private sector involvement in crisis prevention …following this introduction, the first section...

Aprominent feature of the internationalfinancial landscape in recent years hasbeen the occurrence of several finan-cial crises—Mexico in 1994–95, Asia in1997, Russia in 1998, and Brazil in 1998–99. Incontrast to regular modulations in the volume ofcapital flows in response to changes in underly-ing economic fundamentals, a distinguishingcharacteristic of these crises was that markets be-came quickly and completely one-way. Private in-vestors wanted, and attempted, to withdraw fromthese countries at the same time, with the dy-namic very much resembling that of a run by de-positors on a bank. Once sentiment soured, thedynamic became self-fulfilling in that once a crit-ical mass of investors rushed to withdraw theirclaims, it became rational for everyone else todo the same.

The “rules of the game” for international lend-ing and investment, especially as they apply tosovereign borrowers, have always been, and re-main, quite different from those that applywithin domestic boundaries. Two particularlynoteworthy differences are the absence of aclearly established international lender-of-last-resort to sovereigns and the absence of a bank-ruptcy court for sovereigns. By contrast, in thedomestic context, in most countries the centralbank stands ready to step in at short notice tosupport a domestic bank facing liquidity prob-lems, significantly mitigating the risk of a crisis.Similarly, there are well-defined bankruptcy pro-cedures for corporates in many countries thatprovide discipline on both creditors and debtors,allowing for an orderly workout while preservingthe value of an enterprise. The recent emergingmarket crises provided a sobering reminder thatthe absence of clear rules of the game in the in-ternational context combined with sharp swingsin international investor sentiment leaves the in-ternational financial system vulnerable to a vari-ety of inefficiencies and costs:

• A massive withdrawal of capital typically im-poses a severe liquidity squeeze on the debtorcountry, which can result in tremendous over-reaction of asset prices and subsequently sub-stantial output losses.

• The losses can also be substantial for creditors.Moreover, they could be arbitrarily distributedbetween creditor groups. In a context of lim-ited resources, the early exit of some investorswill necessarily be at the expense of those re-maining.

• Coordination process problems between credi-tors, or between creditors and debtors, canprevent markets from quickly and efficientlyresolving payments problems

• The event risk such as defaults associated withcrises in emerging markets, once viewed andpriced in as recurrent and disorderly by in-vestors, will raise the cost of financing foremerging market entities.

• As the crisis in Russia demonstrated, there canbe a widespread spillover that can disrupt oreven threaten the stability of the internationalfinancial system.Clearly, it is in the interest of both the private

and public sectors to have an international fi-nancial system that minimizes the above ineffi-ciencies and the damage from instability. It is inthe context of these market or systemic failuresthat the official community’s efforts at crisis pre-vention and resolution need to be viewed. Oneway of doing this is to adopt measures that re-duce the frequency with which crises occur, andthis has been a key objective of the official com-munity. These measures cannot be foolproof,however, and crises will occur. Indeed, crises pro-vide one mechanism for discipline in financialmarkets. More controversial have been the ef-forts to develop measures for, and the experi-ence with, the resolution of crises. Internationalliquidity support for countries facing external fi-nancing difficulties—often mischaracterized as

115

CHAPTER V

PRIVATE SECTOR INVOLVEMENT IN CRISISPREVENTION AND RESOLUTION: MARKET VIEWS AND RECENT EXPERIENCE

“bailouts”—are one potential solution.1 How-ever, as is well understood, such support can giverise to concerns about moral hazard—both forthe debtor and the creditor. Moreover, as a prac-tical matter, the extent of public resources isclearly limited. Some of the lessons for the inter-national financial system from orderly domesticrestructurings are that a well-functioning systemimposes discipline on both debtors and credi-tors, and that it of necessity involves the privatesector.

Private sector involvement (PSI) in the resolu-tion of crises has been a contentious issue andthere has been much rhetoric. It is important tonote that PSI in crisis resolution is not new. Ithas always been an inherent component, but theextent to which this has been the case has beena function of the nature of capital flows. Foreigninvestors who made direct or portfolio equity in-vestments in the crisis countries clearly sufferedcapital losses as the prices of these assets fell withmarket conditions. By the market’s reduction inthe value of their claims in the crisis countries,these investors were, therefore, clearly “involved”in the resolution if they liquidated their posi-tions. As a general matter, the shrinkage of thevalue of these investments reduced pressures, de-creasing the amount that could flow out andthough these large asset price movements im-posed substantial costs, these investments didnot generate the potential for disorderly pay-ments problems.

The most contentious area has been debtflows. For some of these claims, such as thelarge-scale withdrawal of bank loans from Asia,which took place at face value, there was noshrinking in the value of contractual claims inresponse to the deterioration in market condi-tions. For bond claims, while their secondarymarket value also fell with the deterioration ingeneral market conditions, the fact that the

value of these claims is fixed contractually innominal terms meant that maturing debt also re-sulted in withdrawals at face value. Both types ofdebt claims have, therefore, been a key source ofpotentially disorderly outcomes—by severelytightening liquidity and/or by forcing a default.As a result, crisis resolution efforts have focusedon the bank loan and bond markets, areas thathave been the most contentious given that non-or incomplete payment violates the debtor’s le-gal contractual obligation to the creditor.

In the absence of clear rules of the game, themarket’s interpretation of the precedents set bythe experiences with crisis prevention and reso-lution, and PSI in particular, have fundamentaland far-reaching implications in shaping thefunctioning of the international financial systemand the terms and nature of international fi-nancing to emerging markets. This chapter pro-vides a systematic review of recent experienceswith, and market views on, PSI in crisis preven-tion and crisis resolution. The purpose of thechapter is not only to present these views, butalso to assess their substance and relevance. It isimportant to realize that the private sector’sviews are not those of a disinterested party. A pri-vate sector creditor that may lose money due toa policy or action of the official community willtend to react negatively. However, the objectiveof the official community is not to avoid alllosses to private creditors—or to debtors whomay have borrowed imprudently—but rather toseek to ensure the relatively efficient functioningof the international financial system. In pursuingthis objective, the official community needs tobe aware of, and take account of, how the pri-vate sector will interpret and react to its policiesand actions.

The market views presented are based on a se-ries of informal discussions with commercial andinvestment banks, institutional investors, hedge

CHAPTER V PRIVATE SECTOR INVOLVEMENT IN CRISIS PREVENTION AND RESOLUTION

116

1In the domestic context, public support often comes in forms that impose large costs on taxpayers and effectively bailout debtors and creditors at the taxpayers’ expense. International support—especially from the IMF—is usually in the formof loans that must be repaid with interest below market rates. The countries that receive such international support andtheir creditors are not “bailed out” at the expense of taxpayers in the countries that help supply international support.-Whether a domestic or international context is considered, in the end the crisis country’s debtors are the ones carryingthe repayment burden.

funds, market associations, and legal counsels.Among market participants there is no absoluteconsensus with respect to PSI in crisis preventionand crisis resolution; rather, this chapter tries toprovide the staff’s interpretation of the centraltendencies of the views of the “market partici-pants.” An attempt has also been made through-out the chapter to identify larger disagreementsamong various private sector groups whenprominent. The chapter assesses how the recentPSI experiences relate to these broad marketviews and the extent to which the objectives oflimiting the size of official financing packages, ofreducing moral hazard, and of restoringmedium-term external viability of the debtorcountry are seen as having been attained, andhow the recent experience is likely to impact thenature of international capital flows.

Following this introduction, the first sectionreviews market reactions to recent proposals topromote PSI in the prevention of crises, includ-ing those focused on how to increase trans-parency, improve adherence to standards anddata dissemination, strengthen debt manage-ment, enhance debtor-creditor relationships innormal times as well as in crisis periods, pro-mote the role of collective action clauses in in-ternational bonds, implement payments stand-stills and litigation stays, and utilize market-based capital controls. The second section pro-vides some background on the recent PSI initia-tives relating to crisis resolution by briefly dis-cussing the historical experience with the 1980sdebt crisis. The third section presents marketviews on PSI in crisis resolution. The fourth sec-tion discusses the recent experience with the ma-jor concerted rollovers of interbank loans, thatis, Korea, Indonesia, and Brazil. It discusses theextent to which they were voluntary, whether theexperience to date suggests that they con-tributed to PSI in crisis resolution, in what cir-cumstances rollovers are most likely to be useful,

and what the likely impact will be on the leveland maturity of interbank lending to emergingmarkets in the future. The fifth section examinesthe recent experience with external bond re-structuring. It discusses the concepts for deter-mining the type and scope of bond restructur-ing, that is, whether a country is facing aliquidity or solvency problem and whether it hasaccess to international capital markets. It ex-plores the Paris Club’s comparability of treat-ment principle and how it has been interpreted.The chapter concludes with a brief discussion re-lating the recent experience to market percep-tions about likely future cases of private sectorinvolvement.

