector has decided on the level of low access, the debtor and creditors should work out a restructuring on their own with little official sector involvement).

• The amount of access in large, systemic and semi-liquidity cases remains an open issue with actual practice having been biased so far in favor of large access. Financing beyond normal access has not been the exception but rather the rule in these cases.

• Whether and when PSI should be concerted rather than catalytic in these large systemic cases remains an open issue. In such cases, but even more generally in other non systemic cases, the IMF programs have not been very clear about the PSI conditions of these programs and the consequences of failing to implement adequate PSI.

• The distinction between illiquidity and insolvency remains controversial and difficult to implement, especially since most cases remain in the gray middle range where illiquidity and loss of market access is triggered by serious economic and policy shortcomings even when the country is not strictly “insolvent” by some standards.

• The monitoring of PSI in IMF programs has not been adequate and the consistency of programs’ assumptions about private financing with medium term debt sustainability not adequately fleshed out

Thus, the G-7 and the IMF will thus need to address in the next few months a complex set of issues that have still been largely finessed so far:

How much to restrict the access to IMF resources?

When to provide low or high access?

Under which conditions would exceptional official finance be provided?

Should provision of exceptional finance be softly or strictly linked to concerted forms of PSI?

Should the private sector be always involved in a “liquidity” case and if so, how?

When to rely on “catalytic PSI” (large money with soft PSI that restores confidence and market access) and when to rely on “concerted PSI” (more pro-active engagement of the official restore to induce restructuring of claims)?

Is the role of the official sector to provide whichever finance is deemed appropriate and then let debtors and creditors to work out claims on their own, or should the official sector be more involved in ensuring that appropriate forms of PSI and coordination are achieved?

Whether to sanction standstills or not and under which circumstances?

How to ensure that IMF programs and the assumptions about private financing are consistent with medium term debt sustainability?

Should PSI considerations be better incorporated in IMF surveillance and conditionality?

How to strengthen the monitoring of PSI?

What are the consequences of failing to provide appropriate PSI?

To analyze more systematically these issues, I consider next the difficult issues of what to do in large systemic cases where the country may or may not be only illiquid.

How to deal with different cases: first, it depends on whether it is illiquidity or insolvency

A first way to figure out what is the appropriate form of PSI and the amount of access to IMF resources is to decide where a country falls in the continuous spectrum going from pure liquidity cases to pure insolvency cases:

1.In cases of “pure” illiquidity (Mexico is the closest case even if there were significant policy shortcomings even in that case), a pure full bailout solution may be the right one or better a CCL type of facility. However, “pure” liquidity cases are quite rare leaving open the issue of how much official finance to provide to large countries that are illiquid but suffer of economic shortcomings.

2.In cases of pure insolvency (Ecuador, Russia?), debt restructuring in the form of haircuts and debt reduction may be warranted. It is however hard to clearly define when a country is insolvent and figure out who should be making such a determination.

3. In middle cases (Ukraine, Pakistan, …) (significant macro problems but not be insolvency, loss of market access and large lumpy debt payments) debt restructuring (at below market interest rates but no principal haircut) may be the right policy.

• Open issue: how to distinguish between illiquidity and insolvency? Most cases are in a gray region where illiquid countries have serious macro and structural problems and countries that look insolvent may not be so given serious reform and adjustment.

• Also, applying the solutions of debt restructuring and reduction becomes harder in cases where the country is large and systemically important. In these cases, there is a political tendency to consider the country as “illiquid” rather than “insolvent”

Policy for Liquidity Cases

But the right policy for “liquidity” cases is more complex than a full bailout suggested by theory. Suppose there is a pure liquidity case, there is no uncertainty on fundamentals, no risk aversion of investors and no policy problems. Then, a full bailout and a full bail-in solution are equivalent (LOLR vs. bank holiday) as latter solves also the collective action problem. But conceptually a full bail-in solution may even be better than the full bail-out as its threat would enforce good equilibrium without actually imposing it (as in the case of a bank holiday when there is a domestic banking panic)

But real liquidity cases are different from the abstract ideal as there is uncertainty, risk aversion and policy and macro problems in countries subject to a run. Runs usually do not occur in practice in pure liquidity cases, even if theory says otherwise. Illiquid countries are countries that have serious macro and/or structural problems. Run do not occur in a vacuum but there are caused by economic and policy shortcomings. This is why “appropriate” PSI in addition to policy adjustment and “appropriate” amounts of official financing is necessary to address these “liquidity” crises. However, under these real conditions, the threat of a coercive bail-in in “liquidity with problems crises” may lead to a “rush to the exits” and other destabilizing outcomes.

