Massachusetts Institute of Technology
Department of Economics
Working Paper Series
and Bailouts: Policy Remarks and
a Literature Review
Flight to Quality
Ricardo
J.
Caballero
Pablo Kurlat
Working Paper 08-21
October
9,
2008
RoomE52-251
50 Memorial Drive
02142
Cambridge,
MA
paper can be downloaded without charge from the
Social Science Research Network Paper Collection at
This
httsp ://ssm.com/abstract= 1 29745 6
.
Flight to Quality
Policy
and
Bailouts:
Remarks and a Literature Review
Ricardo
J.
Caballero and Pablo Kurlat^
October
9,
2008
Abstract
Flight to quality
weal<nesses
in
episodes involve a combination of extreme
may
in
US
or uncertainty- aversion,
the balance sheets of key financial intermediaries, and strategic or speculative behavior,
that increases credit spreads on
entire
risk-
financial
system
is
all
but the safest and most liquid assets. Unlike previous episodes, the
currently at the center of the trouble, with no safe haven pockets, w/hich
lead to greater real effects. The
US government's
credit
is still
impeccable, w/hich facilitates policies
support of the financial system. Policy must take into account incentives for behavior during the
crisis,
discouraging excessive prudence, w/hich sometimes implies relegating post-crisis moral hazard concerns
to a secondary role.
JEL Codes:
E44,G14,G21
Keywords: subprime
^
MIT and NBER, and MIT;
crisis, liquidity,
bailout, intermediation, credit spreads
respectively. Prepared for the
Meeting on Financial
Stability
the Financial and Real Sectors, Federal Reserve Board of Governors, Washington
D.C.,
and the Linkages between
October
3,
2008
Digitized by the Internet Archive
in
2011 with funding from
MIT
Libraries
http://www.archive.org/details/flighttoqualityb0821caba
Introduction
The term
"flight to quality"
is
used to describe an environment where investors seek to
sell
assets perceived as risky and purchase safe assets instead, leading to widening risk premia and severe
disruptions
credit
in
and other
extremely severe episode of
Financial
2007 that has
Markets
all
accounts, the U.S.
the U.S. are struggling with a chronic
perceived as "unsafe;" the ensuing
to zero and triggered record setting spikes
gold and
in
September
flight to quality
Relative to the rest of the world, the U.S.
episodes for at least three reasons:
risk,
First,
is
significantly
the U.S. as a whole
in
15'^ of this year,
even
brought treasury rates
more
in
world history.
resilient to flight to quality
perceived as a "safe haven," and hence
and banks
in particular,
are net recipients
search of "quality." This good correlation between leverage and the direction of funds
key ingredient for financial
where the banking sector
stability,
is
was
the U.S. typically sees sustained and stabilizing net capital inflows toward
safest assets. Second, the highly leveraged financial sector,
of funds
is
problem since mid
Policymakers, realizing that time
oil.
running out, reacted by announcing the largest financial "bailout" package
its
currently experiencing an
flight to quality
oscillated in intensity but with ever increasing peaks. By
while investors run from
is
this kind.
in
money market funds were
down
By
financial markets.
the
first
which
is
absent
in
many
is
a
other financial systems around the world
and partly due to the
victim of financial panics. Third,
first
reason,
the US has a flexible economic policy framework.
However, one of the main reasons the current episode has become so severe
stabilizing
mechanism has vanished. While deposits have followed the standard
forms of funding are leaving the financial system
financial
over
in
is
that this second
pattern, almost
herds. This process has dried up liquidity
markets and has strained banks' balance sheets from plummeting asset values and hard
The system was simply not ready
liabilities.
triggering
all
manner
for such a dramatic
rise in
Knightian uncertainty. The
first
purpose of
this article
is
in
key
to-roll-
correlations,
in
endogenous tightening
of amplification mechanisms, ranging from
requirements to a sudden
turnaround
other
all
in
margin
to survey the
empirical and theoretical literature describing these mechanisms.
The second purpose of
mechanisms.
In
rather obvious:
to discuss the policy implications of flight to quality
is
brief terms, the general policy
When
last resort facilities
sound and the
this article
flight to quality
is
is
particularly the case
mechanism mentioned above -the
when government instruments
rise in foreign capital
treasuries remains active, as the cost of capital to finance the intervention
to quality
itself (i.e.,
these times
is
there
to transfer
is
sharp
some
existing literature
is
severe, predictable and reliable systemic bailouts and lender of
are highly desirable. This
first stabilizing
message that emerges from the
rise in
demand
for Treasury
Bills
is
brought
are
inflows toward
down
by the
flight
and Bonds). The goal of policy
of the government's liquidity, collateral
and
trust, to
at
the distressed
domestic financial sector. Tiie
last
section of the paper summarizes
some
of these policy implications
and adds a few conjectures.
There
however,
is,
way
a less "linear"
to present the policy results
from the existing theories, and from what these miss. This
of the
common
article
is
some
by providing a
critical
some
description of
features of current bailout proposals. Given the nature of the meeting for which this
being prepared,
we do
this introduction, while others
of
is
and conjectures that flow
this directly here.
may wish
Most readers may wish
to continue
and
to read only up to the end of
paper a brief description
find in the rest of the
much
of
common
a
of the articles that provide the backbone (but not the entire skeleton) for
what
is
said
here.
Most proposals,
one
including the
principles: First, they recognize the
need to
just
approved by Congress, have
recapitalize the financial
of several key asset and insurance markets. Second, there
is
most see moral hazard
at this
time
is
likely
first
two
"principles" but are less
the standard moral hazard view
real
life,
unlike
in
many
when and
if
costly. Speculators
to stabilize
it.
to
let
and
is its
make
all
crisis in
disregard for the incentive problems
types of of strategic decisions within a
go of their assets, knowing that
strategic players have to decide
the
to punish
in particular,
first
likely
as
it
would
place.
generates within
it
crisis.
a miscalculation
when
crises. In
Distressed agents have to
on the
to reinforce a
right timing
downward
tempo may
can be very
spiral,
to wait before intervening, fully
and when
aware that
blown
crisis
of predicting
what
require that a
full
in
the
game
the likelihood of a bailout and the form
it is
expected to take change
for bickering to be put aside. Each of these agents
likely to do. In particular,
the incentives for both distressed firms and speculators within the
both to the resolution of the current
A standard
of us think that almost
more
future crises
caused the
delaying can be counterproductive, but that the political
others are
many
persuaded by the third one. The main problem of
Governments have to decide how long
becomes observable
liquidity
of our models, crises are not an instant but a time period. This time dimension
creates ample opportunity for
decide
In fact,
as a reason to limit the extent of the intervention and,
financial sector to repeat the excesses that
We share the
system and to improve the
to yield a large excess return to the government. Third,
shareholders. Not doing so, the argument goes, would
encourage the
few general
an agreement on the need to protect
taxpayers by giving the government a share of the upside as well.
any reasonable intervention
in
crisis
is
crisis.
