ww. r. LIBRARIES
-
DEWEY
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium
Member
Libraries
http://www.archive.org/details/emergingmarketchOOcaba
hU5 61
•
VE
WHI5
working paper
department
of economics
Emerging Market Crises: An Asset
Markets Perspective
Ricardo Caballero
October 1998
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
WORKING PAPER
DEPARTMENT
OF ECONOMICS
Emerging Market Crises: An Asset
Markets Perspective
Ricardo Caballero
No. 98-18
October 1998
MASSACHUSETTS
INSTITUTE OF
TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASS. 02142
LIBRARIES
Emerging Markets
An
Ricardo
Asset Markets Perspective
J.
-<
Crises:
Arvind Krishnamurthy*
Caballero
This draft: November
5,
1998
Abstract
Although internal policy mismanagements can be cited in most recent emerging
market
crises,
they seldom account fully for the severity of these
crises.
The
reluc-
tance of international investors to provide the resources that would limit the extent of
the reversal, almost invariably plays a key role in bringing a previously (over?)-heated
economy
to a costly halt. Domestic assets experience dramatic depreciation
and other-
wise solvent investment projects and production, especially in the nontradeables sector,
find
no
down
financiers
and are wastefully shutdown. Ultimately, the reason
of a country's access to international capital markets
(real or perceived) of its international collateral.
sufficiency
and
its
We
build
must he
for this break-
in the inadequacy
a framework where
this in-
consequences stem from microeconomic contractual problems. Fire
sales of domestic assets naturally arise as the result of desperate competition for scarce
why the private sector does not take
international collateral when crises are likely. The answer lies
international collateral. This begs the question of
steps to ensure sufficient
in the presence of a pecuniary externality.
We show that contractual problems also lead
to a problem of insufficient domestic collateral, which restricts the transfer of surplus
arising
from the use of international
international collateral.
also sheds light
there
is
The
on when
collateral
between the users and providers of
interaction between domestic
pre-crisis capital flows
Capital flows,
collateral
ought to be regulated and on whether
scope for currency support measures during the
Keywords:
and international
this
fire sales, financial
crisis
or not.
constraints, wasted collateral, ex-
cessive leverage, collateral underprovision, real depreciation, asset deflation, financial
intermediaries, stock market
"Respectively:
and currency support,
MTT, AEI and NBER; Northwestern
participants at the
IMF, MTT, Northwestern and the
suggestions. Ricardo Caballero thanks the
NSF
interest rate policy.
University.
EFG-NBER
for financial
We
thank V.V. Chari and seminar
(conference) for their
comments and
support and the Research Department at the
IMF, where part of this work was accomplished. E-mails: caball@mit.edu, a-krishnamurthydnwu.edu.
draft:
August 1998.
First
Summary
Motivation and
1
Asset markets view.
In contrast with the disagreements on the causes and appropriate
responses to the ongoing Asian
by international
role played
crisis,
there
is
substantial consensus on the mostly negative
capital markets during the crisis
The
itself.
reluctance of inter-
national investors to provide the resources needed to smooth the impact of a combination
of internal
and external
demand fell,
external
deficiencies, dramatically
external credit
—and capital flows
real sector, especially nontradeables, to
preciations
what was
a virtual
halt.
in general
1
— vanished, bringing the
Asset deflation and currency de-
—as symptoms and consequences— took unexpected proportions,
justified
light of the crisis."
is
"well
beyond
by any reasonable reassessment of economic fundamentals, even
in the
2
Although a very dramatic one
Asian crisis
worsened the situation. As internal and
just the
for its relative surprise
and extent of "contagion," the
most recent chapter of an increasing trend toward shifting the "blame"
from current to capital account
issues.
Many think that
this trend is
side effect of increasing globalization of capital markets.
ually tilting policy advice as well.
traditional policy advice
on the
While large
an almost unavoidable
Such change of emphasis
real devaluations
grad-
were central ingredients in
seems to be increasing
face of external imbalances, there
concern with the possibility that "too much of the medicine
is
may
kill
the patient."
Cur-
rency boards and interest rates hikes to shore up the currency in the midst of severe crises
are
common
advice these days, often even coming from the staunchest former advocates of
devaluations. Markets also have spoken loudly at times;
on Black Thursday (September
10,
1998) the Brazilian stock market plummeted by nearly 15 percent. In a desperate measure
to stop the hemorrhage of capital flows, the authorities raised interest rates from 30 to 50
percent...
and the stock market responded with a 13 percent rebound. 3 Largely due to the
central role played
different
by
capital flows, economics of emerging markets
is
sometimes radically
from that of developed economies.
Gradually, the traditional "goods markets" view of external crises and their solution
is
being replaced by an "assets markets" view.
imbalance
is
no longer to simply devalue the
to reallocate consumption
real
The
central issue in resolving
an external
exchange rate and allow the private sector
and production to restore
real activity
and external accounts.
Although certainly an important factor in the long run readjustment of an economy,
1
For the crisis-countries
this
—Indonesia, Korea, Malaysia, the Philippines and Thailand— net capital flows
declined from 73 billion dollars in 1996 to minus 11 billions in 1997.
The
actual decline in net inflows was
probably higher since "errors and omissions" went from -8.5 billions in 1996 to -19.5 billions in 1997.
2
World Economic Outlook,
3
The
May
1998, p.4.
leading international article in the Wall Street Journal
Big in Winning Back Foreigners."
on September 11 was
entitled: "Brazil
Bets
relative price ingredient
subordinate to the short run movements in asset prices and their
is
feedback into real activity - a link which, in our view,
magnitude of
and our
crises
In this paper
Goals.
ability to control
we develop a framework where
answer many of the questions surrounding
is it
them.
is
crises in
at the core of understanding the
4
asset
during
these affect real activity?
why
crises,
why
emerging economies. In particular,
Why
that foreigners are unwilling to finance these crises?
How do
markets take center stage to
do
assets' fire sales occur?
providing the financing
If foreigners are reluctant in
doesn't the domestic private sector take steps to ensure profitability
during these times by provisioning against them?
from the point of view of crisis-prevention?
Is
Should currency and assets be supported during
The inessential and
there a role for regulating capital flows?
crises?
Our framework
the essential.
the productive structure appropriate
Is
reflects
our view that a core component
of the answer to these central questions does not depend on the intertemporal substitution ingredients
economies
—
i.e.
we normally emphasize
in macroeconomics.
The very
fact that emerging
— normally run
economies whose future looks brighter than their present
very small current account deficits
—
to 10 percent of
in the order of
GDP,
rather than
— indicates that they are nor-
150 percent or more, as they should in a neoclassical setup
mally severely financially constrained.
Deep
crises are just times
are overwhelming. Instead, our answers to the questions
when such
constraints
we posed above build around the
simple and sufficient premise that, ultimately, the reason for the breakdown of a country's
access to international capital markets,
must
lie
in the inadequacy (real or perceived) of its
international collateral.
But what
Collateral.
is
international collateral?
can be valued and seized by foreign investors in
sical as well as
the Sovereign Debt literatures
At a basic
lieu of
The
literatures
on debt
crises,
anything that
promised repayments.
The Neoclas-
make some proportion of the present value
of net exports (or exports alone) the central answer. 5 Although
4
level, it is
it
highlights
a fundamen-
currency speculation, and bank-runs, offer important insights on different
aspects of assets markets during crises. Most of the effects in these literatures are either amplified or
more
likely
by the
(1994, 1996);
issues
on debt
we emphasize
crises see, e.g.,
in this paper.
For papers on currency speculation see,
e.g.,
made
Obstfeld
Calvo (1988), Giavazzi and Pagano (1990), Cole and Kehoe (1996);
on bank runs see the canonical Diamond and Dybvig (1983) and' the applications to open economies by
Goldfajn and Valdes (1998) and Chang and Velasco (1998a). Also see Calvo (1998) for an overview and
interesting discussion
on implications and sources of
worth pointing out that, while
insightful
possibility of multiple equilibria,
the
5
crisis
must be
left
and
unexplained
capital
and probably very
market
relevant,
crises in
small open economies.
a central issue in these papers
in so doing this literature admits from the outset that
(i.e.
a
It is
is
the
large part of
to "sunspots").
See Bulow and Rogoff (1989) for a canonical model of international debt capacity based on the amount
of "punishable" future trade flows
and exports.
tal constraint, this
On
scenarios.
answer
one hand,
is
among very
incomplete
central dimensions of
modern
crises
does not capture the idea that an excessive devaluation
it
be harmful for international
collateral.
6
On
may
the other, in today's emerging economies the
private sector carries the lion's share of international borrowing, thus microeconomic rather
than representative-agent
frictions
ought to be at the roots of the problem. Our model gives
market constraints arising from microeconomic contractual problems
central roles to credit
in the process of creation
and aggregation of international
Domestically, collateral
is
created by the corporate sector
claims on firms in the nontradeable
—
understood
limit the
in distress.
The
and tradeable
number of claims that
to obtain external financing
situation
is
when
in need,
by international
International collateral
and shares on
sector, while claims
assumption captures
by Mexico when their
liquidity package it received
investors (e.g.
6
limits
on
is
made of financial
problems
—broadly
firms can issue, affecting both their ability
collateral to other firms
collateral.
when not
8
In most of our analysis
the tradeable sector.
international collateral
is
on both sectors count as domestic
Very
realistic biases.
directly,
that cash revenues from exports can be seized before they
feature used
and
itself,
sectors. Contractual
and to supply
make a simple broadbrush assumption that
stylized, this
7
worse with respect to foreigners, which only accept a subset of
these claims as backed
on the tradeable
collateral.
oil
made
revenues were
during the 94-95
crisis,
real estate investments).
make
it is
it
we
restricted to claims
collateral.
justified
Although
by the
fact
back into the country, a
part of the collateral backing the
and by mandate on foreign institutional
9
More
indirectly, this
assumption
is
a
Quite the contrary, an excessive (real) depreciation should build more international collateral and further
entice foreign investors to finance the crisis.
Of
course,
if
concerned with this specific issue alone, one could enrich the story by adding a financial
sector which
is
over-exposed to the nontradeable sector, and plays the role of an input into production of
tradeables.
Still,
one needs to address the question of why banks are over-exposed to begin with, and how
an excessive depreciation
7
According to the
is
IMF
an equilibrium outcome.
(1998, pp.
63-64),
a distinguishing
purely macroeconomic considerations has played
et
al.
a
lesser role
characteristic of the current crisis is that
than in previous
crises.
Also, see Johnson,
(1998) for stark evidence on the correlation between weak corporate governance and institutional
environment
(i.e.
a weak contractual environment)
in
an economy and the incidence of the current
crisis.
As the corporate finance literature has demonstrated, weak corporate governance and, more generally, weak
hand in hand with credit market
an extensive literature documenting "home bias"
contractual enforcement go
8
There
1991),
is
matched by asymmetric
reasons behind
home
bias, as
constraints.
in asset holdings (see, e.g.,
information explanations (see e.g.
most countries
Zhou
French and Poterba
1998). There are also institutional
limit foreign ownership of domestic companies.
range from constraints on a few "strategic" sectors
(e.g.
oil
and telecommunications), as
These
limits
in the U.S., to
across-the-board constraints, as in pre-crisis Korea.
9
Of course there are important exceptions
(1997) for systematic
(e.g.
"too big to
Japanese evidence showing that,
fail" utilities),
but
see, e.g.,
for small firms, those that are
Kang and
Stulz
export oriented are
very efficient mechanism to capture the central implications of an expanded model where
foreigners' collateral-bias
is
only against small firms and current production of perishable
nontradeable goods (see the appendix).
Financial Transactions.
Financial claims in our economy consist of equity shares issued
by tradeable and nontradeable
be traded
may
freely.
10
own
issue its
Thus, a firm in the nontradeable sector in need of funds, for example,
may have already owned,
it
nontradeables firms' shares in
reshuffling of collateral
its
as well as the proceeds
is
assumed to be
(firms) are subject to aggregate
which accumulate to determine a country's net import needs.
— such needs,
these,
from
possession, as international collateral. This
through stock market transactions
Domestic entrepreneurs
Shocks.
and use
shares in exchange for shares in the tradeable sector
together with any tradeable shares
selling other
Stock markets function perfectly, so these shares can
firms.
costless.
11
and idiosyncratic shocks,
International funds
must
finance
—
lateral.
Although the mechanisms we emphasize magnify the impact of aggregate surprises
—be
a decline in terms of trade, in foreigners'
it
in full or part
ness, or in
banking capital,
surprises are distinct
etc.
brought about by these
break down. First, a country
may fail to
aggregate international collateral properly;
by shocks have severe agency problems and/or scarce
collateral provisions,
be constrained by domestic collateral well before the country has used up
favored by foreign investors. See, e.g. Blommestein (1997), for a discussion of
assets considered highly illiquid or
by foreign
institutional investors.
10
for
how
may
if
firms
they
may
all its
real estate
available
and Other
exposed to exchange risk are generally avoided (sometimes by mandate)
Interestingly, the very
few exceptions to the sovereignty principle, by
which the rating on debt issued by a country's corporate sector
government debt rating, are
we
until the last section of the paper.
