100% found this document useful (1 vote)
194 views105 pages

Stochastic Differential Equations

1) Stochastic differential equations (SDEs) are used to model situations where the local behavior is known but not the exact governing function, like the price of a fluctuating asset. 2) Ordinary differential equations can model a situation accurately if the initial condition and infinitesimal behavior are known. 3) SDEs add a stochastic term to account for unpredictable errors, with the magnitude of the error depending on time elapsed and current price/state. This leads to a general form of SDEs.

Uploaded by

ali.sarir
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
194 views105 pages

Stochastic Differential Equations

1) Stochastic differential equations (SDEs) are used to model situations where the local behavior is known but not the exact governing function, like the price of a fluctuating asset. 2) Ordinary differential equations can model a situation accurately if the initial condition and infinitesimal behavior are known. 3) SDEs add a stochastic term to account for unpredictable errors, with the magnitude of the error depending on time elapsed and current price/state. This leads to a general form of SDEs.

Uploaded by

ali.sarir
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 105

EDS

Introduction

Ito’s lemma

Solutions to
SDEs Stochastic di¤erential equation (SDE) and
Ito’s lemma
80-646-08
Stochastic Calculus I

Geneviève Gauthier

HEC Montréal
EDS

Ordinary di¤erential eq. I


Introduction
Equations
Ordinary
di¤erential
equations
Introduction to
When we model some situations, we don’t know a priori
SDE
which function to use, since we only know the local
Ito’s lemma

Solutions to
behavior of our system.
SDEs
For example, assume that f (t ) represents a commodity
price at time t. We write

f (t + ∆t ) f (t ) = µ ∆t f (t ) where µ 0

in order to mean that the variation f (t + ∆t ) f (t ) of


the commodity price over a time period is is proportional
to the length ∆t of the time period considered as well as
the commodity price f (t ) at the start of the period, i.e. µ
∆t f (t ) , µ being a constant.
EDS

Ordinary di¤erential eq. II


Introduction
Equations
Ordinary
di¤erential
equations
Introduction to
By dividing both sides of the equality by ∆t, we obtain
SDE

Ito’s lemma f (t + ∆t ) f (t )
= µ f (t ) .
Solutions to
SDEs ∆t
Let’s now take the limit when ∆t tends to zero:
d f (t + ∆t ) f (t )
f (t ) = lim
dt ∆t !0 ∆t
= lim µ f (t )
∆t !0
= µ f (t ) .
EDS

Ordinary di¤erential eq. III


Introduction
Equations
Ordinary
Recall that we consider the equation
di¤erential
equations
Introduction to d
SDE
f (t ) = µ f (t ) .
Ito’s lemma dt
Solutions to
SDEs The notation commonly used for di¤erential eqations
allows us to rewrite the equation above as:

d f (t ) = µ f (t ) dt. (1)

Note that, technically speaking, the object d f (t ) is not


well-de…ned. The latter equation is only a notation to
express that ”the derivative of the function is proportional
d
to the function itself”, i.e. dt f (t ) = µ f (t ) .
The unknown, in that equation, is the function f . We are
looking for the functions that satisfy that equality.
EDS

Ordinary di¤erential eq. IV


Introduction
Equations
Ordinary
di¤erential
equations
Recall that we are studying the equation
Introduction to
SDE

Ito’s lemma d f (t ) = µ f (t ) dt. (1)


Solutions to
SDEs
It is possible to show that the function de…ned for all
t 2 R as

f (t ) = ce µt , where c is any constant, (2)

satis…es equation (1).


Indeed, in such a case,
d d
f (t ) = ce µt = µce µt = µf (t ) .
dt dt
EDS

Ordinary di¤erential eq. V


Introduction
Equations
Ordinary
di¤erential
equations The initial condition helps determine the constant c. We
Introduction to
SDE
know commodity price f0 today. As a consequence,
Ito’s lemma

Solutions to 0
SDEs f0 = f (0) = ce µ = c,

which yields that the commodity price at time t is

f (t ) = f0 e µt .

