Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium IVIember Libraries
http://www.archive.org/details/copingwithchiles00caba2
c
L
j
15
oi-
I
3
'-1^
Technology
Department of Economics
Working Paper Series
Massachusetts
Institute of
COPING WITH CHILE'S EXTERNAL VULNERABILITY:
A FINANCIAL PROBLEM
Ricardo
J.
Caballero
WORKING PAPER
01-23
REVISED, January 2002
Room
E52-251
50 Memorial Diive
Cambridge,
MA 021 42
paper can be downloaded without charge from the
Social Science Research Network Paper Collection at
This
http://papers.ssrn.
o m/abstract=274083
Dewey
\(y
Technology
Department of Economics
Working Paper Series
Massachusetts
Institute of
COPING WITH CHILE'S EXTERNAL VULNERABILITY:
A FINANCIAL PROBLEM
Ricardo
J.
Caballero
WORKING PAPER
01-23
REVISED, January 2002
Room
E52-251
50 Memorial Drive
Cambridge, MA 02142
This
paper can be downloaded without charge from the
Social Science Research Network Paper Collection at
http://papers.ssrn. conn/abstract=274083
'.GHUSETTS INSTITUTE
OFTECHMOLOGY
H'AR
7 2002
LiBF?ARiES
Coping with Chile's External Vulnerability:
A Financial
Problem
01/11/2002
Ricardo
J.
Caballero'
Abstract
With traditional domestic imbalances long under control, the Chilean business cycle
Most
driven by external shocks.
A
financial problem.
a decline
in real
importantly, Chile
's
external vulnerability
decline in the Chilean terms-of-trade, for example,
GDP that is many times larger than
one would predict
is
is
is
primarily a
associated to
in the presence
of
perfect financial markets. The financial nature of this excess-sensitivity has two central
dimensions: a sharp contraction
it
needs
it
the most;
and an
borrowers during external
and argue
Chile's access to international financial markets
in
crises. In this
paper I characterize
that Chile's aggregate volatility can
this
financial mechanism
be reduced significantly by fostering the
private sector's development offinancial instruments that are contingent on Chile
external shocks.
As a
when
inefficient reallocation of this scarce access across domestic
first step, the Central
Bank or
IFIs could issue a
's
main
benchmark
instrument contingent on these shocks. I also advocate a countercyclical monetary policy
but mainly for incentive
— that
is,
as a substitute for taxes on capital inflows and
equivalent measures — rather than for ex-post
'
MIT
Banco
Loayza
and
NBER.
e-mail: caball(a)mit.edu; http://web.mit.edu/caball/www.
Central de Chile.
for their
liquidity purposes.
I
comments;
am
to
Prepared for the
very grateful Vittorio Corbo, Esteban Jadresic, and
Adam Ashcraft, Marco
for excellent research assistance;
and
required data. First draft: 03/06/2001.
to
Norman
Morales and especially Claudio Raddatz
Herman Bennett
for his effort collecting
much of the
Introduction and Overview
I.
With
domestic imbalances long under control, the Chilean business cycle
traditional
extemal shocks. Most importantly, Chile's extemal vulnerability
A decline
is
many
Chilean terms-of-trade, for example,
in the
is
is
driven by
is
primarily a financial problem.
GDP that
associated to a decline in real
times larger than one would predict in the presence of perfect financial markets. The
of
financial nature
this excess-sensitivity
has two central dimensions: a sharp contraction in
when
Chile's access to international financial markets
needs
it
it
the most;
reallocation of this scarce access across borrowers during extemal crises.
I
and an
inefficient
argue that Chile's
aggregate volatility can be reduced significantly by fostering the private sector's development of
on the main extemal shocks faced by
financial instruments that are contingent
Bank
step, the Central
—
that
is,
a
first
or IFIs could issue a benchmark instrument contingent on these shocks.
also advocate a countercyclical
incentive
As
Chile.
I
monetary pohcy (also contingent on these shocks) but mainly for
as a substitute for taxes
on
and equivalent
capital inflows
— rather than for
ex-post hquidity purposes.
The essence of the mechanism through which extemal shocks
be characterized as follows:
First, there is
for extemal resources if the real
economy
latter
to continue unaffected, but this triggers exactly the
who puU back
capital inflows. Occasionally, the
occurs directly as part of "contagion" effects. Second, once extemal financial markets faU
accommodate
the needs of domestic fiimis and households, these agents turn to domestic
financial maricets,
and to commercial banks in
matched by an increase
domestic
in supply as
banks
as
—
firms find
large
it
more
financial markets at
crisis,
of
this
any
when major
crisis is
to
is
not
tighten
are high.
Widespread
forced adjustments are needed.
rise
a
Limit
is
seek financing in domestic markets, in
medium
size firms cannot access intemational
in
expected returns,
symptom of these
The sharp
decline in domestic asset prices,
driven by the extreme scarcity in financial
is
Even
financial constraints are binding during the
the praised rise in
fire sales.
The
FDI
Chile's
integration to
that
occurred during the
fact that this investment takes the
control-purchases rather than portfoUo or credit flows simply reflects
that
-
domestic financial system, which reinforces the above
resources and their high opportunity cost.
problems
demand
price.
mechanism
and corresponding
most recent
Again, this increase in
particularly resident foreign institutions
attractive
circumstances that the displaced small and
costs
particular.
opting instead to increase their net foreign asset positions. Third, there
credit,
significant "flight-to-quaUt}^' within the
effects
The
Chilean economy can
a deterioration of terms-of-trade that raises the need
is
opposite reaction fi^om international financiers,
to
affect the
intemational
financial
markets:
weak
corporate
governance and other "transparency" standards. Finally, the costs do not end with the
financially distressed firms are iU
form of
some of the underlying
equipped to mount a speedy recovery. The
latter
crisis,
often
as
comes
not only with the costs of a slow recovery in employment and activity, but also with a slowdown
in the process
of creative destmction and productivity growth. In the U.S., the
account for about
1998),
which
is
30%
latter
may
of the costs of an average recession (see Caballero and Hammour,
probably a very optimistic lower bound for the Chilean economy. Moreover,
number of
the presence of a large
when
financially distressed small firms,
severe enough,
reduces rather than enhances the effectiveness of monetary policy in facihtating the recovery
(see below).
The
distnbutional impact of this type of crisis
is
On
significant as well.
one end,
large firms are
directly affected by external shocks but can substitute most of their financial needs domestically.
They
by demand
are affected primarily
PYMES)
(henceforth,
are
are the residual claimants
factors.
On
the other, small
and medium sized firms
crowded out and are severely constrained on the
of the
financial crunch.
They
financial side.
-
This diagnosis points in the direction of a stmctural solution based on two building blocks. The
first
one deals with the
institutions required to foster Chile's integration to international financial
markets and the development of domestic financial markets. Chile
along this margin through
its
capital
the unavoidable slow nature of the above process,
liquidity
management
management of
strategy.
The
latter
intemational reserves
by
focus on the
latter
is
to design
must be understood
the central bank,
financial instruments that facilitate the delegation
I
is
making
significant progress
markets reform program. The second one, necessary during
an appropriate international
in terms broader than just the
and include the development of
of this task to the private
type of solutions in this paper.^
sector.
view the pohcy problem as one of
I
remedying a chronic private sector underinsurance with respect to external
outlining the
them. The
main sources of such problem and
first
the corresponding solutions,
one seeks to develop a key missing market.
benchmark "bond"
instalment should
that is
made
facilitate
contingent on the
main
I
I
After
crises.'
focus on
two of
propose the creation of a
external shocks faced
by
Chile. This
the private sector's pricing and creation of similar and derivative
contingent financial instruments.
In the second one,
I
discuss optimal monetary
poUcy and intemational reserve management
fi-om
an insurance perspective. Provided that the Central Bank of Chile has achieved a high degree of
inflation-target credibility, the optimal
response to an external shock
of reserves and an expansionary monetary pohcy. The
real
latter,
is
however,
with a moderate injection
is
unlikely to have a large
impact and indeed will imply a sharp short-Uved exchange rate depreciation. The main
benefit
of such pohcy
is
in the incentives
it
provides.
Its
main problem
it
is
time-
mechanism
at
work.
is
that
inconsistent.
Section
2
II
describes the essence of the extemal shocks and the financial
See Caballero (1999, 2001) for a discussion of the structural reforms aimed
the
development of financial markets. More importantly, Chile
is
in the
at
improving integration and
process of enacting a major capital
markets reform.
See Caballero and Krishnamurthy (2001a) for a formal discussion of this perspective.
.
Section in follows with a discussion of the impact of this
economy and
the different
economic
agents. Section
IV
mechanism on
discusses
the real side of the
pohcy options and Section
V
concludes.
The Shock and the Financial System
II.
In this section
economy.
I
illustrate the
mechanism though which
mechanisms around demand and supply
backbone of the Chilean
presented in
I
external shocks aflFect the Chilean
organize the discussion of the role played by these shocks and their amplification
I
Box
factors affecting the domestic
An
financial system.
banking system
overview of the Chilean financial
-
the
institutions is
1
focus on the most recent cychcal episode, as the changing nature of the Chilean financial
system makes older data
//.
1
less relevant.
The External Shock and its Impact on Domestic Financial Needs
The trigger of the mechanism
of-trade
instrument.
It is
apparent that Chile
(b) offers a better metric to
1996
is illustrated
in Figure
1.
Panel
(a)
shows the path of Chile's terms-
and the spread paid by prime Chilean instruments over the equivalent U.S. Treasury
cvirrent
was
severely affected
by both
at the
end of the 1990s. Panel
gauge the magnitude of these shocks. Taking the quantities firom the
account as representative of those that would have occurred in the economy
absent any real adjustment,
it
deterioration in terms-of-trade
and
translates
them
into
the dollar losses
interest rates. In 1998, these losses
of GDP, with similar magnitudes for 1999 and 2000.
associated to the
amounted
to about
2%
Box
The Chilean Banking
1:
Sector
Table A.2: Composition ofloansby use.
Banks
of financing
are an important source
Chilean
for
The
firms.
financing (table A.l)
is
and much
60%
similar to that of
Residential
the maturity
Trade
15%
10%
10%
US,
like the
more
Other
5%
of Chilean bank loans
structure
of total loans
Fracti(>n
Commercial
advanced European economies. Unlike the
latter,
Type of loan
of
composition
is
57%
concentrated on the short end:
Consumption
Source: Superintendencia de Bancos
of
e Instituciones
total
Financieras.
loans,
and
60%
of commercial loans, have a
maturity of less than
1
year.
summary,
In
Chilean banks look like European banks in
terms of their relative importance but
banks
of
terms
in
on
focus
their
like
US
Composition of commercial loans by
firm
size.
short
Chilean banks concentrate their lending
maturities.
on
activity
size
Relative importance: Loans versus
65%
Around
other instruments.
large fums, at least
approximated
is
of
commercial loans
above
Table A.l: The relative importance of bank
US$
different
1
.
loan
The
is
by
firm
size.
volume
the
of
allocated to loans
importance of
relative
sizes
when
loan
is
summarized
in
loans.
Source of financing
Stock
Flows
Loans
43%
54%
51%
33%
3%
16%
Equity
Bonds
Note:
Participations
December 2000.
using
built
Flows
were
stocks
the
computed
as
Table A.3.
Table A.3: The relative importance of
different loan sizes.
at
Loan
in stocks between December
1999 and
December 2000 except for the flow of new equity,
which was built using information on equity placement
during the period.
on
activity
Medium
Micro
Note:
represent around
(see Table A.2).
70%
and
most of
Commercial
60%
of
Adding
total
their
loans
of total bank credit supply
loans above 50,000
US$
UF
(around
of 1999 dollars)); Medium:
1,400,000); Micro: loans
UF (US$
280,000
below 10,000 UF
(US$ 280,000).
Source: Superintendencia de Bancos
e Instituciones
Financieras.
bank loans
is
directed to
Banks versus other
financial
institutions.
Computed using
Source
all
trade loans, about
firms.
''
Large:
1.4 millions (end
loans between 10,000 and 50,000
concentrate
firms.
