Chapter 10

Asian Options

Asian options are particular cases of options on average, and they were first
traded in Tokyo in 1987. Given an underlying asset St with exercise date T
and strike price K, the payoff of the Asisan call option is given by

1 wT
 +
C := St dt − K
T 0

whereas the payoff of the Asian put option is

1 wT
 +
C := K− St dt .
T 0

This chapter covers several probabilistic and PDE techniques for the pricing
and hedging of Asian options. Due to their dependence on averaged asset
prices, Asian options are less volatile than plain vanilla options whose payoffs
depend only on the terminal value of the underlying asset.

Contents
10.1 Options on Averages . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
10.2 The Asian Call Option . . . . . . . . . . . . . . . . . . . . . . . . . 351
10.3 Moment Matching Approximations . . . . . . . . . . . . . . 357
10.4 PDE Method - Two Variables . . . . . . . . . . . . . . . . . . . 362
10.5 PDE Method - One Variable . . . . . . . . . . . . . . . . . . . . 365
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

10.1 Options on Averages

In this case we can take

1 wT
 
C=φ St dt
T 0

347
 N. Privault

where
1 wT
St dt
T 0
represents the average of (St )t∈R+ over the time interval [0, T ] and φ : R −→ R
is a payoff function.

Fig. 10.1: Brownian motion Bt and its moving average Xt .∗

Figure 10.1 shows a graph of Brownian motion and its moving average process
1wt
Xt := Bs ds, t > 0.
t 0

Arithmetic Asian options

The payoff of the Asian call option on the underlying asset St with exercise
date T and strike price K is given by

1 wT
 +
C= St dt − K .
T 0

Similarly, the payoff of the Asian put option on the underlying asset St with
exercise date T and strike price K is

1 wT
 +
C= K− St dt .
T 0

Due to their dependence on averaged asset prices, Asian options are less
volatile than plain vanilla options whose payoffs depend only on the terminal
∗
The animation works in Acrobat Reader on the entire pdf file.

348 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

value of the underlying asset. Asian options have become particularly popu-
lar in commodities trading.

Other types of exotic options include called Asian-American options, or

Hawaiian options, that combine the Asian option payoff with American style
exercise, and can be priced by variational PDEs, cf. § 8.6.3.2 of [Cré13].

An option on average is an option whose payoff has the form

C = φ(ΛT , ST ),

where
wT 2
wT
ΛT = S0 e σBu +ru−σ u/2
du = Su du, T ∈ R+ .
0 0

• For example when φ(y, x) = (y/T − K)+ this yields the Asian call option
with payoff
 w +  +
1 T ΛT
Su du − K = −K , (10.1)
T 0 T
which is a path-dependent option whose price at time t ∈ [0, T ] is given
by
1 wT
" +  #
e −(T −t)r IE∗ Su du − K  Ft . (10.2)

T 0 

• As another example, when φ(y, x) := e −y this yields the price

h rT i
P (0, T ) = IE∗ e − 0 Su du = IE∗ e −ΛT
 

at time 0 of a bond with underlying short term rate process St .

The option with payoff C = φ(ΛT , ST ) can be priced as

wT   #
" 
∗ ∗
e −(T −t)r
IE [φ(ΛT , ST ) | Ft ] = e −(T −t)r
IE φ Λt + Su du, ST  Ft

t 
  wT S ST

u
= e −(T −t)r IE∗ φ y + x du, x
t St St y=Λt ,x=St
  w T −t S ST −t

u
= e −(T −t)r IE∗ φ y + x du, x . (10.3)
0 S0 S0 y=Λt ,x=St

Using the Markov property of the process (St , Λt )t∈R+ , we can write down
the option price as a function

f (t, St , Λt ) = e −(T −t)r IE∗ [φ(ΛT , ST ) | Ft ] = e −(T −t)r IE∗ [φ(ΛT , ST ) | St , Λt ],

" 349
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

of (t, St , Λt ), where the function f (t, x, y) is given by

  w T −t S ST −t

u
f (t, x, y) = e −(T −t)r IE∗ φ y + x du, x .
0 S0 S0

As we will see below there exists no easily tractable closed-form solution for
the price of an arithmetically averaged Asian option.

Geometric Asian options

On the other hand, replacing the arithmetic average

1 wT
n
1 X
Stk (tk − tk−1 ) ' Su du
T T 0
k=1

with the geometric average

n n
!
(t −tk−1 )/T (t −tk−1 )/T
Y Y
Stkk = exp log Stkk
k=1 k=1
n
!
1 X t −t
= exp log Stkk k−1
T
k=1
n
!
1 X
= exp (tk − tk−1 ) log Stk
T
k=1
 w
1 T

' exp log Su du
T 0

leads to closed-form solutions using the Black Scholes formula, cf. Exer-
cise 10.3.

Pricing using probability density functions

We note that the prices of option on averages can be estimated numerically

using the joint probability density function ψΛT −t ,BT −t of (ΛT −t , BT −t ), as
follows:
  w T −t S ST −t

u
f (t, x, y) = e −(T −t)r IE∗ φ y + x du, x
0 S0 S0
w∞w∞  
σu+(T −t)r−(T −t)σ 2 /2
e −(T −t)r
φ y + xz, x e ψΛT −t ,BT −t (z, u)dzdu.
0 −∞

In [Yor92], Proposition 2, the joint probability density function of

350 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

w 
t 2
(Λt , Bt ) = e σBs −pσ s/2
ds, Bt − pσt/2 , t > 0,
0

has been computed in the case σ = 2, cf. also [MY05]. In the next proposition
we restate this result for an arbitrary variance parameter σ after rescaling.
Let θ(v, τ ) denote the function defined as

v e π /(2τ ) w ∞ −ξ2 /(2τ ) −v cosh ξ

2

θ(v, τ ) = √ e e sinh(ξ) sin (πξ/τ ) dξ, v, τ > 0.