In this example, knowing the in…nitesimal behavior of the


commodity price (d f (t ) = µ f (t ) dt ) and the initial
price f0 is su¢ cient to determine accurately the price at
any time.
EDS

Ordinary di¤erential eq. VI


Introduction
Equations
Ordinary
di¤erential
equations
Introduction to
Recall that we are considering the equation
SDE

Ito’s lemma
d f (t ) = µ f (t ) dt. (1)
Solutions to
SDEs
equation (1) is an example of ordinary di¤erential equation
and it behaves very charmingly since there exists at least
one function f which satis…es equation (1) and, in
addition, it is possible to show that this function is
necessarily of the form described in (2) :

f (t ) = f0 e µt .
EDS

Ordinary di¤erential eq. VII


Introduction
Equations
Ordinary
di¤erential
equations There exist ordinary di¤erential equations much less nice.
Introduction to
SDE For example,
Ito’s lemma f (t )
Solutions to
d f (t ) = 2 dt. (3)
SDEs t
The solution to that equation has form
1
f (t ) = ce t where c is a constant.

We must now specify c by using the initial condition.


But f (0) = 0 whatever c, which implies that (3)
possesses an in…nity of solutions when f (0) = 0 and
possesses none when f (0) 6= 0.
EDS

Introduction I
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to
SDEs

Now assume that the stochastic process S = fSt : t 0g


represents the evolution of a risky asset price.
We don’t know, in general, the law that governs such a
process, but we may have an idea of its local behavior.
EDS

Introduction II
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma
For example, over a short time interval of length ∆t, it is
Solutions to
SDEs possible that such a price tends to vary proportionally to
the period length and the asset price at the beginning of
the period. We write, to begin with,

St +∆t St = µ St ∆t.

If, in general, prices increase, then µ is a positive constant


and if the prices tend to decrease, then µ is negative.
EDS

Introduction III
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma St +∆t St = µ St ∆t


Solutions to
SDEs There is however a problem with the latter equation: we
are not certain that the price varies proportionally to the
period length and the asset price, we only claim that it has
a tendency to do so.
Therefore, an unpredictable error needs to be incorporated
into our equation.
We can however control the magnitude of such a random
error.
EDS

Introduction IV
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma For example, we can assume that it depends on the asset
Solutions to
SDEs
price at the beginning of the period.
Indeed, we observe that, the higher the price, the more the
risky asset price can diverge from the trend.
Moreover, the random error must also depend on the
length of the time interval considered: the longer the
interval, the greater the chance that the price diverges
from the trend.
That is why we add a stochastic term to our initial
equation.
EDS

Introduction V
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE The stochastic term to be added to our initial equation
Ito’s lemma leads us to the equation
Solutions to p
St +∆t St = µ St ∆t + σ St ∆t ξ t
SDEs
(4)

where
σ is a positive constant and
ξ t is a random variable with distribution N (0, 1)
independent from fSu : 0 u t g.
The latter condition is important, since we must not be
able to predict the error ξ t from observing the behavior of
the risky asset price prior to date t.
EDS

Introduction VI
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to
SDEs
Such an equation is random and must be satis…ed by
”almost” every ω, which is to say that
hn p oi
Pr ω 2 Ω : S t +∆t (ω ) S t (ω ) = µ S t (ω ) ∆t + σ S t (ω ) ∆t ξ t (ω )

vaut un.
EDS

Introduction VII
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations Withprespect to the magnitude of the random error, note
Introduction to
SDE that ∆t ξ t is N (0, ∆t )-distributed. Moreover,
Ito’s lemma h p i p
Solutions to E σSt ∆tξ t j σ fSu : u 2 f0, ∆t, ..., t gg = σSt ∆tE [ξ t ]
SDEs
= 0,

p 2 h i
E σSt ∆tξ t j σ fSu : u 2 f0, ∆t, ..., t gg = σ2 St2 ∆tE ξ 2t

= σ2 St2 ∆t,

which implies that the


p conditional standard deviation of
the error term is σSt ∆t.
Thus the longer the time interval ∆t or the higher the
security price St , the greater the standard deviation of the
random error.
EDS