In our framework, there are two stages at which the "credit chain"
Bottlenecks.
hit
issues
In order to highlight these differences,
insights.
col-
"risk appetite," in external competitive-
— the hedging and insurance
from our main
assume no aggregate surprises
which foreign investors demand international
for
is
bounded from above by that country's
companies which belong to the export
In this sense, individual foreigners
may
sector.
hold shares on firms in the nontradeables sector, but the total
value of foreigners' holding of shares in tradeables
and nontradeables must not exceed the
total value of the
pledgeable shares in the tradeable sector.
Some words
a
far cry
of interpretation are in order.
Our
from the reality of emerging markets.
usual form of financing in most economies.
focus purely
First,
on
frictionless equity
Qualitatively,
is
like.
is
not an interesting
Second, the reshuffling of collateral in an emerging
more often done by the banking system than
directly
by
firms' trades in
adding richness by modeling the role of intermediaries in these transactions
these transactions are costless in our model, which
implements them.
certainly
debt rather than equity contracts would aggravate many of the effects we describe
by introducing debt overhang problems and the
economy
is
and with a few caveats that are not central
to our story, whether external financing takes the form of equity purchase or debt
distinction. Quantitatively,
markets
debt contracts rather than equity are the more
means that
it is
a stock market. Again,
is
not central to our insights. All
irrelevant
whether a creditor or borrower
international collateral. 12 Second, even
if
aggregation can be done
the international
fully,
collateral creating sector
may
not be able to generate enough pledgeable shares to finance
the
more
likely as contractual
crisis; this
scenario
is
problems in this sector
rise.
We refer to
these as domestic and international collateral problems, respectively. Given institutions
contractual problems, which constraints the
aggregate shocks. For mild shocks
the domestic constraint
one
likely to
is
become
is
economy encounters depends on the
—perhaps "normal" times
likely to bind, while
and
severity of
an emerging economy
for
during deeper recessions the international
active as well.
the latter scenario, with the simultaneously binding constraints, that accounts for
It is
the most dramatic differences between crises in developing and developed economies, and
in the appropriate policy response to these crises.
Equilibrium Prices and Fire Sales
falls
Once
one region or the other depends on the relative price of domestic and interna-
into
tional collateral.
Other things equal, as the relative price of domestic collateral
likelihood of wasting international collateral
equilibrium price depends to a large extent
As the
other.
whether the economy
in a costly crisis scenario,
falls.
On
is
in
one region or the
becomes more binding, the
relative price of
domestic collateral declines rapidly as competition for scarce international collateral
Indeed,
we show
that a fire sale of these shares, by which
below their fundamental value,
constraint.
13
is
the
the other hand, the behavior of this
on whether the economy
international collateral constraint
rises,
we mean a
rises.
decline of their price
a natural consequence of a binding international
collateral
In our model, this sharp decline in the domestic share prices can be thought of
as the combination an "excessive" depreciation of the real exchange rate
and a substantial
widening in the domestic-international interest rates spread.
On
Collateral underprovision.
the face of these potential
fire sales,
the question arises as
to whether they are not sufficient incentive for the private sector to reallocate factors of pro-
duction to collateral creating sectors to benefit from the additional international collateral
value of their shares.
and increase
The
its
We find
that indeed the private sector will value collateral provision
supply, but generally
it will
not do as much of
it
as
is
socially optimal.
source of the externality behind the undervaluation of collateral provision
ing in
itself,
for
it is
is
interest-
the result of the interaction between the domestic and international
credit constraints. Despite fire sales, the domestic credit constraint limits the transfer that
l2
This
equivalent to the "wasted liquidity" in
is
collateral"
Holmstrom and
Tirole (1998), although oui "wasted
has implications for pre-crises policies not present in theirs, which stem from the two-goods
nature of our economy.
l3
This definition of a fire sale
is collateral
assets as
constrained, prices
we assume that
is
similar to that of Shleifer
must
fall.
and Vishny
(1993).
When the domestic economy
However, unlike Shleifer- Vishny there
foreigners will never hold claims
on the nontradeable
is
no outside bidder
firms.
for the
firms in distress can
constrained
demand
make
The
to suppliers of international collateral.
anticipation of this
for international collateral reduces the incentive to create
Contrac-
it.
tual problems in collateral providers generate the fire sale, while contractual problems in
distressed firms act as a disincentive
is
collateral provision.
Even
Overinvestment in nontradeables.
as
on
additional international collateral
if
the case in the milder wasted collateral region,
benefit from reallocating resources
reallocation
we argue
that the
is
economy would
from the nontradeable to the tradeable
would appreciate the long run
exchange
real
not needed,
Such
sector.
rate, boosting the collateral value
of nontradeables shares during the crisis period. Individual investors in the nontradeables
sector do not internalize the negative effect of their investment
These externalities
Capital flows.
may
economies predictably run into systemic
pre-crisis capital inflows
this question
or in the
since individuals
The
crises.
can exacerbate these
fire sale
is
more
We
fire sale
of insufficient international collateral,
all
Policy.
wasted
collateral
and
region,
this
is
collateral aggregation value of
a
real
on the other hand, the problem
is
one
aggravated by pre-crisis capital inflows.
where domestic and foreign return
when this differential is large. The standard neoclassical
is
a natural obstacle to
differential is low, it is
not
arbitrage of relative productivity
overwhelms the collateral provision enterprise, making severe crises and
the more
whether
show that the answer to
likely to fall in the
Although the equilibrium domestic collateral provision premium
differential
is
region. In the former, equilibrium capital flows are insufficient
exchange rate appreciation. In the
so
why emerging
natural question to follow
externalities.
do not take into consideration the
capital inflows in situations
collateral value.
harbor some of the answers on
depends on whether the economy
more severe
on others'
fire sales
likely.
Since the outcomes of our simple structure are certainly rerniniscent of the events
and dilemmas faced by emerging economies,
of discussing policy implications.
of insurance contracts;
we
Many
it is difficult
to resist the naive temptation
of the natural policies in this setup take the form
relegate these to a secondary role because they often have high
informational and enforcement requirements which are not likely to hold well in emerging
economies.
Rather,
we
focus our attention on direct intervention in asset
markets, without attempting to differentiate
among
and exchange
asset holders within each type of asset.
Perhaps the clearest implication of our model in terms of policy
is
during the pre-crisis
period where subsidizing domestic investment in sectors that create international collateral
should be contemplated whenever the likelihood of falling into the wasted collateral or
fire sale
region
is significant.
Advice regarding capital flows
or taxing pre-crisis capital flows depends
international collateral
is
more
likely to
is
more ambiguous;
fostering
on whether the aggregation or the amount of
be the main problem in the event of a
crisis.
What can be done
during crises exacerbated by collateral constraints?
price of domestic shares, or lowering domestic interest rates,
problem
primarily one of insufficient domestic collateral.
is
may be
sures
transfers
if
In the
one of
is
Boosting the
in principle useful
But the
costs of these
high, since arbitrage conditions require persistent interventions
the exchange rate
fire-sales region
is
when
and
the
mea-
implicit
to be maintained.
the story
is
quite different.
insufficient international collateral,
Since the fundamental problem
is
boosting the price of domestic shares without
supplying additional international collateral or resources does not have any positive impact
by
Supplemented and financed with an increase in domestic
itself.
this
is
what
is
real interest rates
required for the government to attract foreign funds at the margin
—
if
— may
instead have favorable effects. 14
The
Collateral in macroeconomics.
international economics literature has long recog-
nized the importance of international collateral and
sector.
15
The Latin American debt
its relation
with a country's tradeable
of the 1980s led to the development of formal
crisis
models of sovereign debt renegotiation, and to the closely related question of what
is it
that
international lenders can threaten sovereign countries with in the event of a default (see
Eaton and Gersovitz
its
costs
is
(1981),
Bulow and Rogoff
and
mostly modeled as a country-wide phenomenon, the question immediately arises
on whether the domestic private
hood
(1989)). Since in this literature default
Bardhan
of these events.
sector internalizes the effect of their actions
(1967), Harberger (1985)
on the
likeli-
and Aizenman (1989) advocated
taxing capital flows on the grounds that individual agents do not pay for the increase in
a country's average cost of capital brought about by their international indebtedness. Our
results
on
capital flows
and
collateral underprovision
of the circumstances under which this problem
ommendation
can be seen as a formal description
is likely
to arise. Similarly, our policy rec-
in favor of subsidizing the tradable sector to increase collateral availability
is
akin to Borensztein and Ghosh's (1989) advocacy for subsidizing investment in the export
sector.
Although we touch on many of the insights of the
eling
and conceptual
closely related to the
differences with
new
literature
and Moore (1997) place the
it
are large.
on the
On
debt-crisis literature,
both our mod-
these aspects, our approach
is
more
role of collateral in macroeconomics. Kiyotaki
collateral role of
a fixed input
—land— and
its
distribution
across agents at the center of their amplification mechanism. Krishnamurthy (1998) shows
14
Generally, these policies have perverse effects
on the
incentives for collateral provision during normal
Krugman
(1998) for an
times. These should be offset with
some form of
account of the recent Asian
terms of a model that highlights the absence of such palliatives, together
crisis in
collateral creation incentive. See
with uncertainty about the government's willingness to provide "insurance."
l5
See, e.g.,
Simonsen (1985).
how a
richer set of contracts
than that allowed by Kiyotaki and Moore
shifts
the relevant
constraint from that which applies to individual agents to the total availability of collateral in the
economy. Holmstrom and Tirole (1998) argue that the stock market alone
provide insufficient collective collateral
if
not aggregated properly.
even when wasted corporate collateral (hquidity)
be required
in the presence of large shocks.
modeling aspects with each of these
with them as a group
is
not an
Although our
articles are plenty,
They
issue, public
also
and an international one
—
-
and
show that
supply of
it
differences in terms of goals
may
and
our central theoretical departure
our focus on the presence of two collateral constraints
is
may
their rich static as well as
dynamic
—a domestic
interactions.
Certain aspects of our model are similar in structure to models of liquidity in the banking literature
(Diamond and Dybvig
(1983)). In this light, underprovision of collateral
is
analogous to the free rider problems and underprovision of liquidity emphasized in that
literature (see Jacklin (1987),
Bhattacharya and Gale (1987)). However, the source of the
externality behind this inefficiency in our
tract laws
liquidity
and
institutions, while theirs is the
A second
by consumers.
believe that
bank
failures
is
the endogenous result of imperfect con-
exogenous divergent ex-post valuation of
departure from these models
rather than on banking arrangements
we do
model
and
their
fire sale
in our focus
on markets
key sequential servicing constraint. While
and runs can aggravate
the basic features of crises and their prelude arise even
Moreover, the
is
crises,
when
our model points out that
there are only stock markets.
mechanism we emphasize boosts the
likelihood of banking crises
themselves (see the appendix).
Outline.
the
Section 2 follows this introduction with a description of the model, starting from
crisis period.
It
analyzes and characterizes the critical wasted collateral
regions. It continues with
as well as with
an
and
fire sale
a discussion of the dangers of a weak domestic banking system,
illustration of the fragility
brought about by excessive leverage. Section
3 takes a step back and analyses the history preceding
crises.
In particular,
it
illustrates
the basic externalities leading to private undervaluation of collateral provision investment
and to overinvestment
by
in nontradeables. It
pre-crises capital flows. Section
4 discusses insurance markets and the shock structure.
Section 5 discusses policy implications
2
The
2.1
more generally and
concludes.
An
appendix
follows.
Phase
Basic Setup
Overview.
neurs
Crisis
then shows how these externalities are affected
who
We model the core of our economy as comprised by ex-ante identical entrepreuse their resources to begin companies in a tradeable and
a nontradeable
sector,
and to buy shares of other companies. Contractual problems
neurs' leverage, so that they
must retain a
fraction of their
limit the extent of entrepre-
own
shares. Entrepreneurs
do
not value diversification per-se, as they are risk neutral with respect to idiosyncratic shocks;
however they have a strong incentive to hold others' shares as
ing an idiosyncratic
collateral in the event of fac-
In equilibrium, residents are indifferent on whether to hold this
crisis.
form of shares on firms on the tradeable or nontradeable sectors (T- and
collateral in the
To start, we assume
N-shares, respectively).
N-shares as individual
collateral,
foreign investors are not; although they accept
they promptly swap these for T-shares in the market. 16
In emerging economies, marginal investment (and consumption)
We
capture this feature in a simple environment.
demand
their net
for tradeable goods.
Firms are
At the aggregate
existing shareholders. If the firm in need
to
be exchanged
collateral at its avail,
which also need to be swapped
very similar, although
the fully flexible period
structure,
and
fina.nr.ing
own
it
of which
The problem
economy
governance problems are
may be
may be
obtained by diluting
17
in the
new
shares have
The
rest has to
form of N-shares
of a firm in the tradeable sector
Time
is
indexed as
t
—
0,1,2.
Date
is
agents design the productive structure, ownership
Date
1
is
the
crisis period.