64%
14%
22%
Large
US$
Composition of Loans.
volume of
commercial loans
the
difference
Chilean banks
Fraction of the
size
the stock of loans in
Superinlendencia
de
Bancos
December
e
1999.
Jns/i/imones
Financieras.
In Germany, for example, more that 70% of the
commercial loans are long term. In contrast, in the US
short term loans account for about 60% of non^
residential loans.
Banks in Chile are important when
compared with other financial institutions
as well. Table A.4 compares the total
assets
of banks,
pension
fiinds,
and
represent
Banks'
companies.^
insurance
60%
of
assets
total assets.
Table A.4: Banks and other financial
institutions.
Banks
Pensio
Insurance
n
companies
Funds
Relative
60%
30%
10%
asset level
Note: Ratios based on the stock of
assets
at
December 2000.
Sources:
Superintendencia
Bancos
de
e
Instituciones
Financieras, Superintendencia de Valores
y Seguros.
Domestic and Foreign banks.
banks
Foreign
presence
very
is
Table A.5 shows that loans
significant.
by foreign banks represented around
40%
of
total
loans
in
1999.
It
is
important to note that Chilean Banking
Law
does not recognize subsidiaries of
banks
foreign
as
part
of
main
their
headquarters.
Table A.5: Domestic and Foreign Banks
Relative importance
Domestic
Foreign
58%
42%
80%
11%
67%
(%) total loans)
Portfolio composition
Loans
Securities
Foreign assets
Reserves
14%
15%
6%
3%
4%
Note: Ratios based on December 1999 values.
Source: Superintendencia de Bancos
e Instituciones
Financieras.
Total assets
is
the
sum of financial and
fixed assets.
Fixed assets are not included for insurance companies.
Nevertheless, fixed assets represent a very small
fraction of financial institutions' assets.
'
Figure
(a) TefTTis ot
1:
External Shocks
Trade and Sovereign Spread
(b)
Terms
of
Trade and Interest Rate
effect (fixed quanlities)
I1U1MJ
~>
\
—
^
'
Vy
—
/
—
V
.
V V ^s
\
;\/ "
'
_
~
_
-
•^
<•
^
Hi
A
fi
Hi
1
y
-4
/
._ ^ •
.
.
^
/',
f
<
^'"'
i
1
I
i-
1
1.
i
TemaoltfjdB
[
\
'^
I
'^
\
i
s
s
1
i
i
1
?
s
1
— -St»B3dEntnaEr»tea|
H TT eHecl
li.rnri.sl
ram BffT^
Notes: Preliminary data used for 1999 and 2000. The yearly Figure for 2000 was computed multiplying the S quarters
cumulative value by 4/3. (a) Terms-of-trade
is
the ratio
of exports price index and import price index computed by the
ENERSIS-ENDESA corporate bonds. The spread data
Central Bank. The sovereign spread was estimated as the spread of
for 1996 was estimated by the author using information from the Central Bank of Chile,
computed as the difference between the actual terms-of-trade and the terms-of-trade
effect was computed in similar fashion
at
(b)
The terms-of-trade
1996
quantities.
effect
was
The interest rate
Source: Banco Central de Chile.
How
should Chile have reacted to this sharp decline in national income, absent any significant
financial friction (aside
The
from the temporary increase
thin line depicts the path
in the spread)?
Figure 2 hints the answer.
of actual consumption growth, while the thick
hypothetical path of Chile's consumption growth if
it
line illustrates the
were perfectly integrated to
international
markets and terms-of-trade the only source of shocks.' There are two interesting
financial
features in the figure. First, there is a very high correlation
shocks to
its
terms-of-trade. This correlation
is
between Chile's business cycle and
not observed in other
commodity dependent
economies with more developed financial markets, such as Australia or Norway (see Caballero
1999 or 2001). Second, and more importantly for the argument in
response of the
economy
measured on the
it is
is
being measured on the
right axis. Since the scale in the
apparent that the
economy
left
is
being
ten times larger than that of the
latter,
over-reacts to these shocks
by a
significant margin.
shocks to the terms-of-trade are simply too transitory, especially
demand
of the hypothetical
is
axis while that
former
this paper, the actual
when
as in the recent crisis, to justify a large response of the real side
In practice
they are driven by
of the economy.^ The
Consumption growth is very similar to GDP growth for this comparison. Adding the income effect of
shocks would not change things too much as these were very short-lived.
I
neglect the
presence of substitution effects because there is no evidence of a consumption overshooting once
interest rate
international spreads
come down.
income
measured
effect that is
of Figure
in panel (b)
practice, translate almost entirely in increased
Figure
2:
1
should in principle, but
it
does not in
borrowing from abroad.
Excess Sensitivity of Real Consumption
Growth
to
Terms-of-Trade Shocks
13BS
1SB7
PV^g^
af.Brtif<Wl7W.m«in)
I
Notes: consumption growth from IFS, copper prices (London
Metal Exchange) from Datastream, copper exports from Min.
de Hacienda.
Returning to the
late
1990s
crisis,
panel (a) of Figure 3
illustrates that international financial
markets not only did not accommodate the (potential) increase in demand for foreign resources
but actually capital inflows declined rapidly over the period. Panel (b) documents that while the
central
bank
clearly
was not
offset part
of
this decline
by
injecting
back some of
its
international reserves,
it
nearly enough to offset both the decline in capital inflows and the rise in external
needs. For example, the figure shows that in 1998, the injection of reserves amounted to
approximately
US$2
and
in terms of trades
inflows that
'
The
bilhon dollars. This
came with
is
comparable to the
but
rise in interest rates,
it
direct
income
effect
of the decline
does not compensate for the decline in capital
these shocks.
price of copper has trends and cycles at different frequencies,
some of which
are persistent (see
Marshall and Silva, 1998). But there seems to be no doubt that the sharp decline in the price of copper
during the late 1990s crisis was mostly the result of
and Russian
crises.
demand shock,
copper
1
would argue
demand shock brought about by the Asian
that the current shock was a transitory
the univariate process used to estimate the present value impact of the decline in the price of
in Figure 2, overestimates the extent
of this decline. The lower decline
with this view. The variance of the spot price
Moreover, the expectations computed from the
is
in future prices is consistent
6 times the variance of 15-months-ahead future prices.
AR process track reasonably well the expectations implicit in
future markets but for the very end of the sample,
play.
a transitory
on the information
that conditional
when
liquidity
premia considerations
may have come
into
Figure 3: Capital Flows and Reserves
(a) Capital
Account except Reserves
(b)
BabncE
of
Payments
ecoD
son
r
'i^'\
c-^
I
"EE
•1C00
-2.000
Source: Banco Centra! de Chile
How
severe
was
mismatch between the increase
the
financial resources? Figure
in
needs and the availability of extemal
4a has a back-of-the-envelope answer.
account (light bars) and the current account with quantities fixed
the trough, in 1999, the actual current account deficit
what one would have predicted using only
at
bars),
in panel (b),
at
1996 levels (dark
bars).
At
the actual change in terms-of-trade. Moreover, this
dollar adjustment underestimates the quantity adjustment behind
and the current account
graphs the actual current
was around US$4 bUhon smaller than
it,
of-trade typically worsens the current account deficit for any path
of this price-correction can be seen
It
as the deterioration of terms-
of quantities. The importance
which graphs the actual current account (hght
1996 prices (dark
bars).
Clearly, the latter
shows a
significantly larger adjustment than the former.
Figure 4: The Current Account
(a)
CunentAccajnt and CuiTent Account
at
(b)CunGnt Account
1996 quantities
(actual
and
h
19g6pnc8s}
S^
t599
M CjTtrt /•a
Notes:
(a)
USS\» Qjignt Arc |m« USl
2000
96il
The current account at 1996 quantities was computed by multiplying the 1996 quantities by each year's price
indices for exports, imports
(b)
Invi
The current account
and
in 1
interest payments
.
The other components of the current account were kept at current values,
the 1 996 prices by each year 's quantities. Source:
996 prices was computed by multiplying
Banco Central de Chile
Figure 5 reinforces the mismatch conclusion by reporting increasingly conservative estimates of
the shortage of extemal financial resources
by the non-banking
sector.
Panel
(a) describes the
path of the change in potential financial needs stemming from the decline in temis-of-trade and
capital inflows,
by
and the increase
in spreads. Panel (b) subtracts
the central bank, while panel (c) subtracts
banking
sector.
We
can see from
from about $2 bilhons
to just
from
financial system.
extensively,
show
The
(a) the injection
dechne
above $7
billions.
needs for 1999 ranges
Regardless of the concept used, the mismatch
demand
for resources
from the domestic
difference between panel (b) and (c), as well as the next section
that the latter sector not only did not
accommodate
this increase in
but also exacerbated the financial crunch.
Figure
(s)
5:
The Increase
in Potential Financial
CuTprfiensTi/e
Needs
(b) {a)
minus reserves accumulation
1
1
6.0X
-S?:;-:C
4,000
S
w
^^^^^^^^^^
-2.000
(c) (b)
10
of reserves
in capital inflows to the
this figure that the increase in financial
large, creating a potentially large surge in the
is
(b) the
from
minus flews
to
banking setior
more
demand,
corresponds to the sum of the terms-of-trade effect, the interest rate effect, and the decline in capital
(a)
Both the terms-of-trade and the interest rate effects are measured at 1996 quantities. The decline in capital flows
Notes: Panel
inflows.
corresponds
(a) the net
sector.
To
to the
change
difference of the capital account except reserves with respect to
in
Central Bank reserves. Panel
(c)
subtracts
from panel
its
1996
value.
(b) subtracts
flows
to the
from
banking
Source: Banco Central de Chile
conclude
smooth
demand
the
Figure 6
this section.
illustrates
other dimensions that could have, but did not,
from resident banks. Panel
for resources
"good" news. The former panel shows
that issues
and
(a)
(b) carry the relatively
of new equity and corporate bonds did not
decline very sharply, although they hide the fact that the required retum
rose significantly.
The
latter
panel
illustrates that
while the
flmds to foreign assets during the period, this portfoho
instruments.
some
Panel
(b) the net increase in
However,
this decline
APPs
on these instruments
increased their allocation of
occurred mostly against public
shift
probably translated into a large placement of public bonds on
other market or institution that competes with the private sector, like banks (see the
discussion below). Panel
(c),
on
the other hand, indicates that retained earnings, a significant
source of investment financing to Qiilean firms, declined significantly over the period.^ Finally,
panel (d)
illustrates that
workers did not help accommodating the financial bottleneck
either, as
the labor share rose steadily throughout the period.
Figure
{a)
Change
in
the stock of corporate debt
_
199G
>998
Before the
firm.
in
AFP
holdings as fracfion of
GDP
|_
1999
crises, in 1996, the stock
50%
(b)
m PubBe •
oulsUntHno corporaia dflbl
The difference between
represented around a
11
ChanuB
Other Factors
equity
I
1997
IB New equJN
and new
6:
profits
Financial
a
I
20% of total assets for the median
measure of the flow of retained earnings,
of retained earnings represented
and dividend payments,
a ComMriw B FCTCtiji
of total capital expenditures for the median firm
in 1996.
Retained Earnings
(c)
I
I
M
10%'
=
.1
IH
'
'^
;;;^;,
'
iDFtow/Opcfabon
Notes:
(a)
-.
i_
iSiodt/ToB)
|
Source: Bolsa de Comercio de Santiago and Superintendencia de Valores y Seguros.
New
equity
issuance price, and deflated using CPI. The stock of corporate debt outstanding was also deflated by CPI. The
valued at
December
1998 exchange rate was 432 pesos per dollar, (b) Source: Superintendencia de AFP. The values are expressed as a fraction
of the year 's GDP. (c) The data on retained earnings are from Bolsa de Comercio de Santiago (Santiago Stock Exchange)
and correspond to the stocks of retained earning reported in the balance sheets of all the societies traded in thai exchange.
They exclude investment societies, pension funds, and insurance companies. The flow value was computed as the difference
in the (deflated)
The bottom
stocks at
line
I
and
t-l.
(d)
of this section
Source: Banco Central de Chile.
is clear:
dxiring external
cnmches, firms need substantial additional
financial resources fi-om resident banks.