Private Sector Involvement in CrisisPrevention

Since the Asian crisis, there have been a multi-tude of initiatives for enhancing involvement ofthe private sector in crisis prevention. Most no-tably, work has been done by the IMF, theFinancial Stability Forum, the World Bank, theCouncil on Foreign Relations, and many others.Since many of these initiatives have been dis-cussed and summarized in other IMF docu-ments,2 this section focuses on analyzing marketparticipants’ views.3

Increased Transparency, Adherence to Standards,and Data Dissemination

The increased importance attached by marketparticipants to greater transparency in standardsof policymaking and improved data dissemina-tion is one of the legacies of the recent crises.Inadequate knowledge about the foreign cur-rency exposure of the banking system (as hap-pened in Korea), or about the level of a centralbank’s forward foreign exchange position (ashappened in Thailand), is widely regarded as

PRIVATE SECTOR INVOLVEMENT IN CRISIS PREVENTION

117

2IMF (1999c), provides an extensive treatment on official community initiatives already under way for dealing with crisisprevention.

3Table 5.1 provides a comparison of some market views with those of the official community regarding the various offi-cial initiatives for PSI in crisis prevention and crisis resolution. The table reflects the view of the author (Chase Manhattan,2000) on the IMF/official community’s position on various issues, and market associations’ reactions.


118

Tabl

e 5.

1.Pr

ivat

e Se

ctor

Invo

lvem

ent i

n Cr

isis

Res

olut

ion-

Sum

mar

y of

Pos

ition

s of

G-7

/IMF,

Inst

itute

of I

nter

natio

nal F

inan

ce, a

nd E

mer

ging

Mar

kets

Tra

ders

Ass

ocia

tion

I. Ar

eas

of B

road

Agr

eem

ent

Prop

osal

Colo

gne

Prin

cipl

es/IM

FIn

stitu

te o

f Int

erna

tiona

l Fin

ance

Emer

ging

Mar

kets

Tra

ders

Ass

ocia

tion

Burd

en s

harin

gPr

ivat

e cr

edito

rs s

houl

d be

ar c

onse

quen

ces

of

Agre

e.Ag

ree.

inve

stm

ent d

ecis

ions

.Bo

nds

not n

eces

saril

y ex

clud

ed fr

om fu

ture

re

stru

ctur

ing.

Rule

s vs

. cas

e-by

-cas

eG-

7 fra

mew

ork

offe

rs g

uide

lines

, but

flex

ibili

ty

Case

-by-

case

ess

entia

l, ta

king

into

acc

ount

Ca

se-b

y-ca

se a

ppro

ach

mus

t be

guid

ed b

y cl

eare

r and

re

tain

ed fo

r diff

eren

t cas

es.

mar

ket e

xpec

tatio

ns b

ased

on

past

m

ore

cons

iste

nt p

rinci

ples

.ex

perie

nce.

IMF

Cont

inge

nt C

redi

t Lin

eIn

crea

ses

focu

s on

cris

is p

reve

ntio

n.Po

tent

ially

impo

rtant

veh

icle

for a

ssur

ing

Prov

ides

ext

ra in

cent

ive

to p

rom

ote

stan

dard

spr

ivat

e fin

ance

.

Prac

tical

que

stio

ns n

eed

to b

e a

ddre

ssed

.

Debt

or/c

redi

tor d

ialo

gue

Enco

urag

es d

ebto

rs to

est

ablis

h m

echa

nism

s

Prop

oses

maj

or b

orro

win

g co

untri

es to

ado

pt

Play

s a

faci

litat

ing

role

in c

onsu

ltatio

ns b

etw

een

for m

ore

syst

emat

ic d

ialo

gue.

pr

oact

ive

stra

tegi

es o

f inv

esto

r rel

atio

ns.

Ecua

dor a

nd it

s cr

edito

rs.

Prop

oses

to e

stab

lish

a Pr

ivat

e Se

ctor

Adv

isor

y Co

unci

l for

the

IMF.

Mar

ket-b

ased

tool

sFa

cilit

ates

acc

ess

to th

e in

tern

atio

nal m

arke

ts

Cred

it en

hanc

emen

ts s

uch

as W

orld

Ban

k in

tim

es o

f ins

tabi

lity.

“p

olic

y-ba

sed

guar

ante

es”

are

a he

lpfu

l ste

p He

lps

prev

ent l

iqui

dity

cris

is.

that

can

lend

cre

dibi

lity

on p

olic

y re

form

s,

and

can

cata

lyze

new

priv

ate

sect

or fl

ows.

II. K

ey D

iffer

ence

s in

Pos

ition

s

Burd

en s

harin

gG-

7 fra

mew

ork:

pro

mot

e m

ore

orde

rly c

risis

Bu

rden

sha

ring

is fl

awed

con

cept

—ch

ange

s Re

cogn

izes

legi

timac

y of

goa

l, bu

t bel

ieve

s ap

proa

ch

reso

lutio

n in

hel

ping

deb

tors

and

cre

dito

rs

in “

grou

nd ru

les”

cou

ld u

pset

bal

ance

is

flaw

ed. F

orce

d re

sche

dulin

gs w

ill d

isco

urag

e fin

d co

oper

ativ

e so

lutio

ns.

in s

over

eign

wor

kout

s.ca

pita

l flo

ws.

Mar

ket-o

rient

ed a

ppro

ach

urge

d.Pr

ivat

e le

nder

s ha

ve d

iffer

ent o

bjec

tives

In

vest

ors

need

ass

uran

ce th

at e

mer

ging

mar

kets

, (s

hare

hold

er v

alue

) tha

n of

ficia

l len

ders

.an

d of

ficia

l sec

tor i

n su

ppor

t of t

he c

ount

ries,

are

Di

fficu

lt to

find

evi

denc

e of

“ba

iling

out

”; in

com

plet

ely

com

mitt

ed to

hon

orin

g de

bts.

med

ium

-term

con

text

cre

dito

rs h

ave

prov

ided

sub

stan

tial n

et fi

nanc

ing

desp

ite

larg

e lo

sses

.Fe

w p

rivat

e le

nder

s/in

vest

ors

can

pre-

com

mit

to fi

ll fin

anci

ng g

aps.

Invo

lunt

ary

appr

oach

es w

ill d

elay

eve

ntua

lre

stor

atio

n of

nor

mal

mar

ket a

cces

s.Cr

itica

l to

ensu

re c

risis

reso

lutio

n at

vol

unta

ry

end

of s

pect

rum

.


119

Paris

Clu

b “c

ompa

rabi

lity

of

Appl

ies

to a

ll ca

tego

ries

of c

redi

tors

oth

er

Rigi

d ap

plic

atio

n of

“co

mpa

rabi

lity

of tr

eatm

ent”

Paris

Clu

b go

vern

men

ts d

id n

ot p

artic

ipat

e in

deb

t tre

atm

ent”

than

IFIs

.co

uld

dam

age

futu

re m

arke

t acc

ess.

writ

e-do

wns

con

curr

ently

with

Bra

dy p

lan.

No c

laim

s sh

ould

be

auto

mat

ical

ly e

xem

pt.

Bond

s sh

ould

not

aut

omat

ical

ly b

eex

empt

.No

sou

rce

of fi

nanc

ing

shou

ld b

e au

tom

atic

ally

Pa

ris C

lub

shou

ld ta

ke fl

exib

le a

ppro

ach,

taki

ng

Mar

ket-b

ased

sol

utio

n ch

osen

by

the

debt

or

exem

pt fr

om b

urde

n sh

arin

g, b

ut n

ot a

ll eq

ual.

into

acc

ount

size

and

impo

rtanc

e of

cla

ims.

is b

est.

Lack

of P

aris

Clu

b tra

nspa

renc

y is

serio

us

obst

acle

to a

chie

ving

com

para

ble

treat

men

t in

cur

rent

cas

es.

Stan

dstil

lsCa

pita

l or e

xcha

nge

cont

rols

impo

sed

in

Litig

atio

n no

t sig

nific

ant o

bsta

cle

to o

rder

lyex

cept

iona

l cas

es in

con

junc

tion

with

IMF

.w

orko

uts

whe

re a

utho

ritie

s ar

e im

plem

entin

gsu

ppor

tcr

edib

le a

djus

tmen

t pro

gram

s.

Abili

ty to

litig

ate

unde

rpin

s al

l cro

ss-b

orde

rflo

ws

of fi

nanc

e.

Stay

s on

litig

atio

n w

ould

impe

de p

rodu

ctiv

eflo

ws

of p

rivat

e ca

pita

l.

IMF

lend

ing

into

arr

ears

May

be

appr

opria

te if

deb

tor s

eeki

ng

Like

ly to

dis

cour

age

priv

ate

finan

cing

unl

ess

coop

erat

ive

solu

tions

with

cre

dito

rs.

exce

ptio

nal i

mpe

dim

ents

exi

st a

nd a

bro

ad

spec

trum

of p

rivat

e c

redi

tors

sup

port

such

act

ion.