Introduction VIII
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to
SDEs

This implies that the values that the random error may
take are more dispersed around its expectation (which is
zero).
EDS

Introduction IX
Introduction Stochastic di¤erential equations
Equations
Ordinary
Recall that
di¤erential p
St = µ St ∆t + σ St ∆t ξ t .
equations
Introduction to
SDE
St +∆t (4)
Ito’s lemma
Let’s rewrite equation (4) for the next period:
Solutions to
SDEs p
St +2∆t St +∆t = µ St +∆t ∆t + σ St +∆t ∆t ξ t +∆t .
If we don’t want to be able to predict the error ξ t +∆t , the
latter must be independent from
fSu : u 2 f0, ∆t, ..., t + ∆t gg.
For that reason, we introduce the Brownian motion since
it is a Gaussian process, the increments of which are
mutually independent:
St +∆t St = µ St ∆t + σ St (Wt +∆t Wt ) . (5)
Note
p that the law of Wt +∆t Wt is the same as the law
of ∆tξ t : both are N (0, ∆t )-distributed.
EDS

Introduction X
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to Let W = fWt : t 0g be a Brownian motion built on a


SDEs
…ltered probability space (Ω, F , F, P) such that the
…ltration F is the one generated by the Brownian motion,
plus it includes all zero-probability events, i.e. for all t 0,

Ft = σ (N and Ws : 0 s t) .
EDS

Introduction XI
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma The risky asset price today (t = 0) is known with


Solutions to certainty. S0 is thus (?, Ω) measurable and therefore
SDEs
F0 measurable. Let’s go back to equation (5) and let’s
set t = 0.
S∆t = S0 + µ S0 ∆t + σ S0 (W ∆t W0 )
|{z} | {z } |{z} | {z }
F0 measurable F0 measurable F0 measurable F∆t measurable
indpendent from F0
| {z }
F∆t measurable

We observe that S∆t is F∆t measurable.


EDS

Introduction XII
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
We can show by induction that Sn ∆t is
equations
Introduction to Fn ∆t measurable, for any positive integer n.
SDE

Ito’s lemma
Indeed, let’s assume there exists k 2 f0, 1, 2, ...g such that
Solutions to
Sk ∆t is Fk ∆t measurable. Then
SDEs
S (k +1 ) ∆t

= S k ∆t + µ S k ∆t ∆t + σ S k ∆t W (k +1 ) ∆t W k ∆t
| {z } | {z } | {z } | {z }
Fk ∆t measurable Fk ∆t measurable Fk ∆t measurable F (k +1 ) ∆t measurable
independent from Fk ∆t
| {z }
F (k +1 ) ∆t measurable

which implies that S(k +1 ) ∆t is F(k +1 ) ∆t measurable.


Our process S lives on the same …ltered probability space
as the Brownian motion W that we have used to build
that process.
EDS

Introduction XIII
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma Let’s go back to equation (5)


Solutions to
SDEs
St +∆t St = µ St ∆t + σ St (Wt +∆t Wt ) .

When time intervals of length ∆t become of in…nitesimal


length, we obtain an equation of the type

dSt = µ St dt + σ St dWt .
EDS

Introduction XIV
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma
The latter equation is an example of stochastic di¤erential
Solutions to equation and it should raise a few questions:
SDEs
1 the term σ St dWt is not well de…ned, particularly if we
recall that the Brownian motion paths are nowhere
di¤erentiable!
2 does a solution to that equation exist? and
3 if that solution exists, is it unique and how can it be found?
Note that the solution to a stochastic di¤erential equation
is not, as is the case for ordinary di¤erential equations, a
function, but rather a stochastic process.
EDS