These incentives are central,
as well as for the severity of the next crises.
advice stemming from the moral hazard
camp
is
to subject shareholders to
exemplary punishnrient (the words used by Treasury Secretary Paulson during the Bear Stearns
intervention). This
dimension,
all
is
sound advice
in
the absence of a time dimension within crises. With no time
shareholders were part of the
takes place the
crisis is
boom
over; the next concern
is
that preceded the
crisis
and as soon as the bailout
not to repeat the excesses that led to the
crisis.
means punishing those
Punishing shareholders
learn the lesson
sooner rather than
However,
this advice
less
crisis,
and
it
is
better that they
later.
can backfire
when we add back
that shareholders will be exemplarily punished
needed
that led to the current
if
the
crisis
the time dimension. Now, the expectation
worsens delays
in
the decision to inject
As a concrete example, sovereign wealth funds are
capital by stabilizing investors.
much
now much
eager to inject equity into the U.S. financial system than they once were. Conversely, destabilizing
speculators and shortsellers see the value of their strategy reinforced by the policy of exemplary
punishment. For both reasons, crises become more acute, as the equity market becomes extremely onesided
when
uncertainty and
turns into a
during bouts of panic and confusion. The anti-moral hazard strategy
enzyme.
crisis
This
risk rise
theme
also highlights an aspect that
being minimized
is
component. One of the puzzling behaviors during the current
remained on the
side, despite
obvious
fire sales.
crisis
is
in
current proposals: the strategic
why
so
much informed
capital has
For relatively small investors, or those constrained by
regulation on the size of their position, the fear of exemplary punishment should another bout of panic
take place
most
likely
may be
the answer. However, for large investors,,
answer
is
In this
role,
which
is
Of course, there
may
crisis. In
is
crisis
such as a thing as too
reaction complicates the crisis resolution as
its
markets become
is
illiquid
to wait or pull out
such a context, a potential bailout plays yet
and hence to reduce the predator's
much
it
If
of a
good
thing, as the anticipation of a bailout
the anticipation turns out to be incorrect, this
delays external capital injections until
refusal to accept an offer by J.C. Flowers the
example often used to support
rises,
to unfold.
increase the incentive to wait by the seller.
The case of AIG and
uncertainty
to increase the perception of competition
incentive to wait for the full-blown
the potential to gain control, the
context, the optimal strategy of the predators
resources from the target, waiting for a deeper
another
When
predatory and strategic behavior.
and market power develops.
who have
it
may be too
weekend before
However, one should note that
its
demise
late.
is
an
problem caused by
this
is
a
the unpredictability of policy, not by the predictable component of this policy.
If
the government has
ample access to
liquidity,
this position.
and balance sheets are being destroyed by the reinforcing feedback of acute
mispricing and predatory actions,
"good moral hazard," as
it
it is
important that the private sector can count on this
would ultimately be too
liquidity. That's
costly for the private sector to hoard liquidity for
such episodes.
Along the same vein, when Knightian uncertainty
too much private
risk taking.
Within the
crisis,
is
prevalent, the main problem
markets are on the other side of
risk
is
too
little
taking relative to the
conventional moral hazard concern, and hence inducing a more aggressive use of private liquidity
positive rather than a negative development. That
is,
while
it
is
not
is
a
true that excessive risk taking prior to
the
crisis
can be a source of trouble, once a
crisis is
reached, the greater concern
is
insufficient risk
taking and explicit public support can encourage rather than prevent desirable private sector behavior.
These concerns lead to several observations regarding how the
many
details of the bailout,
of
which are yet to be determined, should be arranged. One objective must be to signal to signal the
strategic investors waiting
in
the sidelines that prices
will
stop
falling
and thus discourage speculative
waiting. Speculators will not expect that prices of securities will be lower than those established at the
Treasury's auction
(if
indeed an auction
is
used), at least
in
the period that immediately follows the
auction. Thus the date of the auction provides a clear deadline to any speculative waiting.
timetable for purchasing a given
list
of securities
may
Announcing
a
therefore have a salutary effect on prices even
before the actual purchases take place.
To the extent
of mispricing
possible, the first securities to be purchased should be those
greatest.
is
For instance, certain AAA-rated tranches of subprime mortgage backed
securities
have been trading at prices that are hard to
market.
these securities were
If
first
on the Treasury's
the possible gains from speculative waiting
One
risk
is
that
lead to fire-sale prices
if
some
when
amount
distress will have
on auction
in
shareholders.
this
would send
the Treasury purchases the securities. To
this purchase.
would harm other security-holders,
sufficiently large
equity stake
list,
except by the extreme
illiquidity of
the
a signal to speculators that
of the holders of a particular security are especially distressed, this
market their remaining holdings. One way to
Finally,
justify
soon disappear.
will
by the profits the taxpayers would make on
excessively low prices
where the evidence
Still,
if
some
there
is
a
extent, this risk
is
may
mitigated
concern that purchases at
nothing else, from having to mark-to-
partially avoid this situation
is
to
commit
to purchasing a
of each security to minimize the impact that any particular security-holder's
prices.
the Treasury's plan contains as-yet-unclear provisions for giving the government an
the companies
One way
it
assists.
Presumably
this will involve diluting
the holdings of current
to take into account the within-crisis incentives this policy generates
would be to
give special consideration (for instance, lower dilution) to firms that raised fresh capital since the start of
the
crisis,
and to those that experienced the most extreme predatory attacks which cannot be
justified
on the grounds of fundamentals.
To be
Instead, our
clear,
our position
argument
is
that
it
is
not that the standard moral hazard concerns should be disregarded.
is
important that
mindful of the perverse incentives that they
approach
is
one example of
a
may
when
designing policies to address
trigger within crises.
misguided policy along these
been another one, but there are many
hazard without backfiring during the
crisis.
lines, letting
post-crisis regulatory
it,
we
are
more
The "exemplary punishment"
Lehman Brothers
fail
may have
responses that could deal with moral
Evidence
A
defining feature of a flight to quality
Depending on the
elasticity of supply, this
may
is
decrease
a
materialize as a
the relative dennand for risky assets.
in
supplied, or in both.^ For already-issued securities, which at least
flight to quality
may respond
show/s up as a change
in relative prices.
their relative price, in the quantity
fall in
in
the short run are
in
fixed supply, a
For newly issued securities or loans, quantities
as well. Accordingly, different studies have focused on either quantities or prices
when
exploring evidence on flights to quality.
Obviously there are more than two possible levels of riskiness ("risky" and "safe") and a
quality
may
may
involve shifts
in relative
behavior toward any group or subgroup of assets. For instance,
involve a preference for bonds over stocks, for
corporate bonds, or
many
AAA bonds
it
over junk bonds, for treasuries over
of these at once. Different studies have looked at the evidence on the basis of
one or several of these pairings of
Finding 1: The
flight to
assets.
amount and composition of banl< loans
reveal that flight to quality
is
countercyclical. This
pattern exacerbates the business cycle by depriving lower quality borrowers from financial resources
during contractions.