Brad

y Bo

nd re

stru

ctur

ing

(onl

y)No

cat

egor

y of

cla

ims

inhe

rent

ly s

enio

r.Br

ady

bond

s em

body

pas

t deb

t red

uctio

n.M

isgu

ided

bec

ause

:Ad

vers

e im

pact

of r

estru

ctur

ing

1) p

revi

ous

hairc

ut;

Brad

y bo

nds

in b

road

em

ergi

ng m

arke

t ass

et

2) v

iola

tes

soun

d pr

inci

ples

of w

ork

outs

abs

ent

clas

s/fu

ture

a

com

preh

ensi

ve p

lan;

Brad

y-st

yle

issu

es s

houl

d be

giv

en

3) d

oes

not r

esol

ve m

oral

haz

ard;

appr

opria

te w

eigh

t by

offic

ial l

ende

rs.

4) n

ot c

ompa

rabl

e;5)

mis

use

of in

tere

st c

olla

tera

l;6)

terr

ible

pre

cede

nt.

Colle

ctiv

e ac

tion

clau

ses

Enco

urag

e as

bes

t pra

ctic

e.At

tem

pts

to m

ake

qual

ified

resc

hedu

ling

Initi

ativ

e m

ust n

ot u

nder

min

e le

gal r

espo

nsib

ility

of

Cons

ider

incl

udin

g in

G-1

0 bo

nds

(at l

east

FX)

cl

ause

s m

anda

tory

are

cou

nter

prod

uctiv

e.de

btor

s.an

d as

a c

onsi

dera

tion

for a

cces

s to

IMF’s

CCL

Po

tent

ially

frui

tful a

rea

for p

ublic

/priv

ate

Shar

ing

clau

ses

obje

ctio

nabl

e-th

reat

en th

e le

gal r

ight

of

or IM

F co

nditi

onal

ity.

sect

or c

oope

ratio

n.cr

edito

rs to

enf

orce

cla

ims.

Maj

ority

vot

ing

of a

bout

90

perc

ent a

ccep

tabl

e, b

ut

shou

ld b

e vo

lunt

ary.

Trus

tees

Enco

urag

e w

ider

use

.Ad

d lit

tle in

sov

erei

gn c

onte

xt. P

rovi

sion

s th

at

wou

ld in

fring

e on

indi

vidu

als’

righ

ts (i

.e.,

de fa

cto

shar

ing

clau

ses

not w

elco

me)

.

Sour

ce: P

ublis

hed

in C

hase

Man

hatta

n (2

000b

). Ba

sed

on B

ank

of E

ngla

nd (2

000)

having provided a case for improved trans-parency and data dissemination. An improvedflow of information is viewed as vital if marketsare to appropriately assess and price risks and ifthere is to be effective market discipline that en-courages governments to undertake more open,predictable, and responsible macroeconomicand financial policies.

Given this consensus that there is a need toimprove data dissemination and standards forpolicymaking, it is somewhat surprising that thestaff’s discussions with officials and market par-ticipants indicate a “profound” lack of awarenessin many markets regarding recent efforts by theofficial community to improve the quality andquantity of information available to the privatesector. This is worrying from two different pointsof view. First, the lack of awareness may reflectthe fact that the data actually provided underthese initiatives are not market relevant,4 or, assome market participants pointed out, theirquality is not reliable enough for them to use.Second, there is the concern that, if creditorsare not interested in the information that is be-ing developed, there will not be much emphasisplaced on observance on the part of borrowers.

Improved Debt Management

Market participants regard improved domesticand external debt management as one of thekeys to mitigating the severity and frequency ofcrises. Past crises have demonstrated that privateand official foreign currency–denominated debtcan create additional sources of vulnerability.From a debt management perspective for boththe sovereign’s borrowing and the general bor-rowing done by its private sector, the key issuesare typically seen as avoiding both a bunching inmaturities and inclusion of derivatives in bondcontracts (such as put options) that exacerbatedebt service payments during crises.5 It is impor-tant for the sovereign to ensure prudent debtmanagement by its private sector, especially in a

fixed exchange rate regime, to stop private sec-tor debt from exacerbating the country’s vulner-ability to a crisis. This may imply a lower level ofprivate sector foreign borrowing, but it wouldstrengthen the country’s overall ability to with-stand external shocks. Otherwise, the sovereignmay find itself in a situation where it has to stepin and socialize the private sector’s foreign bor-rowings to resolve the crisis. Throughout, a chal-lenge is to convince any emerging marketdebtor that the up-front cost savings achieved byissuing short-term bonds or bonds with put op-tions always have a cost—one that is harder toquantify—in terms of the additional vulnerabilityto shocks. If a sovereign not only successfullymanages its own debt but also oversees the bor-rowing of its private sector it is well on its way toimproving its ability to withstand future externalshocks.

Improving Debtor-Creditor Relations inNormal Times

Improved debtor-creditor dialogue is perhapsthe one proposed initiative that gains universalapproval by both the official and private sectors.A continuing dialogue with the borrowing coun-try is viewed by many market participants as away to strengthen trust in the lending relation-ship and greatly facilitate negotiations in theevent of debt-servicing difficulties. The benefitof such a dialogue is seen to be, in part, a re-moval of the uncertainty premium that emerg-ing market borrowers have to pay when a lack ofreliable information leads investors to assumethe worst.

The main obstacles to better debtor-creditordialogue, according to market participants, arethe lack of preparedness and public relationsskill of some sovereign borrowers, the sensitivityof some country information, the preference ofsome major individual investors for individualmeetings, and the desire of some bondholdersto remain anonymous.


120

4Timeliness of the data disseminated may also affect its relevance for market participants.5See IMF (1999a) for an overview of the importance of debt management for crisis prevention.

Debtor-Creditor Relations in Crises:Creditor Committees

There are several ways debtor-creditor rela-tionships can be organized in advance of a crisis.One early suggestion was the setting up of astanding creditor committee,6 which could playan important role in facilitating informationsharing across creditors and negotiate on theirbehalf. However, some market participants dis-like standing committees since they want to ne-gotiate on their own behalf; they also fear thatthe committee may not be representative as theholders of the debt likely shifts when there issecondary market trading, and they have a con-cern that a standing committee may reduce thecost of default for the sovereign borrower, bymaking it more orderly, and hence increase theprobability of default. Due to these and otherprivate sector concerns, the official communityhas taken the position that standing creditorcommittees are “generally not considered practi-cal”7 and has since sought to lay down someguiding principles in advance for the formationof “ad hoc” creditor committees, which mainlywould serve the purpose of coordinating andsharing information across creditor groups. Thisapproach is receiving some support from the pri-vate sector through discussions under the aus-pices of the Council of Foreign Relations.However, such a committee raises the difficultquestion of how to handle confidential informa-tion and whether the various different types ofcreditors would maintain adequate “Chinesewalls” to avoid trading on the basis of confiden-tial information. While some market participantsare optimistic that the Chinese wall problem canbe overcome, others have pointed out that a bet-ter approach would be to recognize that theproblem could never be satisfactorily solved andthe committee should only deal with simultane-ously published information.

Some market participants take the view thatthe recent bond exchanges in Pakistan, Ukraine,

and Russia show that there is no real need forsetting up formal creditor committees. In fact,they additionally point to what they regard asthe negative experience of Ecuador with itsConsultative Group. This group, set up by thesovereign, and including only holders of interna-tional bond debt, was widely seen as a noncon-structive point of contact between creditors andthe debtor, tending to “radicalize” creditors’ po-sitions and potentially serving as a coordinatingmechanism against the sovereign. Mostly re-cently, Ecuador has decided to meet with majorholders of its external debt on an individualbasis.8

Collective Action Clauses

Collective action clauses (CACs) are clausesthat can be incorporated into an internationalbond’s legal documentation to facilitate the re-structuring of that bond if needed. The clausescommonly referred to as CACs are a majority ac-tion clause (allowing a qualified majority ofbondholders to bind a minority); a sharingclause (stating that any funds received through,for example, litigation by one bondholder haveto be shared with the other bondholders basedon their share of the outstanding bond); and acollective representation clause (allowing atrustee, for example, to represent bondholdersat bondholders meetings facilitating majority ac-tions). The presence of CACs could facilitatecreditor-debtor negotiations following a crisis,since they reduce both the threshold from the100 percent needed for achieving a restruc-turing agreement (the majority action clause)and the potential threat of litigation from “hold-out” creditors (reducing their incentive to liti-gate through the sharing clause or implicitlyfrom the inclusion of a trustee). Bonds issuedunder English law typically include CACs explic-itly or implicitly. Since these clauses are not reg-ularly contained in bonds issued under New


121

6Eichengreen (1999).7IMF (1999c).8Bloomberg (2000).