11.2
The (Supply) Response by Resident Banks
How
did resident banks respond to increased
demand? The
short
answer
is
that they
exacerbated rather than smooth the external shock.
Panel
(a) in
late 90s,
Figure 7 shows that domestic loan growth actually slowed
even as deposits kept growing. This tightening can also be seen
through the sharp
While spreads
makes
12
it
down
rise in the loan-deposit
started to fall
by 1999,
in prices in panel (b),
spread during the early phase of the recent
there
difficult to interpret this decline as
sharply during the
was a strong
substitution
crisis.
toward prime firms
a loosening in credit standards (see below).
that
1
1
Figure?: The Credit Crunch
{a)
(b) Inlerest
Loans and Deposits: Banking Syslem
*»
/'""•'
'
Rate Spreads (Annual %}
"
s
-—^->'
1
""^
-=^
i
a
/^
_.^--^^
s • s $ s s s
i
• 1
1 2
5
S s
8 8 8
HI 11 it H U 1
^ 1
5
Jl
-l.CWC'---DB'OSns
Notes:
(a)
Loans: commercial, consumption, trade, and mortgages. Deposits: sight deposits and time deposits. The values
were expressed in dollars of December 1998 using the average December 1998 exchange rate (472 pesos/dollar), (b) The
30-day interest rates are nominal while the 90-365 are real (to a first order, this distinction should not matter for spreads
calculations ).
Sources: Banco Central, and Superintendencia de Valores y Seguros.
The main
substitutes for loans in banks' portfolios are public debt
alternatives are
toward external
domestic
shown
in Figure 8a, with the clear conclusion that
assets. Interestingly, the central
interest rates
of 1998). Instead, the
and external
was only temporarily
rise in interest rates
succeeded mostly in slowing
down
bank policy of fighting
assets.
banks moved
These
their assets
capital outflows
with high
successful (see the reversal during the last quarter
during the Russian phase of the
the decline in banks' investment in
crisis
seems
to
have
pubhc instmment, and
encouraged substitution away fi-om loans. Banks, rather than smooth the loss of external
funding, exacerbated
illustrates that the
it
by becoming
panel (c) reveals that
much of
of the
the net outflow
credit lines or inflows, but rather
13
part
capital outflows.
banking sector also experienced a large
due
to
capital
was not due
In fact, while panel (b)
outflow during this episode,
to a decline in their international
an outflow toward foreign deposits and
securities.
Figure
(a)
Inveslmanl
in
8: Capital
Outflows by the Banking System
Domestic and External Assets: Banking System
(b)
CZoomwIic
(c)
Notes:
(a)
assets are
Composition
of
—
Net capital rifbws to tjanking sector
Ettemall
Foreign Assets: Banking System
Domestic assets include Central Bank securities with secondary market and intermediated
the
sum of deposits abroad and foreign
securities. Deposits
securities. External
abroad are mostly sight deposits
in
foreign (non-
resident) banks. Foreign securities are mostly foreign bonds.
Sources: Banco Central and Superintendencia de Bancos.
While most
resident banks exhibited this pattern to
had the most pronounced portfoUo
and
(b) in Figure 9,
shift
assets, foreign
degree,
towards foreign
which present the investment
domestic (private) banks, respectively.
some
While
in
all
assets.
it
is
resident foreign banks that
This
is
apparent in panels (a)
domestic and foreign assets by foreign and
banks increased
their positions in foreign
banks' trend was substantially more pronounced. Moreover, domestic banks
"financed" a larger fiaction of their portfoHo
loans. This conclusion
is
ratios for foreign (dashes)
confirmed by panel
shift
(c),
by reducing other investments
rather than
which shows the paths of loans-to-deposit
and domestic (sohd) banks.'"
This figure excludes two banks (BHIF and Banco de Santiago) that changed firom domestic to foreign
ownership during the period. The change
BHIF, and
place in
in July
May
were going
to
in statistical classification
took place in
November 1998
for
Banco
1999 for Banco de Santiago. In the case of Banco de Santiago, the takeover operation took
1999 by Banco Santander (the largest foreign bank
in Chile).
Assuming
that the
two banks
merge, the Chilean banks regulatory agency (Superintendencia de Bancos) ordered Banco de
Santiago to reduce its market presence (the two banks had a joint presence equivalent to 29% of total loans).
The banks finally decided not to merge and therefore their joint market presence has remained unaltered so
far. Most likely the confusion created by the potential participafion limits did not help the severe credit
crunch that Chile was experiencing at the time.
14
-
Figure 9: Comparing Foreign and Domestic
(a)
Investment
in
Domestic and External Assets Foreign Banks
(b)
Bank Behavior
Investment
in
'•.
Domestic and External Assets: Domestic Banks
'"-
.'•
.
'"
/'•'..-:'
-./vv
—
A
/\.
.
^
J\y^^
'
^^s__—x^-.-^-^
SSsS3SSSSSgsS?|SS
4 1 J - 1 1 i s 1 1
Exlomrf -
[
(c)
Loans over Deposits domestk: and foreign banks except BHIF
-
•
DomesBJ
arxJ
Santiago
I
I
=
5
S
I
3
I
=
«
=
I
S
Notes: Domestic banks do not include
I
5
I
Banco
del Estado.
(a), (b)
Domestic assets include Central Bank securities with
secondary market and intermediated securities. External assets are the sum of deposits abroad and foreign securities.
All constant 1998-peso series were converted to dollars using the average exchange rale during December 1998 (472
pesos/dollar),
Loans-to- deposits ratios are normalized to one in
(c)
March
Loans: commercial, consumption, and trade loans, and mortgages. Banco
changed from domestic
these two banks
The question
is
from
arises
to foreign
the Bolsa
whether
1996. Deposit: sight
BHIF and Banco
during the period, were excluded from the sample. The information used
it is
not the nationality but other factors
banks did not experience a
account for additional hedging.
significantly sharper rise in
by a
15
to
separate
(e.g., risk characteristics)
is
explored more thoroughly, the evidence in Figure 10 suggests that this
that foreign
deposits.
de Valores de Santiago (Santiago Stock Exchange).
the banks that determined the differential response. While this
shows
and time
de Santiago, which
rise in net foreign
certainly a
is
tighter capital-adequacy constraint in panel (c).
is
indicated
to
be
not the case. Panel (a)
currency liabiUties that could
When compared with domestic banks, they
non-performing loans as
theme
of
by panel
did not experience a
(b),
or were affected
Figure 10: Risk Characteristics of Foreign and Domestic Banks
(a)
Foreign Currency Assels and
Foreign Banks
Liabilities:
(b)
Non performing
loans as percenlage ot
total loans:
domestic and
foreign tianks
^^
1
rJ
™n
f
rl
,—y~^^~^
/s^^
_^--N
p-.
••
,
1
-•.-•-.....,..--•
1
i 2 s 1
1
1 i 1 3 1
—
r
1
S g s S S
s
1
(c) Capital
ini
i 1
-AMfltt
5
S
ill!?!!
5 5 5 1 1
S
=
•UnbBifiei
adequacy
ratio
,^i^
TTT
"TTTTTTTT
Notes:
(a)
Foreign
assets: foreign loans, deposits
abroad and foreign
abroad and deposits in foreign currency, (b) ROE: ratio of
Domestic private banks do not include Banco del Estado.
securities.
total profits
Foreign
liabilities:
funds borrowed from
over capital and reserves (as a percentage).
Source: Superinlendencia de Bancos e Instituciones Financieras.
Foreign banks usually play
many
useful roles in emerging economies but they
helping to smooth extemal shocks
(at least in the
most recent
crisis). It is
do not seem to be
probably the case that
these banks' credit and risk strategies are being dictated from abroad, and there
is
no
particular
reason to expect them to behave too differently from other foreign investors during these times.
In summary, during 1999 -perhaps the worst fuU year during the crisis
growth by approximately
US$2
banking private sector was about
billions,"
US$2
while the increase in financial needs by the non-
billions as well (see figure 5c).
general equilibrium issues ignored in these simple calculations,
US$4
"
it
is
Although there are
biUions (about
5-6% of GDP)
large,
perhaps around
in that single year.
A number close to two billion dollars is obtained from the difference between the flow of loans
1999 (computed as the change
flow of loans
16
in
1
in the stock
996 (computed
many
probably not too far-fetched
add them up and conclude that the financial crunch was extremely
to
— banks reduced loan
during
of loans between December 1999 and December 1998) and the
in a similar manner).
The Costs
III.
The impact of the
and medium
financial crunch
on the economy
is
correspondingly large, especially on small
size firms.
The Aggregates
111.1
Figure 2 already summarized the first-order impact of the financial mechanism, with a domestic
business cycle that
many times more volatile
is
than
it
would be
if financial
markets were perfect.
This "excess volatility" was particularly pronounced in the recent slowdown, as
panel (a) in Figure
1 1
move
only appear to
together, but the latter adjusts
largely transitory terms-of-trade
unemployment
sharp rise in the
(a} National
shock
is
11
:
more than
by
(dashes) not
implying that the
the former,
not being smoothed over time. Panel (b) illustrates a
rate that has yet to
Figure
95000
is illustrated
The paths of national income (soUd) and domestic demand
.
be undone.
Output, Demand, and Unemployment
Income and Domestic Demand
(b)
Unemployment
rate
.
'~
-
'
-
^
.
-<.^
BOOOO
.•"
1
M
,
^/
L-^-^
,
^
•
^^.^
^^
-^^T"^.^^-^
E
65000
.
60000
.
5O0O0
,
.
-'(OQl-'inQ
5
-
Notes:
(a)
-
3^mDl-'in05
-•
c.tH;
Series are seasonally adjusted
and annualized. Source: Banco Central de
Chile, (b) Source: INE.
Aside fi^om the direct negative impact of the slowdown and any additional uncertainty that
may
have been created by the untimely discussion of a new labor code, the build up in
unemployment and
persistence can be linked to
its
mechanism described above.
exchange
this,
rate
of Figure
financial resources
(see below).
While
line) suffered a
The
larger real
12.
additional aspects
of the financial
I
falls
on
creation, a
crisis
phenomenon
and
a result of
into the recovery, the lack
that is particularly acute in the
do not have job creation numbers, panel
(b)
deeper and more prolonged recession than the
exchange
As
the labor-intensive non-tradable sector, as illustrated
Second, going beyond the
hampers job
and Krishnamurthy (2001a).
17
two
the presence of an external financial constraint the real
needs a larger adjustment for any given decline in terms-of-trade.'^
a big share of the adjustment
in panel (a)
'^
First, in
rate adjustment corresponds to the dual
shows
rest
of
PYMES
that investment (solid
of domestic aggregate
of the financial constraint. See Caballero
demand (dashed
the
line).
This
is
important beyond
its
impact on unemployment, as
slowdown of one of the main engines of productivity growth:
U.S.
is
any
optimistic
indication of the costs associated to this
lower bound for the Chilean costs
productivity loss that amounts to over
slowdown
—
this
it
also hints at
the restmcturing process. If the
in restmcturing
—
possibly a very
mechanism may add a
30% of the employment cost of the recession.
significant
'^
Figure 12: Forced and Depressed Restructuring
(a)
GDP
tratjeable
and norvlradeable
(b) Evohjlion of
t-
.
flomestrc
RCS1
demand
—=FBgl
Series are seasonally adjusted and normalized to one in March 1996. Tradable sector: agriculture, fishing,
and manufacturing. Non-lradable sector: electricity, gas and water, construction, trade, transport and
telecommunications, and other services, (b) Series are seasonally adjusted, and annualized. Source: Banco Central de
Notes:
(a)
mining,
Chile.
