Lemma

Ito
 Ag
By def I BsdBs ⼆ 点 BED
⽚ 乱旺叮
feed 三星是 ⼀則
⼀

Q Easier way
A Yes 關

How Chain rule 㽰 Itis Lemma

df Bt f Bdd Bttf BD dat 2

dBE
⼆
2Btd Bttdt

4x x f x 2x fizj

o.BE BE2I Bsd Bstt

ff Bsd B 全 BED a
Chapter 7

The Ito’s Formula Fundamental Th

ofstochy
We will derive Ito’s lemma, i.e., the chain rule in the Ito’s integrals. It is one of the most
fundamental results in Stochastic Calculus, and is a very useful rule in calculating Ito’s integrals.

Despite many similarities, calculus and stochastic calculus have fundamental di↵erences.

• In calculus, we start with di↵erentiation, and then move onto integration. The two
concepts are linked by the fundamental theorem of calculus, from which we derive chain
rule, integration by parts.
• In stochastic calculus, however, we have no di↵erentiation theory, only integration theory.
We have Ito formula (i.e., fundamental theorem of stochastic calculus), from which we
derive corresponding chain rule, integration by parts.

7.1 Review on Riemann integrals

We don’t use the definition to calculate Riemann integrals. Instead, we use the fundamental
theorem of calculus, which relates the di↵erentiation and integration, the two central concepts
in calculus.

Theorem 7.1 (Fundamental Theorem of Calculus) f is continuous on [a, b].

Rx
fi deri of

F
1. If F (x) = a f (t)dt, then F 0 (x) = f (x).
Rb
2. If f (x) = F 0 (x), then a f (x)dx = F (b) F (a).
antideri off
Z t
t2 x2
Example 7.1 xdx = F (t) F (0) = , where f (x) = x and F (x) = .
0 2 2

The chain rule and integration by parts formula follow easily from the Fundamental The-
orem of Calculus.

逦
89
Theorem 7.2 (Chain rule or substitution rule)
SBdB or
And
dy dy du
1. [f (g(x))]0 = f 0 (g(x))g 0 (x). Or = , where y = f (u) and u = g(x).
dx du dx
2. Integrating the above and applying Theorem 7.1, we get
Z t Z t Z g(t) !
0 0 0 0
f (g(t)) f (g(0)) = f (g(s))g (s)ds = f (g(s))dg(s) = f (u)du.
0 0 g(0)

ff
Example
Rt
7.2 Take y = u and u = g(x). We have [g 2 (x)]0 = 2g(x)g 0 (x), and g 2 (x)
2
g 2 (0) =
2 0 g(s)dg(s). However, this chain rule fails for Bt as
Z t
2 Bs dBs = Bt2 t 6= Bt2 .
0

(In fact, the LHS is a martingale, but the RHS is not (it is a submartingale.)

Theorem 7.3 (Integration by parts formula)

Z b Z b
0
f (x)g (x) dx = f (x)g(x)|ba f 0 (x)g(x) dx
dg ⼆ a a

BxdBx0
Proof. By [uv] = u0 v + uv 0 and the Fundamental Theorem of Calculus, we have
Z b Z b Z b
0 0
f (b)g(b) f (a)g(a) = [f (x)g(x)] dx = f (x)g(x) dx + f (x)g 0 (x) dx.
a a a

ii
Rx Rx R
Example 7.3 0 x cos(x)dx = 0 xd sin(x) = x sin(x) sin(x)dx = x sin(x) + cos(x).

Example 7.4
R
Integration by parts
Rt
formula fails for RBrownian motion. If not, we
would have 0t Bs dBs = Bt2 B02 0 B s dB t 1 2
s or equivalently 0 Bs dBs = 2 Bt , which is clearly
false.
i
墜 ⾲忙 與 伽 㘜

lòi
90
 BI lim 舙 BF li
7.2 Variation and co-variation processes

Definition 7.1 Let Xt and Yt be two continuous stochastic processes. Let 0 = t0 < t1 <
... < tn = t, tk = tk tk 1 , and k X = Xtk Xtk 1 .
Q i.
• The covariation process of X and Y is: hX, Y i0 = 0
X
hX, Y it = lim ( k X)( kY ) ( limit in probability).
tk !0
tk t

• The quadratic variation process of Xt is: hXi0 = 0 and

X
2
hXit =: hX, Xit = lim ( k X) ( limit in probability).
tk !0

nIi
tk t

Remark 7.1 It is easy to see that

1
p
fnI4ax t iii 1
hX, Y it := (hX + Y it hX Y it ) = (hX + Y it hXit hY it ) ,
4 2
which is similar to relationship between variance and covariance:
1 1
cov(X, Y ) = (var(X + Y ) var(X Y )) = (var(X + Y ) var(X) var(Y )) .
4 2

Theorem 7.4 X, Y , and Z are continuous stochastic processes. Then,

1. hXit ".

