Preliminary draft
Do we need a new international bankruptcy regime?1
Comments on Bulow, Sachs and White
by
Nouriel Roubini
Stern School of Business
New York University, NBER and CEPR
April 4, 2002
1
Discussion of papers by Jeremy Bulow, Jeffrey Sachs and Michelle White for the Brookings Panel on
Economic Activity “Symposium: A Bankruptcy Court for Sovereign Debt”, April 5, 2002, Washington
D.C. The usual disclaimer applies.
Recently, the debate on the reform of the international financial architecture has
centered on the issue of the appropriate mechanism or regime to be developed to ensure
orderly sovereign debt restructurings. While recent sovereign bonded debt restructuring
cases (Pakistan, Ecuador, Ukraine and Russia) have been successfully completed with the
use of unilateral exchange offers (at times complemented by a system of carrots and
sticks such as exit consents to ensure successful deals), many have expressed
dissatisfaction with this “market-based” status quo approach. Also, the recent default by
Argentina suggests that we need to reconsider the issue of optimal debt restructuring
regimes. And recently, Anne Krueger, the First Deputy Managing Director of the IMF,
has proposed the creation of an international debt restructuring mechanism (SDRM) that
would have many of the features of an international bankruptcy regime2. The three
papers by Jeremy Bulow, Jeffrey Sachs and Michelle White are all interesting
contributions to this debate on the appropriate sovereign debt restructuring regime.3 They
all address the question of whether we need an institutional change in the international
financial system that would lead to a new way to provide for orderly sovereign debt
restructurings/workouts when they become necessary.
So, the difficult policy question to address is: when sovereign debt
restructuring/reprofiling/reduction becomes necessary and unavoidable, what is the
appropriate regime that provides an orderly restructuring while safeguarding the balance
of rights of both the creditors and the debtor?4 Is it better to continue with the “marketbased” status quo regime where exchange offers have been used to do bonded debt
restructurings? Or should we move to the wholesale introduction and use of collective
action clauses (a “contractual approach”)? Or should we consider creating an
international bankruptcy mechanism (as “statutory approach”) such as the one proposed
by the IMF.5
The issue in this debate is which restructuring regime is most efficient in cases in
which the country’s debt path is deemed to be clearly unsustainable (cases of
“insolvency”). Subject to the caveats that the concept of insolvency is problematic in the
sovereign context (as “inability to pay” may be combined with “unwillingness to pay”)
and that the assessment of unsustainability is always probabilistic (as a sharp primary
adjustment could in principle make an unsustainable debt path sustainable), there is a
general consensus that in cases of “insolvency” further official finance is not warranted
2
Krueger (2001a, 2001b, 2002)
Sachs (1995) was an early advocate of an international bankruptcy court for sovereign debtors while his
current contribution concentrates on the debt crisis and the debt reduction needs of low income countries.
In my remarks here I will concentrate on the Sachs (1995) arguments in favor of an international
bankruptcy regime and the related issues in his most recent paper. See Rogoff and Zettelmeyer (2002) for a
survey of the literature on sovereign bankruptcy ideas.
4
When debt becomes unsustainable and the country has to restructure its sovereign (and possibly private
sector) external liabilities, it is in principle in the interest of all parties to have an orderly debt restructuring
process, one that can minimize losses of value that are socially inefficient and allow the country to adjust
and return to a sustainable debt path. The need for an orderly restructuring derives from the observation
that a “disorderly” default can impose losses that are socially inefficient and thus can hurt both the debtor
and the creditors. Thus, subject to the caveat that defaults should not be too easy (to prevent opportunistic
defaults), an orderly debt restructuring should be the objective of an international regime that allows
countries with unsustainable debt profiles to restructure their liabilities.
3
5
I discuss these issues in more detail in Roubini (2002).
and that the sovereign should suspend debt payments and restructure/reduce its debts,
while at the same time undertaking serious and credible domestic fiscal adjustment and
structural economic reform.6 The debate is not on whether the sovereign should
restructure its debt in these unsustainable situations. It is instead on the appropriate
regime for orderly sovereign debt restructurings in these unsustainable debt cases. There
are essentially three alternative options/regimes that one may want to implement and use.
First, continue with the current “market-driven” status quo regime where
sovereign bonded debt restructurings (Pakistan, Ecuador, Ukraine and Russia) have been
successfully completed with the use of unilateral exchange offers (at times complemented
by the use of exit consents),
Second, move to a “contractual approach” regime where collective action clauses
(CACs) are introduced in most bond (and possible other debt) contracts and used to
achieve debt restructurings.7
Third, design a new “statutory regime” where an international bankruptcy regime
for sovereigns is created and used to achieve sovereign debt restructurings. The latter
regime has been recently re-proposed by the Anne Krueger, the Deputy Managing
Director of the IMF.
So, when sovereign debt restructuring/reprofiling/reduction becomes unavoidable,
what is the most efficient regime that provides orderly restructuring while safeguarding
the balance of rights of both creditors and the debtor?
Each of these three approaches to sovereign debt restructuring has some pros and
cons. One way to think about the relative merits of these three regimes is to first ask what
are the market failures that may prevent and orderly and efficient restructuring of
sovereign debt when such orderly restructuring is beneficial to both debtors and creditors.