York law,9 proposals have been made to includesuch covenants in all sovereign bonds issued in-ternationally going forward, in an effort to facil-itate orderly restructuring. (An alternativemechanism to CACs would be the use of exitconsents, which have the advantage of also be-ing applicable to bonds issued under New Yorklaw; see Buchheit and Gulati, 2000). To supportthe adoption of CACs by emerging market sov-ereigns, the United Kingdom has recently incor-porated CACs in its own “Euro Treasury Note”program.10 Canada announced in April that inthe future its international bond issues will in-clude CACs.

Market participants have argued that the offi-cial sector has exaggerated the importance ofCACs. Indeed, bond trustees have indicated thatthey would not call a bondholders meeting un-less they had already reached a prior agreementwith the required majority of creditors for ap-proval of a restructuring. Without such a priorapproval, creditor meetings are viewed as likelyto produce “strange and exotic” outcomes, mostof which would be bad for the debtor (for exam-ple, the bond holders’ meeting could be used tocollect enough votes to accelerate the bond).While market commentary on the impact foremerging market borrowers on the wider adop-tion of CACs in bond contracts is mixed, empiri-cal work11 suggests collective action provisionstend to reduce the cost of borrowing for themore creditworthy issuers, who benefit from anorderly restructuring process. In contrast, lower-rated borrowers are shown to pay a premium ifCACs are included. This suggests that for alower-rated borrower, creditors would view theadoption of CACs as a signal of additional de-fault risk that would more than offset the advan-tages the provisions afford by facilitating a moreorderly restructuring process. The additional

cost paid by lower rated borrowers and the im-plicit reduction in international debt financingavailable to them may also avoid the buildup ofpotentially destabilizing debt flows in the firstplace. In general, including these provisions inbond documentation going forward is widely re-garded as a step in the right direction (see Table5.1), but their inclusion will do little to help inthe restructuring of already outstanding bonds.

Standstills and Stays

Standstills and stays12 are primarily an issuefor crisis resolution, but they clearly have ex anteeffects in potentially limiting the number ofcrises that do occur. The analogy with domesticlender-of-last-resort facilities and bankruptcyprocedures has led to the suggestion that intimes of crisis a payments standstill should be de-clared. The issue then arises of whether thestandstill should be voluntary or mandatory, howlong it should last, and who should initiate it.Within the official community, there have beensome arguments made that a sanctioned stand-still could be seen as a last-resort measure de-signed to contain broadly based capital outflowsfrom a country. However, so far no consensushas been reached within the public and privatesectors, but some forums, such as the Council onForeign Relations,13 have recognized, that in ex-ceptional circumstances, a payments standstillmay be desirable and necessary because paniccan take hold and official financing resourcesare limited. Stays on litigation are, in this con-text, seen as another useful instrument to keepthe resulting crisis resolution negotiationsorderly.

Market participants are generally against in-voluntary standstills (see Table 5.1) and arguethat stays and standstills would infringe on their


122

9A bond issued under New York law will generally not permit an amendment to its payment terms without the agree-ment of every bondholder; other terms could be changed. See Buchheit (2000).

10See H.M. Treasury (2000).11Eichengreen and Mody (1999). See also Dixon and Wall (2000).12Refers to standstills on external debt service payments and an imposition of a stay on litigation otherwise likely to re-

sult from the declaration of a standstill.13Council on Foreign Relations (1999).

fundamental contractual rights to receive pay-ments and they would litigate in the event of apayment default. Furthermore, in their view, theway in which an involuntary standstill is imposedvery much affects their reaction to it. It was ar-gued that an involuntary standstill that is im-posed while keeping the private sector creditorsinformed and looks temporary in nature wouldengender much less of a negative reaction thana blanket decree imposing a standstill for alonger period. Seeing a limited role for volun-tary standstills, Corrigan (2000) makes the pointthat the only successful standstills have been therolling over of interbank loans. Therefore, hequestions whether voluntary standstills wouldconstitute a “viable and generalized approach”to deal with other potential outflows. With re-gard to forced standstills, Corrigan sees them—especially if sanctioned by the official commu-nity—as a “clear and present danger” for the“culture of credit” in the international capitalmarkets.

Nevertheless, involuntary standstills are likelyto remain “a fact of life” of cross-border lending,and emerging market borrowers will make useof this instrument if they have to. Indeed, defacto standstills have been a feature of many ofthe recent debt-service crises (see, for example,Indonesia).

Market-Based Capital Controls

Marked-based capital controls are a potentialcrisis prevention instrument that has received in-creased attention following the Asian crisis. Theyaim to avoid destabilizing capital inflows in thefirst place. Chile’s introduction of market-basedcapital controls was motivated both by pruden-tial considerations (including limiting the stockof short-term foreign currency obligations) anda desire to moderate the appreciation of the realexchange rate.14 Views on the effectiveness ofthe Chilean capital controls in achieving those

objectives, and the overall usefulness of market-based capital controls have been mixed.15

Background on Private SectorInvolvement in Crisis Resolution

PSI has been an integral part of all crisis reso-lution efforts and is not new. At the time of theresolution of the Latin American debt crisis ofthe 1980s (see Box 5.1), for example, the officialcommunity had many of the same objectives ithas today, limiting the size of official packages,reducing moral hazard in the private sector’slending decisions, and restoring the external via-bility of the country in crisis. Some (for exam-ple, see Dooley, 1994) see the lending precedingthe debt crisis of the 1980s as raising charges ofmoral hazard.

By the late 1980s, however, the improved fi-nancial positions of major international banks asmeasured by their developing country loan expo-sures relative to their capital and the continuingpoor economic performance of many emergingmarkets led to the adoption of the Brady plan,which involved substantial write-downs (meas-ured in net present value (NPV) terms) of devel-oping country syndicated loans. Indeed, thelosses experienced by banks on medium-termsyndicated lending to developing countries in the1980s are regarded as a key factor in the decisionof banks to shift away from syndicated lending tosovereigns toward shorter-term interbank lendingin the 1990s. (Other important factors were thewithdrawal of favorable tax treatment in creditorcountries and the introduction of the BaselAccord in 1988.) Large official financing pack-ages in the 1990s, starting with Mexico (1994–95)and then in Asia (1997), were also seen by manyobservers as increasing the private sector’s expec-tation of being rescued should it be confrontedby an imminent credit event. This sentimentlikely peaked in the run-up to the Russian defaultin August 1998 as the country was widely viewed

BACKGROUND ON PRIVATE SECTOR INVOLVEMENT IN CRISIS RESOLUTION

123

14For a more extensive discussion on Chile’s experience with capital controls see IMF (1995a) and for a broader discus-sion of Chile and other cases see Ariyoshi and others (2000).

15See, for example, Gallego, Hernández, and Schmidt-Hebbel (1999).


124

On August 20, 1982, Mexico’s FinanceMinister announced that Mexico was no longerable to service its external obligations. The an-nouncement created concerns that other devel-oping countries might also be unable to meettheir debt-service obligations. At the time, largeinternational and regional banks, especiallyfrom the United States (see table), had loan ex-posures to a relatively concentrated set of devel-oping countries that were large relative to thebanks’ capital positions. Widespread defaults onthese loans would thus pose a threat not only tothe stability of the banks themselves but also tothe international financial system.

How did this concentrated exposure arise? Inreviewing the lessons from the 1980s debt crisis,some observers (notably Dooley, 1994) havestressed the role of moral hazard considerationsin influencing intra-creditor dynamics betweenindustrialized country governments and the in-ternational banks that extended credit to thedeveloping countries. During the 1970s andearly 1980s, it was argued that the authorities inthe industrial countries, again especially in theUnited States, expressed their satisfaction with,and implicit support for, the way in whichinternational banks were recycling petrodollarsinto investments in developing countries.Banks were seen as willing to undertake thislending, at relatively low interest rate spreads,because of the belief that, if losses on suchlending were to occur, they would receive assis-tance from the official sector. From this per-spective, banks had strong incentives to expandtheir lending to developing countries, even to

the point where debt-servicing problems couldcause widespread solvency problems for banks(Rhodes, 1989).

Others have seen the 1980s debt crisis as re-flecting a much more complex set of macroeco-nomic and financial factors. While anticipationof support of banks by G-10 authorities in thecase of debt-servicing difficulties for developingcountries may have influenced the willingness ofbanks to participate in the recycling process,other macroeconomic, financial, and policy con-siderations were seen as playing an equally im-portant role. Cline (1984), for example, arguedthat the large balance of payments deficits andthe associated accumulation of large externaldebts by developing countries reflected the com-bination of a sharp rise in oil prices, lower nega-tive real interest rates in the 1970s (whichencouraged borrowing), unsustainable macro-economic and financial policies in developingcountries, and a combination of high real inter-est rates and a recession in the industrial coun-tries in the early 1980s.

