You are on page 1of 30

Submitted to Operations Research

manuscript (Please, provide the mansucript number!)


Closed-Form Solutions for Option Hedging with
Market Impact
Tianhui Michael Li*
Princeton Bendheim Center for Finance and Operations Research and Financial Engineering
Robert Almgren
Quantitative Brokers, LLC and NYU Courant Institute
We present a highly-intuitive closed-form delta-hedging result for a large investor whose trades generate
adverse market impact. Unlike the complete-market or proportional-transaction-cost cases, the agent no
longer nds it tenable to be perfectly hedged or even within a xed distance of being hedged. Instead, he
may nd himself arbitrarily mishedged and optimally trades towards the classical Black-Scholes delta, with
trading intensity proportional to the degree of mishedge and inversely proportional to illiquidity. When
viewed in light of a recent result of Garleanu and Pedersen (2009), this suggests that delta-hedging can
be thought of as a Merton problem where the Merton-optimal portfolio is the Black-Scholes delta hedge.
Both the discrete-time and continuous-time problems are solved. We discuss a number of applications of our
result, including the implications of our model on intraday trading patterns and stock pinning at options
expiry. Finally, numerical simulations on TAQ data based on intraday hedging of call options suggest that
this strategy is able to signicantly reduce market impact cost with respect to Black-Scholes delta hedging.
1. Introduction
The solution to hedging an option in a complete market is well known (Black and Scholes 1973,
Merton 1973). However, the assumption of frictionless trading that underlies complete markets is far
from reality. Practioners who trade large positions are familiar with the concept of market impact
when trading in real markets with limited liquidity. For instance, a buyer who lifts orders from the
oer eats up liquidity and pushes his execution price ever higher. Liquidity of the underlying often
dries up during nancial crises such as the 1987 Crash, the LTCMs collapse, the recent subprime
debacle, or the London Whale scandal and this has only further heightened investor concerns about
liquidity.
We are interested in the optimal hedging of an option by a large risk-averse investor in an illiquid
market. There is a large literature on trading under transaction costs. Part of the literature is
interested in super-replication (Cetin, Soner, and Touzi 2010, Soner, Shreve, and Cvitanic 1995).
Our paper relaxes this requirement by having a nite penalty for being mishedged. Our paper is
more closely linked to these using a utility-based framework (Cvitanic and Wang 2001, Davis and
Norman 1990, Hodges and Neuberger 1989, Janecek and Shreve 2004, Leland 1985, Shreve and
Soner 1994).
Trading frictions have been modeled via various mechanisms. One strand of the literature is cen-
tered around illiquidity (Cetin, Jarrow, and Protter 2004). These authors hypothesize a stochastic
supply curve whose steepness coincides with illiquidity. Similar work was done by Bank and Baum
(2004). The major drawback of both their theories is that liquidity costs can be completely avoided
if trades are nite-variation. Hence, the cost of replicating an option is asymptotically zero as the
writer can take nite-variation approximations to the classical Black-Scholes -hedging technique.
* This research is based upon work supported by the National Science Foundation Graduate Research Fellowship
under Grant No. DGE-0646086 and the Fannie and John Hertz Foundation Graduate Fellowship. The author would
like to thank both organizations for their generous support.
1
Author: Option Hedging with Market Impact
2 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
To get around this problem, Cetin, Jarrow, Protter, and Warachka (2006) impose an articial
constraint on the minimum time between discrete trades.
Another strand of the literature models trading frictions as a cost proportional to trade size,
typically interpreted as arising from the bid-ask spread. Mathematically, these problems exhibit
a scaling which allows the authors to simplify their formulas. This branch of the literature uses
singular control and the optimal solution is typically in the form of a tracking band. As the portfolio
exits this band, the trader makes singular corrections to his holdings to keep it strictly within the
limits of the band (Cvitanic 1999, Cvitanic and Karatzas 1996). This is analogous to the results of
Davis and Norman (1990), Shreve and Soner (1994) for the Merton Problem. Typically, there is no
analytic form for these no-trade regions and much work has been done on their numerics (Clewlow
and Hodges 1997, Davis, Panas, and Zariphopoulou 1993) or asymptotics (Soner and Barles 1998,
Whalley and Wilmott 1997).
Yet another strand of the literature models the full order book. Alfonsi and Schied (2009),
Alfonsi, Fruth, and Schied (2007), Predoiu, Shaikhet, and Shreve (2010) model a resilient order
book where levels gradually replenish after large trades. They generally derive optimal strategies
in which the broker makes large block trades at the beginning and end of the trading period with
many small trades in between. These models often allow for general order book shape and are
mathematically rather challenging to work with.
The nal strand of related literature focuses on modeling eective market impact. Rather than
dwelling on the the details of microstructure, it models eective microstructure. The literature
began with Bertsimas and Lo (1998) and Almgren and Chriss (1999) who decompose price impact
in terms of both temporary and permanent price impact. The model was subsequently rened in
Obizhaeva and Wang (2005) by the addition of a transient price impact term. While earlier works
in the literature concentrated on the liquidation a xed number of shares in a nite period, these
impact models were applied to portfolio allocation (Garleanu and Pedersen 2009) and options
hedging (Rogers and Singh 2007), the liquidation of portfolios (Schied and Schoneborn 2007),
trading in both lit and dark venues (Kratz and Schoneborn 2010), and multiple-agent competitive
liquidation (Schoeneborn and Schied 2010).
In contrast to proportional-transaction-cost models, the source of incompleteness for our model
is price impact: the more aggressively one trades, the more impact is incurred proportionally.
The cost is no longer linear in the shares purchased and the problem no longer exhibits linear
scaling. We readily conceded that an ideal transaction cost model would naturally take into account
both bid-ask spread and market impact. However, such a model presents a dierent set of (very
intriguing) mathematical challenges that are beyond the scope of this paper. Nonetheless, we argue
that neglecting an explicit spread term is a more realistic model of the eective market impact
than including only the spread. The following cases serve to illustrate this point.
The spreads-only cost model assumes all orders are active (at the oer for purchases, at the
bid for sales). Most executions in US equities occur through electronic brokers who execute both
passively and aggressively. If the passive/active fraction is 50%, then the average execution is at
the mid price, i.e. the orders incur zero market impact (this assumes that all passive and active
orders execute at the top of the book). Most brokers are able to attain near-mid execution for
non-urgent orders.
For more urgent orders, the fraction of aggressive lls typically increases. For example, lls
that are 80% aggressive and 20% passive result in an eective market impact of 30% of the bid-ask
spread (again, assuming top-of-the-book execution). We can observe that the eective market-
impact is actually a smooth function of urgency and not realistically modeled by a xed, discrete
transactions cost as in the spread-only model.
For super-urgent orders, the top-of-the-book lls assumption is no longer realistic. A spread-
only model limits the market impact to the half-spread and is completely unable to account for
execution deeper into the book.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 3
In order to realistically account for the last point in a microsturcture sensitive way, we would
need to directly model the order book and both active and passive orders. This is addressed in
another strand of the literature (Alfonsi and Schied (2009), Predoiu et al. (2010)) by explicitly
modeling the order book. However, doing so poses a dierent set of mathematical challenges that
are beyond the scope of this article. Our eective model is chosen to judiciously capture the
eective trading costs while minimizing mathematical fuss. Yet, we are still able to place this model
within a tractable framework to obtain a highly intuitive closed-form trading solution.
Despite this mathematical parsimony, our solution more realistically handles the case when the
portfolio is far from being hedged, for example due to a sudden movement in underlying price.
Under only proportional transaction costs, the optimal solution in this situation is to cross the
spread with market orders in order to re-enter the tracking band. In real markets, such an order
would likely eat several levels into the order-book, incurring signicant market-impact, not modeled
by the naive, proportional-transaction-costs model. Under our market-impact model, the optimal
solution in this situation is to trade aggressively towards being hedged, in a way that takes into
account both available liquidity and the degree of the mishedge. Our trading strategy is much
smootherindeed, we are approximating the impact-free Black-Scholes Delta, an innite variation
process, with trading positions that are dierentiable. A comparison is given in Figure 1 to illustrate
the point.
Figure 1 Comparison of the proportional-transaction-cost fxied-tracking-band strategy (top) and our dynamic
strategy (bottom). In the former, our (green) trading position changes only when the (blue) target leaves
the (gray) trading-band. In the latter, our (green) trading position smoothly adjusts to the same (blue)
target. Compare the smooth trading ow and position of the latter strategy to the abrupt trading of
the former.
We separate market impact into temporary and permanent impact terms following the framework
developed by Almgren and Chriss (2001). The temporary impact aects only the execution price

P
t
Author: Option Hedging with Market Impact
4 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
but has no eect on the fair value or fundamental price P
t
. In contrast, the permanent impact
directly aects P
t
while having no direct eect on the execution price