Introduction XV
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to
In order to answer those questions, let’s consider a
SDEs
stochastic di¤erential equation in a more general form

dX (t ) = b (X (t ) , t ) dt + a (X (t ) , t ) dW (t ) . (6)
| {z } | {z }
drift coe¢ cient di¤usion coe¢ cient

where the functions a : R [0, ∞) ! R and


b : R [0, ∞) ! R are measurable functions.
EDS

Introduction XVI
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma What do we mean by a (X (t ) , t ) dW (t )? We haven’t


Solutions to
de…ned that term. Actually, Equation (6) is the di¤erential
SDEs form of the integral equation:
Z t Z t
X (t ) = X (0 ) + b (X (u ) , u ) du + a (X (u ) , u ) dW (u )
0 0

and we now know the meaning of the term


Z t
a (X (u ) , u ) dW (u ) .
0
EDS

Introduction XVII
Introduction Stochastic di¤erential equations
Equations
Ordinary
di¤erential
equations
Introduction to
SDE

Ito’s lemma

Solutions to
SDEs There doesn’t always exist a solution to that equation and
we will see a few results which give the conditions that b
and a functions must meet for a solution to exist.
To answer questions (2) and (3), we need a few additional
tools.
EDS

Fundamental theorem of calculus I


Introduction

Ito’s lemma
Fundamental th. Ito’s lemma is the stochastic equivalent of the
of calculus
Itô (version 1) fundamental theorem of calculus. It will allow us to
Itô (version 2)
Example determine the stochastic di¤erential equation satis…ed by
Ito process
Quadratic
variation
some given stochastic processes.
Itô (version 3)
Example
Itô (version 4)
Taylor
Theorem
Solution to an
SDE The fundamental theorem of calculus stipulates that, if
Quadratic

dx : R ! R represents the derivative of the function


covariation df
Multiplication
rule
Example f : R ! R, then
Multidimensional
Ito Z b
Ito process df
Quadratic
covariation
f (b ) f (a ) = (x ) dx.
Ito’s lemma a dx
Example

Solutions to
SDEs
Richard R. Goldberg, Theorem 7.8A, page 205.
EDS

Fundamental theorem of calculus II


Introduction

Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example. if f (x ) = x 2 , a = 0 and b = t, then
Example
Ito process Z t
t2 =
Quadratic
variation
Itô (version 3)
2x dx.
Example
0
Itô (version 4)
Taylor
Solution to an
Is such a rule still valid in the context of stochastic
SDE
Quadratic calculus? Is Z t
covariation
Multiplication
rule Wt2 = 2Ws dWs ? (7)
Example 0
Multidimensional
Ito
Ito process We have observed, when constructing the stochastic
Quadratic
covariation
Ito’s lemma
integral, that the paths of the process
Example

Solutions to
SDEs
EDS

Fundamental theorem of calculus


Introduction III
Ito’s lemma
Fundamental th.
of calculus nR o
Itô (version 1) t
Itô (version 2) Ws dWs : 0
0
t T could be negative at some
Example
Ito process
Quadratic
times
variation
Itô (version 3)
Example 0,6
Itô (version 4) 0,4
Taylor 0,2
Solution to an 0
SDE
Quadratic -0,2 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

covariation -0,4
Multiplication -0,6
rule
Example -0,8
Multidimensional -1
Ito -1,2
Ito process
-1,4
Quadratic
covariation t
Ito’s lemma
Example

Solutions to
SDEs
EDS

Fundamental theorem of calculus


Introduction IV
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Recall: is Z t
Wt2
Example
Itô (version 4) = 2Ws dWs ? (7)
Taylor 0
Solution to an
SDE
Quadratic
covariation
But the left side of equality (7) is necessarily non-negative,
Multiplication
rule
while the right side can take negative values. There is a
Example
Multidimensional
contradiction and we conclude that equation (7) is false.
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