2. hX, Y it = hY, Xit

3. hX + Y, Zit = hX, Zit + hY, Zit

4. hX X0 , Zit = hX, Zit

5. haX, bY it = abhX, Y it , where a, b 2 R

6. haXit = a2 hXit , where a 2 R

圝
Proof. Left as an exercise.

幽 之 比非
o 螽
tis
中
Ǜǜ
與

91
Roof of 30

2_Ylqz
地刮缸
It

In it 三継 红
tp
Ǜzst 42 名

i
Theorem 7.5 At and Xt are both continuous, and At is of bounded variation. Then

(a) hAit = 0 a.s.

㬘
(b) hA, Xit = 0 a.s.
(c) hA + Xit = hXit a.s.

Proof. (a) follows from (c) by taking X = 0. (b) also follows from (c): hA, Xit = 14 (hX + Ait hX Ait
1
4
(hXit hXit ) = 0. Thus, it suffices to prove (c) below.

Denote kY = Yt k Ytk 1 , where either Y = A or Y = X. Then, the key step is:

X EIGAI.lt !0
@
X Tkil
1
A
Eh Al
( k A)( kY )  sup | kY | | k A| ! 0, (2.1)
tk t k tk t

as the continuity of Y implies supk | kY | ! 0 while the bounded variation of A implies

P
tk t | k A| < 1. Therefore,
X X
| k (X + A)|2 = | kX + k A|
2

tk t tk t
X X X
2 2
= ( k X) + ( k A) +2 ( k X)( k A)
tk t tk t tk t
! hXit + 0 + 0, in probability. (from (2.1))

On the other hand, the LHS ! hX + Ait in probability by definition. The result then follows
from the uniqueness of the probability limit.

Part (a) of Theorem 7.5 can be further strengthened.

Theorem 7.6 If At is a continuous function of bounded p-variation with 1  p < 2, then

we have hAit = 0.

Proof. We have
X X
2 p 2 p
| k A| = | k A| | k A|
tk t tk t
0 1
2 p
X
@ pA
 (sup | k A|) | k A|
tk t
2 p
 C (sup | k A|)
! 0.

92
Variation processes of Ito’s integrals

Theorem 7.7 hBit = t.

The proof has been done before. This is used in Theorem 7.15 to characterize Brownian motion:
If Xt is a continuous martingale with hXit = t, Xt must be a Brownian motion.

Rt Rt
Theorem 7.8 Let Xt = 0 Cs dBs and Yt = 0 Ds dBs . Then
Z t Z t
hXit = Cs2 dhBis = Cs2 ds
0 0
Z t Z t
hX, Y it = Cs Ds dhBis = Cs Ds ds
0 0

Proof. Too long, omitted here.

93
 7.3 Ito’s integration w.r.t. martingales and semi-martingales
R

o
In defining 0t Cs dBs , we have not used properties of Bt much except the Isometry property.
For more general continuous martingales Xt , we could Rt
establish the corresponding Isometry
property. Then the⇣ Ito integral⌘of C w.r.t. X is a r.v. 0 Cs dXs : ⌦ ! R, defined to be the L2
P
limit of ni=1 Cti 1 Xti Xti 1 as the mesh of the partition 0 = t0 < t1 < ... < tn = t of [0, t]
tends to 0.
四
Definition 7.2 Xt is said to be a continuous semi-martingale if

X t = M t + At ⼆
1 些
where Mt and At are continuous martingale and adapted process, respectively. The Ito integral
of C w.r.t. X is defined to be

Eit
Z t Z t Z t
Cs dXs =: Cs dMs + Cs dAs ,
0 0 0

where the two integrals on the RHS are Ito’s and R-S integrals, respectively.

By Doob-Meyer decomposition, sub- and super-martingales are both semimartingales.