One can think of several externalities that prevent orderly restructurings but there are
three of them that are crucial and they all have to do with collective action problems
among creditors8 9:
6
There is an open issue of whether an international debt workout regime should also be used to address
financial crises that have an “illiquidity crisis” nature and other cases where debt restructuring may be
necessary but the country is not obviously insolvent. The IMF thinks of applying the SDRM to insolvency
cases only but, as discussed below, an SDRM could be used to address collective action problems and other
obstacles to orderly workouts in a broader range of crises, including liquidity cases. This application of a
debt restructuring regime also to liquidity cases appears to be the view of Sachs (1995).
7
This “contractual approach” based on introduction and use of CACs has been long supported by a number
of academics (see Eichengreen (1999) and Eichengreen and Portes (1995)). It has recently also received
the support of the U.S. administration (see John Taylor (2002a, b). Support for the progressive introduction
of CACs in bond contracts can also be found in other official reports, like the Rey Report issued after the
Mexican Peso crisis and in several past G7 communiques on how to reform the international financial
architecture.
8
See Sachs (1995) for an early statement of these market failures as the basis for the need for an
international bankruptcy regime.
9
In Roubini (2002) I discuss of a number of other potential market failures in addition to the four discussed
in my remarks here. Specifically, I consider the “rush to the exits” and the “rush to the courthouse” on nonsovereign claims (and the ensuing need for capital and exchanger controls); the risk of debtor actions (such
as preferential treatment of some creditors) that damage creditor interests; the risk of asset stripping by the
debtor; how to provide senior private “new money” (debtor in possession - DIP - financing) during a
default. Sachs (2002) states that, in addition to the collective action problems among creditors, another
motivation of bankruptcy law is to provide a “fresh start” to insolvent debtor, i.e. avoid situations of a “debt
overhang” and provide some fairness, in addition to efficiency, to the workout process. I interpret this
1. The “rush to the exits”: this is the collective action problem of creditors in the run
up to a crisis. As a sovereign debt crisis is unfolding, creditors may try to rush to
the exits and cause a disorderly crisis that has real and avoidable costs; an
example is the one of liquidity or rollover runs. But, as discussed below, a debt
suspension/standstill (including capital/exchange controls and/or deposit freezes)
may avoid such a destructive rush to the exits.
2. The “rush to the courthouse” externality. While a unilateral debt standstill may
take care of the inefficiencies of a “rush to the exits”, such standstill may cause a
“rush to the courthouse”. Creditors may start litigation and this potential
externality can become a serious problem if creditors can attach assets, i.e. have a
first claim on the assets of the debtor if they move first (“rush to the courthouse”
or “grab race”). But as discussed below, there are important differences between
the corporate paradigm and the sovereign one on this matter as the ability of
creditors to seize/attach sovereign assets is very limited.
3. The “free rider” or “holdout” or “rogue creditors” problem (another collective
action problem among creditors). This is an important obstacle to orderly
restructuring. In situations where unanimity may be required to restructure debt,
minority holdout creditors may scuttle a restructuring that is advantageous to the
majority of creditors. While the unanimity problem may be sidestepped with
exchange offers, the holdout problem may potentially remain serious. If an
holdout does not accept the offer and then receives (via post-deal litigation or its
threat) the full amount of his/her claims while those who accept the offer receive a
lower amount than their full claim, there is a strong incentive to hold out (“free
riding”) with the consequence that an otherwise mutually advantageous deal may
fail because many/most creditors will decide to hold out. If this coordination
problem among creditors cannot be solved, a disorderly and costly workout will
be the outcome even if it would have been in the interest of all creditors to
achieve a cooperative solution (i.e. a non-cooperative Nash equilibrium with
holdouts is inferior to the cooperative/coordinated one). In this regard, the ability
to have a restructuring plan approved by a majority of creditors made binding on
the entire creditor body (a “cram-down” or majority enforcement provision)
would solve this holdout externality.
In addition to these three collective action problems among creditors, any efficient
mechanism has to deal with a fourth potential market failure on the side of the debtor:
4. The “rush to default” or the debtor’s incentive to do “opportunistic defaults”. As
the literature on sovereign debt suggests, a default decision may be due not to
“inability to pay” but to “unwillingness to pay”. There is always the possibility of
opportunistic defaults given that a sovereign benefits from (not full but
“fresh start” as the need to provide debt workouts that are beneficial to both debtors and creditors when a
disorderly and costly and lengthy workout would lead to a loss of value that is not beneficial to either side.
Thus, this issue is which regime (statutory, contractual, market-based) can provide such orderly workout.
significant) sovereign immunity and thus attaching/seizing sovereign assets is
difficult. Thus, an efficient international debt workout mechanism needs to
tradeoff two objectives: not to make workouts too costly as default may at times
be due to inability to pay and restructuring can thus benefit both the debtor and its
creditors; and not to make workouts too easy as otherwise the temptation to have
opportunistic defaults may increase.
I will first analyze how the three regimes being considered address the three
collective action problems of creditors. I will at the end of my remarks address the
question of the “rush to default”; this issue is essential to the views of Bulow on the
desirability of an international bankruptcy regime and on his proposal for granting full
immunity to sovereigns.
Let us consider first the pros and cons of each regime in terms of the collective action
problems of creditors.
Supporters of a new statutory regime10, an international bankruptcy mechanism, stress
the fact that, while the above collective action problems always existed, they have
become more severe in the last few years given development in international financial
market.