As
the external constraint tightens,
must
and hquidity
to
domestic assets that are not part of international hquidity
buy them. Figure 13a shows a
market. Panel (b)
increased
assets.
most
all
lose value sharply in order to offer significant excess returns to the
is
perhaps more interesting, as
it
of this v-pattem
However, while FDI
scarce,
its
is
very useful since
presence during the
See Caballero and Hammour( 1998).
it
investment (FDI)
to the fire sale
provides external resources
crisis also reflects the severe costs
heavy discounts.
will
in the Chilean stock
illustrates that foreign direct
by almost $4.5 bilhon during 1999, and created a bottom
constraint as valuable assets are sold at
18
clear trace
few agents with the
of domestic
when
they are
of the external financial
Figure 13: Fire Sales
Value
(a)
of Chilean stocks (index)
(b)
Foreign Drecl Inveslment (FDI)
"^'
i
llf-lMll-ll
—
I
IGPA
Notes: (a)
is
1997
1996
r^AI
the general stock price index
1999
199B
2000
of the Santiago Stock Exchange. Sources: Bloomberg and Banco Central
de Chile
Gross
(b)
III.2
The
FDI
inflows. Source:
Banco Central de
Chile.
The Asymmetries
real
consequences of the external shocks and the financial amplification mechanism are
differently across firms
of different
sizes.
Large firms are
but can substitute their financial needs domestically.
demand
factors.
The PYMES, on
directly affected
They
the other hand, are
and
to a lesser extent, so are indebted
consumers
external shock
are affected primarily
by
price and
crowded out from domestic
markets and become severely constrained on the financial
-
by the
side.
—
They
felt
financial
are the residual claimants
-
of the financial dimension of the
mechanism described above.
Figure 14
illustrates
some
aspects of this asymmetry, starting with panel (a) that
of the share of "large" loans as an imperfect proxy for resident
Together with panel
(b),
which shows a sharp increase
at a substantial reallocation
I
win argue
19
that
of domestic loans toward
some of this
reallocation is Kkely to
be
shows the path
bank loans going
in the relative size
relatively large firms. In the
socially inefficient.
to large firms.
of large loans,
poUcy
it
hints
section
Figure 14: Flight-to-Quality
(a)
Share of brge toans
in totaJ
loans
(b)
Average size
Strml
I
Notes:
average size corresponds
UF;
larger than 50000
Large loans:
(a)
to the stock
of loans
in
Small loan:
greater than
400UF
of smatl
•
and large loarw
- •Largal
less than
but
10000 UF.
(b)
The
each class divided by the number of debtors.
Source: Superintendencia de Bancos e Instituciones Financieras.
The next
figure looks at cxDntinuing firms fixDm the
relatively large
the largest of
that larger firms fared better as
this period,
FECU
is
more
while
the
by
initial assets)
during the recent slowdown.
medium-sized firms saw
importantly, since the
total
acronym
sheet
of these firms being
that
for
every
shows the path of medium and
size ones. Figure 15
sum of loans
loans declined throughout the
particularly significant in smaller firms (those not in the
balance
all
pubhcly traded companies, one can already see important differences between
them and "medium"
firms' bank-Uabihties (normalized
1999. Perhaps
FECUs.''' Despite
their level
to
firm
in
Chile
is
large firms rose during
the contraction
must have been
FECUs).
Ficha Estadistica Codificada Uniforme, which
public
apparent
of loans fiozen throughout
medium and
crisis,
It is
large
required
to
is
report
the
to
name of the
the
standardized
supervisor
authority
{Superintendencia de Valores y Seguros) on a quarterly basis. Our database contains the information on
those standardized balance sheets for every public firm reporting to the authority between the first quarter
of 1996 and the
20
first
quarter of 2000.
.
Figure 15:
Bank Debt
(a)
Bank
of Publicly Traded Firms
liabilibes of
I
medium and
Mft.ftjm-
-
By
Size
large Hrnis
[jmal
Data source: FECU reported by all listed firms to the Superintendencia de
y Seguros. The sample includes all firms that continuously reported information
between December 1996 and March 2000. Within this sample, the medium (since they are
larger than firms outside the sample) size firms are those with total assets below the median
level of total assets in December 1996. The series correspond to the total stock of bank
Notes:
(a)
Valores
liabilities at all maturities
expressed
in
1996 pesos, divided by the
total level
of assets
in
December 1996). The nominal values were deflated using the CPl. The total level of assets
in December 1996 was VS$788 millions for the group of medium size firms, and
US$59,300 millions for the group of large firms
In summary, despite significant institutional development over the last
economy
is
stUl
two decades,
very vulnerable to external shocks. The main reason for
the Chilean
this vulnerability
appears to be a double-edged financial mechanism, which includes a sharp tightening of Chile's
access to international fmancial markets, and a significant reallocation of resources fi^om the
domestic fmancial system toward larger corporations, pubhc bonds and, most significantly,
foreign assets.
21
Policy Considerations and a Proposal
iV.
The previous
especially
IV. 1
diagnostic points at the occasional tightening of an external financial constraint
when terms-of-trade
— as the trigger for the costly financial mechanism.
General Policy Considerations
If this assessment
a)
deteriorate
~
for a stmctural solution
is correct, it calls
Measures aimed
at
based on two building blocks:
improving Chile's integration to intemational financial markets and
the development of domestic financial markets.
b)
Measures aimed
at
improving the allocation of financial resources during times of
extemal distress and across
While
in practice
states
of nature.
most stmctural poHcies contain elements of both building blocks,
as the guiding principle of the
pohcy discussion below because
already broadly understood and Qiile
dimension through
its
capital
akeady making
is
(a), at least
significant
maikets reform program. That
is,
I
I
focus on (b)
as an objective,
progress
is
along this
focus primarily on the
intemational liquidity management problem raised by the analysis above.'^
Intemational liquidity
management
is
primarily an "insurance" problem with respect to those
(aggregate) shocks that trigger extemal crises.
contingent nature.
large share
trade, the
The
first
step
is
to identify
Of course solutions to this problem must have
a — hopefidly small — set of shocks that capture
a
a
of the triggers to the mechanism described above. In the case of Chile, terms-of-
EMBI+, and weather
index chosen
is
variables, represent a
good
starting point, but the particular
a central aspect of the design that needs to be extensively explored before
it is
implemented.
The second
step is to identify the ex-post transfers (perhaps temporary loans rather than
outright transfers) that are desirable.
level these are simply:
From foreigners to domestic agents.
From less constrained domestic agents to more
i)
v)
At a generic
needs and
let
constiained ones.
level these transfers are clear, but in practice there
availability, raising the information
not already doing as
identifies the policy goals
much
as
is
great heterogeneity in agents'
requirements greatly.
the private sector take over the bulk of the solution,
this sector is
''
At a broad
is
It is
thus highly desirable to
which begs the question of why
is it that
needed. The answer to this question most Ukely
with the highest returns.
See Caballero (1999, 2001) for general policy recommendations for the Chilean economy, including
measures type
(a).
My objective in the current paper is instead to focus and deepen the discussion
subset of those policies, adding an implementation dimension as well.
22
of a
The main
suspects can be grouped into
Among the
a)
supply factors, there are
two
types: supply
at least three
and demand
factors.
prominent ones:
Coordination problems. The absence of a well-defmed benchmark around which the
market can be organized.
b)
These
Limited "insurance" capital.
financial constraint (in the sense that
by
coordination aspects as well,
and
missing
is
that
affect the payoffs
by
respect to aggregate shocks.
The
latter
of
b)
Sovereign problems. Imphcit
c)
Behavioral problems. Over-optimism.
at least three sources
The bulk of
factor (a)
of underinsurance:
depresses effective competition for domestically
crises,
reducing the private (but not the social)
'^
(free) insurance.
view, domestic financial underdevelopment
demand
and
is
still
to supply factor (b) as
a problem in Qiile, which gives
it
reduces the size of the effective
the solution to this problem should be pursued through financial market
reforms. In the meantime, however, an adequate use of monetary and reserves
poUcy may remedy some of the underinsurance impUcations of this
the general features of this
in Caballero
Similarly,
demand
in the next section.
deficiency. I briefly discuss
A more extensive discussion can be found
The
is
factor (c)
seems to be a pervasive human phenomenon (see
whose macroeconomic impUcations can
latter
problem, which
and neither
poUcy
management
and Krishnamurthy (2001 c).
Shiller 1999),
policy.
is
must be done with great care not
for
example
also be partly remedied with the above
to generate a sizeable
demand-(b) type
otherwise not likely to be present in any significant amount in the case of Chile
supply-(c).
This leaves us with supply-(a) as the main focus of the poUcy proposal, which
extensively in section TV. 3.
" See, e.g., Tirole (2000).
" See Caballero and Krishnamvirthy (200
23
parties,
of these contracts.'^
valuation of international liquidity during these times.
market.
"
This leads to a private undervaluation of insurance with
available external liquidity at times
my
due to a
be credibly pledged
A closely related problem, but with
the "insurers.
one should consider
Financial underdevelopment.
relevance to
assets that can
an insufficient number of participants in the market
collateral risk faced
Among the demand problems,
In
to structural shortages or
is
(dual-agency) problems. While contracts are signed by private
Sovereign
government actions
a)
is
the "insurer," rather than actual funds during crises).
raises the hquidity
c)
may be due
what
1
a, b).
I
discuss
IV.2
Monetary Policy as a Solution
to the
of an external
Figure 16 provides a stylized characterization
described up to now.
Underinsurance Problem
The main problem during an emerging market
the presence of a "vertical" external constraint. That
margin,
it
becomes very
financial markets at
of the
crises
any
difficult for
As
price.
is,
crisis is
to gain
any access to
this, international liquidity
domestic competition for these resources bid up the "doUaf' -cost of
by prime
international interest rate faced
crunch
is
a sharp
fall in
when,
at
the
international
becomes scarce and
capital,
(international) Chilean assets,
have
1
well captured by
external crises are times
most domestic agents
a result of
sort that
/*.
above the
f^,
The
dual of this
investment.
Figure 16: External Crisis
t
7*
I^
It is
important to clarify a few aspects of this perspective of
some confusion
a)
Particularly in the case
measure of the
b)
It
of Chile, the
rise in the
rise in
vertical constraint to
scarcity is
seldom the
perceive that there
is
case.
But
it is
literally
enough
Fleming type
;
rather,
it
not a
represents
runs out of intemational liquidity for the
that
some
crisis
may
the intemational liquidity rather than lend
that follows
i^- /*, is
seldom
new
it
lie
ahead, for
domestically,
positive.
vertical but '"diagonal." In that case the
must be blended with
analysis, but the interesting
total
prime firms, banks, and the government
a significant chance that a severe external
in practice the constraint is
of the analysis
f
is
bind for small and less than prime firms. In fact in Chile
to decide to hoard
Of course,
logic
to have caused
may tighten the vertical constraint.
even when the current domestic spread,
c)
seem
sovereign (or prime firms') spread
domestic (shadow or observed) rate
needs not be the case that the country
them
crisis that
in the past:
the rise in /*, which in turn
24
Investment
insight is
that
still
of the standard Mundell-
that
which
related to the non-
,
flat
aspect of the supply of funds.
Ex-post-Optimal Monetary and reserves policy during
crisis
Before discussing optimal policy fijom an insurance perspective,
bank faces during an external
incentives that a central
The main
shortage experienced
by
the country
is
it
is
worth highhghting the
crisis.
one of
international Uquidity (or "collateral,"
broadly understood). Thus, an injection of international reserves into the market
powerful
tool:
stabilizes the
To
see the
there
is
directly relaxes the "vertical" constraint
it
exchange
latter,
credit constraint.
crisis international arbitrage
Domestic
and dollar instruments backed by domestic
arbitrage,
collateral
f'
(1)
where f denotes de peso
rate for next period. In
when
this
a very
it
=f
expectation
is
I
take the
affected
on the other hand, must
hold. Peso
(risk
is:
+(e-Ee).
exchange
interest rate, e the current
what follows,
does not hold since
must yield the same expected return
aversion aside). Thus the domestic arbitrage condition
arise
is
in Figure 16, and
rate.
note that during an external
an extemal
on investment
rate,
by poUcy
and Ee the expected
interest
although interesting interactions
latter as given,
as well (see
CabaUero and Krishnamurthy
2001c).
Rearranging (1) yields and expression for the exchange
e~Ee=
(2)
which shows
that as
given peso interest
The
effectiveness
f
facilitates
H',
of an expansionary monetary pohcy. The reason for
crisis is the
domestic loans and hence the role of
lower peso
this is that
interest rates raise investment
Part of the reason for the "diagonal" shape in practice
is
f
in the investment function in
demand)
constraint.