Date 2 represents the future
summarizes the cost of the crisis and realizes budget constraints.
issues are distinct
from the
chronology, and discuss crisis issues here taking history and
we
raise
issuance can be used directly as international collateral.
when economic
period;
some
lasts three periods.
interrelated, the date
the next section
If
in the nontradeables sector, its
for T-shares.
portfolio allocations.
and "accounting"
Although
its
The world
Basic setup.
by shocks which
for T-shares to serve the role of international collateral.
be financed with the
is
is
financed externally.
these shocks require external
level,
financing, for which foreigners require T-shares as collateral.
not too severe, a substantial fraction of the required
hit
is
discuss history formation.
crisis issues.
its
We
reverse the
"mistakes" as given.
In
Thus, for the purpose of this section, our
starts at date 1.
The domestic economy
is
populated by a continuum of unit measure of agents with
Cobb-Douglas preferences over date 2 consumption,
Where,
q
is
consumption of tradeable (T) goods at date 2 and c„
tradeables (N) at date
as e.
16
2.
2 consumption only,
of the roles the domestic financial sector plays, in our view,
is
we remove from our
Whether the borrower or
lender implements this transaction
10
analysis
in evening this asymmetry.
return to this issue below.
17
consumption of non-
We denote the date 2 exchange rate of T-goods in units of N-goods
By specifying preferences over date
One
is
is irrelevant.
We will
standard intertemporal substitution and smoothing arguments which are not central to our
approach.
On the production side,
how much
care
output at date 2
is
given to the existing productive structure at date
is
the total amount of capital devoted to sectors
from date
0.
the result of investments
At date
1,
production in a fraction p of the firms
goods at date
imports.
20
output of
Rnln
1 to contribute
19
at date 2.
at date
<
is
and
Let kn and kt denote
1.
N and T at the beginning of date
the N-sector, firms are required to contribute In
in order to realize
made
inherited
1,
interrupted by a shock. 18 In
kn additional units of the tradable good
The domestic economy
will
have no tradable
and the shock can only be accommodated by an increase
in
Besides heterogeneity, the two essential ingredients of our modeling of the shock
are an import need coupled with a margin
on which to adjust production. Because having
a second reinvestment opportunity does not add substantial additional insights,
we model
the shocks to the T-sector more simply as just resulting in lost output; firms that are hit
with a shock lose
by the shock
is
all
output at date
2.
Production per unit of capital in T-firms not affected
Rt.
Firms' investments in the two technologies are held as N- and T-shares. These shares can
be bought and sold at
all
dates subject to one important constraint. Because of an agency
problem (contractual problem) between an entrepreneur/firm and
each firm must hold at least a fraction 6n and
only the complements, 1
— 6,
6t
respectively of its
#o,t
As we
worth to
be the fractions that each firm holds in
These fractions are the
outside claimants,
own
shares.
21
Thus,
can be transferred within the agents in the economy. This
fraction constitutes domestic collateral. It has equal
Let #o,n and
its
result of date
decisions
and
its
all
agents in the economy.
own
technologies at date
1.
constraints.
said, foreign investors are willing to receive N-shares in
exchange for funds but
transform them immediately into T-shares. Thus international collateral constitutes only
the T-shares. These investors are risk neutral and have preferences only over the consump18
We
assume that the identity of firms receiving
this idiosyncratic
shock
is
not
verifiable.
This rules out
insurance markets against the shock. See section 4 for details.
19
Production has a time-to-build aspect. Investments are
possibility of scrapping the project or investing
made
at date 0; date 1 leaves firms with the
more resources in order to
realize
some of the output at date
2.
20
In the appendix we introduce production and consumption at date
with this extension and
obtain our
we can drop
fire sale results (see
1.
Our conclusions remain unchanged
the assumption that foreigners do not accept N-shares as collateral to
below).
The formulae
is
obviously more cumbersome, thus
we
leave
it
for the
appendix.
21
We
appeal to agency conflicts in making this "high-level" assumption. See the appendix for a deeper
justification of the assumption. Also, see
La Porta et
correlation between corporate governance problems
higher concentration
is
al.
(1998) for worldwide evidence
and ownership concentration
needed to prevent managers from
11
stealing).
on the high
positive
(in their interpretation,
tion of T-goods. Unlike domestics, foreign investors have preferences over
all
may be
dates because they
called
endowments of T-goods at
large
on to supply tradeable goods at date
count the future so that the preference structure pins
at zero. For now,
we assume foreign
investors have
They have
1.
we assume that they do not
dates. Additionally,
all
consumption at
down
dis-
the international interest rate
no positions
in
terms of domestic shares
at the beginning of the crisis period.
At the beginning
Initial conditions.
technologies, while the remaining 1
of date 1 firms hold #o,n shares in their
— #o,n
p of the N-firms are hit with a shock at date
replenish production
each firm that
is
up
to In
<
k„,
left
of the option to replenish production.
hands of the manager of the
their shares are worthless.
1 that leaves
— #n )(l — p)Rn,
Firms hold
We
assume that
of this at date
all
this option value rests purely in the
firm. Outside shareholders are fully diluted
shares in their
1.
shareholders of
upon a shock and
22
or each share repays (1
#o,t
The
to
with no value in terms of goods but value in terms
Given the shock, the shares on the market of the N-sector
(1
fraction
them with the opportunity
with an input of In tradable goods.
with a shock are
hit
A
held by other firms (the market).
is
own N-
own
The remaining
1
aggregate
will payoff in
— p)Rn.
Firms that receive the shock lose
T-technologies.
— #o,t
held in the market and repays per share of
is
(i-p)RtThus the
by a shock.
(1
total resources available to
A distressed firm
— p)(l — 6o,t)kt
T-shares at date
(D)
is left
a firm at date
with (1
shares of the T-sector. Let
1.
Then the date
1
depend on whether or not
— p)(l — #o,„)fcn
shares of the N-sector
a£
= (1 - p)(l -
o ,n)
1 collateral (in T-shares) of such
Decisions and constraints.
and wf
= (1 - p)(l - 60it
—
of shares to raise (1
— 9)In P
<
fcn-
Then
a
firm,
wd
,
is
(1)
).
it
can issue against this a fraction
shares of T-sector. However this share issue
to the constraint that the firm retains at least
>6n
of the firm also include its shareholdings in other
"Some
it
a
is
subject
fraction of the firm because of the
agency problem. In addition to the proceeds from the share
to convert a fraction
and
A firm in distress needs to raise funds to finance continuation.
Suppose the firm chooses to finance In
1
hit
P be the price of an N-share in units of the
w^KPuZ + ktcof
where,
it is
issue,
the available resources
N- and T-firms. Suppose that
it
chooses
of the N-shares of this holdings into T-shares so that these also
degree of dilution of external claims'
is
obviously
realistic.
We
have adopted an extreme form of
to exclude "debt overhang" problems from our discussion. In any event, expected dilution will be built
into prices at date 0.
12
constitute resources for reinvestment. Then, the total resources of the firm at date 1
The
this
firm chooses to sink In from these resources back into production.
reinvestment
first
is
on
= OlnRn + c(((l - 0)In P + OcknPui + k,f4)Rt - In)
term represents the
second term
profit
is,
7T
The
The
is,
fraction, 0, of
date 2 output that the firm retains.
the amount of resources that the firm
is
left
with after contributing In
The
firm maximizes this profit subject to budget
einRn
+ e(((l - d)In P + aKP^i + k u>f)R - /„)
towards saving production.
The
and
collateral
constraints,
max
S.t.
t
t
(1)
((l-eVnP + aknPcjZ + ktvfiRt-In^O
(2)
e
>
(3)
e
<i
(4)
/„
(5)
a<
en
<
K
1
Forming the Lagrangian,
l = einRn + ie+xMai-e^p+w^Rt-ij + Mie-On)-
M9 - 1) - Hln -K)- A
where the A's are multipliers on each of the
Let In , 6 and
by these firms
is
a
be solutions to
p((7n (l
5 (a
- 1)
five constraints
this problem.
Then the
the agency problem
of N-shares to
is
buy to
is
total
all
greater than zero.
number of N-shares sold
— 6) + cuJ^kn).
A firm that does not receive a shock is sound (S)
these N-shares. This
and are
paid for with
its
at date 1
and may choose to purchase
available shares of the T-sector,
the fraction (1— 6t) of its T-share holdings.
It
solve,
max
s.t.
e(Rtkt
- XPRt) + XRn
xpKktVo.t-et+ii-eojO.-p))
x>o
13
which because of
chooses
X, the number
Let
X{P) be
the solution to this program.
Equilibrium.
In equilibrium, the
number of N-shares
purchased by the sound firms. Thus, market clearing
(1
This condition gives the date
Date 2 exchange rate and
date
sold by the distressed firms
must be
is
- p)X{P) = p((l - B)In + auZkn)
1
equilibrium price, P, of an N-share in units of the T-shares.
collateral
premium.
Consider
now
the spot exchange rate at
Given the Cobb-Douglas preference structure, the exchange rate will only depend
2.
on aggregate consumption. The
first
order condition for a representative consumer yields,
Cn
Given /„,£„,&*,
= (l-p)Rn kn + pRn In
C„
C = (l-p)R kt-pIn
t
As we know,
of
T
t
P is the price to exchange claims on R„,
N for
units of
at date 2. Consider however the simple exchange at date 2 of
R units
of N for Rt
claims on
Rn goods
t
goods of T. Let the price of this exchange be P, which we refer to as the fundamental price.
Then,
P=
When
P y£ P,
there
is
a
collateral
—
eRt
premium governing exchange
at date
1.
Let
L
be
this
premium,
-I
we show
The
to the
in the next section that equilibrium will require that
strict inequality
rises
collateral.
Equilibrium requires that the return on holding
above that of holding international
collateral.
interpreted exclusively as a decline in the relative price of N-shares.
a
[fictional]
This need not be
To see
this, let
us define
date 1 exchange rate, ci, as the price to exchange 1 unit of the T-good for e\
units of the N-goods at date 1.
1 holding
1.
P < P holds if there is scarcity of international collateral relative
amount of domestic
domestic collateral
L>
Now
consider the investment decision of an agent at date
one unit of T-good and choosing between investing this in the domestic or the
international economy. Investing
the international interest rate
is
abroad just means buying international
collateral. Since
pinned down to be one, the return on this investment
14
is
one
as well. Investing domestically
of L.
The domestic
and an investment
means purchasing domestic
collateral,
which yields a return
transaction can also be implemented via an exchange rate transaction
at the domestic interest rate.
units of N-goods at date
The agent can
convert the T-good into e\
which can then be invested domestically at an interest rate of
1,
Rd, yielding Rd&\ N-goods at date
The
2.
return on this investment must be equal to L, or,
rAe = l
Thus, when
L>
1,
the domestic economy
must
these resources, domestic investment
>
we
In the next section
Wasted
2.2
Preliminaries.
w
which
d
is
(1)
need of foreign resources. To attract
1,
come
or in the form of an expected real exchange
and equilibrium
characterize decisions
in
more
detail.
and Fire Sales
Return now to the optimization problem of the distressed
>
and
0)
(4)
does not (A4
firms. Analysis
We focus our discussion on the case
depend on which constraints bind.
binds (Ai
either in the
23
Collateral
of this problem will
in
l.
in great
yield a high return, which can
Rd >
form of a high domestic interest rate,
rate appreciation, ei/e
is
= 0).
Constraint (1) will not bind only
if
very large, so domestic firms are sufficiently well capitalized that they are financially
unconstrained.
markets
like
While perhaps useful in classifying countries with well developed capital
the US, this constraint
is likely
the focus of this paper. Constraint (4)
net-present-value (henceforth,
constraint only binds
N-sector and
is
if
NPV)
is
to bind in the emerging economies which are
similar, in that
projects that the
it
represents the
economy
is
able to undertake. This
the economy has sufficient resources to save
thereby able to finance
all
of its positive
Differentiating the Lagrangian with respect to
amount of positive
all
production in the
NPV projects.
In we obtain that constraint
,
(1)
binds
as long as,
6(Rn
-e)+ e(l - 0)(RtP - 1) >
Saving a unit in distress in the N-sector is a project that creates
at a cost e, the value of
Rn/e which we assume
that
Rn/e >
one unit of tradeable. The social return on
is
greater than
l.
24
We
will shortly
then Ai
> 0.
show that RtP
Consider next the derivative with respect to
^ = PR (e(L-l)-\
t
1
this project is therefore
>
1,
implying
1 is sufficient for (1) to always bind.
If (1) binds,
23
Rn units of nontradeables
The model does not pin down Rd and ei/e because
6,
1)
in our formulation there
is
no goods market at date
through which an exchange rate can be established. In the appendix we introduce this market.
See appendix for assumptions on primitives that yield this result.
15
Our concern
is
mostly with regions where this expression
straint (1) binds 'Voluntarily;"
will
not be willing to use up
this region at the
end of
if
is
negative, for otherwise con-
the equilibrium price of N-shares
is
too low, entrepreneurs
N-shares to finance reinvestment.
all their
this section; for
now,
In
let
6
= 6n
and
a=
1.