First, in the 1980s most of sovereign debt was held in the form of medium-long term
syndicated bank loans; the covenant of these loans included sharing clauses and other
limits to initiation of litigation that made the “rush to the courthouse” less serious. They
also had implicit or explicit majority clauses that allowed to deal with holdout banks; also
moral suasion, deriving from repeated interaction among banks was more likely to rein in
holdouts.
Second, in the 1990s most of the flows to emerging markets sovereign have taken the
form of bonds. The number, heterogeneity and differences of interest of this wider group
of creditor makes the holdout problem much more severe.
Third, the emergence of new bond creditors with no ongoing relations with the debtor
or other creditors, suggests that the presence of aggressive holdouts, i.e. “vulture”
creditors, who are willing to hold out and aggressively pursue their claims in court may
have increased. Indeed, the recent Peru-Elliott case is seen as major threat to orderly debt
restructuring as the creditor successfully pursued a litigation strategy and ended up being
paid in full.
In summary, the variety of claims (bank loans of various maturity, different types of
bonds under different legal jurisdiction and with or without collective action clauses) and
types of creditors (retail investors, investment and commercial banks, real money fund,
hedge funds and other highly leveraged aggressive creditors, dedicated emerging market
funds and cross-over investors) makes the collective action problem of coordinating the
interests and action of such an heterogeneous world of claims and claimants almost
impossible: investors may rush to the exits in a destructive panic; they may rush to the
courthouse and start litigation if the debtor suspends payments and they may free ride
/holdout even if a majority of creditors could reach an agreement advantageous to all. If
this view is correct a new international bankruptcy mechanism could allow an orderly
restructuring. The main advantage of such a mechanism is that it would solve the three
collective action problems by:
10
See Krueger (2001a,b, 2002).
1. allowing the imposition of a suspension of debt payments that stops the “rush to
the exits”;
2. imposing a “stay of litigation” following the debt suspension that is legally
binding on all creditors and thus prevents disruptive litigation (the grab race);
3. allowing for a majority vote on a restructuring agreement that is binding on all
creditors, thus eliminating the “free riding” or “rogue creditor” problem.
Supporters of the second option11, the “contractual approach (the universal
introduction and use of collective action clauses in bond and debt contracts) would argue
that most of the benefits of the “statutory approach” could be obtained with the use of
collective action clauses (CACs). Such clauses usually do not allow individual
bondholders to start litigation (litigation has to be voted by a majority of creditors) and/or
include sharing clauses that reduce the benefits of being an holdout and litigate. Also,
CACs would include majority "cram-down clauses" so that an agreement reached by a
majority of creditors can be binding on all holdouts, thus solving the free rider problem.
Thus, all the collective action problems that prevent an orderly restructuring could be
solved with the use of CACs. And relative to an international bankruptcy regime that
could potentially give new judicial powers to the IMF and/or to some “bankruptcy court”,
the contractual solution could be more market friendly and rely on an agreement to be
reached between the sovereign debtor and its creditors.
However, it is important to note that the sovereign debt restructuring regime proposed
by the IMF (at least its last incarnation as in Krueger (2002)) would not be substantially
different from a contractual approach as it would be “creditor-centered” rather than being
“IMF-centered”: specifically this latest SDRM proposal would provide to creditors all the
rights related to approving by a majority vote an initial stay of litigation (and its
continuation) and approving by majority a restructuring deal that would be binding on
minority holdout creditors. Thus, this “IMF-lite” SDRM would not be substantially
different from a beefed-up contractual approach in terms of being creditors, rather than
“court”, driven.
Moreover, supporters of the “statutory” approach would counter-argue that the
statutory solution is superior to a contractual regime for several reasons.
First, there is a transitional problem as many outstanding bonds, those issued
under New York law, do not have CACs. So, even if new bonds included them, the past
stock of outstanding bonds would not have them.
Second, under traditional CACs the vote to start litigation or cram down by
majority an agreement is taken bond-by-bond rather than by a majority of all creditors in
the asset class (all bondholders). So, holdout problems and litigation problems may
reemerge if a majority of bondholders in one issue decides not to cooperate. While one
may think of some super-clauses that would imply a super-majority vote by all creditors
in a particular credit class (all bonds), these clauses do not exist so far and are not likely
to be introduced in a uniform way any time soon.
Third, while collective action clauses could be eventually included in all bond
covenants, many other claims on the sovereign (banks loans, various other credit
instruments) may or may not have them. Also, over time financial innovation may lead to
11
See Eichengreen (1999), Taylor (2002a, b).
the creation of new financial instruments, such as various credit derivatives, that may not
include such clauses. The statutory approach has the advantage that, regardless of what
current and other future claims on the sovereign are, they would all be included in the
same restructuring mechanism and be subject to the same overall majority vote to initiate
or withhold litigation and vote to approve a restructuring agreement.
Fourth, achieving uniformity of CACs (their wording and interpretation) issued in
very different legal jurisdictions may be very hard to achieve. Messy, costly and
protracted legal issues of interpretation and adjudication may result. While an uniform
international bankruptcy regime would codify a standard set of rules, case law and
interpretations.