Note that most of the shortcomings of standstills – discussed below - are particularly serious in large systemic cases where international contagion is likely, but they also apply to non-systemic small countries cases.

Pros and cons of coercive bail-ins and “standstills” on debt payments

The main argument in favor of coercive bail-ins and “standstills” on external debt payments is that they solve the “collective action problem” and may prevent a sudden panic and capital flight.

But such “standstills” have also several destabilizing shortcomings:

1. They may lower long run lending and capital flows to emerging markets

2. They may lead to a “rush to the exits” (as in the case of anticipated capital controls)

3. They may lead to international financial contagion (see the Russia/Malaysia contagion to emerging markets in the summer of 1998)

4. Partial standstills on sovereign claims may not be enough as private claims may run too. Thus, need to lock in the latter too if they start to run.

5. Need for exchange controls and capital controls if private claims run too and there is generalized capital flight.

6. Standstills on private claims are hard to arrange and you also have the risk of “asset stripping” (as in Indonesia)

7. There are also a number of complex legal issues:

14 The IMF’s Article VIII.2.b is not likely to be amended given current G-7 views on this issue;

15 a court enforced “stay of litigation” after a debt suspension is unlikely to occur in the absence of such amendment even if in principle one could argue that a debtor that is dealing in good faith with its creditors should be granted such a stay;

16 the IMF’s “Lending into Arrears” policy is valid and appropriate but it will not formally prevent litigation if creditors decide to take their case to court.

Big open issue: what to do in large country systemically important cases?

The hardest open issue in PSI policy is what to do when a large systemically important country gets in trouble? Ideally, a combination of policies would be the appropriate response:

1. Policy adjustment on the part of the country especially when this is not a pure liquidity case

2.Large but not systematically exceptional official financing (to prevent moral hazard)

3. “Appropriate” forms of PSI

In the best cases, one would hope that the catalytic approach would work but this is not going to be always the case especially when the crisis country has serious policy problems and uncertain policy credibility. Thus, more concerted forms of PSI may become necessary. Moreover, when large systemic countries suffer of macro and policy problems, the issue emerges of how large access to IMF resources should be and whether large access should be conditional on concerted PSI.

Large systemically important liquidity cases

While many episodes where a large systemically important country got into trouble had elements of a liquidity crisis (as investors panicked), in almost all of these episodes there were serious policy shortcomings, macro weaknesses and/or structural problems. In other terms, there are almost never “pure” liquidity cases. Even country close to being illiquid (rather than insolvent) become so because of some fundamental weakness. This is why a “full bail-out” solution is not optimal in these “liquidity cases”; PSI and policy adjustment should be part of the solution. There is also a growing consensus among the G-7 (both the new US administration and the Europeans) that large official packages should become exceptional rather than the rule. However, there is a big gap between the public rhetoric about “no more big bail-outs” and the reality of specific cases.

Political pressure to bail-out systemically important countries

When a large systemically important country gets in trouble, the political pressure to bailout this country (with an exceptionally large package) is common. The recent episodes in Argentina and Turkey confirm that bailouts are more common than bail-ins. All these programs were long in official support and quite short in their PSI elements. Note that based on standard measures of debt sustainability, Argentina and Turkey were in worse conditions than, say, Ukraine or Pakistan. While in Ukraine and Pakistan, a debt restructuring at below market rates (10% rather 20%) was forced, in Argentina and Turkey, there was no meaningful PSI as the Argentine mega swap occurred at market rates while in Turkey large official support allowed a rollover of the domestic debt at very high market-determined real interest rates.

This leaves open the question of whether the bar to declare a large country as insolvent (or as requiring a forced/involuntary restructuring of its external and domestic debt) has been set too high? The answer is probably yes as a smaller country would have been “forced/induced” to restructure rather than being provided with a large amount of official resources.

The incentives to bail-out large countries stems from several factors:

These countries tend to be systemic and there is concern about contagion (Argentina, Brazil).

Often they are subject to a liquidity run, in spite of also having fundamental weaknesses; thus, some exceptional package may be part of the optimal policy response.

They are often geo-strategically, politically and militarily important (Turkey, Korea, Argentina, …)

Thus, in reality actual packages for large countries have been long in official financing (exceptional financing) and skimpy in their bail-in components (soft forms of PSI, reliance on the catalytic role of IMF financing as a way to restore confidence and market access).