P
t
. Thus we can think of
the temporary impact as connected to the liquidity cost faced by the agent while the permanent
impact as linked to information transmitted to the market by the agents trades (see, for example,
Back (1992), Kyle (1985)).
Previous work involving the Almgren-Chriss model has involved trading in the presence of Dark
Pools (Kratz and Schoneborn 2010), portfolio liquidation (Schied and Schoneborn 2007), and com-
petitive liquidation (Schoeneborn and Schied 2010). Hernandez-Del-Valle and Pacheco-Gonzalez
(2009) extends the work to Geometric Brownian Motion and Horst and Naujokat (2008) extends
it to derivative valuation. Rogers and Singh (2007) also examine the delta hedging problem but for
agents with dierent risk preferences and were only able to obtain analytic expressions asymptot-
ically. Our work diers from theirs in that we derive closed-form expressions by using simpler risk
preferences. The paper most closely related to ours is Garleanu and Pedersen (2009). They solve
the innite-horizon Merton Problem under only temporary market-impact assumptions. As in
our setup, they use a linear-quadratic objective rather than the traditional expected utility setup
of the classical Merton Problem. They nd that trading intensity at time t (Proposition 5) is given
by

t
=h
_
X
t
target
t
_
1/

where X
t
is the number of shares, is an urgency parameter with units of inverse time, h >0 is
a dimensionless constant of proportionality related to the convexity of the continuation value, and
the target portfolio target
t
is related to the frictionless Merton-optimal portfolio. That is, with
market-impact costs, it is no longer optimal to hold the instantaneous Merton-optimal portfolio but
instead, the agent trades towards a new target portfolio (which is a new Merton-optimal portfolio).
The intensity of trading
t
is proportional to the distance between the current holdings and target
(X
t
target
t
) and is inversely proportional to the square-root of the illiquidity parameter .
We use an analogous nite-horizon setup on [0, T] with both temporary and permanent impact
and obtain that the trading intensity is

t
=h((T t))
_
X
t
target
t
_
where the target portfolio target
t
is now related to the frictionless Black-Scholes delta-hedge and
h() is a positive dimensionless function, which comes from the nite-horizon nature of the setup
(compare to (15) with =0 so K =1). Hence, as in Garleanu and Pedersen (2009), an agent facing
market illiquidity no longer maintains the zero-liquidity-cost optimal portfolio but instead trades
towards a new target to correct this mishedging. Furthermore, also as in Garleanu and Pedersen
(2009), the agents trading intensity
t
is proportional to the dierence between his current holdings
and the optimal no market-impact portfolio (X
t
target
t
) and inversely related to the market-
impact cost . If the option as well-approximated by having a constant gamma, we nd that target
t
is actually the Black-Scholes hedge (Theorem 1). This similarity to Garleanu and Pedersen (2009)
suggests that we can think of delta-hedging in an illiquid market as a Merton optimal investment
problem where the Merton portfolio is the Black-Scholes hedge portfolio. For more general options,
the target target
t
has an extra term accounting for the non-zero third derivative with respect to
spot of the Black-Scholes option price (Theorem 2).
The subsequent sections are as follows. We motivate our assumptions and formally set up the
problem in Section 2. Closed-form solutions for two general cases are presented in Section 3. In
Section 4, we give a number of applications and implications of our result. In Section 5, we give
a discrete-time formulation of the problem. This is necessary for performing discretized numerical
simulations on TAQ data. The discrete-time solution is used in the hedging simulation in Section 6.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 5
2. Problem Setup
In this section, we rst give a heuristic justication for our objective (6). The resulting optimal
policy for this objective is then rigorously proved in Section 3. We give the market impact model,
our model for a European option, and our formal objective in the next three subsections.
2.1. Market Impact Model
Let (, T, P) be a complete probability space with T
t
being the ltration generated by W
t
. Consider
the trading model
X
t
=X
0
+
_
t
0

u
du
where X
t
is the number of shares held by the agent and
t
is the intensity of trading. The funda-
mental price is given by
P
t
=P
0
+
_
X
t
X
0
_
+W
t
(1)
where >0 is the coecient of permanent impact, >0 is the absolute volatility of the fair value,
and W
t
is a standard one-dimensional Brownian Motion. Using a linear Brownian Motion rather
than a geometric one is appropriate over the short time horizons considered in the paper and leads
to dramatic simplications in our result. Throughout this paper we neglect interests rates, which
is appropriate for the short time-periods under consideration.
To model temporary impact, we assume that the
t
th share (at time instant t) will cost a linear
premium
t
over the fair value

P
t
(
t
) P
t
=
t
where > 0 is the market impact parameter. Therefore, the total temporary impact (rate) of
buying
t
shares is
_

t
0
[

P
t
() P
t
]d =
_

t
0
d =

2
t
2
. (2)
We can think of temporary impact as coming from a limit-order book with constant depth 1/
and instant resilience (see, for example, Alfonsi and Schied (2009), Predoiu et al. (2010)). In such
a model, purchasing
t
shares would consume all the shares priced from P
t
to P
t
+
t
on the book,
thus pushing the execution cost of the last share up by
t
. The limit orders that were eaten up
are replaced immediately after execution. Thus our temporary impact coecient is related to
the corresponding term in Almgren and Chriss (2001) by =2.
2.2. European Option
We think of hedging a European contingent claim over a nite time horizon [0, T]. For a hypothetical
small trader whose execution has no price-impact and trades in a complete market, the value of
the options price g
_
t, P
t
_
is a function of the time t and the price P
t
of the underlying asset,
g : [0, T] RR.
Hence, the agents total portfolio value at time t is given by X
t
P
t
+g(t, P
t
). We assume the price
at t =T is given by g
0
: RR and the value for t [0, T) is prescribed by Feynman-Kac to be the
solution of the PDE
g(t, p) +

2
2
g

(t, p) =0 for t, p [0, T) R and g(T, ) =g


0
. (3)
Here g represents derivatives with respect to t and g

and g

represent derivatives with respect to


p. We can view this as the Black-Scholes option-pricing PDE for the arithmetic Brownian Motion
Author: Option Hedging with Market Impact
6 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
P
t
and zero interest rates so that g has the interpretation of the option price in the corresponding
(no impact-cost) complete-market. For notational simplicity, we dene
(t, p) =g

(t, p)
to be the gamma of the option.
For a large trader who does face market-impact, the terminal wealth (or revenue) R
T
is the sum
of the options value, the stocks value, and the cost of acquiring the position,
R
T
= g(T, P
T
)
. .
T option value
+ X
T
P
T
. .
T stock value

_
T
0
_

t
0

P
t
()d
. .
acquisition cost
dt . (4)
Using integration by parts, the stock dynamics (1), the total temporary impact (2), and Feynman-
Kac (3), we may rewrite the terminal wealth as
R
T
=R
0
+
_
T
0
_
X
t
+g

(t, P
t
)

dP
t

2
_
T
0

2
t
dt ,
where
R
0
=g(0, P
0
) +X
0
P
0
.
It is now convenient to introduce the variable
Y
t
=X
t
+g

(t, P
t
)
with the interpretation as the portolios net delta exposure. This variable replaces net shares
held X
t
as a more relavent variable. Y
t
also has the interpretation of distance from being delta-
hedged as the ideal net delta-exposure is 0. Hence, we may decompose our wealth process into
three intuitive parts
R
T
=R
0
+
_
T
0
Y
u
dW
u
. .
Intraperiod
Fluctuation
+ Y
u

u
du
. .
Permanent
Impact

2
u
du
. .
Temporary
Impact
. (5)
If we interpret the interval [0, T] as the trading day, we have the wealth R
T
as the sum of the
uctuation during the trading day and the liquidity cost from permanent and temporary impacts.
Finally, we make the assumption
Assumption 1. Let g(t, p) have a gamma (t, p) that is bounded everywhere and Lipschitz in p
with a Lipschitz constant that is independent of t.
The above assumption holds, e.g., for the Delta and Gamma of calls and puts except at expiry and
so this formula will be valid for hedging except on the expiration day of the option. The case of
hedging at expiry is well-known for being dicult because of the divergence in Gamma and these
complications are beyond the scope of this paper.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 7
2.3. Objective
Assumption 2. Assume and P
T
are L
2
random variables satisfying
1. the conditional distribution P[ [ T
T
] depends only on P
T
.
2. the conditional distribution P[P
T
[ T
T
] is independent of T
T
and has mean zero.
We dene our linear-quadratic objective to be
inf

E
_

2
(Y
T
P
T
)
2
. .
Terminal
Mishedging
+
_
T
0

2
2
Y
2
u
. .
Running
Mishedging
Y
u

u
. .
Permanent
Impact
+

2

2
u
..
Temporary
Impact
du
_
, (6)
where and P are to be dened later and the control set is given by
:=
_
predictable : E
_
T
0

2
s
ds < and
t
C(1 +[Y
t
[) a.s. for all t
_
.
This objective is standard in the literature and can be thought of as a nite-horizon adaptation
of Garleanu and Pedersen (2009). The relationship to utility optimization in the portfolio theory
literature is well-known and given in Fabozzi (2007, Chapter 2). The rst term in (6) gives a
terminal (time t = T) penalty for being mishedged. We allow this terminal penalty to be quite
exible so that it can accommodate multiple interpretations (through multiple intepretations of
and P: see Examples 1, 2, and 3). The running terms penalize for intraperiod mishedging
uctuations, and permanent and temporary impact. Equation (6) has a natural interpretation
of balancing the temporary and permanent liquidity costs with the penalty for variance of the
intraperiod and terminal mishedging errors.
The state variables have dynamics
dP
t
=
t
dt +dW
t
(7)
dY
t
=(1 +(t, P
t
))
t
dt +(t, P
t
)dW
t
. (8)
(Recall that is the gamma of the option). It is easy to see that P
t
is a continuous semimartingale
and hence predictable. We see from (1) that since is bounded that Y
t
is well-dened for all
. We see from (7) that every share purchased
t
pushes P
t
up by because of the permanent
impact. We also see from (8) that every share purchased
t
increases his net delta position Y
t
by
1+(t, P
t
): 1 for the increase in the stock position and (t, P
t
) for the permanent impacts eect
on the options delta.
Example 1. We could choose to interpret T as the maturity of the option. In this case, the
mishedging penalty is proportional to the time to maturity (i.e. the time remaining to potentially
accumulate mishedging). Hence, one might choose to not include a terminal term (P
T
= 0 and
=0) and no additional penalty is incurred for being mishedged at maturity
E
_