But semimartingales are much more general processes.

fly ⼆ fytfly fykfytfyxttif

yytifhlyj
Put to
⼆
Yt y ⼆ Xttot

Xtkfngffj
94
 CXfdtA
7.4 One-dimensional Ito’s formula

Theorem 7.9 If Xt is a continuous semi-martingale, and f 00 is continuous, then

Z t
ihTF.li
1 Z t 00
f (Xt ) f (X0 ) = f 0 (Xs )dXs + f (Xs )dhXis .
0 2 0
Formally, we could write this in “di↵erential form”:
I
SDEO
1
df (Xt ) = f 0 (Xt )dXt + f 00 (Xt )dhXit .
2

in
Remark: Intuitively, Ito’s formula is just a local two-term Taylor expansion in that
1
f (Xt+ t ) f (Xt ) ⇡ f 0 (Xt ) (Xt+ t Xt ) + f 00 (Xt ) (Xt+ t Xt )2 .
2

Proof. We will consider the case where Xt is a continuous martingale, from which
the case for continuous semi-martingale follows easily. First we have
1
f (b) f (a) = (b a)f 0 (a) + f 00 (c(a, b)) (b a)2 .

to
2
where c = c(a, b) in between a and b. Let 0 = t0 < t1 < ... < tn = t. Therefore,
n
X
f (Xt ) f (X0 ) = [f (Xti ) f (Xti 1 )]
i=1
ˋ

o
Xn ⇣ ⌘ 1X n ⇣ ⌘⇣ ⌘2
0
f 00 c(Xti 1 , Xti ) Xti

凹
= f (Xti 1 ) Xti X ti 1
+ X ti 1
i=1 2 i=1
1
O
=: In1 + In2 .
2
ˋifd Bs
R
• ClearlyR In1 !p 0t f 0 (Xs )dXs as n ! 1. By taking subsequences we can suppose that
In1 ! 0t f 0 (Xs )dXs , a.s.

• For In2 , we can rewrite it as In2 =

Z t

0
Gns dAns , where is_sn
⇣ ⌘
Gns = f 00 c(Xti 1 , Xti ) , when s 2 (ti 1 , ti ]
= f 00 (Xt ) , when s t

and dh lt
X⇣ ⌘2
Ans = X ti X ti 1
.
tk s

Now Ans !p hXis . By taking subsequences we can suppose that, as n ! 1,

Ans ! hXis , a.s. (4.2)

豳 95

Xti FK.tk
 Next let us look at the integrand Gns . Observe that the uniform continuity of f 00 implies
Gns ! f 00 (Xs^t ) uniformly in s. Since Xs is continuous, it is bounded on [0, t], i.e.
|Xs |  C for s 2 [0, t]. Therefore, |c(Xti 1 , Xti )|  C. Since f 00 is assumed continuous,
thus |f 00 (r)|  C0 for |r| 2 [0, C], implying that

DCTO 7
|Gns |  C0 . (4.3)

In view of (4.2) and (4.3), for almost all !, we can apply the bounded convergence theorem
to obtain
Z t Z t

斷
In2 = Gns dAns ! f 00 (Xs )dhXis .
1 0 0

Remark 7.2 点点 Huno

1. Martingales (semi-martingales)
Rt
are closed under the Ito’s integration. (If Xt is a martin-
gale (semi-martingale), so is 0 Ct dXt .)

2. By Ito’s lemma, martingales are NOT closed under smooth transformation, but semi-
martingales are.
Rt 0
3. If Xt = At (B.V.), then hAit = 0, we have f (At ) f (A0 ) = 0f (A s )dAs . In fact, this
holds if f 0 is continuous.
⇣ ⌘
4. Since i B ⇠ N (0, i t), we have i B = Op ( i t)1/2 or ( i B)
2
= Op ( i t) . The fol-
lowing heuristics are often helpful in formal derivations:

dBt = (dt)1/2 , dBt · dBt = dt,

dt · dBt = (dt)3/2 = o(dt) = 0, dt · dt = (dt)2 = o(dt) = 0.

Mii
Why dfy f Mtjd Mtt If Mt day

f M AM
Áiyntifǖbd修
斎 3

96
主 _MTG
 Theorem 7.10 At is a continuous function of bounded variation, and f 0 is continuous, then
Z t
f (At ) f (A0 ) =
0
f 0 (As )dAs .
0
n_nProof.
f (At ) f (A0 ) =
n
X
[f (Ati ) f (Ati 1 )]
mi
i=1
Xn ⇣ ⌘⇣ ⌘
= f 0 c(Ati 1 , Ati ) At i Ati 1
i=1
Xn ⇣ ⌘
= f 0 (Ayi ) Ati Ati 1
i=1
(9yi 2 [ti 1 , ti ] s.t. Ayi is in between Ati 1
and Ati as At is continuous.)
Z t
! f 0 (As )dAs .
0

Rt
Example 7.5 Find 0 Bsm dBs where m 1.