A
series of early studies focused
financing varied with
on quantities, examining how the composition of
macroeconomic conditions. Kashyap,
tightening of monetary policy there
Stein,
were systematic increases
paper compared to bank lending. The underlying view
is
&
in
firms' external
Wilcox (1993) found that following
the relative quantity of commercial
that large firms have access to a commercial
paper market whereas small firms are more dependent on bank lending, perhaps because bank
monitoring
essential to
is
overcome informational asymmetries. Thus,
a relative reduction in
bank
lending can be interpreted as a flight to quality. Similarly, Gertler and Gilchrist (1993) and Oliner and
Rudebusch (1993) found that the
relative proportion of loans to large corporations increases in these
episodes, and Lang and
Nakamura (1995) found that the
prime+1% (which they
interpret as relatively safe)
monetary
policy.
fluctuations
as
much
is
at rates
below
countercyclical and rises after tightenings of
Commerce's Quarterly
Financial Report of manufacturing firms,
Furthermore, they compare the
a greater proportion of small (manufacturing) firms.
the growth rates of sales, inventories, and debt
in
small and large firms and conclude that
The broad conclusion of these studies
is
that financial
The terms supply and demand are subject to some ambiguity. A decrease
claims
made
as one third of aggregate fluctuations could be explained by the differences
large firms.
^
in
loans
Bernanke, Gertler, and Gilchrist (1996) offer evidence consistent with these findings
using data from the Department of
which includes
is
new
fraction of
customarily described as a decrease
in
demand
in
in
between small and
constraints for lower-quality
investors' desires to acquire risky
the context of securities, and a decrease
the context of bank loans. This essay follows this customary use unless there
is
risk
of confusion.
in
supply
in
y,
'
1,
borrowers tighten
real
in
periods of recession or tight
money and
that they have quantitatively important
consequences.
Finding 2: During nnild flight to quality episodes, funds flow toward banks. However, banks' safe haven
status
weakens for severe flights
to quality,
where only the safest assets experience
Gatev and Strahan (2006) study the other side of banks' balance sheets.
they find that
when
the spread between Treasury
bills
Inflows.
In
data for 1988-2002,
and high grade commercial paper Increases,
banks (but not other financial intermediaries) tend to experience inflows of deposits and a decreased
cost of funding. This suggests that banks tend to be seen as safe havens
in
periods of turmoil. Gatev and
Strahan attribute this advantage to implicit government backing and suggest that this
why banks
reasons
U.S.
for 1998,
banks grew 2.5%
in
the
flights
shift
their
toward banks. Nontransactional deposits
the second half of the year, whereas they had grown 5.2%
CD
in
toward banks rather than from banks.
however, show a very small
Furthermore, although LIBOR and
widened
one of the
are better placed than other institutions to offer liquidity insurance. At least
sample, flights to quality seem to have been
The data
is
in
the
in
first half.
rates decreased, their spreads with respect to Treasury rates
substantially, as did other indicators of risk such as the VIX. Overall, the evidence suggests that
flight to quality also
involved a worsening
in
the relative position of banks compared to the very
safest assets. As the figures show, however, the episode
both spreads and the VIX were within a normal range.
was very short
lived.
By the end of the year
6.5
LIBOR
3
month CD
3
month CP
rate
^^^ 3 month Treasury
V.
5.5
4.5
3.5
Jan98 Jan98 Mar 98 Apr 98 May 98 Jun 98
Figure
1.
After the Russian default ttiere
Jul98
was a decrease
Aug98Sep98 Oct98 Nov 98 Dec 98
in interest rates,
Jan 99
including interest rates paid by
banks on CDs and LIBOR
Figure
term
2.
AAA
The "TED spread" increased, as did
bonds. There was
tiie
some degree of flight
spread on high grade commercial paper and on long
to guality even within the safest assets. This lasted a
very short time.
50
•VIX
45
40
35
30
0)
25
20
Russian default
15
10
5
Jan98 Jan98 Mar98 Apr98 May98 Jun98
Figure
3.
Aug 98 Sep98 Oct98 Nov98 Dec98 Jan99
Jul98
Other indicators of risl(, such as the
VIX,
increased sharply but
briefly.
Both stock market crashes and large reallocation of funds towards bonds are
Finding 3:
increases the likelihood of the other. The
US bond market
also serves as a safe
but each
rare,
haven for international
equity market crashes.
Both
Bemanke
et. al.
times and use the expression
in
and Gatev and Strahan study long data
"flight to quality" in studies of systematic,
the availability of financing for different kinds of firms.
they describe constitutes a
flight to quality in
environment use. Other studies, looking
that explicitly refers to
crashes
in
stock
more extreme
the
bond
and bond markets
between
but not necessarily large, shifts
not clear, however, that the
in
than quantities, have focused on a definition
G5 countries (between 1987 and 1999) tend
flight
once every 30
different markets at these
phenomenon
that observers of the current financial
to
quality
a crash as
pattern,
years).
an episode where there
They
whether
to
occur
with stock market crashes
is
stock prices (estimated to occur once every 39 years) or a drop of
prices (estimated to occur
"normal"
events. Hartmann, Straetmans, and de Vries (2004) study
accompanied by bond market booms. They define
in
It is
same sense
at prices rather
simultaneously or instead tend to follow a
more than 20%
series that include mostly
find evidence of
a
weekly drop of
more than 8%
in
strengthened linkages
extreme values, but these are approximately
just as likely to
go
in
either direction,
especially likely
i.e.
co-crashes are approximately as frequent as flights to quality. Flight to quality
toward the US bond market:
it
is
estimated that 4.6% of US, 7.9% of German, 7.7% of
French, 8.3% of UK, and 3.0% of Japanese stock market crashes
boom
coincide with a
will
lower, reflecting the greater perceived safety and liquidity of
for short
identifiable events) the correlation
and
this
this as
was
especially the case
evidence of a
in
much
US bonds. Gonzalo and Olmo (2005)
find a
It is
the Asian
find evidence that during crises periods (defined by
crisis
of 1997 and the Russian
flight to quality effect. Overall, this literature
news there
is
of 1998. They interpret
crisis
tends to confirm the view that
often a flight from risky assets
like
in
stocks to less
bonds.
worth emphasizing
that, unlike
to quality have not generally been
the typical emerging markets experience,
accompanied by generalized
in flights
accompanied the stock market crash of 1987, net foreign inflows
increased from 0.16% to 0.22% of US GDP;
increased from 0.31% to 0.38%, and
Finding 4: Flight to quality
is
in
in
the brief
in
capital flights, indeed,
case that non-residents are net purchasers of government bonds
quality that
of distress, but only
between stock and bond markets becomes stronger and negative,
periods of uncertainty or distress or bad
risky assets like
US bond
is
between stock returns and bond returns during periods
term bonds. Baur and Lucey (2008)
in
booms
markets. The likelihood of stock market crashes coinciding with non-US bond market
similar negative association
is
it
to quality.
into
the U.S. flights
the
In
flight to
US government bonds
before the Gulf war
flight to quality
usually the
is
1990, they
in
1998 from 0.06% to 0.15%^
perceptible even across nearly equivalent assets.