Current G-7 views suggests a potential shift away from large packages but recent cases, and most likely future ones, will challenge this new doctrine. It is still to be seen whether a collapse in a country like Argentina or Turkey – more likely now given the renewed recent turmoil in these countries - will lead to further official support (multilateral and/or bilateral) or whether the G-7 will allow these countries to restructure on involuntary terms (“default”) their domestic and external debt.

PSI versus Bailout in Argentina and Turkey

The lack of meaningful PSI in the recent Argentina and Turkey programs is particularly disturbing as exceptional official finance was used to try to stem the crisis faced by these two countries.

The Turkish bailout

In Turkey, the only PSI measure was of the softest nature, a mild monitoring of international exposure and a vague commitment of some international banks to maintain interbank exposure. This was a milder form of PSI than even the mild Brazilian interbank rollover. And it became soon quite leaky as some banks (especially U.S. ones but partly the German ones as well) decided to roll-off their positions rather than roll them over, in face of the unwillingness of U.S. regulatory bodies to suggest to their banks to maintain exposure. The ensuing cutback in exposure (up to almost 50% of the stock of U.S. banks claims held in December 2000) was a source of pressure on the liquidity of domestic Turkish banks and played a role in the Turkish government decision to intervene in forex market, thus contributing to the significant further loss of forex reserves after the February 2001 devaluation.

Moreover, in Turkey the unwillingness of the IMF (eventually supported by most G-7 governments) to consider a semi-involuntary restructuring of the domestic public debt was also a failure to consider more meaningful PSI. The objections to such involuntary restructuring have some merit (risk of causing a greater loss of confidence by investors, losses inflicted to already financially weak domestic bank). But the consequences of the decision to follow a market solution (market based rollover of domestic debt and official financing of the fiscal gap) were troublesome: in face of significant policy uncertainty, the scenario for a stabilization of the debt to GDP ratio was quite unrealistic as it implied very large primary surpluses, a sharp fall in real interest rates, a rapid recovery in growth and significant privatization receipts.

While a forced restructuring of domestic debt on highly coercive terms would have been destabilizing for the domestic financial system, some middle solutions that were not fully voluntary may have worked better: for example one could have administratively capped the real interest rate received by the holders of domestic debt to positive but not prohibitive rates. This semi-involuntary rollover of debt, as well as some semi-involuntary lengthening of the very short maturity of such debt, may have allowed to minimize the losses (or minimize the gains to bondholders from artificially high real interest rates) to financial institutions while allowing a more sustainable debt dynamics. While lack of policy performance on the program would have eventually doomed both the voluntary rollover path that was chosen and the involuntary rollover path that was not chosen, the latter may have better controlled real interest rates and the debt dynamics even in the presence of policy uncertainty while at the same minimizing the risks of an excessively large program to official resources.

Thus, a meaningful PSI policy in Turkey would have implied a strong commitment to reform and adjustment, the provision of less official resources and a combination of a binding interbank rollover agreement and a semi-involuntary restructuring of domestic debt. Instead the chosen solution amounted to a semi-full bailout of the country (as the new official resources committed after the devaluation covered the full expected fiscal gap) in exchange for a strong policy adjustment. The PSI components of the program were so mild to be effectively fictional.

This semi-full bailout solution was particularly disturbing as it was based on political considerations (the geo-strategic interest of the US that saw Turkey as a “strong friend and ally” and the financial and regional interest of some European countries - namely Germany and other EU countries) rather than on sound economic reasons. In the U.S. the unwillingness of Treasury to advocate more meaningful PSI (semi-involuntary interbank and domestic debt rollover) left open the door for political geo-strategic interest groups to advocate and force a full bailout in face of a large fiscal gap that could not be otherwise filled. In Europe, the rhetoric in favor of limiting exceptional access gave in to the fact that Turkey’s stability, its trade links to Europe and European financial institutions exposure to the country also made the country of strategic importance. The arguments used in the U.S. to justify the Turkish bailout (that the problem had been inherited by a previous administration, that the program required more prior actions and structural reforms and that the program provided no bilateral money) did not conceal the basic fact that the bailout was more based on political than economic considerations and signaled a huge gap between official views on limiting large IMF packages and actual practice. Indeed, market participants' reactions to the Turkish bailout suggest that the market interpreted this bailout as the death of meaningful PSI for large, systemically or geo-strategically important countries.