2π 3 τ 0
(10.4)
Proposition 10.1. For all t > 0 we have
w 
t 2 σt
P e σBs −pσ s/2 ds ∈ dy, Bt − p ∈ dz
0 2
1 + e σz 4e
   σz/2 2 
σ −pσz/2−p2 σ2 t/8 σ t dy
= e exp −2 θ , dz,
2 2
σ y σ2 y 4 y

y > 0, z ∈ R.
The expression of this probability density function can then be used for the
pricing of options on average such as (10.3), as
  w T −t S ST −t

v
f (t, x, y) = e −(T −t)r IE∗ φ y + x dv, x
0 S0 S0
= e −(T −t)r
w∞  2
 w T −t S 
v
× φ y + xz, x e σu+(T −t)r−(T −t)σ /2 P dv ∈ dz, BT −t ∈ du
0 0 S0
σ −(T −t)r+(T −t)p2 σ2 /8 w ∞ w ∞  σu+(T −t)r−(T −t)σ 2 (1+p)/2

= e φ y + xz, x e
2 0 −∞
! !
2 2
1 + e σu−(T −t)pσ /2 p 4 e σu/2−(T −t)pσ /4 (T − t)σ 2 dz
× exp −2 − σu θ , du
σ2 z 2 σ2 z 4 z
w∞w∞  
−(T −t)r−(T −t)p2 σ 2 /8 2 (T −t)r−(T −t)σ 2 /2
= e φ y + x/z, xv e
0
0 
1 + 2
4vz (T − t)σ 2
 
v dz
×v −1−p exp −2z θ , dv ,
σ 2 σ 2 4 z

which actually stands as a triple integral due to the definition (10.4) of θ(v, τ ).
Note that here the order of integration between du and dz cannot be ex-
changed without particular precautions, at the risk of wrong computations.

10.2 The Asian Call Option

By repeating the argument of (10.3) for φ(x, y) := (x − K)+ , the Asian call
option can be priced as
" 351
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

1 wT
" +  #
e −(T −t)r IE∗ Su du − K  Ft

T 0 
wT +  #
" 
1

∗
= e −(T −t)r
IE Λt + Su du − K  Ft

T t 
w +  #
" 
1

T Su
= e −(T −t)r IE∗ y+x du − K  Ft

T t St 
x=St , y=Λt
w
"  + #
1

T −t Su
∗
= e −(T −t)r
IE y+x du − K .
T 0 S0
x=St , y=Λt

Hence the Asian option can be priced as

1 wT
" +  #

f (t, St , Λt ) = e −(T −t)r IE∗ Su du − K  Ft ,

T 0 

where the function f (t, x, y) is given by

w T −t S
"  + #
1

u
f (t, x, y) = e −(T −t)r IE∗ y+x du − K
T 0 S0
"  + #
1

x
= e −(T −t)r IE∗ y+ ΛT −t − K , x, y > 0. (10.5)
T S0

Bounds on Asian option prices

We note (see Lemma 1 of [KV90] and Exercise 10.5 below for the discrete-time
version of that result), that the Asian call option price can be upper bounded
by the corresponding European call price using convexity arguments.
Proposition 10.2. Assume that r > 0. We have the bound

1 wT
" + #
e −rT IE∗ 6 e −rT IE∗ (ST − K)+ .
 
Su du − K
T 0

Proof. By Jensen’s inequality for the uniform measure with density (1/T )1[0,T ]
on [0, T ] and for the probability measure P∗ , we have

wT wT
" + # " + #
du du
e −rT IE∗ Su−K = e −rT IE∗ (Su − K)
0 T 0 T
w 
T du
6 e −rT IE∗ (Su − K)+
0 T

352 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

w 
T du
= e −rT IE∗ ( e −(T −u)r IE∗ [ST | Fu ] − K)+
0 T
w i du 
T h
+ 
∗ ∗
6 e −rT IE IE e −(T −u)r
ST − K  Fu (10.6)
0 T
wT h h  ii du
∗ ∗
6 e −rT IE IE (ST − K)  Fu + 
(10.7)
0 T
wT  du
= e −rT IE∗ (ST − K)+

0 T
= e −rT IE∗ (ST − K)+ ,
 

where from (10.6) to (10.7) we used the fact that r > 0. 

More generally, given that

1 wT
" +  #

∗
f (t, St , Λt ) = e −(T −t)r
IE Su du − K  Ft ,

T 0 

where, from Proposition 10.2,

w T −t S
" + #
1
 
u
f (t, x, y) = e −(T −t)r IE∗ y+xdu − K
T 0 S0
"  + #
1

x
= e −(T −t)r IE∗ y+ ΛT −t − K
T S0
" + #
∗ y x
= e −(T −t)r
IE −K + ΛT −t
T T S0
" + #
(T − t)x −(T −t)r ∗ yS0 KT S0 ΛT −t
= e IE − +
T S0 (T − t)x (T − t)x T − t
" + #
(T − t)x −(T −t)r ∗ yS0 KT S0
6 e IE − + ST −t
T S0 (T − t)x (T − t)x
" + #
y (T − t)xST −t
= e −(T −t)r IE∗ −K + , x, y > 0,
T T S0

we find the bound

1 wT
" +  #

∗
f (t, St , Λt ) = e −(T −t)r
IE Su du − K  Ft

T 0 

1 wt
" +  #
T −t
6 e −(T −t)r IE∗ Su du + ST − K  Ft

T 0 T

" 353
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

at time t ∈ [0, T ]. See also Proposition 3.2-(ii) of [GY93] for lower bounds
when r takes negative values. We also have the following bound which yields
the behavior of Asian option prices in large time.
Proposition 10.3. The Asian option price satisfies the bound

1 wT e −(T −t)r w t
" +  #
∗ 1 − e −(T −t)r
e −(T −t)r
IE Su du − K  Ft 6 Su du+St ,

T 0  T 0 rT

t ∈ [0, T ], and tends to zero (almost surely) as time to maturity T tends to

infinity:

1 wT
" +  #!
lim e −(T −t)r IE∗ Su du − K  Ft = 0, t ∈ R+ .