Solution. Take Xs = Bs and f 0 (x) = xm . Then f (x) = xm+1 /(m + 1), f 00 (x) = mxm 1 , and
hXis = hBis = s. Applying the Ito’s formula, we get
Z t
Btm+1 B0m+1 m 1Z t
= Bs dBs + mBsm 1 ds
m+1 m+1 0 2 0
Hence,
Z t
B m+1 mZ t m 1
Bsm dBs = t B ds. (4.4)
0 m+1 2 0 s
Rt
In particular, if m = 1, we have 0 Bs dBs = (Bt2 t)/2.
R B m+1
Alternative solution: We guess that 0t Bsm dBs should be roughly like m+1t
, which we can
m+1
apply Ito’s lemma to find out the exact relationship. (Let f (x) = x /(m + 1) and take
x = Bt .)

fties

tof Bid Biijmb

then
Xt Bt⼆

家譽 器 器 fhii
97

findBs ⼆
器 _jns
Miywe guess isi

By Ito
leg 她
與 鼠瓷器品

dfy f Bid It If Btldt

ǛZ
d B ⼆
lmtl Bid It 池州
mid d

B
TdBttjid
1
o fpidB.tt Bids
same solution as above

lvagMedni.ae
7.5 Multi-dimensional Ito’s formula

Theorem 7.11 Let Xt = (Xt1 , ..., Xtd ) and f : Rd ! R. Then,

d Z t
X 1 X Zt
f (Xt ) f (X0 ) = fi (Xs )dXsi + fij (Xs )dhX i , X j is
i=1 0 2 1i,jd 0

In di↵erential form, this is

X 1X
df (Xt ) = fi dXti + fij dhX i , X j it .
i 2 i<j

Proof. The proof is similar to that in one-dimensional case, and hence omitted.

Corollary 7.1 In Theorem 7.11, if Xt1 , ..., Xtc are continuous semi-martingales, and Xtc+1 , ..., Xtd
are of bounded variation, then we have
d Z t
X 1 X Zt
f (Xt ) f (X0 ) = fi (Xs )dXsi + fij (Xs )dhX i , X j is .
i=1 0 2 1i,jc 0

Corollary 7.2 If Xt is a continuous semi-martingale, then

Z t Z t
1Z t
f (t, Xt ) f (0, X0 ) = f1 (s, Xs )ds + f2 (s, Xs )dXs + f22 (s, Xs )dhXis .
0 0 2 0
or
1
df (t, Xt ) = f1 (t, Xt )dt + f2 (t, Xt )dXt + f22 (t, Xt )dhXit .
2

Rt
Example 7.6 Find 0 sdBs .

Solution 1. Take Xs = Bs and f2 (x, y) = x. Then f (x, y) = xy, f1 (x, y) = y and f22 (x, y) = 0,
and hXis = hBis = s. Applying the Ito’s formula, we get
Z t Z t
tBt 0= Bs ds + sdBs + 0.
0 0
Rt Rt
Hence, 0 sdBs = tBt 0 Bs ds.

Solution 2. Applying Ito’s formula to f (t, Bt ) = tBt , where f (x, y) = xy, we have
1
d(tBt ) = df (t, Bt ) = f1 dt + f2 dBt + f22 dhBit = Bt dt + tdBt + 0.
2

98
 Z dI

dfx.lk fdyfdyt
合 fdcxiifdyl f.dk 只
⼀

zgi
fyyjfk. E

fizttātfiyddi
Ìfi 邈 ⽚ Ìfydyj
ff.tl 㖄 品 咧
fdktfzdyt
Ìfdsttffdqtfzd 化
只
 3 dto

nre cl

fy 到
⼆
fdytfdyttfd 不

Ìfikxztifd 傩 ⼗ 点 fda
tfzdaytffxytf KY 2点
 eg dflt 到
1 ⼆

Sh
df 則 fdttfzd 只

tfndttfikBIO_tfda.BG
ǒ it
ǒ
三
fdttfd Bttifndt ˊ

f 主 的 dttfd 是
nerrrrn
fnzs
are t

ftyfhzffktifzy
0t
佐與 女
 7.6 Integration by parts formula

Theorem 7.12 Xt and Yt are continuous semi-martingales, then,

Z t Z t
X t Yt X 0 Y0 =
0
⼀
Ys dXs +
0
Xs dYs + hX, Y it .

Proof. Let Xt = (Xt , Yt ), f (x, y) = xy. Hence

f1 = y, f2 = x, f11 = f22 = 0, f12 = 1.