The Argentina Bailout

The Argentina bailout package also included extremely limited PSI components. It is true that the program was underfunded relative to the fiscal and external financing needs of the country; thus, the program implied that the country had to rely on commitments from domestic financial institutions to roll over the domestic debt of the country as well as some return to limited international market access some time in 2001. But these were the most mild, light and voluntary forms of PSI that one could think of.

For one thing, if PSI is defined as involvement of international investors in crisis resolution, there was little PSI as the program size allowed almost a complete roll-off of external debt coming due in 2001 thus allowing the exit of external investors unwilling to rollover the external debt coming to maturity in 2001.

It is true that some of the domestic financial institutions committed to rollover the domestic debt were foreign owned, thus implying some form of PSI of foreign investors. But the nature of this rollover “commitment” was as voluntary as possible: banks and pension funds committed to purchase domestic debt but the price/yield at which they would buy it was not decided in advance, leaving open the possibility that a spike in rate would lead to a burdensome rollover. Moreover, the legal nature of this commitment was effectively non-existent with only heavy-handed - but eventually not long lasting - “moral suasion” by the government inducing such institutions to do the rollover.

Finally, the June 2001 mega swap, that had some elements of PSI of domestic and international investors, was done all on purely voluntary terms leading to very high spreads that worsened the debt sustainability of the country. Thus, this was not meaningful PSI as there was not NPV haircut. Note that based on standard measures of debt sustainability, Argentina was in much worse conditions than Ukraine or Pakistan. While in Ukraine and Pakistan, a debt restructuring at below market rates (10% rather 20%) was forced on investors, in Argentina the mega swap occurred at market rates that were clearly unsustainable.

Given that the country was probably borderline insolvent, had had three years of negative GDP growth, an overvalued currency, an unsustainable peg and very high real interest rates, a solution that relied mostly on the catalytic effects of official money and policy adjustment was likely to fail. Concerns about the balance sheet effects of devaluation and confidence effects of semi-involuntary restructuring of debt were relevant, as well as the risks of international contagion. But, as in previous episodes, the official sector tendency was to front-load official support, rely on catalytic restoration of market confidence and access and minimize the use of meaningful PSI. But the results of this approach may end up being worse than an earlier and more concerted form of PSI if the country were to end up defaulting on its external debt.

While one could argue that both in Turkey and Argentina the catalytic approach had to be given a chance before more coercive forms of PSI were to be attempted given the risks of causing a worse loss of confidence and triggering massive international contagion, one can otherwise argue that failure to provide more meaningful and concerted PSI early on made things eventually worse. Take the case of Argentina: faced with a political inability to solve its fiscal and structural problems and facing a large debt burden, the country is now teetering on the verge of a default that may imply massive losses to investors as well as significant disruption of economic activity in the country. An earlier semi-coercive restructuring of a larger fraction of the domestic and external debt (say on Ukraine or Pakistan terms with no principal haircut but significantly below market interest rates and significant maturity lengthening) could have created the conditions for a better debt dynamics and justified the political economy costs of painful fiscal adjustment. The same could be said of Turkey: a more concerted form of PSI with a strictly monitored commitment to maintain interbank lines and a semi-involuntary restructuring of the public debt may have contributed to reduce the bleeding of forex reserves and the unsustainable debt dynamics observed after the devaluation of the currency. Thus, both in Turkey and Argentina, the decision to avoid concerted forms of PSI probably ended up being a mistake.

How to address the open controversial issues in PSI: some suggested ideas

Standstills.

Given the problems with standstills that were discussed above, supporters of these solutions need to address more carefully the risks of a systematic use of standstills. While standstills may become necessary in some extreme cases (one can even interpret some concerted PSI solutions such as the interbank rollover in Korea as being conceptually close to a standstill) and they have been officially sanctioned as a tool of last resort (“in certain extreme cases, a temporary payments suspension or standstill may be unavoidable”, IMF Communiqué, September 2000), their use should be infrequent and not linked via a mechanical rule to the provision of official finance. Otherwise, the risk of a rush to the exits would be serious. But while rigid rules specifying ex-ante when standstills should occur may end up being destabilizing, their ex-post discretionary use may be appropriate at times.

Degree of coercion in PSI.