First version I
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example
Itô (version 4)
Taylor There exists, in the framework of stochastic calculus, an
Solution to an
SDE equivalent of the fundamental theorem of calculus that
Quadratic
covariation
Multiplication
was established by K. Itô. Note that a term is added.
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

First version II
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus Theorem
Itô (version 1)
Itô (version 2)
Example
Ito’s lemma (…rst version). Let W be a Brownian motion
Ito process
Quadratic
built on the …ltered probability space (Ω, F , F, P) and let
variation
Itô (version 3) f : R ! Rbe a function, the …rst two derivatives of which exist
Example
Itô (version 4) and are continuous. Then 80 t T ,
Taylor
Solution to an Z t Z t 2
SDE
P a.s. df 1 d f
Quadratic
covariation f ( Wt ) f ( W0 ) = (Ws ) dWs + (Ws ) ds.
Multiplication 0 dw 2 0 dw 2
rule
Example
(8)
Multidimensional
Ito In its di¤erential form, equation (8) is written
Ito process
Quadratic
covariation
Ito’s lemma df 1 d 2f
Example df (Wt ) = (Wt ) dWt + (Wt ) dt.
Solutions to
dw 2 dw 2
SDEs
EDS

First version III


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
For example, if f (x ) = x 2 , then
Itô (version 1)

f (Wt ) = Wt2 and f (W0 ) = W02 = 0


Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3) and
Example
df d 2f
Itô (version 4)
(Ws ) = 2Ws and (Ws ) = 2.
Taylor
Solution to an dw dw 2
SDE
Quadratic
covariation
By substituting into Ito’s equation, we obtain
Multiplication
rule Z t Z t Z t
Example P a.s.
Multidimensional
Ito
Wt2 = 2 Ws dWs + 1ds = 2 Ws dWs + t
Ito process 0 0 0
Quadratic
covariation
Ito’s lemma which implies
Example Z t
Wt2 t
Solutions to Ws dWs = .
SDEs
0 2
EDS

Idea of the proof I


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) Let’s assume a division of time 0 = t0 < t1 < ... < tn = t.
Itô (version 2)
Example One can assume that ti ti 1 = t/n.
Ito process
Quadratic
variation
By expanding f as a Taylor series about point x0 , one …nds
Itô (version 3)
Example
1
Itô (version 4)
Taylor f (x ) f (x0 ) = f 0 (x0 ) (x x0 ) + f 00 (ξ ) (x x0 )2
Solution to an 2
SDE
Quadratic
covariation
Multiplication
where min (x0 , x ) ξ max (x0 , x ) .
rule
Example
By applying that result to random points
Multidimensional
Ito
(x = Wti and x0 = Wti 1 ),
Ito process
Quadratic 1
covariation f (W t i ) f (W t i 1 ) = f 0 (W t i 1 ) (W t i Wti 1 ) + f 00 (ξ i ) (Wti W t i 1 )2
Ito’s lemma 2
Example

Solutions to where min (Wti 1 , Wti ) ξi max (Wti 1 , Wti ) .


SDEs
EDS

Idea of the proof II


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
f ( Wt ) f ( W0 )
Ito process n
∑ ( f ( Wt )
Quadratic
variation
Itô (version 3)
= i f (Wti 1 ))
Example i =1
Itô (version 4)
n
Taylor 1 n 00
Solution to an
SDE
= ∑ f 0 ( Wt i 1 ) ( Wt i Wt i 1 ) +
2 i∑
f ( ξ i ) ( Wt i Wt i 1 ) 2
Quadratic
covariation
i =1 =1
Multiplication

By taking the limit when n ! ∞ on both sides,


rule
Example
Multidimensional
Ito
Ito process n
∑ f 0 ( Wt
Quadratic
covariation f ( Wt ) f ( W0 ) = lim i 1 ) ( Wt i Wt i 1 )
Ito’s lemma n !∞
Example i =1
n
Solutions to 1
SDEs + lim ∑ f 00 (ξ i ) (Wti Wt i 1 ) 2 .
2 n ! ∞ i =1
EDS