⼀⼀
Then applying the multivariate Ito’s formula, we get

dfti_d(X Y ) tn
t t = df (Xt , Yt )
= f1 (Xt , Yt )dXt + f2 (Xt , Yt )dYt
1
x t , Yt )dhXit + f12 (Xt , Yt )dhX, Y it + 1 f22 (Xt , Yt )dhY it
fykxy
57__ㄨ
+ f11 (X
2 2
= Yt dXt + Xt dYt + dhX, Y it .

Rt
Example 7.7 Find 0 g(s)dBs , where g is di↵erentiable.

Solution. By integration by parts formula,

Z t Z t Z t
0
g(s)dBs = g(t)Bt Bs g (s)ds hg, Bit = g(t)Bt Bs g 0 (s)ds.
0 0 0

In particular, when g(x) = x, this has been done in Example 7.6 by Ito’s formula. But it is
much simpler to use integration by parts formula here.

99
7.7
Sisco
Ito exponential
SGdkz.EE 4
Rt B
One can easily check by Ito’s formula that eBt 1 6= 0e s dB s (LHS is a submartingale while
RHS is a martingale). However, we have

dfiǜi
Z t
Theorem 7.13 We have eBt t/2
1= e Bs s/2
dBs .
0

Trueifps
For a general martingale X, we have

Theorem 7.14 Xt is a continuous martingale. Then

Z t
1 1
hXit hXis
e Xt 2 1= e Xs 2 dXs . (7.5)
0

1
hXit
Consequently, eXt is a martingale.

die
2

Proof. Let Yt = Xt
1
hXit , and f (x) = ex . Then, by the Ito’s lemma,
焱
2
⼀ ✓ ◆
1 1 1
df (Yt ) = f 0 (Yt )dYt + f 00 (Yt )dhY it = f (Yt )d Xt hXit + f (Yt )dhXit = f (Yt )dXt .
2 2 2

1
Now eXt 2 hXit is called the Ito’s exponential w.r.t. Xt . The theorem will prove very useful
⼀
in the change of measure (Girsanov theorem). It is also very useful result in characterizing BM
as shown below. 0

MTG .Xt.is e
X convex

efissub
MTGMIgtB.n.li
ge Doob

蕋不⼆ fist is MTG

100
Proof Yt
Xttiue
de
fy
⼆
dfy fmyyyfyufy_
下
世
f x d

ědyttiétckx

⼆
eidd 北非 與 不 ⼆
位

ědxifetdytiehkxz
ědxt
ettstdxt I

Know Bt is B.M o B z t
只 is cts MTG
7.8 Levy characterization of Brownian motion

Theorem 7.15 (Levy characterization of Brownian motion) Wt is a Brownian motion

if and only if

• Wt is a continuous martingale with W0 = 0;

• hW it = t.

Proof. The “=)” part is trivial. Let us check the “(=” part. We must check

(a) Wt Ws ⇠ N (0, t s),

(b) Wt Ws ? Fs ,
(c) Wt is continuous. 但 ⼆ 比

準
We only need to check (a) and (b) as (c) is assumed. From Theorem 7.14,
✓ ◆ ✓ ◆
1 1 2
Mt =: exp ✓Wt h✓Wt i = exp ✓Wt ✓ t
2 2
is a martingale. That is

E(Mt |Fs ) = Ms
✓ ✓ ◆ ◆ ✓ ◆
1 2 1 2
() E exp ✓Wt ✓ t |Fs = exp ✓Ws ✓ s
2 2
✓ ◆
1
() E (exp (✓[Wt Ws ]) |Fs ) = exp ✓2 (t s)
nrr
2
tz ✓
1 2
◆
() E (exp (✓[Wt Ws ])) = E (exp (✓[Wt Ws ]) |Fs ) = exp ✓ (t s) ,
2
which implies (a) and (b).

o fiabtside vfEexpldwt.ws
中 up 住 024⼀⾏
⼆
E 胎
mgfofhf yz.lt 㝤 mgf of NO ts

101
7.9 Geometric Brownian motion (GBM)

Bachelier (1900) used Brownian motion to model stock markets, but the shortcoming of this
approach is that Brownian motion could be negative. To avoid this problem, Samuelson (1964)
introduced geometric Brownian motion (GBM) to model stock markets.

2 /2)t+
Example 7.8 The GBM Xt = e(µ Bt
satisfies
Z t Z t

SDE Xt = X0 + µXs ds + Xs dBs .