What form of private sector involvement (PSI)was utilized to help manage this crisis?Following Mexico’s announcement, the officialcommunity had the twin objectives of contain-ing the threat to the international banking sys-tem while at the same time trying to improvethe economic position of the debtor countriesto prevent a default on their debt-service obliga-tions. Throughout, three basic principles under-pinned the handling of the crisis by both the of-ficial community and the private sector (seeRhodes, 1989):

Box 5.1. The 1980s Debt Crisis and Private Sector Involvement

Exposure of the Nine Largest U.S. Banks to Non-Oil Developing Countries Relative to Capital(In percent)

1977 1978 1979 1980 1981 1982

Total 188 190 206 221 240 235

Brazil 42 42 40 39 41 46Mexico 33 30 30 38 44 44Eastern Europe1 25 24 24 22 20 14Other non-oil LDCs1 88 94 112 122 135 131

Source: Milivojević (1985).1As defined by the source.

as “too-big- or too-nuclear-to-fail” and wouldtherefore receive the support of the official com-munity no matter what.

Others have seen the crises of the 1980s and1990s as arising out of a much more complex setof macroeconomic and financial factors andhave argued that there needs to be a more nu-anced view of the extent and potential sourcesof moral hazard. It has also been argued thatmoral hazard in the international financial sys-tem can potentially arise from a number ofsources including the official safety net that un-derpins all banking systems and the lending ac-tivities of international financial institutions.

The official safety net underpinning the bank-ing system is typically designed to ensure theoverall stability of the domestic financial systemand to protect the domestic payments system. Itis widely recognized that the knowledge that abank is “too big to fail” can lessen the incentivesto impose both market and managerial disci-pline. Domestic bank bailouts costing the sover-eign the equivalent of 10–20 percent of GDPhave not been uncommon and clearly have animpact on the expectations for future bailouts bythe domestic banks as well as the expectations ofinternational banks providing financing to thedomestic banks.


125

The approach was balanced, in that all partiesaccepted a need for burden sharing within aconcerted effort. It was case-by-case, in recog-nition of the fact that each of the restructur-ing countries had its own particular circum-stances. And it also had to be flexible, giventhe many political and economic variables atwork internationally and within each restruc-turing country.

Initially, the heavily indebted developingcountries were viewed as facing a liquidity crisis,and forbearance by their private and publiccreditors combined with new money was seen asa means of meeting this liquidity shortage andthereby facilitating continued debt-service pay-ments to the creditor commercial banks.Although bank exposures increased with theprovision of new money, banks were graduallyable to build up both provisions against trou-bled loans and bank capital. In the mid-1980s,banks also continued negotiations with debtorcountries to restructure payment profiles (oftenthrough multiyear rescheduling agreements)and to achieve some debt reduction throughmarket based mechanisms such as debt-for-eq-uity swaps. The Baker Plan of 1985 continuedthe strategy of coordinated lending by the offi-cial and private sector while at the same timeshifting the focus of the debt strategy fromshort-term balance of payments adjustment to-

ward long-term structural changes. Moreover, by1987–88, there was also increasing emphasis onthe “menu approach” to mobilizing bank lend-ing in order to address the divergent needs andobjectives of the various classes of banks. Thesedivergences arose out of differences in regula-tory and accounting practices, as well as the con-flicting objectives of those banks that wanted towithdraw completely from lending to developingcountries and those that had a longer-term in-terest in remaining active lenders. These diver-gent interests, as well as the large-scale loan-lossprovisions by major banks during 1987, ledbanks to take a tougher negotiation stance andthey expressed much less interest in extendingmore new money. Moreover, some banks beganto avoid further involvement by selling their de-veloping country loans on an increasingly liquidsecondary market for distressed loans. By thelate 1980s, the combination of continuing weak-ness in the economic performance of thedebtor countries and a strengthening of the fi-nancial condition of the creditor banks led to anew approach based on voluntary, market-baseddebt reduction (the Brady plan of 1989). In re-turn, banks received a “sweetener” in the formof more liquid collateralized bonds, which wereregarded as harder to restructure in the futureand as capable of being sold to a broader in-vestor base than the existing syndicated loans.

While the moral hazard effects of the officialsafety net underpinning national banking sys-tems are a constant feature of the global finan-cial system, the potential moral hazard effects oflending by international financial institutionswill be influenced by both the scale and timingof such lending. As noted above, such lending isregarded by market participants as having hadits most significant effect on creditors’ expecta-tions during the run-up to the Russian default inAugust 1998. Nonetheless, there remains consid-erable disagreement between those that seelending by international financial institutions ashaving a “first-order” effect in creating moralhazard and those that view such lending as hav-ing a much smaller and episodic effect. Haldane(1999) argues that IMF lending to date has insome instances increased lender moral hazard,arguably by the IMF providing resources to acountry in crisis that are used to bail out the pri-vate sector. The so-called Meltzer report(International Financial Institution AdvisoryCommission, 2000) also argues that the expecta-tion of future IMF support packages has helped“fuel the volatile short-term capital flows thathave played a key role in recent crises” (seeMeltzer, 2000) and that the “importance ofmoral hazard cannot be overstated.” However,empirical evidence of lending from the interna-tional financial institutions, and in particularfrom the IMF, causing lender moral hazard ismixed. Krugman (2000) argues that there “is noshred of evidence, for example, that the in-vestors who poured money into Asia before itsrecent crisis thought at all about the possibilityof future IMF bailouts.” Instead, Krugman ar-gues that the main driving factor behind thelarge flows of credit to Asia was rather motivatedby “irrational exuberance.” The view that lendermoral hazard did not drive the decision of credi-tors to lend to Asia receives support from Mussa(1999), who argues that many of the argumentsbehind the view that lending by the interna-tional financial institutions contributes to debtorand/or lender moral hazard are “simplistic andfundamentally wrong.” The fact that some credi-tors ex post come out of a crisis unscathed does

not necessarily provide evidence that ex ante thelending decision was based on a hope of beingbailed out. Furthermore, many observers forgetthat international support packages to countriesin crisis are in the forms of loans and not grants,and that they involve significant conditionality,ensuring that debtor moral hazard is limited.

Whatever the conclusion on the likely signifi-cance of moral hazard arising from official inter-national support to countries facing external fi-nancing difficulties, the fact is that the scale ofsuch support is limited. When there is a mean-ingful risk that a country may be insolvent andtherefore incapable of timely repayment ofemergency official assistance, the official com-munity typically refrains from providing such as-sistance except on the condition that otherclaims against the country be rescheduled andwritten down to an extent that ensures thatemergency official assistance can be repaid.These are situations where, like it or not, thecreditors of a country (its sovereign, its bankingsystem, or its private sector) will unavoidably be“involved” in the resolution of the country’s fi-nancial difficulties. More broadly, when a coun-try faces a huge outflux of capital that threatensto swamp that country’s own resources plus anyplausible level of emergency assistance from theofficial community, and when efforts to resolvethe crisis through policy adjustments, limited of-ficial assistance, and a spontaneous restorationof confidence fail, the creditors of that countrywill also face “involvement” in the resolution ofthat country’s financial difficulties on terms andconditions not contemplated in their credit in-struments. In these situations, private sector in-volvement in crisis resolution is, and always hasbeen, a fact of life.

In designing and implementing policies con-cerning private sector involvement, the officialsector has—and is perceived in private marketsto have—several, not necessarily consistent, ob-jectives. One is burden sharing. Because of con-cerns about moral hazard and for other reasons,the official community wants to keep its emer-gency support limited. It also wants to ensurethat private creditors play—and are seen to


126

play—an appropriate role in resolving crises thattheir lending has helped to engender. Whenlosses need to be absorbed—especially in situa-tions of insolvency—the official sector wants toensure that private creditors do not escape byimposing losses they should bear onto others. Asecond broad objective is limiting the damagedone by the crisis, both to the country primarilyinvolved and to the world economy more gener-ally. Sometimes, especially in cases of insolvency,this may mean that creditors should absorblosses (also part of burden sharing). It alsomeans, especially in cases of illiquidity, seekingto restore external viability and market access asrapidly as possible following the resolution of acrisis—something that may not be facilitated byefforts to impose substantial short-term losses oncreditors. The third broad objective of the offi-cial community—although this may not be fullyappreciated in private markets—is to preserveintegrity and reasonable efficiency in the func-tioning of international credit markets. Thismeans that debtors should not be allowed to es-cape from servicing their obligations when theyhave the capacity to do so. It also means thatcreditors who undertake risks should expect tosee those risks sometimes materialize into actuallosses. Symmetrically, for the official sector, itought to mean not using private sector involve-ment as a mechanism for off-budget foreign aidto countries in distress or for pursuing policy ob-jectives unrelated to the integrity and efficiencyof the international financial system.