The G7 PSI doctrine has stressed the importance of voluntary, rather than coercive, solutions to crises whenever these are feasible and the importance of good ongoing relations between debtors and creditors. Some go as far as saying that there should never be coercion in the approach to PSI but this option is not realistic. While in some cases voluntary approaches and catalytic official financing will be successful in restoring investors’ confidence and market access, in many other cases (both in small and large/systemic countries) this will be not enough. Thus, the idea that there should be no coercion in PSI – both in systemic and non-systemic cases – is not realistic. Experience shows that market access may not be restored especially when a country with significant problems, policy uncertainty and lumpy external debt payments gets in trouble. Thus, more concerted forms of PSI that imply effectively some degree of coercion will become necessary. Hoping otherwise is not realistic.

High/Low Access for large systemic countries.

Concerns about systemic effects and the geo-strategic importance of large countries have led to a bias in favor of large access/catalytic solutions in these cases. But moral hazard considerations as well as the risk that the chosen approach will eventually fail, put the Fund in defensive lending mode and create political backlashes against large bailout should moderate the tendency of considering all large systemic countries as liquidity problems. Concerted and coercive forms of PSI may become necessary and the international contagion effects of letting a large country go may be better addressed with appropriate ring-fencing of “well behaved” neighbors subject to contagion (CCL for Brazil?).

On the other hand, rigid mechanical rules automatically linking large access (when deemed necessary) to concerted forms of PSI should also be avoided as they may trigger the exact destabilizing reaction (a rush to the exits) that one would like to avoid with large access. Usually, large access should be limited to those truly exceptional cases when there is a good chance that the catalytic approach will work in which case softer forms of PSI are warranted. A tough assessment should be made ex-ante of the chances that the catalytic approach will work and large access should be truly and appropriately justified.

When large access cannot be justified, the provision of limited finance in large systemic cases would obviously risk to trigger a run on the currency and the debt of the country. Thus, the official sector should be involved in facilitating the orderly workout that becomes necessary in these cases.

Note that there are currently biases in the behavior of principals and agents towards treating large systemic cases as liquidity problems requiring large access: the principals - G-7 shareholders - have political, geo-strategic, military and financial interest in bailing out large friends, allies and systemically important countries. The agent – the IMF – has a similar bias towards treating large cases as liquidity cases as it is at times “captive” to the client country while the risks of being blamed for failure of small sized programs leads it to prefer the overwhelming use of large amounts of official money. Given these biases, a framework where the conditions for large access are spelled in more detail may be useful in limiting the use of excessive discretion in providing large access in systemic cases. This need to limit discretion has, however, to weighed against the risk that excessively rigid rules (like automatic linking of large access to concerted PSI) would have destabilizing effects. Some degree of “constructive ambiguity” can and should be designed to limit large access while allowing its use when appropriate. The current G-7 consensus, however, still remains too fuzzy, discretionary and open on the issue of when large access should be provided. A more constrained discretion is needed.

Note also that, in some situations (say Korea for example), large access may be warranted given broader systemic concerns but the IMF may make the determination that, even in the presence of large official money, market access is not likely to be restored (as in December 1997 when interbank lines were being rolled off in Korea). Thus, there will be cases where large access will and should be associated with concerted forms of PSI. In other terms, large access should not be restricted only to cases where the catalytic approach is likely to succeed. But if determination is made that a large access that does not fill entirely the financing gap is not likely to restore market access, concerted solutions a la Korea may become necessary and should be considered in advance. Going for catalytic PSI and then switching to concerted PSI if the former fails may not be the appropriate solution when it is clear that catalytic official money will not work. A clearer spelling of when large access should come with catalytic PSI and when it should come with appropriate forms of concerted PSI should be attempted while leaving some degree of flexibility to address each case as appropriate. While automatic rules linking large access to concerted PSI should be avoided, the current bias in favor of catalytic large access solutions should be restricted by clarifying in detail the assumptions that lead the IMF to believe that such catalytic solutions are likely to succeed.

Active, concerted involvement of the official sector in PSI solutions versus ex-ante determination of official access and allowing laissez-faire market solutions to the debt workout process.

As argued above, a laissez-faire approach where the official sector decides how many resources to provide and let debtors and creditors work out the remaining gap may not be appropriate especially in large systemic cases where provision of low access will trigger a run. If determination was made that catalytic large access will not allow resumption of market access and restoration of confidence and thus limited access is appropriate (or large access is appropriate but not in amounts filling the entire financing gap), one should also prepared for the necessary and difficult debt workout process that will ensue. As the Korea case shows, an involvement of the official sector in concerted forms of PSI may become necessary to resolve collective action problems and allow orderly workouts. Similarly, in cases where bonded debt restructuring becomes necessary, the official sector has an important role to play for a number of reasons and laissez-faire solutions are not appropriate.