We
will return to
Thus,
= jR wd
t
where,
lf
-i-(i-en )PR >0
t
and
7-Rf
Wasted
is
the return on internal collateral of a firm that
Region
collateral.
1 is
is
hit
with a shock.
the case in which the domestic collateral constraint binds
while the international collateral constraint does not, which happens as long the following
inequality holds:
p P ((i - en 1 R w d + Kwi) <
t
)
The left hand side
(henceforth,
LHS)
The
right
hand
side (henceforth,
There
is
off selling
financial constraints
factors govern
market
RHS)
is
(2)
for
T shares from firms
collateral in the
hands of distressed
the
demand
the supply of T-shares from the rest of the
an excess supply of T-shares, the price governing
must simply
for T-shares
is
- p))-
reflect the
exchange rate at date 2
(i.e.
P).
wasted international collateral in this region because, while the economy as a whole
would be better
Two
is
o , t )(i
t
of this inequality
domestic economy. As long as there
exchange of N-shares
- p)kt(0o,t -e + (i-
amount of domestic
that receive a shock. It reflects the
firms.
(i
is
more of
this collateral in order to finance this shock, domestic
on the firms that
poor
receive the shock prevent this collateral aggregation.
collateral aggregation. First, if
6n
is
high, the domestic financial
poor so that firms are reliant on their own resources. Scarcity of these resources
(low vf1 ) compounds this problem.
As aggregate
be
easily seen for
conditions deteriorate, the
an increase
in foreign cost of capital.
economy
into the
Fire sales.
25
in
p but
2,
of (2) rises relative to the
RHS;
this
can
can also be shown for a decline in Rt or an increase
Eventually, such decline in aggregate conditions brings the
more severe fire
In Region
it
LHS
sale region.
the inequality (2)
is
violated at the (fundamental) price P.
Excess demand for T-shares requires that the price of these shares in terms of N-shares
rises (i.e., that
fire sale region
25
P falls),
resulting in
a depressed value of N-shares.
We
refer to this as the
because the shock results in firms selling N-shares at prices depressed beyond
A decline in foreigners
"risk appetite"
can be easily built into our framework and has consequences very
similar to an increase in Ot; see below.
16
fundamentals. Investment (/„)
economy
is
primarily curtailed by scarce international collateral.
The
as a whole has T-shares of,
{i- P){i-et )kt
the international collateral constraint binds,
If
must be that
it
to foreigners to finance the shock so that reinvestment
In
of these shares are sold
all
by each firm
is,
= hRt—il - St)
P
However,
this equality derives
from general equilibrium considerations. Each firm chooses
P as given. Thus, the equilibrium value of P must fall until yRtWd falls to satisfy
P instead of P. We can write this equilibrium condition in the fire sale region as
r
In taking
(2)
with
1
=
-&Pu>Z+u?)
n
tJ
i-(i-en )PRt K h
where we have made one simplifying assumption and
^(l-6
(3)
t ),
P
let #o,t
= St
(we later show that date
optimality will ensure this relation.)
A
Equilibrium price and collateral premium.
conducted from condition
LHS downward
(kn/kt) will shift the
LHS
a
First,
(3).
down, leading to a decrease in
rise in
so that
prices.
number
of comparative statics can be
the ratio of N-production to T-production
P
falls.
Second, a
fall in
6n also shifts the
Finally, a rise in 6t will cause the supply of
26
Figure 1 illustrates shifts on
T-shares to shift inward resulting also in a decrease in P.
LHS
(upward sloping curves) and
investment and prices in the
The main consequences
RHS
fire sale
(vertical lines), together with their consequences for
region.
of negative aggregate shocks, in terms of moving the
into or along the fire sale region, are well captured
economies as their St increases. This
decreasing curve of
economy has
Thus, in the
less
what we do
and
cross sectional comparison of
in figure 2.
Raising 6t traces out a
Additionally, a large 6t will
in the fire sale region.
less international collateral
This will lead to
e).
P
is
by the
is less
economy
able to fund
its
mean
that the
shock in the N-sector.
production of N-goods at date 2 and an appreciating currency (lower
fire sale
region
eP
is
uniformly downward sloping.
In the wasted collateral region, domestic firms that do not receive a shock hold both
The return on N-shares
N-shares and T-shares.
eRt. Since firms are holding both
Rn/P
is
N and T shares,
while the return on T-shares
is
arbitrage will dictate that,
Rn/P = eRt
26
Note that an increase
in
by the productivity growth
Rt lowers
P but by less than the increase in the former.
in tradeables, the relative price
17
of N-shares
rises.
That
is,
once correced
I./K&
Figure
Equilibrium in the Fire Sale Region
1:
or,
The
first
region of this graph
is
enough that the only constraint that
and there
is
wasted international
sufficient T-shares.
The second
the case in which
t
6 FS .
<
limits investment is the
In this case 6
region
is
where, 6t
insufficiency of international collateral leads
>
.
both to a
small
domestic collateral constraint
collateral. Prices in this region are
6 FS In
is
simply
this case
fire sale
eP
P since there
falls
is
because the
and an appreciated currency.
Then, the collateral premium,
eP
which
is
6t
clearly increasing in
.
As the
collateral
premium
rises,
the cost of financing
production in the N-sector becomes higher as issuance and sales of N-shares occurs at
sale prices.
is
The
collateral
premium
is
fire
financed with the surplus from reinvestment, which
gradually transferred to T-shareholders. There
is
a point where
this transfer is complete,
and from then on firms bit by shocks withdraw voluntarily as they prefer to cut reinvestment
rather than "give away"
all
their N-shares at
region of the diagram, where &t
Returns and transfers.
2 both domestic
and
>
extreme
fire-sale prices; this is
Q IC , snd can be shown that
the right-most
P = 1/Rt-
In Region 1 only the domestic constraint
is
binding, while in region
international constraints bind. In region 3 only the latter constraint
18
\\
\
\
X
"\^
e/R,
^
1
6K
Figure
binds. This observation
which ones do not, and
is
2:
Regions
important for understanding which arbitrage relations apply and
will play
a key
role later
on when we analyze the
social inefficiencies
associated to each region.
The
sells
is
return to purchasing N-shares to an outside investor depends
one share on the T-sector which
is
Thus, the external return on investment at date 1
ext
P.
The
investor
a claim on T-goods at date 2 totalling Rt. This
used to purchase 1/P units of N-shares which each return
r
on
Rn
units of date 2 N-goods.
is,
= Rn = L,
PeRt
while the social or total return
is
r
tot
= Rn
L
L
e
Figure 3 depicts these returns as a function of 0*
L —
1,
the external return of T-shareholders
capital. Collateral providers,
is
.
In the wasted collateral region,
equal to one, the international cost of
which are in excess supply in equilibrium, bid up
P in their
attempt to capture part of the surplus of reinvesting in a constrained firm. In equilibrium,
all
the surplus goes to the constrained firms.
supply shortage of international
collateral.
19
The problem
is
one of demand rather than
eR
a*
Figure
In the
since
all
and
in so doing bid
the collateral premium, L, above
1.
collateral
becomes binding
Constrained firms compete for the few units of
the pledgeable shares are used.
international collateral,
Returns
on the other hand, the supply of
region,
fire sale
3:
e,
As
down the
this
relative price of N-shares,
and
raise
happens, the domestic collateral constraint
becomes binding, hmiting the amount of surplus that constrained firms can transfer to
collateral providers.
Eventually, as the increase in
t
keeps reducing the amount of pledgeable T-shares,
reach a point where the entire surplus
is
transferred to external investors (region 3); from
then on the price of N-shares stabilizes at
2.3
Fire Sales
P = 1/Rt, which sets r^ = r*°*.
1
and the Banking System
Domestic financial systems play many useful
collateral.
we
Most of these
roles
functioning stock markets.
roles in the allocation
we have already circumvented
One we have
not, however,
is
and aggregation of
in our assumption of perfectly
the capacity of banks to
smooth
the distinction between domestic and international collateral in the eyes of foreign investors.
In emerging economies, banks are largely in the business of intermediating foreign funds
into nontradeable sectors.
In our setup, since the tradeable sector has direct access to
international financiers, banks have
an "arbitrage" opportunity
20
in the
nontradeable sector
and naturally borrow abroad to lend to
This intermediation
this sector.
when domestic
is
highly beneficial,
especially in the event of
a
crisis,
of future T-shares which
is
not part of international collateral due to governance problems
in that sector.
a
However, the cost of this activity
-
fragile capital structure
and
collateral
T-liabilities).
A fire
and
assets
is
that
it
leaves the
banking system with
that are de-facto mismatched (N-assets
liabilities
sale in this context will deteriorate banks' balance sheets rapidly
by lowering the value of assets (N shares) while leaving the
create a host of problems
can be used to access that part
by
itself,
either
or simply through the possibility of a
substantial withdrawal at
due to
unchanged. This can
rigid institutional factors
bank run by
fire sale prices
liabilities
foreign investors
(BIS constraints)
which
would dry up the banks resources.
adequacy
the reason for the failure of the banking system
is
straightforward but important recommendation
that closing banks using
as a reference is not a
good
is
amounts to a
idea. It
reduce international collateral at the wrong time.
rigid capital
self inflicted
realize that
27 ' 28
Indeed
ratios,
a
if
a very
fire sales prices
bank run, which can
in turn
29
Foreign Leverage
2.4
Opening the
capital account at date
domestic firms to increase their scale of
will allow
production by selling some of their T-collateral to foreigners and using these proceeds to
finance
have
more production. At the aggregate
level,
less T-collateral to sell to foreigners at
aggravate the price decline in the
We shall
analyze
how
fire sale
date
the consequence of this
1.
We
show
is
that firms will
in this section that this will
region.
equilibrium relation, (3), in the
4
,
u..d\-
l
fire sale
region,
~Pi
changes with the introduction of foreign leverage.
Changes
in foreign leverage
First, if firms sell
have
effects
on both the
reversal in cross-board interbank lending,
Most of the
and
LHS
a fraction Of of their T-shares to foreigners at date
"See, for example, Diamond and Dybvig (1983)
28
Indeed, the bulk of the capital flows reversal in the
in 1997.
RHS
pre-crisis loans
crisis
of this expression.
0,
the
RHS
of this
economies of Asia came in the form of a sharp
which went from 41
billion dollars in
1996 to minus 32
billions
were transformed into unhedged domestic currency loans or foreign-
denominated loans to the N-sector (primarily
real estate), in
which case exchange-rate risk came in the form
of credit risk. Banks' realization of the risky nature of their unhedged positions after Thailand's devaluation
was a big factor behind the regional currency crises that followed; see IMF (1998).
29
As for real or self inflicted bank runs, they can be avoided by boosting the price of N-shares. In a
multiple equilibria scenario, this can be done by just credibly announcing a transfer to "patient" foreign
investors in the event of a run.
As
always,
if
credible this policy will cost nothing.
21
expression will become,
l—^ktil-Ot-Of)
P
Selling shares at date
to date
1.
The per
cjf , falls since
also reduces holdings of T-shares as internal collateral
on the distressed entrepreneur's holdings of T-shares,
capital return
a fraction of
from date
this is sold to foreigner. Thus,
u;?
= (i-p)(i-e -ef
,
t
)
Substituting these into the equilibrium relation yields,
i-(i-en )PR
(knPu}»
+ ** (1 ~ 6t ~
t
The key
expression to be analyzed
is
kt(l
W = ^y-
—
9t
— 6f). On
from foreigners allows financing more production, so that
increased production
is
- 0t ~ W-
the one hand, raising funds
kt rises.
On
the other hand, this
held mostly as illiquid shares by domestic firms so that they hold
of the fraction of aggregate collateral at date
less
h{l
shares issuance are used towards an increase in
1.
Suppose that
all
of the proceeds of
so that collateral creation
A^,
is
maximized.
If
the firm had originally invested k\ in the T-sector, opening the capital account will allow
it
to scale this investment
up
to,
kt
i-R (i-p)ef
t
Then,
d
89f
The numerator must be
i-e -ef
t
1 - R
less
t
l- (i e )Rt(i
1 - Rt(l - p)9f
t
(l - p)9f
than zero or production
reflecting
P must fall.
resulting in
down
would be a money
— 9t — 9f)
falls
with a rising
First, the vertical curve shifts to
a contraction in the aggregate supply of collateral at date
curve for T-shares shifts
that
two ways.
<
in the T-sector
machine (which we must rule out by assumption). Since kt(l
9f, the curves in figure 1 are affected in
p)
reflecting the fall in
domestic
1.
Second, the
collateral.
The
the
left
demand
net of these
is
Thus, aggregate collateral in the economy declines with foreign leverage
a steeper price decline in the
fire sale
region.
A puzzling feature of crises is that both flows and stocks, rather than only stocks, seem
to contribute to their unraveling
and
likelihood.
Our model
offers
one possible explanation.
Flows matter primarily through distressed firms which in principle can renegotiate the
stock of debt but need the
new
flows to survive.
Stocks,
on the other hand, primarily
matter though sound firms, by curtailing their role as suppliers of international collateral.