Stocks and bonds are coarse categories of assets and there
is
what can be learned by
a limit to
studying asset prices at that level of aggregation. Several studies have studied the behavior of
narrowly defined assets
order to determine exactly what happens during
in
more
flights to quality. Longstaff
(2004) studied the spreads between bonds issued by Refcorp and comparable US Treasury bonds.
Refcorp
is
a
US government agency whose
risk in its liabilities
is
liabilities
are guaranteed by the Treasury, so legally the credit
identical to that in Treasury bonds. Nevertheless, the yield
on average between 1991 and 2001, between 10 and 16
on Refcorp bonds was,
on Treasury bonds
basis points higher than
(depending on maturity). This difference accounted for as much as 10% to 15% of the value of
Treasuries. Longstaff labels this spread a flight-to-liquidity premium.
a series of time-varying
associated with
(a)
^
measures of investor sentiment. He finds that increases
drops
market mutual funds
He then regresses
in
consumer confidence,
(since these are
The comparisons are are the
against the third quarter and
last
one
the
amount
the
premium are
of funds held
of the safer asset classes, an inflow of funds
3 quarters of 1987
2"" half of
(b) increases in
in
premium on
this
in
money-
presumably
against the previous 3 quarters; the last quarter of 1990
1998 against the l"
half respectively, in
all
cases seasonally adjusted.
and
indicates a flight to quality")
their
premium tends
(c)
Treasury buy-backs (since these make Treasury bonds more scarce,
when comparing
to increase). Krishnamurthy (2002) finds similar results
the yields
on on-the-run and off-the-run Treasury bonds. He finds that higher spreads on off-the-run bonds are
associated with higher spreads between commercial paper and Treasuries.
Finding 5: The relative importance of liquidity over credit quality rises during flight to quality episodes.
In
both of these studies the authors identify premia that increase
be associated with credit quality but rather with
and hence
closely related
it
and Kavajecz (2008) address
direct
measures of the
it
is
liquidity are
that investors seek. Beber, Brandt,
question by taking advantage of the fact that the credit quality and
this
bonds from different Euro-area countries have
liquidity of
times of distress but do not seem to
Of course, credit quality and
liquidity.
hard to determine exactly what
is
in
credit quality of different countries
a slight negative correlation.^
They have
from spreads on credit default swaps and
construct measures of liquidity on the basis of data on bid-ask spreads and the depth of limit order
books. They find that both credit quality and liquidity are significant determinants of bond yields.
normal times, credit quality plays a greater
VSTOXX options
index), or
when
order flows), the importance of
Overall, the evidence
premium on
role;
however, at times of uncertainty
measured by the
there are flights into or out of the bond market (as measured by net
liquidity for explaining
the cross-sectional variation of yields increases.
from these studies suggests that
at times of uncertainty,
investors place a
assets that are not just safe but have very low transaction costs.
Finding 6: The macroeconomic cost of recent flight to quality episodes
although
(as
In
this
may
well be the result of aggressive
in
the U.S. has been limited,
and successful policy responses.
Asset pricing evidence indicates that from investors' point of view the possibility of flights to
quality
is
a real concern.
From
a policy point of
episodes lead to consequences
part
due to the
may be
in
what constitutes
associated with either a drop
in
the evidence
is
to
what extent these
tentative at best,
toward firms with more
solid
in
a flight to quality. Conceivably, a flight to quality
less
balance sheets. Bernanke, Gertler, and Gilchrist (1996)
in this
direction, but the
magnitude of the efficiency
due to misallocated investment cannot be ascertained.
In
the turmoil that followed the Russian default and the collapse of
1998, the U.S.
This statement
^'
this point
is
aggregate investment or a redirection of investment toward
and related studies contain suggestive evidence
losses
On
the real economy.
difficulty in defining
risky projects or
view the more important concern
economy
did not
credit qualities.
to suffer greatly. Real
may have become outdated
For instance, Italian bonds are
I
seem
among
GDP grew
since Longstaff' s paper
the most liquid
in
was
at
LTCM
in
the
summer
an annualized rate of 3.6%
of
in
written.
Europe despite the
fact that Italy has
one of the lowest
'.'<:
'
the
first
half of
resilience,
1998 and 5.4%
in
the second
half,
investment at 7.2% and 9.8% respectively. This
however, may be due to the relative brevity of the scare. As shown
back within a normal range by the end of the year after a sharp spike
increased their volume of loans by 5.0%
half),
in
haven status of the
What do
in
U.S. as a
these findings
tell
in
2,
spreads were
September. Furthermore, banks
issues of corporate debt
economy. The Fed's aggressive easing
by 75 basis points to 4.75%
Figure
the second half of the year (compared to 3.9%
which partly compensated the decrease
credit to the real
in
in
of
the second half of 1998)
monetary
may
whole further lowered the cost of
in
the
first
and preserved the flow of
policy (the target rate
was lowered
also have played a role. Finally, the safe
capital.
and nature of the current crisis ?
us about the severity
The current episode has many of the features of previous
flights to quality. Firstly,
there have
been sharp and opposite movements of bond and stock markets. The S&P 500 index was 27% lower
than
in
June 2007, while the Lehman Brothers Aggregate bond index was more than 10% higher. This
the kind of co-movement studied by Hartmann
with the pattern of previous episodes
monetary
policy) has led to sharp
(in
drops
1
et. al. (2004), albeit at a
lower frequency.
In
is
keeping
particular 1998), investors' flight to safe assets (plus the Fed's
in yields.
LIBOR
3
month CD
rate
----3 month CP
3 month Treasury
Mar 07
Dec 06
Figure
4.
Jun07
Safe interest rates have gone
down
also noted
flight
in
to quality
is
JunOS
Sep 08
since the beginning of the current turmoil in the
2007, as they did
The
Mar 08
Dec 07
Sep 07
in
summer of
1998.
evident even within traditionally very safe asset classes. This feature
previous episodes but
its
magnitude and persistence
is
far greater this time.
was
LIBOR spreads
over Treasuries have been between 100 and 200 basis points for most of the past year, reaching peal<s
of 280 basis points
that only lasted a
flight
in
September.
In
1998 the peak spread was approximately 160 basis points and even
few days. The force of the
flight
from banks
towards banks. This may yet translate into the
especially
if
real
the credit channels identified by Bernanke
is
proving stronger than the force of the
economy
et. al.
in a
way
that
(1996) react strongly.
was not seen
in
1998,
"•i:i|
'>.>
400
LIBOR spread
350
3
(b.p.)
month CP spread
(b.p.)
