/
_
DEWEY
MIT LIBRARIES
i
^-gS)
"'
111
!l
1
'
lllllll"
'I'
'
H'
!
3 9080 03317 5966
Massachusetts Institute of Technology
Department of Economics
Working Paper Series
The "Other" Imbalance and
Ricardo
J.
the Financial Crisis
Caballero
Working Paper 09-32
December 29, 2009
RoomE52-251
50 Memorial Drive
Cambridge,
This
MA 02142
paper can be downloaded without charge from the
Network Paper Collection at
httsp://ssrn.com/abstract=l 529764
Social Science Research
,
The "Other" Imbalance and the Financial
Ricardo
J.
Crisis
Caballero 1
December
29,
2009
Abstract
One
of the main
economic
before the
villains
crisis
was the presence
of large "global
imbalances." The concern was that the U.S. would experience a sudden stop of capital
which would unavoidably drag the world economy into a deep recession.
flows,
However, when the
crisis finally
come, the mechanism did not
did
feared sudden stop. Quite the opposite, during the
were
stabilizing rather
a
imbalance was of
crisis
than a destabilizing source.
The
a different kind:
I
at
all
resemble the
net capital inflows to the U.S.
argue instead that the root
entire world had an insatiable
demand
for safe
debt instruments that put an enormous pressure on the U.S. financial system and
incentives (and this
was
result of the negative
by regulatory mistakes).
facilitated
feedback loop between the
initial
The
tremors
in
its
was the
crisis itself
the financial industry
created to bridge the safe-assets gap and the panic associated with the chaotic
unraveling of this complex industry. Essentially, the financial sector was able to create
"safe" assets from the securitization of lower quality ones, but at the cost of exposing
the
economy
to
a
systemic panic.
governments around the world
This
explicitly
structural
absorb
a larger
problem can be alleviated
share of the systemic
options for doing this range from surplus countries rebalancing their portfolios
riskier assets, to private-public solutions
good parts of the
banks'
if
The
toward
risk.
where asset-producer countries preserve the
removing the systemic risk from the
securitization industry while
Such public-private solutions could be designed with fee
balance sheets.
structures that could
incorporate
all
kind of too-big- or too-interconnected-to-fail
considerations.
Keywords:
systemic
Global imbalances, financial
fragility, panic,
crisis,
safe assets shortage, securitization,
complexity, Knightian uncertainty, contingent insurance, TIC,
contingent CDS
JEL Codes:
1
E32, E44, E58, F30, G01,
MIT and NBER. Prepared
2009.
I
G20
for the Paolo Baffi Lecture delivered at the
Bank of
Italy
on December 10*
thank Francesco Giavazzi, Arvind Krishnamurthy, James Poterba and seminar participants
of Italy for their
comments, and Fernando Duarte, Jonathan Goldberg, and Matthew Huang
research assistance.
First draft:
November
24, 2009.
at
the Bank
for outstanding
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium
Member
Libraries
http://www.archive.org/details/otherimbalancefiOOcaba
/.
Introduction
One
main economic
of the
imbalances," which
refer
villains
before the
the
to
was the presence
crisis
massive and
persistent
of large "global
current account deficits
experienced by the U.S. and financed by the periphery. The IMF, then
new mandate
search for a
imbalances as a paramount
many around
that would justify
risk for
in a
desperate
existence, had singled out these
its
the global economy. That concern was shared by
the world and was intellectually grounded on the devastating crises often
experienced by emerging market economies that run chronic current account
The main
trigger of these crises
deal with a
sudden reversal
in
expansion and current account
is
the abrupt macroeconomic adjustment needed to
the net capital inflows that supported the previous
deficits
(the so called
concern was that the U.S. would experience a similar
drag the world
economy
However, when the
into a
deficits.
"sudden stops"). The global
fate,
which unavoidably would
deep recession.
crisis finally did
come, the mechanism did not
feared sudden stop. Quite the opposite occurred: During the
crisis
at
all
resemble the
net capital inflows to
the U.S. were a stabilizing rather than a destabilizing source. The U.S. as a whole never
experienced, not even remotely, an external funding problem. This
observation to keep
least not
I
in
mind
as
it
hints that
it
is
an important
not the global imbalances per-se, or at
is
through their conventional mechanism, that should be our primary concern.
argue instead that the root imbalance was of a different kind (although not unrelated
to global imbalances, see below):
investors, but also
many
The entire world, including foreign central banks and
U.S. financial institutions,
had an insatiable demand
debt instruments which put enormous pressure on the
incentives (Caballero and Krishnamurthy 2008).
interaction
between the
initial
The
tremors (caused by the
financial industry created to supply safe-assets
U.S. financial
crisis itself
rise in
for safe
system and
was the
its
result of the
subprime defaults)
in
the
and the panic associated to the chaotic
was able
unraveling of this complex industry. Essentially, the financial sector
to create
micro-AAA assets from the securitization of lower quality ones, but at the cost of
exposing the system to a panic, which
In this
view, the surge of safe-assets-demand
and macroeconomic
U.K.,
Germany, and
profits
of
its
finally did
risk
a
concentration
is
take place.
in financial institutions in
few other developed economies), as these
generated from bridging the gap between
natural supply (more on this later on).
2
this rise in
complex
leverage
rise in
the U.S. (as well as the
sought the
institutions
demand and
In all likelihood,
also a central force behind the creation of highly
linkages,
behind the
a key factor
the expansion
the safe-asset shortage
financial instruments
is
and
which ultimately exposed the economy to panics triggered by Knightian
uncertainty (Caballero and Krishnamurthy 2007, Caballero and Simsek 2009a, b).
This
is
not to say that the often emphasized regulatory and corporate governance
weaknesses, misguided homeownership
policies,
and unscrupulous lenders, played no
role in creating the conditions for the surge in real estate prices
However,
it
is
to say that these
were mainly important
resistance path for the safe-assets imbalance to release
structural sources of the dramatic recent
Similarly,
it
is
in
its
and
its
eventual crash.
determining the
minimum
energy, rather than being the
macroeconomic boom-bust
not to say that global imbalances did not play
3
cycle.
a role.
Indeed, there
is
a
connection between the safe-assets imbalance and the more visible global imbalances:
The
latter
were caused by the funding
their ability to
countries'
produce them (Caballero et
al
demand
for financial assets
2008a, b), but this gap
is
in
excess of
particularly acute
for safe assets since emerging markets have very limited institutional capability to
Moreover, Caballero and Krishnamurthy (2008) show
than
rises as
generates
the
demand
a stable
for
that, paradoxically, the risk
AAA-assets grows at the margin. The reason for
income flow and
Acharya and Schnabl (2009) document that
it is
precisely those developed
that the
drops rather
for safe assets
ABCP and
economies where banks could
exploit
ABCP sought by money market funds around the world, and
that experienced the largest stock-market declines during the
countries' issuance of
is
initially
demand
positive wealth effect for the underlying asset producers.
regulatory arbitrage through conduits that issued the
between
this
premium
crisis.
Moreover, they show that there
their current account deficits.
is
no relation
produce these assets.
greatly
added
to the U.S.
NASDAQ
4
financial institutions.
point,
in
however,
latter provides a
is
mutual funds, insurance companies,
for banks,
This safe-asset excess
to,
that the gap to focus on
demand was exacerbated
is
parsimonious account of
many
—something the
of the main events prior to, as well as
global (current account) imbalances
view
unable to do.
demand
for safe assets
began to
rise
above what the
mortgage-borrowers naturally could provide,
mechanisms
U.S.
By 2001, as
corporate world and safe-
financial institutions
began to search for
to generate triple-A assets from previously untapped and riskier sources.
Subprime borrowers were next
loans, "banks" had to create
large
we
Shifting the focus to the
Within this perspective the main pre-crisis mechanism worked as follows:
the
by the
not along the external dimension
but along the safe-asset dimension.
during, the onset of the crisis
is
a variety of collateral,
developed economies.
are so accustomed
alone
from the periphery
for safe-assets
crash which re-alerted the rest of the world of the risks inherent to the equity
market even
The
demand
economy's own imbalance caused by
mandated requirements
regulatory, and
and other
Thus, the excess
all
sorts of
but
in
order to produce safe assets from their
complex instruments and conduits that
numbers and tranching
securitization of
in line,
of their
liabilities.
Similar instruments
payment streams, ranging from auto
relied
on the law of
were created from
to student loans (see
Gorton and Souleles 2006). Along the way, and reflecting the value associated with
creating financial instruments from them, the price of real estate and other assets
short supply
rose
A
sharply.
positive
appreciation of the underlying assets
derivative
See
CDOs and
Krishnamurthy
related products.
and Vissing-Jorgensen
quantitatively very significant and that
iucid description of
(beyond
some
risk aversion).
it
in
feedback loop was created, as the rapid
seemed
to justify a large triple-A tranche for
Credit rating agencies contributed to this loop,
(2007)
for
affects macro-prices
persuasive
evidence that safe-asset
demand
and quantities. See Gorton and Metrick (2009) for
of the special features of safe debt (repos,
in
particular) that justify
its
high
is
a
demand
and so did greed and misguided homeownership
the main structural causes behind the
From
a
boom and
policies,
of triple-A assets
than the traditional single-name highly rated bond. As Coval et
al
was much
quality underlying assets will tend to default, in fact
This
means
even
that,
between these complex assets
distress
if
it
riskier
(2009) demonstrate,
for a given unconditional probability of default, a highly rated tranche
systemic event.
they were not
likely
bust that followed.
new found source
systemic point of view, this
but most
made
of lower
can (nearly) only default, during a
correctly rated as triple-A, the correlation
and systemic
distress
is
much
higher than for
simpler single-name bonds of equivalent rating.
