fi
-
mtm
MIT LIBRARIES
3 9080 03317 5909
Massachusetts Institute of Technology
Department of Economics
Working Paper Series
Sudden Financial Arrest
Ricardo
J.
Caballero
Working Paper 09-29
November 8, 2009
RoomE52-251
50 Memorial Drive
Cambridge,
This
MA 02142
paper can be downloaded without charge from the
Paper Collection at
httsp://ssrn.com/abstTact= 504985
Social Science Research Network
1
?fV.'.ws;
Sudden Financial Arrest
Ricardo
J.
Caballero^
November
08,
2009
Abstract
There are
striking
and
terrifying similarities
that of a financial system.
sudden cardiac
In
failure of a heart
the medical literature, the former
In
arrest (SCA). By analogy,
financial arrest (SFA).
between the sudden
this article
I
I
refer to
its
is
referred to as a
financial counterpart as a
describe SFA and
its
and
sudden
treatment guided by
its
medical counterpart.
Keywords:
Financial
crisis,
panic,
defibrillation,
complexity, Knightian uncertainty,
moral hazard, financial network, capital requirements, contingent
insurance, IMF
JEL Codes:
1
.
for the
G20
Mundell-Fleming Lecture delivered
Blanchard, Morris Goldstein, Bengt Holmstrom, Pablo Kurlat, Juan
comments, and Jonathan Goldberg
for
Tenth Jacques Polak Annual
at the
Research Conference, IMF, November 5-6, 2009 and the IMF Economic Review.
their
contingent
_'
E32, E44, E58, F30, GOl,
MIT and NBER, Prepared
capital,
thank
Viral
Acharya, Olivier
Ocampo and conference
participants for
outstanding research assistance.
I
First draft;
October
22,
2009.
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium IVIember Libraries
http://www.archive.org/details/suddenfinancialaOOcaba
Introduction
/,
_,,'-^.
"Sudden cardiac arrest (SCA)
^_^^-
'
a condition
is
and unexpectedly stops beating. When
to the brain
i.
and other
this
in
happens, blood stops flowing
SCA usually causes death
vital organs....
treated within minutes...." (NHLBI/NIH)
There are
striking
and
terrifying similarities
that of a financial system.
In
which the heart suddenly
,
.
I
refer to
between the sudden
its
not
.
the medical literature, the former
cardiac arrest (SCA). By analogy,
if It's
failure of a heart
referred to as a
is
financial counterpart as a
and
sudden
sudden financial
arrest [SFA).
When
an
investors
economy enters an episode
of SFA, panic takes over, trust breaks
and creditors withdraw from their normal
financial
down, and
transactions.
These
reactions trigger a chain of events and perverse feedback-loops that quickly disintegrate
the balance sheets of financial
institutions,
eventually dragging
down even those
institutions that followed a relatively healthy financial lifestyle prior to the crisis.
article
I
draw on the
parallels
between SCA and SFA
a
component
of systemic insurance than
willing to
to characterize the latter
and to
much
larger
pragmatic policy framework to address SFA requires
argue that
In this
most policymakers and
a
politicians are currently
go along with.
An important
risk
prevention for
factor behind
CAD
is
a
SCA
healthy
is
coronary artery disease (CAD), and the front
lifestyle.
However, the medical profession
is
line
keenly
aware that people make poor choices regardless of warnings, and that even those who
do adopt
fatal
a
healthy lifestyle and have no
known
SCA episode. The pragmatic response
risk
conditions
to these facts of
may
life
is
still
to
experience
a
complement
preventive healthy lifestyle guidelines and advise
w/ith
an effective protocol to prevent
death once SCA takes place. The main (and perhaps only) option to treat SCA once
triggered
the use of a defibrillator. Moreover, the window/ of time for this treatment
is
to be effective
is
readily available
very narrov^, just a few minutes, making
in
as
many
places as
is
it
crucial to
have defibrillators
economically feasible.
Unfortunately, the pragmatic approach followed by the medical profession
the
risk of
in
reducing
death associated with SCA contrasts sharply with the stubborn reluctance to
supplement the
financial equivalent of
CAD-prevention type policies (mostly regulatory
requirements) with an effective //nonc/o/ defibrillator mechanism. The main antidote to
SFA
is
massive provision of credible public insurance and guarantees to financial
transactions and balance sheets.
In this
analogy, these are the financial equivalent of a
defibrillator.
,
,
.
The main dogma behind the great resistance
.
the policy world to institutionalize a
in
moral hazard argument:
public insurance provision
is
were
an economy, the argument goes, banks and their creditors
to
be implanted
would abandon
all
in
a fuzzy
forms of healthy financial
lifestyle
If
the financial defibrillator
and would thus dramatically
increase the chances of an SFA episode.
This
moral hazard perspective
a
sudden urge
to
the equivalent of discouraging the placement of
because of the concern
defibrillators in public places
would have
is
consume cheeseburgers
chances of surviving an SCA had risen as
But actual behavior
is
less
a result of
that,
upon seeing them, people
since they
would
the ready access to defibrillators.
forward looking and rational than
is
implied by that
People indeed consume more cheeseburgers than they should, but
independent of whether
all
this
is
defibrillators are visible or not. Surely, there
advocating healthy habits, but no one
making
realize that their
in
their right
a
is
a
or less
need for
mind would propose doing so by
available defibrillators inaccessible. Such policy
an incentive mechanism, and
more
logic.
would be both
human tragedy when an episode
of
SCA
ineffective as
occurs.
By the same token, and with very few exceptions (Fannie and Freddie?), financial
institutions
dreams
and investors
Nothing
(financial) death,
their mind,
is
is
if
which are driven by
further from these investors' minds than the possibility of
and hence they could not ascribe meaningful value to an
meant
for
someone
This
else.
compression and undervaluation during the
that
portfolio decisions
not by distant marginal subsidies built into financial
of exorbitant returns,
defibrillators.
mode make
in bullish
one believes
in
is
boom
simply the other side of the
the near-
in
irrelevance of anticipated subsidies during distress for private actions during the
Of course, once the
crisis sets in,
risk-
phase. Logical coherence dictates
then one must also believe
this undervaluation,
aid which, in
insurance acquires great value and leads to
boom.
more
risk-
taking and speculative capital injections into the financial system, but by then this
mostly desirable since the main economy-wide problem during
little,
not too much, risk-taking.
panic,
indeed
and shortsellers to
fail
from
The
feast, as
last
thing
in
at this
time
and
liquidity dry-ups, for
middle of
a
too
for creditors to
which there
is
no
place aside from ill-timed "market discipline" or a
high-fatality risk surgery. Indeed, attempting to "resolve" a large
institution in the
is
is
is
they suddenly realize that financial institutions can
self-fulfilling runs, fires sales,
counteracting policy framework
we need
a financial panic
.
panic,
when
and interconnected
asset prices are uninformative and hence
"resolution" decisions are largely arbitrary, carries the serious
risk
of adding fuel to the
fire (panic)."
One way
to get a sense of
how much
the market values the "too big to
fail"
insurance provided by the
government is to compare the cost of funding for small and large banks. Baker and McArthur (2009)
compare the average costs of funding for banks with more than $100 billion in assets to the average costs
for banks with less than $100 billion. They find that between the first quarter of 2000 and the fourth
quarter of 2007, the large banks' costs were 0.29 percentage points lower than the small banks, averaging
across time.
Between fourth quarter 2008 and second quarter 2009, the spread increased to 0.78
percentage points. Clearly, there are many reasons why larger and smaller banks can have different costs
of funding: different types of assets, different amounts of leverage, and so on.
Baker and McArthur
(2009) take the difference between these spreads, 0.78 minus 0.29, as a crude upper-bound on the
would
subsidy associated with the solidification of the "too big to fail" policy after Lehman's collapse.
I
suggest an alternative interpretation: During
little
importance, while during the
stability.
crisis,
it
boom
times, the "too big to
fail"
insurance was there but of
became much more important and probably
a
source of
In
any event, when SFA does take place,
policymal<ers that there
is
it
becomes imnTediately apparent
to pragmatic
no other choice than to provide massive support to distressed
and marl<ets, but since the channels to do so are not readily available,
institutions
precious time and resources are wasted groping for a mid-crisis response (recall the
many
during the early stages of the
flip-flops
(which
I
am
TARP implementation).
not) that hubris plays only a small role during the
about incentive problems due
If
one
boom and
is
of the
instead
it
view
is
all
then one must
to anticipated subsidies during distress,
believe that savvy bankers and their creditors anticipate intervention anyway. Hence the
incentive benefit of not having financial defibrillators readily available does not derive
from the absence of
a
that improvised ex-post interventions
from SFA. This
logic
framework but from the
well designed ex-ante policy
seems contrived
may
fail
to be deployed in time to prevent
at best as the
foundation for
a policy
that does not include readily available financial defibrillators.
In
summary,
innovation
much
in
is
it
a fact
of
life,
policymakers
regulatory creativity
defibrillator
is
a
very
weak
economic tragedy during
:
incentive
will
framework
'
,
continue to happen regardless of
may muster. The absence
of
mechanism during the boom phase, and
a financial crisis.
We
need
a
a
how
financial
a potential
more pragmatic approach
than the current monovision CAD-style, hope-for-the-best, approach.
endow
death
and of cognitive distortions, financial complexity, and
SFA episodes
particular, that
real risk
We
to
SFA
need to
the policy framework with powerful financial defibrillators.
Modern economies
already count on one such device
in
the lender of last resort
facility
(LOLR) housed at the central bank, but this has clearly proven to be insufficient during
the recent
•
crisis.
I
discuss three supplements to this facility:
Self insurance,
this
is
which
reflected
in
is
a
where policymakers'
call
for
instinct
lies.
In
the current context
higher capital adequacy ratios,
systemically important financial institutions.
especially for
•
Contingent capital injections, which
basic idea
is
where most academics'
differ
specific or systemic events,
and on whether the source of
in
perceived probabilities of
a
is
is
capital
enormous
that the
is
external to
distortion in
catastrophe during panics can be put to good use
Providing these can be as effective as capital injections
at a fraction of the
expected cost (when assessed
panic-driven probabilities of
practice, there are
good reasons
to have
it
insure. For large shocks, there
always
is
in
at
dealing with the panic
reasonable rather than
catastrophe).
a
mechanisms. For normal shocks,
is
in
place
some
of each of these types of
probably easiest to have banks
events, public). However, the large panic
of guarantees,
self-
fundamental component, which
a
best addressed immediately with contingent capital (private at
amounts
not
economic agents greatly overvalue public insurance and guarantees.
since
In
is
the most cost effective mechanism for
The basic idea
of SFA.
when
The
the debt-convertibility proposals).
Contingent insurance injections, which
component
lies.
on whether the contingency depends on bank-
needed. Proposals primarily
the panic
instinct
to reduce the costs associated with holding capital
the distressed bank or internal (as
•
is
component
of an
first
and
and crossis
probably
in
extreme
SFA episode requires large
which would be too costly and potentially counterproductive
they add to the fear of large dilutions) to achieve through capital injections. For
component,
One
a
contingent-insurance policy
particularly
flexible
form of
Insurance Credits (TlCs) proposed
contingent
(on
systemic
events)
a
in
is
(if
this
the appropriate response.
contingent insurance
program
is
the Tradable
Caballero and Kurlat (2009a). These TICs act as
CDS
to
protect
banks'
uncertainty. They are a (nearly) automatic, pre-paid, and
assets
against
spikes
mandated mechanism
in
to ring-
fence assets whose price
some
Citibanl<
is
severely affected by SFA, as
of
emerging markets which has
a
within developed economies.