T →∞ T 0 
Proof. We have the bound

1 wT
" +  #
0 6 e −(T −t)r IE∗ Su du − K  Ft

T 0 
w
"  #
1 T 
6 e −(T −t)r IE∗ Su du  Ft

T 0 
w 1 wT
"  # "  #
1 t  
= e −(T −t)r IE∗ Su du  Ft + e −(T −t)r IE∗ Su du  Ft
 
T 0  T t 
1 w t 1 w T
= e −(T −t)r IE∗ Su | Ft du + e −(T −t)r IE∗ Su | Ft du
   
T 0 T t
w
1 t 1 w T
= e −(T −t)r Su du + e −(T −t)r e (u−t)r St du
T 0 T t
1 −(T −t)r t w w
St T −(T −u)r
= e Su du + e du
T 0 T t
1 −(T −t)r t w 1− e −(T −t)r
= e Su du + St .
T 0 rT


Pricing by the Hartman-Watson distribution

First we note that the numerical computation of Asian option prices can be
done using the probability density function of
wT
ΛT = St dt.
0

354 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

From Proposition 10.1, we deduce the marginal probability density function

of wT 2
ΛT := e σBt −pσ t/2 dt,
0

also called the Hartman-Watson distribution [BRY04], as follows:

w 
T 2
P e σBt −pσ t/2 dt ∈ du
0

σ p2 σ2 T /8 w ∞
2
! 2
!
1 + e σv−pσ T /2 p 4 e σv/2−pσ T /4 σ 2 T
= e exp −2 − σv θ , dvdu
2u −∞ σ2 u 2 σ2 u 4
2 2
w∞ 
1 + v2
 
4v σ 2 T

du
= e −p σ T /8 v −1−p exp −2 2 θ , dv ,
0 σ u σ u 4
2 u

u > 0. From this expression we get

w 
T
P(ΛT /S0 ∈ du) = P St dt ∈ du (10.8)
0

2 2
w ∞

1+v 2
 
4v σ T
2

du
= e −p σ T /8 v −1−p exp −2 2 θ , dv ,
0 σ u σ2 u 4 u
2
where St = S0 e σBt −pσ t/2 and p = 1 − 2r/σ 2 . By (10.5), this probability
density function can then be used for the pricing of Asian options, as
"  + #
1

∗ x
f (t, x, y) = e −(T −t)r
IE y+ ΛT −t − K (10.9)
T S0
w ∞ y + xz
  +
= e −(T −t)r −K P(ΛT −t /S0 ∈ dz)
0 T
σ 2 2
w ∞ w ∞  y + xz +
= e −(T −t)r e −(T −t)p σ /8 −K
2 0 0 T
1 + 2
4v (T t)σ 2
   
v − dz
×v −1−p exp −2 2 θ , dv
σ z 2
σ z 4 z
1 2 2
w∞ w∞
= e −(T −t)r−(T −t)p σ /8 (xz + y − KT )
T 0∨(KT −y)/x 0
1 + v2 4v (T − t)σ 2
   
dz
× exp −2 2 θ , dv
σ z 2
σ z 4 z
4x −(T −t)r−(T −t)p2 σ2 /8 w ∞ w ∞ 1 σ 2 (KT − y)
 +
= 2 e −
σ T 0 0 z 4x
1 + v2 (T − t)σ 2
   
dz
×v −1−p
exp −z θ vz, dv ,
2 4 z

" 355
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

cf. Theorem in § 5 of [CS04], which is actually a triple integral due to the

definition (10.4) of θ(v, t). Note that since the integrals are not absolutely
convergent, here the order of integration between dv and dz cannot be ex-
changed without particular precautions, at the risk of wrong computations.

Time Laplace transform

The time Laplace transform of the rescaled price

1wt
" + #
C(t) := IE∗ Su du − κ , t ∈ R+ ,
t 0

as
r K/2 √ √
w∞ e −x x−2+(p+ 2λ+p )/2 (1 − 2Kx)2+( 2λ+p −p)/2 dx
2 2

e −λt
C(t)dt = 0
,
λ(λ − 2 + 2p)Γ − 1 + p + 2λ + p2 /2
p


0

with here σ := 2, and Γ (z) denotes the gamma function, see Relation (3.10)
in [GY93]. This expression can be used for pricing by numerical inversion
of the Laplace transform using e.g. the Widder method, the Gaver-Stehfest
method, the Durbin-Crump method, or the Papoulis method. The follow-
ing Figure 10.2 represents Asian option prices computed by the Geman-Yor
[GY93] method.

30
25
20
15
10
5
0
95

Underlying 90
85
300 350
150 200 250
80 50 100
0 Time in months

Fig. 10.2: Graph of Asian option prices with σ = 1, r = 0.1 and K = 90.

We refer to e.g. [CS04], [Duf00], and references therein for more results on
Asian option pricing using the probability density function of the averaged
geometric Brownian motion.

Figure 7.6 presents a graph of implied volatility surface for Asian options on
light sweet crude oil futures.

356 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

10.3 Moment Matching Approximations

Lognormal approximation

Other numerical approaches to the pricing of Asian options include [Lev92],

[TW92] which rely on approximations of the average price distribution based
on the Lognormal distribution. The lognormal distribution has the probabil-
ity density function
1 2 2 dx
g(x) = √ e −(µ−log x) /(2η ) , x > 0,
η 2π x

where µ ∈ R, η > 0, with moments

2 2
IE[X] = e µ+η /2
and IE[X 2 ] = e 2µ+2η . (10.10)

The approximation is implemented by matching the above first two moments

to those of time integral
wT
ΛT := St dt
0
of geometric Brownian motion
2
St = S0 e σBt +(r−σ /2)t
, t ∈ [0, T ],

as computed in the next proposition, cf. also (7) and (8) page 480 of [Lev92].
Proposition 10.4. We have

e rT − 1
IE∗ [ΛT ] = S0 ,
r
and 2
r e (σ +2r)T
− (σ 2 + 2r) e rT + (σ 2 + r)
IE∗ (ΛT )2 = 2S02
 
.
(σ 2 + r)(σ 2 + 2r)r
Proof. The computation of the first moment is straightforward: we have
w  w wT
T T e rT − 1
IE∗ ΛT = IE∗ Su du = IE∗ Su du = S0 e ru du = S0
   
.
0 0 0 r

For the second moment we have, letting p := 1 − 2r/σ 2 ,

2  wT wT 2 2
IE∗ = S02 e −pσ a/2−pσ b/2 IE∗ e σBa e σBb dbda
  
ΛT
0 0
wT wa 2 2 2 2
= 2S02 e −pσ a/2−pσ b/2 e σ (a+b)/2 e bσ dbda
0 0
wT 2
wa 2
= 2S02 e −(p−1)σ a/2 e −(p−3)σ b/2 dbda
0 0