2
(Y
T
P
T
)
2
_
=0.
However, the individual trader may choose to incorporate a penalty because he nds it too risky
to be too mishedged even at expiry (see Remark 1).
Example 2. For general intraday trading, T would represent the close of trading. However, there
is an unhedgable potential overnight jump and trading does not resume until the next morning.
Therefore, we might choose to benchmark the objective to the opening of trade the next morning.
The choice of benchmarking to the morning is natural because the objective can be reset for the
Author: Option Hedging with Market Impact
8 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
next trading day. (Of course, the agent may choose to not incorporate a terminal penalty, even on
the day of the options maturity, as in Example 1).
For practical applications, adding a terminal mishedging penalty is more realistic. Without one,
the agents incentive to be hedged would (unrealistically) vanish as he approaches the closing bell.
This extra penalty reects the need to be well-hedged for the overnight jump, during which, he is
unable to trade.
To make this more precise, let T

be the opening of trading the next morning, T

T, and let
T = T

T denote the overnight time interval. Between T and T

, the agent is not allowed to


trade but the underlying price continues to evolve. From (4), we obtain
R
T
=g(T

, P
T
) +X
T
P
T

_
T
0
_

t
0

P
t
()ddt
=R
T
+g(T

, P
T
) g(T, P
T
) +X
T
(P
T
P
T
)
=R
T
+
_
T

T
[g

(t, P
t
) g

(T, P
T
)] dP
t
+Y
T
(P
T
P
T
) .
Then we can set the terms in the terminal penalty to
=
_
T

T
[g

(t, P
t
) g

(T, P
T
)] dP
t
P
T
=P
T
P
T
.
and the terminal penalty reduces to the quadratic (in Y
T
)
E
_

2
(Y
T
P
T
)
2
_
=

2
E
_

2
T
Y
2
T
2E[P
T
[ T
T
] Y
T
+E
_

2
[ T
T

_
where
2
T
=
2
T. Observe that the distribution of does not depend on P
T
and so satises
Assumption 2.
Example 3. One highly illustrative subcase of Example 2 is when we can approximate the option
as having a constant gamma, i.e.
g

(t, P
t
) g

(t, P
0
) +(P
t
P
0
) with R constant .
In other words, we assume (t, p) . Under this assumption, the terms in the terminal penalty
simplify down to
=
_
T

T
(P
t
P
T
) dP
t
P
T
=P
T
P
T
.
and so we can use ItoCalculus to derive
E[P
T
[ P
T
=p] =E
_

_
T

T
g

(u, P
u
) g

(T, P
T
) du [ P
T
=p
_
=0
(see Subsection 3.3 for the full calculation). Following directly from Example 2, the terminal objec-
tive would be
E
_

2
(Y
T
P
T
)
2
_
=

2
E
_

2
T
Y
2
T
+
2

.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 9
3. Continuous-time Solution
Let J(t, p, y) be a C
1,2,2
([0, T] RR) function. Then from the Martingale Principle of Optimal
Control, we have the HJB equation is
0 =inf

2
2
y
2
y +

2

2
+

J +(1 +)
y
J +
p
J
+
2

p
J +
()
2
2

2
y
J +

2
2

2
p
J
=
1
2
[ (y
p
J) (1 +)
y
J]
2
+

2
2
y
2
+

J
+
2

p
J +
()
2
2

2
y
J +

2
2

2
p
J . (9)
We will not solve the equation in full generality but will stick to two major sub-cases. For these, it
suces to to make the Ansatz,
J(t, p, y) =
A
2
(T t)
2
y
2
+A
1
(T t, p)y +A
0
(T t, p) . (10)
We will later verify that this Ansatz holds in the proofs of Theorem 1 and Theorem 2. Observe
that A
2
has no dependence on p. This requires Assumption 2 and will be key to our decomposition.
Then the equation breaks up into the following three equations

A
2
=
2

[(1 A

1
) (1 +)A
2
]
2
(11a)

A
1
=

2
2
A

1
+
1

[(1 A

1
) (1 +)A
2
] [A

0
+(1 +)A
1
] (11b)

A
0
=
1
2
[A

0
+(1 +)A
1
]
2
+A

2
+

2
2
_

2
A
2
+A

. (11c)
(Here A

and

A denote the derivative with respect to p and t, respectively). Observe that the
equation for

A
2
(11a) cannot hold in general: A
2
does not have a dependence on p but A
1
does.
The equation will nonetheless hold in the two cases of interest. These equations are subject to the
boundary conditions
A
2
(0) =E
_
P
2
T

A
1
(0, p) =

2
E[P
T
[ P
T
=p]
A
0
(0, p) =

2
E
_

2
[ P
T
=p

.
The trading strategy is then

t
=
Y
t
(1 +(t, P
t
)) (A
2
(T t)Y
t
+A
1
(T t, P
t
)) (A

1
(T t, P
t
)Y
t
+A

0
(T t, P
t
))

. (12)
Hence, the optimal policy comes from solving for A
1
and A
2
. However, they exhibit closed-form
solutions in two important broad cases. Before we move on to these two cases, we rst dene
K =1 + =
_

d =
[K[

2
T


K
_
. (13)
Author: Option Hedging with Market Impact
10 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
3.1. Constant Gamma Approximation
On short time scales, we can safely assume that the option has a constant gamma. The approx-
imation says that for small intraperiod uctuations of P
t
P
0
, the options delta varies linearly
with the stock price. We make the the same approximation as in Example 3
g

(t, P
t
) g

(t, P
0
) +(P
t
P
0
) with R constant . (14)
As in Example 3, the terminal penalty will be
E
_

2
(Y
T
P
T
)
2
_
=

2
E
_

2
T
Y
2
T
+
2

and does not depend on P


T
. Furthermore, the dynamics of (8) are now independent of P
t
,
dP
t
=
t
dt +dW
t
dY
t
=(1 +)
t
dt +dW
t
where is now a constant. The assumption considerably simplies the problem by dropping the
dependence on the state variable P
t
and eliminating the function A
1
.
In this subsection, we will make the additional assumption
Assumption 3. Assume that either
1. d 1 holds or
2. d <1 and [K[T +arctanh(d) 0.
Theorem 1. Let Assumption 3 hold. Then the optimal trading intensity is given by

t
=h(K(T t))Y
t
(15)
where the function h is dened by
h(x) =
_

_
tanh
_
x+arctanh(d)
_
[d[ <1
1 d =1
coth
_
x+arccoth(d)
_
[d[ >1.
(16)
Under the optimal trading strategy, Y
T
,= 0 a.s. Indeed, P[[Y
T
[ > M] > 0 for all M > 0. That is,
because of the market-impact costs, the position is not perfectly delta hedged, even at the terminal
time and there is a chance (albeit small) that Y
T
is far away from 0.
Proof of Theorem 1 The Martingale Principle of Optimal Control tells us that
M
t
=
_
t
0

2
Y
2
u
Y
u

u
+

2

2
u
du+J(t, Y
t
)
is a submartingale for all
t
and a martingale under the optimal control. The HJB equation (9)
simplies to
0 =

2
2
y
2
+

J(t, y) +
()
2
2
J

(t, y)
1
2
_
y KJ

(t, y)
_
2
(17)
(t, y) =
1

(y KJ

(t, y)) (18)


wehre J

and J

are derivatives with respect to y. The ansatz (10) then simplies to


J(t, y) =A
2
(T t)
y
2
2
+A
0
(T t) .
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 11
and (11a) and (11c) simplify to

A
2
=
2

( KA
2
)
2
,

A
0
=
()
2
2
A
2
.
We can see that

A
2
is a parabola opening downward in A
2
with vertex at the point (

K
,
2
) and
left and right critical points are A

2
and A
+
2
, respectively where
A

2
=

K


[K[
.
If A
2
(0) =
2
T
(A

2
, A
+
2
) then

A
2
(T t) >0 and so A
2
(T t) A
+
2
as T t . This corresponds
to the case when 1 <d <1 where the solution is given by
A
2
(T t) =

K
+

[K[
tanh([K[(T t) +arctanh(d))
A
0
(T t) =
()
2
2K
_
(T t) +

K
_
log cosh
_
[K[(T t) +arctanh(d)
_
+
1
2
log(1 d
2
)

_
.
(19)
Similarly, if A
2
(0) =
2
T

_
A

2
, A
+
2
_
then A
2
is xed at the critical point. This corresponds to the
case [d[ =1 and the solution is given by
A
2
(T t) =

K
+
d
[K[
=
2
T

A
0
(T t) =

2

2
2

2
T
(T t) .
(20)
Finally, if A
2
(0) > A
+
2
then

A
2
(T t) < 0 and so A
2
A
+
2
. This corresponds to the case d > 1
where the solution is given by
A
2
(T t) =

K
+

[K[
coth([K[(T t) +arccoth(d))
A
0
(T t) =
()
2
2K
_
(T t) +

K
_
log sinh
_
[K[(T t) +arccoth(d)
_
+
1
2
log(d
2
1)