The systemic
once
fragility of
risk in itself
share of them was kept within the financial system rather than sold to
a significant
final investors.
these instruments became a source of systemic
Banks and their SPVs, attracted by the low capital requirement provided
by the senior and super-senior tranches of structured products, kept them
(and issued short term triple-A
liabilities
to fund
in their
books
them), sometimes passing their
(perceived) infinitesimal risk onto the monolines and insurance companies (AIG,
particular).
Schnabl
The recipe was copied by the main European
2009).
interconnected
Through
in
this
increasingly
process,
the
core
of
financial centers (Acharya
the
complex ways and, as such,
it
financial
system
in
and
became
developed vulnerability to
a
systemic event.
The triggering event was the crash
mortgage defaults that followed
in
5
it.
the real estate "bubble" and the rise
But this cannot be
all
of
it.
in
The global
subprime
financial
system went into cardiac arrest mode and was on the verge of imploding more than
once, which seems hard to attribute to a relatively small shock which was well within
the range of possible scenarios. Instead, the real
damage came from the unexpected
and sudden freezing of the entire securitization industry. Almost instantaneously,
The bubble was
of the indirect
in itself a
reflection of the chronic shortage of assets (Caballero 2006) and, as
demand stemming from
securitization
and complex safe-asset production.
mentioned
earlier,
confidence vanished and the complexity which
bread" during the boom, turned into
imaginary.
more
possible the "multiplication of
source of counterparty
risk,
Making matters worse, banks had to bring back
of this
both
real
and
even senior and super-senior tranches were no longer perceived
Eventually,
as invulnerable.
sheets
a
made
new
risk
into their balance
from the now struggling SIVs and conduits (see Gorton
2008). Knightian uncertainty took over, and pervasive flights to quality plagued the
financial system. Fear fed into
transactions, even
among
more
worsened
deficit of safe assets that
in
herd of
these
a
new
new
boon
for
clients. In
clients that
money market
in distress.
had
than their usual
a higher risk-tolerance
in
demand
for
the flight to
funds, which suddenly found themselves facing a
in
demand from
clients,
some money
short-term commercial paper issued by the investment
This strategy backfired after Lehman's collapse,
Primary Fund "broke-the-buck" as a result of
bankruptcy.
Perceived complexity reached a
private funds
were no longer immune
it
levels. Initially,
order to capture a large share of this expansion
market funds began to invest
banks
was behind the whole
perceived uncertainty further increased
these assets. Safe interest rates plummeted to record low
was
financial
newly found source of triple-A assets from the securitization
as the
industry dried up, and the spike
quality
in
the prime financial institutions.
Along the way, the underlying structural
cycle
engage
fear, causing reluctance to
its
new
to contagion.
losses
level as
when
the Reserve
associated with
Lehman's
even the supposedly safest
Widespread panic ensued and were
not for the massive and concerted intervention taken by governments around the
world, the financial system would have imploded.
Global imbalances and their feared sudden reversal never played a significant role for
the U.S. during this deep
crisis. In fact,
the worse things became, the
foreign investors ran to U.S. Treasuries for cover and treasury rates
more domestic and
plummeted (and the
in
matched the downgrade
Instead, the largest reallocation of funds
dollar appreciated).
5
perception of the safety of the newly created triple-A securitization based assets.
Note also that global imbalances per-se were caused by
financial assets
more broadly (Bernanke 2007 and Caballero
main consequence (and
has) the recurrent
al
developed world
until
it
began to
drift
7
was probably the
This drift
NASDAQ crash and
demand
al
the increase
in
it
was
a
It
for
2008b) which had as a
2006
of bubbles (Caballero
toward safe-assets.
system became compromised, as
financial
emergence
et
2008a), but this was not a source of systemic instability
and Caballero et
process.
still
demand
large excess
a
the
in
was only then that the
required input to the securitization
result of the rise in risk
awareness following the
the relative importance of global public savings
the
in
for financial assets.
One approach
explicitly
bear
to addressing these issues prospectively
a
greater share of the systemic
On one hand, the
approach.
risk.
would be
There are two prongs within this
surplus countries (those that on net
assets) could rebalance their portfolios
toward
governments to
for
riskier assets.
On
demand
financial
the other hand, the
asset-producer countries have essentially two polar options (and a continuum
in
between): Either the government takes care of supplying much of the triple-A assets or
it
lets
the private sector take the lead role with government support only during
extreme systemic events.
If
the governments
in
asset-producing countries were to do
have to issue bonds beyond their
risky assets
fiscal
needs, which
themselves. From the point of view of
a
in
it
directly,
turn would require
securities
a crisis
seemed
more information
From
even controlling for
to have
less
was not the case
RMBS
risks
buy
to
across
reflected their relative
for triple-A securities. Investors in these
informed about the quality of the securitized assets than investors
in riskier,
sensitive, securities.
this perspective,
banks were not
been
rating, this
them
balanced allocation of
Adelino (2009) documents that while the issuance prices of lower rated
exposure to
then they would
the reason the NASDAQ-bubble crash did not cause a major systemic event
a central input into
securitization process
the process of bubble-creation.
and they did not
diversify
away enough
In
is
because
contrast, banks played a central role in the
of the risk created by that process.
the world, this option appears to be dominated by one
countries
(e.g.,
China) choose to
demand
in
which sovereigns
in
surplus
themselves.
riskier assets
Rather than discussing these purely public options, which are relatively straightforward
to envision (although their implementation
constraints), in this paper
will
I
focus on the
is
subject to a large
more cumbersome but
public-private option within asset-producing countries.
is
that the main failure during the
triple-A assets
mechanism
creation activity
size shocks)
was not
The reason
possible to preserve the
It is
to relocate the systemic risk
away from the banks and
and the government
ante basis and for
a
fee,
fair
this
potentially larger
is
an option at
in
good aspects of
this
process while
component generated by
This transfer can be
this asset
and medium
done on an ex-
which can incorporate any concerns with the
complexity, and systemic exposure of specific financial institutions. There are
options to do so,
all
of which
amount
to
all
the systemic vulnerability
into private investors (for small
(for tail events).
of political
the private sector's ability to create
in
through complex financial engineering, but
created by this process.
finding a
crisis
number
some form
of partially
size,
many
mandated governmental
insurance provision to the financial sector against a systemic event.
The
rest of the
highlighted
paper goes into
in this
a
more
detailed analysis of the three
introduction: The prelude of the
crisis,
the
crisis,
and
components
a discussion of
policies that can play the dual role of alleviating the safe-asset shortage while increasing
the resilience of the financial system to panics.
//.
The Prelude
The most
large
visible
anomaly
in
international financial markets prior to the crisis
and sustained current account
danger was that
between the
it
deficit of the U.S. Paradoxically,
distracted attention from a
global
demand
for safe assets
more
and the
perhaps
was the
its
biggest
serious and critical imbalance, that
ability of
the U.S. private sector to
generate these assets without over-stretching
develop
II.
A
this
argument
in
more
In
this section
I
detail.
Global Imbalances
Prior to the crisis, there
was
widespread concern with "global imbalances," which
a
and persistent current account
essentially refer to the massive
the U.S. and financed by the periphery (see Figure
Figure
financial system.
its
deficits
experienced by
1).
1.
Current Account
1.5
1.1
0.7
o.
Q
U
|
0.3
-0.1
._*=
t$
o
s?
-0.5
f
\r*
>»«>»
\
-0.9
USA
-1.3
-•-EU
—•— Rest of World
-1.7
..
-2.1
1990
1992
1996
1994
1998
2000
2002
2004
2006
2008
Source: IMF International Financial Statistics, author's calculations.
In
Caballero et
al
(2008a)
we argued thatthe emerging market
crises at the
end of the
1990s, the subsequent rapid growth of China and other East Asian economies, and the
associated
emerging
In
rise
in
markets
commodity
prices
toward the
in
United
Caballero and Krishnamurthy (2007)
we
recent years reoriented capital flows from
States.
8
In
effect,
emerging markets and
described the emerging markets side of the story.
We
argued
there that the financial underdevelopment of these economies naturally led to the formation of domestic asset
commodity producers
in
need of sound and
newfound wealth turned
liquid financial
instruments to store their
to the U.S. financial markets and institutions, which
perceived as uniquely positioned as providers of these instruments.
developed
in
A
were
related story
was
Bernanke's (2007) famous savings glut speech.