- -— - AAA 30-year spread
300
(b.p)
250
o
°-
200
150
100
50
Dec 06
Figure
5.
Mar 07
Jun07
Sep 07
Dec 07
Mar 08
JunOS
Sep 08
Spreads between Treasuries yields and other safe assets have remained persistently high for
over a year.
i
The VIX index, which was
more than
much more
threefold since then.
It
at comparatively
low levels at the beginning of 2007, has increased
has not reached the peak levels of 1998 but the increase has been
persistent, evidence of a lasting pullout
from
volatility.
70
•VIX
60
50
40
X
u
•a
c
30
20
10
Mar07
Dec 06
Figure
Jun07
6.
Dec 07
Sep 07
MarOS
JunOS
Sep 08
The VIX index has also remained high for over a year.
Theory
A number
of authors have attributed the existence of flights to quality to various institutional
features of financial markets, which lead to feedback effects between asset prices and investors'
preference for
liquidity.
Proposition 1:
When
asset price volatility
rises, illiquidity risk rises
and
this
feeds back into an increase
in
effective risk aversion.
Vayanos (2004) presents
times of market
a
increase
in
there
one safe asset and many
is
stochastic process.
The
volatility,
risky assets.
The
volatility of
premia
illiquidity
flight to quality. In
the risky assets
is
itself
may
the model,
an exogenous
each carries a different (exogenous and constant)
assumed to be fund managers whose incentives are governed by the
fee they are paid. This
is
proportional to assets under
management
liquidated, to the liquidation value of the fund (the fund will be liquidated
fixed threshold, as clients
and
risk
which are precisely the features of a
risky assets are illiquid in that
transaction cost. Investors are
management
model that may explain why both
withdraw their money).
When market
volatility
is
if
or,
if
performance
low,
the fund
falls
is
below a
managers are not very
concerned with withdrawals and therefore do not care about each asset's transaction
when
the volatility of the market
is
high, the probability that the fund's
threshold increases. This brings two effects.
value of each asset, so
illiquidity
First,
cost.
However,
falls
below the
performance
managers place greater weight on the
premia increase. Secondly, each
liquidation
risky asset's contribution to
and
likelihood of the fund's liquidation increases, so effective risk aversion increases
risk
the
premia
increase.
Proposition 2:
Agency problems
specialists' capital
is
Building on
pro-cyclical
amount of
limit the
and hence
Holmstrom and
uninformed investors can bear, which means
risk
triggers flight to quality during severe contractions.
Tirole's
(1997) analysis of the role of intermediary capital
in
connecting uninformed investors to projects. He and Krishnamurthy (2008) study a model related to
that of Vayanos. While the latter takes the fact that that poorly performing fund managers will face
withdrawals as a given feature of the environment, they instead model the relationship between
investment managers and investors
skills
explicitly. In their
model, investment
in a risky
provided by a specialist. Both specialists and uninformed investors contribute capital to an
how much each
intermediary institution (such as a bank) and write contracts that govern
as a function of the return on the investment portfolio, which
two
asset requires specific
limits to
what the
is
managed by the
investors can get the specialist to do on their behalf.
First,
to exert effort, which lowers expected returns. To motivate effort, investors must
specialists a function of realized returns.
chosen by the
specialist
(i.e.
what
fraction
a portfolio balance that optimizes his
party receives
There are
specialist.
the specialist
may
fail
make the payment
to
Second, investors cannot monitor the portfolio composition
is
own
invested
in
the risky asset). Hence the specialist
desired exposure to market
upper bound on the proportion of aggregate
risk
risk.
will
choose
These two forces place an
that non-specialist investors can bear:
If
they wish to
increase their exposure to risk they must persuade the specialist to increase the riskiness of the portfolio
by reducing sensitivity of
from shirking places
his
a limit
payments to
on how much they can do
such that specialists are willing to bear
wealth there
will
this
where intermediary
theory focuses on
how
why
this.
minimum
be a negative relation between
the model offers a rationale as to
periods
realized returns;
however, the need to prevent the
Therefore the equilibrium price of
proportion; since
specialists' capital
risk
aversion
and the price of
is
specialist
risk will
decreasing
risky assets.
in
Thus
effective risk aversion (and therefore risk premia) increases in
institutions
have suffered
delegation of investment
losses. In
common
management may
with Vayanos (2004), this
create flight to quality patterns
asset prices at times of distress. By modeling the structure of the agency relationship explicitly.
Krishnamurthy highlight the role of
be
in
He and
specialists' capital.
Proposition 3: The tightening of margin requirements during periods of high price volatility reinforces
flight to quality, as funds are reallocated from
more
to less volatile assets.
Krishnamurthy (2008) explores a
play a role
in flights
to quality.
In his
model there are two
(perhaps a bank) issues a claim at an
(limited) liquid funds.
pays a
fair price;
if
If
channel by which intermediaries' capital
slightly different
Initial
layers of intermediation.
date and commits to repurchase
few investors request repurchasing, the bank's
instead
funds of the bank. Investors
many
ask for repurchasing, the price
The feedback mechanism works
as follows: Since
in
try to resell the asset to the bank. Since
funds are sufficient and
governed by the amount of
meet margin
some
will
that the inefficient equilibrium here
is
work through market
model by Brunnermeier and Pedersen
in
supply and
decrease
in
the
Thus they
will
like
failure.
Diamond and
The difference
know the fundamental
(2008). In
demand and
prices.
it,
Margin requirements play
speculators buy and
their
own
value of the assets and therefore
overall riskiness
value at
may
risk. Crucially,
financiers
fundamentals. Furthermore, this implies that
liquidity
flights to quality will
Proposition 4:
When
risk)
as an
liquidity spirals as
and prices deviate more from
occur since the assets that are least
be subject to lower margin requirements, and therefore their prices
fundamentals at times of
do not
misinterpret price deviations from
and thus increase margin requirements. This may lead to
margins means that speculators cannot provide as much
volatile will
sell risky
thus providing liquidity to asset
fundamentals (which would lead to arbitrage opportunities for speculators and thus reduce
rising
is
keeping prices close to fundamental values). Speculators have limited capital and borrow
from financiers who set margin constraints to control
in
liquid
not triggered by fear of what others are doing but by institutional
smoothing temporary imbalances
increase
it
lower the equilibrium price
of the features of models of bank runs
features, in particular margin requirements that
(i.e.
a
constraints.
Dybvig (1983), such as the possibility of multiple equilibria due to coordination
markets
its
the asset to a fixed multiple of their
the bank has limited funds, this
of the asset even further. The model has
assets,
intermediary
from investors using
hedge funds are leveraged,
price of the asset will necessitate a decrease in their holdings to
a similar role in the
first
the bank are the second intermediaries (perhaps hedge funds) and they
in
are subject to margin constraints that limit their investment
equity.
is
liquid
it
The
may
will
be closer to
illiquidity.
marl<ets are relatively new,
they are subject to Knightian uncertainty.