22
The
3
Pre-Crisis Phase
The Model
3.1
Introduction.
in
Of
advance of the
course,
many
crisis itself.
of the conditions that lead to a crisis are created well
The
analysis thus far has taken these initial conditions as
given and illustrated the types of problems that they might lead
international collateral
question that arises
is
is
what
largely
why
to.
As a shortage
of
responsible for triggering a fire sale, a natural
is
the private sector does not anticipate this and take steps to
invest in highly profitable international collateral. In order to address these issues in our
model, we add a history to our economy; a date
in
which agents have
full flexibility.
In particular this section addresses two sets of choices: production and financing.
On
the
production side, we study the difference between private and social incentives to allocate
benefits
and
On
and nontradeables.
factors of production to tradeables
costs of foreign leverage that results
the financing side,
we study
the
from opening the capital account at date
0.
The
must
Firms at date
basic problem.
first
how much
decide
N- and T-sectors. Production requires resources at date
finance this production
— how many T- and N-shares
the complement of this financing decision
to produce in each of the
and firms must decide on how to
to issue against production. Finally,
an investment decision
is
— how many T- and
N-shares of other firms in the economy should be purchased.
We focus
first
on the case in which there
total resources of the
economy are
fixed,
kn
is
no date
we must have
borrowing from abroad. Since the
that,
+ kt = w.
While the incidence of crises must surely by tied to aggregate shocks, our model allows
us to highlight the main issues leading to a
crisis in
an environment with only idiosyncratic
shocks. Given wealth at date 2 (in units of N-good) of t«2, a firm solves
—
Since uncertainty
is
-1 ' 2
= max{tzd
idiosyncratic, e
is
s.t.
Cte
deterministic
+ Cn= W2-}
and firms are
risk neutral.
taking e as given, simply maximizes the expected value of wealth at date
Let
Vn
Each
firm,
2.
be the marginal value of a unit invested in the N-sector and Vt be the marginal
value of a unit invested in the T-sector.
max
s.t.
Then the
firm solves,
Vn kn + eVth
kn
+ kt = w
23
Investing in tradeables.
Consider each of
V
t
Vn
and
A
in turn.
unit of date
physically invested in the T-sector generates returns in two components.
keeps a stake in
its
own
enterprise of
o ,t
> &t-
This
illiquid
component
is
resources
First, the firm
worth
(in
date 2
T-goods),
eQit Rt(l-p).
The complement
price of q.
The
(1
— #o,t)
is
The market
sold to the market.
firm raises from this issue T-goods at date
discounts these flows at a
totaling,
{l-p)(l-8Qtt )Rt q.
Since these T-goods at date
this
can always be reinvested
in the T-sector, the date 2 value of
is,
{l-p){l-eQ t)Rt V q.
t
Thus,
V=
t
,
t
Rt(l
-p)
+ (l-p)(l- e
0it
)Rt qVt
.
Rather than making a physical investment, a firm can also choose to simply purchase the
T-collateral of another firm.
it
can use
Doing
this collateral to finance
in this case
is
Q^Rnje.
this
produces two benefits.
the shock with foreigners.
If the firm
If
the firm receives a shock,
The value of a unit of collateral
does not receive a shock, the collateral can be sold to
another firm which does. In this case the collateral generates a value of L. Thus
we
define
the value of collateral provision as,
<)
j
since the cost of
a unit of
= penl
^ + (l-p)L>l;
e
collateral is q, the return to this investment is simply,
4>jq.
Holding provisions of collateral from date
to date 1
(precautionary savings on the part of firms).
insurance, however the shares
or the (1
— 6)
is
insurance against entering distress
Firms can hold either N- or T-shares as
must be saleable in the market
at date
1.
Thus, only collateral
share of production can satisfy this insurance need.
In equilibrium, the return on holding a unit of collateral,
return on undertaking physical investment, Vt. This
is
<f>/q,
because
if
must be equal to the
Vt
<
<f>/q,
no firm
will
undertake physical investment and hence there will be no collateral for any other firm to
purchase.
On the other hand, if Vt >
Since neither of these
is
<f>/q,
each firm will only undertake physical investment.
consistent with equilibrium,
we have
the above condition.
Then,
V = (l-p)(eQ ,tRt + (l-eo,t)Ri4>).
t
24
(4)
now the
Consider
Since
<fr
>
1, it is
choice of
The
#o,<-
max
jt
s.t.
e0>t
easy to see that
firm chooses this to maximize Vt
RtO.-p)
>
et
#o,t
+ {l-p)(l-Qojt)Rt
4>
.
=
&t-
In other words, each firm will only hold the
minimum possible of its own shares. Holding its own shares leaves the firm subject
to shocks
which reduce the resources of the firm precisely when they are most needed. Holding shares
market
in the
is
partial insurance against these shocks.
Turning now to the
Investing in nontradeables.
region where In
reinvest
<
kn.
N
sector,
Then, adding a unit of capital to
by expanding the capacity constraint at date
1.
N
A
we continue working
in the
does not affect the option to
similar analysis to the
above
yields that,
Vn = (1 - p)(0n Rn + e(l - en )PR
t <f>).
In equilibrium firms must be indifferent between investing in
Equilibrium.
Equating
Vn
and eVt
this
is
+ (i-en )4>/L) = o + (i-e
t
an equilibrium
relation, shifting
sheds light on private incentives for production.
or a
Rn/Rt
a
rise in
3.2
fall
(5)
fundamentals and the collateral premium
A rise in P, either through an increase in
On the other
hand,
the collateral premium, L, increases the incentive to shift resources to the T-sector.
Externalities in Collateral Creation
While the
reallocating production
is insufficient
as,
t )4>.
in e, increases the return to production in the N-sector.
Issue and setup.
To
and T.
yields:
P(0n
While
N
collateral
and Nontradeables Investment
premium provides a
clear private incentive for
from nontradeables to tradeables, we show in
to align private
and
simplify the analysis, let 6n
this section that this
social incentives.
— 6t = 6.
The problem of a
firm at date
can be written
30
max
j
u
=
i
\CnCt)
2
kn,kt
We have written the optimization problem as that of a representative agent over aggregate consumption.
There are some
this case
issues of aggregation that
have been suppressed
because of the Cobb-Louglas preference structure.
Thus, the objective function
in writing this optimization. It is valid in
A firm at date
is,
maxp(4c?)±+ (1 -/»)««*)*
However, aggregate resource constraints
mean
(*£
that,
+ (1 - P)<£ = Cn
25
1 is either distressed
or sound.
= Rnknil - p) + pRnln
Cn
S.t.
= Rth{l ~ p) ~ pin
Ct
- 0)(1 - pXknP + kt)
In
= 7R
kn
+ kt = w.
t
(l
Consider the derivative of the objective with respect to a perturbation of a
in kt
and
A decrease in
A
increase
A^.
dud
5A
/
1
dCn
dct\
+C
2^( Q aA "5Aj
,A=0=
where,
and
dCt
Substituting these into the above
0,
T> fi
we
9I"
\
arrive at the first order condition for a firm at date
31
Consider next the
Social versus private problems.
planner at date
0.
This
differs
from
(6) in
first
order condition for a central
only one way. Changes in
A affect investment at
date 1 through their effect on asset prices, that private decisions do not take into account.
This difference
is
reflected in the
or
term
qj
soda!
At the
private level,
private
^
whereas at the social
^.
=7^(1- 0(1 -p)(l-P)
level this is,
sp
q
Private incentives do not take into account the effect of a change in
collateral at date 1. Since firms at
2, increases in collateral
A
aj
private
on the value of
date 1 are financial constrained in both regions
values allow for
more
positive
NPV
investments (Rn
>
1
and
e) to
be
undertaken.
p4 + (1 - pK = <*
where, individual consumption
satisfies, -g-
=
-£. These relations
and the
fact that preferences are
Cobb-
Douglas imply that the objective can be written to maximize preferences over aggregate consumption.
3I
A little algebra verifies that
this condition is equivalent to (5).
26
Wasted
In the wasted collateral region, increasing kt causes the pro-
collateral region.
duction of T-goods at date 2 to rise relative to that of N-goods resulting in an appreciated
At date
currency.
1, this
raises
meaning that the
positive,
In the
Fire sale region.
P, the domestic value of N-collateral. Both g^ and |£ are
social value of increasing
fire sale
region, this effect
and
relaxes the international collateral constraint
easily understood.
N
economy
is
always greater than zero since Rn/e
is
would not
at date
>
affect the
investment decision, /„, at date
is
(i.e.
6n
=
or
and private and
1,
wd
high),
social incen-
closely related to the financial constraint
differs across region 1
and
In the wasted collateral region, shifting production to the T-sector results in a
2.
is
reflected at date 1 as
value of collateral, which loosens the financial constraint.
take this into account because there
is
the collateral premium (L) at date
1.
The
no extra price which
duction decisions on firms in distress. In the
fire sale
only L. In region
diminishes as
2,
Rn/e
<
region, there
The source of the
L, while in region
private
3,
an increase
in the
economy does not
internalizes the effect of prois
externality
not high enough. Socially, the premium to collateral creation
it is
Since the
1.
investment was not financially constrained
If
permanent appreciation of the currency. This
is
— p)(l — 6)Rt-
always be used to fund reinvestment
However, the mechanism underlying the externality
1.
is
return to this reinvestment gives,
would be aligned. Thus, the externality
region
P. 32 The intuition behind this
Suppose that a central planner allocated one unit of resources away from
The
The mechanisms.
tives
reinforced because the change in k t
raises
in the fire sale region, this collateral will
in the N-sector.
prices
is
into T. This generates aggregate collateral at date 1 of (1
and
which
A is higher than the private value.
Rn/e
is
such a price, namely
is
that this premium
Rn/e, however privately
= L.
Thus, the externality
we move through region 2. An immediate implication of this is that any date
1
policy of supporting prices in the fire sale region has the ex-ante consequence of diminishing
private incentives to create collateral.
Foreign Leverage
3.3
The
The
issue.
answer in
this asset
question of whether capital flows should be regulated or not finds one
markets view:
it
depends on whether they
collateral constraints during the crisis phase; ultimately,
during the
do not
I.e.
crisis.
and 7
on whether they
raise or lower
As in our discussion of externalities in production decisions,
fully internalize
P
affect positively or negatively
rise
P
private agents
the consequences of their decisions on others' collateral value at date
by more with A.
27
1.
In the wasted collateral region, the chief factor behind
P
is
the long run real exchange
rate; as capital inflows appreciate the latter, they help aggregating collateral
In the
fires sales region,
on the other hand, the main factor
capital inflows increase this
premium and lower P,
of international collateral at the time of
Private tradeoffs.
for
during
crises.
P is the collateral premium L;
as they reduce the country's availability
crisis.
Borrowing resources from foreign investors allow firms to expand
production at date
However, in return firms must
0.
would otherwise yield a
collateral
premium
a discount rate of one (the international
sell
some
of their collateral which
at date 1. Foreigners always
interest rate),
buy T-shares using
which means firms borrow
if,
V- >
t
<f>
or,
<?>!•
(7)
While, foreigners use a discount factor of one to buy T-shares, domestics discount flows at
q. If
<
q
1,
firms will always prefer to sell shares to foreigners
Expanding the expression
their domestic production.
Q
which shows that as Rt(l — p)
q
falls
^+
(l- p)Ri
for q, yields
(1
-e ))
t
the neoclassical advantage of borrowing from abroad to
invest in high return projects rises. If this
(<p),
in
1
=
rises,
and invest these proceeds
below one and firms find
it
high relative to the value of collateral provision
is
profitable to
borrow from foreigners and take on
leverage.
Wasted
In the wasted collateral region, boosting the domestic exchange
collateral region.
rate raises the value of N-collateral
Foreign inflows are one
at date
way
and loosens the domestic
financial constraint at date 1.
to accomplish this since they appreciate the currency. If firms
are able to leverage their production in the T-sector with respect to foreigners,
the effective productivity of the T-sector
appreciates the currency. This
Fire sale region.
fire sale region.
The
is
rises.
In turn, the increased production of T-goods
the benefit of opening the capital account at date
social benefits of
0.
opening the capital account are reversed in the
Foreign leverage lowers the amount of international collateral available to
the domestic economy and therefore lowers
reduces the amount of investment at date
generates T-goods of Rt(l
P (see section 2.4).
1.
On
fall
in collateral values
the one hand, selling a T-share at date
— p), which are each worth Vt
28
This
at date 0.
On the other
hand, this
.
V.'
R./e
y^
*
•
eK
Figure
4:
Social
and Private Values
reduces the amount of collateral at date 1 by Rt(l
generate date 2 flows of
Rn/e per
unit.
in the Fire Sale
— p),
Thus, leverage
Rn/e
e,
is
Region
which could otherwise be used to
socially beneficial
if,
<V
t.
Figure 4 illustrates the divergence between private and social values of opening the
capital account in the fire sale region.
abroad, while
region 2 while
<f>
is
it is
Rn/e
is
the social cost of selling a unit of collateral
the private cost of this transaction. It
equal to Rn/e in region
3.
Thus the
is
easy to show that
private cost
is
<f>
< Rn/e
in
always less than the
social cost.