While some of these difficulties could be surmounted in a contractual approach
via the use of superclauses, arbitration and other meta-clauses, such a beefed-up
contractual approach ends up becoming very close to a creditor-centered statutory one.12
Supporters of the status quo regime13 start from the observation that, while ideally
a “statutory approach” or a “contractual approach” would solve these collective action
problems and thus be welcome institutional developments, they are both unlikely to
emerge for a complex set of political-economy issues. The U.S. will not agree to have an
international legal regime overrule US security laws and its protection of creditor rights;
many emerging markets may resist the bankruptcy regime based on a concern that it
would make it easier for the IMF to cutoff lending to crisis countries, thus “dumping”
them into the bankruptcy court; and the other G7 (while being in principle more
sympathetic to the idea of an international bankruptcy regime)14 will not aggressively
push for it. A “contractual approach” is also unlikely to see light as, while rhetorically
supported by the G7/G10 since the times of the Rey Report in 1996, there is no system of
carrots and sticks to ensure that both creditors and debtors include them in new bond
issues. Thus, progress on adoption of CACs has been dismal so far.15 Thus, if neither the
statutory nor the contractual approach are likely to see light in the foreseeable future, one
has to try to make the most of the status quo regime to achieve orderly restructuring.
In this regard, recent experience suggest that bonded debt restructurings are
feasible and have been successfully achieved, even in the presence of hundreds of
thousands of heterogeneous creditors, with the use of unilateral exchange offers (cum exit
consents when available) in several episodes: Pakistan, Ukraine, Ecuador, Russia.
Moreover, the collective action problems emphasized by many may be
exaggerated in reality. First, any sovereign faced with a “rush to the exits” can stop it
with a unilateral debt suspension; thus, this collective action problem has already a
solution available in the current status quo: a unilateral debt suspension/standstill/default.
The main difference remains that a statutory approach requires an amendment of the IMF’s Articles of
Agreement or an international treaty while the contractual approach could evolve over time without such a
radical institutional change. But even a contractual approach would require changes in legislation in some
major legal jurisdictions, thus taking some institutional/statutory features rather than being purely marketdriven.
13
Roubini (2002).
14
See the very cautious endorsement of the idea of an international bankruptcy regime in the latest
(February 2002) communiqué of the G7 Finance Ministers.
15
Even the latest remarks by Taylor (2002) in favor of the contractual approach appear to be so far
toothless in terms of the system of incentives (carrots and sticks) that may lead to real progress towards the
wholesale adoption of CACs.
12
It is true that this debt suspension, in the absence of a stay of litigation, may lead to a
“rush to the courthouse”. But the collective action problem of a “rush to the courthouse”
is not severe in the case of sovereign debtors. This market failure (also referred to as the
“grab race”) is certainly important and severe in the corporate bankruptcy context where
rushing to litigate may allow a creditor to attach assets. Thus, bankruptcy regimes such as
Chapter 11 or 7 prevent such a grab race via a stay of litigation once the debtor has
applied for the bankruptcy protection. And the stay is mostly about protecting creditors
rights (i.e. avoid the unfairness of some creditors attaching assets to the disadvantage of
other creditors) in a corporate context.
In the sovereign context, the rush to the courthouse is much less of a problem as
sovereign immunity implies that creditors have trouble finding assets worth rushing to
claim. The ability to attach asset via early litigation is severely limited. In fact, there is a
scarcity of assets under the jurisdictions of foreign courts that can be potentially available
for creditors to seize/attach. And, indeed, there is little evidence of a rush to litigate in
sovereign debt crisis when a country suspends debt payments. For example, in the recent
case of Argentina, creditors threatened litigation but they have not so far started it.
Then, if the rush to the exits and the rush to the courthouse are not real issues in the
current status quo, one is left with the “holdout /free rider” problem as the main collective
action problem that may not be as easily solved in the absence of majority cram-down
clauses. But, in the reality of the current status quo regime, even the free rider problem
(and the related litigation threat) has not been as severe as initially thought. There are
plenty of sensible ways to overcome and minimize the rogue creditor problem even
without majority cram-down clauses. Here are ten reasons why the holdout problem is
not a big problem in practice.
First, the unanimity problem (as when bond contracts do not have majority cramdown clauses) can be bypassed with the use of unilateral exchange offers. While these
offers do not eliminate the holdout problem, they allow for a great majority of
cooperative bondholders to accept new bonds with new payment features even when the
old bond required unanimity to change their terms. And indeed, in cases where there were
thousands of bondholders (Ukraine, Pakistan, Ecuador, Russia) such unilateral exchange
offers have had overwhelming success with 99% plus of creditors accepting the offer.
Second, as it is well known “exit consents”, to change by majority vote the nonfinancial terms of the bond covenant, have been successfully used (see Ecuador) to dilute
the benefits of being a holdout.
Third, a system of carrots (sweeteners in the form of cash, collateral release,
seniority upgrade) and sticks (threat of default, ex-post use of CACs, exit consents) has
been used and can be used to ensure a successful completion of deals.
Fourth, the “free rider” or “holdout” problem is predicated on the assumption that, in
a debt restructuring, a creditor that holds out would receive a financial benefit that is
greater than the one he/she would receive by participating in an exchange offer. But this
assumption is flawed in a number of ways that significantly reduce the risk of having
many rogue creditors and the litigation risk. Note that the simple assumption that a
holdout would automatically receive the full value of his/her claims rather than the value
offered in an exchange is false. Moreover, in any market based exchange offer, any markto-market investor should accept the offer (or be at the margin be indifferent) – rather
than holdout - as long as the value of the new claims is at least as large as the value of the
market value of the old claim. Since no creditor would want to accept an exchange offer
where there is not at least mark-to-market neutrality, a successful exchange offer must
provide a new claim whose market value is at least as high as that of the exchanged
instruments. And indeed in all previous debt exchanges (Pakistan, Ecuador, Ukraine,
Russia) creditors have enjoyed mark-to-market gains (20-30% on average); such gains
increased the likelihood that the offer will be accepted by a majority of creditors.