If
0 0

its Yt
dxt ⼆
Mkdt 5Xtd Bt
2 x
Proof. Let Xt = e = f (Yt ), where Yt = (µ /2)t + B and f (x) = e . Then,
t

dXt = df (Yt )
1
= f 0 (Yt )dYt + f 00 (Yt )dhY it
2
不⼆ Bfft
⇣ ⌘ 1
2
= f (Yt )d (µ /2)t + Bt + f (Yt )d( 2 t)
2
= f (Yt )d (µt + Bt )
= µXt dt + Xt dBt .

xh dk

Mkdt 5Xtd Bt
7
⼤哭 Mdt 千 5d Bt

在 只⼀則 ii NG.at 5到
等 ⼆
Not
5

Tonti 102

Ǖ
Is GBM appropriate?

Empirical evidence shows and most economists agree that the relative returns (or log returns)
of stock markets for the same time length can be regarded as constants and the relative returns
over non-overlapping intervals remain roughly uncorrelated. As a first approximation, we
could describe this phenomenon by
Xt+ t Xt
⇡ µ t + (Bt+ t Bt ), (9.6)
Xt
Note that
✓ ◆ ✓ ◆
Xt+ t Xt Xt+ t Xt 2
E = µ t, var = t,
Xt Xt

hence µ and 2 represent the average rate and variance (or volatility) of relative return per
unit. Letting t ! 0 in (9.6), we formally get dX Xt
t
= µdt + dBt , or equivalently, dXt =
2
µXt dt + Xt dBt . From the last example, Xt = e(µ /2)t+ Bt .

Empirical evidence also shows that, although the relative returns over non-overlapping
intervals are roughly uncorrelated, the square of the relative returns over non-overlapping in-
tervals are correlated. This means that the model (9.6) is not really appropriate to describe
stock markets. This implies that the stock price is a GBM in practice. Despite this, GBM
o↵ers a very good starting point.
Noīhn
One way to improve the GBM is to introduce the so-called general Ito process:

dXt = µ(t, Xt )dt + (t, Xt )dBt .

The shortcoming is that we no longer have simple interpretations for µ(t, Xt ) and (t, Xt )
anymore.

103
7.11 Multi-factor SDE

Combine several factors into one

A process S may be “driven” by several Brownian motions. The next theorem shows that it
may be expressed in a form “driven” only by one Brownian motion. We illustrate this by a
two-factor SDE.

Theorem 7.16 Let

dSt = a1t (!)dB1t + a2t (!)dB2t ,
where B1t and B2t are standard BMs with dhB1 , B2 it = dt. Assume that hSi0t = a21t (!) +
2 a21t (!)a22t (!) + a22t (!) > 0 a.s. Then, we have
q dSt q
dSt = hSi0t q =: hSi0t dWt ,
hSi0t
Z t Z t
dSt a1t (!)dB1t + a2t (!)dB2t
where Wt = q = q is a standard BM.
0 hSi0t 0 hSi0t

Proof. Noting that St is a continuous martingale, we can easily show that

• Wt is a continuous martingale with W0 = 0;

R t dhSit Rt
• hW it = 0 hSi0t = 0 1dt = t for t 0.

By Theorem 7.15, Wt is a Brownian motion.

Remark 7.3 Let B1t , B2t be two standard Brownian motions with dhB1 , B2 it = dt. Note that
corr(B1t , B2t ) = E(B1t B2t )/t = EhB1 , B2 it /t, we have

1. If = 0, we have hB1 , B2 it = 0, and hence corr(B1t , B2t ) = 0. Therefore, B1 and B2 are

orthogonal (i.e., uncorrelated) hence independent.
2. If = ±1, we have hB1 , B2 it = ±t. Thus, corr(B1t , B2t ) = ±1, resulting in B1t = ±B2t .

Therefore, is simply the correlation of the two Brownian motions.

Remark 7.4 We assumed that dhB1 , B2 it = dt, where is assumed to be free of t. This is
a simplification. In general, however, may depend on t. For example, in Theorem 7.16, for
two Brownian motions B1 and W , we have
1
dhB1 , W it = q (a1t + a2t ) dt.
hSi0t

105
 (i)
Corollary 7.3 Let {Bt , t 0} be two Brownian motions with the correlation coefficient ⇢.
Then, B̃t is a Brownian motion, where
(1) (2)
a1 B t + a2 B t
B̃t = 1/2
.
(a21 + a22 + 2⇢a1 a2 )

Proof. This follows from the last theorem.

(1) (2) ai
Direct proof. Write B̃t =: c1 Bt + c2 Bt , where ci = 1/2
. Then,
(a21 + a22 + 2⇢a1 a2 )

1. B̃t is a Gaussian process.

⇣ ⌘
2. E B̃t = 0, and cov B̃s , B̃t = min(s, t).