This does not mean, of course, that capital flows should be stopped altogether. The
benefit
from borrowing from abroad
which shows that when Rt
and social values
is
high, the business of borrowing is very
h
While Vt
is
a function of
are less interested in
good and both private
dictate taking on foreign leverage. This case corresponds to the
horizontal line in the graph. 33
33
is:
its
6t, it
shape than
medium, and low values of
On the other hand, when Rt
has been drawn as a horizontal
its
dependence on Rt. This
Vt.
29
is
is
line.
uppermost
low, private costs of leverage
For the purpose of our
the parameter that
is
figure,
we
shifted to give high,
overwhelm the
at.
date
to take
rises,
0. It is in
V™ in between Rn/e and
the intermediate range,
on foreign leverage when
and leverage
be
will
rise,
that firms will choose
<p,
As aggregate
social incentives dictate otherwise.
both private and social costs of leverage
V™ in the figure,
of
and firms do not access international capital markets
benefits of borrowing
eventually turning
V
h
t
leverage
into the equivalent
socially inefficient.
Insurance Markets and Aggregate Shocks
4
If distressed
firms could arrange in advance to receive financing from
we have
vorable levels the inefficiency that
identified
- contracts which our
fa-
would disappear. However, doing so
would require introducing insurance contracts contingent on the date
ization of firm type
sound firms at
1 idiosyncratic real-
analysis has implicitly ruled out.
To
see
why
the introduction of such contracts would eliminate the inefficiency, consider the following.
Note that the externality
arises in the fire sale
because at date 1 distressed firms are un-
able to transfer the full value of their surplus to sound firms because of a binding domestic
collateral constraint. If these firms could arrange in
distressed they
slack,
advance that in the event of becoming
would receive a payment to exactly leave the domestic
collateral constraint
then they would always have sufficient resources to transfer the surplus to the sound
firms. Ex-ante, since all firms are identical, the
no transfers and
is
purchase of this insurance contract involves
therefore clearly feasible as long as the firm type
is verifiable.
While insurance markets against idiosyncratic shocks can be defensibly suppressed, those
against aggregate shocks need
is
that
we
be considered.
and
are able to describe crises
First,
a simplifying feature of our analysis
identify the collateral undervaluation externality
without introducing aggregate shocks. However, the incidence of a
aggregate shocks, and as such
we now extend the model
crisis is
surely tied to
to include aggregate shocks
and
insurance markets against these shocks.
At date
1,
given date
on (p,6n ,6t,Rn,Rt)1,
decisions of
Let s G {so,
with corresponding probabilities
...
(fcji,fct,0o,n 3 #o,t)j
=
,Sj}
7r(s).
«S
the extent of a
pG
{0,/>i},
where p(«o)
where the economy
is
in
=
a
is
all
depend
represent the state of the world at date
Then, in each state
s there are realizations of
(p(s),6n (s),0t(s), Rn(s),Rt(s)) that represent the aggregate shock.
on the case where s € {so,Si} and
crisis will
aggregate uncertainty
is
To
simplify let us focus
captured by variation in
a state where no constraints bind and p(si)
= p\
is
a state
fire sale.
Let us close the capital account at date
provide insurance against shocks. This
is
0, so that
kn
+ kt — w,
captured by letting
#o,t
depend on
to the restriction that,
l-6 =
t
7T
(1
-
M
s o))
30
+ tti(1 -
M
but allow foreigners to
s i))
s,
but subject
At date
is
0, 1
arrived at
— Qt
is
by imagining that firms
state contingent
The
the fraction of T-shares that can be sold to foreigners.
sell all
restriction
of this to foreigners and then purchase back a
number of shares. The quantity
of insurance
is
then denominated in units
of T-shares.
Now,
in state sq the
if
economy
firms encounter no shock, there
insurance against this state, so that 1
— 6o,t(so) — 0.
is
no reason to purchase
34
This implies that,
^
i-M^i) = 1
Return to the question of the choice of
k„,
and
(8)
kt.
(8) reflects all
of the insurance
decisions that firms will take in response to aggregate shocks. Then, at date 0, the
of kn and kt
-
kt is just
is
similar in character to that of the case in which there are
a scale factor on the amount of T-shares available in
sj.
underprovision for the
production and the decentralized economy will
fire sale.
Final Remarks, Caveats
In terms of our model, the
particular
and hence command
However, for the same logic as in the previous section the premium will be
insufficient to incentivize efficient collateral
5
no aggregate shocks
While in state so the T-
shares have no collateral value, in state sj, T-shares are very valuable
a premium.
problem
first
and Policy Considerations
phase of the current
crisis
—Thailand
(mid 1997)
— can be seen as an accumulation of aggregate shocks, leverage and weakening
nancial institutions that eventually ran the
the devaluation of the baht,
ent practices
and
these countries
came
institutions,
economy
fi-
into the fire sale region.
Together with
than
fully transpar-
investors' learning of
a
series of less
which they extrapolated to neighbor countries.
—Indonesia, Malaysia and the
in
Effectively,
— experienced a
Philippines, in particular
sharp increase in perceived #'s and perhaps a decline in the fraction of the available shares
which were thought to be appropriate international
international collateral
As the
stiff
crisis
collateral.
This perception of declining
was exacerbated by a widespread weakening of financial
has spread, beginning with the regional consequences of
institutions.
Hong Kong's
defense of their currency board after the Taiwanese devaluation, continuing with the
severe distress of
Korean and Japanese banks, and reaching a climax
in Russia's collapse,
the problem has shifted in nature from domestic to one of shortages in the international
supply of funds. Even relatively solid countries within Latin America have been severed
from international capital markets. Although the world as a whole seems to have plenty of
liquidity,
34
those that have the
One can think
know-how
to invest in these countries have seen their funds,
of this as a line of credit that the central bank takes out from foreign banks.
31
capital
and mandate dry up. Marginal
constrained ones.
investors in the region are either uninformed or very-
Both considerations have
led to
an increase
in the required return
on
these countries. Equivalently, the "flight to quality" and avoidance of exchange rate risk of
these investors has led to a reduction in the fraction of domestic shares that are acceptable
by foreigners
widespread
— a negative aggregate shock from the point of view of
fire sales
collateral.
As
before,
are the natural consequence of these shocks. 35
Before discussing policy aspects of our problem,
it
remind the reader of the narrow goal we have pursued
seems particularly appropriate to
which
in this paper;
to highlight
is
the core features of a world where asset market and financing considerations take center
stage. In so doing,
we have removed many
aspects of the problem which obviously need to
be considered when designing an actual policy package; traditional nominal and real wage
rigidities,
intertemporal smoothing and aggregate
Even within the domain of an
what we view
and
as the essence
asset
demand
name a
factors, to
few.
markets perspective, we have limited ourselves to
novel, leaving aside important issues
which have been
already discussed in the literature and for the most part, would only reinforce our conclusions.
firms
is
We
have assumed, for example, that
full dilution of
smooth and uneventful. The prolonged and
claim holders on distressed
costly nature of the Latin
American
debt overhang problem during the 1980's attests to the fact that the process of dilution
is
a sluggish and imperfect one. 36 Adding such considerations would reduce even further
the collateral available to distressed firms, and
since
commitments to
foreigners are often
it
made
would
also reduce international collateral
—
directly or through domestic
banks
terms of T-shares. Similarly, except for our brief discussion of the banking system,
—
in
we have
shutdown the most powerful dynamic multiplier mechanisms discussed in the credit constraints literature.
mechanisms a
Adding a few periods to our model would give us dynamic amplification
la Kiyotaki
and Moore
may be the result of today's distress,
exacerbating the extent of the
crisis
(1997),
where the anticipation of a
which
fire sale,
further feeds back into today's decline in stock prices,
and the
effects
we have
discussed.
37
Throughout our analysis we have taken contractual problems as given, and have explored
the implications of these for crises scenarios and their likelihood. Whether a country
to exhibit a crisis
35
—and the particular form
it
— depends to a
takes
large extent
a
fixed
number
of T-shares
is
on whether
relaxed. Financial panic does
its damage, we argue, precisely by reducing the number of shares that are acceptable
36
See, e.g., Krugman (1985a,b) and Sachs (1984) on disorderly workouts.
last
likely
See the section on financial panic as well as the appendix for a discussion of situations where the strict
allocation of international collateral to
37
is
The main reason we do not have
period there can be no
fire sale
this effect in
and the date 2
underinvestment problems. By adding more
our two periods (dates
real
1
and
much
of
collateral to foreigners.
2)
model,
is
that in the
exchange rate always dampens rather than amplifies
periods, the equilibrium real exchange rate constraint can be
postponed and replaced, along the transition, by a dominant collateral premium
32
effect.
individual entrepreneurs can
commit
effort
and output
we have
the direct consequence of misguided policies,
at relatively low cost.
given a more active role to the private
sector in generating the conditions for crises in emerging economies.
is
more
indirect,
and
it lies
in not providing
and commitment mechanisms. Johnson et
recent crisis.
Rather than
The government's
fault
economic agents with appropriate enforcement
al (1998) find strong
They document that corporate governance
support for this view in the
factors
seem to do substantially
better in explaining the cross sectional variation in aggregate distress
stock market decline, exchange rate depreciation
and
—
as
interest rates hike
—
measured by
in
emerging
economies, than do traditional macroeconomic variables.
Unfortunately, while removing misguided policies should be "relatively" easy, creating
appropriate institutions to protect investors and financiers from dishonest behavior
tainly a
more complex
rent-seeking cultures.
38
task, as
We
is
cer-
often involves changing deeply ingrained customs and
are not nearly as ambitious in our discussion; instead
on the policy implications of our
approach
it
is
analysis, given imperfect institutions. Consistent
we
focus
with this
our omission of insurance, complicated microeconomic contracts, and narrowly
targeted transfers, from our discussion of potential remedies.
The most immediate implication of our model
in terms of policy
is
during the pre-crisis
period, where subsidizing domestic investment in the tradeable sector (perhaps financed
with a tax on nontradeable investment) should be contemplated whenever the likelihood of
faffing into
wasted collateral or fire sale region is significant. Although the bottom line is not,
the rationale for such reallocation policy
region the goal
is
is
different in each region. In the
one of long run appreciation of the real exchange
relative price of N-shares.
In the event of a
crisis, this
rate,
wasted collateral
which raises the
higher price facilitates collateral
aggregation and thus reduces the extent of the waste. Policy incentives are needed because
the private sector does not take into account the negative consequences that overinvesting
in nontradeables has
In the
fire sale
on others'
region,
collateral shortage,
collateral.
on the other hand, the most acute problem
which
is
is
one of international
obviously alleviated by an expansion of domestically
owned
tradeable sector. Incentives to create collateral are inefficiently low in the private sector,
because collateral providers anticipate a relatively depressed
due to limited domestic
collateral.
international collateral constraints,
It is
demand
for this collateral
the dynamic interaction between domestic and
which underlies the externality that justifies government
intervention.
On
pre-crisis capital flows,
sale is likely or not. If
38
our recommendation varies depending on whether a
wasted collateral
is
fire
the primary concern, then capital flows should
Which, in turn, axe sometimes the consequence of misguided
33
policies.
be
The
fostered.
long term real exchange rate appreciation which results from such active
policy would revalue assets in the nontradeable sector, boosting domestic collateral values. 39
However,
are the primary concern, then the opposite
if fire sales
entrepreneurs consider the effect of foreign leverage on their
negative impact of their leverage on aggregate
course of action.
As the
amount
—
fragility,
and
on what actions to take during
market intervention and
principle milder
the suggested
do not always take the appropriate
— may be large enough to make
crises.
— wasted
In particular, and motivated by policies
crisis,
how should one think about
interest rate hikes within
an asset markets view?
But
relaxes the financial constraint of distressed firms.
rates,
In the
crises.
collateral scenario, boosting the price of N-shares
an appreciation of future exchange
The question
crises unavoidable.
important to distinguish between different types of
it is
is
in those circumstance that they do, aggregate shocks
advocated and implemented during the current Asian
it
taxing capital flows
real or speculative
Again,
but disregard the
fragility,
'
of preventive measures,
direct stock
true. Since individual
40 41
recent events so vividly highlight, countries
arises, then,
own
is
unless this
is
is
—in
useful because
accompanied by
such policy will depress the expected return on
holding N-shares and create arbitrage opportunities across N-shares and T-shares. In other
words, devised only as a temporary measure, and not supplemented by structural changes
supporting a more appreciated currency in the future, this policy can be prohibitively
expensive in terms of the resources needed to
In the
fire sale
region the story
is
make
quite different
it
succeed.
and more
relevant for current events in
emerging economies. Boosting the price of N-shares without adding available international
collateral or resources has
no
effect
by
itself.
It is at this
juncture that raising domestic
a high return
—not in the sense of pure monetary policy but by
For
to happen, the government
instrument to foreigners— may have a favorable
offering
interest rates
effect.
must be able to
this
partially circumvent foreign investors concerns with holding N-shares
43
promising a high return on the bonds issued.
39
42
Alternatively, with
The
by
international collateral so generated
a well capitalized domestic banking system,
this incentive
may
not be needed since
the N-sector also becomes the recipient of capital flows in that case.
