American Economic Association

Liquidity and Risk Management

Author(s): Nicolae Grleanu and Lasse Heje Pedersen
Source: The American Economic Review, Vol. 97, No. 2 (May, 2007), pp. 193-197
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/30034445
Accessed: 20-03-2015 12:33 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/30034445?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The American Economic
Review.

http://www.jstor.org

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions

FINANCIAL
SEARCH-AND-MATCHING
MARKETSt

andRiskManagement
Liquidity
By NICOLAEG^XRLEANUAND LASSE HEJE PEDERSEN*
This paperprovidesa model of the interaction
between risk-managementpractices and market
liquidity. Our main finding is that a feedback
effect can arise. Tighterrisk managementleads
to marketilliquidity, and this illiquidity further
tightens risk management.
Risk managementplays a centralrole in institutional investors' allocation of capital to trading. For instance, a risk manager may limit a
trading desk's one-day 99 percent value at risk
(VaR) to $1 million. This means that the trading desk must choose a position such that, over
the following day, its value drops no more than
$1 million with 99 percent probability. Risk
management helps control an institution'suse
of capital while limiting defaultrisk, and helps
mitigateagency problems.Phillipe Jorion(2000,
xxiii) states that VaR "is now increasinglyused
to allocate capital across traders,business units,
products,and even to the whole institution."
We do not focus on the benefits of risk managementwithin an institutionadoptingsuchcontrols,but, rather,on the aggregateeffects of such
practiceson liquidityandasset prices.An institution may benefitfromtighteningits risk management and restrictingits securityposition,but as a
consequenceit cannotprovideas much liquidity
to others.We show that,if everyoneuses a tight
risk management,then marketliquidity is lowered in that it takes longer to find a buyer with
unusedrisk-bearingcapacity,and, since liquidity
is priced,prices fall.
tDiscussants: Dimitri Vayanos, London School of
Economics; Neil Wallace, Pennsylvania State University;
Manuel Amador, Stanford University.
* Gdrleanu:Wharton School,
University of Pennsylvania, 3620 Locust Walk, Philadelphia,PA 19104-6367, and
National Bureau of Economic Research, and Centre for
Economic Policy Research (e-mail: garleanu@wharton.
upenn.edu);Pedersen: Stern School of Business, New York
University, 44 West Fourth Street, Suite 9-190, New York,
NY 10012-1126, NBER, and CEPR (e-mail: lpederse@
stern.nyu.edu).We are grateful for helpful conversations
with Franklin Allen, Dimitri Vayanos,and Jeff Wurgler.

Not only does risk managementaffect liquidity; liquidity can also affect risk-management
practices.Forinstance,the Bankfor International
Settlements (2001, 15) states, "Forthe internal
risk management,a number of institutions are
exploring the use of liquidity adjusted-VaR,in
which the holding periods in the risk assessment
are adjusted to account for market liquidity,
in particularby the length of time required to
unwindpositions."For instance, if liquidationis
expected to take two days, a two-day VaR might
be used instead of a one-day VaR. Since a security's risk over two days is greaterthan over one
day, this means a tradermust choose a smaller
position to satisfy his liquidity-adjustedvalue at
risk (LVaR)constraint.One motivationfor this
constraintis that, if an institutionneeds to sell,
its maximum loss before the completion of the
sale is limited by the LVaR.
The main resultof the paperis that subjecting
tradersto an LVaRgives rise to a multipliereffect.
Tighterrisk managementleads to morerestricted
positions, hence longer expected selling times,
implying higher risk over the expected selling
period, which furthertightensthe risk management, and so on. This feedbackbetweenliquidity
andriskmanagementcanhelpexplainwhyliquidity can suddenlydrop.We show that this "snowballing"illiquiditycan arise if volatilityrises, or
if more agents face reducedrisk-bearingcapacity-for instance, because of investor redemptions, losses, or increasedrisk aversion.
Our link between liquidity and risk management is a testable prediction.While no formal
empirical evidence is available, to our knowledge, our predictionis consistentwith anecdotal
evidence on financial marketcrises. For example, in August 1998 several traderslost money
due to a defaultof Russian bonds and, simultaneously,marketvolatility increased.As a result,
the (L)VaRof many investmentbanks and other
institutionsincreased.To bringrisk back in line,
many investmentbanks reportedlyaskedtraders
to reducepositions,leading to falling prices and