SFA adds
rather than
own
its
ingredients.
in
in
the U.S. for
the U.K.^
focus on the problem for
1
close parallel with the issues faced by the financial sector
For emerging
becomes entangled
itself
international
was done ex-post
and Bank of America assets and was offered more broadly
The international dimension
sovereign
it
the
in
domestic
(just)
market economies are acutely aware of
markets,
as
crisis
liquidity.
main shortage
the
Most policymakers
danger, which
this
often the case that the
is
it
is
one of
is
emerging
in
one of the main reasons
they accumulate large amounts of international reserves. However, large accumulations
of reserves are the equivalent to self-insurance for domestic banks, and as such are
costly insurance facilities.
many
For this reason,
of us
have advocated the use of
external insurance arrangements, and the IMF has spent a significant
amount
attempting to design the right contingent credit
paper
line facility.
In
this
I
of time
propose
creating a system akin to the TICs but aimed at supporting the value of emerging market
new and
legacy emerging debt during global SFA episodes.
as E-TICs
and envision them as being controlled by the IMF rather than by the
'
other developed economies' governments.''
In
the rest of the paper
I
develop
this line of
the analogy between SCA and SFA. Section
IV
It
1
.
'
in
greater detail. Section
II
covers
focuses on financial defibrillators. Section
adds an international dimension to the analysis, and Section V concludes.
turns out that the Bank of America deal
contain the panic.
optimal because
it
The
U.K.
system was
was never
signed, but the perception that
less successful in
terms of the takers than
appropriately dealt with the
on
a
it
it
had been was enough to
would have been
socially
was voluntary and very expensive. Both aspects would be improved by the TICs framework.
For developed economies, the international liquidity shortage problem
least
U.S. or
.
argument
III
refer to these instruments
swap arrangements between major
contingent (to SFA) basis.
A more
delicate
problem
is
much
less
significant
central banks. These should remain
for these countries
and
in
it
was
place, at
stems from the high degree of
cross-borders interconnectedness of their financial institutions and the potential arbitrage and free riding issues
that
may emerge from
coordination issues which
differences
I
in
don't develop
the type of financial defibrillators available. This raises international
in this
paper but that obviously need to be addressed.
11.
An Overview of SFA
In this
section
I
SCA. This sets up the case for
in
""'
the next section.
What
a. A.
._
•
.
;
.
characterize SFA, drawing a close financial-analogy with the health-
equivalent issues
in
.
.
,
SFA and what are
Sudden financial arrest (SFA)
system suddenly vanishes.
is
a condition
When
discussion of financial defibrillators
•
'
'
is
a
this
its
immediate causes?
in wliicli trust witliin
and toward
and wealth destruction
financial
happens, the financial system freezes and credit
stops flowing within the financial system and to the real economy.
contractions
tlie
in
SFA causes severe
the absence of an immediate systemic policy
response.
The most immediate cause
is
a truly
unexpected event that triggers enormous confusion
(see Caballero and Kurlat 2009a). Often there
is
an observable shock
(e.g., a real
estate
crash, a sovereign default, the uncovering of a significant fraud, or a sharp decline in
terms of trade) but
this
financial institutions
fulfilling
is
confusion:
shock
is
small relative to the observed response
and markets.
Why
such
It
a large
is
this
in
some key
apparent over-reaction that triggers
a self-
response? Are things worse than they seem?
Who
compromised? Which instruments? What are the transmission channels? How
others react to this over-reaction?
How
will
the government react?
These doubts by themselves bring about reluctance to engage
and are triggers of good old-fashioned financial
enough
multipliers,
in
financial transactions
which are often sufficient
to cause a real contraction.
The now-classic contributions of Bernanke and Gertier (1989), Bernanke
Kiyotaki
and Moore (1997), and others
illustrate
between asset markets and the macroeconomy.
firm decrease
sell
will
in
value, the net
financial assets or
worth of the firm
decrease investment or
the
When
is
hiring.
et
a!
(1999),
powerful feedback channels
the assets held by
a
leveraged
quickly eroded, forcing the firm to
These actions may
in
turn further
Brunnermeier and Pedersen (2009a, 2009b)
decrease the value of the asset.
type of feedback
market
liquidity dries
trying to
sell.
Loss spirals
a "loss spiral."
Such
may
be aggravated by
fire sales, in
fire
been extensively documented,
sales have
e.g.,
airliners (Pulvino
which
Pedersen also emphasize "margin
borrow against an asset leads
spirals,"
to sales.
v^hich
in
Increases
role of
these models, the supply and
demand
but also equilibrium leverage.
News
in
firms'
the present
ability to
of equity
as
crisis,
Gorton
Geanakoplos (2003, 2009), and Fostel and
and Metrick (2009) have documented.
Geanakoplos (2008) highlight the
in
1998), and the
- the amount
haircuts
in
- have been dramatic
to finance a given asset
reduction
a
the equity
in
Brunnermeier and
convertible bond market (Mitchell, Pedersen and Pulvino 2007).
shift
this
up because the natural holders of an asset are simultaneously
markets (Coval and Stafford 2007), the market for
needed
call
margins
for loans
theory of "leverage cycles."
in a
In
determines not only the interest rate,
that causes increases
in risk
the equilibrium haircut, with large effects on asset prices.
or disagreement can
,:.
However, there are two key additional and related ingredients which have the potential
to
leverage the consequences of these mechanisms to SFA level: Complexity and
Knightian uncertainty.
Reality
post,
it
is
is
immensely more complex than models, with
millions of potential
easy to highlight the one that blew up, but ex-ante
market participant and policymaker knows
their
own
local
is
weak
a different
links. Ex-
matter. Each
world, but understanding
all
the possible linkages across these different worlds (which are mostly irrelevant except
during a severe
from irrelevant to
uncertainty
when they
crisis
critical
turn
linkages,
(when the unknowns
critical)
triggers
shift
is
too complex.
massive
This change
uncertainty,
in
paradigm,
indeed
Knightian
from known to unknown), and unleashes
destructive flights to quality.
In
Caballero and Simsek (2009a)
followed by widespread panic
in
we
capture the idea of
the financial sector.
In
a
sudden
rise
in
complexity,
our model, banks normally collect
basic information about their direct trading partners which serves to assure
them
of the
soundness of these relationships. However, when acute
parts of the financial network,
partners, but
it
for the
banks to learn about the health of their
order to assess the chances of an indirect
in
in
not enough to be informed about these direct trading
becomes important
also
trading partners,
it is
emerges
financial distress
hit.
And
as conditions
continue to deteriorate, banks must learn about the health of the trading partners of the
trading partners, of their trading partners, and so on. At
information gathering becomes too large and banks,
choose to withdraw from loan commitments and
ensues, and the financial
this
crisis
facing
enormous uncertainty,
positions.
illiquid
A
flight-to-quality
Caballero and Simsek (2009b)
In
the cost of
point,
we show how
complexity mechanism interacts and greatly amplifies the collateral and
mechanisms.^
In
spreads.
now
some
we
Caballero and Krishnamurthy (2008a)
amplification
of
role
Knightian
defibrillation" in this context).
We
sales
'
-
-
fire
.•
illustrate
(and
uncertainty
with a model and examples the
the
pointed out that most
triggered by unusual or unexpected events. The
,
effectiveness
of
flight to quality
common
"financial
episodes are
aspects of investor behavior
across these episodes —re-evaluation of models, conservatism, and disengagement from
risky
activities—
immeasurable
on
tail
indicate
that
these
episodes
and not merely an increase
risk)
outcomes and worst-case scenarios
in
involved
in risk
framework.
We
exposure. The extreme emphasis
we
place the origins of the current
crisis in this
argue that financial instruments and derivative structures underpinning
the recent growth
in
in
credit markets
were complex.
the financial landscape over the
CDOs, CLOs, and the
like.
last five
Indeed, perhaps the single largest
years was
in
complex credit products:
Because of the rapid proliferation of these instruments, market
Haldane (2009) masterfuliy captures the essence of the counterparty uncertainty problem that can
complex modern
(i.e.,
•
Caballero and Krishnamurthy (2008b)
change
uncertainty
agents' decision rules suggests uncertainty
aversion.
In
Knightian
financial
dimension Sudoku
network: "...Knowing your ultimate counterparty's
risk
then becomes
like
arise in a
solving a high-
puzzle...."
10
couldn't
participants
refer
to
historical
a
record
measure how these
to
structures would behave during a time of stress. These
two
of history, are the preconditions for rampant uncertainty.
factors, complexity
When
was the
Early
losses experienced by
became uncertain about
on
in
the
AAA subprime
their investments.
subprime mortgage investments, given the
and lack
subprime
defaults on
mortgages occurred, many market participants were taken by surprise
investments were reacting.
financial
how
at
their
the most prominent example of this
crisis,
tranches.
It
was
at this point that investors
Had the uncertainty remained confined to
relatively small size of the
the financial system could have absorbed the losses without too
subprime sector,
much
dislocation.
However, investors started to question the valuation of the myriad other credit products
— not
just
mortgages—
The
investments.
market.
The
that had been structured
was uncertainty and
result
policy response to this
initial
in
much
the
same way
freezing up across the entire credit
a
freezing
was
timid,
which kept the stress on
the financial system alive until the latter eventually gave up and a
developed
In
(after
absence of SFA. Traditional
fundamental shocks, but
it
panic and confusion that
a
may go through
resources
tl.B.
Every
in
a
factors are often the triggers, and
economy
is
rich financial
real
damage
is
is
often contained
good
in
the
impact of
only after they are themselves amplified by widespread
deep
crisis
is
Absent the panic, the
likely to set in.
financial
few scary arrhythmias, but there are too many safety valves and
at risk?
financial
markets for
modern
dynamic process of
network.
it
to
come
to a sustained halt.
'
with a reasonably
inextricably tied to the
a
is
.
hence there
financial multipliers can greatly increase the real
modern developed
Who
blown SFA episode
,
reason to try to manage them, their potential for
heart
full
Lehman's demise).
summary, while fundamental
subprime
as
financial
system
is
financial innovation
at risk of SFA.
and
It
is
a risk
to the complexity of
As highlighted by Cabailero and Krishnamurthy (2008a),
builds over time as untested financial innovations
an unavoidable side effect of
a
grow
in
relative importance,
which
it
is
period of prosperity.
11
Of course, there are factors that elevate SFA
risk in
addition to the
mere passage of
(innovation) time. There are the usual bad habit suspects, such as a high concentration
of real estate loans
in
leveraged
financially
appreciations,
banks' balance sheets, currency and liquidity mismatches
institutions,
regulatory and
and
lending
supervisory
consumption
negligence,
booms,
sharp
in
asset
weak corporate governance,
sustained appreciations and current account deficits, excessive exposure to terms of
trade shocks, low levels of international reserves, and so on. But unfortunately these
factors have very limited predictive
power
for SFA, especially in sophisticated financial
markets (see Caballero and Kurlat 2009a).