Concern about these "imbalances" was
intellectually
grounded on the devastating
crises
often experienced by emerging market economies that run chronic current account
The main
deficits.
needed
trigger of these crises
to deal with a
sudden reversal
in
is
the abrupt macroeconomic adjustment
the net capital inflows that supported the
previous expansion and current account deficits (the so called "sudden stops").
Figure 2 from Calvo and Talvi (2005) and Calvo, Izquierdo and Talvi (2006) captures the
dramatic events during
the end of the 1990s.
a
In a
sudden stop such
as the
matter of months, net capital flows as
by double-digits, with a corresponding adjustment
demand. The exchange
one that affected South
in
share of
a
East Asia at
GDP
declined
the current account and aggregate
rate collapsed (and interest rates spiked),
and along with
it
so
did national income.
The global concern by the mid-2000s was that the
U.S.
would experience
which unavoidably would drag the world economy into
when
the
crisis finally
did
come, the mechanism did not
sudden stop. Quite the opposite occurred, during the
U.S.
were
a stabilizing rather
than
deep
a
at
crisis
a destabilizing source.
all
a similar fate,
recession.
However,
resemble the feared
net capital inflows to the
The
experienced, not even remotely, an external funding problem.
U.S. as a
whole never
Moreover, during the
bubbles during periods of large net capital inflows. The crash of these bubbles lead to large capital flow reversals.
The events of the
late
1990s corresponded to
a
coordinated crash
funds toward the U.S. large enough to lead to bubbles
in
in
emerging markets and hence
a reallocation of
the U.S. and other developed economies (Caballero et
al
2008b).
10
early phase of the
crisis,
Sovereign Wealth Funds were the primary providers of fresh
capital to distressed U.S. banks.
Figure
9
2.
Sudden Stop
in
South East Asia*
Boom and
Capital Flows
(in
Ml
of
US
dorian, taat fbut quartern)
Cm
Bust in Capital Flows
% of GDP, laal four quarters)
Boom
jfifstflililllii
Bust
Aroragi 1934.08
paak199G
paaa1996
trough 1998
Thailand
v
422,9
-11*
Indotntla
2,6
21.9
Jfl
Koran
*<*
-14,6
3*
Philip pints
Bfl
10,3
-14,8
Malaysia
B,1
*,T
-10 J)
CEO
C3Bg>
<3E>
SEA-6
pa a k 1996
Real Exchange Rate
Cun-ont Account
(»%of GOP)
SSKSSISSBSUSESSE
;~s;s;t;ssssrrrs
>0>»a, Matayaia. PMIppinaa
ztul
I
II
I 1
II IEI
ly
jgj
Thailand
Source: Calvo and Talvi (2005) and Calvo, Izquierdo and Talvi (2006).
Figure 3
shows that the
U.S. current
around since 2006, and that
it
account
deficit
was already experiencing
a turn-
indeed dropped sharply after the Lehman episode. But
the key contrast with emerging markets experiencing large external adjustments
is
that
the U.S. dollar appreciated sharply (and- U.S. interest rates plummeted) during this
episode. That
due
is,
the adjustment was the result of a contraction
to domestic financial
They did so
until
problems rather than due to
a
in
aggregate
demand
sudden shortage of net external
Secretary Paulson decided to exemplary punish shareholders during the Bear Stearns
intervention.
11
B
funding resulting from a
observation to keep
least not
Figure
in
flight to quality
mind as
it
away from the
hints that
it
is
U.S.. This
is
an important
not the global imbalances per-se, or at
through their conventional mechanism, that should be our primary concern.
3.
US Current account and dollar-euro exchange
"—Current
-3.0
-
-4.0
-
rate
l^"\
Account
[\
O
a.
1.4
a
3
QJ
U
\D
Q.
o
*
-5.0
TO
-
-
1.3
O
O
111111111
OOOOCDOOOOOOOOOOO
i
i
R)
i
i
LOu~iLTit-DU3i-D'>or^r^-r^r^-
oo
oo
do
do
=
~
—J^
o.
<z
~
=
*->
a.
<c
w
—=
o —
"
Jjj
>
a.
<c
'
LJ
O —
ra
ex
<E
—=
'
<->
o
>
i
c>
1
on
C3
'O—
a.
to
1
1.1
o~i
O
=
i
Source: IMF International Financial Statistics
The Critical Safe Assets Imbalance
II.
Since the 1980s, there has been a surge
those for
risk "diversification").
than 160 percent of
GDP
Debt plays an important
September
11,
2001 and
in
in financial
Figure 4
shows that
assets to store value
U.S. financial assets
1980 to almost 480 percent
in
addition to
grew from
less
the third quarter of 2007.
role in this surge, especially in the
NASDAQ crash
(in
1980s and
in
the post
period.
12
Figure 4,
Assets
U.S. Financial
600
500
Jan-
Jan-
Jan-
Jan-
52
57
62
67
Jan- Jan-
Jan-
Jan-
Jan-
Jan-
Jan-
Jan-
82
87
92
97
02
07
77
72
I
Debt
Equity
Source: Federal Reserve, CRSP.
The growth
in
the ratio of debt to
GDP was due
debt outstanding rose from 12 percent
5).
At the
percent to
same
15
time, the
percent.
in
to an increase
The domestic
private sector, especially the financial sector.
U.S.
10
1980Q1
to
in
financial sectors' share of
34 percent
government share of outstanding
(Unsurprisingly,
this
trend
debt generated by the
U.S.
2007Q3
in
debt
fell
(Figure
from 24
been reversed of
has
late.)
Households and non-financial firms, which are the residual, were each responsible for a
relatively flat share of the
A fundamental
growing debt-outstanding
driver of this increase
safe debt instruments. This
also
from many
not be readily
Debt
is
the
U.S. financial institutions.
met by
sum
liabilities
demand came from
of debt outstanding for
was the
insatiable
demand
for
foreign central banks and investors, but
The demand
existing sources of triple-A debt.
governments, the federal government and domestic
of the NYSE,
debt
in
pie.
households,
for safe
Only a
debt instruments could
sliver of
corporate debt,
businesses (non-financial), state and
financial sectors.
Equity
is
the
sum
of
local
market capitalizations
Nasdaq and Amex exchanges.
13
for example, carries a triple-A rating.
safe-assets imbalance, which
Figure
is
not
Thus, the world
likely to
economy experienced
a
massive
financial
system.
go away any time soon.
5.
Government and
financial sector debt
0.4
0.35
D
0.3
e
ID
i
0.25
O
S
a
o
a
n
0.2
0.15
£
CO
0.1
0.05
Jan-80
Jan-84
Jan-92
Jan-
Jan-96
Jan-00
Source: Federal Reserve
This imbalance
presented
a
system and
its
incentives.
this
The
demand
Jan-08
li
large profit opportunity for the
However, creating safe assets to meet
financial
Jan-04
Government
Domestic financial sector
U.S.
put enormous pressure on the U.S.
U.S. financial
system created safe assets from
unsafe ones by pooling assets and issuing senior claims on the payoffs of the pools. The
senior claims are protected against losses because
before the senior claims.
more
junior claims absorb losses
The claims on these pools of assets are
called collateralized
debt obligations, or CDOs.
Global issuance of
CDOs grew from $185
billion in
2000 to $1.3
trillion in
share of these assets carried triple-A ratings, at least before the
Government
crisis.
includes federal, state, and local government. The denominator
is
2007.
12
A
large
As of June 2007,
total
households, businesses (non-financial), federal, state and local government, and the domestic
debt outstanding by
financial sector.
14
Fitch,
agency, gave a triple-A rating to almost 60 percent of
ratings
a
structured products, according to a Fitch Ratings (2007)
and Stafford (2009).
13
Benmelech and Dlugosz (2009)
collateralized loan obligations (CLOs) issued
issuance was rated triple-A, a percentage that
In
contrast, Fitch
(in
June 2007) and S&P
(in
document
fairly stable
is
throughout the period.
early 2008) gave a triple-A rating to less
in
2000, the volume of
debt issuance; by 2006,
Corroborating
government
CDO
this ratio
triple-A assets
was only 18 percent
for
these
instruments,
fast as did
of the
had risen almost four-fold.
demand-pull
the
and CDOs. Corporate
of U.S. issuance of corporate debt
issuance
issuance
in
the issuance of CDOs.
debt issuance grew between 2000 and 2006, but not nearly as
In
sample of
find that, in a large
between 2000 and 2007, 71 percent of
was slower than the increase
compares the volume
Figure 6
global
cited in Coval, Jurek
than one percent of single-name corporate issuers. Moreover, the increase
of corporate debt
all
CDO
volume
issuance.
of corporate
14
the
as
quantity
non-
of
expanded, the yields required to hold them tightened.
In
January 2002, the spread between Moody's triple-A seasoned corporate bond rate and
a
30-year U.S. Treasury was 1.19 percentage points.
15
began, the spread reached 0.55 percentage points.