193

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions

AEA PAPERSAND PROCEEDINGS

194

lower liquidity.These marketmoves exacerbated

the risk-managementproblems,fueling the crisis
in a similar mannerto the one modeledhere.
We capture these effects by extending the
search model for financial securities of Darrell
Duffie, Garleanu,andPedersen(2005, forthcoming, henceforthDGP). This frameworkof timeconsuming search is well suited for modeling
liquidity-basedrisk managementas it providesa
naturalframeworkfor studyingendogenousselling times. While DGP relied on exogenousposition limits, we endogenize positions based on a
risk-managementconstraint,and considerboth a
simple and a liquidity-adjustedVaR. Hence, we
solve the fixed-pointproblemof jointly calculating endogenouspositionsgiven the risk-management constraintand computingthe equilibrium
(L)VaRgiventhe endogenouspositionsthatdetermine selling times and price volatility. PierreOlivier Weill (forthcoming) considers another
extensionof DGP in which marketmakerliquidity provision is limited by capital constraints.
Our multipliereffect is similarto thatof Markus
K. BrunnermeierandPedersen(2006) who show
that liquidity and traders'margin requirements
can be mutuallyreinforcing.
I. Model
The economy has two securities: a "liquid"
security with risk-free return r (i.e., a "moneymarket account"),and a risky illiquid security.
The risky security has a dividend-rateprocess X
and a price P(X), which is determinedin equilibrium. The dividend rate is L6vy with finite
variance. It has a constant drift normalized to
zero, E, (X(t + T) - X(t)) = 0, and a volatility

ax > 0, i.e.,
(1)

var,(X(t + T) - X(t)) = a2xT.

Examples include Brownian motions, (compound) Poisson processes, and sums of these.
The economy is populatedby a continuumof
agents who are risk neutraland infinitely lived,
have a time-preferencerateequal to the risk-free
interest rate r > 0, and must keep their wealth
boundedfrombelow.Each agentis characterized
by an intrinsictype i E {h, 1},which is a Markov
chain, independentacross agents, and switching
from 1 ("low") to h ("high")with intensity A,,
and back with intensity Ad. An agent of type i

MAY2007

holding 8, shares of the asset incurs a holding

cost of 8 > 0 per shareand per unit of time if he
violates his risk-managementconstraint
(2)

vart (Ot[P(Xt+,)

P(Xt)])

(o'i)2

where a' is the risk-bearingcapacity,definedby

o"h= &"> 0 and o-T= 0. The low risk-bearing
capacity of the low-type agents can be interpreted as a need for more stable earnings,hedging reasons to reduce a position, high financing
costs, or a need for cash (e.g., an asset manager
whose investorsredeem capital).'
We use this constraint as a parsimonious
way of capturing risk constraints, such as the
very popular VaR constraint,2which are used
by most financial institutions. Our results are
robust in that they rely on two naturalproperties of the measure of risk: the risk measure
increases with the size of the security position,
and the length of the time period r over which
the risk is assessed. While the constraintis not
endogenized in the model, we note thatits wide
use in the financial world is probably due to
agency problems, default risk, and the need to
allocate scarce capital.
We consider two types of risk management:
(a) "simplerisk management,"in which the variance of the position in (2) is computed over a
fixed time horizon r; and (b) "liquidity-adjusted
risk management,"in which the variance is
computed over the time requiredfor selling the
asset to an unconstrainedbuyer,which will be a
randomequilibriumquantity.
Because agents are risk neutral and we are
interested in a steady-state equilibrium, we
restrict attentionto equilibria in which, at any
given time and state of the world, an agentholds
either0 or0 unitsof the asset,where0 is thelargest
1 An interesting extension of our model would consider
the direct benefit of tighter risk management,which could
be capturedby a lower Ad.
2 A VaR constraint stipulates that Pr(-O[P(X,,,)
P(X,)] VaR) <:7r for some risk limit VaR and some confidence level
7r.If X is a Brownian motion, this is the same
as (2). We note that ratherthan considering only price risk,
we could alternativelyconsider the risk of the gains process
(i.e., including dividend risk) G,,, = P(X(t + 7)) - P(X(t))
+ fX(s) ds. This yields qualitatively similar results (and
quantitativelysimilar for manyreasonableparameterssince
dividend risk is ordersof magnitude smallerthan price risk
over a small time period).