The most
visible of
these factors, the macroeconomic ones, play
emerging markets, since the strong
likelihood that large
macroeconomic shocks
lessons have been
largely
However,
this tight
and developed
financial markets. Reinhart
Of
and Reinhart (2008)
the
These
not
did
a
banking
crisis
is
is
weak
in
deep
find that, for low
and
higher during episodes of
(The probability conditional on a bonanza
14 percent unconditional probability.)
in
increases
why emerging markets
connection between macroeconomic events and SFA
"capital flow bonanza."
absent
flows
capital
in
crisis.
middle income countries, the probability of
a
detectable role
quickly translate into SFA episodes.
learned, which explains
implode during the recent global
of
pro-cyclicality
a
is
21 percent, versus
For high income countries, this connection
is
their sample.
course
SFA
is
more
macroeconomic shock,
place during
a
in
likely
the
to
take
same sense
stressful situation, but this
macroeconomic imbalance.
Instead, the
over time and becoming deeply
place
that an
the
midst
SCA episode
of
is
a
contractionary
more
likely to
does not mean that the cause of SFA
main SFA
embedded
in
in
risks
are
more
silent, building
the financial network. These
take
is
a
slowly
risks are
very
12
hard to detect
time to prevent an SFA. Going back to
in
again from the NHLBI/NIH
"The major
risk
analogy with SCA,
tine
quote
I
site:
factor for SCA
is
undiagnosed coronary artery disease (CAD). Most
people who have SCA are later found to have some degree of CAD. Most of these
people don't know that they have CAD
has no signs or symptoms. Because of
it
is,
detected
it.
Most cases of SCA happen
no known heart disease prior
Yes,
this
we
SCA
until
all
knew
in
people who have
of
that there could be a large correction
subprime mortgages. Instead,
developed
interconnections
financial
silent
CAD and who have
to SCA.
would complicate subprime mortgages and lenders
amount
"silent"— that
is
doctors and nurses have not
this,
in
it
was
first,
in
and that
a
recession
would
price of real estate or
the particulars of the complex
process
the
estate markets and that
in real
was not the
surely follow a crash. But the hidden financial-CAD
the
CAD
occurs. Their
new
creating
of
financial
instruments and products. There are, of course, ample anecdotes and media-created
gurus, but these are mostly stories,
which
long and
is
There
way
is
it
crisis
of
them,
should) claim
to
have
(or
•'
'
,.
.
-,
,
.:''
-
•
;
;
an economic parallel to the earlier hidden-CAD quote, found
learning literature.
over time
in
Like
plaque
in
in a
a
in
catastrophe.
bad news before "business as usual" turns
do with physics, but for financial-CAD,
this
into crisis.
can occur
it
the social
Indeed, the financial
state of "business as usual" even as increasing
people receive bad news about fundamentals. As with plaque,
v.
news can accumulate
one's circulatory system, bad
the financial system without causing
system can continue
of
were never part
the precise channels through which complexity would turn
__
did.
crucial details
No one can
conveniently forgotten.
understood before the
out the
where the
can take
numbers
a critical
of
mass
For CAD, the reasons have to
when bad news
knowledge. For example, as shown by Caplin and Leahy (1994),
if
is
not
common
private information
is
13
revealed through irreversible actions, agents
low threshold before acting upon them;
over time without any
numbers
visible
if
the bubble's existence
Brunnermeier (2003). Such behavior
CEO Chuck
is
Prince observed
in
is
not
still
expectations cross a
even as large
persist
coordination
if
common
is
needed
knowledge, as
in
to burst
Abreu and
not inconsistent with rationality. As the then
July 2007,
"When
the music stops,
things will be complicated. But as long as the music
dance. We're
may
bubbles
occurring,
is
until their
Vi/ait
can allow bad information to accumulate
Similarly,
signs.
of agents learn that a bubble
the bubble and
this
v;/ill
is
in
terms of
liquidity,
playing, you've got to get up
dancing." Even with the rough contours of the
crisis clear in
mind, he kept dancing, because the timing and exact contours of the
crisis
Citi
and
the CEO's
were not
clear.
Diagnosis
!!.C.
Diagnosis
it
is
is
the key element
difficult to predict
are machines that do
for the untrained
in
deciding
how and when
the occurrence of SCA
it
its
to react. In health matters, while
diagnosis
is
not hard, and
fact there
in
nearly automatically (more specifically, defibrillators designed
check for severe arrhythmias before applying an electric shock to the
heart).
In
economic matters, diagnosis
a
is
much
perhaps because even when
trickier task,
violent, events are less abrupt than cardiac arrest. Usually
warning
sign. In contrast,
there
is
almost always
a
SCA happens without
narrow window of opportunity to
react to early warnings and prevent a full-blown SFA episode. Unfortunately,
much
is
of this advantage
sufficiently clear to
needed.
than
in
is
wasted as
a conclusive diagnosis
persuade reluctant
politicians that
By then, solving the problem often
is
economic
is
delayed
until
in
practice
the evidence
emergency treatment
for
SFA
is
an order of magnitude more expensive
an early intervention. The following quote from
assistant secretary for
a clear
Phillip
L.
Swagel (2009), then
policy, vividly captures this reality:
14
"Political constraints
were an important factor
forward proposals
to put
address the credit
TARP were first
that later turned into the
2008: buy assets,
to
them,
insure
inject
in
the reluctance at the Treasury
early
crisis
written
down
2008. The options
in
at the Treasury
in
March
or
capital into financial institutions,
massively expand federally guaranteed mortgage refinance programs to improve
asset performance from the bottom up.
saw
But we at the Treasury
little
prospect of getting legislative approval for any of these steps, including a
massive program to avoid foreclosures. Legislative action would be possible only
when Treasury Secretary Henry Paulson and Federal Reserve Chairman Ben
Bernanke could go
even though by then
The bickering early on
excesses
in
Congress and attest that the
to
could well be too late to head
it
is
often Schumpeterian
financial balance sheets
The problem
is
and
it
is
in
was
crisis
it
nature.
healthy that
at the doorstep,
off." (Swagel, p. 4)
A boom accumulates many
some
of
them
that the ratio of inefficient panic-contagion-driven destruction to healthy
cleansing can rise quickly. This
is
the policy conundrum, where to draw the line and,
even more importantly, to muster the support to act quickly once the
intervention takes place too early, there
there
In
is
great
practice,
are cleansed.
risk of
some
is
no healthy cleansing.
line
is
crossed.
If
takes place too late,
If it
an SFA episode.
objective
measurement
is
needed. For SCA, electrocardiograms (EKGs)
can be used to detect and locate the source of several heart problems. For SFA, there
are
many
financial
The blue
EKG-equivalents. Spreads of
markets and
line in Figure 1
TED spread, the
is
difference
both since the 1990s
financial
institutions.
until
all
sorts can
be used to spot dysfunctional
'
.,
the VIX at market close, from the CBOE.
,
The red
line
is
the
between 3-mo LIBOR and 3-mo constant maturity Treasury,
the present.
turmoil. The graph
It is
apparent that both capture well episodes of
shows the unprecedented degree
of financial-markets
15
was considerable well
dislocation during the current crisis and that the level of unease
before Lehman's collapse. The TED spread
in late
2007 and early 2008 exceeded
levels
associated with the Persian Gulf War, the LTCM/Russian default episode, and 9/11.
There
is
also a subtle but important difference
probably
a
good warning
SFA
for
occurrence of an SFA event by
picks up distress
in
proclivity but
itself.
This
is
it
is
indicators.
The VIX
is
too sensitive to determine the
not surprising since by construction the VIX
the riskiest segments of assets (equity). The TED, on the other hand,
reflects trouble in the safest counterparties,
that financial chaos
between these two
is in
full
1.
it
is
a
more
reliable indicator
"
force.
Figure
and hence
•
Indicators of Panic since
1990
5,0
4.5
80
4.0
70
60
50
40
30
w
20
10
i*§*^*^^
J
o
m
ro
o>
c
O)
in
Gi
ai
m
en
T-
CT>
CD
T—
T—
CD
T-
CD
T—
en
CD
1—
T—
hcn
en
1—
OJ
fN
rj
fsj
CNJ
CN
CN
OvJ
rsi
a>
CO
en
en
T—
en
en
en
T—
CJ
OJ
-VlX{lettaxis)
O
ro
^
uo
CN
CN
CM
(N
CNJ
rvj
fN
rj
o
o
o
o
o
CM
CM
fM
^
CN
TED
O
o
o
o
o
o
o
tJD
o
o
og
o
o
m
o
o
CM
£M
CN
S
S
h-
CNI
0.0
en
o
(Right a)OS)
Sources: CBOE, Datastream, Federal Reserve
Figure 2 offers a bird-eye view of the crisis from the perspective of the
indicators of panic. Note that the
TED spread soared the week
of
same two
Lehman's collapse and
16
retreated very sharply as the Capital Purchase Program (CPP) and Temporary Liquidity
Guarantee Program (TLGP) were announced on October
Figure
14, 2008.
Indicators of Panic During the Crisis
2.
90
5,0
80
70
60
50
40
30
20
10
>~WvJ
^^f'-'v^Af^^bi.
-i-
ooooooooooooooooooooooo
rsjtNtNCNJCNiCNOJrsJfNjrsjCslfsjCMCNCNCNCNrMCNJCNCNCNCM
0,0
tDtDiotDuDtDr^r^r^i^r-r^coxicocDoott'ajQicnoicD
CO
^
CO
CO
CO
LO
CO
ri
en
r-
CO
CO
to
'-
V-
fO
CO
in
-VIX
h-
co
en
(left
axis)
CO
CO
CO
to
CO
CO
-r-
T-
CO
UO
r^
-TED
CO
Ol
CO
CO
1-
>^
co
CO
CO
un
indication of the extent of the
line in Figure 3
is
"New
come by
The new issuance
series
is
the
real estate; autos; credit cards;
real time,
in
damage being caused by
sum over
O)
come from
JP
the following categories of ABS:
in
%.
•
but they offer a clear
the underlying
Issuance of Asset-Backed Securities
Adrian and Shin (2009); the data originally
(O
r^
(Right axis)
Sources: CBOE, Datastream, Federal Reserve
Quantity indicators are often harder to
CO
crisis.
The blue
Previous 3 Months," from
Morgan Chase. ^ The
"home
crisis in
equity (subprime)"; commercial
student loans; non-US residential mortgages; and other. The data were provided
by Tobias Adrian.
17
this
market
apparent from the disappearance of new issuances. The pink
is
"impiied spread on the 2006-1
default by
AAA
2006 vintage."
the
crisis
quantity
AAA ABX, which measures
is
the
the cost of insuring against
tranches of subprime mortgage-backed securities of the first-half-of-
The spread data are from JP Morgan Chase and not only corroborates
impression from the quantity side but also makes
is
line
demand
,
rather than supply driven.
Figure
3.
SCA
in
it
clear that the collapse in
;.