_
.
(21)
These three cases cover case 1 of Assumption 3. Under case 2 of Assumption 3, equation (21) also
holds and A
2
(T t) is well-dened for t [0, T] since [K[(T t) +arccoth(d) 0 for all t [0, T].
Thus we have a solution J to the HJB equation (18) and we see that M
t
has non-negative drift
for all and is a local martingale for the optimal given in (15). To conclude the proof, we need to
show that M
t
is a submartingale for all and a martingale for the optimal . Thus, it suces
to show E[M]
T
< for all . Observe that the function A
2
(t) is uniformly bounded by some
C >0 independent of t [0, T] so that
E[M]
T
=E
_
T
0

(t, Y
t
)

2
dt
C
2
E
_
T
0
[Y
t
[
2
dt . (22)
Author: Option Hedging with Market Impact
12 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
Itos Lemma gives us
EY
2
t
=Y
2
0
+E
__
t
0
2Y
s
dY
s
+d[Y ]
s
_
=Y
2
0
+E
__
t
0
C

s
Y
s
+
2

2
ds
_
Y
2
0
+E
__
t
0
C

[Y
s
[
2
+
2

2
ds
_
for some C

, C

>0 independent of t [0, T]. Thus, a standard application of Gronwalls Lemma


yields E[M]
T
< and we have (15) is the optimal policy.
The second part also follows from the fact that Y
T
is given by a linear SDE. Its solution
Y
t
=e

t
0
Kh(K(Ts)) ds
_
Y
0
+
_
t
0
e

s
0
Kh(K(Tu)) du
dW
s
_
has a density that is non-singular with respect to the Lebesgue measure on R as h is bounded.
In Almgren and Chriss (2001) is an urgency parameter which dictates the speed of liquidation:
the higher the faster the initial liquidation. In Garleanu and Pedersen (2009), this term gives
the relative intensity of trading
1
, that is, the higher , the faster the agent trades towards the
Merton-optimal portfolio.
The agents trading target for shares X
t
is the Black-Scholes delta hedge,
target
t
=g

(t, P
t
) .
He constantly trades towards this target but is prevented from holding the exact Black-Scholes
delta hedge by the convex temporary impact stemming from limited liquidity. Hence, the trading
intensity
t
is proportional to the degree of mishedge Y
t
and the urgency parameter . There is
a greater penalty to being mishedged with higher underlying volatility and risk aversion so
these parameters increase urgency. Similarly, a more illiquid market (higher ) makes trading more
costly, which decreases trading intensity.
The solution falls into three cases depending on whether h(), the trading intensity proportion,
is decreasing, at, or increasing in time, i.e. whether the agent trades less intensely, is at, or more
intensely (respectively) towards the end of the trading period. This, in turn, depends on the value
of a constant d which gives the relative size of the overnight jump P
T
versus permanent impact
and noise dP
t
.
The interpretation of d is clearer when there is no permanent impact ( = 0) and we rst
handle this case. In this case, the cost of being mishedged comes purely from running and terminal
mishedging terms. Intraperiod trading can be thought of as primarily hedging out the intraperiod
uctuations dP
t
while trading near the end of the period is primarily for hedging the terminal
jump P
T
. When d <1, the majority of the uctuations occur during the day, and the solution is
marked by t h(K(T t)) decreasing as t approaches T. That is, the agent is more concerned
about hedging intraperiod uctuations dP
t
and relaxes hedging intensity to reduce hedging cost
as the end of the period approaches. This is the case in Example 1 where d =0. The second case
is when d =1 and hedging risk for the intraperiod and end-of-day are weighted equally. Then the
agents trading intensity is constant in time. The nal case is when d >1 and the terminal jump
is large compared to the daily uctuations. So the agent trades more intensely with time and
t h(K(T t)) increases towards d as t approaches T. This case could occur in Example 2.
1
The a/ in Proposition 5 of Garleanu and Pedersen (2009) is equivalent to h in our setup
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 13
Units of Fundamental Quantities
Y
t
sh 1/$ $/sh
2
P $/sh sec $/sh
2
sh
2
/$
sh/sec $/sh/

sec J $
Units of Derived Quantities
K 1 d 1 1/sec
Table 1 Units for constants: sh is a share, sec is a second, and $ is a dollar.
Remark 1. If we consider the case when [0, T] is the trading day, the last case (21) when d >1
is the most realistic. Options market-makers typically increase their hedging towards the close
of trading to minimize overnight exposure. The trader may choose a lower d as options expiry
approaches, selecting d = 0 on the day of options expiry. However, this strategy may be deemed
too risky and the trader may choose d >0 even at expiry to avoid the risk of being mishedged.
When there is permanent impact ( >0), the running penalty comes from two sources in (6), the
running mishedging term and the permanent impact. The running mishedging term is as above but
the permanent impact term is more subtle. If the trader is in a net long delta position, he sells the
position to remain hedged, incurring a capital loss from the downward pressure on his inventory.
A buy to hedge out a net short delta position incurs a similar loss and so the permanent impact
creates a cost for being mishedged in either direction. The terms for A
2
in (19), (20), and (21) have
an extra term which accounts for the permanent impact. The size of d balances the running costs
(both from running uctuations and the permanent impact eect) with the terminal cost of being
mishedged.
3.2. Comparative Statics
We give some comparative statics for our solution. The eect of the temporary impact term
is the easiest to understand. As the cost of hedging decreases, it becomes optimal to trade more
aggressively to minimize mishedging error (see Figure 2c). The eect of risk aversion is similar.
As increases, so does the penalty for being mishedged and there is a stronger incentive to trade
more aggressively (see Figure 2e). To understand the eect for intraperiod volatility , recall that
the agent is trying to hedge both intraperiod volatility (via intraperiod trading) and volatility
from the terminal jump
T
(via trading near the end). As intraperiod volatility increases while
the terminal jump size
T
remains constant, the agent optimally trades more aggressively initially
while still ending at the same trading intensity at time t =T (see Figure 2b). The eect of increasing
terminal jump volatility
T
while keeping intraperiod volatility constant is roughly the opposite
(see Figure 2f): it increases trading near t =T while trading during the rest of the period remains
relatively unaected. Note that with higher
T
, there is a need to trade slightly more intensely
even at the beginning of the period t =0 as this impacts the mishedging at the end of the period
t =T.
and the permanent impact have more complex eects, which are qualitatively similar to
each other. There are two competing eects at play, whose importance diers with the time to
maturity. Firstly, there is an adverse-impact eect stemming from the permanent impact. To
see this, let Y
t
<0 so the agent optimally wishes to purchase shares. This purchase will push up
each shares value by , hence resulting in a decrease of Y
t
<0 in the stock value of the agents
portfolio. In general, rebalancing under permanent impact always hurts his existing position, giving
the agent an incentive to trade less aggressively in the hope that subsequent price-uctuations will
automatically rehedge his portfolio. The second eect is dubbed the leverage-eect. With =0,
the purchase of a single share increases the traders net delta position Y by 1. However, if >0
Author: Option Hedging with Market Impact
14 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
(a) varying (b) varying
(c) varying (d) varying
(e) varying (f) varying T
Figure 2 Comparative statics of (KA2(T t) )/ as a function of t. Parameters (unless otherwise specied):
=5.0, =0.2, =0.2, =0.2, =2.0, T =0.4, T =1.0.
then his net delta position actually increases by K =1 + due to the permanent impacts eect
on the options price . When >0, the eect of a purchase on the portfolios net delta position
is leveraged by the permanent impact on the options delta. We can think of this as eectively
lowering the cost of trading, (i.e. lowering the cost of achieving a xed change in delta). Hence,
the agent optimally increases the intensity of trading. When <0, the eect of a sale or purchase
is diminished, if not reversed. We can think of this as reducing the eectiveness of trading on the
net delta position or increasing the cost of rebalancing. Either way, the agent optimally decreases
the intensity of trading.
The relative importance of the two eects can be seen in the expression

t
=
1

_
Y
t
KJ

(t, Y
t
)

.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 15
The leverage-eect acts through KJ

while the adverse impact acts through Y


t
. In Figure 2,
d >1 and the pressure to be hedged (KJ

) increases towards the end of the period. Hence, near


the beginning of the period, the adverse impact is the overriding concern and trading intensity is
less aggressive with higher because of the adverse-impact eect (Figure 2d) Near the end, the
agent is more concerned about hedging the terminal jump. This combined with the ease of hedging
from the leverage eect acting through KJ

(t, Y
t
) implies the agent trades more aggressively with
higher near the end. The eect of is similar except that it only directly aects the leverage
eect. With higher , the increase in the leverage eect allows the agent to optimally trade more
aggressively, particularly near the end of the day (Figure 2a). However, like varying
T
, because
aects mishedging near the closing bell, it has an indirect eect on the trading intensity during
the middle of the day as well.
3.3. General Options, No Permanent Impact
We can relax the constant gamma assumption and allow for general options. However, to make the
solution tractable, we need to dispense with permanent impact ( =0). We are still able to obtain
a fairly explicit solution for the control in (12) through A
2
and A
1
. Recall that we have made
the assumptions Assumption 1 on our option.
Theorem 2. The optimal control is given by

t
=
A
2
(T t)Y
t
+A
1
(T t, P
t
)