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions

VOL.97 NO. 2

LIQUIDITYAND RISKMANAGEMENT

position that satisfies (2) with o"'= 6 taking the

prices and search times as given.3Hence, the set
of agent types is T = {ho, hn, lo, ln}, with the
letters"h"and "'"designatingthe agent'scurrent
intrinsicrisk-bearingstateas high or low, respectively,and with "o"or "n"indicatingwhetherthe
agent currentlyowns 0 shares or none, respectively.We let ug(t)denotethe fractionat time t of
E T . These fractionsadd up to
agents of type G
1 and marketsmust clear:
(3)

(4)

1 = tho

hn+

ln,

,Llo

0 = 0(-ho ++ lto),

where 0 > 0 is the total supply of shares per

investor.
Central to our analysis is the notion that the
risky security is not perfectlyliquid, in the sense
that an agent can trade it only when she finds
a counterparty.Every agent finds a potential
counterparty,selected randomlyfrom the set of
all agents, with intensity A, where A > 0 is an
exogenous parameter characterizing the market liquidity for the asset. Hence, the intensity
that is, the
of finding a type-f investor is
search intensity multiplied by A/zt,
the fraction of
investors of that type. When two agents meet,
they bargain over the price, with the seller having bargainingpowerq E [0, 1].
This model of illiquidity directly captures
the search that characterizes over-the-counter
(OTC) markets. In these markets,tradersmust
find an appropriatecounterparty,which can be
time consuming. Tradingdelays also arise due
to time spent gathering information, reaching trading decisions, mobilizing capital, etc.
Hence, trading delays are commonplace, and,
therefore,the model can also capturefeaturesof
other markets such as specialist and electronic
limit-order-bookmarkets, although these markets are, of course, distinctfrom OTC markets.
II. EquilibriumRisk Management,Liquidity,
and Prices
We now proceed to derive the steady-state
equilibrium agent fractions Ax,the maximum3 Note that the existence of such an equilibriumrequires
that the risk limit & not be too small relativeto the total supply 0, a condition that we assume throughout.

195

holding 0, and the price P. Naturally,low-type

ownerslo want to sell and high-typenon-owners
hn want to buy, which leads to
(5)

0 =

-2Ap.Lhn(t)p.Llo(t)
-Auplo(t)

+ Ad
ho(t)

and three more such steady-state equations.

Equation(5) statesthatthe changein the fraction
of lo agents has three components,corresponding to the three terms on the right-hand-sideof
the equation.First, whenevera lo agent meets a
hn investor,he sells his asset and is no longer a
lo agent.Second, wheneverthe intrinsictype of a
lo agentswitchesto high, he becomes a ho agent.
Third, ho agentscan switch type and become lo.
Duffie, Garleanu,andPedersen(2005) showthat,
taking 0 as fixed, there is a uniquestable steadystate mass distributionas long as 06 0. Here,
agents' positions 0 are endogenous and depend
on 1a,so that we must calculatea fixed point.
Agents take the steady-statedistributiong as
fixed when they derive their optimal strategies
and utilitiesfor remaininglifetime consumption,
as well as the bargainedprice P. The utility of an
agent dependson his currenttype ((t)E T (i.e.,
whetherhe is a high or a low type andwhetherhe
owns zero or 0 shares),the currentdividendX(t),
and the wealth W(t)in his bank account:
(6) V,(X(t),W,)= Wt+ 1(~E{ho,,lo})X(t)/r+ovy,
where the type-dependent utility coefficients
vg are to be determined.With q the bargaining
power of the seller, bilateral Nash bargaining
yields the price
(7)

PO = (Vlo

Vt) (1 - q) + (Vho- Vhn q.