Policies to pursue these different objectives in-teract and potentially conflict not only in deal-ing with a specific crisis but also dynamically, asthe private sector reacts to the policies of the of-ficial sector and the official sector, in turn,adapts its policies. This phenomenon is clearlyapparent in evolution of international creditarrangements over the past two decades.Medium-term loans from large syndicates ofcommercial banks to developing country sover-eigns and public sector entities were a dominantform of international capital flow before thedebt crisis of the 1980s. An important part of themechanism that the official sector used to deal

with that crisis involved the concerted rolloverand subsequent restructuring and write-down(in present value terms) of syndicated bankloans. Bonded debts of the sovereigns of the af-fected countries generally escaped restructuringon the grounds that the amounts were small andthat these instruments (held by widely diversi-fied creditors) were difficult to restructure. Themarket adapted. Medium-term syndicated bankloans to developing country sovereigns largelydisappeared in the 1990s. Banks shifted to inter-bank loans of much shorter maturity.International borrowing by sovereigns took pre-dominantly the form of bonded debts. The shiftsin the form of international credit flows posednew challenges in efforts to resolve the financialcrises of the 1990s. Lenders to emerging marketswere either thousands of individual bond hold-ers whose actions were difficult to concert, orbanks with short-term facilities that could easily“cut and run” in a crisis. As described below,mechanisms for private sector involvement inthe crises of the 1990s had to adapt to these newrealities.

The fact that the private sector will adapt tothe official sector’s policies and practices with re-spect to private sector involvement is not neces-sarily bad. For example, although they are prob-ably unwelcome to the potential debtors,policies that raise the cost and diminish theavailability of international credits to someemerging market borrowers may be desirable ifthey reflect a more appropriate pricing of risksand serve properly as a deterrent to imprudentborrowing. Alternatively, policies that encouragelonger-term securitized borrowing (which is pre-sumably limited by available collateral) may con-tribute to the avoidance or more efficient resolu-tion of crises because such loans are hard torestructure. Longer-term loans are likely to beless dangerous in a potential crisis than an equiv-alent volume of short-term loans; and creditorswho believe that they have secure collateralshould be less prone to panic than those that donot. On the other hand, a country that has al-ready encumbered most of its liquid assets and agood deal of its future export earnings may find


127

itself in a very difficult situation in the event of afinancial crisis. The point is that in consideringvarious policies and practices with respect to pri-vate sector involvement, it is critical to be awareof how the private sector is likely to adapt tothese policies and practices and to the difficul-ties or opportunities that these reactions willgenerate.

Market Views on Private SectorInvolvement in Crisis Resolution

Not surprisingly, market participants have ingeneral reacted negatively to the official commu-nity’s initiative to increase PSI in the resolutionof crises. Some market participants have ques-tioned the need for any PSI, arguing the initia-tive and uncertainty surrounding its applicationhave built into emerging market asset prices apremium that is eventually being borne by bor-rowing countries and is limiting private capitalflows, in fact raising the eventual extent of offi-cial funding necessary to finance emerging mar-kets’ investments. The majority of market partici-pants, however, recognize the need for PSI incrisis resolution. They recognize that interna-tional lending—even to sovereigns—does entailrisks, which are compensated to creditors exante through the risk premiums charged onloans. In an effectively functioning financial sys-tem, ex ante risks sometimes materialize into expost losses for creditors, with larger losses gener-ally accruing to those who take larger risks.Nevertheless, these market participants have ex-pressed several other concerns with the ap-proach taken to date.• Most market participants remain highly uncer-

tain, if not outright confused, about the offi-cial community’s approach to achieving itsgoals. The experiences of Pakistan andUkraine indicated to many market partici-pants that the official community was increas-

ingly focusing on restoring medium-term ex-ternal viability rather than on “burden shar-ing”16 with the private sector. The interpreta-tion of more recent initiatives in Nigeria have,however, once again raised questions aboutthe relative importance the official communityattaches to the two objectives.

• While it is generally recognized that there willnot be a detailed rules-based approach to PSI,there is a desire for a “framework” that wouldlimit arbitrariness and provide market partici-pants with some understanding of when PSIwill be invoked, what will determine the scaleof PSI, and whether PSI will be on a voluntary(i.e., negotiated between the debtor and credi-tors) or involuntary basis. Market participantsargue that the embryonic framework to date(as represented by recent statements by the G-7 and the International Monetary andFinancial Committee (IMFC)) amounts to nomore than a codification of the past few years’debate, and a summary of current practices.The framework is seen as remaining ad hocuntil disagreements within the IMF’s member-ship regarding the “rules of the game” for PSIare resolved. In this context, most expect that“bigger” countries would be treated differentlybecause of their systemic importance unlikesmaller countries.

• Regarding the relationship between claims ofthe official community and those of the privatesector, most market participants accept that theclaims of the multilaterals should be senior.They are not persuaded of similar treatmentfor the Paris Club, however, whose membersare seen as having often lent based on non-commercial, and sometimes political, consider-ations. Many market participants, therefore,take the view that “comparability of treatment,”for fairness reasons, should, if at all, work inboth directions, and the Paris Club’s claimsshould not be treated as senior. Furthermore,


128

16The official community’s insistence on burden sharing is frequently viewed by the private sector as reflecting an arbi-trary decision by the official community to fill a short-term financing gap with disproportionate private sector resources orforbearance to limit the size of official financing packages. If the objective of crisis resolution is to restore medium-term ex-ternal viability for a country’s balance of payments, the private sector’s reaction has generally been more supportive, assuch an objective is seen as being in the interest of both the private sector and the official community.

market participants expressed concern that thelack of information regarding how the ParisClub assesses “comparable treatment” intro-duces additional uncertainties and costs for theemerging market sovereign borrower.

• Many market participants have expressed con-cerns about the potentially ad hoc treatmentof different creditor classes, arguing it is notthe role of the official community to establishimplicit seniority hierarchies in intra-privatesector claims. For example, in the case ofEcuador there was a perception by market par-ticipants that holders of eurobonds and dollar-denominated domestic debt were initiallyabout to receive a more favorable treatmentthan Brady bondholders with the implicit sup-port of the official community.

• Looking ahead, market participants expressedconcerns about the implications of the officialcommunity’s PSI initiative for the structure ofcapital flows. One implication of the resolutionof the 1980s Latin American debt crisis is seento be the sharp drop-off in syndicated banklending to sovereigns, encouraging the takeoffof the emerging bond market. The present lessfavorable treatment of unsecuritized instru-ments, such as eurobonds, is seen as potentiallyleading to a greater use of harder to restruc-ture securitized bonds and loans.

Interbank RolloversThe rollover of interbank lines can be seen as

one instrument to mitigate crises primarily by re-ducing capital outflows. The usefulness of the in-strument depends on whether interbank lines,especially their nonrenewal, is an important fea-ture of the balance of payments crisis facing theparticular country. Hence, it could be arguedthat the applicability of this instrument has littleto do with whether a country faces a crisis due tosolvency or liquidity concerns, and one would

reasonably expect to see rollovers used in bothcases. The terms of the rollovers would, however,reflect the market’s assessment of whether it isasked to roll over its exposure to an insolventcountry or an illiquid country. This section willdiscuss the recent major rollovers of interbankdebt in Korea, Indonesia, and Brazil.17

A key challenge for the official community ishow investors can be convinced to maintaintheir exposure in times of stress. This is tricky tofinesse in a crisis situation where there is an in-centive to run preemptively at the first hint ofdifficulty. An example of the latter occurred inthe case of Brazil18 (see Box 5.2) where interna-tional banks aggressively cut the credit lines totheir local operations. An important reason wasthat the local operations had large holdings ofBrazilian domestic bonds and would have suf-fered if the rumored forced restructuring ofBrazil’s domestic debt had occurred. Hence, in-ternational banks with local operations judgedthat they were significantly more exposed to thedomestic restructuring risk than internationalbanks without local operations. To manage theirrisk from both a coerced rollover of interbankdebt à la Korea and a rumored forced restruc-turing of government domestic debt, the inter-national banks’ headquarters encouraged theirlocal operations to reduce their holdings of do-mestic debt while at the same time the creditlines extended to them by their headquarterswere cut. In Indonesia, the fact that a lot of ex-posures of international banks were toward thecorporate sector, which had stopped servicingtheir foreign currency debt, meant exposure wasmaintained whether creditors wanted it or not.

Comprehensive Versus Limited Approach

The cases discussed in Box 5.2 suggest thatthe instruments to address a particular balanceof payments crisis need to be tailored to the

INTERBANK ROLLOVERS

129

17Steps were also taken in the case of Thailand to encourage the rollover of a significant part of the short-term foreigncurrency debt, but a significant proportion of this comprised international banks’ (mostly Japanese) lending to their ownThai branches and the experience raises slightly different issues than those treated above. (For a more extensive treatmentof the Thai rollover, see Lane and others, 1999.)