3. B̃t is continuous.

The proofs for parts 1 and 3 are trivial. Let us prove part 2. Suppose s  t, then
⇣ ⌘
E B̃t = E c1 Bs(1) + c2 Bs(2) = 0,
(2) (2) (2)
E(Bs(1) Bt ) = EE(Bs(1) Bt |Fs ) = E[Bs(1) E(Bt |Fs )] = E[Bs(1) Bs(2) ] = ⇢s,
(1) (1) (1)
E(Bt Bs(2) ) = EE(Bs(2) Bt |Fs ) = E[Bs(2) E(Bt |Fs )] = E[Bs(1) Bs(2) ] = ⇢s,
⇣ ⌘⇣ ⌘
(1) (2)
E[B̃s B̃t ] = E[ c1 Bs(1) + c2 Bs(2) c1 B t + c2 B t ]
⇣ ⌘
(2) (1)
= c21 s + c22 s + c1 c2 E[Bs(1) Bt ] + E[Bt Bs(2) ]
⇣ ⌘
= c21 s + c22 s + 2⇢c1 c2 s = c21 + c22 + 2⇢c1 c2 s
= s.

dXt (1) (2)

Example 7.9 Let = cdt + 1 dBt + 2 dBt , where dhB (1) , B (2) it = ⇢dt. Then, Xt is a
Xt
geometric Brownian motion.

Proof. Now
dXt (1) (2)
= cdt + 1 dBt + 2 dBt
Xt
⇣ ⌘
(1) (2)
= cdt + d 1 Bt + 2 Bt
0 1
⇣ ⌘1/2 (1) (2)
2 2 1 Bt + 2 Bt
= cdt + 1 + 2 + 2⇢ 1 2 d@ 1/2
A
2 2
( 1 + 2 + 2⇢ 1 2)

=: cdt + ˜ dB̃t ,
where
(1) (2)
2 2 2 1 Bt + 2 Bt
˜ = 1 + 2 + 2⇢ 1 2, B̃t = 1/2
.
2 2
( 1 + 2 + 2⇢ 1 2)

Therefore, Xt is a geometric Brownian motion driven by B̃t . Hence,

0.5˜ 2 )t+˜ 2 B̃t
Xt = X0 e(c .

106
Example 7.10 Define
(1) (1)
dXt = µ(t, Xt )dt + (t, Xt )dBt =: µt dt + t dBt ,

(2) (2)
dYt = ⌫(t, Xt )dt + ⌘(t, Xt )dBt =: ⌫t dt + ⌘t dBt ,
Rt (2)
It can be shown hX, Y it = 0 s ⌘s dhB , B (2) is , and hB (2) , B (2) it = ⇢t. Then,

d(Xt Yt ) = Xt dYt + Yt dXt + dhX, Y it

= Xt dYt + Yt dXt + t ⌘t dhB (1) , B (2) it
= Xt dY
⇣ t
+ Yt dXt + ⇢⌘ t ⌘t dt⇣ ⌘
(2) (1)
= Xt ⌫t dt + ⌘t dBt + Yt µt dt + t dBt + ⇢ t ⌘t dt
⇣ ⌘ ⇣ ⌘
(2) (1)
= ⌫t Xt dt + ⌘t Xt dBt + µt Yt dt + t Yt dBt + ⇢ t ⌘t dt
⇣ ⌘
(2) (1)
= (⌫t Xt + µt Yt + ⇢ t ⌘t ) dt + ⌘t Xt dBt + t Yt dBt .

In particular, if µt = µXt , t = Xt , ⌫t = ⌫Yt , ⌘t = ⌘Yt , then,

⇣ ⌘
(2) (1)
d(Xt Yt ) = (⌫t Xt + µt Yt + ⇢ t ⌘t ) dt + ⌘t Xt dBt + t Yt dBt
⇣ ⌘
(2) (1)
= (⌫Xt Yt + µXt Yt + ⇢ ⌘Xt Yt ) dt + ⌘Xt Yt dBt + Xt Yt dBt
⇣ ⌘
(2) (1)
= (⌫ + µ + ⇢ ⌘) Xt Yt dt + Xt Yt ⌘dBt + dBt .

That is,

d(Xt Yt ) ⇣ ⌘
(2) (1)
= (⌫ + µ + ⇢ ⌘) dt + ⌘dBt + dBt .
X t Yt
One can then apply an earlier result to turn this into a SDE driven by one BM.

107
 7.12 Exercises choose59
1. f is Lipschitz-continuous on [0, T ] if there exists a constant C > 0 such that for any
t, t0 2 [0, T ], |f (t) f (t0 )| < C|t t0 |. Show that a Lipschitz-continuous function is of
bounded variation and its quadratic variation over [0, T ] is zero, i.e. , hf iT = 0.