We conjecture,and later confirm, that the equilibriumasset price per shareis of the form
(8)

P(X(t))= X(t)/r+ p,

for a constant p to be determined. The valuefunctioncoefficientsvyandp are given by a set of

Hamilton-Jacobi-Bellmanequations, stated and
solved in the Appendix availableat www.e-aer.
The Appendix
org/data/may07/p07048_app.pdf.
contains all the proofs.
PROPOSITION1: If the risk-limit &ris sufficiently large, there exists an equilibrium with

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions

196

AEA PAPERSAND PROCEEDINGS

holdings 0 and &that satisfy the risk management constraint (2) with equality for low- and
high-type agents, respectively.Withsimple risk
management,the equilibriumis unique and
=

(9)

r-x

1
\/

With liquidity-adjustedrisk management,0

fractionof potential
dependson theequilibrium
buyersAIhnandsatisfies
(10)

= r

hn

In both cases, the equilibriumprice is given by

X,
(11) P(Xt) = r
r(1 q) + Ad + 2AI/.o(1 q)
r r + Ad+ 2Al.ro(1 q) + Au+
2A-hnq'

where the fractions of agents gt depend on the

type of risk management.
These resultsareintuitive.The "positionlimit"

8 increases in the risk limit &-and decreases in

the asset volatility and in the squareroot of the

VaR period length, which is r under simple
under liquidrisk management and
In the lattercase,
(2Ahhn)-1
ity-adjustedrisk management.
position limits increase in the search intensity
and in the fraction of eligible buyers
I-hn.
The price equals the present value of dividends, Xt/r, minus a discount for illiquidity.
Naturally,the liquiditydiscountis largerif there
are more low-type owners in equilibrium(-tlois
larger)and fewer high-type nonownersready to
buy (Ihn is smaller).
Of the equilibria with liquidity-adjustedrisk
management,we concentrateon the ones thatare
stable, in the sense that increasing0 marginally
would result in equilibriumquantitiesviolating
the VaR constraint (2). Conversely,an equilibrium is unstable if a marginal change in holdings that violates the constraintwould result in

MAY2007

the equilibriumadjustingso that the constraint

is not violated. If an equilibriumexists, then a
stable equilibrium exists. Indeed, the equilibriumwith the largest0 is stableand has the highest welfare among all equilibria.
The main resultof the papercharacterizesthe
equilibrium connection between liquidity and
risk management.
PROPOSITION 2: Suppose that &"is large
enough for the existence of an equilibrium.
Consider a stable equilibrium with liquidityadjustedriskmanagementand let 7 = 1(2AUhn),
which means that the equilibrium allocations
and price are the same with simple risk management. Consider any combination of the
conditions (a) higher dividend volatility
(b)
orx,
lower risk limitC;(c) lower meeting intensity
A,
(d) lower switching intensity Auto the high riskbearing state, and (e) higher switching intensity
Ad to the low risk-bearing state. Then, (i) the
equilibriumposition 0 decreases, (ii) expected
search timesfor selling increase, and (iii)prices
decrease. All threeeffects are larger with liquidity-adjustedrisk management.
To see the intuitionfor these results, consider
the impact of a higher dividend volatility. This
makes the risk-managementconstraint tighter,
inducing agents to reduce their positions and
spreading securities among more agents, thus
leaving a smallerfractionof agents with unused
risk-bearingcapacity.Hence, sellers' searchtime
increases and their bargainingpositionworsens,
leading to lowerprices.This price dropis due to
illiquidity,as agents are risk neutral.4
With liquidity-adjustedrisk management,the
increased search time for sellers means that the
risk over the expected liquidationperiod rises,
thus further tightening the risk-management
constraint,reducingpositions, increasingsearch
times, and so on.
This multiplier also increases the sensitivity of the economy with liquidity-adjustedrisk
managementto the other shocks (b)-(e). Indeed,
a lower risk limit (b) is equivalent to a higher
4 In a Walrasianmarketwith immediate trade, the price
is the present value of dividends X/r when (0/0) < Ah/
(Au + Ad), a condition that is satisfied in our examples
below. (When /0 > A,/(A, + Ad),the Walrasianprice is
(X-6)/r .)