" 357
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

4S02 wT 2 2
= e −(p−1)σ a/2 (1 − e −(p−3)σ a/2 )da
(p − 3)σ 0
2

4S02 wT 2 4S02 wT 2 2
= e −(p−1)σ a/2 da − e −(p−1)σ a/2 e −(p−3)σ a/2 da
(p − 3)σ 2 0 (p − 3)σ 2 0
8S02 2 4S02 wT 2
= (1 − e −(p−1)σ T /2 ) − e −(2p−4)σ a/2 da
(p − 3)(p − 1)σ 4 (p − 3)σ 02

8S02 2 4S02 2
= (1 − e −(p−1)σ T /2 ) − (1 − e −(p−2)σ T )
(p − 3)(p − 1)σ 4 (p − 3)(p − 2)σ 4
2
r e (σ +2r)T
− (σ 2 + 2r) e rT + (σ 2 + r)
= 2S02 ,
(σ 2 + r)(σ 2 + 2r)r

since r − σ 2 /2 = −pσ 2 /2. 

By matching (10.4) with the moments of Proposition 10.4 we estimate µ̂, η̂
as
1 E[Λ2T ] 1 1
 
η̂ 2 = log ∗ and µ̂ = log IE∗ [ΛT ] − η̂ 2 .
T (IE [ΛT ])2 T 2

2.5
exact expression
lognormal approximation

2
probability density

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3
x

Fig. 10.3: Lognormal approximation for the probability density function of ΛT .

As a consequence, from Lemma 6.5 we find the approximation

1 wT
" + #
∗ 1 2
e −rT
IE St dt − K ' e (µ̂+η̂ /2)T Φ(d1 ) − KΦ(d2 ), (10.11)
T 0 T

where
√
log(IE∗ [ΛT ]/(KT )) T µ̂T + η̂ 2 T − log(KT )
d1 = √ + η̂ = √
η̂ T 2 η̂ T
and √
√ log(IE∗ [ΛT ]/(KT )) T
d2 = d1 − η̂ T = √ − η̂ .
η̂ T 2

358 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

The next Figure 10.4 compares the lognormal approximation to a Monte

Carlo estimate of Asian option prices with σ = 0.5, r = 0.05 and K/St = 1.1

1
lognormal approximation
Monte Carlo estimate
stratified lognormal approximation
0.9

0.8

0.7
asian option price

0.6

0.5

0.4

0.3

0.2

0.1
0 2 4 6 8 10 12
time t

Fig. 10.4: Lognormal approximation to the Asian option price.

Figure 10.4 also includes the stratified approximation

1 wT
" + #
e −rT IE St dt − K (10.12)
T 0
e −rT w ∞  −p(z/x)σ2 (z/x)T /2+σ2 (z/x)T /2 
' e Φ(d+ (K, z, x)) − KT Φ(d− (K, z, x))
T 0
×dP(ST 6 z | S0 = x),

of [PY16], where
√
1 2x(bT (z/x) − (1 + z/x)aT (z/x))
 
σ(z/x) T
d± (K, z, x) := √ log ±
2σ(z/x) T σ2 K 2 T 2 2

and
1 2 bT (z)
   
 σ 2 (z) := log −1−z ,




 T σ aT (z) aT (z)
2



1 log z 1√ 2 log z 1√ 2

     
aT (z) := 2 Φ √ + σ T −Φ √ − σ T ,

 σ p(z) σ2 T 2 σ2 T 2



1 log z √ log z √

     
 bT (z) := 2 + σ2 T − Φ √

Φ √ − σ2 T ,


σ q(z) σ2 T σ2 T
and
1 1
e −(σ T /2+log z ) /(2σ 2 T )
e −(σ T +log z ) /(2σ 2 T )
2 2 2 2
p(z) := √ , q(z) := √ .
2πσ 2 T 2πσ 2 T

" 359
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

Pricing Asian options by conditioning on the geometric mean

price

Asian options on the arithmetic average

1 wT
St dt
T 0
have been priced by conditioning on the geometric mean price
 w   w
1 T 1 T 1 wT
 
G := exp log St dt 6 exp log St dt = St dt
T 0 T 0 T 0

in [Cur94], as

1 wT
" + #
e −rT IE∗ Su du − K
T 0
w∞ 1 wT
" +  #
= e −rT IE∗ Su du − K  G = x dP(G 6 x)

0 T 0
wK 1 wT
" +  #
= e −rT IE∗ Su du − K G = x dP(G 6 x)

T 0

0

w∞ 1 wT
" +  #
+ e −rT IE∗ Su du − K  G = x dP(G 6 x)

K T 0
= C1 + C2 ,

where
wK 1 wT
" +  #
C1 := e −rT IE∗ Su du − K G = x dP(G 6 x),

T 0

0

and
w∞ 1 wT
" +  #
C2 := e −rT IE∗ Su du − K  G = x dP(G 6 x)

K T 0
w∞  w
1 T  
= e −rT IE∗ Su du − K  G = x dP(G 6 x)

K T 0
e −rT w ∞ ∗ w T w∞
  
= IE Su du  G = x dP(G 6 x) − K e −rT dP(G 6 x)

T K 0 K

e −rT ∗ w T
 
= IE Su du1{G>K} − K e −rT P(G > K).
T 0

360 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

The term C1 can be estimated by a lognormal approximation given that

G = x. As for C2 we note that
 w
1 T

G = exp log St dt
T 0
 w 
1 T σ2 t
 
= exp µt + σBt − dt
T 0 2
 
T σ 2
 w
σ T

= exp µ− + Bt dt ,
2 2 T 0

hence
T σ wT
log G = (µ − σ 2 /2) + Bt dt
2 T 0
is Gaussian N T (µ − σ 2 /2)/2, σ 2 T /3 with mean T (µ − σ 2 /2)/2, and vari-

ance
wT w w
" 2 # 
T T
IE Bt dt = IE Bs Bt dsdt
0 0 0
wT wT
= IE [Bs Bt ] dsdt
00
wT wt
=2 sdsdt
0 0
wT
= t2 dt
0
T3
= .
3
Hence we have

P(G > K) = P(log G > log K)

σ2 σ wT
   
T
=P µ− + Bt dt > log K
2 2 T 0
w
σ2
   
T T T
=P Bt dt > − µ− + log K
0 σ 2 2
 √  
3 2
 
T σ
=Φ √ µ− − log K .
σ T 2 2

Basket options

Basket options on the portfolio

N
(k)
X
AT := αk ST
k=1

" 361
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

have also been priced in [Mil98] by approximating AT by a lognormal or a re-

ciprocal gamma random variable, see also [DLV04] for additional conditioning
on the geometric average of asset prices.