Moody's
triple-A
IMF Global
seasoned corporate bond rate
Due
to the
crisis,
What
Credit Ratings
issuance of
CDOs
Mean,"
the corporate debt market
in
percentage of corporate debt issuance volume
2002.
In
early 2007,
it
shows the spread
These numbers include both CDOs and
Fitch Ratings,
in
is
for
first five
CDO A 2.
August 2007.
first five
the newly thawed debt markets, issuing 465
the same period. Thus, for the
The spread between Moody's
Figure 7
collapsed to less than 24 billion dollars during the
2009, while corporations rushed to secure financing
in
early 2007, before the crisis
relative to a 20-year Treasury bond.
Financial Stability report (2009), Figure 2.2.
"Inside the Ratings:
In
months
of 2009,
CDO
months
of
billion dollars
issuance volume as a
only five percent.
triple-A rate
and
a
20-year Treasury was 0.9 percentage points
in
January
reached 0.45 percentage points.
15
Figure
6.
A
2001
2000
2002
-
shift
2004
2003
CDO issuance
toward CDOs
2006
2005
2007
2008
2009"
as a share of corporate debt issuance
16
Source: SIFMA, IMF Global Financial Stability report (2009)
Coval, Jurek
and Stafford (2008)
CDX tranches -
find that
with credit default swaps (CDS) as assets
-
carried similar yields to single-name
similar loss rates ("despite their highly dissimilar
analysis
is
September 2004
triple-A corporate
to
September
bonds carried over
for safe
assets.
became very complex,
2009a, b).
Also,
retained a large
2007'.
economic
risks").
number
CDS with
Their period of
This suggests that the tight pricing of
in financial institutions
helped meet the global
However, these securities and the linkages among firms
leaving the system vulnerable to panic (Caballero
senior
CDO
into the structured-finance market.
The creation of CDOs and the leveraging
demand
essentially, tranches of a
CDO
tranches concentrate macroeconomic
of these assets.
risk,
and Simsek
and banks
Thus, the response of the financial system to
U.S. corporate bond issuance includes "all non-convertible debt, MTNs and Yankee bonds, but excludes
CDs and federal agency debt." Both investment-grade and high-yield issues are included. Corporate debt issuance
volumes are from the Securities Industry and Financial Markets Association (SIFMA) and based on Thomson
Reuters data. CDO issuance includes the issuance of CDOs and CDO A 2, from the IMF Global Financial Stability
report (2009). The 2009 data covers through the end of June.
16
produced conditions
the safe-assets imbalance
uncertainty could do significant
Figure
damage
in
which an episode of Knightian
(Caballero and Krishnamurthy 2007).
7.
AAA corporate bond spread
2
1
1
-
5
o.
a
o>
a
c
0)
u
Ai
k
'
I
l
1
'
0}
Q.
0.5
-
14-D SC-01
14-Jun-03
14-Dec-04
14-Jun-06
14-Dec-07
Sources: Federal Reserve
II.
C
From Global
to Safe Assets
14-Jun-09
17
Imbalances
Having downplayed the conventional concern with global imbalances,
note that there
global
is
a
imbalances: The latter were caused
This figure plots the spread
is
a small
is
important to
connection between the safe-assets imbalance and the more visible
by the funding countries'
financial assets in excess of these countries' ability to
There
it
between Moody's
triple-A rate
demand
for
produce them (Caballero et
al
and a 20-year constant maturity Treasury.
maturity mismatch because "Moody's tries to include bonds with remaining maturities as close as
possible to 30 years." However, there
is
no 30-year constant maturity Treasury bond
since the Treasury stopped issuing 30-year debt for a period.
Thus, the spread
for a
shown
good part of the 2000s,
in
the figure
is
a noisy
measure of the maturity-matched spread.
17
2008), but this gap
is
particularly acute for safe assets since
limited institutional capability to
produce them.
assets from the periphery greatly
added
emerging markets have very
demand
Thus, the excess
to the U.S.
for safe-
economy's own imbalance caused
by a variety of collateral requirements and mandates for mutual funds, insurance
companies, and others. The point, however,
external dimension
we
is
that the gap to focus on
FDI.
owned by
This
is
assets and
figure
18
in
However,
11,
emerging markets
liabilities
this
throughout
after the
1990.
this period,
and FDI
was
liabilities
a
sharp
as foreign
board and particularly high return risky
U.S. This
is
partially a price effect; but
Cumulated
whereas flows
flows
into U.S.
into equity
We
debt
debt
into
see this
liabilities
on the safe tranches of the asset
in
it
also reflects
Figure
whole
which
and equity and FDI
and FDI reached an
risk in U.S.
inflection point
After the crash the
assets as well and decided
distribution.
This phase resembled the Japanese (real estate) asset shopping spree during the 1980s, which
to safe assets or their safe tranche but to
9,
were growing exponentially
attacks, tapering off thereafter.
world came to realize that there was substantial
18
of the late 1990s, there
toward debt and away from equity.
dot-com crash and 9/11
to refocus
of the
trend turned around sharply after the dot-com crash and the
shows the cumulated flows of foreign assets
since
crisis
to U.S. equity
U.S. assets across the
2001 attacks on the
a shift in allocation
liabilities
liabilities
according to careful estimates by Gourinchas and Rey (2007). The
the ratio of U.S. debt
September
the U.S., the
Figure 8, which gives the ratio of debt to equity and FDI for U.S.
after the
demanded
investors
assets.
shown
liabilities,
in
owned by
the rest of the world are skewed toward debt, rather than equity and
shows that
decline
not along the
are so accustomed to, but along the safe-asset dimension.
Relative to the liabilities of the rest of the world
U.S.
is
was not targeted
assets.
18
Figure 8.
Ratio of debt to equity and FDI
1990q1
1992q1
1994q1
•
Gross
1996q1
liabilities
of
1998q1
US
2000q1
Gross assets
2D02q1
of
20Q4q1
US
Source: Gourinchas and Rey (2007).
Figure
9.
Cumulated
U.S. liabilities flows since
1990
g
E
S
1500
1990q1
1992q1
1
994q1
1996q1
Debt
1998q1
2000q1
2002q1
2004q1
Equity and FDI
Source: Gourinchas and Rey (2007).
19
D
Complexity and Systemic Risk Build-up
II.
"Banks" generated safe assets from risky ones by creating complex instruments that
pooled assets, such as subprime mortgages, and creatively divvied up the flows from
those assets.
Figure 10
two
shows
the balance sheet of a CDO.
properties: (1) pooling of assets; and (2) tranching of
of the figure
the
a caricature of
CDO
is
shows
a
individual assets
RMBS
or
owned by
CMBS CDO,
A CDO can be defined by
liabilities.
here
is
where the tranching takes
sometimes referred to
CDO
to
CDO,
a
the more junior
the blue boxes are bundles of residential
It
is
right (liability) side has the liabilities of the
place.
The bottom tranche
is
liabilities
liability is
liabilities
Similar instruments
or
CDO;
it is
the equity tranche,
from
protected from loss because equity and
have to be wiped out before the given
if
here on the asset
as "toxic waste." Although the implementation details vary
given non-equity
These more junior
flows.
The
side
left (asset)
the CDO, represented by blue boxes;
commercial mortgages. These blue boxes are the CDO's assets.
side that pooling takes place.
The
liability
all
suffers any loss.
serve as a buffer. The arrow shows the direction of the cash
were created from
securitization of
all
sorts of
payment
streams, ranging from auto to student loans.
Figure 10. The Balance Sheet of a
Assets
CDO
Liabilities
AAA tranche
AA tranche
A tranche
BBB
tranche
BBB- tranche
Toxic waste
20
Consider
a
CDO
each mortgage
with two $1 mortgages as assets. Suppose the probability of default for
is
10 percent and that
in
if
default, the
mortgage
is
worthless. Suppose
further the defaults are uncorrelated. There are (essentially) three states of the world:
no defaults; one default; and two defaults. What
consisting of
liabilities
assets?
Then the $1
two $1
liability
liabilities
would pay
cents 18 percent of the time; and
individual
mortgage
is
individual mortgages,
it
CDO had
the
if
that each received equal payoffs from the
in full
81 percent of the time;
diversification.
But
CDOs
junior tranche and a $1 senior tranche.
is
The senior tranche pays
created. However,
very exposed to "systematic"
risk;
(Consider an additional asset
in
it
would pay 50
"safer" than the
Suppose there are
new
when both
in full
99 percent of the
CDOs "economic catastrophe bonds."
when
During
asset defaults only 1 percent of the time, whereas the
explored by Coval, Jurek, and Stafford (2009a, b),
catastrophes
is
of the mortgages default.
the economy, which also pays
senior tranche defaults 100 percent of the time.)
like
$1
99 percent of the
in full
time; suppose the asset's payoffs are uncorrelated with the mortgage payoffs.
the global downturn, the
a
key to note that the senior tranche
it is
only defaults
is
create even safer assets by
dividing their liabilities into tranches with different seniorities.
time. Thus, a "safe" asset
it
would be worthless one percent of the time. An
worthless ten percent of the time, so this asset
because of
only one class of
who
This point has
been rigorously
structured finance products
call
precisely during
economic
defaults on the asset side of the balance sheet are large
enough to
By design,
it
is
eradicate the buffer protecting the senior tranches.