(23)
with
A
2
(T t) =
_

_
tanh((T t) +arctanh(d)) d <1

2
T
d =1
coth((T t) +arccoth(d)) d >1
and
A
1
(T t, p) =
_

_
1

1 d
2
sech((T t) +arctanh(d)) E[F(p +B
Tt
)] d <1
exp
_

2
T
(T t)
_
E[F(p +B
Tt
)] d =1
1

d
2
1
csch((T t) +arccoth(d)) E[F(p +B
Tt
)] d >1
where we assume B is a new Brownian Motion, the constants and d are dened by
=
_

d =

2
T

=T
and the function F() is dened by
F(p) =

2
E
_
P
T

P
T
=p
_
.
Proof In this case, the equations (11) then reduce to

A
2
=
2

A
2
2

A
1
=

2
2
A

A
2
A
1

A
0
=
A
2
1
2
+
2
A

1
+

2
2
_

2
A
2
+A

.
Author: Option Hedging with Market Impact
16 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
The solution for A
2
then follows immediately. (Observe that the range of tanh includes (0, 1) and
the range of coth includes (1, )). Feynman Kac then gives us
A
1
(T t, p) =E
_
exp
_

_
T
t
A
2
(T s)

ds
_
F(p +B
Tt
)
_
(24)
where B is any Brownian Motion. We have by Assumption 1 that g

is bounded hence g

is
Lipschitz. Since we have
F(p) =

2
E[P
T
[ P
T
=p]
=

2
E
_

2
_
T

T
g

(u, P
u
) g

(T, P
T
) du [ P
T
=p
_
we can show that F is Lipschitz as well. Hence A
1
given by an expectation in (24) is well-dened.
Observe that
sech(arctanh(d)) =

1 d
2
d <1
csch(arccoth(d)) =

d
2
1 d >1
from which our expressions for A
1
then follow. Finally, since the PDE for A
0
is linear, it is straight-
forward to verify it exists and so we may dene J as in (10).
For the verication argument, dene
M
t
=

2

2
Y
2
u
Y
u

u
+

2

2
u
du+J(t, P
t
, Y
t
)
Since J solves the HJB equation (9), we have that M
t
has non-negative drift for all and zero
drift for the optimal policy given in (23). Again, we shall show that for all , the expected
quadratic variation E[M]
T
<. We have
E[M]
T
=E
_
T
0

A
2
(T t)Y
t
+A
1
(T t, P
t
)

2
dt
C

E
_
T
0
[1 +Y
t
+P
t
[
2
dt , (25)
where C

is a positive constant. Since A


1
is bounded and A
2
is Lipschitz in p, we have for all
that
E
_
T
0
[1 +Y
t
+P
t
[
2
E[1 +Y
0
+P
0
[
2
+2E
_
T
0
(1 +Y
s
+P
s
) [1 + (1 +(s, P
s
))]
s
+
2
(1 +(s, P
s
))
2
ds
E[1 +Y
0
+P
0
[
2
+CE
_
T
0
[1 +Y
s
+P
s
[
2
ds (26)
by the denition of and the fact that (t, p) is bounded (1). A standard application of Gronwalls
Lemma yields E[M]
T
<.
The optimal control diers from the constant gamma case in that the target is no longer the
delta-neutral portfolio. There is a new term A
1
(T t, P
t
), which biases the target portfolio to
account for the third-derivative of the option price, or the delta of the gamma. If this quantity
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 17
vanishes, then F(p) vanishes and so does A
1
. To see this, consider the setup of Example 2 and
observe that by Itos Isometry that
F(p) =

2
E
_

2
_
T

T
g

(u, P
u
) g

(T, P
T
) du [ P
T
=p
_
.
F has an interpretation as the average change in delta of the option over the period T to T

. We
saw in Example 3 that F and hence A
1
vanish for options with constant gamma.
More generally, if the option has vanishing delta of gamma (that is, g

vanishes) then for t >T,


we have by Itos Lemma that
E[g

(t, P
t
) g

(T, P
T
)] =E
__
t
T
g

(u, P
u
)dW
u
+
1
2
g

(u, P
u
)
2
du
_
=0.
Hence, F and A
1
also vanish. For an option with positive delta of gamma (g

>0), an agent who


is delta-neutral at the close T will actually have a positive expected overnight delta exposure F.
He corrects for this by adding a negative bias to his trading target for shares X
t
is
target
t
=g

(t, P
t
)
A
1
(T t, P
t
)
A
2
(T t)
.
The eect is the opposite if g

<0. Observe that for constant p, the magnitude of A


1
increases as
t T (see Figure 3). This follows since the bias corrects for an overnight delta slippage and is less
relevant far away from the closing bell.
We plot the oset A
1
in Figure 3. Compared with Figure 3 we omit the plots for (as = 0)
and (as can only aect A
1
through a nonzero ). As expected, a higher reduces the tracking
bias A
1
as it increases the penalty for not being delta-hedged, although this dierence disappears
as t T (see Figure 3a.) Varying and have inverse eects: increasing is similar to increasing
(increasing the penalty for mishedging) while increasing has the inverse eect (increasing the
market impact costs). The results are plotted in Figure 3b and Figure 3c. Finally, increasing
T
also increases the running costs of being mishedged (it increases A
2
in Figure 2f). This in turn
decreases A
1
(see (24) and Figure 3d).
4. Asymptotics and Applications
4.1. Small Market-Impact Cost Limit
We are interested in the limit as both the temporary and permanent impacts vanish. For instance,
in the constant gamma case (Theorem 1) the trading intensity
t
is proportional to the distance
from being delta hedged Y
t
=X
t
+g

(t, P
t
) as given in

t
=coth
_
K(T t) +arccoth(d)
_
Y
t
.
We observe in Figure 2d that the prole of trading intensity attens as a function of time with
decreasing and from Figure 2c, we see that the trading intensity shifts upward with decreasing .
Hence, we observe the trading intensity both increasing and attening as market-impact decreases.
In the limit as 0 and 0, we have K 1, . If
T
> 0 then d as well. Therefore
A
2
(T t) and
t
/Y
t
and the process Y
t
is driven back to zero very quickly. That is,
the agent aggressively drives X
t
towards g

(t, P
t
). Hence, we recover the Black-Scholes equation in
the limit of vanishing transaction costs. Similarly, for general options and no permanent impact
(Theorem 2), we still have A
2
(T t) and A
1
(T t, p) 0 for all p and t < T and so again

t
/Y
t
. The following theorem makes this idea precise.
Author: Option Hedging with Market Impact
18 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
(a) varying (b) varying
(c) varying (d) varying T
Figure 3 Comparative statics of A1(T t) as a function of t. Parameters (unless otherwise specied): =0.2, =
0.2, =2.0, T =0.4, T =1.0. We normalize E[F(p +BTt)] to be 1.
Theorem 3. Assume that the assumptions of either Theorem 1 or Theorem 2 hold. In the limit
of vanishing market-impact costs, we recover the Black-Scholes delta hedge. That is |Y | 0 as
, 0 where | | is the L
2
(P)-norm and is the Lesbegue measure on [0, T].
Proof If we make the assumptions of Theorem 1, we have from (26) that
E
_
Y
2
t

_
Y
0
+
2

2
T
_
exp
_

_
t
0
Kcoth
_
K(T s) +arccoth(d)
_
ds
_
.
and the result then follows from noting that , K 1, and d in our limit. In the setup of
Theorem 2, we the result follows similarly by (26) and the fact that A
2
(T t) and A
1
(T t, p)
0 in the limit.
4.2. Small Intraperiod Mishedging Penalty Limit
We consider the case when the penalty for the terminal jump is large compared to the intraperiod
hedging penalty. Mathematically, this is the limit when 0 and and
T
are held constant.
The objective in this limit is given by
inf

E
_

2
T
2
Y
2
T

_
T
0
Y
u

u
+

2

2
u
du
_
(27)
where we have removed the intraperiod mishedging penalty. We include this result because (28)
clearly demonstrates the increasing nature of the functions
t
/Y
t
as t T in terms of algebraic,
rather than transcendental, functions.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 19
Theorem 4. We assume that the option has constant gamma as in Theorem 1. Then for small
intraperiod mishedging, the optimal trading intensity
t
can be written as

t
=
d
1 +Kd(T t)
Y
t
. (28)
Proof If we make the Ansatz (10) then again A
1
= 0 and we separate the resulting HJB by
powers of y. This yields
(t, y) =
1

( KA
2
(T t))y
and

A
2
=
1

_
(1 +)A
2

2
A
2
(0) =
2
T

A
0
=

2

2
2
A
2
A
0
(0) =0.
By a similar argument as in Theorem 1, we obtain that
A
2
(T t) =
1
K
_
+
d
1 +dK(T t)
_
A
0
(T t) =

2

2
2K
_
(T t) +

K
log
_
1 +dK(T t)
_
_
.
The result then follows.
For a sense of how the execution of the policy in Theorem 4 compares to the full case in Theo-
rem 1, see Figure 4. Observe that without an intraperiod mishedging penalty, the trading intensity
ratio
t
/Y
t
(see (28)) is lower than with the intraperiod penalty (see (15)), especially during the
beginning of the period. The extra trading comes from hedging out uctuations during the period.
Intuitively, the optimal investor trades less aggressively when only faced with a terminal and no
running mishedging penalty. As in the d >1 case in Theorem 1, trading becomes more aggressive
towards the end of the period.
If we take the interpretation of [0, T] as the trading day, then this approximation is valid when
the the overnight jump is large compared to the daily uctuations. This may be the case in advance
of a major post closing-bell announcement.
Figure 4 Comparison of trading intensity with and without intraperiod mishedging penalty. The parameter values
are dened as in Figure 2
.
Author: Option Hedging with Market Impact
20 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
4.3. Restriction on the direction of trading
If executed through a broker, the delta-hedging problem is more challenging. Regulatory policy
mandates that brokers can only either buy or sell for any given client order. This is to prevent
market-manipulation and to protect clients from potentially unscrupulous dealers. We solve the
optimal delta-hedging problem under these constraints.
Without loss of generality, we impose a buy-only restriction on trading. That is, we add the
constraint
t
0 a.s. to the minimization of our objective (6),
inf