Fifth, a creditor may decide to hold out, even if the offer has neutral value (mark-tomarket neutrality) or a mark-to-market gain if the risk-adjusted expected discounted net
value of its original claim, if he/she holds out, is greater than that of the new claim. But
all the underlined elements are fundamental in determining whether a creditor will hold
out and litigate. Litigation is costly (especially for small creditors); some creditors (the
small retail ones) are more risk averse than others and the outcome of litigation is
uncertain; some have a high rate of time preference and may not want to wait for the
delay costs of protracted litigation. Thus, a majority of creditors are likely to rationally
accept an offer that is mark-to-market neutral or slightly positive, rather than holdout and
incur the costs and risks of litigation.
Sixth, large financial institutions that have ongoing business relations with a
sovereign debtor (through the franchise value of their commercial banking operations in
the debtor country and/or the fees/commissions from their investment banking services to
the debtor) are unlikely to hold out and fight. They may actually be the leader/catalytic
agents that would contribute to coordinate the actions of the many creditors, apply moral
suasion on holdout and, if necessary, bribe them into accepting a deal. The desire to gain
the large fees/commissions involved in a successful deal leads the intermediaries to
design workout packages that minimize such “deal risk”.
Seventh, the holdout problem can be minimized through side payments (“bribes”)
offered by creditors who have a lot to gain from a successful deal; or by the debtor (that
“ex-post” buys out a limited number of holdouts); or official creditors (via extra amounts
of official finance that provide enhancements and/or sweeteners to a deal).
Eighth, the Elliott-Peru decision was, from a legal standpoint, highly controversial
and unusual and, most likely, its logic would not stand if challenged in other legal cases.
A legal doctrine that interprets “pari passu” allowing to block payments to creditors that
have accepted an exchange offer is highly controversial and very likely to be successfully
challenged in court.
Ninth, creative variants of the status quo regime of exchange offers can be designed
to provide market based orderly restructurings that reduce the risks of litigation and/or
free riding.16
Tenth, rogue creditors and vulture are often part of the solution rather than part of the
problem. Low risk aversion vultures tend to buy low, when default has occurred and debt
prices have collapses and get large mark to market gains from a successful deal; thus,
they may accept an exchange offer rather than litigate. For example, the infamous Elliott
who successfully sued Peru was holding Ecuador debt and decided, with 99% plus of
creditors, to accept the exchange offer rather than holdout as the offer provided
significant mark to market gains. Morover, even “rogue creditors” who will eventually
sue will not jeopardize the completion of an exchange offer: their incentive to start
litigation are triggered by a successful, not a failed offer; only after a majority of creditors
16
Se for example the recent JPMorgan proposal by Bartholomew and Stern (2002).
accepted a deal, a rogue will have the incentive to litigate and attempt to obtain his/her
full claim.
Thus, while one cannot fully solve the free rider problem in the absence of a majority
cram-down clause (that is available in the contractual and statutory regimes), there are
creative ways to minimize its risks and consequences in the current market based status
quo. And, indeed, recent experience has shown that holdout problems have not prevented
the successful achievement of orderly bonded debt restructurings. In most cases, the
status quo may still work and allow successful exchange offers where the holdout
problem becomes only a post-deal nuisance.
This view in favor of the market-based status quo does not mean that a contractual or
statutory regime would not be beneficial if implemented; the latter regimes would
provide cleaner and potentially more efficient ways to solve the free rider problem. But,
as long as these regimes are not in place and unlikely to be in place in the foreseeable
future, my argument is that one can creatively use the market-driven status quo to achieve
orderly restructurings.
“The Rush to Default” problem
I will finally consider the “rush to default” problem on the part of the debtor as this is
a main concern of those, like Bulow, who worry about debtor moral hazard. In a world
where countries benefits of sovereign immunity and creditors have very limited ability to
attach/seize sovereign assets, there is always a possibility that a sovereign would
“opportunistically” default, i.e. default may be driven by unwillingness to pay rather than
inability to pay. And indeed, the large literature on sovereign default has studied the
incentives of a sovereign to default and the factors (costs/punishments) that may limit the
temptation to have such opportunistic defaults.
Of course, since sovereign debt has always the potential to lead to “opportunistic
defaults” (unwillingness to pay as opposed to inability to pay), a restructuring that is too
“easy” or “orderly” (i.e. with little or no cost to the debtor) may not be socially efficient.
Indeed, given the pervasiveness of sovereign immunity, the appropriate costs (in terms of
loss access to international capital markets, output and trade losses) that creditors can
impose on the debtor are an important component of a well balanced regime that
minimizes the moral hazard of opportunistic default. But while “default” that is too
“easy” may not be efficient, a “disorderly” default (triggered by an inability to pay) can
impose losses that are socially inefficient and thus can hurt both the debtor and the
creditors. Thus, subject to the caveat that defaults should not be too easy (to prevent
opportunistic defaults), an orderly debt restructuring should be the objective of an
international regime that allows countries with unsustainable debt profiles to restructure
their liabilities.
How would the three restructuring regimes deal with the “rush to default” issue?