As private-label securitizations increased,, the price of
short supply rose sharply,
in
real estate
and other assets
in
part reflecting the value associated with creating financial
instruments from them. A positive feedback loop was created, as the rapid appreciation
of the underlying assets
seemed
to justify a large triple-A tranche for derivative
and related products. Credit rating agencies contributed to
this loop,
CDOs
and so did greed
21
and misguided homeownership
but as stated earlier: They were not the root
policies,
cause.
shows the
Figure 11
this period.
risk in itself
19
fast
growth of global private-label securitization issuance during
The systemic
once
than sold to
a significant
final
these instruments became
fragility of
a
source of systemic
share of them was kept within the financial system rather
Banks and their SIVs, attracted by the high
unleveraged investors.
return and low capital requirement combination provided by the senior and supersenior tranches of structured products, kept
their (then perceived as) infinitesimal risk
(AIG, in particular).
20
them on
The
complex ways and vulnerable to
light-blue bars represent
increased from $25 billion
use of
CDO A 2
in
money
became interconnected
a systemic event.
CDOs
that have other
billion in
2006.
in
According to Acharya
triple-A asset-backed securities
all
CDO A 2, CDOs
2000 to $338
quickly absorbed by
institutions with insatiable appetite for safe assets.
and Schnabl (2009), "about 30% of
19
was
process, the core of the financial system
this
increasingly
and indirect hoarding by issuing
this direct
highly rated short-maturity commercial paper, which
Through
sometimes passing
onto the monolines and insurance companies
They funded much of
market funds and the many other
their books,
remained within
Global issuance of
as their assets.
CDO A 2
As Coval, Jurek, and Stafford (2009a,b) point out, the
only increases the correlation between default and systemic events.
An important reason
(2009): "In hindsight,
was regulatory and
it is
for the creation of SIVs
stemmed from
regulatory arbitrage.
From Brunnermeier
one distorting force leading to the popularity of structured investment vehicles
arbitrage. The Basel Accord (an international agreement that sets guidelines for bank
clear that
ratings
I
regulation) required that banks hold capital of at least eight percent of the loans on their balance sheets; this
capital
requirement
capital charge at
(called a 'capital charge')
all
was much lower for contractual credit lines. Moreover, there was no
- noncontractual liquidity backstops that sponsoring banks
for 'reputational' credit lines
provided to structured investment vehicles to maintain their reputation. Thus, moving a pool of loans into offbalance-sheet vehicles, and then granting a credit
line to that
pool to ensure a AAA-rating, allowed banks to
reduce the amount of capital they needed to hold to conform with Basel
I
regulations while the risk for the bank
remained essentially unchanged. The subsequent Basel
Accord, which went into effect on January 1, 2007 in
Europe but is yet to be fully implemented in the United States, took some steps to correct this preferential
treatment of noncontractual credit lines, but with little effect. While Basel implemented capital charges based on
II
II
asset ratings, banks
were able to reduce
Because of the reduction of idiosyncratic
their capital charges by pooling loans
risk
better rating than did the individual securities
overall rating
downgraded
in
off-balance-sheet vehicles.
through diversification, assets issued by these vehicles received
in
the pool.
In
addition, issuing short-term assets
a
improved the
even further, since banks sponsoring these structured investment vehicles were not sufficiently
one reason the large banks kept exposure
for granting liquidity backstops." According to Tett (2009),
to the super senior tranches on their books
on these tranches, eventually decided
it
was that
AIG, which
in
the early stages of the
had too much exposure and stopped providing
boom
provided insurance
this insurance.
22
—
the banking system, and
fraction
rises
if
one includes ABCP conduits and SIVs that had recourse,
50%." A November 2007 announcement by Citigroup shows
to
investment bank had $43
billion
in
exposure
in
super senior tranches of ABS
"primarily" backed by subprime residential mortgages. Notably,
total direct
of
of
its
exposure to
U.S.
subprime mortgages was $55
Citi's
this
its
CDOs
investment bank's
so almost 80 percent
billion,
exposure was through super senior tranches. By 2007, the Federal Reserve Bank
New
trillion,
York calculated that SIVs and similar vehicles had combined assets of $2.2
more than the
assets of hedge funds ($1.8
assets of the five largest broker dealers ($4
trillion),
trillion) (Tett
and more than half the total
2009).
Figure 11.
Global private-label securitization issuance by type
4500
-
3500
-
o
+
TH+
Mil
iiiii
1
i
1000
500
-
-
-
2000
2001
2002
2003
ABCP
2004
~~1~
—
—
2005
2006
,
1
1
,
2007
2008
1
1
2009
MBS BABS bCDO «CD02
Source: IMF, Global Financial Stability Report, Figure 2.2.
///.
The Crisis
The conditions were thus
far
ripe for a severe systemic event,
more force than anyone had
anticipated. There
estate prices and the corresponding rise
in
which eventually came with
was the shock from
declining real
subprime defaults, but even under the most
23
pessimistic scenarios, these shocks pale in comparison with the
that eventually followed once the panic set
III.
A
magnitude of the
crisis
in.
The Shock
The triggering event was the crash
mortgage defaults that followed
system went into cardiac arrest
the real estate "bubble" and the
in
But this cannot be
it.
mode and was on
all
of
it.
rise in
subprime
The global financial
the verge of imploding
more than
once, which seems hard to attribute to a relatively small shock such as the real-
estate/subprime
In
combo
(see Caballero 2009).
Caballero and Kurlat (2009),
we
constructed an estimate of
mortgage-related losses were amplified by the
estimate,
we computed
crisis
how much
banks'
that these losses sparked.
initial
For this
the evolution of the market value (equity plus long term debt)
of the major U.S. banks since January 2007,
which yielded an estimate of
the right side of these banks' balance sheets.
21
Absent any feedback
should be equal to the losses suffered by the assets on the
sheets. However, as illustrated
in
Figure 12,
we
find that losses
left
total losses
on
effects, these losses
side of the balance
on the
right side are
on
the order of three times the IMF's (evolving) estimates of losses related to mortgage
assets accruing to U.S. banks.
The procedure
22
for estimating this
was
as follows: For equity,
bank's market capitalization, excluding increases
in
we
simply tracked the evolution of each
the market 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
the 10-K statements as of December 2007, assuming the interest rate was equal to the rate on 10-
year Treasuries plus the spread on 5-year CDS for each bank, obtained from JP Morgan. Assuming an unchanged
maturity
profile,
we
then tracked the changes
basis of the evolution of the
in
the implied market value of each bank's long term debt on the
CDS spread. The banks included
in
the calculation are the 19 banks that underwent
the "stress tests" plus Lehman, Bear Stearns, Merrill Lynch, Wachovia, and Washington Mutual.
The IMF uses
a projection of
macroeconomic variables and default
market values to estimate losses on subprime-related
overreacted due to
fire sales,
securities.
rates to estimate losses on loans,
and
To the extent that market prices of securities
our procedure understates the multiplier.
24
Beginning
in
2008, and increasingly after the
fall
of Bear Stearns, the overall loss in
market value became larger than the losses from subprime assets alone. The market
began to price
recession,
its
from the overall disruption of
losses
and losses on other types of assets which
from the mortgage market
FiRure 12. Losses
far
financial markets, the severe
exceeded the estimated losses
itself.
from mortgage
assets, total loss of
market value and
—
—
J 000
multiplier.
TauHoitolnuikflniiiicleqjnYpiuideBllQf
US binki
u eiiimiltO
by IMF
Jfe_
1500
rV
1000
1
sao
7Z
r07
„„.,
hi 07
S*p07
4av07
anOS
,ii'
as
Mir oa
ui 01
Oct OS
Dae OS
Mi' 09
Miy 09
MO
Sources: IMF Global Financial Stability Reports, banks' financial statements and JP
Morgan. From Caballero and Kurlat (2009).
II
LB
The
The Panic
real
damage came from the unexpected and sudden
securitization industry.
Securities
in
line
in
Figure 13
is
"New
The new issuance
23
series
The
is
crisis in this
the
market
sum over the
real estate; autos; credit cards;
is
entire
Issuance of Asset-Backed
Previous 3 Months," from Adrian and Shin (2009); the data originally
from JP Morgan Chase.
commercial
The blue
freezing of the
come
apparent from the disappearance of
following categories of ABS:
"home
equity (subprime)";
student loans; non-US residential mortgages; and other. The data were
provided by Tobias Adrian.
25
new
The red
issuances.
line
is
the implied spread on the 2006-1
measures the cost of insuring against default by
mortgage-backed securities of the
JP
f irst-ha
lf-of-2006 vintage.
Morgan Chase and not only corroborates the
but also makes
it
triple-A
is
tranches
ABX, which
of
subprime
The spread data are from
impression from the quantity side
crisis
clear that the collapse in quantity
AAA
demand
rather than supply driven.
Figure 13.