First, restructuring deals should be consistent with medium term debt sustainability; failure to ensure that would jeopardize the program and official resources.

Second, since official support is always at stake and since programs often requires restructuring of bilateral Paris Club claims, the official sector cannot ignore the process, terms and outcomes of a private workout process.

Third, collective action problems are prevalent both in bonded debt and bank rollover cases; the official sector may contribute to solve such problems in constructive ways.

Fourth, the official sector has to decide when lending into arrears is appropriate or not; this amounts to an effective determination that a formal or informal standstill or debt payment suspension is appropriate or not. Thus, the official sector cannot just pretend that its role is to determine the amount of official finance and then let the debtor and creditor work out their claims.

The above analysis thus suggest that IMF programs should be clearer about the PSI conditions of these programs and the consequences of failing to implement adequate PSI. The monitoring of PSI in IMF programs should be more systematic and the consistency of programs’ assumptions about private financing with medium term debt sustainability should be appropriately fleshed out.

Risks of litigation and how to address them.

Defaults, debt payments suspensions, concerted debt restructurings are all subject to the risk of litigation. Since there is no political willingness among the G-7 to consider amending the IMF’s Article VIII.2.b to provide to the IMF the power to impose a legally sanctioned suspension of payments, litigation risk will remain and IMF’s lending into arrears policy, while useful, does not rule out the possibility of such litigation. It is also unlikely, based on case law that a court would enforce a stay of litigation even when the debtor has suspended payment but is cooperating with the IMF and its creditors to work out its debts.

But so far, litigation risk has been limited in actual PSI cases; even in Ecuador there was acceleration of claims but no litigation. The reasons for this lack of litigation so far are several: 1. CACs were used ex-post to bind-in holdouts & cram down new terms (Ukraine); 2. generous deal terms and sweeteners reduced holdout problems in bonded debt restructuring cases; 3. exit consent clauses used to cram down new terms even in the absence of formal CACs (Ecuador).

But litigation risks may be higher in the future for three reasons:

• The Elliott-Peru case showed that an aggressive creditor may induce a debtor to pay. However, the legal issues of this case are not settled as the debtor eventually preferred to pay rather than pursue a court challenge

• A recent legal case where a creditor sued the Republic of Congo and was able to find a US court imposing an effective “sharing clause” on the payments made by the sovereign to other creditors creates another precedent where litigation may be successful

• If Argentina were to suspend payments on its external debt, it is likely that several investors will try to pursue their claims in court with the risk that an orderly workout of Argentina’s debt may be disrupted by litigation.

There is thus room for the official sector to take a stance on these difficult issues. While it may not be appropriate to limit the legal rights of private claimants to pursue in court their legitimate claims, the official sector will have to become more alert of the fact that, in the future, litigation may become more likely, thus jeopardizing the chances of orderly debt workouts.

Conclusions on large liquidity cases and PSI in IMF programs

Liquidity cases in large systemic countries are always difficult and complex

Ideally, a combination of policy adjustment, large but not exceptional financing in most cases and appropriate forms of PSI should restore confidence and market access

Large catalytic official money may be better when the country is closer to the pure liquidity case and large access highly likely to restore confidence and market access but these cases may be the exception rather than the rule.

Smaller official money packages and concerted PSI may be better when macro problems are more severe and prospects of restoration of investors’ confidence and market access are not high. In some cases, large access may be warranted even if restoration of market access is unlikely. But in these cases, large access should be ex-ante associated with concerted PSI.

Some degree of “constructive ambiguity” will be necessary to avoid moral hazard and “too big to fail” distortions

It is very hard to have mechanical rigid rules in these complex cases

Lack of rules may lead to “destructive ambiguity” but rigid rules (“always PSI when exceptional money is provided”) may be even more destabilizing causing “rush to exits”

But given the political biases to provide large access to official resources in large systemic cases, the conditions under which such exceptional access will be provided and whether PSI should be catalytic or concerted should be spelled out more clearly than in current doctrine

IMF programs should be clearer about the PSI conditions of these programs and the consequences of failing to implement adequate PSI. The monitoring of PSI in IMF programs should be more systematic and the consistency of programs’ assumptions about private financing with medium term debt sustainability should be appropriately fleshed out.

-----------------------

[1] This section is partly based on an excellent summary of the debate by Brad Setser (“A brief history of the PSI debate inside the official sector”, April 2001).

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download