2. Derive dXt and check which ones are martingales, where Xt is

(a) et/2 cos(Bt ),

0Use Ito
a fdttfd
d
Z t
2
(b) t Bt 2 sBs ds,
0
⇣ ⌘3
(c) a1/3 + Bt /3 ,
Xt ⼆ Bt
gO.se (d) exp{ Bt },
(e) Bt4
(f) tBt
6tBt2 + 3t2 ,
Rt
0 Bs ds,
Xt MTG Mt
⼆ 0
1
(g) (Bt + t) exp{ Bt 2
t},
Bt
(h) tBt e ,
Z t
dBs
(i) (1 t) , where 0  t < 1,
0 1 s
(j) (Yt ), where Yt = pBT t t , 0  t  T , and (x) is the standard Normal d.f.

3. Find f (x) such that f (Bt + t) is a martingale.

Rt
dflB.tt ⼝ dtt ⼝ d Bt
4. Show that Xt = 0 sign(Bs )dBs is a Brownian motion, where

sign(x) = 1 if x  0,
1 if x > 0.

5. Let dXt = a(Xt )dt + (Xt )dBt , X0 = 0, where (t) > 0. Find an increasing
function f (·) such that Yt = f (Xt ) satisfies

dYt = b(Yt )dt + dBt .

R Xt
6. Let Xt = eBt and Yt = 1 ln zdz.

(a) Find dXt and dhXit .

(b) Find dYt and hence show that Yt can be expressed as dYt = a(Bt )dt + b(Bt )dBt .
Specify these two functions a(x) and b(x).

7. Define k (t) = EBtk ; k = 0, 1, 2, ...; t 0.

Z t
1
(a) Use Ito’s formula to show that k (t) = k(k 1) k 2 (s)ds; k 2.
2 0
(b) Use (a) to deduce that E[Bt4 ] = 3t2 and also find E[Bt6 ].
(2k)!tk
(c) Show that E[Bt2k+1 ] = 0, and E[Bt2k ] = ; k = 1, 2, ...
2k k!
Z t
8. Let Yt := Su du, where dSt = St dBt with S0 = 1. Find EYt and V ar(Yt ).
0

108
2 h
dk Btétd 1社 沚
與
火 pssossortnrtorrtretheonntt
就 領 ⼀ 悜
.de df BD fi
HBHdBtt fydt
⼆
愛 髮
h
i

eitjengf

.cl

Bt ⼆ I d Bt

dē 是 iédttid Bt

dhēf édt

X.DE
ti
IS Dsds
dlt En
Btēk

_jdt

Ì Btē_ē to

漸
 9. (The Ornstein-Uhlenbeck process) The Ornstein-Uhlenbeck process, {Xt , t 0}, is the
unique solution to Langevin’s equation,

dXt = ↵Xt dt + dBt , X0 = x.

Show that Z t
↵t ↵t
Xt = e x+e e↵s dBs ,
0
and use this expression to calculate the mean and variance of Xt .
(This equation was originally introduced as a simple idealized model for the velocity of
a particle suspended in a liquid. In finance it is a special case of the Vasicek model of
interest rates.)
MLEC
10. The Cox-Ingersoll-Ross model of interest rates assumes that the interest rate, r, is not
deterministic, but satisfies the stochastic di↵erential equation ⼀
p
drt = (↵ rt )dt +
p
rt dBt . -
0
(This process is known as a squared Bessel process.) Taking ↵ = 0 above, find the
中 與
p
stochastic di↵erential equation followed by { rt , t 0}.
(k) (k) (k)
11. Let dXt = Xt Vt dBt , where k = 1, 2. Using Ito’s lemma, it can be shown that this
is equivalent to
Z t
(k) (k) 1 Z t (k) 2 (k) 1 (k)
Xt = X0 exp{ Vs(k) dBs V ds} =: X0 exp{Zs(k) hZ is }
0 2 0 s 2
(1) (2)
Is Yt =: Xt Xt a martingale? If not, how do you turn it into one?

12. For a continuous semimartingale Z, denote E(Zt ) as the stochastic exponential of Z, i.e.,
E(Z) satisfies
dE(Zt ) = E(Zt )dZt , and E(Z0 ) = 1.
For any two continuous semimartingales X and Y , show that

(a) E(Xt )E(Yt ) = E (Xt + Yt + hX, Y it ) .

1
(b) (E(Xt )) = E ( Xt + hXit ) .

109