Freezing of the Asset-Backed Securities (ABS) market
700
o
o
o
o
5
o
a
o
o
m
m
_
-^
£J
m
£
m
jn
'""
en
m
u\
2O06-1AAAABX
Sources: JP
IS.
3
-
New
o
a
o
o
O
O
m
m
m
r^
in
fN
o>
Morgan Chase, Adrian and
bread" during the boom, turned into
a
o
o
cr.
ssua nee of Asset-B scked Securlt ei
Confidence vanished and the complexity which
(see Gorton
o
made
Shin (2009).
possible the "multiplication of
source of counterparty
2008). Senior and super-senior tranches
risk, real
and imaginary
were no longer perceived
as
invulnerable, and making matters worse, banks had to bring back into their balance
sheets
more
of this
new
risk
from the now struggling SIVs and conduits.
26
In
December 2007,
the SIVs' $49
assets
were
Citigroup provided a guarantee facility to
billion in assets
onto
its
balance sheet.
financial-institutions debt.
percent) assets.
(6 percent);
One hundred percent
SIVs, essentially bringing
About 60 percent of the
Thirty-nine percent
SIVs'
were structured-finance
and non-U. S. (12 percent) residential
assets, including: U.S. (7 percent of total assets)
MBS; CBOs, CDOs, and CLOs
24
its
student loans
(5 percent),
and credit card
of the structured-finance assets carried
(5
Moody's
triple-A ratings.
Knightian uncertainty took over, and pervasive flight to qualities plagued the financial
system.
Fear
fed
transactions, even
into
more
among
fear,
and caused
reluctance
to
engage
in
financial
the prime financial institutions (see Figure 14).
Figure 14.
TED Spread
Source: Global Financial Data.
"Citi
Commits Support
November 2008,
to Purchase All
Citi
directly
Facility for Citi-Advised SIVs," Citigroup press release,
purchased the remaining assets of
its
SIVs.
Remaining Assets," Citigroup press release, November
"Citi Finalizes SIV
December
13, 2007.
In
Wind-down by Agreeing
19, 2008.
27
As the crises spread there was a sharp
shift of
the maturity structure of asset-backed
commerciai paper (ABCP) toward very short-term maturities. At the height of the
crisis,
nearly 80 percent of the asset-backed paper issued had a maturity of only one to four
an increase of more than 40 percentage points since January 2004.
days,
shortening of the maturity structure was
decline
in
September 2008.
ABCP declined from 17 days
in
a
pronounced
Figure 15, the average
in
August 2008, to 9 days
in
has long been understood that funding through short-maturity
It
generates rollover
emerging market
was
gradual, but there
September 2008, when Lehman collapsed. As shown
maturity of newly issued
liabilities
fairly
This
crises
risk,
but, as
by Broner et
al
it
has been
documented
in
the context of
do not have many other
(2008), borrowers
options during severe financial crises.
Figure 15.
Average maturity of newly issued ABCP (estimate)
Jan-06
Jan-07
Jul-06
Jul-07
Jan-08
Jul-08
Jan-09
Jul-09
Sources: Federal Reserve, author's estimate.
In
25
the months following Lehman's collapse, there was also a sudden
short-term maturities
in
the CP market for financial companies, as
The weighted average maturity
80] day CP
is
is
estimated as follows:
[1-4]
day CP
is
move toward
shown
in
very
Figure 16.
26
treated as (4+l)/2 = 2.5 days; [41-
treated as (80+41)/2 = 60.5 days; etc. For [80+] day CP, a value of 90 days
was used.
28
The decline was precipitous because market participants were surprised by the depth of
and the chaotic aftermath of Lehman's collapse, whereas
financial firms' difficulties
participants
fact,
as
knew
(since at least
problems festered
of the
2007) about problems
in
the
ABS market.
the ABS market during the early phases of the
in
financial firms conservatively
summer
lengthened their CP maturities.
In
in
one
crisis,
August 2008, 50 percent
commercial paper issued by financial companies matured
November, 87 percent of the paper issued matured
In
in
one to four days. By
to four days.
Figure 16.
Average maturity of newly issued
financial
CP (estimate)
25
20
^~^
Q
A
ILL/ \/^
10
5
Jan-06
Jan-07
Jul-06
Jul-07
Jan-08
Jul-08
Jan-09
Jul-09
Sources: Federal Reserve, author's estimate.
Along the way, the underlying structural
cycle,
worsened
deficit of safe assets that
as the newly found source of triple-A assets
industry dried up, and the spike
Safe interest rates
plummeted
There was also substantial
in
uncertainty further increased
was behind the whole
from the securitization
demand
for safe assets.
to record low levels (see Figure 17).
volatility of in
the maturity structure during this period,
likely
the result of a
desire of financial firms to push out the maturity structure and a willingness of investors to absorb longer
maturities that varied with volatile market conditions and expectations of government involvement.
29
Figure 17.
3-Month Treasury
Bill
rate
Source: Federal Reserve Bank of
Initially,
the flight to quality was a boon for the
found themselves with
demand from
a
clients.
money market
In
investment banks
in
distress.
Reserve Primary Fund
bankruptcy.
to
invest
than their usual
clients,
This strategy backfired after Lehman's collapse,
"broke-the-buck"
Perceived
as
a
complexity
result
in
some
commercial paper from the
short-term
in
funds, which suddenly
order to capture this expansion
clients having a preference for riskier assets
money market funds began
Lehman's
new
herd of
St. Louis.
of
reached
its
a
losses
new
when the
associated with
level
as
even
the
supposedly safest private funds were no longer immune to contagion. Widespread panic
took over and had
it
not been for the massive and concerted intervention taken by
governments around the world, the
financial
system would have imploded
27
Reserve Primary Fund had invested $785 million
assets.
Immediately after Lehman
filed for
in
Lehman
27
debt, which constituted about 1.2% of
bankruptcy, the fund suffered
a
massive run, with over $30
its
billion in
30
While the end of the panic has removed some of the pressure on the safe-assets world,
the
crisis
destroyed a significant share of the financial industry created by the private
demand
sector to satisfy the large
triple-A assets
gap
now worse
than
it
around the world. The shortage of
was before the
crisis,
and unless
a solution to this
found soon, many of the weaknesses that were created before the
is
reemerge
IV,
is
for safe assets
in
same
the
crisis will
mutated form.
or a
The Policy Options
The core
policy question
is
how
to bridge the safe-asset gap without over-exposing the
financial sector to systemic risk.
One approach
explicitly
bear
to addressing these issues prospectively
greater share of the systemic
a
risk.
would be
for
There are two prongs within this
approach. On one hand, the surplus countries (those that on net
assets) could rebalance their portfolios
toward
governments to
riskier assets.
On
demand
financial
the other hand, the
asset-producer countries have essentially two polar options (and
a
continuum
in
between): Either the government takes care of supplying much of the triple-A assets or
it
the private sector take the lead role with government support only during
lets
extreme systemic events.
Should the governments
would have
to
buy
to issue
risky assets
the
flight to quality
asset-producing countries choose to do
bonds beyond
their fiscal needs,
which
themselves. From the point of view of
redemption requests (about
a.m. the following day.
in
half of
its
Money market
total assets) before
it
a
in
directly,
they
turn would require
them
it
balanced allocation of
risks
stopped accepting redemption requests at $1 at 11
funds had been considered extremely safe, and had indeed benefited from
billion (34%) since mid 2007. The drop in the
caused investors to question the safety of the entire industry. There were net
during the previous year, growing by about $850
Reserve Primary Fund's
NAV
billion during that week, as well as a large shift from prime funds towards funds
government debt. In order to stem the panic, on September 19 the U.S. Treasury
guarantee program that would compensate investors if the NAV of participating funds fell below $1.
redemptions for about $170
investing exclusively
announced
a
in
31
across the world, this option appears to be dominated by one
surplus countries
(e.g.,
demand
China) choose to
riskier assets
Rather than discussing these purely public options,
cumbersome
but
potentially
in this
in
themselves.
paper
option
public-private
larger
which sovereigns
in
will
I
focus on the
within
more
asset-producing
countries.
The reason the public-private venture
the main failure was not
complex
in
is
an option despite the recent
in
that that
the private sector's ability to create triple-A assets through
financial engineering (although rating agencies
the process), but
crisis, is
may have
excessively facilitated
the systemic vulnerability created by this process. Public-private
ventures preserve the successful parts of this asset creation activity while finding
mechanism
to reallocate the systemic risk
sheet to private unlevered investors
government
component
(for
it
small and
creates from the banks' balance
medium
size
shocks) and the
(for tail events).
Just raising capital requirements achieves the goal of reducing vulnerability but
it
does
so at the cost of exacerbating the structural problem of excess safe-asset demand.
this sense,
a
it is
In
not a stand-alone policy.
There are two broad categories of recent proposals to reduce
crisis
risk
without
excessively limiting the financial sector's ability to bridge the safe-assets gap:
•
Pre-paid/arranged contingent capital injections, and
•
Pre-paid/arranged contingent asset and capital insurance injections.