18See IMF (1999b).


130

Korea

In December 1997, following intense pressureon reserves, four to five notch downgrades bythe major rating agencies that triggered addi-tional outflows, and the realization that much ofKorea’s foreign exchange reserve was actuallyunavailable, Korea was close to facing severedebt-servicing problems. With only $6 billion inreserves, interbank claims amounting to $28 bil-lion to be settled before end-February, andlosses of $1 billion a day, Korea faced a dauntingchallenge. Short-term interbank credit lines, themain source of external credit exposure forKorea, as local banks acted as intermediaries forother domestic debtors, had been subject tocutbacks, which accelerated in November. Asthe macro situation worsened further, the with-drawal of credit lines rapidly accelerated, caus-ing liquidity to dry up in the Korean interbankmarket. With interbank rollover ratios falling tonear zero, capital flight rampant, the IMF-supported program failing to restore confi-dence, and the $20 billion sovereign guaranteeof the Korean banking system proving ineffec-tive, the international official community wastrying to stabilize the situation. However, put-ting in the limited available official resourceswould have done little to ameliorate the situa-tion as these resources would have quicklyflowed out to pay down maturing interbanklines.

On December 22, 1997, the Federal ReserveBank of New York called a meeting to convincekey U.S. banks that a rollover of their maturinginterbank lines was in their own interest as notall of them could exit at the same time.Furthermore, it was highlighted that failure toroll over enough of the credit lines wouldclearly pose a systemic risk to the world’s finan-cial system and that, if agreement was reached,the official financial community would provideadditional accelerated resources. Represen-

tatives of the U.S. banks agreed on ChristmasEve to maintain their interbank lines at currentlevels for at least a week, while a longer-term so-lution could be found. Similar meetings tookplace in all the major financial centers withJapanese, German, English, French, and othercentral banks trying to convince their respectivecommercial banks to agree to a rollover. Afterfierce negotiations, broad international agree-ment was reached on January 16, 1998 to rollover creditor bank exposure until March 31,1998. A key component to enforce the agree-ment was a debt-monitoring system, set up byIMF and Bank of Korea staff, which helped solvethe inherent collective action problem2 presentin any rollover operation.

Rollover ratios quickly recovered in the sec-ond half of January 1998, reaching an averageof 80 percent for the month. The subsequenttwo months saw some leakage compared to theJanuary agreement, as some small creditorbanks exited and the approval of new creditlines suffered delays. Still, rollover ratios re-mained in the neighborhood of 90–95 percent,but borrowing rates widened substantially.

In addition to maintaining interbank lines,there were some efforts under way to find mech-anisms to maintain trade credit and derivativesexposure. However, rolling over trade creditswas deemed counterproductive, as it was likelyto delay an export-led recovery, and the mainte-nance of derivatives exposure was later droppedas OTC derivatives and other potential “blackholes” proved to be of less importance thananticipated.

Toward the end of January and beginning ofFebruary, signs emerged that the sharp depreci-ation in the Korean won had led to a rapid turn-around in Korea’s trade balance, stabilizing themacro situation and allowing Korea and its pri-

Box 5.2. Major Interbank Rollovers to Date1

1Main sources for this box, if not otherwise men-tioned, are IMF (1998b, 1999a, 1999b), plus variouspress releases, Bloomberg wires, and articles fromInternational Financing Review, Euroweek, and Euromoney.

2To improve the success of the rollover agreements,one country’s banks had to be assured that their for-bearance by maintaining their credit lines was beingmatched by other countries’ banks as well. The moni-toring system allowed for daily reports of renewal andmaturities loan-by-loan and bank-by-bank (see IMF,1999a).

INTERBANK ROLLOVERS

131

vate creditors to discuss a longer-term solutionfor roughly $24 billion worth of debt. After sev-eral proposals were considered, it was agreedthat interbank loans would be restructured intoone-, two-, or three-year loans paying LIBOR +225 basis points, 250 basis points, or 275 basispoints, respectively, and that these loans wouldenjoy a sovereign guarantee while keeping theborrower of record intact. The final agreementalso included call options for the debtors, allow-ing them to prepay the longer maturity debt ifthey so wished at a predetermined price.

Indonesia

While the sources of the immediate pressureon Korean foreign exchange reserves were fewand relatively clear, Indonesia provided a com-pletely different challenge. Starting inNovember 1997, the source of pressure on for-eign exchange reserves stemmed from broadlybased bank runs and domestic capital flight,causing Indonesian corporates to default ontheir foreign currency loans to internationalbanks as the Indonesian rupiah plummeted.Only a fairly small portion of external debt wasinterbank debt3 and hence the subsequentrollover exercise focused on maintaining tradelines while a de facto standstill occurred vis-à-viscorporate debtors, leading to a buildup of ar-rears (forcing international banks’ exposure tothese corporates to be maintained). Despite sub-stantial differences in country exposures acrosscreditor banks, which were estimated at $23 bil-lion for Japan, $5.6 billion for Germany, $4.8billion for France, and $4.6 billion for theUnited States (see BIS, 1998), no clear leader-ship emerged among the commercial banks asJapanese banks were seen as failing to take aleading role. Moreover, a sovereign guaranteewas not forthcoming on corporate debt.

At end-January 1998, the Indonesian authori-ties declared a “pause” for private sector foreign

currency–denominated debt service, while simul-taneously guaranteeing the obligations ofIndonesian commercial banks. This sanctionedthe de facto standstill, but may, as was argued bysome observers, have aggravated creditor rela-tions and delayed the formation of any coordi-nating mechanisms among commercial banksuntil the “pause” ended. A bank steering com-mittee was eventually set up and agreement on ascheme to handle corporate debt was finallyreached in June. While the main challenge wasto set up a private sector debt restructuringscheme backed by a preferential government for-eign currency guarantee, creditor banks alsoagreed to maintain their trade and interbanklines. The latter agreements had to be reachedbefore the clearance of all corporate arrears sothat the clearing of the arrears did not spark an-other round of pressure on the currency. A debt-monitoring system, which was put into operationin March 1998, again played an important rolein ensuring that the agreed-to international bankexposure to Indonesia was maintained.

Brazil

After managing to steer through the conta-gion from the Russian default in August 1998, inspite of an attack on its currency, and with thehelp of the IMF program, Brazil appeared capa-ble of preserving its currency regime. However,throughout September to early November, inter-national banks were actively cutting their expo-sures toward the country, reaching a cumulativereduction of $5.7 billion in the second half of1998. After a calmer period following the an-nouncement of the IMF-supported program,new pressures emerged in late December 1998,as the congress failed to pass key reform legisla-tion and investors increasingly questioned thecredibility of the exchange regime.

International banks with local operations wereespecially early and aggressive in cutting theirinterbank lines4 in the hope of avoiding both aforced rollover of their local operations’ hold-ings of domestic debt and interbank loans made

3Market estimates at the time suggested that 90 per-cent of Indonesia’s $65 billion external debt was owedby 1,000–1,500 companies, the remaining 10 percentreflected short-term interbank debt. 4See IMF (1999b).

specifics of the country as the key channels ofcapital flight will differ. In Korea, maturing inter-bank lines were seen by the official communityas the core issue that needed to be resolved toenable other crisis resolution mechanisms towork. In Indonesia, on the other hand, inter-bank lines played a much smaller role and werepart of a comprehensive rollover package, in-

cluding trade lines and corporate credits, sincethe resolution of one of the three credit sourceswas dependent on simultaneously achieving amaintenance of exposure for the others.Furthermore, in the case of Indonesia, privatecreditors were clear in expressing that theywould have shown little forbearance if they feltanother creditor group with a different or simi-


132

to their subsidiaries. As the foreign exchangemarket became increasingly one-sided and adiscrete devaluation became a near certainty, in-terbank rollover ratios fell as low as 20 percent.

Following the devaluation on January 12,1999, rollover ratios recovered to 60–70 percent,as a key macro uncertainty had been resolved.In the weeks leading up to the mid-March vol-untary agreement (formalized on March 31 andlaunched together with the strengthened IMFarrangement) between the international banksand the government to maintain exposures atend-February levels for six months, rollover ra-tios stabilized at a high level. After agreementwas reached, all maturing interbank debt wasregularly rolled over. In April, strong indicationsthat the Brazilian economy was on track for re-covery in the absence of a systemic banking cri-

sis in the aftermath of the devaluation, and asmost local commercial banks posted record firstquarter profits, some international banks startedto increase their exposures above what was re-quired under the agreement. The overall reduc-tion in international banks’ exposure towardBrazil reached $10.6 billion during the first halfof 1999,5 with the total exposure bottoming outat $50.8 billion for the banks domiciled in G-10countries.

In Brazil, a debt-monitoring system was alsoput in place to track the rollover ratios of inter-bank lines and later corporate credit lines. Thesystem played an important surveillance role forthe official community throughout the crisis andits resolution.