Supporters of the status quo regime would argue that the “rush to default” is not a big
issue in the first place. In this view, even in the current regime with limited sovereign
immunity, sovereign have strong incentives not to opportunistically default as such action
has severe reputational costs and punishments in the form of protracted loss of access to
international capital markets, output and trade losses and other punishments, including
litigation and attempts to seize assets, that creditors can impose on the sovereign.17 In this
view, a healthy and balanced regime is similar to the current one where the incentives of
sovereign to opportunistically default are already limited by the consequences and costs
of such default. Thus, making it easier for the debtor to default via a statutory regime that
provides greater legal protection against creditors’ actions in case of default may tip the
balance in favor of debtors and trigger opportunistic default that would ultimately reduce
the ability of emerging markets to access capital markets.
At the other extreme, a well designed SDRM regime would provide legal protection
of the sovereign against creditor litigation but it would have safeguards against the abuse
of this protection by opportunistic debtors. In one variant of the SDRM, access by the
debtor to this SDRM legal protection would be conditional to an IMF assessment that the
country has an unsustainable debt position; without such sustainability test, the country
would not receive such legal protection. In another variant, closer to a majority
enforcement clause, a majority of creditors would take the decision on whether approve
and extend a stay or instead to start litigation. Thus, as in the current regime,
opportunistic defaults could be dealt with the threat of litigation if creditors believe that
the debtor is behaving in “bad faith”.
In the contractual approach, opportunistic defaults would again be addressed by the
threat of litigation on the part of a qualified majority. Unlike the current status quo where
any creditor can start legal action if he/she desires so, in the contractual approach the
decision to start litigation would be made by a majority of creditors (to avoid disruptive
litigation by a small minority of creditors).
An interesting but radical view of debtor “moral hazard” and the merits of a
bankruptcy regime is presented by Jeremy Bulow in his paper. Bulow starts from the
view that debtor moral hazard is pervasive in two ways: first, emerging market policy
makers have a bias towards socially inefficient budget deficits (as the are “corrupt” or
“malevolent” policy makers that will borrow for inefficient reasons and transfer the debt
burden to future policy makers and taxpayers); second, the “unwillingness to pay”
problem (rather than the “inability to pay”) is severe in emerging market economies
where a sovereign (benefiting from partial sovereign immunity) has a strong incentive to
do opportunistic defaults. So, the problem of emerging market sovereigns is that they
borrow too much to begin with and the ability to borrow internationally distorts this
initial bias towards budget deficits and debt accumulation.18
Thus, Bulow believes that the way to reduce or eliminate this bias is,
paradoxically, to provide sovereigns full sovereign immunity, rather than the partial
sovereign immunity coming from debt issuance in major financial centers; sovereign
debtors should be allowed to borrow only in their own legal jurisdictions where sovereign
immunity is close to full. In his view, the partial legal protection of creditor rights in
international jurisdictions exacerbates the debt/deficit bias of reckless emerging market
policy makers. Since their ability to issue debt (purchased by foreign investors) when
there is full sovereign immunity in local jurisdictions would be severely limited, this
17
There is a broad literature on whether reputational mechanisms and which type of costs of default are
able to sustain an equilibrium without opportunistic defaults. See for example Bulow and Rogoff (1989)
and Wright (2001).
18
See Corsetti and Roubini (1999) for a model of how a political bias towards budget deficits is
exacerbated when policy makers can borrow in international capital markets.
reform may altogether shut down or severely restrict the ability of “reckless” sovereign
debtors in to borrow internationally. Only good and responsible sovereign policy makers
following sound policies and devoid of the temptation to default would be able convince
foreign investors to lend them in securities issues in domestic jurisdictions. If, after this
reform, most sovereign debtors are unable to borrow from international investors, this
would be a better world in Bulow’s view.
Then, again paradoxically, in the Bulow view the only potential benefit of an
international bankruptcy court is that, by making it easier for a sovereign to default and
restructure its debts (assuming that the regime is designed in such a way that the relative
power of the debtor is increased), it would severely shrink the amount of international
capital lending to emerging market sovereign debtors, an outcome that he finds to be
socially efficient. So, the international bankruptcy court is good as it would effectively
destroy the market for international issuance of new sovereign emerging market debt.
While granting full sovereign immunity would be his first best policy to achieve this
goal, the international bankruptcy court could be a second best policy that would reduce
the initial excessive borrowing distortion of debtors.
I am not convinced by his arguments for a number of reasons.
First, reputational mechanisms and output/trade costs of default do significantly
restrict the willingness, even of otherwise “malevolent” policy makers, to default.
Governments try to avoid as much, and as long as possible, defaults as they are
politically, socially and economically costly.
Second, the empirical evidence on moral hazard (both debtor’s and creditors’) in
international lending is extremely thin; for example, Jeanne and Zettelmeyer (2001) show
that domestic tax payers, rather than the IMF/IFIs (i.e. the international tax payers) or
creditors, pay for the costs of official support packages. Thus, the idea that emerging
market economies borrow too much and follow reckless economic policies in
expectation of being bailed out by the IMF has little basis. The idea that countries would
willingly follow policies that lead to currency crises, banking and financial crises and
possible default in expectation of bail-out is not supported by evidence. The costs of
crises are severe and crises lead (good and bad, democratic and autocratic) policy makers
to be booted out of power.19
Third, a side implication of the point above (i.e. IMF support is not in reality a
“bail-out” of the debtor as it is an unsubsidized loan, not a grant) is that Bulow’s aversion
towards lending by the IMF/IFIs (the IMF should not make loans because it leads to
gaming between debtors and private creditors to extract resources from the official
sector) does not have a strong empirical basis. Also, there are many other arguments in
favor of IMF loans and conditionality and against the “aid but no loans” view.20 Even the
authoritative views of Jeffrey Sachs on the issue of IMF loans appear to have changed
over time. While in his 1995 paper first he made the argument that the IMF should
become an international lender of last resort (ILOLR) to deal with liquidity crises, he
then argued that, even better than large bailout packages, liquidity runs could be
addressed by turning the IMF into an international bankruptcy court with the power to
A side implication of this observation is that Bulow’s aversion towards lending by the IMF/IFIs (the IMF
should not make loans) as a source of gaming between debtors and private creditors does not have a strong
factual basis
20
See the critique of similar views by the Meltzer Commission in Frankel and Roubini (2001).