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions

VOL.97 NO. 2

LIQUIDITYAND RISKMANAGEMENT

0.015
0.014

197

9.9
VaR
Liquidity
SimpleVaR

VaR
- Liquidity
' SimpleVaR

9.88

0.013

(years)

9.86

0.011
times

9.84

0.012

Price

sale0.01

9.82

0.009
9.8
0.008

Expected
0.007
0.2

9.78
0.2

0.3
0.35
0.25
Dividend volatilityox

0.35
0.3
0.25
Dividend volatilityox

0.4

FIGURE
1
Note: The effects of dividend volatility on equilibrium seller search times (left panel) and prices (right panel) with simple
(dashedline) and liquidity-adjusted(solid line) risk management,respectively.

dividendrisk. The "liquidityshocks" (c)-(e) do

not affect the equilibriumposition 0 with simple
risk management,but they do increase the sellers' searchtimes andreduceprices. With liquidity-adjusted risk management, these liquidity
shocks reduce securitypositions,too, because of
increased search times and, as explained above,
a multipliereffect arises.
The multiplier arising from the feedback
between trading liquidity and risk management clearly magnifies the effects of changes
in the economic environmenton liquidity and
prices. Our steady-state model illustrates this
point using comparative static analyses that
essentially compare across economies. Similar
results would arise in the time series of a single
economy if there were random variationin the
model characteristic,e.g., parametersswitched
in a Markov chain as in Duffie, Garleanu,and
Pedersen (forthcoming).In the context of such
time-series variation, our multiplier effect can
generatethe abruptchangesin prices and selling
times that characterizecrises.
We illustrate our model with a numerical
example in which A = 100, r = 0.1, Xo 1,
Ad = 0.2, Au = 2, 8 = 3, q = 0.5, 0 = 1, and
S= 1. Figure 1 shows how prices (rightpanel)
and sellers' expected search times (left panel)
depend on asset volatility.The solid line shows
this for liquidity-adjustedrisk managementand
the dashed line for simple risk management

with 7 = 0.0086, which is chosen so that the

risk managementschemes are identicalfor ox =
0.3. Search times increase and prices decrease
with volatility. These sensitivities are stronger (i.e., the curves are steeper) with liquidityadjustedrisk managementdue to the interaction
between marketliquidity (i.e., searchtimes) and
risk management.
REFERENCES
Brunnermeier, Markus K., and Lasse Heje
Pedersen.2006. "MarketLiquidityand Funding Liquidity."Unpublished.
Duffie,Darrell,NicolaeGirleanu,and LasseHeje
Pedersen.2005. "Over-the-Counter
Markets."
Econometrica,73(6): 1815-47.
Duffie,Darrell,NicolaeGirleanu,and LasseHeje
Pedersen.Forthcoming. "Valuationin Overthe-Counter Markets."Review of Financial
Studies.
Bank for InternationalSettlements.2001. "Final
Report of the Multidisciplinary Working
Group on EnhancedDisclosure."http://www.
bis.org/publ/joint0l.htm.
Jorion,Phillipe.2000. Valueat Risk. New York:
McGraw-Hill.
Weill, Pierre-Olivier. Forthcoming. "Leaning
against the Wind." Review of Economic
Studies.

This content downloaded from 210.160.35.61 on Fri, 20 Mar 2015 12:33:38 UTC
All use subject to JSTOR Terms and Conditions