+
E
_

2
(Y
T
P
T
)
2
+
_
T
0

2
2
Y
2
u
Y
u

u
+

2

2
u
du
_
, (29)
where
+
= , 0 a.s.. While the problem is still Markovian, a closed-form solution is no-
longer readily available. We use a policy-improvement algorithm that assumes a Markovian optimal
policy
t
=(t, Y
t
) to solve this problem. As a check, the same policy-improvement code was used
to solve the Simplied Model in Section 4.2 and compared against the analytic solution found in
Section 4.2. It obtained continuation-value functions J accurate to within .1%.
A comparison of the restricted and unrestricted cases are plotted in Figure 5. Observe the
asymmetry in trading policy (t, y) in Figure 5c. For negative values of y, (t, y) is positive to
reduce the delta-hedging error, that is the agent still purchases stock as in the unrestricted case
(compare with Figure 5a). For positive values of y, an unrestricted agent would sell shares but a
restricted agent cannot do so, and the binding constraint forces to be zero (see Figure 5c). There
is a no-trade curve y(t) 0 such that the agent trades if and only if he his current net delta position
is below this level, i.e.
t
=0 when Y
t
y(t) and
t
>0 when Y
t
<y(t). This curve is marked in the
plot in Figure 5c. Observe that y(0) <0 and t y(t) is an increasing function. Hence, far from the
terminal time (t T), the agent does not purchase stocks even when he is net short delta. The
result is interpreted thus: a selling-restricted agent is hesitant to purchase stock only to see his
delta position become positive before T due to stock-price uctuations. However, y(T) =0 so this
eect disappears as t T: as the time remaining for Y
t
to uctuate above 0 runs out, the agents
rule becomes buy if I am net short delta.
In the restricted case, the asymmetry in execution strategy for translates into an asymmetry for
the value function J (compare the restricted case Figure 5d with the unrestricted one Figure 5b).
For negative values of Y
t
, J behaves similarly in both the selling-restricted and non-restricted cases
as the optimal strategies are similar. For positive values of Y
t
, J is signicantly higher in the
selling-restricted case as the control cannot be exercised to reduce the hedging error. The dierence
between the two cases represents the premium of being able to sell.
4.4. Stock Pinning
The increase in trading intensity for high gamma near the expiry is the mechanism behind stock
pinning. A put or call option that is about to expire with a large outstanding institutional interest
and with a stock price near a strike level can induce a so-called pinning eect whereby the stock
price at the close of trading on expiry is pinned to the strike level. Empirically, this manifests
itself in the observation that the distribution of end-of-day stock prices on these days deviates
substantially from the distribution of end-of-day stock price on any other days. That is, the stock
price clumps around the strike level at the close of such an option expiry (Avellaneda and Lipkin
2003).
A call or put option near expiry with the underlying prices near its strike exhibits a large positive
gamma. When institutional investors are net short gamma (perhaps because they collectively short
the put or call) option market makers are net long gamma. The market-makers will hedge their
position by buying stock as the price dips and selling it as the price increases, thus stabilizing
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 21
(a) Unrestricted (t, y) (b) Unrestricted J(t, y)
(c) Restricted (t, y) (d) Restricted J(t, y)
Figure 5 Plots of the continuation-value function J and the optimal policy for the unrestricted case (Figure 5b
and Figure 5a) and the case when trading is restricted to purchases 0 (Figure 5d and Figure 5c). In
the restricted case, we also plot the no-trade boundary in black, above which (t, y) =0 (see Figure 5c).
Parameters: =5., =.4, =.2, =.2, =2., T =.4, T =1.
the price. Then as Figure 2a shows, a large gamma implies the agent trades intensely as expiry
approaches and as his trades are stabilizing, the permanent impact pins the stock price at the
strike. Since the gamma peaks near expiry with the underlying price near the strike, the eect
manifests itself as pinning near the strike.
Our model provides some insight into this model. Observe that for a call option, the delta is
given by
C

(t, p) =E[1
P
T
>0
[ P
t
=p] =N
_
p

T t
_
and so the gamma is given by
C

(t, p) =
1
_
2
2
(T t)
exp
_

p
2

2
(T t)
_
. (30)
For small uctuations of p, the gamma appears to be at. Let g =NC where N is a large positive
number representing the number of call options held. We make the constant gamma assumptions
of Theorem 1, where now the gamma will be NC

(t, p) for the gamma of the entire position.


Author: Option Hedging with Market Impact
22 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
It turns out that this assumption is valid for large N and moderately large values of T t. For
simplicity, we take d =1 so that
t
=Y
t
and we rewrite (7) and (8) as
dP
t
=
1

(dY
t

t
dt)
dY
t
=(1 +)Y
t
dt +dW
t
.
Hence, we have that the stationary distribution of Y
t
is
Y
t
A
_
0,
()
2
2(1 +)
_
.
That is, both
t
and Y
t
are distributed with standard deviation proportional to

N in the limit as
N . Hence, we have that the standard deviation of P
T
1/

N as N . In other words, if
P
0
starts close to the strike then P
t
will remain not deviate far and the price is pinned to the strike.
We can see from (30) that near the strike (p =0) the gamma of the call option is at in moneyness.
Hence, our constant gamma approximation is valid in the limit as N as the pinned stock only
sees a at gamma surface.
4.5. Intraday Trading Patterns
If we continue with the interpretation of the trading period as the trading day, we see that it is
consistent with the intraday trading pattern. The typical daily prole is U-shaped, with the most
intense trading occurring during the opening and close (see Figure 6a). Compare Figure 6a with
the graph of the average trading volume from delta hedging in Figure 6b as predicted by our model,
E[
t
[
with Y
0
N(0,
0
). The distributional assumption on Y
0
comes from the overnight uctuations
in the underlying aecting the delta of the option. The expression for E[
t
[ is given in (31). The
parameters for the plot are chosen to reect realistic values and are explained in Example 4.
We see that our model predicts a trading rate that is roughly 20% of the trading for (NYSE:BA).
More importantly, the trading volume prole roughly corresponds to the double-humped trading
pattern that is observed in the empirical data. Near the open, the high E[
t
[ comes from the
high initial variance in Y
t
as the trader trades down his net delta position from the previous
evenings jump. Near the close, the high E[
t
[ comes from the increase in h near t = T, which
represents aggressive trading in anticipation of the coming evenings overnight jump. Clearly the
intraday trading volume is determined by many additional factors. However, on stocks with a large
outstanding interest, Figure 6 illustrates how delta hedging may account for a signicant fraction
of the trading.
Example 4. We choose =.04 based on the corresponding value for Boeing (NYSE:BA) in Fig-
ure 2 (see the chart for units). We set
0
= (1.5
T
) with
T
from Figure 2. This corresponds
to the change in the options delta from an overnight jump in the stock of size 1.5
T
. If the book
were perfectly hedged the night before, we would see a jump of size
T
in the option value. The
factor 1.5 accounts for the book not being perfectly hedged at the close of the previous evening.
We set T =6.5 hours, the length of the trading day. =10
4
corresponds to a temporary impact
of 10 cents for trading 20% of ADV or 32 shares per second. We choose the permanent impact
=10
5
corresponds to a permanent impact of 1 cent for trading 32 shares. For the computation
of d, we choose
T
= 25 to penalize not just for the unhedgable overnight uctuation but for the
hedging costs to be incurred for future days starting the next morning. This penalty accounts for
We choose = 10
5
to corresponds to the gamma position of the entire book. In the expression
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 23
for the gamma of an option (30), (3.5 hours, 0) .1 is the gamma of an At-The-Money option
on Boeing (NYSE:BA) about 3 hours from expiry. Hence, for nearest month options we expect
the gamma of an individual option gamma to be between .01 and .1 (compare with Figure 2).
Therefore, our assumption on the net book corresponds to a position with one to ten million
At-The-Money contracts. Finally, we choose 10
8
. This can be thought of as corresponding to
a relative risk aversion of 1 for an agent with a portfolio of a million dollars.
We now give the derivation of E[
t
[ based on the assumptions of Theorem 1. It is clear that Y
t
is normally distributed with mean zero. From (8) and (15), we obtain a PDE for the variance of
Y
t
as
d
dt
EY
2
t
=2Kh(K(T t))EY
2
t
+()
2
.
We are interested in the more realistic case when d >1 (see Remark 1) whose solution is given by
EY
2
t
=e
2H(t)
(
T
)
2
+()
2
e
2H(t)
_
t
0
e
2H(s)
ds
where
H(t) =kK
_
t
0
coth(kK(T s) +arccoth(d)) ds
=log [sinh(kK(T t) +arccoth(d)) csch(kKT +arccoth(d))] .
Since we have
_
t
0
e
2H(s)
ds =sinh
2
(kKT +arccoth(d))
_
t
0
csch
2
(K(T s) +arccoth(d)) ds
=
sinh
2
(kKT +arccoth(d))
K
[coth(K(T t) +arccoth(d)) coth(KT +arccoth(d))]
we can compute the variance of Y
t
as
EY
2
t
=
sinh
2
(kK(T t) +arccoth(d))
sinh
2
(kKT +arccoth(d))