19
declare standstills and restructure sovereign debts and thus avoid the destructive effects
of a “rush to the exits”. But his later analysis of the Asian crisis as being driven mostly
by self-fulfilling liquidity runs suggested again that large IMF liquidity packages would
be necessary to deal with such destructive liquidity panics and runs. Thus, his later
support, within the work of the Meltzer Commission, of turning the IMF into a quasi
ILOLR that would lend very large amounts to well-behaved countries that experience
liquidity runs, panics and contagion.21 Also, while some (as Sachs in 1995) may argue
that, even in the case of liquidity runs one could use - as an alternative to large official
support - capital controls, standstills, debt suspensions and debt reprofiling/restructuring
as a way to prevent the effects of such panics, this approach would be seriously
counterproductive and destabilizing in practice. In a world with uncertainty, risk aversion
and imperfect policy credibility expectations of standstill to solve liquidity runs may
trigger an early and destructive “rush to the exits” that would have serious consequences
even if all international financial transactions (including sovereign and private ones) were
subject to a standstill.22 Thus, at least for cases closest to illiquidity runs, there is a broad
intellectual and policy consensus that large IMF loans, rather than standstills and defaults,
may be the way to resolve such crises. This is also way central banks mostly use lender of
last resort liquidity support, rather than bank holidays (deposit freezes), to deal with pure
liquidity runs and panics.
Fourth, restricting the ability of sovereign policy makers to borrow in foreign
jurisdictions may not restrict significantly their ability to borrow and the risk of crises. Of
the financial crises of last decade (Mexico, Korea, Thailand, Indonesia, Russia, Brazil,
Turkey, Argentina) only one (Argentina) had to do with the foreign issued debt of a
sovereign.23 In some crises (Korea, Thailand, Indonesia), private rather than public
liabilities, were the source of the vulnerabilities. Even in crises where public debt was at
the source of the liquidity or currency mismatch that triggered a crisis, most of the
borrowing that was at risk was issued at home (Tesobonos in Mexico, GKOs in Russia,
domestic local and foreign currency short term debt in Turkey and Brazil, domestic debt
of Argentina).24 Even in these sovereign crises, the foreign currency liabilities of a
private financial system were often subject to a run or rollover crisis, on top of the run on
sovereign claims.
Fifth, as long as the ability of the private sector to borrow internationally is not
restricted, restricting the ability of a sovereign to borrow internationally will not affect its
ability and cost of accumulating debt: the sovereign will borrow at home and the private
sector will in turn borrow abroad to indirectly finance the borrowing needs of the
sovereign.25 For example, in many recent crises (Mexico, Russia, Brazil, Turkey,
Argentina) a large fraction of the government debt was issued domestically and
purchased by domestic banks; these, in turn, borrow short term and in foreign currency
from abroad to indirectly finance the government’s budget needs.
21
This is in principle the role of the current CCL facility within the IMF.
See Frankel and Roubini (2001), Roubini (2000, 2002b) for various arguments against standstills as a
tool to prevent runs.
23
Ted Truman has recently stressed this point.
24
See Roubini (2001) for a discussion of how to treat “domestic” and “external” claims in sovereign debt
restructurings.
25
See Corsetti and Roubini (1997) on this conceptual point.
22
Sixth, while it is conceptually obvious that the existence of a third player, the IMF
or official sector providing funds, may lead to a delay game between the debtor and its
creditor aimed at extracting further official resources, the empirical relevance on this
problem is very weak. As long as the subsidy component of IMF loans is small (as shown
by Jeanne and Zettelmeyer (2001)) and as long as IMF loans are senior to private claims
(as they are), such gaming would not be beneficial to the debtor (whose tax payer
eventually repays those loans) nor to the creditors (as further debt senior to private claims
at some point may end up hurting the servicing of private claims). Thus, this specific
critique by Bulow of IMF loans is conceptually correct but policy-wise of minor
relevance.
Seventh, the international bankruptcy court, as designed and proposed by the IMF
(see Krueger (2002)) would not provide any new powers to the IMF or to the debtor
country, relative to the current status quo or a contractual regime with collective action
clauses. Thus, it will not tip the balance in favor of the debtor relative to the current status
quo and thus it will not affect the debtors’ incentives to default and, thus, it would not
affect in principle the amount of flows to emerging market sovereigns. Thus, as proposed
by the IMF, the international bankruptcy regime would not severely restrict (as otherwise
hoped by Bulow) the amount of lending to emerging market governments.
Thus, in conclusion, Bulow’s view of the evils of international sovereign
borrowing and of IMF lending seems extreme and not warranted by the facts. And his
solution to this alleged problem/distortion would not, most likely, solve the distortions
that it is meant to address.