As
but, to a first order,
25
it
a result, an expansionary
that the currency depreciates as
I
rises. If
exports are an important dimension of the country's intemational liquidity, then a depreciation increases
intemational collateral.
rate
lack of intemational not domestic hquidity.
does not relax the binding intemational financial
'*
and protecting the exchange
injections in boosting investment
main problem during an extemal
Figure 16 (given
f
rate.
of reserves'
Domestic hquidity
-
with the reserves' injection, the exchange rate appreciates for any
falls
contrasts sharply with the impact
the
f'
rate:
monetary policy
impact
is
equation
not effective in boosting real investment in equilibrium. Instead,
is
(2), the latter
offset the reduction in
f
In conclusion, a central
have much
means
that the
real benefits
exchange
rate depreciates sharply, as
(the standard channel) but also absorb the rise in
bank
that has
may have on
the exchange rate
it,
an inflation target and
will rather tighten
during the
crisis
if the central
and can lead
tighten during a
bank could commit ex-ante
The answer
crisis.
is
no.
is
of the
crisis
during the
but
is
it
boom
not ex-ante.
years,
It
to a
The reason
concemed with
exchange
to a sharp
the impact that
rate depreciation.
monetary policy, would
it
for this is that the primitive
crisis
how much
sectoral
international
allocation
still
choose to
problem
is
the
The amount of
may be exogenous
at the
time
borrowing occurred
on the maturity stmcture and denomination of
contingent credit lines contracted, on the
By
crisis
an external
depends on
f'.
must not only
t'.
private sector's underinsurance with respect to aggregate external shocks.
international Uquidity that the country has during
it
monetary policy. Doing otherwise does not
Ex-ante-Optimal Monetary and reserves policy during
But
and hence
to raise domestic competition for the limited international liquidity,
main
its
that debt,
on the
of investment, and so on.
Underinsurance means that these decisions did not fully internalize the social cost of sacrificing a
unit
of international
hquidity. Figure 17 captures this
Figure 16. The gap A-
i'
captures the undervaluation problem. In other words, had the private
sector expected value of a unit
more
gap by adding a social valuation curve to
of hquidity being
A
as
opposed
to
i^
right.
Figure 17: Underinsurance
Social
,•*
26
,
intemational Uquidity and the vertical constraint during the crisis
10^^)
they would have hoarded
would have
shifted to the
Thus, while an expansionary monetary pobcy during a
anticipation
make
If private agents anticipate such policy, they will expect a higher
is.
ex-ante decision
What
in line with those of the social planner.
one of incentives
is
an external
scarcity during
i^
more
of monetary poUcy
role
unable to boost investment,
crisis is
and hence will
main
In this context, the
one of Uquidity provision, for the main
rather than
crisis is international rather
f'
its
than domestic hquidity.
about intemational reserves' injections? They are
optimal ex- ante, however the
stUl
they bring about have a perverse incentive effect that also needs to be offset
fall
by
in
the
expansionary monetary policy.
Before concluding
a)
There
this section, I shall
mention two caveats:
a related but distinct type of monetary policy ineffectiveness that arises dviring
is
the recovery phase of an external crunch.
WhUe
intemational Uquidity
the binding constraint in this phase, the lack of domestic collateral
during the intemational crunch
lost
injects resources into the
binding constraint
b)
Finally,
the
is
commitment
is.
An
Ee
much
this is the case
does not reach the
as given throughout. If this
PYMES
this
—
case
since the
not warranted, in the sense that
is
policies during the extemal crunch, then
an available incentive mechanism. Having
resort to
it
and
to a post-crisis inflation target is not sufficiently credible
compromised by
PYMES
by the
expansionary monetary policy in
banking system, but
longer be
not the availability of loanable funds.
have taken
I
—
may no
is
seriously
may not be
may have to
monetary poUcy
lost the latter, the authorities
costher incentive mechanisms such as capital controls.
I
do not beheve
of Chile today.
IV.3 Creating a market: Issuing a bencfimarl<-contingent-instrument
As
I
mentioned before,
I
suspect another key aspect behind underinsurance
coordination problem. While there are
many ways
to hoard intemational liquidity
they tend to be cumbersome and costly for most, especially
insurance.
The
basic proposal here
is
when
issuer, this
one
in section IV.l. Ideally,
bond should be
free
the goal
extremely simple: The Central
the IFIs should issue a financial instrument -a "bond" for
identified in step
of other
risks.
now—
Bank
is
desired insurance
to create the market.
With
may be
It is
its
The
issue should
achieved directly by
own
and hedge,
to obtain long
or, preferably,
term
one of
and hence the role played by a reputable
be
significant
bond,
this
the basic contingency well priced,
for the private sector to engineer
simply a
contingent on the shocks
the participation of intemational institutional investors and hence generate
some of the
is
it
its
its
first
enough to
attract
Uquidity.
WhUe
main purpose
is
simply
should be substantiaUy simpler
contingent instruments.
important to realize that the contingency generated via this
mechanism addresses only
the
expected differential impact of the aggregate shocks contained in the index. Individual agents
27
hedging through
will generate heterogeneity in their effective
their net positions rather than
through a change in the specific contingency." Other, more idiosyncratic underinsurance
problems are also welfare reducing but are
the concern in this paper
While
connected with aggregate stabihzation, which
is
of how responsive should be the contingency to the underlying
in principle the question
insurance factors or indices
is
considerations suggest otherwise.
bond contingency very
be able to develop the
irrelevant since the private sector should
derivative markets that can generate
the
less
.^^
any slope they
With the
latter
"steep," so as to
required to generate an appropriate
may
need, financial constraints and Uquidity
considerations in mind,
it is
desirable to
make
minimize the leverage and derivative instruments
amount of insurance
for large crises.
The
counterpart will
probably be a very limited insurance of small and intermediate shocks, which should be fine
since these shocks seldom trigger the financial amplification
How much
many
mechanism highhghted
insurance does the country need? This question
is
in this paper.
not easy to answer as
it
involves
general equilibrium considerations, but a very conservative answer for a full-insurance
upper bound should not be very
trigger the
from the
partial
equihbrium answer. Crises deep enough to
complex scenarios experienced by Chile recently are probably once
and according to our estimates earUer on the
events,
dollars a year, lasting for at
of the
far
damage
created
by the lack of
is
in a
decade
about 5 biUion
most
likely the result
insurance, rather than the direct effect
of the external
most two years (recessions longer than
But the required insurance
shocks).
of resources
is
shortfall
this are
only a fiaction of this shortfall since
all that is
needed
is
be lent, not transferred. If we conservatively assume that the average
(shadow - not the sovereign) spread on Chilean external debt rises by 600 basis points, the
that these resources
required insurance
line).
Gathering
Of course
is
all
effectively
The mapping from
is that
total
" See
ultimate
focus
is
the amounts iavolved are not large.
amount of insurance required
issues.
Of course, the
work on promoting
28
to
(flill
issues required depends
insurance
is
the creation of macro-markets as a
I
on the equilibrium (macro)-benefits of microeconomic
rather than
on the
more expensive
is
direct
more
on the
highly unlikely to be
it
carmot be
outset.
his reasoning applies in the context
Also, insurance against aggregate shocks
opposed
bond
benchmark bond should represent only a small
needs to be Uquid enough from the
While much of
macroeconomic shocks,
msurance
20
it
Shiller's (1993) seminal
risks.
but the point of this "back-of-the-
amount, as hopefully the private sector will follow behind. But
too small either, since
microeconomic
fair,
the above to the size of the contingent
and other design
of the
year (with the loan commitment or credit
these factors puts the fair price of flxU-insurance at around 60 rmUions a year.
slope of the contingency,
fraction
crisis
prices for this type of insurance are never
envelope" calculation
optimal),
300 milHons per
mechanism
for insuring
have discussed as well,
insurance
with
respect
my
to
microeconomic welfare enhancing features of such
likely to support a
market instrument as
(less liquid) individually tailored insurance contracts.
its
solution, as
AH
of the above refers to the insurance between foreigners and domestic agents, but there
also a
need
to create a substantial
amount of domestic
is
insurance. Large corporations with better
access to intemational financial markets should be able to profit firom arbitraging their access to
smaller domestic firms, rather than
of distress, as
it
happened
move inward
in the recent episode.
natural institution to administer this side
regulatory changes to
accommodate
crunches as the index
moves around.
The
to
borrow in the domestic markets during time
Probably the domestic banking system
of the contingent
this
new
strategy,
which would probably require
on whether the development of such contingency has any obvious advantage
question arises
bonds denominated in Chilean pesos. This
World Bank issuance
support of a recent
indexing the contingency will in
all
is
for
disadvantage in them:
The Chilean
105 milhon.^' In
likelihood put pressure
on
exchange
but
rate,
it
may
bond
view, while the shocks
the exchange rate,
and hence the
have proposed, there
I
is
a clear
of the peso, and therefore
much
holders) as
as an exogenous
somewhat paradoxically, a peso-denominated bond may not only
be subjected to a higher premium due
the authorities decide to
my
authorities largely control the value
unlikely to appeal the insurers (or
is
contingency. Moreover, and
of placing abroad
is,
the strategy currently being pursued, with the
US$
peso-bonds serve a role similar to the contingent instmments
contingency
the
role without causing undesirable domestic credit
over the admittedly simpler strategy of "internationalizing the peso;" that
this
is
to the risk
of Central Bank's manipulation of the
also not prove a very effective insurance
dampen an exchange
mechanism
rate depreciation that is risking
if,
an
for example,
inflation target
(see previous section).^^
The disadvantage of
the contingent bond, on the other hand,
implement and hence the
carefiil
of not finding the market for
risk
planning not to design an instrument that
contingencies are not fliUy transparent.
it
is
it
is
is that
it
is
more
non-negligible.
difficult to
It
difficult to
will require
comprehend, or whose
The experiences of previous placements of commodity-
indexed bonds should be studied carefiiUy, and a few references to these experiences are
summarized
Box
in
2 below and described in more detail in the appendix.
There are also plenty of lessons to leam fi^om the recent securitization of catastrophe-insurance
contracts in the
US.
Box
recently attempting to
the
first
3 describes the instruments employed by the insurance industry in
hedge catastrophe
risk through capital markets.
placements of these instruments were
plentiful,
The
understanding of their contingencies to the legal obstacles to overcome due
great confusion
^'
The bond was
on whether
floated in
May
3
to classify these instruments as
1
,
obstacles faced
by
ranging fi-om investor's lack of
to, for
example, the
bonds or insurance, and hence on
2000, for an amount of $55,000 million of Chilean pesos. Tlie bonds have
4"", 2005), a coupon of 6.6%, and are indexed to Chilean inflation.
was acquired by Chilean institutional investors (75%), while the rest was placed among
intemational investors. Chase Manhattan Intemational Ltd, New York, acted as director bank.
^'
The so called "fear of floating" (Calvo and Reinhart 2000). In principle, one could think of an optimal
amount of external peso debt so that the moral hazard problem is exactly offset by the fear of floating effect.
In practice, such mechanism may introduce significant policy uncertainty to investors.
5 years of maturity (to
Most of the
29
issue
be repaid on June
deciding which institution should regulate them. Another lesson from the catastrophe risk
experience
should
that these
is
come down
bonds
will probably
pay a
premium
significant
"war-of-attrition" in the private placement of these bonds,
it
process needs to be started by the Chilean authorities or an
IFI.^''
V. Final
early on, but this
reinforces the conclusion that the
Remarks
no reason
There
is
many
contracts can naturally arise indexed to
to stop at financial markets.
removed much of the
index) once
it.
Once
As
the contingent index has been created,
the U.F. (unidad de fomento
uncertainty created
microeconomic agents, the new contingency
may be
-
e.g.,
minimum wages and temporary
institutionalized to
index to ease firms' financial
sinularly
Finally,
be indexed
it
is
difficulties
contracts
during extemal
-
do something similar
that includes other
contingency
^^
For example, labor
crises.