•
the Asset-Backed Securities (ABS) Market
700 T
350
600
300
500
400
300
200
100
8
r^
r^
8
CO
oD
CO
00
(X)
cMCMrMCNr\lcMr>JojojCNjrNtNtNcN(>jcsj
c^
S
iS
•2005-1
AAA ABX
Sources: JP
n
ft
New
LD
r-
O)
Issuance of Asset-Backed Securities
Morgan Chase, Adrian and Shin (2009)
To conclude, perhaps one of the best indicators of an imminent SFA episode
mismatch between the
size of the
value-destruction that takes place
shock that
in
is
is
the large
"credited" for the contraction and the
banks' balance sheets. Such mismatch
only to cause, but also to reflect confusion and fear
and Kurlat (2009a) we constructed an estimate of
in
this
financial markets.
mismatch, which
I
is
In
bound not
Caballero
reproduce
in
18
Figure
For
4.
this,
we computed
term debt) of the major
U.S.
the evolution of the' market value (equity plus long
banks since January 2007.^ From
this
we obtained an
estimate of total losses on the right hand side of these banks' balance sheets. Absent
any feedback
effects, this should
be equal to the losses suffered by the assets on the
hand side of the balance sheet. However,
on the
right
as illustrated in Figure 4,
we
left
find that losses
hand side are on the order of three times the IMF's (evolving) estimates of
losses related to
mortgage assets accruing to
Figure 4. Losses from
Mortgage
U.S. banks.
assets. Total Loss of
Market Value and
- Total
(OSS of
Multiplier.
market value {eouitv plus debt) of
banks
'
Total losses
US banks
Jari07
Tine
Mar 07
tvlayO?
procedure for estimating
market
Jul
this
07
Sep 07
was
capitalization, excluding increases
Nov 07
Jan 08
Mar 08
tulav08
as follows; For equity,
in
Jul
08
we
Aug 08
Oct 08
from mortgage assets accruing to
as estimated bv
Dec 08
'MF
Mar 09
Way 09
simply tracked the evolution of each bank's
the m.arket cap due to issues of
new
shares. For debt,
we
estimated
the duration of each bank's long term debt (including any preferred shares) from the maturity profiles described
in
December 2007, assuming the interest rate was equal to the rate on lO-year Treasuries
the spread on 5-year CDS for each bank, obtained from JPMorgan. Assuming an unchanged maturity profile,
the 10-K statements as of
plus
we then
tracked the changes
in
the implied market value of each bank's long term debt on the basis of the
evolution of the CDS spread. The banks included
tests" plus
Lehman, Bear Stearns,
The IMF uses
a projection of
in
the calculation are the 19 banks that underwent the "stress
Merrill Lynch, V^achovia,
macroeconomic
and V^ashington Mutual.
variables and default rates to estimate losses on loans, and market
values to estimate losses on subprime-related securities. To the extent that market prices of securities overreacted
due to
fire sales,
our procedure understates the multiplier.
19
Sources: IMF Global Financial Stability Reports, banks' financial statements and
JPMorgan, From Caballero and Kurlat (2009a).
Beginning
in
2008 and increasingly
after the
fall
in
from the overall disruption of
losses
and losses on other types of assets w/hich
mortgage market
Most
of
itself.
„
assets.
SFA was
They
high.
also
episode was taking place.
-
The market began
financial markets, the severe recession
exceeded the estimated losses from the
.
,
.
these financial distress indicators were available
thereafter. By the end of the
for
-
far
,.,
of Bear Stearns, the overall loss in
market value becomes larger than the losses from subprime
to price
,
summer
in
real
time or shortly
of 2007, they clearly reflected that the potential
showed by the end
Policymakers
at
of the
summer
of 2008, that a severe
SFA
the Fed and Treasury were acutely aware of
these scenarios, but the treatment options at their disposal were limited by
political
constraints and lack of ex-ante preparation.^
ii.D.
Treatment
'
,
SCA requires immediate treatment with
a defibrillator.
That
is,
an electric shock to the
heart, to restore
normal rhythm. This must be done quickly as with every passing
moment
there
steep decline
Although
less
is
a
dramatic
SCA and the urgency
of
in its
its
in
the chances of
early stages,
treatment
is
SCA
survival.
once SFA reaches
accurate.
days that followed the Lehman and AIG events,
This
was
full
force, the analogy with
clearly the case during the
when suddenly even Money Market
funds became entangled and experienced panic-driven withdrawals that clearly had not
been anticipated by policymakers and bankers. Swagel recounted
See Wessel (2009) for
a fascinating
to
Bloomberg that on
account of some of the constraints faced by the Fed to act quickly during the
crisis.
20
Monday, September
15, the
day of Lehman's
collapse,'
"The general feeling was things
were working." And even
as officials
money market funds and
the drying up of the commercial-paper market
and market participants witnessed the run on
market redemption requests were $33.8
week-
the previous
2009).
the
In
the
full
extent of the
same Bloomberg
that Tuesday,
billion
crisis
was
Mohamed
article,
compared
-money-
to $4.9 billion
hard to understand (Stewart
still
El-Erian
recounted "Monday and
Tuesday, people didn't quite see what was happening. ..You had to be on the desk
payments and settlements system, cash and
failures
and
a
collateral, to start seeing
cascading market
trust.
Essentially, financial defibrillators are large public
guarantees of different components
were many
of the financial system's balance sheet. There
crisis.
of these put
in
For example, to staunch outflows from
place
program
commercial-paper
directly
from
for
facility
eligible
money-market funds
to
and
the
Fed
a
created
an ad-
temporary
backstop
a
purchase highly rated three-month commercial-paper
Through the Term
issuers through a special-purpose vehicle.
Asset-Backed Securities Loan
in
money market
funds and stabilize the commercial-paper market, the Treasury created
insurance
the
""^^
complete erosion of
hoc basis during the recent
in
Facility (TALF)
funding -the non-recourse element
is
program, the Fed provides non-recourse
key- for purchases of asset-backed securities.
Deposit insurance on bank accounts (FDIC) was extended (temporarily) from $100,000
to $250,000.
Stanley and
The Fed created
Goldman
facilities for
Sachs, the
become bank-holding companies
lending to investment banks.
Later,
Morgan
two remaining investment banks, were allowed
to reassure creditors and counterparties.
to
Moreover,
the FDIC, through the Temporary Liquidity Guarantee Program, provided guarantees of
newly issued senior unsecured debt of financial
10
Ivry,
Bob, Mark Pittman and Christine Harper. 2009.
Missing $63
Billion, "
Bloomberg.com, September
institutions.
"
_^-
^
Sleep-At-Night-Monev Lost
in
Lehman Lesson
8.
21
Europeans also implemented large public guarantees and interventions.
Scheme
Protection
the
in
UK backed more than
half a trillion
pounds
assets for RBS and Lloyds. The maturity of discount-window borrowing
the
the U.K., and the E.U.
U.S.,
The UK asset purchase
facility
In
large capital injections by
the
U.S.,
but after
the
initial
post-haircut
in
was extended
in
acquired commercial-
paper and corporate bonds and the ECB purchased covered bonds.
were
The Asset
In
addition, there
governments across the globe.
response was ad-hoc, inconsistent and unpredictable. Eventually,
much had gone wrong, there was convergence
to a consistent policy that
placed panic and the systemic nature of the problem at the core. According to the IMF,
between summer 2007 and June 2009, 49 front-page
in
the U.S., 18
the U.K., 49
in
Switzerland (IMF 2009).
in
policy
the Euro area, and 37
announcements were made
more
in
Japan,
Sweden
These measures included: interest rate changes;
or
liquidity
support measures such as changing reserve requirements or widening collateral rules;
capital injections;
right
debt guarantees; asset purchases; and asset guarantees. This was the
response and the steps that followed along the same direction eventually stopped
the economy's free
fall
and made
a
recovery possible.
The unfortunate aspect of the process that got us
to the right
answer
exasperating
is its
slow speed. For SCA, cardiopulmonary resuscitation (CPR) should be given to
having SCA
is
until defibrillation
can be done. The equivalent to CPR
in
a
person
the context of SPA
the ad-hoc and idiosyncratic (as opposed to systemic) policy response to the early
For emerging market economies,
stages of SPA.
that
is
available until an external rescue
obstacle
is
lack of political
apparent that
in
is
much time
is
all
is
mounted. For developed economies the main
support for activating
practice too
sometimes the CPR-equivalent
spent
a financial defibrillator.
in
Either way,
financial-CPR-mode, delaying
a
it
is
much
needed shock treatment.
Another important lesson from SCA treatment
term treatment.
The Lehman episode
is
that this
is
not the time to
illustrates that similar patience
is
initiate
long
advisable for
22
SFA. Once
fail,
more quoting Swagel
(2009),
when
the U.S. government allowed
Lehman
to
"The view at Treasury. ..was that Lehman's management had been given abundant
time to resolve their situation by raising additional capital or selling off the firm, and
market participants were aware of
and had time to prepare."
this
However,
it
took
"breaking the buck" at just one large money-market fund that was over-exposed to
Lehman
causing
for panic to
a
ensue among investors
money market mutual
in
funds,
drying up of short-term financing for banks and corporations.
highlights the
unreasonable
extreme
risk
of a financial system on the verge of
fragility
in
turn
This episode
SFA and the
associated with attempting financial surgery during these episodes."
Prevention
!LE.
For low-risk populations, the main preventative prescription
address the main factors
particular, there are
two main types
of financial crises: control of
of preventive policies often discussed
macroeconomic imbalances that can
market crashes; and regulatory controls to
in financial
a healthy lifestyle to
Economics has an analogous protocol.
CAD.
behind
is
limit
excessive aggregate
in
In
the context
lead to large asset
risk
concentration
balance sheets.
For example, the Treasury department
imbalances through
its
monitoring by the IMF.
liquidity buffers,
Framework
It
also has
and for imposing
and most interconnected
is
currently seeking to control
for Sustainable
macroeconomic
and Balanced Growth, which
proposed requiring banks to hold higher
stricter
firms.
It
requirements on the
has also
proposed
a
largest,
more
calls for
capital
and
most leveraged
unified
regulatory
approach that would merge the Comptroller of the Currency (OCC) and the Office of
The Lehman episode
institution
when
highlights the difficulty of predicting the fallout
the possibility of SFA
is
impending.
from the resolution of
a
major financial
There have been several proposals from academia and
policymakers to create mechanisms for orderly resolutions, but resolving a major financial institution, especially
during times of panic, inevitably involves an
consequences.
amount
of uncertainty that can
easily
generate catastrophic
.,_-
23
new
Thrift Supervision (OTS) into a
agency
for
consumer protection.
Continuous progress
ending
there
,.
made
being
is
job, as the financial
is
National Banl< Supervisor and create
,
,,
,
along these dimensions; however, this
an aspect of SFA prevention that has received disproportionally
For SAC, there
In
is
substantial research effort
known
that
became more
profession
in
developing
new
CRP was an
CRP was discovered
indicator to bear
mind, but
in
precise, the useful predictive features of
believes
CRP reading
a
in
the
in
1930 and doctors had
when
CRP became
upper-third
versus
when
a
fact
it
must be viewed
are quite familiar with
in
financial equivalent of
new connections
as only
one piece of
its
financial
better
clear.