40
The same argument
rationalizes, within
can be used as international
our model, the practice of government's reserve hoarding which
collateral in the event of
a
crisis. It is
straightforward to
show
that, because of
the externality, the private sector will not undo one for one the government's reserve accumulation.
41
In agreement with our policy proposition and the factors behind it, IMF (1998, p.7) conjectures that in
the absence of go jd financial supervisory and corporate governance, taxing short term capital inflows
may
be appropriate.
42
Of course, the
interest rates is
43
initial
shock could be the withdrawal of existing foreign resources, in which case raising
a mechanism to reduce the outflow.
A possibility not available to individual entrepreneurs.
Cavallo (1996) describes Central
Bank
policy in
a currency board system as the "normal ones" of a Central Bank but where funds must be obtained from
34
must be channeled back
their shares,
A
into N-share holders in distress, possibly
which otherwise would be hurt by the high interest
brief formal
ernment can
issue
make payments
model highlights the
Suppose the gov-
limits of such interest rate policy.
The government
of T-goods at date 2.
investors, with R(0)
=
for these
and R'(B)
1,
R{B)
bonds.
>
0.
premium given B. Domestics
is
B
issues
will
bonds and faces a down-
and allows
for
sale of
B
bonds in the
more investment. L(B)
=
R(B). Raising
R
return,
is
raised
beyond
this point, the
and actually the economy
domestic stock market starts to
raises less funds
from abroad.
In closing, we highlight a few extensions which
we have not paid much
The same mechanism
domestic investors also run.
We
It follows
that
we have
fall
is
better
to equalize the
45
left for
future work.
is
First,
the maturity
leading to underprovision of collateral in
our model should encourage excessive short term borrowing as well.
crises
the
is
R(B). This
economy
attention to a central vulnerability issue, which
structure of foreign debt.
fire
interest rates as long as these returns
are less than the return on domestic holding of N-shares assures that the
off. If
>
not buy bonds as long as L(B)
because they prefer to hold the domestic stock market at these prices.
the optimal bond issue sets L(B)
will
it
the required return by international
The proceeds from the
sale region reduces the extent of the fire sale
is
rates.
bonds to foreigners at date 1 against an imperfect promise that
ward sloping demand curve
collateral
by boosting the price of
view this as a dynamic
issue,
Second, in actual
which
reflects
informational advantage these investors have, and therefore their ability to anticipate the
the
fire
sale scenario. After the capital loss occurs, they are likely to return before foreign investors
do to reap the
benefits of fire sales. Lastly,
we have
discussed wasted collateral
and
fire sale
regions almost in parallel. In reality, they are likely to correspond to different phases of
development and medium frequency events. For example, an early wasted collateral region
can gradually turn into a
fire sale
region as capital inflows accumulate without the parallel
development of an adequate enforcement and regulatory environment.
willing lenders rather than from printing
money.
Argentina's issuance of collateralized (by domestic mortgages) repos,
is
a very
direct
example of N-shares
turned into international collateral through government intermediation.
44
Unlike the wasted collateral region, there
post-crisis scenario, since the increase in
is
P just
no need to promise a higher price of N-shares during the
reduces the extent of the
fire sale,
so there
is
no need to
worry about arbitrage across N- and T-shares.
To shorten our
problem.
discussion,
we have disregarded the monopoly aspects of the government bonds issuance
Taking into account the monopoly problem, we get that the optimal
rain{L(B),Rn/e(l
+ (i)}, where y. is a monopoly markup.
35
B
is
such that
R(B)
=
Appendix
6
Regions
6.1
We have discussed fire sales and wasted collateral regions in the text without demonstrating
that both of these regions necessarily arise in the model.
to find conditions
on parameters such that these regions
Take the case where 6n
= 6 = 6.
t
P(0 +
Consider
first
(1
Then date
The purpose
of this subsection
is
exist.
optimization means that,
- 6)4,/ L) =0 + (\-6)<j>
the wasted collateral region. In this case,
L=
1,
so that
4 = P=1.
eR
Prom the aggregate
t
constraints at date 2,
Rnkn (l -p)+ pRnln _ i^
Rt kt(l-p)-pln
Rt
solving,
In (Rt
+ 1) = R^—^ih - K).
Now,
Jn
where the
= 7# (l - 6)(1 - p^KP + h)= 7-Rt(l -fi)(i - P)w
t
last equality follows
from the
fact that
+ kt = w.
kn
Substituting this above and solving for kt, yields
w
f p(Rt
+ !)(!- 6) + !-(!- 6)Rt \
^~2V
We
i-(i-e)Rt
are in the wasted collateral region where
In
L=1
if
)
w
r
two conditions are
satisfied: (1)
if,
< ^—^-ORth
P
and
(2)
if,
In
<K
Substituting in for kt and rewriting inequality (1),
<
l-(l-6)Rt ~
The RHS
than
varies
zero.
from
Thus there
to oo as
p
falls
from
exists p* such that
p
1 to 0.
< p*
p> p*, gives us the fire sale region.
36
1-P
p
The LHS
is
bounded and always greater
implies that (1) holds.
The complement,
Both region 2 and region 3 are
possible to
jump from
possible
when p >
region 1 to region 3, so that the
optimality condition at date
p*.
We
fire sale
show next that
not
it is
of region 2 can occur.
The
is,
P6 + (1 -6)P) =
+ (1 -0)4,
rewriting,
0(P-i)
It
easy to show the following:
<f>
the
fire sale region.
The
to
are continuous,
a continuous, increasing function of
!?;
is
P and P
= (i-0)4,(i-p)
collateral
jump from
premium
is
L=
region 1 to region 3,
^°-.
p. In region
This ensures that equilibrium
3, this
p
enough shocks
<
p* ensures that
(e.g.
p
= 0),
6.2
Returns
Some
basic assumptions
To ensure
we
are not in the
may be
inequality (2)
condition for (2) to always hold
is
L = Rn/e >
means that
1 to
1.
unique in
In order
Rn/e. However,
this is not possible.
fire sale region.
violated.
is
However, for small
A sufficient,
but not necessary,
that,
on returns which we have not stressed
that investments be positive
NPV at date 0,
be greater than one.
We
on
be bounded. In other words,
internal collateral
becomes
we must have that L jump from
continuity of the optimality condition with respect to p
Finally,
monotone decreasing functions of
also require that,
when
in the
both jR„(1
paper are required.
— p) and Rt(l — p)
must
firms are able to borrow, the multipliers
l-(l-0n )Rn >O
and,
i-(i-e )Rt (i-p)>o.
t
We
unit
is
have also assumed throughout the paper that Rn/e
always a positive
NPV project.
In the wasted collateral region
we have
that,
eRt
rewriting,
^ = Rt>l
€
37
>
1
so that saving a distressed
which
verifies
In the
the assumption.
fire sale region,
we must have
that,
£(P-l) = (l-fiMl-P)
Since
P > P, it must be that P >
1
P<
and
^>
1.
Rt
e
which again
verifies
Therefore,
>
1
the assumption.
Incentive Constraints
6.3
We have
appealed to agency conflicts in order to justify the assumption that entrepreneurs
be required to retain a fraction of their shares. One possible formulation to justify
illiquidity
assumption
is
as follows.
Consider a contract between a firm and a lender at date
Now
[0,oo).
can observe
R
yields date 2 flows of
one unit at date
t
where E[Rt]
assume that the realization of Rt
this
this
payment only by paying a
is
0.
Suppose that a firm investing
=
Rt,
and the support of Rt
A lender
private information of the firm.
cost of c (in units of T-good).
Then,
is
it is fairly
standard to show that an optimal contract will be a debt contract (see Gale and Hellwig
(1985)).
The
contract will specify a face of /;
otherwise, the firm defaults
Assume that each
Each project
is
is
m
= E[Rt-f\R
t
-
Then
>f]Prob[Rt>f]
- c\R\ <
f]Prob[Rt
— c.
i.i.d.
return of Rt.
define,
<1
<f]+ fProb[Rt >
< f\Prob[Rt < f) _ Prob[Rt < f]
(
Lenders receive
f]
Rt
^>
=
^^
Proftfc
the component of firms that can be held by outsiders. In our formulation
suppressed the cost
holds
and receive Rt
is,
Rt
is
cost
of,
scaled by Rt this
This
and lenders pay the audit
individually financed via a debt contract.
E[Rt
-
t
the share of each firm that the entrepreneur necessarily holds.
expected flows
E[Rt
R > f the firm makes the repayment of /;
firm has a continuum of such projects each with
et
This
if
itself.
c,
so that this component
is
-ft*
we have
simply the complement of what the firm
However, this makes no qualitative difference to the problem.
38
< /]
Domestic and International Collateral
6.4
Throughout the paper, we have assumed that international
on the T-sector.
While
justified in
many
the most prominent example being the
question to pose
is
how
instance, this assumption
ADRs
on Telecoms and
restricted to shares
is
Utilities.
Two basic results have been stressed the paper.
of international collateral generates
a
interpreted the fire sale in two ways:
fall in
is
often violated
-
Thus, a natural
sensitive are our results to the assumption that all N-shares are
purely domestic collateral.
and a
collateral
fire sale at
date
a shortage
In terms of asset prices,
1.
an excessive date
First,
1 real
we have
exchange rate depreciation
we have shown that there
the value of non-tradeable shares. Second,
externality that leads to underinvestment in the tradeable sector
-
or too
little
is
an
international
collateral creation.
The purpose
of this section
is
two-fold. First,
we introduce a small "dividend" on pro-
duction at date 1 and a corresponding consumption decision at this date in order to pin
down
the extent of exchange rate depreciation and rise in domestic interest rates. Second,
we break
the strict connection between domestic collateral and N-shares by assuming that
a fraction of the
N sector
is
also international collateral.
domestic firms") remain purely domestic collateral.
sumption, the
fire sale
region naturally yields a large
large real exchange rate depreciation.
ation
is
preserved.
46
The important
The remaining N-shares
We
show that even under
fall in
("small
this as-
the price of N-shares and a
In addition, the basic externality in collateral cre-
auxiliary assumption that
is
needed
is
an asymmetry in
the goods market: foreigner never accept N-goods as payment, but always accept T-goods.
In other words, at date 1 N-goods count only as domestic collateral while T-goods count as
both domestic and international
Consumption at Date
collateral.
1
Extend preferences to include consumption at date
As
before, given aggregate domestic
1,
consumption at dates d
=
1,
2 of
Cn< d and Ct d, we have
t
that,
Cn
,d
^t,d
Suppose that given production choices of
k„., kt,
N-goods and Dtkt T-goods to each firm.
46
A
trivial
extension
is
to
is
generated at date 1 of
Dn kn
assume that these goods are perishable.
add "small" T-firms, whose shares do not constitute international
Indeed, by introducing enough of these,
to the N-bias
We
a dividend
we can go back and
rely only
on the "home-bias"
collateral.
—as opposed
— finding as our primitive assumption driving the wedge between domestic and international
collateral.
39
However, while N-goods must be consumed by domestics, T-goods can be sold to foreigners
to generate resources to finance the production shock.
Consider
now
the problem of a sound firm.
It
has resources of goods at date
shares at date 2 in each sector. Let the fraction of N-shares that
be
/.
Then
resources at date 1 in units of T-goods
its liquid
ws =
kn (Dn /e 1
and
illiquid
n )(l
- f)PR
E\ and £2
1
t
t
t
)R
t)
is
the
at each date to maximize,
1
s
t
t
for this
this,
E\e\
s.t.
>
+ h(D + (1 - p)(l - e
/e 2 )
+ E2 e%
Ei + E2/L = w + {6n Rn kn /e 2 + e R k )/L
E2 >en Rn kn /e2 +e R k
max
where A
is,
wealth at date 2 of N-goods of OnRnkn and T-goods of 6t Rth- Given
firm chooses expenditures of
The F.O.C.
international collateral
is
+ (l-p)(l-en )fRn /e2 +
- p)(l -
(1
and
1,
t
t
t
t
is,
the multiplier on the second constraint.
(t>L = 1
When L > 1
In the wasted collateral region,
we know
not bind. Thus
e\
ei
> ^2,
it
must be that
or the real exchange rate
= e2
is
.
that
firm
is
suitable modifications of liquid
and
illiquid resources.
The amount
in the fire sale region,
excessively depreciated at date
The problem for a distressed
collateral constraint.
and that the second constraint
we have
will
that
1.
similar to the case of only date 2 consumption, with
Consider, however, the aggregate
of international collateral must be modified to include
date 1 T-goods as well as claims on date 2 N-shares which are international collateral. Thus,
In
The economy can
at date 2.
< hD + (1 - P)(l - Qt)hRt + (1 - P)(l - 6n)fknRn/e 2
t
finance a production shock
Suppose
it sells
Jni i at date
1
and In>2
In ,i
and
selling resources at date 1
at date 2, then (l),
and resources
47
< hD
t
(2),
/n,2
7
by
We
<
(1
- />)(* " et)hRt + (1 - />)(! " Gn)fknRn/e2
have suppressed the aggregate resource constraint at date
40
2,
that I„_2
<
Rtkt(l
— p).