Idea of the proof III


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic The objective is to show that
variation
Itô (version 3)
Example n Z t
∑ f 0 ( Wt i 1 ) ( Wt i f 0 (Ws ) dWs
Itô (version 4)
Taylor lim Wt i 1 ) =
Solution to an n !∞ 0
SDE i =1
Quadratic
covariation
n Z t
Multiplication
rule
et lim
n !∞
∑ f 00 (ξ i ) (Wt i Wt i 1 ) 2 =
0
f 00 (Ws ) ds.
Example i =1
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Idea of the proof IV


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
First, Z t Z ti
of calculus
n
Itô (version 1)
Itô (version 2)
0
0
f (Ws ) dWs = ∑ f 0 (Ws ) dWs
Example i =1 t i 1
Ito process
Quadratic
variation which implies that
Itô (version 3)
Example
Itô (version 4)
Z t n
Taylor
Solution to an
SDE 0
f 0 (Ws ) dWs ∑ f 0 ( Wt i 1 ) ( Wt i Wt i 1 )
i =1
Quadratic
covariation n Z ti n Z ti
∑ 0
∑ f 0 (Wti 1 ) dWt
Multiplication
rule = f (Ws ) dWs
Example
Multidimensional i =1 t i 1 i =1 t i 1
Ito
Ito process
n Z ti
Quadratic
covariation
= ∑ f 0 ( Ws ) f 0 (Wti 1 ) dWs
Ito’s lemma i =1 t i 1
Example
n !∞
Solutions to ! 0.
SDEs
EDS

Idea of the proof V


Introduction Ito’s lemma
Ito’s lemma
Fundamental th. Second,
of calculus
Itô (version 1) Z t n
Itô (version 2)
Example
0
f 00 (Ws ) ds ∑ f 00 (ξ i ) (Wt i W t i 1 )2
Ito process i =1
Quadratic
variation n Z ti n
Itô (version 3)
Example
= ∑ f 00 (Ws ) ds ∑ f 00 (ξ i ) (Wt i W t i 1 )2
Itô (version 4)
i =1 t i 1 i =1
Taylor n Z ti
Solution to an
SDE = ∑ f 00 (Ws ) f 00 (ξ i ) ds
Quadratic i =1 t i 1
covariation
Multiplication n Z ti n
rule
Example
+∑ f 00 (ξ i ) ds ∑ f 00 (ξ i ) (Wt i W t i 1 )2
Multidimensional i =1 t i 1 i =1
Ito
n Z ti

Ito process
Quadratic = f 00 (Ws ) f 00 (ξ i ) ds
covariation
Ito’s lemma
i =1 t i 1

Example n
Solutions to ∑ f 00 (ξ i ) (W t i W t i 1 )2 (ti ti 1)
SDEs i =1
EDS

Idea of the proof VI


Introduction Ito’s lemma
Ito’s lemma R ti
Fundamental th. The …rst term ∑ni=1 ti 1 (f 00 (Ws ) f 00 (ξ i )) ds converges
of calculus
Itô (version 1)
Itô (version 2)
to zero since min (Wti 1 , Wti ) ξ i max (Wti 1 , Wti )
Example
Ito process
implies that ξ i Ws ! 0. f 00 being continuous, we
Quadratic
variation deduce that f 00 (ξ i ) f 00 (Ws ) ! 0.
Itô (version 3)
Example
Itô (version 4)
The second term
2
Taylor
Solution to an
∑ni=1 f 00 (ξ i ) (Wti Wti 1 ) (ti ti 1 ) tends also to
SDE
Quadratic 0 since
covariation
Multiplication
rule
h i
Example E (Wti Wti 1 )2 (ti ti 1 ) = 0
Multidimensional
Ito h i
Var (Wti Wti 1 )2 (ti ti 1 ) = (ti ti 1 ) ! 0.
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
Such a proof is only roughly sketched since we have
SDEs
omitted to specify some technical details. (ref. R. Durrett)
EDS