The medical
lower-third
the
of
However, some
over-emphasize CRP,
clinical puzzle, a situation
•
hidden-CAD derives from the
particulars of the evolution of
within the financial network. Tracing and increasing the transparency
to
expand our
toolkit.
and other variables
measurements
an October
13,
variables as
we
into our
2008 WSJ
assistant Treasury secretary for
critical
measurement
in
limiting the possibility of panics.
We
have not found the financial equivalent of CRP testing, but early attempts are
underway
In
complex
economics.
of these connections should be a high priority
still
•
.
tests to detect hidden-
led doctors to
The
never
attention:
.
worry that the useful qualities of CRP testing has
we
little
,;
,
distribution implies a relative risk of 2.0 for a future coronary event.
in
a
is
the 1990s, doctors found that elevated levels of the C-reactive protein, or CRP,
strongly predict future coronary events.
long
.,,.
,
system adapts to new controls and regulations. Moreover,
Hidden financial-CAD detection mechanisms.
CAD.
,.,
.
dedicated
a
As with CRP, making more precise measurements of
already
know
to be important, and integrating these
models can help us
find predictors of financial distress.
(Summers Outlines Risks to Recovery) there
economic policy, who pointed to a near-complete
article
hedge fund
positions,
the interconnections
among
financial
is
a
quote to
Alan Krueger,
lack of information
institutions
on such
and mortgage-
24
One
interesting
value-at-risl< of tlie
being
step
first
in distress.
whole
is
Adrian and Brunnermeier's (2009) CoVaR measure of the
financial
system conditional on
a
given financial institution
Adrian and Brunnermeier argue that capital requirements should be
tighter for institutions that can be predicted to have a higher systemic risk contribution
in
more maturity mismatch,
the future; they find that higher leverage,
A
higher book-to-market predict future CoVaR.
measure proposed by Acharya
shortfall"
shortfall
is
a
(2009a, b).
al
A
is
the "marginal expected
firm's marginal
mean
The "protein" that CoVar and marginal expected
between the returns on
shortfall
seek to measure
different firms' assets.
of the response of the firm-j or the system to a shock to firm-i.
expected
shortfall
a
measure
the
is
a
IVlarginal
of the response of firm-i to a systemic shock.
Drehmann (2009) pursue
test for the joint
is
CoVaR
measure
is
is,
stock return for a given firm conditional on a lower-tail market
conditional correlations
to
expected
Acharya and co-authors apply the method using stock prices; that
calculating the
Borio and
and
measure of losses experienced by the firm conditional on aggregate losses
being large.
return.
et
related test
larger size
a different
type of early-warning system. They seek
presence of two different proteins: aggregate balance-sheet
imbalances and asset-price bubbles.
They develop
a binary indicator of risk
based on
deviations from trend of the quantity of private-sector credit and the prices of equity
and property.
Acharya et
al
(2009a, b) find that the marginal expected shortfall of a firm measured
during June 2006 to 2007, before the
Borio and
Drehmann
crisis,
predicts the firm's return during the
(2009)'s binary indicator
would have
prior to the current crisis only for versions that sufficiently
Their credit-to-GDP gap had exceeded a
refinancing activity. He argued that "...the economic
strict
crisis
crisis.
reflected heightened risk
emphasize property
prices.
threshold since 2001, but their equity
has given an unintended stress test of our economic and
financial indicators."
25
price gap
was below even
a
weak threshold
search for the C-reactive protein of financial
On
a
(pejoratively
institutions
financial
interconnected-to-fail)
the
is
of small institutions they
if
referred
stages. ^^
what matters
is
probably be more, not
less
transactions
.
of the current concern with
to
as
their
many
limit
too-
or
too-big-to-fail
linkages. This
less
complex and more transparent.
that
for
complex, since
are
macroeconomic and
I
think this
it
financial stability,
is
the risks
some
within
internalized
private-sector party that has
embedded
player would do so
if
all
in
and through
larger
some information and
these internal linkages.
linkages
become
would
would take many linkages to perform some
institutions.
Moreover, despite the many internal control problems that large institutions
least there
It
is
decentralized.
exhibit, at
incentives to
unlikely that an equivalent
In
summary, detecting the
gradual buildup of hidden-CAD probably would be harder rather than simpler
financial
network with no giant components and many more small nodes.
Finally, as a positive
powerful treatment
treatment
in.
Monetary
is
in
The
these linkages instead were implemented by thousands
would be much
whole network, which
the
much
crisis.
composition. Indeed, each individual small bank would be simpler, but the
a fallacy of
of
stillin its early
is
enormous complexity caused by
view presumes, however, that
is
CAD
cautionary note about hidden-financial-CAD,
large
decade before the
for almost a
observation, since panic
is
preventative
in
is
itself.
,
in
a
.
at the core of SFA, the very existence of a
1
turn to
a
detailed discussion of this
the next section.
Financial Defibrillators
policy
is
the
first line
systemic and quick, and
" See Segoviano and Goodhart
it is
of defense
when
distress arises in the financial system.
It
an effective instrument for macroeconomic shocks that do
(2009) for another interesting attempt to measure banking interconnectedness and
stability.
26
not trigger significant panics. The next step
financial
defibrillators:
the
mechanism whose primary
Bagehot advocated
in
The LOLR provides
a
role
1873 that
a
resort
last
facility
(LOLR).
to prevent depositors' runs
LOLR should lend
is
in a crisis
one of the oldest
It
insurance
an
is
from commercial banks.
to "solvent but illiquid"
collateral.
channel and infrastructure that can be used even for situations that
do not exactly match
its
original intent.
directly to primary dealers
Facility
is
policy escalation
'^^
adequate
institutions with
lender of
in
In
the current
crisis,
the Fed extended credit
and investment banks through the Primary Dealer Credit
(PDCF) and the Term Securities Lending
Facility (TSLF).
The former
lends
facility
against collateral, using haircuts to assure the safety of the loan and applying an interest
rate
above the rate prevailing under normal market conditions.
succeeded as
mechanism
a built-in exit
for the
program. The
This penalty rate has
latter facility
is
the primary-
dealer analog to the Term Auction Facility (TAF) that serves commercial banks.
TSLF,
the
Fed
Under
determines an amount of funding to be provided and an auction
determines the recipients and price of the funding. The Fed also established programs to
serve as a lender of
programs were
largely
last
resort, albeit indirectly, for
money-market funds, but these
superseded by the Treasury's guarantee program.
,
;.
l
Aside from the specific value of these extensions of the LOLR, the general principle that
wish to highlight
financial
is
that having readily available channels to different segments of the
system can save precious time
arguments supporting the creation
The solvency requirement
is
nice
in
v^/hich
the midst of a
crisis,
and
is
one
of the central
of a broader class of financial defibrillators.
illiquid
turns the Bagehot principle
another source of uncertainty.
in
theory but a real practical problem during an SFA episode.
the distinction between a solvent and an
uninformative,
I
institution
(i.e.,
how
is
highly arbitrary as
most asset
In
such instance
prices
become
regulators determine which firms are insolvent)
in
yet
-^"
27
Unfortunately,
as
starkly
by
illustrated
the
interpretation and use of the conventional LOLR,
recent
crisis,
even
with
when
can be insufficient
it
extends beyond commercial banks and to assets and
liabilities
flexible
a
the panic
other than deposits.
In
on banks balance
particular, systemic panic temporarily destroys the value of assets
sheets and, more broadly, the collateral of most financial institutions. At distorted asset
and margin
prices, capital-adequacy ratio constraints
sales)
and feed back
system
add
into panic itself.
Absent
a clear
calls trigger costly
framework
to the
-
crisis.
been several recent proposals
to provide
some form
Essentially, there are three
I
of the proposals
review some of these
in
is
a
the following sub-sections.
broad categories of proposals to reduce
crisis risk:
•
Pre-paid/arranged contingent capital injections,
•
Pre-paid/arranged contingent asset and capital insurance injections.
is
have
key distinction
Self-insurance through higher (and cyclical) capital-adequacy ratios,
This
the
do not distinguish
•
III. A.
in
traditional LOLR, there
systemic event and an individual bank problem, which
).
further
of protection to financial institutions
and markets during systemic events (although many
from an SFA perspective
ilk
.
To address these extreme scenarios, not covered by the
a
all
and discourage any private sector participation
uncertainty-fuel,
solution to the
(e.g., fire
to support the financial
such scenarios, speculations and misguided policy proposals of
in
between
actions
Self-insurance
where policymakers'
increase their war chests.
It is
instinct
lies.
It
essentially consists of requiring banks to
the analogue of requiring that every individual carries an
ICD (implantable cardioverter defibrillator) to prevent SCA.
The Group of Central Bank Governors and Heads of Supervision, the oversight body of
the Basel Committee on Banking Supervision, has agreed that banks should hold more
28
This Basel group also favors a shift
Caruana 2009).
capita! (BIS 2009,
The new/
"quality" capital; for example, favoring equity over preferred shares.
requirements
metric
is
liquidity
capital
be set with respect to banks' leverage ratios; the coarseness of this
v^/ould
here seen as
The group also favors countercyclical
a virtue.
The
standards for funding operations.
recommendations, highlighting an interest
conservative capital, liquidity and
whose combination
financial stability
toward higher
to
it
fail
(U.S.
in
more
financial
and interconnectedness could pose
and
similar
the U.S., stricter and
management standards on any
risk
of size, leverage,
were
Treasury has offered
U.S.
imposing,
in
capital buffers
firm
threat to
a
Treasury 2009).
This category of proposals has the virtue of being the simplest to implement, but has the
drawback
of being the costliest.
The reason
for the latter
protection against extreme macroeconomic events,
which vastly exceeds what they need
for their
management, operations.
..
Adding
a
-
.
assets
fragile
improvements
to the
most
relative
mechanism.
•;,..../-
-m
risk
.•
basic self-insurance systems.
how much
fragile
However,
at the
sound
are
assets,
end of the day,
insurance can be mandated through
_
.
,
.
,
.
Continaeut cuDftaJ injections
Ili.B.
is
order to increase
normal, mostly microeconomic
idiosyncratically
to
the costly nature of self-insurance caps
This
in
forces banks to freeze capital
.
,
that
procyclical capital clause, as well as increasing the capital-requirements for
systemically
this
,
it
is
where most academics'
associated with holding capital
instinct
when
approaches recognize that access to
advance, since
kind differ
in
it
is
it
lies.
is
The basic idea
is
to reduce the costs
not needed. However, and centrally, these
capital during crises
needs to be arranged
often hard to raise capital during a severe
the source of this contingent capital,
sector and the government. Within the former,
in
in
crisis.
particular
some proposals
in
Proposals of this
between the
private
the contingent funds
29
come
through contingent debt/equity swaps) while
However, outsiders' commitments problems
can serve as the source of
Holmstrom and
made one
in
others the funds
limit
come from
(e.g.,
outsiders.
the extent to which the private sector
during extreme events, a point highlighted by
this capital
and AIG (and the monolines)
Tirole (1998) in theory
Fiannery's (2002) proposal
his
"new money"
primarily from existing stal<etiolders and hence require no
in
practice.
of the first significant steps in this direction with
proposal for "reverse convertible debentures." Such debentures would convert to
equity whenever the market value of a firm's equity
One problem
addressed
in
article
that
The Kashyap
idiosyncratic shocks.
calls for
is
made no
it
a
et
al
negative systemic shock.
and place the amount insured
(2008) proposal deals with this distinction and
Private investors
into a "lock
box" invested
are themselves subject to capital requirements
insurance. The insurance would be triggered
of quarters
not be included
in
a certain threshold.
between aggregate and
distinction
banks to buy capital insurance policies that pay off
experiences
number
below
from the point of view of the systemic problems
of this early proposal
this
falls
when
the banking sector
would underwrite the
in
Treasuries.
would not be allowed
when aggregate bank
exceed some significant amount; losses
determining whether the insurance
is
at the
policies
Investors
who
to supply this
losses over a certain
covered bank would
triggered.