The
stmctural fiscal surplus could
to free financial resources to the private sector during those times.
important to consider regional interactions.
and Mexico.
inflation
inflation to
could be indexed to the contingent
It
may
well be worth giving
the sf)ecific tailoring of the instrument to the Chilean needs, for a
bond
- an
by high and unstable
with the uncertainty that financially amplified extemal shocks generate.
markets
premium
rapidly as well.^^ Since this declining path creates a natural incentive for a
Ucpoid contingent
advanced emerging economies of the region, such as Argentina, Brazil
Or perhaps even
needs
much more
up some of
not
search for other co-issuers beyond Latin America, with
too
distant
from
those
of
Chile.
cat-bonds were placed in 1 997 and initially paid a premium over nine times the expected loss. By
premium had steadily declined by nearly forty percent.
There is also some evidence that CAT-Options paid a lower initial excess-premium than CAT-Bonds. See
Cummins, Lalonde and Phillips (2000).
The
first
1999, that
^''
30
Box
Examples
of
Commodity-Indexed Bonds
2:
with
price
months
maturities
10
to
Industries
indexed
copper-
issued
bonds. See the boxes below for more
Bonds with warrants on the issuer's shares
have become fairly common financial
instruments, while bonds embedded with
options on commodity prices are relatively
Sunshine Mining issued bonds
These commodity-indexed bonds have
would
to
like
participate
commodity
in
purchase
While these
directly.
indexed
valued)
as
securities.
sum of
the
Some
the greater of
$1000 and
silver.
If the
amount is greater than
company had the right to deliver
principal
all
silver to
down
cases they can be broken
in
at
between 50 and 58 ounces of
$1000, the
can be quite complex,
instruments
1980 due
At maturity or redemption, each bond
was payable
options but cannot for regulatory or other
reasons
in
1995 and 2004 incorporating an option on
silver.
generally been attractive to investors that
detail.
SUver.
in
rare.
3
Copper
United
years.
also
from
staggered
bondholders in satisfaction of the
(and thus
The coupon
indexed principal amount.
rates
standard
fairly
on these bonds varied
recent examples include:
percent.
fi"om 8 percent to 9.75
These bonds also had sinking fund
covenants requiring the
company
to call a
Oil.
Standard
In 1986
issue
OU
fraction
of Ohio created an
payment was a
Natural Gas.
the value of 170 barrels
Interest
Indexed
Debentures. This note was intended
coupon
a base
$25 and price ceiling
Gas
of Natural
issue
of Light Sweet Oklahoma Crude Oil with a
price floor at
Forest Oil proposed a 12-year
in addition to a guaranteed
amount -
principal
oil
In July 1988
~
note received
of
fiinction
The holder of a 1990
prices at maturity.
by the company.
where the
of Oil Indexed Notes
principal
of the original indenture and an
option of recall
rate
basis
points
for
pay
every 6 months plus 4
$40.
at
to.
every
$0.01
by
which
Investors essentially held standard four-year
bond while being long
average gas spot price exceeds $1.76 per
in
a
call
option with
million
exercise price of $25 and short in a call
option with
Mexican
exercise
state
oil
issued bonds in
price
at
The
$40.
Oil
PEMEX
company
and Currency.
also
1973 including a forward
contract on oil
Smith
Magma Copper
$200
issued
million of 10-year junk bonds with quarterly
coupon
rate
was indexed
While
copper price.
fixed at
1
8%
for the
the
first
to
average
the
coupon
rate
six months,
it
was
was
indexed thereafter to average copper prices
in the
preceding quarter. This rate had both
a ceiling at $2 of
12%.
Magma
21%
and a floor
president
J.
at
31
the
feasibility
A
of
20
call
five-year
options having an exercise price of
$17 with staggered maturity and a five-year
forward contract on yen with an exercise
price of 140 Yen/$.
$.80 of
Burgess Winter
mind expensive money if the
copper price stays high." Embedded in this
10 year bond are 40 options on the copper
said "I don't
discusses
bond would pay a
semiannual coupon of two barrels of Sweet
Light Crude with a price floor of $34 and
principal of 140,000 yen at maturity. This is
simply a standard five-year coupon bond plus
bond.
1988,
(1995)
issuing an oil interest-indexed dual currency
Copper.
hi
BTUs.
Source: Smith and Smithson (1990), Smith
(1995) and
SEC lOK
filings
of Magma
Copper and Sunshine Mining
32
Box 3: The Catastrophe
(CAT) Securities Market
these bonds contingent on losses exceeding
Experience
first
over
bilHon
$1
were
Investors
in
months.
80 percent of the
liable for
$500 million
12-18
next
the
USAA
with
losses,
taking a 20 percent stake to reduce any
The pooling of catastrophe
of development
stages
early
marked increase
in
over
two
the
risk is
last
in the
still
despite
the
catastrophe
insurance
decades.
While
remaining agency problems.
These bonds
have yielded between 275-575 basis points
LIBOR
over
depending on the principal
^protection purchased.
prospective event losses could easily exceed
$50
and surplus of the
billion, the total capital
U.S. insurers
is
only about $240
The
billion.
insurance industry has historically tried to
reduce
exposure
reinsurance
such
to
capitalized firms, but
separately
reinsurance capital
has been extremely limited and
insufficient to
let
through
risk
with
contracts
fmance exposures
in the
alone the rest of the world.
prompted insurers
US,
This has
Capital
risk.
markets plausibly represent a solution as the
$50-$ 100
billion
could be
was
and
in
financed
bear
equivalent
risk
exceeding $25
industry
to
losses
The success of
billion.
this
issue set the stage for later issues covering
earthquake exposure in California and Japan.
Approximately
$1
bonds was issued
of
billion
in
catastrophe
each of 1997 and 1998.
industry exposures that
are
about equal to a
Nationwide Mutual
million
Investors purchase
the
company has
of catastrophe.
provide
new
a
of
source
in
surplus
notes.
Treasury bonds but
the right to liquidate these
exchange company notes
bonds
As
these risks are "zero-beta" events, they
US
$400
issued
recently
contingent
in
normal day's fluctuation in equity markets.
potentially
and
oversubscribed,
actually
demonstrated a willingness by investors to
to look to other investors
fmance catastrophe
in order to
largely
is
This issue has been a remarkable success
in event
This instrument infuses the
company with cash when times
are bad and
pays investors a premium over a normal
diversification for investors.
Treasury.
CAT Bonds and Notes.
Overall,
Guy
Carpenter, J.P.
have
institutions
God" bonds
Morgan
recently
and other
issued
"act-of-
as a source of financing for
been
have
there
catastrophe
bond
compared
totaling about $3 billion,
$120
in
billion
volume
annual
20
about
with
issues
international reinsurance market.
proceeds
to about
on
the
There
is
insurance companies. These instruments are
evidence that they are gaining acceptance by
subject to reductions in principal in the event
the investment community.
As
of a catastrophe loss to the insurer.
these
types of risks
are
acts-of-God not
subject to manipulation, they
from
severe
adverse
hazard problems.
substantial
do not
selection
or
in
1997
initially
year
moral
by
later that
premium had
33
USAA
issued
premium
but two
steadily declined
nearly forty percent.
10%,
One
important issue with
$400 million of
CAT
bonds
is
that
investors are not only taking on catastrophe
risk but also credit risk
1997-98
loss,
cat-
first
at
above Treasuries of similar maturity.
hi
paid a
over nine times the expected
suffer
These bonds traded
spreads, however, almost
bonds placed
The
when buying
these
instruments, implying that adverse selection
and moral hazard problems could
limit their
(in
contrast
CAT
to
bonds),
creates
this
company making
usefiibess to companies seeking to hedge by
basis risk for the
significantly raising their cost.
options less useful as a hedging instrument.
these
Despite this recent success, there remains a
CAT Options.
great
deal
of
The Chicago Board of Trade
(CBOT)
of an
but withdrew these contracts in 1995 due to
disaster
securities.
a lack of interest by investors. Currently the
generally
not
demand
compensation
exchange trades
call
option spreads on nine
the
risk, the lack
of
standardizafion in measuring losses, and the
absence
introduced catastrophic loss fiitures in 1992,
about
uncertainty
assessment of catastrophe
institutional
structure
for
These
markets
are
very
liquid
and
for
investors
this
risk.
catastrophe indices constructed by Property
Competition firom the reinsurance market as
Claims Services (one national, five regional
well as the tax and accounting advantage of
and three
state).
However, these options
trade less fi^equently than other opfions at the
CBOT
and
less than
reinsurance
are
hurdles
development of CAT
for
fiirther
securities markets.
$100 million has been
placed in any quarter in these instruments.
Source: Froot (1999); Cummings Lalonde,
While agency issues mentioned above are
and
ameliorated through the use industry index
34
Phillips (2000)
Appendix
Hedging
A.I:
at
Magma
Magma Copper
had
copper
a
instituted
price
protection program to ensure, regardless of
the copper price, adequate cash flow for the
Magma Copper Company
was a
fully-
integrated producer of electrolyctic copper
and was one of the
largest
US
producers in
completion of several major capital projects
that
important to
are
projects, Tintaya,
These
fiiture.
its
Kalamazoo and Robinson,
The
company owned and operated copper mines
at San Manuel, Superior, and Pinto Valley
and owned a mine in Pern through its
production
Tintaya subsidiary.
copper price put options or forward sales of
1980s and early
the later
1990s."
were
intended
included hedging using either the purchase of
copper
The copper
prices established
on the two
The Commodity
Exchange, Inc. (Comex) in New York and
the London Metal Exchange (LME) major metals exchanges
a fixed price depending on copper
at
market
conditions
copper supply and demand. The
Magma's
operations
upon
dependent
largely
market price for copper.
change
in the
A
profitability
was
the
obviously
worldwide
$0.01 per
pound
average price realized for the
company's 1994 output would have affected
pre-tax income
As
illustrated
by an estimated $6.0
in
the
table
million.
below, copper
prices have historically been subject to
factors, including international
conditions,
political
demand, the
substitutes,
levels
availability
volatihty.
Copper Price
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
Low
Average
$0.66
$0.56
$0.61
$0.69
$0.57
$0.62
$1.46
$0.60
$0.78
$1.63
$0.88
$1.15
$1.59
$1.00
$1.25
$1.38
$0.96
$1.19
$1.20
$0.96
$1.05
$1.16
$0.94
$1.03
$1.07
$0.72
$0.85
$1.40
$0.78
$1.07
Due
to these hedging strategies.
average copper price realized was nine cents
and cost of copper
and international exchange
^^
rates.
than the average
was nine
Mineralsacquired the company
in 1995.
Comex
cents per
price for 1993 and
pound lower (98 cents
versus $1.07) than the market price in 1994.
With the recent purchase of Tintaya, and
construction of the Kalamazoo and Robinson
mines, the copper price protection program
was extended
into early 1997.
The
of pertinent copper industry data contains
low and average copper prices per pound
and illustrates the historic voladlity of copper prices.
table
high,
Price data are given for the
Comex
standard-grade
In contrast to
1994, which included a significant amount of
forward sales entered into
Comex
Magma's
per pound greater (94 cents versus 85 cents)
degree, inventory carrying costs (primarily
BHP
(1 lb.)
High
and
of supply
copper producers and others and, to a lesser
^^
and
economic and
inventory levels maintained by
interest charges)
price
wide
and are affected by numerous
fluctuations
including
-
broadly reflect the worldwide balance of
of
The program
pound.
per
cost
overall
Magma's
increase
to
and ore reserves and reduce
1993,
the
program
for
at
1995
the end of
and
1996
consists almost entirely of copper price put
options which lock in a copper price floor
copper contract for the years 1985-1989. Thereafter,
and operating cash flow while retaining the
prices are given for contracts for delivery of high-grade
upside potential of higher copper prices.
copper cathode, which replaced the standard-grade
contract effecdve January 1, 1990.
35
As
a result of this program, which created a
copper price floor of 87 cents and 93 cents
for 1995
is
strategic
its
Magma
and 1996 respectively,
confident that
would be able
growth
with
opportunities
moderate outside financing and maintain
its
strong financial position and flexibility.
company had experimented
with commodity-indexed bonds as a means
of hedging risk created by volatility in copper
In the
1
prices.