Box 5.2 (concluded)

Claims of Banks in BIS-Reporting Countries on Selected Emerging Market Countries(In billions of U.S. dollars)1

All BIS-Reporting United United Countries Japan Kingdom States Euro Area2 France Germany

KoreaJune 1997 103.4 23.7 6.1 10.0 30.3 10.1 10.8December 1997 94.2 20.3 6.9 9.5 26.8 11.1 9.6June 1998 72.4 18.9 5.6 7.4 22.5 7.9 8.4

IndonesiaJune 1997 58.7 23.2 4.3 4.6 17.9 4.8 5.6December 1997 58.4 22.0 4.5 4.9 18.8 4.8 6.2June 1998 50.3 19.0 4.0 3.2 18.0 4.0 5.9

BrazilJune 1998 84.6 5.2 5.8 16.8 37.6 7.9 12.8December 1998 73.3 4.2 6.5 12.7 36.7 6.1 11.3June 1999 62.3 3.1 4.2 12.8 30.2 5.8 9.5

Sources: BIS; and IMF staff calculations.1On-balance-sheet claims, excluding claims on offshore centers.2Data are for Austria, Belgium, Finland, France, Germany, Ireland, Italy, the Netherlands, and Spain. Data are not reported for Greece

and Portugal.

5See BIS (1999b).

lar type of claim would have used the resourcesgenerated by the initial creditor’s forbearance toexit. For example, since international banks defacto maintained their exposure to most corpo-rates through mounting arrears especially afterthe government sanctioned such an approach,they argued that they would not have agreed tothe “Frankfurt agreement” if it would have beenthe case that other creditors did not at least rollover their claims for some time period.

From an analytical perspective, the actual expe-rience in these crises as well as the views of mar-ket participants suggest the general conclusionthat a crisis resolution package needs to be com-prehensive in the forms of debt and capital flightthat it covers when there are several importantchannels of capital outflows. This would be thecase for most countries with an open capital ac-count. As the experiences with capital controlsshow, capital that wants to leave will do so unlessvery strict and comprehensive controls are put inplace. Therefore, in cases where creditor equalityconcerns are important and a broadly based hem-orrhaging of capital occurs, the rollover of inter-bank lines will have to form an important part ofany crisis resolution package together with otherinstruments to stabilize the crisis situation.

Were the Rollovers Voluntary?

Rollover experiences of interbank and tradecredit lines to date suggest that the rollovers in-ternational banks engaged in were more or lessvoluntary depending on the circumstances.Market participants argue that in Korea the offi-cial community’s approach was perceived asheavy handed in recognition of the fact that aKorean default had the potential of posing athreat to the global financial system. FollowingKorea, market participants have drawn the les-son that only if a country’s crisis is of a nonsys-temic nature, monetary authorities of the maincreditor banks’ home countries will not step inas urgently to encourage a rollover of their inter-national banks’ exposures. In this, these authori-ties face a trade-off between trying to help acountry in crisis and ensuring the credit quality

of their own banking systems. However, banks in-volved in the Korean rollover exercise did re-ceive several sweeteners to encourage theiragreement. All loans with an original maturity ofless than one year, about $24 billion, were givenan explicit sovereign guarantee and, hence, whatinitially was a claim of an international bank ona Korean bank became a claim on the sovereign.This sweetener that in the end, according tomost of the banks involved, clinched the dealmay not be available for other countries facing aKorean-style crisis for the following reasons:• Offering a sovereign guarantee may for most

countries not be seen as credible as their fiscalsituation and outstanding debt may already beunder the specter of default. In Korea, anOECD country, the explicit sovereign guaran-tee provided additional value in that it allowedthe international creditor banks to free up cap-ital set aside against the loans, as current Baslerules stipulate that no reserve capital be ap-plied to a loan extended to an OECD sover-eign. Market analysts argued at the time thatthis feature was especially attractive to Japanesebanks, which were trying to raise their capitaladequacy ratios. However, the drawback of ex-tending a sovereign guarantee included the pos-sible moral hazard implications for the Koreanbanking system and international banks, whichmainly interpreted the guarantee as a form ofgovernment bailout of their claims.

• If most of a country’s external debt stemsfrom corporate sector borrowing abroad, asovereign guarantee may not be attractive fora sovereign-in-crisis to extend. Most creditorbanks therefore did not expect a guarantee oftheir corporate Indonesian exposure.However, later in the crisis the obligations ofIndonesian banks were guaranteed by the gov-ernment, at the same time an overall “pause”in corporate debt service was in place as thepayment system was freezing up. At that time,neither interbank lines nor overnight creditwas available to most Indonesian banks.Before guarantees covering deposits in com-mercial bank and payments were extended,the central bank had to take an active role in

INTERBANK ROLLOVERS

133

the normal operations of the payments’ sys-tem to ensure that economic activity did notstop completely.A country in crisis may see a mandated

rollover of interbank claims as a last resort, asthe coercive nature of such an instrument wouldcause reputational damage going forward. InBrazil, the outflows generated by the cutting ofinterbank lines were only a relatively small partof total outflows at the time. According to mar-ket participants, the defense of the currency bythe sovereign aided by the official community re-sulted in most of the capital outflows. By thetime the currency peg was abandoned, it wasclear to all that the provisioning of fresh inter-bank lines and trade credits by internationalbanks would play a crucial part in an export-ledrecovery. In the market’s assessment, it wastherefore important for the authorities to ensurethat throughout the crisis these banks were notdiscouraged from supporting the future recov-ery, by not coercing them to maintain their ex-posure at the time. In the context of thestrengthened IMF program in March 1999, cred-itor banks participated in a truly voluntaryrollover. However, by that time the creditorbanks that had wished to exit had already cuttheir credit lines and rollover rates were basicallyall the way back to 100 percent as the Brazilianrecovery looked increasingly likely to take hold.

Extent of Private Sector Involvement

For the three recent major rollover experi-ences, the record with regards to their success insecuring private sector involvement and restor-ing external viability is seen by market partici-pants as being mixed.

Korea

In the Korean case, thanks to the rollover, theinternational banks stayed involved after beingconvinced it was counterproductive to all run forthe exits at the same time. Furthermore, Koreadid have the potential of being a good futurecustomer of the international banking commu-nity once external viability was restored and for-

bearance now could potentially lead to morebusiness later. While all Korean commercialbanks enjoyed a government guarantee, most in-ternational banks tried their best to “trade up”their exposure—that is, while maintaining theiroverall exposure, they tried to shift to betterquality banks. The few Korean banks that satis-fied the Basel minimum capital adequacy rules(a requirement under the IMF program) hadseveral banks offering them credit lines at rela-tively favorable spreads. However, weaker banks,despite the rollover agreement, suffered a liquid-ity shortage and had to continue to look to theKorean government for liquidity support.

In return for their forbearance until March 31,1998, international banks received a sovereignguarantee and their loans were rolled into one-to three-year claims and refinanced at rates thatwere about 150–200 basis points higher than pre-crisis lending rates. Given the quick export-ledrecovery of the Korean economy, the banks thatdid roll over their claims into these loans madegood profits. This meant that PSI was to some ex-tent maintained, but at a price in terms of theupgrade in the creditor. In retrospect, it is clearinternational banks’ initial exposures were ex-plicitly upgraded to a sovereign level and interna-tional banks did not suffer losses. At the time,however, the terms of refinancing were disputedamong private creditors as the quick economicrecovery was not obvious in January 1998 andmany international banks believed they wererolling over at below-market rates.

Indonesia

Of the three cases discussed in Box 5.2, theIndonesian case was not so much a question ofinternational banks being convinced of the ben-efits of maintaining their exposures, as this hadalready been de facto ensured by corporatedebtors running arrears with their creditors.Instead, the challenge was to handle a system-wide bankruptcy in the corporate sector andsubsequently in the financial sector, due to thecollapse of the rupiah. The Indonesian experi-ence of rollovers of interbank and trade lineswas seen by markets as having to do with facili-


134

tating crisis resolution, while burden-sharingagreements were much less of an issue.Following Indonesia’s Paris Club deal inSeptember 1998, due to comparability of treat-ment, additional private sector assistance to thesovereign itself took place through more tradi-tional means, such as the rescheduling, by ex-tending maturity, of a single $350 million com-mercial loan (led by Bank of Tokyo-Mitsubishi)at the previously contracted interest rate (seeStandard & Poor’s, 1999b).

Brazil

In the most recent use of rollovers to facilitatecrisis resolution, Brazil asked its creditor banksin March 1999 to maintain their interbank andtrade lines at end-February levels. As noted pre-viously, this was a voluntary agreement betweenBrazil and its major creditor banks. At the timeof the agreement, as discussed in Box 5.2, inter-national banks were already rolling over close to100 percent of their

private sector involvement in crisis prevention …following this introduction, the first section...

Documents