Asian basket options

Moment matching techniques combined with conditioning have been applied

to Asian basket options in [DDV10]. See also [DB02] for the pricing of Asian
basket options using quasi Monte Carlo simulation.

10.4 PDE Method - Two Variables

The price at time t of the Asian option with payoff (10.1) can be written as

1 wT
" +  #
f (t, St , Λt ) = e −(T −t)r IE∗ Su du − K  Ft , t ∈ [0, T ].

T 0 
(10.13)
Next, we derive the Black-Scholes partial differential equation (PDE) for the
price of a self-financing portfolio. Until the end of this chapter we model the
asset price (St )t∈[0,T ] as

dSt = µSt dt + σSt dBt , t ∈ R+ ,

where (Bt )t∈R+ is a standard Brownian motion under the historical proba-
bility measure P.
Proposition 10.5. Let (ηt , ξt )t∈R+ be a portfolio strategy such that
(i) (ηt , ξt )t∈R+ is self-financing,

(ii) the value Vt := ηt At + ξt St , t ∈ R+ , takes the form

Vt = f (t, St , Λt ), t ∈ R+ ,

for some function f ∈ C 1,2,1 ((0, ∞)3 ).

Then the function f (t, x, y) in (10.13) satisfies the PDE

362 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

∂f ∂f ∂f 1 ∂2f
rf (t, x, y) = (t, x, y) + x (t, x, y) + rx (t, x, y) + x2 σ 2 2 (t, x, y),
∂t ∂y ∂x 2 ∂x
t, x > 0, under the boundary conditions
 +
f (t, 0+ , y) = lim f (t, x, y) = e −(T −t)r y − K ,

(10.14a)






x&0 T
f (t, x, 0+ ) = lim f (t, x, y) = 0, (10.14b)
 y&0

 y +
f (T, x, y) = (10.14c)

−K ,


T
0 6 t 6 T , x, y ∈ R+ ,

and ξt is given by
∂f
ξt = (t, St , Λt ), t ∈ R+ . (10.15)
∂x
Proof. We note that the self-financing condition (5.8) implies

dVt = ηt dAt + ξt dSt

= rηt At dt + µξt St dt + σξt St dBt (10.16)
= rVt dt + (µ − r)ξt St dt + σξt St dBt , t ∈ R+ .

Since dΛt = St dt, an application of Itô’s formula to f (t, x, y) leads to

∂f ∂f
dVt = f (t, St , Λt ) = (t, St , Λt )dt + (t, St , Λt )dΛt
∂t ∂y
∂f 1 2
∂ f ∂f
+µSt (t, St , Λt )dt + St2 σ 2 2 (t, St , Λt )dt + σSt (t, St , Λt )dBt
∂x 2 ∂x ∂x
∂f ∂f
= (t, St , Λt )dt + St (t, St , Λt )dt
∂t ∂y
∂f 1 ∂2f ∂f
+µSt (t, St , Λt )dt + St2 σ 2 2 (t, St , Λt )dt + σSt (t, St , Λt )dBt .
∂x 2 ∂x ∂x
(10.17)

By respective identification of the terms in dBt and dt in (10.16) and (10.17)

we get

" 363
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

 ∂f ∂f ∂f
 rηt At dt + µξt St dt = (t, St , Λt )dt + St (t, St , Λt )dt + µSt (t, St , Λt )dt
∂t ∂y ∂x



1 ∂2f


+ St2 σ 2 2 (t, St , Λt )dt,


2

∂x






 ξ S σdB = S σ ∂f (t, S , Λ )dB ,


t t t t t t t
∂x
hence
1 2 2 ∂2f

∂f ∂f
 rVt − rξt St = ∂t (t, St , Λt ) + St ∂y (t, St , Λt )dt + 2 St σ ∂x2 (t, St , Λt ),




 ξt = ∂f (t, St , Λt ),



∂x
i.e.
 ∂f ∂f ∂f
 rf (t, St , Λt ) = (t, St , Λt ) + St (t, St , Λt ) + rSt (t, St , Λt )
∂t ∂y ∂x



1 ∂2f


+ St2 σ 2 2 (t, St , Λt ),


2

∂x






 ξ = ∂f (t, S , Λ ).


t t t
∂x

When Λt /T > K, by Exercise 10.6 we have

1 wT
" +  #
f (t, St , Λt ) = e −(T −t)r IE∗ Su du − K  Ft

T 0 
1− e −(T −t)r
 
Λt
= e −(T −t)r − K + St , (10.18)
T rT

t ∈ [0, T ], hence in this case the Delta ξt is given by

∂f 1 − e −(T −t)r
ξt = (t, St , Λt ) = , t ∈ R+ . (10.19)
∂x rT
In addition we have ξT = 0 at maturity from (10.14c) and (10.15).

Next, we examine two methods which allow one to reduce the Asian option
pricing PDE from three variables (t, x, y) to two variables (t, z). Reduction
of dimensionality can be of crucial importance when applying discretization
scheme whose complexity are of the form N d where N is the number of

364 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

discretization steps and d is the dimension of the problem (curse of dimen-

sionality).

10.5 PDE Method - One Variable

(1) Time-independent coefficients

Following [LL96], page 91, we define the auxiliary process

1 1 wt 1 Λt
   
Zt = Su du − K = −K , t ∈ [0, T ].
St T 0 St T

With this notation, the price of the Asian option at time t becomes

1 wT
" +  #
e −(T −t)r IE∗  Ft = e −(T −t)r IE∗ ST (ZT )+  Ft .
  