Second version I
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example
Itô (version 4)
Taylor The proof given by K. Itô applies to much more complex
Solution to an
SDE situations that the one described above. Here is a …rst
Quadratic
covariation
Multiplication
generalization:
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Second version II
Introduction Ito’s lemma
Ito’s lemma Theorem
Fundamental th.
of calculus
Itô (version 1)
Ito’s lemma (second version). Let W be a Brownian motion
Itô (version 2)
Example
built on the …ltered probability space (Ω, F , F, P) and let
Ito process
Quadratic
f : R [0, ∞) ! R be a function, the …rst and second partial
variation
Itô (version 3) derivatives of which exist and are continuous. Then
Example
Itô (version 4) 80 t T ,
Taylor
Solution to an
SDE
Quadratic f ( Wt , t ) f ( W 0 , 0 )
covariation
Z t
Multiplication
rule ∂f 1 ∂2 f
Example = ( Ws , s ) + ( Ws , s ) ds
Multidimensional 0 ∂t 2 ∂w 2
Ito
Ito process
Z t
∂f
Quadratic
covariation + (Ws , s ) dWs
Ito’s lemma 0 ∂w
Example

Solutions to In its di¤erential form, we have


SDEs
∂f 1 ∂2 f ∂f
df (Wt , t ) = (W t , t ) + (W t , t ) dt + (Wt , t ) dWt .
∂t 2 ∂w 2 ∂w
EDS

Example I
Introduction The Black-Scholes model
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) The stochastic process S = fSt : 0 t T g represents
Itô (version 2)
Example the price evolution of a risky asset where
Ito process
Quadratic

σ2
variation
Itô (version 3)
Example St = S0 exp µ t + σWt ,
Itô (version 4) 2
Taylor
Solution to an
SDE
Quadratic µ and σ being constants, and W represents a standard
covariation
Multiplication
rule
Brownian motion.
Example
Multidimensional Since Wt is N (0, t )-distributed,
Ito
Ito process
Quadratic
covariation σ2 σ2
Ito’s lemma µ t + σWt follows a distribution N µ t, σ2
Example 2 2
Solutions to
SDEs
and St is lognormally distributed.
EDS

Example II
Introduction The Black-Scholes model
Ito’s lemma An intermission: the moments of a lognormally-distributed random
Fundamental th.
of calculus variable
Itô (version 1)
Itô (version 2)
If Z represents a standard normal random variable and if a and b are
Example constants then
Ito process
a2
Quadratic
variation
E [exp [b + aZ ]] = exp b + .
Itô (version 3)
2
Example
Itô (version 4) So, if S (0) is independent from the Brownian motion,
Taylor
Solution to an
SDE σ2
Quadratic E [S (t )] = E [S (0)] E exp µ t + σWt
covariation 2
Multiplication
rule = E [S (0)] exp [µt ]
Example
Multidimensional
Ito
Ito process h i h i σ2
Quadratic E S 2 (t ) = E S 2 (0) E exp 2 µ t + 2σWt
covariation
2
Ito’s lemma
Example
h i h i
2 2
= E S (0) exp 2µt + σ t
Solutions to
SDEs
h i h i
Var [S (t )] = E S 2 (0) exp 2µt + σ2 t E2 [S (0)] exp [2µt ] .
EDS

Example III
Introduction The Black-Scholes model
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Using Ito’s lemma, it is possible to show that the
Ito process
Quadratic
stochastic process S de…ned as
variation
Itô (version 3)
Example σ2
Itô (version 4)
St = S0 exp µ t + σWt
Taylor
Solution to an
2
SDE
Quadratic
covariation
Multiplication
is a solution to the integral equation
rule
Example
Multidimensional
Z t Z t
Ito
Ito process
St S0 = µ Su du + σ Su dWu .
Quadratic 0 0
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Example IV
Introduction The Black-Scholes model
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) Recall that
Itô (version 2)
Example
Ito process
σ2
Quadratic
variation St = S0 exp µ t + σWt .
Itô (version 3) 2
Example
Itô (version 4) h i
Taylor
σ2
Solution to an
SDE
Indeed, if f (w , t ) = S0 exp µ 2 t + σw , then
Quadratic
covariation
Multiplication f ( Wt , t ) = S t
rule
Example f ( W0 , 0 ) = S 0
Multidimensional
Ito ∂f σ2
Ito process ∂t (w , t ) = µ 2 f (w , t )
Quadratic
∂f
∂w (w , t ) = σf (w , t )
covariation
Ito’s lemma
∂2 f
(w , t ) = σ 2 f (w , t ) .
Example