This
uncertainty has the potential to explain the extreme withdrawal fron) risk-sharing during severe flight to
quality episodes.
A second
class of
"Knightian" uncertainty.
In
models of
resort to
decision
make
rules
when
based on a distinction between
argument goes, market participants
that seek to optimize worst-case-scenario outcomes, with
flights to quality.
cash and have to decide
is
and Krishnamurthy (2008) model how
They study an economy where
to
risk
and
lack the
precise probabilistic judgments about future events. Instead, they
destabilizing aggregate consequences. Caballero
behavior affects
quality
certain circumstances, the
information or experience to
may
flights to
consume
it.
identical agents
The economy may be
hit
have an
this
possibly
form of
endowment
of
by liquidity shocks, whereby
some agents have
a
sudden need for cash. Conditional on
a first (aggregate) liquidity shock, there
is
a
probability that a second shock, affecting the agents spared by the first shock, also takes place. Agents
write insurance contracts that dictate transfers of cash to
by either a
first
shock or a second shock. The efficient allocation
against being hit by the
event.
likely
market
If
one another
first
shock than against being
hit
agents knew the probabilities of being
is
the event of one of thenn
by the second, simply because
hit
be either shock, this
shocks but have Knightian uncertainty about whether they
be
will
in
is
first
by the
hit in
a
more
probabilities of
in
outcome, which
the second wave, and are not willing to commit enough of their capital to insuring those
wave. There
first
is
or second wave. Thus
free insurance market they seek to insure themselves against the worst possible
being
this
the allocation a free
know aggregate
the
thus an inefficient
is
flight to quality:
hit
is
more insured
such that agents are
insurance would produce. By assumption, however, agents
in
in
a
is
hit
Out of fear of being part of a second wave
of shocks, agents prefer to hoard the safest asset (cash) instead of offering insurance contracts against
first
wave
shock, which as a result, they are underinsured against. Put differently, private liquidity
freezes too soon.
Brock and Manski (2008) also study the role of Knightian uncertainty
in a
model of
a
market for
Lenders must decide what fraction of their assets to allocate to loans and what fraction to a
risky loans.
safe asset. At
some
point there
is
an unexpected shock that lenders do not know
rules following the
possible decision
how to
interpret.
Three
shock are considered: a standard one where lenders place
subjective probabilities on the possibility of repayment;
one where they seek to maximize
their payoffs
under the worst possible repayment scenario (maxmin), and one where they minimize the
maximum
possible regret from their decision (minmax-regret).
show
the
to quality effect (increases
flight
amount
equilibrium
A question
is
what
in
In
contractual interest rates on loans and decreases
of loans) can be greater under the
left
calibrated numerical examples, they
maxmin
liquidity shocks. Caballero
the
in
or minmax-regret criteria.
open by theories that appeal to Knightian uncertainty to explain
max-min decision making by investors
exactly triggers robust or
that
in
flights to quality
response to aggregate
and Krishnamurthy (2008) argue that unfamiliar contexts, often related to
recent financial innovation, are prone to this kind of behavior. As an example, they contrast the market's
reaction to the demise of
same
in
1998 to
its
reaction to the losses suffered by
Central's default
is
in
1970 to
their reaction to
Mercury Finance's default
Evidence on the overall relevance of Knightian uncertainty
Wang
in
2006.
In
the
in
1997.
In
each case the
that market participants' increased familiarity, with the operations of hedge funds and with
commercial paper respectively, accounts for the calm with which the
and
Amaranth
Krischnamurthy (2008) contrasts the reaction of commercial paper investors to Penn
line,
argument
LTCM
(2005).
They
prices far better than
find that a
model allowing
one with pure
of-the-money put options.
in financial
markets
for uncertainty aversion
risk aversion. In particular,
it
latter
fits
is
event was received.
provided by
Liu,
Pan,
the evidence on option
accounts for the premium on far-out-
When markets
Proposition 5:
turn illiquid
develops and this leads to strategic
One important
is
and important players become
power
constrained, pricing
exacerbation.
illiquidity
feature of the Knightian uncertainty nnodel of Caballero and Krishnamurty (2008)
that market participants
to pool their liquid assets efficiently, hoarding
fail
them
in
fear of worst-case
scenarios instead of insuring each other against intermediately-bad shocks. Acharya, Gromb, and
Yorulmazer (2008) provide an alternative explanation for the
specific context of the interbank loan market,
stemming from
failure of private
liquid
coinsurance
banks exercising monopoly power
may
over banks that have liquidity needs. They model a bank (bank A) which needs cash and
either of
two ways. One
enough stake
must
retain a large
of
assets. Alternatively,
its
raise
more cash
specificity
by borrowing from a liquid bank (bank
is
since
in its
A does
B);
there
sell
some
of
its
obtain
a limit to this
is
asset portfolio to have incentives to engage
bank A may simply
it
in
because
A
costly monitoring
in
assets to bank B; by assumption, this can
not need to retain a stake; however, the assets have varying degrees of
and B cannot obtain as much value from them as A would. Under perfect competition,
interbank borrowing would always be the preferred option and asset sales would only be used
maximum
cash that can be raised by borrowing were insufficient to meet A's cash needs,
some degree
sales.
the
in
of
monopoly power, the only way
The mechanism
in
to transfer value from
the paper resembles a flight to quality
A
to B
is
if
the
if
instead B has
through inefficient asset
that banks hoard liquidity rather than
in
lend to each other. However, they do so for opportunistic and strategic rather than precautionary
One
reasons.
of the reasons for the current high rates observed
in
the interbank loan market
may be
that they are not being set competitively as liquid banks speculate with the possibility of purchasing
assets at distress prices.
Brunnermeier and Pedersen (2005) also explore strategic considerations and show how they
may
lead to "predatory trading" during a flight to quality. They model a situation
known
(to a limited
The market
number
for this asset
impact on the
price.
is
of speculators) that a trader needs to liquidate his position
not perfectly
meaning that the distressed
liquid,
The optimal reaction by the informed speculators
is
order to profit from buying
in
exacerbate the problem.
if
his
wealth
falls
price, forcing
risk
speculators
back at a lower price
know
a given threshold,
the prices of
some
they
may
try to
risky illiquid assets
flights to quality
episodes. While the specifics are different,
in
that a given trader
may
asset.
to trade
in
the
down
same
direction
the price of the
Furthermore, margin constraints
will
become
provoke
fall
some
this
may
distressed and need to
sell
down
the
by
selling to drive
later. This
may
explain
why
in
a
more sharply than one would expect
considerations alone.