The basic purpose of the former, contingent capital
associated with the holding of capital
this
when
approach recognizes that access to
advance, since
kind differ
in
it
is
it
is
injections,
is
not needed. However, and centrally,
capital during crises
needs to be arranged
often hard to raise capital during a severe
crisis.
their sources of this contingent capital, in particular,
sector and the government. Within the former,
in
to reduce the costs
in
Proposals of this
between the
private
some proposals the contingent funds
32
come
primarily from existing stakeholders (e.g., through contingent debt/equity swaps)
while
others the funds
in
problems
extreme events,
a point highlighted
theory and AIG (and the monolines)
Flannery's (2002) proposal
his
outsiders.
However, outsiders' commitments
the extent to which the private sector can serve as the source of this
limit
capital during
come from
in
made one
by Holmstrom and Tirole (1998)
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
One problem
of this early proposal
idiosyncratic shocks.
calls for
The Kashyap
equity
a firm's
that
is
et
al
it
falls
made no
below
a certain
distinction
threshold.
between aggregate and
(2008) proposal deals with this distinction and
banks to buy capital insurance policies that pay off when the banking sector
experiences a negative systemic shock. Private investors would underwrite the policies
and place the amount insured
into a "lock box" invested in
US Treasuries. Investors
who
are themselves subject to capital requirements would not be allowed to supply this
insurance. The insurance would be triggered
when aggregate bank
number
amount;
of quarters exceed
not be included
in
some
significant
determining whether the insurance
losses over a certain
losses at the covered
is
bank would
triggered.
Combining both contributions, the Squam Lake Working Group on Financial Regulation
has
a
proposal (2009) similar to Flannery'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
is
for the insurance policy to
regulator, instead of the firm.
Under
amount
would be proportional to an estimate
risk?
of insurance required
posed by the bank,
systemic
in
this proposal,
pay out to the
by Acharya and others (2009), the
of the systemic
order to discourage firms ex-ante from taking on excessive
risk.
33
Hart and Zingales (2009) advocate an alternative approach;
CDS
rise
above
issue equity
in
its
CEO
is
it
at risk.
in
of time to
the bank
is
Although
If
the regulator determines the bank's debt
is
not at
the bank by lending to the bank; otherwise, the regulator
with a trustee,
the bondholders.
If
a bank's
CDS spread, the regulator reviews the bank's books and determines
the regulator invests
component,
window
order to bring the CDS spread back below the threshold.
whether the bank's debt
replaces the
allows the bank a
a certain threshold, a regulator
unable to reduce
risk,
when spreads on
who
approach does have
this
also relies heavily
the bank and pass the proceeds to
will liquidate
on the resolution of
a
contingent capital-injection
financial firms,
which can be
a
useful disciplinary device during normal times but can be highly counterproductive
during a systemic episode.
More
the
generally, the contingent capital approach
crisis
a central
may
is
mostly one of fundamentals. However,
feature of most financial crises, then
most
is
it
if
is
likely to
restore stability
the panic component
is
significant,
not the most cost-effective, and
it
well trigger further panic as fear of dilution and forced conversion increases.
This takes us to the second set of proposals, contingent insurance injections.
idea of this approach
is
that the pure panic
costly capital injection to subside. All that
will
when
is
component
needed
be available should conditions worsen. Despite
cost of such a policy
is
low because
underlies the panic. That
is,
the
it
derives
enormous
its
is
its
a
The basic
of a crisis does not require a
broad guarantee that resources
high notional value, the expected
power from the very same feature that
distortion
in
perceived probabilities of a
catastrophe also means that economic agents greatly overvalue public insurance and
guarantees. Providing these can be as effective as capital injections
in
dealing with the
panic at a fraction of the expected cost (when assessed at reasonable rather than panicdriven probabilities of a catastrophe).
In
Caballero and Krishnamurthy (2007)
uncertainty, a
we showed
that during an episode of Knightian
government concerned with the aggregate
will
want
to provide insurance
34
against extreme events even
sector.
The reason
fears itself to be
is
in a
it
has no informational advantage over the private
that during a panic of this kind, each individual bank and investor
worse than the average, an event that cannot be true for
situation
the collective. By providing
to react
if
broad guarantee, the government leads the private sector
a
more than one-for-one
since
it
also closes the
gap between the true average
and the average of panic-driven expectations.
During the current
argument
panic
is
for
why
crises,
it
developed
may
there were
many
Krishnamurthy (2008b).
institutions face a constraint such that value-at-risk
equity.
In
normal times,
these are "multiplied"
28
The
be optimal to support assets rather than inject capital during a
Caballero and
in
asset-insurance injection proposals.
this structure
many
times
in
must be
less
practice,
In
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.
In
and complex
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
this
is
the government can pledge a
capital (Caballero 2009a).
same
principle of insurance-injection can be used to
the chosen solution. Rather than directly injecting capital,
minimum
This
future price guarantee for newly privately raised
mechanism
is
very powerful because private investors
overvalue the guarantee, and because the recapitalization
event
that
less likely. Caballero
once the equilibrium
mechanism
significantly
and Kurlat (2009a) quantified
response
of
equity
prices
itself
this
is
makes
a
catastrophic
mechanism and showed
taken
into
account,
reduces the effective exposure of government
this
resources
relative to a public equity injection.
2S
See, e.g., Caballero (2009a,b), Mehrling
and Milne (2008) and Milne (2009)
for proposals.
35
Many
of the actual programs
implemented during the
had elements of guarantees
crisis
rather than being pure capital injections. Perhaps the clearest case of this approach
announced
that followed by the UK. Their asset protection scheme,
provided insurance against 90 percent of losses above
portfolios of corporate
a
"first-loss"
is
provided
in
exchange
for a fee.
RBS and Lloyds Banking Group, with
respectively.
The main
January 2009,
The APS covered £552
a first-loss
criticism to the U.K.'s
CDO
amount
approach
In
billion portfolios
of
of £19.5 billion and £25 billion,
is
that they charged such a high
economy
to their failure than socially optimal.
we
Caballero and Kurlat (2009b)
guarantee access to insurance
flexible
The
obligations.
fee for the insurance that most banks chose not to engage, leaving the overall
more exposed
on
threshold
and leveraged loans, commercial and residential property loans,
and structured credit assets such as RMBS, CMBS, CLO, and
insurance
in
is
manner
discussed a policy framework which would not only
the event of an SFA episode, but
in
that integrates the role of the
government
it
would do so
in a
as an insurer of last resort
with private sector information on the optimal allocation of contingent insurance.
In
that framework, the
government would
would be purchased by
financial
issue tradable insurance credits (TICs)
institutions,
some
of
which
which would have minimum
each TIC would entitle
its
holder to
attach a government guarantee to newly-issued and legacy securities.
All
regulated
holding requirements. During a systemic
financial institutions
would be allowed
crisis,
to hold
funds, corporations, and possibly hedge funds.
flexible
and use
In
TICs, as well as private equity
principle, TICs could be
and readily available substitute for many of the
during the
crisis.
that
were created
The basic mechanism would consist of attaching them to
variants could include attaching
particular
facilities
them
needs of the distressed
assets, but
to liabilities and even equity, depending on the
institutions
as collateral-enhancers for discount
used as a
and markets, and they could also operate
window borrowing.
36
TICs are equivalent to
CDS during systemic
TICs are contingent-CDS. They
become
normal times. That
crises but not during
activated only
when
a systemic crisis arises.
is,
By
targeting the event that needs protection, this contingent feature significantly lowers
the cost of insurance for financial institutions and therefore raises the
protection they can be required to hold for unit of systemic
Note also that
TICs' tradability
system
would
be
Lehman
better
misbehaving institution
To conclude,
is
to
V.
remove
of the
Final
most
fails,
did
and
failed), at
protected
as they
against
dire need.
And
if
distressed
risk their survival for a
higher
the very least the rest of the financial
the
turmoil
would be holding the
that
could
arise
if
the
TICs.
important to highlight that the point of these insurance arrangements
it is
(for a fee)
institutions, while
many
in
chose to not seek to stock up on TICs and
return (as probably
they choose to hold.
would allow private agents to use markets to reallocate
the access to insurance toward financial institutions
institutions
risk
amount of
the systemic
risk
from leveraged and interconnected financial
they continue to produce the triple-A assets whose shortage
main global macroeconomic phenomena of the
last
is
behind
two decades.
Remarks
The world entered the current
emerged from
it
financial crisis with a shortage of safe financial assets
with an even more acute
deficit.
The
crisis itself
was the
and
result of the
collapse of the financial industry created to bridge the original gap, and of the severe
panic caused by the chaotic unraveling of this complex industry.
with this dual problem
—
the shortage of safe assets and the financial
by the private sector's solutions to
absorb
a larger
resort,
and not
it
share of the systemic
would include modifying
One approach
—
is
risk
their portfolios
for
fragility
governments around the world
(and be compensated for
it).
to deal
created
explicitly
This approach
and becoming the provider of insurance of lost
just the lender of last resort, for
widespread panics.