Combining both contributions, the Squam Lake Working Group on Financial Regulation
has
a
proposal (2009) similar to Fiannery's except that conversion from debt to equity
is
triggered only during systemic events and only for banks that violate certain capital-
adequacy covenants.
Yet another variant on capital insurance
regulator, instead of the firm.
of insurance required
by the bank,
in
Under
is
for the insurance policy to pay out to the
this proposal,
by Acharya and others, the
would be proportional to an estimate of the systemic
amount
risk
posed
order to discourage firms ex-ante from taking on excessive systemic
risk.
30
Hart and Zingales (2009) advocate an alternative approach;
when spreads on
CDS
bank
rise
above
issue equity
in
a
certain threshold, a regulator allows the
order to bring the CDS spread back below the threshold.
unable to reduce
its
be
it
If
of time to
the bank
is
the regulator determines the bank's debt
is
the bank by lending to the bank; otherwise, the
a trustee,
Although
proceeds to the bondholders.
component,
in
CEO with
regulator replaces the
injection
at risk.
is
the regulator invests
risk,
If
bank's
CDS spread, the regulator reviews the bank's books and determines
whether or not the bank's debt
not at
window
a
a
this
also relies heavily
who
will
bank and pass the
liquidate the
approach does have
a
contingent capital-
on the resolution of financial firms, which can
device during normal times but can be highly counterproductive
a useful discipline
during an SFA episode (especially because CDS prices are severely affected by the panic
itself).
More
generally, the contingent capital
mostly one of fundamentals.
However,
feature of an SFA episode, then
is
it
approach
if
the right one
is
the panic
component
All
that
is
needed
worsen. Despite
derives
its
enormous
is
its
a
distortion
tlie
very
same feature
perceived
in
tliat
probabilities
public insurance
be as effective as capital injections
in
(when
assessed
at
may
:
is
well trigger
-
'
'-
will
be available should conditions
high notional value, the expected cost of this policy
power from
crisis
a costly capital injection to subside.
broad guarantee that resources
economic agents greatly overvalue
cost
it
rises.
Contingent insurance injections
The pure panic component of SFA does not require
the
significant, a central
not the most cost-effective, and
further panic as fear of dilution and forced conversion
la.C.
is
when
of
is
low because
underlies the panic. That
a
catastrophe also
is,
it
the
means that
and guarantees. Providing these can
dealing with the panic at a fraction of the expected
reasonable
rather
than
panic-driven
probabilities
of
a
catastrophe).
31
In
Caballero and Krishnamurthy (2007)
government or
uncertainty, a
we showed
and investor fears to be
in a
that during
is
a
if it
between the true average and the average
why
crisis,
argument
for
panics
developed
is
it
may
there were
panic of this kind, each individual bank
this
In
normal times,
these are "multiplied"
government gets the private
guarantee since
also closes the
it
asset-insurance injection proposals. "^^
gap
The
be optimal to support assets rather than inject capital during
Krishnamurthy (2008b).
must be
institutions face a constraint such that value-at-risk
equity.
want to
of panic-driven expectations.
many
Caballero and
in
will
worse than the average, an event that cannot be
situation
more than one-for-one with
During the recent
an episode of Knightian
has no informational advantage over
true for the collective. By providing a broad guarantee, the
sector to react
in
bank concerned with the aggregate
central
provide insurance against extreme events even
the private sector. The reason
that
this structure
many times
in
less
In
practice,
financial
than some multiple of
speaks to the power of equity injections, since
relaxing the value-at-risk constraint.
In
contrast,
insuring assets reduces value-at-risk by reducing risk directly, which typically does not
involve a multiplier.
assets, such as
However, when uncertainty
CDOs and CDO-squared, can
is
rampant, some
illiquid
reverse this calculation.
and complex
In
such cases,
insuring the uncertainty-creating assets reduces risk by multiples, and frees capital,
more
effectively than directly injecting equity capital.
Moreover,
it
turns out that the
recapitalize banks
when
that's the
the government can pledge
capital
same
(Caballero 2009a).
chosen solution. Rather than
minimum
a
This
principle of insurance-injection can be used to
directly injecting capital,
future price guarantee for newly privately raised
mechanism
is
very powerful both because private
investors overvalue the guarantee, and because the recapitalization
catastrophic event less
likely.
Caballero and Kurlat (2009b) quantified this
and showed that once the equilibrium response of equity prices
See,
e.g.,
itself
is
makes the
mechanism
taken into account,
Caballero (2009a, b); Mehrling and Milne (2008) and Milne (2009) for real time proposals.
32
this
mechanism
significantly
reduces the effective exJDOsure of government resources
relative to a public equity injection.
Many
of the actual
programs implemented during the
crisis
had elements of guarantees
rather than being pure capital injections. Perhaps the clearest case of this approach
that follovi/ed by the UK. Their asset protection scheme, announced
provided insurance against 90 percent of losses above
insurance
is
credit assets such as
provided
exchange
in
RBS and Lloyds Banking Group,
respectively.
The main
in
January 2009,
"first-loss"
threshold on
commercial and residential property loans,
portfolios of corporate and leveraged loans,
and structured
a
RMBS, CMBS, CLO, and CDO
The APS covered £552
for a fee.
amount
w/ith a first-loss
criticism to the U.K.'s
approach
billion portfolios
of £19.5 billion
is
In
and £25
guarantee access to insurance
•flexible
manner
we proposed
in
a policy
billion,
that they charged such a high
economy
framework which would not only
the event of an SFA episode, but
government
that integrates the role of the
of
"
to their failure than socially optimal.''^
Caballero and Kurlat (2009a)
The
obligations.
fee for the insurance that most banks chose not to engage, leaving the overall
more exposed
is
it
would do so
in a
as an insurer of last resort
with private sector information on the optimal allocation of contingent insurance.
Under our proposal, the government would
would be purchased by
financial
institutions,
holding requirements. During a systemic
attach a government guarantee to
some
would be allowed
TICs,
to hold
and corporations as
well.
available substitute for
16
This
is
a
reminder that
(Caballero 2009c). Also,
hence be
less affected
this reason, a
accept
it
and use
In
many
issue tradable insurance credits (TICs)
of
its
entitle
its
holder to
and possibly hedge funds, private equity funds,
principle, TICs could
be used as
a
flexible
were created during the
of the facilities that
important to note that each bank
reject an
which would have minimum
assets. All regulated financial institutions
a taxpayers' "deal mentality" during
is
of
each TIC would
crisis,
by Knightian uncertainty with respect to
bank may
some
which
crisis.
an SFA episode can be highly counterproductive
is
its
expensive insurance deal even
and readily
more
own
when
likely to
know
its
own
financial health
and
financial health than are their creditors. For
outsiders faced with the
same option would
it.
33
There are
•
five
features of the TIC-framework which are worth highlighting at this point;
automatic. While the precise threshold that defines the
It is
practice
crisis
is
fuzzy, as in
that which determines the trigger for the Fed's unusual and exigent
is
circumstances clause, once triggered the actions are well defined, broad, and
easy to communicate to the public and financial institutions
•
it
advance/^
addresses directiy the muitipie-effects of panic on asset values and financial
balance sheets. While the conventional LLOR
ability to
generate
liabilities,
the type of assets
extremely useful but
on
not
accepts
it
simultaneously.
crises
in
its
crisis
a
TIC
it
window
was
operations. This
would
have
addressed
both
.
crisis.
TICs are equivalent to
but not during normal times. That
become
at protecting the banks'
the Fed relaxed the constraints on
discount
Instead
contingent on a systemic
It is
aimed
only addressed the effect of uncertainty on funding but
it
constraints.
capital
is
the TIC framework protects assets and through
example, during the recent
capital. For
•
in
activated only
needs protection,
when
this
a
systemic
contingent
is,
TICs are contingent-CDS.
crisis arises.
feature
CDS during systemic
They
By targeting the event that
significantly
lowers
the
cost
of
insurance for financial institutions.
•
It
Which asset
flexible.
is
classes TICs are attached to
largest panic discounts take place
concerns that banks
will
on
a specific crisis. This
If
there
is
a
concern
tiiat
in
FDICIA
FDIC has to sign
off).
(i.e,
I
activation
the President has to sign
owe
this
may
raise
knowledge of
adverse selection
TICs will be activated too frequently without sufficient evidence of systemic
one could contemplate making
exemption
feature
select their assets based on insider's
expected losses rather than on panic discounts, but
risk,
depend on where the
this
subject to several "keys," as with the systemic risk
off,
2/3 of the Fed board has to sign
off,
and 2/3 of the
example to Morris Goldstein.
34
problem can be limited with methods
Management System
(CMS)
for
similar to those used by the Collateral
discount
supplemented
borrowing,
with
Representations and Warranties clauses.
•
markets to reallocate
TICs are tradable. This feature allows private agents to use
the access to insurance toward financial institutions
most
in
dire need.
And
if
distressed institutions chose to not seek to stock up on TICs and risk their
survival for a higher return (as
Lehman probably
the rest of the financial system would be
that could arise
if
did
and
failed), at
the very least
better protected against the turmoil
the misbehaving institutions
fail
as they
would be holding the
TICs.
Of course some of these basic properties can be modified depending on the specific
circumstances and complementary measures available. Tradability clearly has pros and
cons, and
it
could be done without
it.
Also, while the basic
attaching TICs to assets, variants could include attaching
equity,
them
depending on the particular needs of the distressed
and they could
is
deemed
be controlled by narrowing the
to be
flexibility
overwhelming
in a
consist of
to liabilities and
institutions
also operate as collateral-enhancers for discount
the adverse selection effect
iV.
mechanism would
even
and markets,
window borrowing.
specific context, this
on the assets than can be protected by
If
can
TICs."^^
The International Dimension
The international dimension
of
emerging markets which has
a close parallel
SFA adds
its
own
ingredients.
I
focus on the problem for
with the issues faced by the financial sector
within developed economies.
In
Caballero and Kurlat (2009a)
recent
crisis
and provided an
we argued
illustrative
that TICs would iiave been an effective policy tool to address the
example of how these could have been used.