_
(
T
)
2
+
()
2
K
sinh
2
(kKT +arccoth(d))
_
coth(K(T t) +arccoth(d)) coth(KT +arccoth(d))
_
_
.
Since
t
is normally distributed, we have that
E[[
t
[] =
_
2

coth(K(T t) +arccoth(d))
_
EY
2
t
. (31)
This is plotted in Figure 6b with parameters chosen as in Example 4.
5. Discrete-Time Formulation and Solution
In this section, we assume a constant model as in Theorem 1. We need a discrete-time formulation
and solution to the hedging problem in order to perform the simulations on TAQ data in Section 6.
The simplest discrete strategy would be to evaluate the continuous-time strategy at each time
point in question, buying
t
t shares over the interval [t, t +t]. In eect, this is a forward Euler
Author: Option Hedging with Market Impact
24 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
(a) Average Minutely Trading Volume of Boeing (NYSE:BA) in the
period April 3rd, 2000 to March 30th, 2001.
(b) Intensity of trading E|t| for one trading day with parameters given
in Example 4. In our example, the trading predicted by our model
accounts for 20% of trading ow.
Figure 6 Intraday trading intensity, actual and projected in our model from delta hedging.
discretization of the underlying dynamics. Unfortunately, this strategy results in unrealistic and
costly overshooting.
To illustrate this, consider the mean value of Y
t
, Y
t
= E[Y
t
]. Under the continuous-time policy
of Theorem 1, it would behave as
Y
t
=Y
0

_
t
0
Kh(K(T s))Y
s
ds
and converges exponentially towards 0 but remains positive if Y
0
>0. On the other hand, under a
discrete Euler policy where we purchase
0
t shares, Y
t
would behave as
Y
t
=Y
0
(1 Kh(KT)t) .
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 25
Depending on the size of t, this may not be positive for xed Y
0
> 0. So while this strategy
converges to the continuous strategy as t 0, for xed t, it may overshoot zero, a result which
is known in the literature (see, for example, Ascher and Petzold (1998)). Indeed, simulations using
the discretized Euler-method exhibited this overhedging. In order to obtain well-behaved discrete
time solutions we must pose the discrete-time problem directly. The overshooting is so sever as to
lead to unstable solutions! It is especially prominent when temporary impact is small compared to
K
2
A
2
(T t) (see the denominator of discrete-time solution (33)). The continuous-time solution
is important as it is closed-form and provides important intuition for understanding the problem
dynamics. The discrete-time solution resembles the discretized continuous solution but makes an
important correction for overshooting (33) that is important for practical applications.
5.1. Discrete-Time Formulation
We will discretize the dynamics by imposing a lattice T
N
of N points of width t = T/N. That
is T
N
=
_
T
N
(Z [0, N))
_
. At each time t T
N
, the agent sets his strategy for the entire timestep
[t, t +t). Under such conditions, it is only necessary to consider the implied discrete-time process.
Hence, the agents stock position is given by
X
t
=X
0
+

uTn,u<t

u
while the dynamics for the stock price becomes
P
t
=P
0
+X
t
+

uT
N
,u<t
W
u
where W
kT/N
=W
(k+1)T/N
W
kT/N
. Then the new discrete-time value function becomes
J(t, y) =inf

E
_

uT
N
,ut
_

2
2
Y
2
u
Y
u

u
+

2

2
u
_
t +

2
T
2
Y
2
T

Y
t
=y
_
.
5.2. Problem Solution
Again, we make a quadratic Ansatz that for T t T
N
,
J(T t, y) =A
2
(T t +t)
y
2
2
+A
0
(T t +t) . (32)
Theorem 5. The discrete-time optimal trading intensity
t
=(t, Y
t
) is given by
(t, y) =
KA
2
(T t)
+K
2
A
2
(T t)t
y (33)
and the continuation value J is given by (32) where
A
2
(T t +t) =A
2
(T t) +
2
t +
_
KA
2
(T t)
_
2
+K
2
A
2
(T t)t
t A
2
(0) =
2
T
A
0
(T t +t) =A
0
(T t) +
1
2

2
A
2
(T t)t A
0
(0) =0.
Proof The discrete-time HJB equation becomes
J(T t, y) =inf

t
_

2
2
y
2
y
t
+

2

2
t
_
t +
A
2
(T t)
2
_
_
y +K
t
t
_
2
+()
2
t
_
+A
0
(T t) =0.
Separating both sides by powers of y and using (32) yields the desired result.
Author: Option Hedging with Market Impact
26 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
P
T

T

BA 53.27084 0.0433451 0.8270762 0.0515023
BAC 49.5763872 0.0243846 0.7866773 0.0780991
MSFT 64.9043047 0.0122172 1.4496147 0.1173921
PFE 43.5416817 0.0164636 0.6695676 0.1008441
WMT 52.5306878 0.0208141 0.9813519 0.0773531
Table 2 Summary Statistics of Stocks: the average closing stock price PT , absolute volatility in dollars per share
per square-root seconds , the average overnight volatility in dollars per share T , and the average of a
call-option that matures in 5 days and was At-The-Money at the previous close.
6. Simulation Using TAQ Data
To test the hedging strategy, we use stock TAQ Data to simulate delta hedging a call option
using our trading strategy and compare that to the benchmark Black-Scholes strategy. We choose
ve companies: Boeing (NYSE:BA), Bank of America (NYSE:BAC), Microsoft (NASDAQ:MSFT),
Pzer (NYSE:PFE), and Wal-Mart (NYSE:WMT), each representing a dierent sector of the
economy. Daily NBBO data was collected for the 251 trading days from April 3rd, 2000 to March
30th, 2001 (inclusive) and the mid-price prevailing at each minute of each trading day (9:30AM to
4:00PM) was struck (computed). Summary statistics of the data are given in Figure 2.
We use the price data to calibrate our the discrete-time model (see Section 5) under minutely
rehedging. We then use the data to simulate delta hedging a derivatives position with a at
gamma. We chose to corresponded to that of one million call options that were At-The-Money
at the previous trading days close and mature 5 days after that days close. These quantities are
computed using the Black-Scholes formula.
We can think of the at gamma derivative as approximating that of an At-The-Money call option
under Approximation (14). A call-options gamma prole becomes very narrow in moneyness near
the options expiry so that small uctuations in the stock price will greatly aect the gamma, thus
violating the assumptions of the model. This feature is not unique to our setup but stems from
the kink in the hockey-stick shape payo of the call and the associated diculty of hedging
At-The-Money call options near expiry is a widely recognized problem among practitioners. We
sidestep this issue in our simulation by choosing a call option that expires in 5 days.
We set the initial position X
0
so that the option is initially delta hedged at the previous close.
This (optimistically) simulates the position of a trader who hedges an option across multiple days.
His initial position in the morning may not be hedged due to the overnight uctuation.
We follow the logic of Example 4 in choosing our parameter values. We choose 10
6
. We
assume no permanent impact = 0 and a temporary impact of the form =
p
where
p
> 0
is a proportionality constant and is the absolute volatility of the stock. This accounts for the
well-known stylized fact that, caeteris paribus, market impact is higher for stocks with greater
volatility (Ane and Geman 2000, Cont 2001, Jain and Joh 1988). We choose
p
= 10
2
so that
10
4
as in Example 4. The parameter was chosen based on empirical values while
T
was
chosen to be 10 times its empirical value to induce the agent to hedge aggressively towards the
close. Given our choice of parameter values, the characteristic time-scale at which trading intensity
peaks near the market close is
1

which is on the order of 10 to 20 minutes. In other words, for our liquid stocks, the agent trades
with a constant proportional intensity until the last hours of the trading day.
We track the performance of two dierent agents, one using the proposed hedging intensity of
Section 5 (Model) and another whose strategy is to maintain a Black-Scholes delta hedged (BS).
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 27
The Black-Scholes trader will not be perfectly hedged due to uctuations in the subsequent minute
interval before rehedging. However, (BS) will likely maintain a tighter hedge on average as he
would be perfectly hedged if the stock did not uctuate in the subsequent interval whereas the
other trader (Model) is not even hedged at the start of the interval to begin with.
The dollar terminal exposure, dollar running exposure, and dollar impact costs are given in
Figure 3. The formulas for these quantities are given by
terminal exposure =[
T
Y
T
[
intraday exposure =