The various theories of
element
later.
the trader into distress and profiting from buying back
flight to quality,
based on
below
If
it
in
becomes
it
trader's sales will have an
as the distressed trader as fast as possible (attempting to "front run"), driving
asset
where
all
the theories
is
suggest several mechanisms that can be at play
some common themes emerge.
an actual or feared weakness
in
In
particular, a
the balance sheet of
in
these
common
some market
participants,
who by
eitiier tlieir specialized
sl<ill
or information play a key role
the determination of
in
many
asset prices, risk premia, or the allocation of funds. This observation informs
recommendations that the theories have
of the policy
inspired.
Policy
Walter Bagehot famously argued
be ready not only to keep
others.
security
it
in
1873 that
in a
panic "the holders of the cash reserve must
advance
for their ow/n liabilities, but to
They must lend to merchants, to minor bankers, to
is
good."
A
Principle 1:
l<ey
What do modern
is
to "issue a
guarantee"
be acted upon aggressively, even
private sector about asset values
if
the
government has
way
that
it
prices.
A
a lender or a
(at least certain kinds of)
when
in
clear
agreem.ent with
similar effect can be achieved by public purchases of illiquid assets.
would prevent bank
is
information than the
less
Bagehot's
a fair price, eliminating the equilibrium with mis-coordination
market maker of
last resort
runs. Similarly, Brock,
could allay the fears that lead investors to act
government
extreme flight
Krishnamurthy's (2008) model, a loan to intermediaries would enable them to
In
bank acting as either
that man,' w/henever the
.
honor their promise to purchase assets at
and depressed
man and
to tlie private sector tliat
The implications of the models discussed above are
recommendations.
freely for the liabilities of
theories have to add to or subtract from this recommendation?
aspect of intervention
to quality events will
'this
most
it
investment.
A
in
would prevent the
with this
last
approach
is
is
a
that
not subject to the kind of uncertainty faced by the private sector.
private investors are being excessively prudent
central
the
same
and Manski (2008) argue that the government
max-min fashion by guaranteeing
difficulty
fire sales in
A
minimum
it
return on
assumes that the
In practice,
determining
not to ascertain..
Caballero and Krishnamurthy (2008) consider a form of intervention that can be useful even
the government
is
not better informed than the private sector.
because agents, fearful of being
other against a
first
hit
is
model, the inefficiency arises
by a second liquidity shock, are excessively unwilling to insure each
shock and prefer to hoard
form of intervention
In their
If
liquidity rather
to provide insurance against a second
agents to insure each other against a
first
than pool
wave
it.
They show that the optimal
of shocks, which persuades private
shock. By acting as a lender of truly last resort, the central
bank may help overcome private banks' reluctance to act as each other's lenders of "intermediate
resort."
crisis
A
practical difficulty with
implementing
this policy
is
how
that the private sector should be encouraged to deal with on
warrant intervention.
to distinguish intermediately severe
its
own
with true catastrophes that
moral hazard concerns are important for regulatory purposes but
Principle 2: Conventional
so for
less
interventions during crises.
A major concern, known
moral hazard.
If
financial firms expect to receive
an incentive to take on excessive
of excess investment
illiquid
the days of Bagehot and studied extensively since,
in
risk.
assets.
in illiquid
This risk
Due to
emergency financing should they require
may
take the form of excessive leverage or,
the issue of
it,
they have
more
subtly,
arbitrageurs' limited resources, each firm that has to
assets depresses prices for everyone else, but does not take this effect into account
decision making. This concern
institutions.
The current
one of the
is
rationales for the
may be
a case for subjecting
are imposed on banks, such as limits on leverage and
(2008) argues
and may
in this direction,
in
its
has highlighted that these externalities can be created by institutions
crisis
asset markets, there
sell
prudential regulation of financial
other than commercial banks, such as hedge funds. Given their prominent role
many
is
them
in
supplying liquidity to
to the kinds of prudential regulations that
minimum
requirements. Krishnamurthy
liquidity
cautioning however that these regulations
necessarily be a blunt tool
will
distort decisions in the (usual) non-crisis states of the world.
when
Referring back back to Caballero and Krishnamurthy (2008),
prevalent, the main problem
is
too
aggressive use of private liquidity
little
is
a
once
a crisis
is
private risk taking,
positive rather than a negative
highlights an important policy concern. While
a source of trouble,
much
not too
it is
true that excessive
reached, the greater concern
is
risk
Knightian uncertainty
and hence inducing
a
is
more
development. Their model
taking prior to the
crisis
insufficient risk taking
can be
and public
support can encourage rather than prevent desirable private sector behavior.
Principle 3: In the presence of speculative
and predatory behavior, there
intervention tools, including equity support
and shortselling
If
the worry
order. Acharya,
is
is
space for a wide range of
constraints.
about opportunistic or speculative behavior, several possible
Gromb, and Yorulmazer (2008) point out that
if
policies
may be
monopolistic behavior by liquid banks
in
is
make
what prevents the smooth functioning of interbank markets,
a central
loans to troubled banks could improve their outside option
bargaining, leading to less inefficient asset
sales without the
need to ever disburse emergency
by predatory traders, as
in
in
the
list
may prove
useful.
on shortselling give some support to
financial institutions
was
If
instead the concern
is
willing to
about front running
Brunnermeier and Pedersen's (2005) model, unorthodox measures
trading halts or limits on shortselling
restrictions
loans.
in
bank that stood
The share
price responses to recently
this idea. Interestingly,
when
like
announced
shortselling of several
restricted last July, the share prices of other financial institutions not included
reacted similarly to those that were included. This
is
probably partly due to the fact that the
were aimed
shortselling constraints
at stabilizing the core of the financial system,
which maximizes
positive feedbacl<s/
•Restricted July 15
Not restricted
July
15
S&P500
80
3
60
I
40
20
ni
l-Jun-08
l-Aug-08
l-Jul-08
l-Sep-08
were announced on July 15 and
announcement was met by a sharp increase in share prices, but
for the companies included in the restricted list and for other financial companies (the graph shows the
averages of the companies incuded in the ban and that of the five largest financial companies not
included in the ban). The reintroduction of a wider ban on September 19 was also accompanied by a
Figure
were
in
Restrictions
7.
on short-selling
place until August 12. The
sliares in certain financial firms
initial
sharp
We
close with a brief discussion of important implementation issues which are not well covered
by existing theories but rather by only occasional
Conjecture
without
^
rise.
1:
The
"political
politicians. This
hints.
tempo" of intervention
is
significantly slower than that implied
by the models
negotiated delay exacerbates uncertainty and flight to quality.
See also Caballero (1999) for
a discussion of
the value of equity market interventions
Kong's stock market intervention to fight a speculative attack during the Asian
crisis in
the
in
the context of Hong
late 1990s.