35
Financial crises in
caused, by
This
emerging markets are greatly amplified, indeed
sudden stop and reversal
a
phenomenon enormously complicates
liquidity,
which means that the
of financial-defibrillator
Most policymakers
which
is
one
when coping
policy
we
"sudden stops").
problem, since rather than
becomes entangled
often
instances are
in
the
one of international rather than
government may not have access
(just)
In
crisis.
problem
refer to this as the dual-liquidity
domestic
to the right kind
with large capital outflows."'^
emerging market economies are acutely aware of
in
of the
local
is
the
itself
Caballero and Krishnamurthy (2001, 2006),
emerging markets; The shortage
many
of capital inflows (the so called
providing the solution, the government
of
in
main reasons they accumulate
large
amounts
this
danger,
of international
Large accumulations of reserves are the equivalent to self-insurance for
reserves.
domestic banks, and as such are costly insurance
facilities
(although less so
are the side effect of exchange rate intervention policies). For this reason,
when they
many
of us
have advocated the use of different external insurance arrangements, and the IMF has
spent
a significant
amount
of time attempting to design the right contingent credit line
facility.
19
While European banks also experienced a currency mismatch problem at some stage during the recent
this
was expediently solved by
liquidity
swap
lines
a
swap mechanism between the Fed and the ECB. The Fed
with the central banks of Australia,
Korea, Mexico,
New
National Bank
were established on December
Brazil,
Canada, Denmark, the United Kingdom (BOE), Japan,
Zealand, Norway, Singapore, Sweden, and Switzerland. The swap lines with the ECB and Swiss
12,
2007, and the others were established subsequently.
addition, the Fed established foreign-currency liquidity
swap
lines with
National Bank, should the Fed need to borrow foreign currency
these economies
is
the
political
concerns that arise from
significant cross-border interconnectedness
among
problem which further delays an already slow
in
utilizing
the future. A far more problematic issue for
domestic financial-defibrillators
their financial systems.
policy response during
most other macroeconomic
In
the BOE, the ECB, the BOJ, and the Swiss
In this
context there
when
is
there
is
a free-rider
SFA episodes. This coordination problem can
be reduced through different international conferences such as the G7 or others, which
is
the mechanism used for
policy coordination issues. This process can be further facilitated by having cross-
pledges proportional to these cross-border exposures. That
were these to be
crisis,
also established dollar
is,
the U.K. would stand behind a share of the U.S. TICs,
activated, and vice-versa.
36
Since 1999 three different contingent lending
facilities
IMF: the Contingent Credit Lines (CCL), the Short
Flexible Credit Lines (FCL).
requirements ex-ante
allovued to expire
FCL
in
in
in
They
share the
all
Term
same
have been introduced by the
Liquidity Facility (SLF),
basic premise; increase eligibility
order to provide reduced conditionality ex-post. The CCL
2003 and the
SLF, introduced in
March 2009. Both the CCL and the SLF
failed to attract a single
and
the
commitment
were
fee
was further reduced.
conditionality
simplified to
were
also
borrower request.
to join these facilities resulted in
consecutive re-designs that provided more appealing terms.
CCL's conditions for activation
v^/as
October 2008, was replaced by the
The reluctance of the targeted emerging economies
charge
and the
November 2000, the
In
enhance automaticity; the
reduced.
Under
The FCL increased the
the
SLF
of the
size
rate of
ex-post
committed
resources to 10 times the member's quota (compared to 3-5 under the CCL and SLF) and
extended the repayment period to 3.5 to
It is
not clear whether
of the global crisis
Mexico ($47
enrolled
in
it
was the redesign
when they
billion),
5 years
is
3
months under the
or the fact that countries
signed on, but by
Poland ($21
the FCL. The FCL
(compared to
billion),
now
several of
and Colombia ($11
the right kind of
facility for
capital injection but a contingent insurance injection
(it
were
the middle
them have
billion),
SFA, since
in
it is
SLF).
joined:
are currently
not a contingent
gives the right to a credit line
during crises).
The IMF's FCL
not
all
is
the equivalent to a LOLR for sovereigns. As such,
it
addresses some but
the effects of a widespread panic on the country's ability to tap global capital
markets during
a global
SFA episode. There
At the beginning of the millennium
1
is
scope for complementary
was asked
facilities.
to write an article on "The Future of the
IMF" for an AEA session (Caballero 2003). Uncomfortable with the broad mandate,
I
began by changing the meaning of the IMF acronym to that of "International Markets
Facilitator." Basically, the idea
was
to have a division of the (original)
IMF play the
role
37
of a (well capitalized) AIG or monoline
would
debt. Countries
main
their
real
and
insuring
CDOs
of contingent emerging
would
issue debt with contingencies that
financial external into a
insulate
CDO, and the IMF would add insurance to
market the soundness of the underlying assets and
effective cost for
market
them from
The IMF's wrapping would serve both the
the senior tranches of these CDOs.
signaling to the
in
it
role of
would lower the
emerging markets to buy external insurance. "° This would be
a sort
of
PPIP for sovereigns, which would leverage IMF's anti-crisis resources.
The sudden stop phenomenon
mechanism
that can
is
just
one more manifestation
of the flight of quality
be triggered by conventional financial multipliers and greatly
exacerbated by SFA. As such,
framework but customized
would make sense to develop
it
a
system akin to the TIC
on the value of emerging
to contain the effect of panic
market new and legacy emerging debt held by investors during global SFA episodes.
There
is
extensive evidence that severe non-idiosyncratic sudden stops are caused by
runs against emerging markets as an asset class (see,
insuring these assets during a
limit
crisis,
they would
TICs)
would
typically be activated at the
when both
same time
as
decouple. For this reason they
be other instances
should be controlled
by a global institution such as the
IMF rather than by the
of the U.S. or other developed economies.
Conclusions
To recap, the absence of systemic insurance
the
2008). By
part of "quality" and hence
TICs, but there could
governments
'"
become
al.
the run against them.
These E-TICs (emerging markets
V.
Broner et
e.g.,
boom and
is
a
very
weak
an economic tragedy during a financial
incentive
crisis.
mechanism during
A pragmatic
anti-crisis
See Cabaliero and Panageas (2007, 2008) for models and calibrations of the hedging gains for emerging markets
subject to sudden stops
in
net capital inflows and terms of trade shocks.
38
system
needs
to
complement
guidelines) during the
boom
regulatory
(healthy
constraints
financial
lifestyle
with an institutionalized systemic insurance arrangement for
crises (a readily available financial defibrillator).
The standard LOLR should be complemented with three kinds of
In
anti-crisis
mechanisms:
•
Self-insurance through higher (and cyclical) capital-adequacy ratios,
•
Pre-paid/arranged contingent capital injections, and
•
Pre-paid/arranged contingent asset and capital insurance injections.
practice, there are
For small shocks,
always
a
it is
good reasons
to have
some
of each of these types of
mechanisms.
probably easiest to have banks self-insure. For large shocks, there
fundamental component, which
is
is
probably best addressed immediately with
contingent capital. However, the large panic component of an SFA episode requires large
amounts
of guarantees,
which would be highly impractical and costly to achieve through
capital injections. For this
component,
''
response.
The same
contingent-insurance policy
"i-
.>
•
/
':
is
The IMF has an important
economies.
itself
role to play
is
at risk of
becoming embroiled
in facilitating
the appropriate
;;-..,-'?-'
principles carry over to the international arena, especially for
economies, where the sovereign
for these
a
emerging market
in
an SFA event.
contingent insurance arrangements
,
39
Vi.
References
Abreu,
Dilip,
Vol. 71, No.
.
and Markus
1,
K.
Brunnermeier. 2003. "Bubbles and Crashes," Econometrica,
pp. 173-204. January.
Markus
Adrian, Tobias and
Adrian, Tobias and
K.
'
V.,
'
,
Brunnermeier. 2009. "CoVaR," mimeo., August.
Hyun Song
Shin. 2009.
"Money,
American Economic Review Papers and Proceedings,
Acharya, Viral
,
Lasse H. Pedersen,
Thomas
Liquidity,
and Monetary
Policy."
Vol. 99, No. 2..
Philippon and
Matthew Richardson. 2009a.
"Regulating Systemic Risk: Academic Paper," mimeo, April.
Acharya, Viral
V.,
Lasse H. Pedersen,
Thomas
Philippon and
Matthew Richardson. 2009b.
"Regulating Systemic Risk," presentation, August.
Baker,
Dean and
Travis McArthur. 2009. "The Value of the 'Too Big to
Fail'
Big
Bank
Subsidy." Center for Economic Policy and Research Issue Brief. September.
Bank for International Settlements. 2009. "Comprehensive Response to the Global
Banking
Crisis."
Press release,
Bernanke, Ben and
Fluctuations," The
Mark
September
Gertler,
7.
1989.
"Agency Costs, Net Worth and Business
American Economic Review,
Vol. 79, No. 1, pp. 14-31,
March.
40
Bernanke, Ben, Mark Gertler and Simon
Gilchrist, 1999.
Quantitative Business Cycle Framework."
Handbook of Macroeconomics,
Bernheim,
and
B.
In
"The Financial Accelerator
in a
John Taylor and Michael Woodford, eds.,
pp. 1341-1393.
Douglas and Rangel, Antonio. 2005. "Behavioral Public Economics: Welfare
Non-Standard Decision-Makers," NBER Working Paper No.
Policy Analysis with
W11518.
Berg,
Andrew, Eduardo Borensztein, and Catherine
Warning Systems: How Have They Worked
CDS
Boberski, David. 2009.
Borio, Claudio
in
2004. "Assessing Early
Pattillo.
Practice?"
IIVIF
Working Paper 04/52.
Delivery Options. Bloomberg Press, NY.
and Mathias Drehmann. 2009. "Assessing the
risk of
banking crises -
revisted," 6/5 Quarterly Review, pp. 29-46, March.
Borensztein, Eduardo and Mauro, Paolo, 2004. "The Case for GDP-lndexed Bonds."
Economic
Policy, Vol. 19, No. 38, pp.
165-215,
'-•-April.
:
Brunnermeier, Markus
Liquidity,"
K.
K.
'
'
-
-
Vol. 22, No. 6, pp.
2201-2238.
and Lasse Heje Pedersen. 2009. "Deciphering the
Credit Crunch 2007-08," Journo/ 0/ f conom/c Perspecf/Ves, Vol. 23, No.
Caballero, Ricardo
Crisis ."
J.,
,
and Lasse Heje Pedersen. 2009. "Market Liquidity and Funding
Reweiv o/F/nanc/o/Sttvd/es,
Brunnermeier, Markus
;-
"Why Do Emerging
Broner, Fernando, Guido Lorenzoni, and Sergio Schmukier. 2008.
Economies Borrow Short Term?" MIT mimeo, October.
T
2009a.
"
Dow
Boost and
a (Nearly) Private
1,
Liquidity
and
77-100.
Sector Solution to the
VOX, February 22.
Caballero, Ricardo
J.,
2009b.
"
A Global Perspective on the Great Financial Insurance
Run: Causes, Consequences, and Solutions ." MIT mimeo, January 20.
Caballero, Ricardo
J.,
2009c.
"
A Deal Mentality
is
Bad Macroeconomics ." VOX, January.
41
Caballero, Ricardo
and Proceeding,
The Future of the IMF ." American Economic Review, Papers
May.
Emmanuel
J.,
Commodity
/Act/V/fy, Fall,
"
93(2), 31-38,
Cabaliero, Ricardo
Crash,
2003.