980s, the
Magma
Copper
year
issued $210,000,000 of 10
1988 - junk bonds
in
underwritten by Drexel
Bumham
The Copper Notes paid an
coupon
increments
in
described
in
the
Commencing on November 15, 1992 and on
each November 1 5 thereafter, the holders of
the
Copper Notes had the option of
requiring the company to repurchase, for
cash, or in certain circumstances, for new
debt securities. Copper Notes in principal
amount equal to 15% of the principal sum at
100%) of the principal amount plus accraed
interest.
Senior
Interest-Indexed
Subordinated Notes
only
indenture.
felt
complete
to
but
Lambert.
18
If the average copper price for the preceding
eleven-month period
percent
pound
Magma
the
is
$ .80 or below per
had the
right to offer to
months
holders in lieu of cash an equal principal
and then had the coupon indexed to the
amount of Senior Subordinated Notes due
average copper price according to the table
1998 with a fixed
below.
the
quarterly
rate for the first six
opinion
such that
interest rate
of two nationally
in
recognized
investment banks chosen by the company
When the
an
interest
average copper price applicable to
payment was $
less,
and the company's
ratio
was
less than
1.5
.80 per
pound or
interest
coverage
to
recent fiscal quarter ending.
option
to
pay
50%
delivery of additional
for the
1
most
Magma had
of such
interest
Copper Notes
the
by
("Pay-
in-Kind Notes") in principal amount equal to
the
amount of such
new notes would have a bid
101% of their principal amount.
The
limitations, limitations
and
dividends
on the payment of cash
distribution
and
on,
repurchases of the company's capital stock.
interest.
Moody's
Quarterly
Average Copper Price
Coupon Rate
(1 lb.)
21%
20%
19%
18%
17%
$2.00 or above
$1.80
$1.60
initially rated
"Bl"
December
these bonds as
1991 as the company lowered
reduced
its
As copper
its
costs
and
debt levels.
prices rebounded fi-om their low
levels
of the 1980s, the
bonds
steadily climbed.
$1.40
interest rate
on these
$1.20
confident that prices
$1.10
through the
As Magma became
would remain steady
1990s, the company eventually
$1.00
decided to
retire
$0.90
interest expense.
$1.30
the
bonds
in
to
lower
$0.80 or below
In
The Copper Notes were redeemable at the
option of the Company at any time or firom
time to time on or after November 15, 1991,
36
Copper
indenture pursuant to which the
Notes were issued contained, among other
but raised the rating to "Ba3" in
16%
15%
14%
13%
12%
value of
the
December 1991,
of the
was
Magma's Copper
company gave
the
there
holders
a defeasance
Notes.
The
notice of the redemption to
of
the
Copper
Notes
on
The company sold
$200,000,000 of 12% Senior Subordinated
December
Notes
1991.
31,
(12%
irrevocably
Notes)
deposited
due
the
additional funds into a trust.
under
Copper
the
agreement,
108% of
the
notes
the principal
2001,
proceeds
As provided
Copper.
37
SEC
and
for
indenture
Notes
were redeemed at
amount thereof plus
accrued interest on January 31, 1992.
Source:
and
10k Filings of Magma
Appendix
important plank in the firm's strategy.
To be
Board considered
successfiil, the
that
sound financial practice and protection from
(UCI)
a downturn in copper prices were essential.
As
Copper
United
formed
Upland project was an
In this regard, the
Hedging
Commodity Price Risk at
United Copper Industries
A.II:
Industries
was
Inc.
merger from two smaller copper
in a
Mesa Copper Industries and
Canyon Copper Corporation in 1976.
company considered
a rider, the
it
was
not in the business of speculating on copper
prices and hence, taking a
view on
short-
term demand.
companies:
The primary business of
UCI was
copper
merger a number of other
mining and the production of refined copper
copper mines that were in poor shape were
cathodes or wirebars. The company also had
Following
this
USA
acquired in the
and Canada, mostly
an
exploration
division
had
that
bought
financed via small equity issues. However,
licenses to seek out reserves in promising
following a proposal from the group's lead
geological areas.
As mentioned
company was not
against using
company
investment bank, the
into a
number of
early
1980s for
limited partnerships in the
more mature mines.
its
Limited partnerships were
way
of
also entered
a
income
passing
tax-efficient
to
investors.
Exploration and development had not been
company formed
ignored either, and the
venture
joint
with
had discovered
Vancouver-based
a
company
exploration
to develop the find
Yukon. To finance
in the
share of the partnership,
a
UCI
it
its
issued a 4.5
per cent copper-index note, together with a
common
placing of
shares and a warrant
issue.
now
the
in
North
top five copper producers in
America.
actively
potential.
and
shares
Its
on
traded
Exchange
UCI was
the
American
sought
it
The company had
clear strategic goals to
to
now
were
Stock
realize
set itself a set
the initial
merger and
acquired thereafter.
close
near
the
in
rose significantiy
three, smaller
would
not
it
UCI's output was
aimed
to
would
become
in
the third largest producer
North America. This probably
entail
a
number
acquisitions in mining, in the
above
all
in
in that
of
significant quantities
to
US, Canada, but
UCI
market would
considered a presence
facilitate sales
and the
negotiation of long term contracts with users.
38
joint venture
start
of copper
The company
producing
in the
also
coming
had yet to
decide whether to proceed with
project,
its Upland
which would more than replace the
Upland had great potential, but
would require a significant investment and it
would be three to four years before the first
if
It
copper prices
remained reasonably buoyant and above the
US$1400/ton
level.
significant
Mexico, given the high potential
of that market.
likely to fall in
The Canadian
expected
financial year.
copper prices
pay the company
poor quality ore and
would also only be viable
It
if
to
nearly
was probably imminent.
a decision to close
result,
mines
revenues were generated from the project.
next five years.
of copper
having
term,
to continue to exploit the
the near term.
company
came from
One of these was due
exhausted the deposit. Only
As a
expertise
the
operated five mines, the two that
closure.
of
At the moment,
discoveries.
its
be realized over the
its
with other companies to exploit
in joining
was
Following these developments,
earlier, the
The
UCI
volatility
of the copper price meant that
had over the years favored hedging
some of
its
output.
Since introducing the
decision to hedge, the firm had expanded the
ways and means
it
had used to eliminate the
commodity price risk from its production.
The company produced just over 400000
tonnes per annum. The firm had initially just
hedged the immediate production when the
outlook seemed to justify
gradually extended
it.
However,
it
had
approach to having
its
rolling, five-year output targets
and a long-
term view on the copper price.
entering the agreement,
to
was
it
in a position
modify the standard terms and conditions
by mutual agreement. Nevertheless, because
of the problem of counterparty
10 percent of
hedged
annual
its
risk, less
than
was
production
way. However, because of the
this
atfractions, the
company was always looking
hedging via
at opportunities to increase its
fixed-price supply agreements.
Equity
who
analysts
followed
stock
the
considered the company's hedging program
to
be one of the prime reasons for investing
in
shares.
its
company was
The
compared
favorably
companies
for the
with
relative
often
mining
other
of
stability
The company
2 per cent of annual capacity in 1997.
UCI was
due
share
futures
to
in
capital
the
previous year (1996) and the downturn in the
industry
started
that
company had been
year
in
forced
(1995),
to
the
reduce
its
dividend from the long-run 60 cents a share
to
48 cents
much
-
a reduction that had led to
adverse analytical
comment and a dive
management was
an active hedger with copper
on
fraded
(LME)
Exchange
regularly
sold
against
fritures
was concenfrated
liquidity
forward
its
current tools used
by
hedge
the firm to
price risks mostly involved forward and
futures confracts.
The firm had hedged a
margin,
commodity swaps
any transactions of this
-
for a part of
type.
as mentioned earlier,
copper-indexed note
at
its
as yet -undertaken
On
it
the liability
had issued a
a time
when
these
were fashionable.
company
particularly
liked
forward
its
price risks
means
the basis risk problems this
as
was
be
to
to post considerable
involving
amounts
internal
resources and tying up borrowing lines.
manage
To
this activity, the cenfral freasury unit
employed two
staff
in
cash
dealers, plus three
back-up
Positions were monitored daily and
adjustments, based on market view and the
production
factored
and
demand outlook
any
into
adjustments.
Currently, about 45 per cent of
was hedged
this
UCI's output
way.
were being hedged.
The forward market was atfractive in two
respects. First, the company could lock in
buyers to their output up to two years hence
and could plan where the delivery was going
to take place. Second, although the company
knew it was taking counterparty risk in
39
all
both
evolving
confracts and used these as the basic
by which
a
to resort to
structural positions. This required
company
were
The
with
The firm was,
had large
the
side,
To hedge
feasible.
company had
expected, an active user of the market and
and had also received a number of proposals
production, but had not
-
entailed.
very small quantity of its output using options
to use
nearby
the
in
stack hedges and rolling hedging positions
friture.
The
production.
months, selling confracts beyond six to nine
keen
in
the
Because of the market structure where
given exposure, the
of the problem
Metal
New
York
Commodity Exchange (Comex). The firm
months was usually not
to avoid a repetition
London
the
and
in the share price. Senior
the
maximum
its
In 1997, however,
increase
number of long-
maturity of seven years, which amounted to
earnings and cash flow.
the
also had a
term confracts with users out to a
Whereas
futures
were used
bulk of the firm's price
and
futures
fianction
aspects
for
fulfilled
the
UCI was
risks,
the
to
manage
the
both forwards
same economic
company.
One of
the
keenest to examine in any
changes that might be contemplated was
ways
improve the firm's approach
to
hedging.
The Upland
opportunity
hedging
the
Furthermore, a number of banks
strategy.
were
project offered just the
re-examine
to
to
keen
promote
to
commodity swap
copper-Hnked
a
as a solution. Others, such
were
as Phibro-Salomon,
interested in selling
company long-dated over-the-counter
the
puts on the copper price.
was
situated
in
issue
was whether
cash flow from the project.
had been quite
the
and
volatile
Copper
that relying
price
on the
market might lead the company
spot
to
experience losses and suffer considerable
from the project
variations in cash flow
hence
project
The second
debacle).
year's
price volatility to an acceptable variability in
-
profits
and
firm,
The Upland mine
foUowang the previous
(particularly
making
to investors.
-
and
less attractive to the
it
An
examination of the
Utah, in the Rockies in a find that offered
showed that the impact of the
copper price on the annual after-tax cash
the prospect of recovering about 10 million
flow during the
forecasts
maximum
extraction phase.
tons of copper over a 20-year period by
The
means of open-cast mining.
To
UCI
bring the project on line,
US$100
about
investing
would be
million
in
site
clearing and preparation, mining equipment
and copper
analysis,
As
refining.
UCI
part of the project
had prepared a
series
of long-
term forecasts on the copper price. These
are
based
three
demand
increased
company
scenarios:
forecast
considered
(assigning
a
to
it
a
million;
bullish
which
most
the
0.6
a
neutral
the
likely
probability
in
NPV
where the project had a
analysis)
$39.3
on
forecast
on
risk
a
since
especially
the
the
in
fall
copper price would also affect other parts of
the firm at the
were
that
same
time. Other alternatives
available
operational
for
or
financial
hedging of the Upland project had
already
been
earlier,
by
UCI
As
proposed.
some of these had
but the
mentioned
already been used
company was
exploring
be more appropriate
in the context
The
expansion.
firm's
of the
alternatives
under
consideration were:
of $21.8
million; the bearish forecast (0. 1 5 probability)
NPV
downside
the
investment,
definite attractions
of
where
with
with a
reducing
were
methods which might
0.25
NPV
in
Board would take
the
felt that
that there
other, as yet untried,
largely constant (given a
a
view
the
demand remained
probability)
president
the
of $10.2 million, envisaging a
Enter into a long-term supply contract at a
fixed price with a consumer;
long-term decline in copper demand.
Sell
forward the copper for an agreed
One investment bank's proprietary
For the three scenarios given above, the
period.