Su du − K

T 0 

Lemma 10.6. The price (10.2) at time t of the Asian option with payoff
(10.1) can be written as

1 wT
" +  #
f (t, St , Λt ) = St g(t, Zt ) = e −(T −t)r IE∗ Su du − K  Ft ,

T 0 
(10.20)
t ∈ [0, T ], with the relation

1 y
 
f (t, x, y) = xg t, −K , x, y ∈ R+ , 0 6 t 6 T,
x T

where

1 w T −t Su
" + #
g(t, z) = e −(T −t)r IE∗ z+ du (10.21)
T 0 S0
" + #
ΛT −t
= e −(T −t)r IE∗ z+ ,
S0 T

and with the boundary condition

g(T, z) = z + , z ∈ R.
Proof. For 0 6 s 6 t 6 T , we have
w
1

t St
d (St Zt ) = d Su du − K = dt,
T 0 T

hence

" 365
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

ST Zt 1 w t Su
= Zs + du, t > s.
Ss T s Ss
Since for any t ∈ [0, T ], St is positive and Ft -measurable, and Su /St is inde-
pendent of Ft , u > t, we have:
" +  #
ST
e −(T −t)r IE∗ ST (ZT )+  Ft = e −(T −t)r St IE∗
  
ZT  Ft

St 
w
" +   #
1 T Su
= e −(T −t)r St IE∗ Zt + du  Ft

T t St 
w
" #
+
1 T Su

= e −(T −t)r St IE∗ z+ du
T t St
z=Zt

1 w T −t Su
" + #
∗
= e −(T −t)r
St IE z+ du
T 0 S0
z=Zt
" + #
∗ ΛT −t
= e −(T −t)r
St IE z+
S0 T
z=Zt
= St g(t, Zt ),

which proves (10.21). 

When Λt /T > K we have Zt > 0, hence by (10.18) and (10.20) we find

1 − e −(T −t)r
g(t, Zt ) = e −(T −t)r Zt + , t ∈ [0, T ]. (10.22)
rT
Note that as in (10.9), g(t, z) can be computed from the probability density
function (10.8) of ΛT −t , as
" + #
∗ ΛT −t
g(t, z) = IE z+
S0 T
w∞ u +

Λt

= z+ dP 6u
0 T S0
2 2
= e −p σ t/8
w∞ u + w ∞ −1−p

1 + v2
 
4v (T − t)σ 2

du
× z+ v exp −2 θ , dv
0 T 0 σ2 σ2 u 4 u
2 2
= e −p σ t/8
w∞  u  w ∞ −1−p

1 + v2
 
4v (T − t)σ 2

du
× z+ v exp −2 θ , dv
(−zT )∨0 T 0 σ 2 2
σ u 4 u
w∞ w∞ 
1 + v2
 
4v (T − t)σ 2

du
−p2 σ 2 t/8
= ze v −1−p
exp −2 θ , dv
(−zT )∨0 0 σ2 σ2 u 4 u

366 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

1 −p2 σ2 t/8 w ∞ w∞ 1 + v2 4v (T − t)σ 2

   
+ e v −1−p exp −2 θ , dvdu.
T (−zT )∨0 0 σ 2 2
σ u 4

The next proposition gives a replicating hedging strategy for Asian options,
cf. [RS95].
Proposition 10.7. Let (ηt , ξt )t∈R+ be a portfolio strategy such that
(i) (ηt , ξt )t∈R+ is self-financing,

(ii) the value Vt := ηt At + ξt St , t ∈ R+ , takes the form

Vt = St g(t, Zt ), t ∈ R+ ,

for some function g ∈ C 1,2 ((0, ∞)2 ).

Then the function g(t, z) satisfies the PDE

1 1 ∂2g
 
∂g ∂g
(t, z) + − rz (t, z) + σ 2 z 2 2 (t, z) = 0, (10.23)
∂t T ∂z 2 ∂z
under the terminal condition

g(T, z) = z + , z ∈ R, (10.24)

and the corresponding replicating portfolio Delta is given by

∂g
(t, Zt ),
ξt = g(t, Zt ) − Zt t ∈ [0, T ]. (10.25)
∂z
Proof. By the Itô formula applied to 1/St we have

1 1
 
= −µ + σ 2 dt − σdBt ,



d
St St

hence
1 Λt
  
dZt = d −K
St T
 
Λt K
=d −
T St St
1 1
   
Λt
= d − Kd
T St St
1 dΛt 1
   
Λt
= + −K d
T St T St
1
 
dt
= + St Zt d
T St

" 367
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

dt
= + Zt −µ + σ 2 dt − Zt σdBt .


T
By the self-financing condition (5.8) we have

dVt = ηt dAt + ξt dSt

= rηt At dt + µξt St dt + σξt St dBt , (10.26)

t ∈ R+ . An application of Itô’s formula to f (t, x, y) leads to

d(St g(t, Zt )) = g(t, Zt )dSt + St dg(t, Zt ) + dSt · dg(t, Zt )

∂g ∂g
= g(t, Zt )dSt + St (t, Zt )dt + St (t, Zt )dZt
∂t ∂z
1 ∂2g
+ St 2 (t, Zt )(dZt ) + dSt · dg(t, Zt )
2
2 ∂z
∂g
= µSt g(t, Zt )dt + σSt g(t, Zt )dBt + St (t, Zt )dt
∂t

∂g 1 ∂g ∂g
+St Zt −µ + σ 2 (t, Zt )dt + St (t, Zt )dt − σSt Zt (t, Zt )dBt
∂z T ∂z ∂z
1 ∂2g ∂g
+ σ 2 Zt2 St 2 (t, Zt )dt − σ 2 St Zt (t, Zt )dt
2 ∂z ∂z
∂g
∂g 1 ∂g
= µSt g(t, Zt )dt + St (t, Zt )dt + St Zt −µ + σ 2 (t, Zt )dt + St (t, Zt )dt
∂t ∂z T ∂z
1 ∂2g ∂g
+ σ 2 Zt2 St 2 (t, Zt )dt − σ 2 St Zt (t, Zt )dt
2 ∂z ∂z
∂g
+σSt g(t, Zt )dBt − σSt Zt (t, Zt )dBt .
∂z
By respective identification of the terms in dBt and dt in (10.26) and (10.17)
we get
∂g ∂g


 rηt At + µξt St = µSt g(t, Zt ) + St (t, Zt ) − µSt Zt (t, Zt )

 ∂t ∂z
1 ∂g 1 ∂2g


+ St (t, Zt ) + σ 2 Zt2 St 2 (t, Zt ),


2

T ∂z ∂z





 ξ S σ = σS g(t, Z ) − σS Z ∂g (t, Z ),



t t t t t t t
∂z
hence
1 ∂g 1 2 2 ∂2g

∂g
 rVt − rξt St = St ∂t (t, Zt ) + T St ∂z (t, Zt ) + 2 σ Zt St ∂z 2 (t, Zt ),




 ξt = g(t, Zt ) − Zt ∂g (t, Zt ),



∂z
368 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

i.e.
1 1 ∂2g
  
∂g ∂g
 (t, z) + − rz (t, z) + σ 2 z 2 2 (t, z) = 0,
2

 ∂t

T ∂z ∂z

 ξ = g(t, Z ) − Z ∂g (t, Z ),



t t t t
∂z
under the terminal condition g(T, z) = z + , z ∈ R, which follows from (10.21).