Note that while
on
it
can
of
sell all
its
resources at date
1, its
depends
sale of date 2 resources
Given these choices,
6.
Cn,i
= Dn kn
Ct,\
= D h — In
t
,i
C„,2
=
{l-p)kn Rn
C
=
(l-p)ktRt-In,2
t fi
+ {In ,l+Infi)kn Rn
In the model with consumption at both dates, the analogy to wasted collateral
D—
nK
-
l
Dth-In,!
while in a
(l-p)ktRt-In,2
fire sale,
(1- p)knRn + (/»,! + In^KRn
(1 - p)hRt ~ Infi
Dnkn
Dth ~ In,l
Consumption smoothing dictates that
T-goods at both dates
is
unable to
is
that ei
>
1
and date
sell sufficient
=
e2, or that the
However,
2.
if 6t is
economy balances
very high (and
collateral.
/
that in the
fire sale
region
we
will
In addition,
it
must be that both the
N
have
of
economy
The
L >
result
1,
since
premium over purely
shares that
collateral as well as the T-shares that are international collateral carry this
fire sale is restricted
its sale
low), the
claims on consumption at date 2 and (2) will bind.
It follows closely
e^.
e\
claims on date 2 consumption that can be sold to foreigners carry a
domestic
that,
- p)kn Rn + (In.l + In^nRn
-(1
= ei — e 2 =
-
is
is
international
premium. The
to domestic collateral, which in this case will be N-shares that are not
international collateral
and the date
1
N-goods (manifest as the exchange rate depreciation
at date 1).
Consider next the price of a claim on the non-tradeable sector.
In this formulation
the N-shares that are international collateral and the domestic N-shares will have different
prices. While, as in the
sale),
this,
1
main
text, it is clear that
with consumption at date
consider a
bond that
is
1,
the international N-shares will also
fall
and 2 of one unit of N-good. The cum-dividend price of such a bond
the economy
is
not in a
fall
in price (fire
in price. 48
To see
issued by such a firm which pays a dividend at each of dates
fire sale,
then e\
will be,
e2
ei
If
domestic N-shares will
— ei = e and,
Pn = e
<8
The
decline in the latter's price
value of dividends.
Still, its
decline
is
is
not a
fire sale,
in the sense that the price
excessive in the Kiyotaki
41
is still
and Moore (1997)
equal to the present
sense.
Now, consider the
constraint (2) on sale of date 2 consumption
fire sale
if
6t
increased further, as the
is
Suppose we increase 6t
following experiment.
economy
is
is
exactly binding.
not able to smooth
its
We can calculate the effect of such an increase on P
across time.
until the point that the
n
A
fire sale
consumption of T-goods
As
such, suppose consumption at date 1
decreased by
is
how a small
be evaluating
reallocation in consumption between these dates will affect the exchange rates
prices.
occurs
dC and
and thereby
that at date 2
is
increased by dC. Then,
pn
= pn_dC(_l
e
If
C2>
Ct,i,
tj
falls in
the
there are
then the expression in parentheses
fire sale region.
more resources
1_\
C
\Ct,i
ti2
is
)
positive so that the price of N-shares
Constraints in an emerging economy will usually
mean
that
than in the present (but liquidity issues
available in the future
prevents financing a consumption increase in the present), thus this inequality will naturally
The economy would
hold.
rather finance a production shock by selling claims
consumption than claims on present consumption.
exchange rate depreciation at date
of these
falls
sale effect
sale of
is
due to
1.
over and above the
fall
do
this causes
a severe
N shares are just N-goods, and the value
Dividend on
this depreciation resulting in
Inability to
on future
a
fall
in share prices.
Note that
this fire
caused by a depreciation in both e\ and e2 due to the
T-goods to finance a shock. 49
Consider next the externality that leads to too
underinvestment in the T-sector. The logic
now is
little
international collateral creation
exactly as before. In the
fire
and
sale region,
provisions of international collateral will receive a yield of L. However, the social return
still
is
&.
is
Divergence in these returns leads to the externality. Investment in N-sector that
international collateral
date 1 N-goods that
is
is
beneficial,
however, this investment
is
packaged with a flow of
only domestic collateral.
Banking
6.5
This subsection sketches more formally the role and potential collapse of a banking system.
Let us introduce banks whose main role
date
and date
1
and
is
that they can invest in the N-sector at either of
raise foreign funds against these assets.
Thus banks are a conduit
through which foreigners can invest in the N-sector. However in order to raise funds from
foreigners at date 0,
we assume that banks must standby to provide liquidity in case
(date 1) withdrawals by foreigner. This
<9
An
a
accomplished by selling foreigners a claim with
extension of the above by adding consumption
KiyotaM and Moore (1997) may amplify
in
is
fire sale, this
in turn will
(the share price) will
mean
of early
this price drop.
and dividends
at dates
The exchange rate
d €
[1,2]
along the lines of
at each date will be depreciated
that dividends will be less valued at each date and their present value
fall.
42
50
One way
of thinking about short versus
the option to redeem this claim at par at date
l.
long term debt, a central issue during crises,
to think of the former as foreign investments
is
with an option to redeem.
The bank
holds kn6B, n shares of the N-sector at date
whose face value
withdraw at date
economy
The
is
RnKi^B^/^
1,
for T-shares
Against
in units of date 2 T-goods.
the bank simply
and
0.
If
this, it issues
claims
any foreigner wishes to
N-shares to another firm in the domestic
sells its
gives these T-shares to the foreigner.
assets of the bank, in units of T-shares is therefore,
Pkn6B n
,
while the liability of the bank at date 1 in units of date 1 T-shares
n
S-k
exit
Btn
is,
= PK6B,n
—.
A
fire sale
occurs
if
P
falls
below P.
Each
he does not withdraw funds, the bank remains solvent, as the
appreciation between date 1 and date
investors
bank
withdraw funds, then he too
fails at
date 1 and
is
2.
will
.
-
.?
foreign investor then has a choice.
However,
if
fall
in
P
greeted by an
the investor conjectures that other
withdraw funds. This has two
therefore unable to provide
is
If
effects. First
any additional funds to the
the
N sector.
Second, the foreign withdrawals imply that banks need to buy T-shares, which further
contracts the supply of scarce international collateral
50
and the
fuels the fire sale process.
These contractual features follow the model of Diamond-Dybvig (1983). See that model
of the optimality of the banking contract.
43
for
a justification
References
[1]
Aizenman, Joshua, "Country Risk, Incomplete Information and Taxes on International
Borrowing," Economic Journal 99, March 1989, pp. 147-161.
[2]
Bardhan, Pranab K., "Optimum Foreign Borrowing,"
the Theory of Optimal
in Karl Shell, ed.,
Economic Growth, Cambridge, Mass.,
MIT
Essays on
Press, 1967, pp.
117-128.
[3]
Bhattacharya, Sudipto, and Gale, Douglas M., "Preference Shocks, Liquidity, and Central
to
Bank
Policy", Barnett, William A., ed.; Singleton,
Kenneth
J., ed..
New pproaches
monetary economics: Proceedings of the Second International Symposium in Eco-
nomic Theory and Econometrics.
Econometrics
series
Cambridge;
International Symposia in Economic Theory and
New York and Melbourne: Cambridge University Press
1987, pp. 69- 88.
[4]
Borensztein, Eduardo, and Atish R. Ghosh, "Foreign Borrowing and Export Promotion,
[5]
IMF Staff Papers,
36-4,
December 1989,
Bulow, Jeremy and Rogoff, Kenneth,
"A
p. 904-932.
Constant Recontracting Model of Sovereign
Debt" Journal of Political Economy 97-1, February 1989, pp. 155-78.
,
[6]
Calvo, Guillermo A., "Balance of Payment Crises in Emerging Markets," mimeo, Maryland,
[7]
January 1998.
Cavallo,
Domingo, "A Comment," in
Latin America, ed. by R.
Volatile Capital Flows:
Hausmann and
L. Rojas-Suarez,
Taming Their Impact on
IDB, Washington DC, 1996,
pp. 45-48.
[8]
Chan-Lau, Jorge A., and Zhaohui Chen, "Financial
Crisis
and Credit Crunch as a
Results of Inefficient Financial Intermediation-with Reference to the Asian Crisis,"
mimeo IMF, August
[9]
Chang, Roberto and Andres Velasco, "Financial Crisis
ical
[10]
Model,"
mimeo Atlanta FED, July
in
Emerging Markets:
1998b.
A Canon-
1998a.
Chang, Roberto and Andres Velasco, 'The Asian Liquidity
FED, July
[11]
1998.
Crisis,"
mimeo Atlanta
''**
'
Diamond, Douglas W. and Dybvig, Philip H., "Bank Runs, Deposit Insurance, and
Liquidity," Journal of Political
Economy
91-3,
44
June 1983, pp. 401-19.
[12]
Cole, Harold L.,
Debt
[13]
Crisis,"
French,
and Timothy
J.
Kehoe, "A
Self-Fulfilling
Model of Mexico's 1994-95
Minneapolis FED, Research Staff Report 210, April 1996.
Ken and James
Poterba, "Investor Diversification and International Equity
Markets," American Economic Review 81, 1991, pp. 222-226.
[14]
Gale, Douglas and Martin Hellwig, "Incentive Compatible
Debt Contracts: The One-
Period Problem" Review of Economic Studies 52(4), October 1985, pp. 647-63.
[15]
and Marco Pagano, "Confidence Crises and Public Debt Manage-
Giavazzi, Francesco
ment," in Public Debt Management: Theory and History, ed Rudiger Dornbusch and
Mario Draghi. Cambridge Univesity Press, p.94-118.
[16]
Han and Rodrigo O.
Goldfajn,
Intermediation,"
[17]
IMF
Valdes,
"The Twin Crises and the Role of Financial
mimeo, August 1998.
Harberger, Arnold, "Lessons for Debtor-Country Managers and Policymakers," in
J.Cuddington and G.Smith
(eds.), International
Debt and the Developing Countries.
Washington: World Bank, 1985.
[18]
Holmstrom, Bengt and
of Political
[19]
Economy
Tirole, Jean, "Private
106-1, February 1998, pp. 1-40.
"Demand
Jacklin, Charles J.,
and Public Supply of Liquidity," Journal
Prescott,
Edward C,
ral trade.
Minnesota Studies
ed.;
Deposits, Trading Restrictions,
and Risk Sharing,"
Wallace, Neil, ed.. Contractual arrangements for intertempoin
Macroeconomics
series, vol. 1
Minneapolis: University
of Minnesota Press 1987, pp. 26-47.
[20]
International
Prospects,
Monetary Fund,
and Key
Policy," in
Charles Adams, Donald
[21]
J.
Markets:
Capital
Developments,
World Economic and Financial Surveys, directed by
Mathieson, Garry Schinasi, and
Bankim Chadha,
1998.
Johnson, Simon, Peter Boone, Alasdair Breach, and Eric Friedman, "Corporate Governance in the Asian Financial
[22]
"International
Kang, Jun-Koo and Rene M.
Portfolio Equity
Crisis, 1997-98,"
Stulz,
MIT
mimeo, September 21, 1998.
"Why Is There Home
Bias?
An Anlysis
of Foreign
Ownership in Japan," Journal of Financial Economics 46-1, October
1997, pp. 3-28.
[23]
Kiyotaki, Nobuhiro
and Moore, John, "Credit Cycles," Journal of Political Economy,
105-2, April 1997, pp. 211-48.
45
MIT LIBRARIES
3 9080 01444 0694
[24]
Krishnamurthy, Arvind, "Collateral Constraints and the Credit Channel," mimeo MIT,
January 1998.
[25]
Krugman,
Paul,
"International
J.Cuddington and G.Smith
Debt
Strategies
(eds.), International
in
an
Uncertain
World,"
in
Debt and the Developing Countries.
Washington: World Bank, 1985a.
[26]
Krugman, Paul, "Prospects
for International
Debt Reform,"
in International
Monetary
and Financial Issues for the Developing Countries, Geneva: UNCTAD., 1985b.
[27]
Krugman, Paul, "What Happened to Asia?"
[28]
La Porta,
Political
and Robert Vishny, "Law
Economy, forthcoming, 1998.
Radelet, Steven and Jeffrey Sachs, "The Onset of the East Asian Financial Crisis,"
mimeo HUD, February
[30]
mimeo, 1998.
Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer
and Finance," Journal of
[29]
MIT
1998.
Sachs, Jeffrey, Theoretical Issues in International Borrowing. Princeton Studies in International Finance, #54, Princeton University, July 1984.
[31]
Simonsen, Mario H., "The Developing-country Debt Problem," 1985.
[32]
Shleifer,
Andrei and Robert Vishny, "Liquidation Values and Debt Capacity:
A Market
Equilibrium Approach," Journal of Finance 47-4 September 1992.
May
[33]
World Economic Outlook, IMF,
[34]
Zhou, Chunseng, "Dynamic Portfolio Choice and Asset Pricing with Differential Infor-
1998.
mation," Journal of Dynamics and Control, July 1998.
46
'
/
'
C
3
8
Date Due
Lib-26-67