37
Bibliography
Acharya,
Viral
and
V.
"Do
2009.
Schnabl.
Philipp
Banks
Global
Spread
Imbalances? The Case of Asset-Backed Commercial Paper During the Financial
Global
Crisis of
2007-09." November.
Acharya, Viral V. and Philipp Schnabl. 2009.
Chapter 2
Stability:
Acharya, Viral V. and
in
How
John Wiley
&
"How Banks
Played the Leverage 'Game'",
Matthew Richardson,
to Repair a Failed System,
New
eds.,
Restoring Financial
York University Stern School of Business,
Sons.
Acharya, Viral
V.,
Lasse Pedersen,
"Regulating Systemic Risk."
Thomas
NYU mimeo,
Philippon and
Matthew Richardson. 2009.
January.
Adelino, Manuel. 2009. "Do Investors Rely Only on Ratings? The Case of Mortgage-
Backed
Scurities."
MIT mimeo, November.
Adrian, Tobias and
Hyun Song
Shin. 2009.
"Money,
American Economic Review Papers and Proceedings,
Liquidity,
Brunnermeier, Markus
Press, NY.
and Lasse Heje Pedersen. 2009. "Deciphering the
K.
Credit Crunch 2007-08," Journal of Economic Perspectives, Vol. 23, No.
Caballero, Ricardo
Crisis ."
2009a.
J.,
"
Dow
Boost and
Policy."
Vol. 99, No. 2.
CDS Delivery Options. Bloomberg
Boberski, David. 2009.
and Monetary
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
The
of Money:
Role
J.,
2006.
"
On the
Macroeconomics
Money and Monetary
Policy
in
of
the
Asset
Shortafies ."
Twenty-First
In
Century
38
The Fourth European Central Banking Conference 9-10 November, Andreas Beyer and
Lucrezia Reichlin, editors. Pages 272-283.
Caballero, Ricardo
2003.
J.,
"
The Future of the IMF ." American Economic Review, Papers
and Proceeding, 93(2), 31-38, May.
Caballero, Ricardo
Crash,
Commodity
Activity, Fall,
Emmanuel
J.,
Prices,
Farhi
and Pierre-Olivier Gourinchas, 2008a.
"
Financial
and Global Imbalances ." Brookings Papers on Economic
pp 1-55.
Caballero,
Ricardo
Equilibrium
Model
J.,
Emmanuel
Farhi
of "Global Imbalances"
and Pierre-Olivier Gourinchas, 2008b.
"
An
and Low Interest Rates ." American Economic
Review, 98:1, pgs 358-393.
Caballero, Ricardo
Fragility, "
and Arvind Krishnamurthy, 2009.
J.
"
Global Imbalances and Financial
American Economic Review, May.
Caballero, Ricardo
a Flight to Quality
and Arvind Krishnamurthy, 2008a.
J.
"
Collective Risk
Episode ." Journal of Finance, Vol. 63, Issue
Caballero, 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,
Caballero,
ManaRement
November
its
24.
Bubbles and Capital Flow
Causes and Risk Management ." Journal Monetary of Economics, 53(1), pp. 35-
53, January.
Caballero, Ricardo
J.
and Arvind Krishnamurthy, 2001.
Collateral Constraints in a
Model
of Emerging
Market
"
International and Domestic
Crises ." Journal of
Monetary
Economics 48(3), 513-548, December
Caballero, Ricardo
J.
and Pablo
Kurlat, 2009.
"
Public-Private Partnerships for Liquidity
Provison ." MIT mimeo, March.
39
Caballero, Ricardo
Remarks and
Caballero,
and Pablo
J.
a Literature
Ricardo
Flight to Quality
and Bailouts: Policy
Review ." MIT mimeo, October.
and
J.
"
Kurlat, 2008.
Stavros
Panageas,
"
2008.
Hedging
Sudden Stops and
Precautionary Contractions ." Journal of Development Economics, 85 pp 28-57.
Caballero, Ricardo
J.
and Stavros Panageas, 2007.
Sudden Stops and External
Caballero, Ricardo
J.
Liquidity
"
A Global
Equilibrium
Model of
Management." MIT mimeo, September.
"
and Alp Simsek, 2009a.
Complexity and Financial Panics ." MIT
mimeo, June.
Caballero, Ricardo
mimeo,
J.
and Alp Simsek, 2009b.
"
Fire Sales in a
Model
of Complexity ."
MIT
July.
Calvo, Guillermo A. 2005. "Crises
in
Emerging Market Economies: A Global Perspective,"
NBER Working Paper No. 11305.
Calvo, Guillermo A.
and
Economic Collpase
Latin
in
Talvi,
Ernesto. 2005. "Sudden Stop,
Financial
Factors
and
America: Learning from Argentina and Chile," NBER Working
Paper No. W11153.
Calvo, Guillermo A., Izquierdo, Alejandro and Talvi, Ernesto. 2006.
Phoenix Miracles
Caplin,
in
Emerging Markets," American Economic Review 96(2),
Andrew and John
May
Leahy, 2004. "Business as Usual, Market Crashes and
after the Fact,"
American Economic Review, pp. 548-565.
Coval, Joshua
D.,
Bonds."
"Sudden Stops and
Jakub W. Jurek and
Erik
Stafford.
2006.
Wisdom
2008. "Economic Catastrophe
HBS Finance Working Paper No. 07-102.
Coval, Joshua D., Jakub
W. Jurek and
Erik Stafford.
2009. "The Economics of Structured
Finance." Journal of Economic Perspectives, Vol. 23, No.
1,
Winter.
40
and
Coval, Joshua,
Erik Stafford.
2007. "Asset Fire Sales (and Purchases)
Markets," Journal of Financial Economics, Vol. 86, No.
2,
in
Equity
November.
Mark, 2002, "No Pain, No Gain? Effecting Market Discipline via 'Reverse
Flannery,
Convertible Debentures'," Working Paper, November.
and Lawrence Summers. 2009. "A
Geithner, Timothy,
Washington
New
Financial
Foundation."
Post, June 15
Gourinchas, Pierre-Olivier and Helene Rey. 2007. "From World Banker to World Venture
US External Adjustment and The Exorbitant
Capitalist:
editor,
G7
Privilege," in
Richard Clarida,
Current Account Imbalances: Sustainability and Adjustment, The University of
Chicago Press, pp. 11-55.
Gorton, Gary
2008. "The Panic of 2007."
B.
Financial System. Proceedings of the
Bank of Kansas
Gorton, Gary
Gorton,
in
a Changing
2008 Jackson Hole Conference, Federal Reserve
City.
and Andrew Metrick. 2009. "Haircuts." NBER
B.,
Gary
Maintaining the Stability
In
B.,
and
Nicholas
Souleles.
2006.
WP #15273,
"Special
Purpose
Securitization." In The Risks of Financial Institutions, edited by
Rene
August.
Vehicles
Stulz
and
and Mark
Carey. University of Chicago Press.
Hart, Oliver
Institutions."
and
Luigi Zingales.
2009. "A
New
Capital Regulation for Large Financial
Harvard mimeo. September
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
Institute.
in
Symposium sponsored by The
Wyoming.
2009. Report of the
Money Market Working Group.
41
Monetary Fund. 2009. Global Financial
International
Stability Report:
Navigating the
Financial Challenges Ahead, October.
Kashyap, Anil
Regulation,"
K.,
Raghuram
G. Rajan
and Jeremy
Maintaining Stability
in
in
C. Stein.
2008. "Rethinking Capital
a Changing Financial System,
Symposium
sponsored by The Federal Reserve of Kansas City at Jackson Hole, Wyoming.
Krishnamurthy, Arvind, and Annette Vissing-Jorgenses, 2007. "The Aggregate
for Treasury Debt," Kellogg
Lehman
Demand
mimeo.
Brothers. 2008. "Residential Credit Losses
-
Going into Extra Innings?"
Mehrling, Perry and Alistair Milne, 20008. "Government's role as credit insurer of
resort and
how
it
can be
fulfilled,"
mimeo
Milne, Alistair. 2009. The Fall of the
Cambridge and
New
last
October.
House of Credit. Cambridge University
Press,
York.
Office of the Special Inspector General for the Troubled Asset Relief Program. 2009.
Quarterly Report to Congress, July.
Wayne W.
Roxburgh, Charles, Susan Lund, Charles Atkins, Stanislas Belot,
Pierce. 2009. "Global Capital Markets: Entering a
New
Era,"
McKinsey Global
report.
September.
Squam
Lake Working Group on Financial Regulation. 2009.
Mechanism
Tett, Gillian.
Hu, Moira
"
S.
Institute
An Expedited Resolution
for Distressed Financial Firms: Regulatory Hybrid Securities ." April.
2009. Fool's Gold:
How
Unrestrained Greed Corrupted a Dream, Shattered
Global Markets and Unleashed a Catastrophe.
Little,
Woodford, Michael. 1990. "Public Debt as Private
Brown.
Liquidity,"
The American Economic
Review, Vol. 80, No. 2 (May), pp. 382-88.
42