When Λt /T > K we have Zt > 0 and (10.22) and (10.25) show that

∂g
ξt = g(t, Zt ) − Zt (t, Zt )
∂z
1 − e −(T −t)r
= e −(T −t)r Zt + − e −(T −t)r Zt
rT
1 − e −(T −t)r
= , t ∈ [0, T ],
rT
which recovers (10.19). Similarly, from (10.24) we recover

∂g
ξT = g(T, ZT ) − ZT (T, ZT ) = ZT 1{ZT >0} − ZT 1{ZT >0} = 0
∂z
at maturity.

We also check that

∂f ∂f
ξt = e −(T −t)r σSt f (t, St , Zt ) − σZt f (t, St , Zt )
 ∂x ∂z

∂g
= e −(T −t)r −Zt (t, Zt ) + g(t, Zt )
∂z
1 1 wt
   !
∂g
= e −(T −t)r
St t, Su du − K + g(t, Zt )
∂x x T 0 |x=St
 w
1 1 t
  
∂
= x e −(T −t)r g t, Su du − K , t ∈ [0, T ].
∂x x T 0 |x=St

We also find that the amount invested on the risk-free asset is given by
∂g
ηt At = Zt St (t, Zt ).
∂z
Next we note that a PDE with no first order derivative term can be obtained
using time-dependent coefficients.

" 369
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

(2) - Time-dependent coefficients

Define now the auxiliary process

1 − e −(T −t)r 1 1 wt
 
Ut := + e −(T −t)r Su du − K
rT St T 0
1
= (1 − e −(T −t)r
)+ e −(T −t)r
Zt , t ∈ [0, T ],
rT
i.e.
e (T −t)r − 1
Zt = e (T −t)r Ut + , t ∈ [0, T ].
rT
We have
1 −(T −t)r
dUt = − e dt + r e −(T −t)r Zt dt + e −(T −t)r dZt
T
= e −(T −t)r σ 2 Zt dt − e −(T −t)r σZt dBt − (µ − r) e −(T −t)r Zt dt
= − e −(T −t)r σZt dB̂t , t ∈ R+ ,

where
µ−r
dB̂t = dBt − σdt + dt = dB̃t − σdt
σ
is a standard Brownian motion under
2 ST ∗
dP̂ = e σBT −σ t/2
dP∗ = e −rT dP .
S0
Lemma 10.8. The Asian option price can be written as

1 wT
" +  #

St h(t, Ut ) = e −(T −t)r IE∗ Su du − K  Ft ,

T 0 

where the function h(t, y) is given by

ˆ (UT )+  Ut = y ,
h(t, y) = IE 0 6 t 6 T.
  

Proof. We have

1 1 wT
 
UT = Su du − K = ZT ,
ST T 0

and
dP̂|Ft 2 e −rT ST
= e σ(BT −Bt )−(T −t)σ /2 = −rt ,
dP∗|Ft e St
hence the price of the Asian option is

e −(T −t)r IE∗ [ST (ZT )+ | Ft ] = e −(T −t)r IE∗ [ST (UT )+ | Ft ]
370 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

"  #
e −rT ST 
= St IE∗ (UT ) +
 Ft
e −rt St 
"  #
dP̂|Ft 
= St IE∗ (UT )+  Ft

dP∗|Ft 
ˆ
= St IE[(UT ) | Ft ].
+


The next proposition gives a replicating hedging strategy for Asian options.
See § 7.5.3 of [Shr04] and references therein for a different derivation of the
PDE (10.27).
Proposition 10.9. Let (ηt , ξt )t∈R+ be a portfolio strategy such that
(i) (ηt , ξt )t∈R+ is self-financing,

(ii) the value Vt := ηt At + ξt St , t ∈ R+ , takes the form

Vt = St h(t, Ut ), t ∈ R+ ,

for some function h ∈ C 1,2 ((0, ∞)2 ).

Then the function h(t, z) satisfies the PDE

2
σ2 1 − e −(T −t)r ∂2h

∂h
(t, y) + −y (t, y) = 0, (10.27)
∂t 2 rT ∂y 2
under the terminal condition

h(T, z) = z + ,

and the corresponding replicating portfolio is given by

∂h
ξt = h(t, Ut ) − Zt (t, Ut ), t ∈ [0, T ].
∂y
Proof. By the self-financing condition (10.16) we have

dVt = rVt dt + (µ − r)ξt St dt + σξt St dBt , (10.28)

t ∈ R+ . By Itô’s formula we get

d(St h(t, Ut )) = h(t, Ut )dSt + St dh(t, Ut ) + dSt · dh(t, Ut )

= µSt h(t, Ut )dt + σSt h(t, Ut )dBt
1 ∂2h
 
∂h ∂h
+St (t, Ut )dt + (t, Ut )dUt + (t, Ut )(dUt )2
∂t ∂y 2 ∂y 2

" 371
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

∂h
+ (t, Ut )dSt · dUt
∂y
∂h
= µSt h(t, Ut )dt + σSt h(t, Ut )dBt − St (µ − r) (t, Ut )Zt dt
∂y
σ2 2 ∂ 2 h
 
∂h ∂h
+St (t, Ut )dt − σ (t, Ut )Zt dB̃t + Zt 2 (t, Ut )dt
∂t ∂y 2 ∂y
∂h
−σ St (t, Ut )Zt dt
2
∂y
∂h
= µSt h(t, Ut )dt + σSt h(t, Ut )dBt − St (µ − r) (t, Ut )Zt dt
∂y
σ2 2 ∂ 2 h
 