J.,
Prices,
,,
Farhi
.
,
,
,
,.
„
.
and Pierre-Olivier Gourinchas, 2008a.
,^
"
.,
Financial
and Global Imbalances ." Brookings Papers on Economic
pp 1-55.
Emmanuel
and
Cabaliero,
Ricardo
Equilibrium
Model of "Global Imbalances" and
J.,
Farhi
Pierre-Olivier
Gourinchas,
Low/ interest Rates ."
"
2008b.
An
American Economic
Reiz/ew, 98:1, pgs 358-393.
Cabaliero, Ricardo
Fragility ,"
and Arvind Krishnamurthy, 2009.
J.
"
Global Imbalances and Financial
American Economic Review, May.
Cabaliero, Ricardo
and Arvind Krishnamurthy, 2008a.
J.
"
Collective Risk
Episode ." Journa! of Finance, Vol. 63, Issue
a Flight to Quality
Cabaliero, Ricardo
J.
5,
Volatility:
Ricardo
J.
"
and Arvind Krishnamurthy, 2006.
in
October.
and Arvind Krishnamurthy, 2008b. "Knightian uncertainty and
implications for the TARP," Financial Times Economists' Forum,
Cabaliero,
Management
November
its
24.
Bubbles and Capital
Flov^/
Causes and Risk Management ." Journal Monetary of Economics, 53(1), pp. 35-
53, January.
Cabaliero,
Ricardo
J.
and Arvind Krishnamurthy, 2001.
Collateral Constraints in a
Model
of
"
International and
Domestic
Emerging Market Crises ." Journal of Monetary
Economics 4S(3), 513-548, December
Cabaliero, Ricardo
Financial Crises:
Symposium on
J.
and Pablo
A Macroeconomic
Financial Stability
Cabaliero, Ricardo
J.
and Pablo
"
Kurlat, 2009.
Policy
The "Surprising" Origin and Nature of
Proposal ." Prepared for the Jackson Hole
and Macroeconomic
Kurlat, 2009.
"
Policy,
August.
Public-Private Partnerships for Liquidity
Provison ." MIT mimeo, March.
42
Caballero, Ricardo
Remarks and
Caballero,
J.
and Pablo
a Literature
Ricardo
J.
"
Kurlat, 2008.
and Bailouts: Policy
Flight to Quality
Review ." MIT mimeo, October.
and
Panageas,
Stavros
"
2008.
Sudden Stops and
fHedRJnR
Precautionary Contractions ." Journal of Development Economics, 85 pp 28-57.
Caballero,
Ricardo
J.
and Stavros Panageas, 2007.
Sudden Stops and External
Caballero, Ricardo
J.
'"
Caballero, Ricardo
July.
Caplin, Andrew/
after the Fact,"
J.
'
'
A Global Equilibrium Model of
Management." MIT mimeo, September.
"
and Alp Simsek, 2009a.
mimeo, June.
mimeo,
Liquidity
"
Complexity and Financial Panics ." MIT
'
^
,.
and Alp Simsek, 2009b.
-
..
"Fire Sales in a
'-'-•.'-: "--
,
•
Model
r'
of Complexity ."
••-„.••
•
MIT
.•;
and John Leahy, 2004. "Business as Usual, Market Crashes and Wisdom
Amer/con fconom/c
Caprio, Gerard and Patrick
Rewei/v,
pp. 548-565.
Honohan. 2008. "Banking
Crises."
IIIS
Discussion Paper No.
242.
Caruana, Jaime. 2009. "Deposit insurance core principles." Keynote address at the "Core
Principles for Effective Deposit Insurance Systems" conference,
Coval, Joshua
D.,
Jakub W. Jurek and
Erik
Stafford.
September 23-24.
2008. "Economic Catastrophe
Bonds." HBS Finance Working Paper No. 07-102.
Coval, Joshua, and
Erik
Stafford.
<
2007. "Asset Fire Sales (and Purchases)
Markets," Journal of Financial Economics, Vol. 86, No.
2,
in
in
Mario
Emerging Markets. Cambridge: M\J
I.
Blejer
Equity
November.
Eichengreen, Barry and Carlos Arteta. 2002. "'Banking Crises
Presumptions and Evidence",
in
and Marko Skreb,
in
Emerging Markets:
eds., Financial Policies
Press.
43
No Gain?
Flannery, Mark, 2002, "No Pain,
Market
Effecting
Discipline via
'Reverse
Convertible Debentures'," Working Paper, November.
Geanakoplos, John and Heraklis M. Polemarchakis. 1986. "Existence, Regularity, and
Constrained
Suboptimality of Competitive
Incomplete,"
in
W.
Heller, R. Starr
Communication: Essays
in l-lonor
of
and
K.
J.
When
Allocations
the
Market
Asset
D. Starret (eds.) Uncertainty, Information,
Arrow, Vol.
Ill,
Cambridge University
and
Press.
Geanakoplos, John. 2003. "Liquidity, Default, and Crashes: Endogenous Contracts
General
Equilibrium."
Applications,
Eighth
In
Advances
World
in
Economics and Econometrics: Theory and
in
Conference,
Monographs.
Is
Vol.
2,
170-205.
^
Econometric
Society
.
Geanakoplos, John. 2009. "The Leverage Cycle." mimeo.
July.
Geanakoplos, John and Ana Fostel. 2008. "Leverage Cycles and the Anxious Economy,"
American Economic Review,
Gilovich,
Thomas, Dale W.
Vol. 98, No. 4, pp.
Griffin
Biases: Ttie Psycliology of Intuitive
Geithner, Timothy,
Washington
Post,
1211-1244.
and Daniel Kahneman, eds. 2002. Heuristics and
Judgment.
Nevi/ York:
Cambridge University
and Lawrence Summers. 2009. "A
New
Financial
Press.
Foundation."
June 15
Gersbach, Hans. 2008. "Banking with Contingent Contracts, Macroeconomic
Risks,
and
Banking Crises," CER-ETH Working Paper 08/93.
Gersbach, Hans. 2009. "Private Insurance against Systemic Crises," CEPR Discussion
Paper No. 7342.
Gorton, Gary and
Andrew
Metrick. 2009, "The Run on
Repo and the Panic
of 2007-
2008," Yale Worl<ing Paper.
44
Gorton, Gary and Andrew Metrick. 2009. "Haircuts," A/Bf/? Working Paper No. 15273,
August.
Gorton, Gary and George Pennacchi. 1990. "Financial Intermediation and Liquidity
Creation." The Journal of Finance, Vol. 45, No. 1 (Mar.), pp. 49-71.
Haldane, Andrew G. 2009. "Rethinking the Financial Network." Speech delivered at te
Financial Student Association,
Hart, Oliver
and
Institutions,"
Amsterdam.
April.
2009. "A
New
Luigi Zingales.
mimeo, Chicago,
Capital Regulation for Large Financial
April.
Holmstrom, Bengt. 2008. "Comment on: 'The Panic of 2007,' by Gary Gorton,"
Maintaining Stability
in
a Changing Financial System,
Federal Reserve of Kansas City at Jackson Hole,
Investment Company
International
Money Market Working
Monetary Fund. 2009. Global Financial
Financial Challenges Ahead, October.
Symposium sponsored by The
Wyoming.
2009. Report of the
Institute.
in
Group.
Stability Report: Navigating the
Preliminary version
downloaded October
15,
"
2009.
Kashyap, Anil
Regulation,"
".
,;:
.
K.,
in
Raghuram
.
G.
.
Rajan and Jeremy
Maintaining Stability
in
;.
/
Stein. 2008. "Rethinking Capital
Changing Financial System, Symposium
a
sponsored by The Federal Reserve of Kansas
C.
:
City at Jackson Hole,
Wyoming.
Kaminsky, Graciela, Saul Lizondo and Carmen M. Reinhart. 1998. "Leading Indicators of
Currency Crises." Staff Papers
1-48.
-
International
Monetary Fund,
Vol. 45, No. 1 (Mar.), pp.
-
Kaminsky, Graciela and Carmen M. Reinhart. 1999. "The Twin Crises: The Causes of
Banking and Balance-Of-Payments Problems." The American Economic Review, Vol. 89,
No. 3 (Jun.), pp. 473-500.
45
Nobuhiro and John Moore. 1997. "Credit Cycles," Journal of
Kiyotaki,
Political
Economy,
Vol. 105, No. 2, pp. 211-248.
Kotlikoff,
Laurence
J.
and John
C.
Goodman. 2009.
"Solving Our Nation's Financial Crisis
with Limited Purpose Banking." Boston University mimeo?.
Lehman
Brothers. 2008. "Residential Credit Losses
Mendoza, Enrique
G.
-
Going into Extra Innings?"
and Marco Terrones. 2008. "An Anatomy of Credit Booms:
Evidence from Macro Aggregates and Micro Data,"
NBER Working Paper 14049.
Mehrling, Perry and Alistair Milne, 20008. "Government's role as credit insurer of
resort and hov^
it
can be
fulfilled,"
mimeo
October.
House of Credit. Cambridge University
Milne, Alistair. 2009. The Fall of the
last
Press,
Cambridge and New York.
Mitchell, Mark, Lasse Heje Pedersen
American Economic Review,
and Todd Pulvino. 2007, "Slow Moving Capital,"
Vol. 97, No. 2, pp. 215-220,
May.
Office of the Special Inspector General for the Troubled Asset Relief Program. 2009.
Quarterly Report to Congress, iuly.
Pulvino,
Todd
Commercial
Reinhart,
View
1998.
"Do Asset
Fire
Sales
Exist?
An
Empirical
Investigation
of
Aircraft Transactions," iourno/ of Finance, Vol. 53, No. 3, pp. 939-978, June.
Carmen and Vincent
of the Past
Reinhart,
C.
Reinhart. 2008. "Capital Flow Bonanzas;
An Encompassing
and Present." NBER Working Paper No. 14321.
Carmen M. and Kenneth
Rogoff. 2008. "Banking Crises:
An Equal Opportunity
Menace." NBER Working Paper No. W14587.
Segoviano, Miguel and Charles Goodhart. 2009. "Banking Stability Measures." IMF
working paper, January.
46
Squam
Lake Working Group on Financial Regulation, 2009.
iViechanism
Distressed
for
Firms:
Financial
"
An Expedited Resolution
Regulatory Hybrid Securities ," Working
Paper, April.
Stewart, James
Swagel,
Phillip.
Economic
B.
2009.
"Eigtnt Days,"
"The Financial
2009.
US
New
Crisis:
Yorker,
An
September
Inside
21.
View". Brookings Papers on
Activity, Spring.
Taylor, John and Michael
UK
The
Treasury. 2009. The
Treasury.
Woodford,
Road
2009.
to the
eds, 1999, l-landbook of
Macroeconomics,
Elsevier.
London Summit, http://www.londonsummit.gov.uk
Financial
Regulatory
Reform:
A
New
Foundation.
http://treasury.gov/initiatives/reguiatoryreform/
Wessel, David. 2009.
In
Fed
We
Trust.
Crown Publishing Group, New
Woodford, Michael. 1990. "Public Debt
as Private Liquidity,"
York.
The American Economic
Review, Vol. 80, No. 2 (May), pp. 382-88.
47