Conclusion
The debate on which one of the alternative debt restructuring regimes is better at
achieving orderly restructurings is still open; all three regimes provide different creative
solutions to the collective action problems inherent in debt restructurings. While the
statutory approach provides the cleanest way to solve in a consistent and coherent way all
the collective action problems involved in an orderly restructuring of sovereign bonds, it
is unlikely to be implemented in the near future. Similarly, the contractual approach has
some appeal, as being more market based than the statutory one, but transitional
problems and incentives to implement it may be insurmountable. Thus, for the time
being, working with the status quo remains the dominant option. And I have argued that
using the current market based regime (or non-regime) thus allow to minimize collective
action problem and provide for orderly restructurings.
In part, the verdict on the appropriate regime will depend on the experience with the
Argentine restructuring. This is a most complex case given the heterogeneity of both the
claims and the claimants. If the Argentine debt restructuring becomes messy, disorderly,
protracted and causes avoidable loss of economic value that hurts both the debtor and the
creditors, the political pressure to reform the current regime and move towards the
adoption of an international bankruptcy regime or a contractual approach will increase. If,
instead, the creative use of exchange offers with various carrots and sticks allows to
achieve an orderly restructuring, the incentive and pressure to create a new statutory or
contractual regime would disappear.
References
Bulow, Jeremy and Kenneth Rogoff (1989) “Sovereign Debt: Is to Forgive to
Forget?” American Economic Review, Vol. 79, No.1, March.
Corsetti, Giancarlo and Nouriel Roubini (1997) “"Politically Motivated Fiscal
Deficits: Policy Issues in Closed and Open Economies," Economics and Politics, March
1997.
Eichengreen, Barry (1999) Toward a New International Financial Architecture
Institute for International Economics, Washington, D.C. 1999.
Frankel, Jeffrey and Nouriel Roubini, “The Role of Industrial Country Policies in
Emerging Market Crises,” NBER WP No. 8634, December 2001; forthcoming in M.
Feldstein (ed.) Economic and Financial Crises in Emerging Market Economies, volume
published by NBER and Chicago University Press, 2002.
Jeanne, Olivier and Jeromin Zettelmeyer (2001) “International Bailouts, Moral
Hazard, and Conditionality,” Economic Policy 16, issue No 33, October.
Krueger, Anne (2001a) “International Financial Architecture for 2002: A New
Approach to Sovereign Debt Restructuring,” IMF November 26, 2001.
(http://www.imf.org/external/np/speeches/2001/112601.htm)
Krueger, Anne (2001b) “A New Approach to Sovereign Debt Restructuring,”
IMF, December 20, 2001.
(http://www.imf.org/external/np/speeches/2001/122001.htm)
Krueger, Anne (2002) “New Approaches to Sovereign Debt Restructuring: An
Update on Our Thinking,” Conference on "Sovereign Debt Workouts: Hopes and
Hazards" Institute for International Economics, Washington DC, April 1, 2002
(http://www.imf.org/external/np/speeches/2002/040102.htm)
Rogoff, Ken and Jeromin Zettelmeyer (2002) “Early Ideas on Sovereign
Bankruptcy Reorganization: A Survey,” IMF Working Paper 02/57, March.
(http://www.imf.org/external/pubs/ft/wp/2002/wp0257.pdf)
Roubini, Nouriel (2000) “Bail-In, Burden-Sharing, Private Sector Involvement
(PSI) in Crisis Resolution and Constructive Engagement of the Private Sector. A Primer:
Evolving Definitions, Doctrine, Practice and Case Law,” September 2000.
(http://www.stern.nyu.edu/globalmacro/psipaper.pdf)
Roubini, Nouriel (2001) “Why should the foreign creditors of Argentina take a
greater hit/haircut than the domestic ones: On the economic logic, efficiency, fairness and
legality of "discriminating" between domestic and foreign debt in sovereign debt
restructurings” December 14, 2001
(http://www.stern.nyu.edu/~nroubini/asia/discriminationforeigndebt.doc)
Roubini, Nouriel (2002a) “Do we need a new international bankruptcy regime?
Alternative regimes for orderly sovereign debt restructurings”, April 2002, New York
University.
Roubini, Nouriel (2002b) “Private Sector Involvement in Crisis Resolution and
Orderly Sovereign Debt Restructurings: The G-7/IMF Framework and the Open
Unresolved Issues,” March 2002, New York University.
Sachs, Jeffrey (1995), “Do We Need an International Lender of Last Resort,”
Frank D. Graham Lecture at Princeton University 8, April 20, 1995 (unpublished
manuscript.
(http://www.ksg.harvard.edu/cid/ciddirector/publicat.html#Working).
Taylor, John B. (2002a) “Grants and Sovereign Debt Restructuring: Two Key
Elements of a Reform Agenda for the International Financial Institutions,” Testimony of
John B. Taylor Under Secretary of the Treasury for International Affairs before the Joint
Economic Committee, February 14, 2002
(http://www.treas.gov/press/releases/po1016.htm)
Taylor, John (2002b) Sovereign Debt Restructuring: A U.S. Perspective” remarks
at Conference on "Sovereign Debt Workouts: Hopes and Hazards" Institute for
International Economics, Washington DC, April 1, 2002.
(http://www.treas.gov/press/releases/po2056.htm)
Wright, Mark (2001) “Reputations and Sovereign Debt,” unpublished, MIT.