∂h ∂h
+St (t, Ut )dt − σ (t, Ut )Zt (dBt − σdt) + Zt 2 (t, Ut )dt
∂t ∂y 2 ∂y
∂h
−σ St (t, Ut )Zt dt.
2
∂y

By respective identification of the terms in dBt and dt in (10.28) and (10.17)

we get
∂h ∂h

 rηt At + µξt St = µSt h(t, Ut ) − (µ − r)St Zt (t, Ut )dt + St (t, Ut )



 ∂y ∂t
σ2 ∂2h

+ St Zt2 2 (t, Ut ),



2

∂y






 ∂h
 ξt = h(t, Ut ) − Zt (t, Ut ),


∂y
hence
∂h σ2 ∂2h

 rη A = −rSt (ξt − h(t, Ut )) + St (t, Ut ) + St Zt2 2 (t, Ut ),
 t t 2

 ∂t ∂y

 ξt = h(t, Ut ) − Zt ∂h (t, Ut ),



∂y
and 2 2
σ 2 1 − e −(T −t)r
 
∂h ∂ h
(t, y) + − y (t, y) = 0,


2

 ∂t
 rT ∂y 2

1 − e −(T −t)r
  
 ∂h
 ξt = h(t, Ut ) + (t, Ut ),

 − Ut
rT ∂y
under the terminal condition

h(T, z) = z + .

372 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options


We also find the risk-free portfolio allocation

1 − e −(T −t)r ∂h
 
∂h
ηt At = e (T −t)r St Ut − (t, Ut ) = St Zt (t, Ut ).
rT ∂y ∂y

Exercises

Exercise 10.1 Consider the short rate process rt = σBt , where (Bt )t∈R+ is
a standard Brownian motion.
rT
a) Find the probability distribution of the time integral 0 rs ds.
b) Compute the price

wT
" + #
∗
e −rT
IE ru du − κ
0

rT
of a caplet on the forward rate 0
rs ds.

Exercise 10.2 Asian call options with negative strike price. Consider the
asset price process
2
St = S0 e rt+σBt −σ t/2
, t ∈ R+ ,

where (Bt )t∈R+ is a standard Brownian motion. Assuming that κ 6 0, com-

pute the price

1 wT
" +  #
∗
e −(T −t)r
IE Su du − κ  Ft

T 0 

of the Asian option at time t ∈ [0, T ].

Exercise 10.3 Compute the price

"  w +  #
1 T

∗
e −(T −t)r
IE exp log Su du − K  Ft , 0 6 t 6 T.

T 0

at time t of the geometric Asian option with maturity T , where St =

2
S0 e rt+σBt −σ t/2 , t ∈ [0, T ].

Hint: When X ' N (0, v 2 ) we have

" 373
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 N. Privault

2
IE∗ [( e m+X − K)+ ] = e m+v /2
Φ(v + (m − log K)/v) − KΦ((m − log K)/v).

Exercise 10.4 Consider a CIR process (rt )t∈R+ given by

√
drt = −λ(rt − m)dt + σ rt dBt , (10.29)

where (Bt )t∈R+ is a standard Brownian motion under the risk-neutral mea-
sure P∗ , and let
1 wt
Λt := rs ds, t ∈ [τ, T ].
T −τ τ
Compute the price at time t ∈ [τ, T ] of the Asian option with payoff
+
(ΛT − K) , under the condition Λt > K.

Exercise 10.5 Consider an asset price (St )t∈R+ which is a submartingale un-
der the risk-neutral measure P∗ , in a market with risk-free interest rate r > 0,
and let φ(x) = (x − K)+ be the (convex) payoff function of the European
call option.

Show that, for any sequence 0 < T1 < · · · < Tn , the price of the average
option with payoff
ST1 + · · · + STn
 
φ
n
can be upper bounded by the price of the European call option with maturity
Tn , i.e. show that

ST1 + · · · + STn
  
IE∗ φ 6 IE∗ [φ(STn )].
n

Exercise 10.6 Let (St )t∈R+ denote a risky asset whose price St is given by

dSt = µSt dt + σSt dBt ,

where (Bt )t∈R+ is a standard Brownian motion under the risk-neutral mea-
sure P∗ . Compute the price at time t ∈ [τ, T ] of the Asian option with payoff

1 wT
 +
Su du − K ,
T −τ τ

under the condition that

1 wt
At := Su du > K.
T −τ τ

374 "
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html
 Asian Options

Exercise 10.7 Pricing Asian options by PDEs. Show that the functions g(t, z)
and h(t, y) are linked by the relation

1 − e −(T −t)r
 
g(t, z) = h t, + e −(T −t)r z , t ∈ [0, T ], z > 0,
rT

and that the PDE (1.35) for h(t, y) can be derived from the PDE (1.33) for
g(t, z) and the above relation.

Exercise 10.8 Hedging Asian options [YEM11].

a) Compute the Asian option price f (t, St , Λt ) when Λt /T > K.
b) Compute the hedging portfolio allocation (ξt , ηt ) when Λt /T > K. When
Λt /T > K we have

1 − e −(T −t)r
 
Λt
ξt = and ηt At = e (T −t)r −K , t ∈ [0, T ].
rT T

c) At maturity we have f (T, ST , ΛT ) = (ΛT /T − K)+ , hence ξT = 0 and

+
e −rT
  
ΛT ΛT
η T AT = AT −K 1{ΛT >KT } = −K .
A0 T T

d) Show that the Asian option with payoff (ΛT − K)+ can be hedged by the
self-financing portfolio

1 1 Λt
     
Λt
ξt = f (t, St , Λt ) − e −(T −t)r − K h t, −K
St T St T

in the asset St and

e −rT Λt 1 Λt
    
ηt = − K h t, −K , t ∈ [0, T ],
A0 T St T

in the risk-free asset At = A0 e rt , where h(t, z) is solution to a partial

differential equation to be written explicitly.

Exercise 10.9 Compute the first and second moments of the time integral
wT
St dt for τ ∈ [0, T ), where (St )t∈R+ is the geometric Brownian motion
τ
2
St := S0 e σBt +rt−σ /2
, t ∈ R+ .

" 375
This version: June 23, 2018
http://www.ntu.edu.sg/home/nprivault/indext.html