tT

2
Y
2
t
t (34)
impact cost =

tT

2
t
t .
Each term has dimension of dollars to make them directly comparable. Hedging according to the
model in Section 5, (Model) saves $10K a day in market impact costs, while taking on a comparable
amount of risk.
7. Conclusion
Market-impact from hedging a non-trivial outstanding position can greatly increase costs. This
paper develops a highly-tractable framework for analyzing optimal hedging of options for traders
with large positions who face market impact. We use a linear-quadratic framework and nd that
the optimal solution for options hedging is for the agent to trade towards a target portfolio.
The trader nds it prohibitively expensive to hold exactly the Black-Scholes hedge portfolio due
to the market-impact costs and can only trade towards it to minimize the turnover cost. We also
show that our results can be interpreted as a kind of Merton problem where the Merton-optimal
portfolio is the Black-Scholes hedge portfolio.
We show that in the limit of vanishing market-impact costs, we recover that the agent engages
in Black-Scholes hedging. We apply our result to stock-pinning and our predictions are consistent
with the U-shaped intraday trading pattern. Finally, we present the results of numerical simula-
tions which demonstrate the practicality of this technique even for real-world intraday stock-price
uctuations.
Author: Option Hedging with Market Impact
28 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
Black-Scholes
Terminal Intraday Impact
BA 9.965e+02 9.661e+03 1.235e+04
BAC 1.283e+03 9.300e+03 1.865e+04
MSFT 5.758e+03 1.387e+04 6.022e+04
PFE 1.736e+03 7.984e+03 1.583e+04
WMT 2.142e+03 1.114e+04 2.564e+04
Model, =0.5 10
6
Terminal Intraday Impact
BA 1.086e+03 2.751e+04 5.036e+02
BAC 1.392e+03 3.095e+04 6.780e+02
MSFT 5.858e+03 4.969e+04 2.019e+03
PFE 1.838e+03 2.572e+04 5.163e+02
WMT 2.242e+03 3.641e+04 9.844e+02
Model =1.0 10
6
Terminal Intraday Impact
BA 1.043e+03 2.359e+04 6.779e+02
BAC 1.339e+03 2.641e+04 8.921e+02
MSFT 5.808e+03 4.267e+04 2.585e+03
PFE 1.789e+03 2.205e+04 6.686e+02
WMT 2.193e+03 3.116e+04 1.281e+03
Model, =2.0 10
6
Terminal Intraday Impact
BA 1.020e+03 2.034e+04 9.253e+02
BAC 1.311e+03 2.258e+04 1.192e+03
MSFT 5.783e+03 3.667e+04 3.401e+03
PFE 1.763e+03 1.894e+04 8.820e+02
WMT 2.167e+03 2.672e+04 1.700e+03
Table 3 Summary Statistics of Simulation
Summary Statistics of Simulation: Average dollar exposure versus dollar market impact by stock
for a simulated Black-Scholes trader and one that follows our Model. See (34) for details of the
terms.
References
A. Alfonsi and A. Schied. Optimal trade execution and absence of price manipulations in limit order book
models. Working Papers, Apr 2009.
A. Alfonsi, A. Fruth, and A. Schied. Optimal execution strategies in limit order books with general shape
functions. Quantitative Finance Papers 0708.1756, arXiv.org, Aug. 2007. URL http://ideas.repec.
org/p/arx/papers/0708.1756.html.
R. Almgren and N. Chriss. Value under liquidation. Risk, Jan 1999.
R. Almgren and N. Chriss. Optimal execution of portfolio transactions. Journal of Risk, Apr 2001.
T. Ane and H. Geman. Order ow, transaction clock, and normality of asset returns. Jour-
nal of Finance, 55(5):22592284, October 2000. URL http://ideas.repec.org/a/bla/jfinan/
v55y2000i5p2259-2284.html.
Author: Option Hedging with Market Impact
Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!) 29
U. M. Ascher and L. R. Petzold. Computer Methods for Ordinary Dierential Equations and Dierential-
Algebraic Equations. SIAM, Philadelphia, PA, 1998.
M. Avellaneda and M. D. Lipkin. A market-induced mechanism for stock pinning. Quantitative Finance, 3:
417425, Sep 2003.
K. Back. Insider trading in continuous time. Review of Financial Studies, Jan 1992.
P. Bank and D. Baum. Hedging and portfolio optimization in nancial markets with a large trader. Mathemat-
ical Finance, 14(1):118, 2004. URL http://ideas.repec.org/a/bla/mathfi/v14y2004i1p1-18.
html.
D. Bertsimas and A. W. Lo. Optimal control of execution costs. Journal of Financial Markets, 1(1):150,
1998. URL http://EconPapers.repec.org/RePEc:eee:finmar:v:1:y:1998:i:1:p:1-50.
F. Black and M. Scholes. The pricing of options and corporate liabilities. The Journal of Political Economy,
Jan 1973.
U. Cetin, R. Jarrow, and P. Protter. Liquidity risk and arbitrage pricing theory. Finance and Stochastics, 8
(3):311341, 08 2004. URL http://ideas.repec.org/a/spr/finsto/v8y2004i3p311-341.html.
U. Cetin, R. Jarrow, P. Protter, and M. Warachka. Pricing options in an extended black scholes economy
with illiquidity: Theory and empirical evidence. Review of Financial Studies, 19(2):493529, 2006. URL
http://ideas.repec.org/a/oup/rfinst/v19y2006i2p493-529.html.
U. Cetin, H. M. Soner, and N. Touzi. Option hedging for small investors under liquidity costs. Finance and
Stochastics, 14:317341, Jan 2010.
L. Clewlow and S. Hodges. Optimal delta-hedging under transactions costs. Journal of Economic Dynam-
ics and Control, 21(8-9):13531376, June 1997. URL http://ideas.repec.org/a/eee/dyncon/
v21y1997i8-9p1353-1376.html.
R. Cont. Empirical properties of asset returns: stylized facts and statistical issues. Quantitative Finance, 1:
223236, 2001.
J. Cvitanic. Theory of portfolio optimization in markets with frictions, 1999. URL http://citeseerx.ist.
psu.edu/viewdoc/summary?doi=10.1.1.40.9078.
J. Cvitanic and I. Karatzas. Hedging and portfolio optimization under transaction costs: A martingale
approach12. Mathematical Finance, 6(2):133165, 1996. ISSN 1467-9965. doi: 10.1111/j.1467-9965.
1996.tb00075.x. URL http://dx.doi.org/10.1111/j.1467-9965.1996.tb00075.x.
J. Cvitanic and H. Wang. On optimal terminal wealth under transaction costs. Journal of Mathematical
Economics, 35:223231, Jan 2001.
M. H. A. Davis and A. R. Norman. Portfolio selection with transaction costs. Mathematics of Operations
Research, 15(4):676713, Nov 1990.
M. H. A. Davis, V. G. Panas, and T. Zariphopoulou. European option pricing with transaction costs. SIAM
J. Control Optim., 31(2):470493, 1993. ISSN 0363-0129. doi: http://dx.doi.org/10.1137/0331022.
F. Fabozzi. Robust portfolio optimization and management. Frank J. Fabozzi series. John Wiley, 2007. ISBN
9780471921226. URL http://books.google.com/books?id=9QnuJqeXSmQC.
N. Garleanu and L. Pedersen. Dynamic trading with predictable returns and transaction costs. NBER
Working Paper, Mar 2009.
G. Hernandez-Del-Valle and C. G. Pacheco-Gonzalez. Optimal execution of equity with geometric price
process, 2009.
R. Hodges and K. Neuberger. Optimal replication of contingent claims under transaction costs. Rev. Futures
Markets, 8:222239, 1989.
U. Horst and F. Naujokat. Illiquidity and derivative valuation. Quantitative nance papers, arXiv.org, 2008.
URL http://econpapers.repec.org/RePEc:arx:papers:0901.0091.
P. C. Jain and G.-H. Joh. The dependence between hourly prices and trading volume. The Journal of
Financial and Quantitative Analysis, 23(3):pp. 269283, 1988. ISSN 00221090. URL http://www.
jstor.org/stable/2331067.
Author: Option Hedging with Market Impact
30 Article submitted to Operations Research; manuscript no. (Please, provide the mansucript number!)
K. Janecek and S. E. Shreve. Asymptotic analysis for optimal investment and consumption with transaction
costs. Finance and Stochastics, 8:181206, Jan 2004.
P. Kratz and T. Schoneborn. Optimal liquidation in dark pools. papers.ssrn.com, Mar 2010.
A. Kyle. Continuous auctions and insider trading. Econometrica, 53(6):13151335, Jan 1985.
H. E. Leland. Option pricing and replication with transactions costs. Journal of Finance, pages 12831301,
1985.
R. Merton. Theory of rational option pricing. The Bell Journal of Economics and Management Science, 4
(1):141183, Jan 1973.
A. Obizhaeva and J. Wang. Optimal trading strategy and supply/demand dynamics. Working Paper 11444,
National Bureau of Economic Research, June 2005. URL http://www.nber.org/papers/w11444.
S. Predoiu, G. Shaikhet, and S. Shreve. Optimal execution in a general one-sided limit-order book.
math.cmu.edu, Feb 2010.
L. C. G. Rogers and S. Singh. The cost of illiquidity and its eects on hedging. Preprint, Aug 2007.
A. Schied and T. Schoneborn. Optimal portfolio liquidation for cara investors. Mpra paper, University
Library of Munich, Germany, 2007. URL http://econpapers.repec.org/RePEc:pra:mprapa:5075.
T. Schoeneborn and A. Schied. Liquidation in the face of adversity: stealth vs. sunshine trading.
papers.ssrn.com, Mar 2010.
S. Shreve and H. Soner. Optimal investment and consumption with transaction costs. The Annals of Applied
Probability, 4(3):609692, Aug 1994.
H. M. Soner and G. Barles. Option pricing with transaction costs and a nonlinear black-scholes equa-
tion. Finance and Stochastics, 2(4):369397, 1998. URL http://ideas.repec.org/a/spr/finsto/
v2y1998i4p369-397.html.
H. M. Soner, S. E. Shreve, and J. Cvitanic. There is no nontrivial hedging portfolio for option pricing with
transaction costs. The Annals of Applied Probability, 5(2):327355, Jan 1995.
A. E. Whalley and P. Wilmott. An asymptotic analysis of an optimal hedging model for option pricing with
transaction costs. Mathematical Finance, 7(3):307324, 1997. URL http://ideas.repec.org/a/bla/
mathfi/v7y1997i3p307-324.html.