Upland project would be
product that
measured by
However,
-
as
net present value (NPV).
its
at prices just
the project
profitable
would be
below US$1400/ton,
unprofitable.
is
available to
a flat-rate forward
UCI
is
knovra as
where the contango
(or
the difference between the spot and forward
price)
was
fixed
regardless
of maturity.
Another product, known as a spot deferred,
The
questions facing
first,
whether the
and
copper
the
UCI's Board were,
project's
sensitivity
break-even
price
to
were
acceptable in terms of future copper price
behavior.
maintain
The Board was
the
investors as a
profit
40
also concerned to
company's
standing
with
mining stock with a
stable
record and a good dividend record
is
a forward contract with a floating copper
price and
more
no
fixed delivery date.
flexibility than a
It
provides
conventional forward
contract but without the upfront cost of using
an option;
Enter into a commodity swap where
UCI
would receive a
given
fixed price
for
a
quantity of copper against paying a variable
The
price.
would be
effect
to synthetically
As
part
had
new
of
copper
would
require
setting
up of
This
the
futures.
a stack hedge since fiitures prices do not
extend beyond about three years. However,
lack of liquidity in the longer contracts means
that only the shorter-dated contracts would
prove
using
by
position
the
how
strategy, the
new hedges
the
be integrated with existing
create a fixed selling price;
Hedge
of the hedging
to consider
instruments
positions, the use
advantages of options), and extending the
hedging period
some of
for
UCI
The
case has looked
strategic business decision
way of saving money
as a
Asian-style
(that
-
these could be
average rate
is,
options
concems
that
All the above measures could be used to
be used
exposure fully or partially and could
tandem. So
in
UCI
could both use a
commodity swap and buy copper puts
to hedge.
knew
that
The
first
a
for
As
or not.
with
issues are
more complex
case where the benefit/cost
receivable
single
more
is
straightforward.
UCI
For
any
-
trade-off firom the hedge-not hedge decision
price
company
and some of the
if this
provided the best alternatives.
The
issues
most business problems there are no hardand-fast rules for deciding when, what or
than in the
its
the
at
managers might have when
deciding whether to hedge
how
over the exercise period).
hedge
group's
the
output with a commodity hedge.
surrounding the hedging or insuring of a
practical;
buy a series of copper puts. These would
need to cover output over a given period and
-
the
particular
in
(
Board
should
hedging
hedge means guaranteeing a
to
on the new
profitability
-
project's output
but
and hence
-
means most
also
likely
decision potentially meant giving up on the
giving up
upside for the copper price and he was
gain fi'om future price increases. There are
concerned
also questions as to the risk appetite of the
that,
given the firm's long-run
assessment
bullish
despite
for
temporary
its
copper,
weakness
prove to be the wrong decision.
to
hedging
might
-
A
decision
UCI
hedge might then be somediing
would
-
regret.
firm
some
and
its
or
all
of the opportunity to
what they
shareholders and
would want UCI's managers to
instruments available also seem
do.
The
less
well
adapted to UCI's needs, or are costly. The
firm really requires a very long-term hedging
insti-ument. This is not readily available
The market
uncertainty surrounding copper,
is
on
the
commodities exchanges, did not make
this
with
the
price
see-sawing
daily
to
decision any easier. This unsettled behavior
made any
likely to involve the
premium
Board
also
more
problematical.
The
wanted
to be cautious
and not
management
approaches,
fiitures
add
to the
problem of the dividend
already caused so
much
in the price
might leave the company exposed
that
they
to
now
such
as
paying a
significant
costs
time
in
and,
the
in
the
Source: Moles (1988)
norm
for
in
impose
and management
analysis,
deliver the promised outcome.
had inappropriately and
UCI
hedges
replication,
skill
final
if
Do-it-yourself
stacking
dynamic
or
unwisely hedged.
41
premium
products.
had
external and internal
comment. However, locking
charge
that
significant
seeks to trade beyond the market's
'no action' recommendation to the
even
in
for protection. Providers are likely
impose a
risk
company
company
and
may
not
Appendix
10 million shares of
Hedging
A.III:
common
The
stock.
were
Debentures
repurchased
utilized
to
Silver Price Risk at Sunshine
satisfy all fliture sinking flind obligations
on
Mining Company
the Debentures.
and natural gas
Starting in 1985, silver, oil
low
suffered a prolonged period of
prices,
and, in February 1991, silver prices reached
a
At December 31, 1991,
17-year low.
were
silver prices
settlement
still
COMEX
reported by
depressed, with a spot
ounce
per
price
of
silver
of $3.88. As a
as
result,
Mining Company reported
the Sunshine
consolidated net losses for each of 1985-
For
1991.
company
year
the
ended
1991,
the
net
loss
of
a
reported
Other than the Debentures the only other
indebtedness
consolidated
outstanding
consisted of the silver-indexed bonds issued
from 1980
to
obligations
of
These bonds were
1986.
Sunshine
subsidiary,
its
Precious Metals, and were not assumed or
guaranteed by
its
Due
parent company.
to
the continued depressed price of silver, the
of
operations
payments of
and
subsidiary
the
interest
the
and principal on the
Bonds were funded from
Silver Indexed
approximately $40. 1 million. The company's
1988, primarily by the advancement of funds
operating losses and cash flow deficiencies
from
were
likely
continue until silver prices
to
Accordingly, the
recover.
company began
financial restructuring plan to reduce
burden and ensure
its
its
a
debt
As of
funds to
future viability.
domestic
other
sold
oil
restructuring
the
to
company
substantially
The
received from this
the
company
parent
its
subsidiary for the payment of
interest
or sinking fund
obligations
the
with respect to the Silver Indexed Bonds.
of
the
The non-payment by Sunshine Precious
net
sale, after
1991,
plan,
all
and natural gas properties of
subsidiaries.
April
determined not to advance any additional
future
Pursuant
parent company.
its
its
proceeds
discharge of
all
Metals of interest or sinking fund obligations
wdth respect to any series of represented an
result,
payment
were
series
of Silver Indexed Bonds.
the
related
approximately
Mining
remaining
costs
of
$60.0
million.
Company
Sunshine
discontinued
sell
due
Approximately
principal face
a tender offer for
2008
(the
$51.8
occurred
on
a
all
The
aggregate amount of accrued and unpaid
interest
along
with
obligations as of
Company
accrued
aggregate
in consideration
principal
interest,
amount
^^
unpaid
December
sinking
fimd
31, 1991,
was
of
common
its
per
of
$1,000
each
new
commodity
of bonds
1995 and 2004 that incorporated an option on
silver prices.
At maturity or redempdon, each
Exploring
a
strategy in diversifying
price risk, subsidiary issued in
due
Reset
Debentures).
million
$600 cash and 200 shares of
aggregate
its
amount of the Debentures was
tendered to the
plus
Mining
Sunshine
Subordinated
Convertible
Debentures
defaults
those assets.
On April 9, 1991,
Company completed
stock,
its
and natural gas operations and
oil
attempted to
disposal,
As
event of default under such series.
of outstanding indebtedness and payment of
in
future
1980 four
series
at the maximum of $1,000 or the
market price of between 50 and 58 ounces of silver.
These bonds had stated interest rates ranging from 8%
to 9.75%, sinking ftmd requirements, an option of recall
by the company, and were collateralized by an interest
in the annual mining production of the Sunshine Mine.
bond was payable
The
bond
including
contained
indentures
covenants applicable to
minimum
certain
restrictive
Sunshine Precious Metals
net
worth
requirements
and
on the payment of dividends or other
distributions on its capital stock, purchases of treasur)'
limitations
Debenture for a
total
aggregate purchase
price of approximately $32.2 million cash and
stock
and
securities.
42
the
issuance
of additional
silver-backed
approximately
$15.6
million
for
all
such
Bond would be
cash
series.
The parent and
subsidiary believed that a
Indexed Bonds
restructuring of the Silver
Sunshine Precious Metals
represented
only viable plan for continuing operations
unless there
of
a significant rise in the price
is
silver in the
December
near term.
parent
the
1991,
Accordingly, in
company
proposed an exchange offer whereby the
Bonds would
amount of a new issue
option of Sunshine Precious
or, at the
Metal and
payable semi-annually in
subject to availability if certain
conditions are met, in shares of the parent
common
company's
stock.
Tlie
8%
Indexed Bonds were redeemable
or
in
part
option
the
at
Precious Metals
at
of Sunshine
any time
in cash or
shares
of the parent company's
stock
at
indexed
the
Silver
whole
in
common
principal
amount
together with accrued and unpaid interest.
holders of the Silver Indexed
receive
of
$800
principal
indexed
silver
Precious Metals
annual rate of
8%
bonds
of
Sunshine
8%
the
1993,
outstanding
8%
principal
Silver Indexed
reduced from $57.2 million to $7.6 million
through various redemption transactions for
Sunshine Mining's
value of $200 for each outstanding $1,000
result
amount of Silver Indexed Bonds
In addition, each tendering holder
tendered.
would receive
Mining's
additional shares of
common
Sunshine
stock in satisfaction of a
portion of accrued and unpaid interest. Each
$ 1 ,000 face
amount of
Bond would
redemption
be
8%
payable
at the greater
43
hiterest
on the
of these
common
As a
stock.
the
parent
company recorded a charge of $12.5
million,
representing
the
transactions,
loss
on
an
induced
conversion, and an extraordinary charge of
$13.6 million relating to the redemption of the
remaining outstanding bonds.
Silver Indexed
at
maturity
or
of $1,000 or the
specified average market price of 85 ounces
of silver,
8%
of
Bonds was
Silver hidexed
at
Bonds), and shares of which have a market
face
amount
an
bearing interest
(the
During
Silver Indexed
Source:
Mining.
SEC 1 OK Reports
of Sunshine
.
References
1)
Caballero, Ricardo
"Structural Volatility in Chile:
J.,
A
Policy Report,"
lADB mimeo,
October 1999.
2)
Ricardo
Caballero,
Macroeconomic
J.,
Volatility
"Reformed" Latin America:
in
Diagnosis and Policy Proposals, lADB, Washington D.C., 2001
3)
Caballero,
R.J.
and M.Hammour, "The Cost of Recessions Revisited:
MIT mimeo,
Liquidationist View,"
4)
Caballero, Ricardo
Constraints in a
J.
A
Reverse
1998.
and Arvind Krishnamurthy,
Model of Emerging Market
"International
Crises,"
and Domestic
Collateral
Journal of Monetary Economics,
December 2001a.
5)
Caballero, Ricardo
and Arvind Krishnamurthy, "Excessive Dollar Debt: Underinsurance
J.
MIT mimeo,
and Domestic Financial Underdevelopment,"
6)
Caballero, Ricardo
Central
Bank
2001b.
and Arvind Krishnamurthy, "A Vertical Analysis of Crises and
J.
Intervention,"
mimeo MIT, October 2001c.
7)
Calvo, G.A. and
CM. Reinhart, "Fear of Floating," mimeo Maryland, 2000.
8)
Cummins,
D. Lalonde, R.D.
J.D.,
Securities,"
9)
Phillips,
"The Basis Risk of Catastrophic-Loss Index
Wharton 00-22-B-B, May 2000.
Froot, K.A., The
Financing of Catastrophe Risk, Chicago: The University of Chicago
Press, 1999.
10) Marshall,
I.L.
and E.S.
Macroeconomico para
11) Moles,
Peter
la
(1988)
SUva,
"Fluctuaciones
del
Precio
del
Cobre,"
in
Informe
Empresa #35, September 1998, pp. 38-60.
[Online
text]
Financial
Management
Risk
2,
available
http://chigley.ebs.hw.ac.uk/finance/fr/index.html.
12) Smith,
C.W. "Corporate Risk Management: Theory and
Derivatives,
Summer
13) Shiller, Robert,
44
J.,
The Joumal of
1995, p. 21-30.
Macro Markets: Creating
Economic Risks, Oxford University
14) Tkole,
Practice",
Institutions for
Managing
Society's Largest
Press, 1993.
"Notes on International Financial Architecture," Toulouse, 2000.
Date Due
^h^l^
Lib-26-67
MIT LIBRARIES
3 9080 02527 8885