The models assume that the conditions under which the government would intervene can be
precisely formulated ex-ante
taken by a
political
and are well understood by market
participants. In practice, decisions are
how
the costs and benefits of intervention
process that
among
are to be allocated
subject to disputes about
is
distressed firms, other market participants, and taxpayers. The process
therefore subject to both delays and uncertainty about outcomes. This
flight to quality since
measures that could be
might not calm investors' fears
An important
if
sufficient
if
is
it
harder to defuse a
they were promptly and credibly announced
they do not know when, whether, or
quality of any intervention policy
may make
is
how
they
will
be implemented.
the promptness and predictability of
political
its
implementation process.
Conjecture 2: Agents learn within a
crisis,
which raises the intervention threshold as time goes
by.
Unfamiliar conditions, financial instruments, or events are particularly susceptible to flights to
quality.
However, what
unfamiliar at the beginning of a
is
may
crisis
rapidly
become
familiar.
Many
market participants were caught unaware by the collapse of Bear Sterns, whereas the collapse of
Lehman Brothers was more widely
anticipated.
progressively better, their assessment of risks should
on worst-case scenarios diminished.
If
this
is
so,
market participants understand the situation
If
become more
firmly
grounded and
their reliance
then private co-insurance should gradually take
precedence over public insurance, and the threshold for intervention should become increasingly
demanding.
Conjecture 3: Conventional lender of last resort interventions are insufficient once capital constraints
become
binding.
A lender
of last resort
is
useful
when
financial institutions
from either meeting
firm facing a
If
institutions
run.
instead the
from taking on
risk,
main usefulness of lending of
availability of cash
form of intervention
is
the binding constraint that prevents
their obligations or extending credit, as
binding constraint
is
then traditional lending of
last resort in this
context
of a traditional bank run. To the extent that this
this
is
is
is
may be
that insufficient capital
last resort will
to reduce
not the main
one
risk
the case with a
prevents financial
not be able to relax
particular source of
this.
risk,
the
The
risk
that investors are worried about,
of limited usefulness.
Conjecture 4: Whether intervention should be done through asset buybacl<s or direct equity injections
depends on which market
is
experiencing the largest distortions.
Intervention should be designed to achieve the greatest possible impact per dollar.
If
there
is
confidence that certain assets held by financial institutions are significantly undervalued, then limited
asset buybacks can
(i)
sustain asset prices, allowing financial institutions to escape the rigors of mark-to-
ri','
1,;
market accounting at
fire-sale prices;
(ii)
deliver a profit to (or
an indirect way of recapitalizing them; and
possible
in
(iii)
ensure a
linnit
the losses
of) financial institutions,
profit for taxpayers. This,
is
only
the cases of securities that are significantly undervalued, which creates a sufficiently large
wedge between current
valuations and fair prices to allow for profits for both the seller and the
taxpayer. The behavior of liquidity premia during flights to quality suggests that this
the most
however,
illiquid securities, including
Assessing
identify obviously
fair
values of
new and
may be
the case for
untried ones.
illiquid securities is a
considerable practical challenge.
If
it
is
hard to
undervalued assets, direct equity injections are a more straightforward approach. This
policy simultaneously capitalizes financial institutions
(see, e.g., Caballero 1999).
and most
likely yields a
high return to taxpayers
References
Acharya, Denis Gromb, and Tanju Yorulnnazer. Imperfect Competition
o
Viral V.
Market
for Liquidity.
o
Dirk Baur
o
IMS, April
An
Empirical Analysis of Stock-
2006.
from the Euro-area bond market. Review of Financial Studies 2008
Ben Bernanke, Mark
Quality. Thie
o
Flight-to-quality or contagion?
Alessandro Beber, Michael W. Brandt, and Kenneth A. Kavajecz. Flight-to-quality or Flight-to-
Liquidity? Evidence
o
the Inter-Bank
Working Paper, London Business School, 2008.
and Brian M. Lucey.
Bond Correlations.
in
Gertler,
and Simon
Gilchrist.
The
and the
Financial Accelerator
Flight to
Review of Economics and Statistics, 78(1):1-15, February 1996.
and Charles
Willian A. Brock
F.
Manski. Competitive Lending
w/ith Partial
Knowledge of Loan
Repayment. Working paper. University of Wisconsin, 2008.
o
Markus
K.
Brunnermeier and Lasse
H.
Pedersen. Predatory Trading.
Tlie
Journal of Finance,
60(4):1825-1863, 2005.
o
Markus
Brunnermeier and Lasse
K.
H. Pedersen.
Market
Liquidity
and Funding
Liquidity.
Review
of Financial Studies, forthcoming, 2008.
An Update. Mimeo, MIT,
o
Ricardo
J.
Caballero. Stock Market Interventions:
o
Ricardo
J.
Caballero and Arvind Krishnamurthy. Collective Risk
July 1999.
Management
in a Flight
to Quality
Episode. Journal of Finance, Forthcoming, 2008.
o
Douglas,
W. Diamond, and
Philip H. Dybvig.
Bank Runs, Deposit Insurance, and
Liquidity.
Journal
of Political Economy, 91(3): 401-419, June 1983.
o
Evan Gatev and
Philip
E.
Strahan. Banks' Advantage
in
Hedging Liquidity
Risk:
Theory and
Evidence from the Commercial Paper Market. The Journal of Finance, 61(2):867-892, 2006.
o
Jesus Gonzalo
Universidad Carlos
o
P.
III,
Hartmann,
and Jose Olmo. Contagion versus
Departamento de Economia,
S.
Flight
to
Quality
in
Financial
Markets.
April 2005.
Straetmans, and C.G. Vries. Asset Market Linkages
in Crisis
Periods. Review of
Economics and Statistics, 86(l):313-326, 2004.
o
Zhiguo
He and Arvind Krishnamurthy. A Model of
Capital
and
Crises.
Working paper.
Northwestern University, 2008.
o
Bengt Holmstrbm and Jean
Tirole.
Financial
Intermediation, Loanable Funds, and the Real
Sector. The Quarterly Journal of Economics, 112(3): 663-691, August 1997.
o
Anil K. Kashyap,
Jeremy
C. Stein,
and David W. Wilcox. Monetary Policy and Credit Conditions:
Evidence from the Composition of External Finance. American Economic Review, 83(1):78- 98, March
1993.
o
Arvind
Krishnamurthy.
Amplification
Mechanisms
in
Liquidity
Crises.
Working
paper.
Northwestern University, 2008.
o
William
W. Lang and Leonard
I.
Nakamura.
Flight to Quality in
Banking and Economic Activity.
Journal of Monetary Economics, 36(1):145-164, 1995.
o
Jun
Liu,
Jun Pan, and Tan Wang. An Equilibrium Model of Rare-Event Premia and
for Option Smirks.
o
its
Implication
Review of Financial Studies, 18(1):131-164, 2005.
Francis A. Longstaff.
The
Flight-to-Liquldlty
Premium
in
US Treasury Bond
Prices.
and the Pricing of
Risk.
The Journal of
Business, 77(3):511-526, 2004.
o
DImltri Vayanos. Flight to Quality, Flight to Liquidity,
Paper, 2004.
NBER Working