Solutions to ∂w 2
SDEs
EDS

Example V
Introduction The Black-Scholes model
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Thus,
Example
Ito process St S0
Quadratic
variation
Itô (version 3)
= f (W t , t ) f (W 0 , 0 )
Z t Z t
Example
Itô (version 4) ∂f ∂f 1 ∂2 f
= (Wu , u ) dWu + (W u , u ) + (Wu , u ) du
Taylor
Solution to an
0 ∂w 0 ∂t 2 ∂w 2
SDE Z t Z t
σ2 1
f (Wu , u ) + σ2 f (Wu , u ) du
Quadratic
covariation = σf (Wu , u ) dWu + µ
Multiplication 0 0 2 2
rule Z t Z t
Example
Multidimensional = σf (Wu , u ) dWu + µf (Wu , u ) du
Ito 0 0
Ito process Z t Z t
Quadratic
covariation = σSu dWu + µSu du.
Ito’s lemma 0 0
Example

Solutions to
SDEs
EDS

Example VI
Introduction The Black-Scholes model
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
We have just proved that the stochastic process
Quadratic
variation
S = fSt : 0 t T g where
Itô (version 3)
Example
Itô (version 4) σ2
Taylor
St = S0 exp µ t + σWt
Solution to an
SDE 2
Quadratic
covariation
Multiplication
rule is a solution to the stochastic di¤erential equation
Example
Multidimensional
Ito
Ito process dSt = µSt dt + σSt dWt .
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Ito process I
Introduction

Ito’s lemma
De…nition
Fundamental th.
of calculus An Ito process is de…ned to be a process
X = fXt : 0 t T g taking its values in R such that:
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
Z t Z t
variation
Itô (version 3)
Xt X0 + Ks ds + Hs dWs (9)
Example 0 0
Itô (version 4)
Taylor
Solution to an
SDE
where K = fKt : 0 t T g and H = fHt : 0 t T g are
Quadratic processes
hR adapted toi the …ltrationh fFt g, satisfying i
covariation
T RT
P 0 jKs j ds < ∞ = 1 and P 0 jHs j2 ds < ∞ = 1.
Multiplication
rule
Example
Multidimensional
Ito
Ito process
Quadratic
Written in its di¤erential form, the Ito process becomes
covariation
Ito’s lemma
Example
dXt = Kt dt + Ht dWt .
Solutions to
SDEs

Damien Lamberton and Bernard Lapeyre, page 53.


EDS

Ito process II
Introduction
Remark. First, note that, if
Ito’s lemma Z T
Fundamental th.
of calculus EP jHs j2 ds < ∞, (10)
Itô (version 1) 0
Itô (version 2)
Example
Ito process then the process M = fMt : 0 t T g where
Quadratic
variation
Itô (version 3)
Z t
Example
Itô (version 4)
Mt = Hs dWs
Taylor 0
Solution to an
SDE
Quadratic is a (fFt g , P) martingale.
covariation
Multiplication
rule
Remark. As a consequence, if the drift coe¢ cient is zero,
Example
Multidimensional
i.e. Kt = 0 for all 0 t T , then the Ito process X
Ito
Ito process where Z t
Quadratic
covariation
Ito’s lemma
Xt X0 + Hs dWs
Example 0
Solutions to
SDEs
is a (fFt g , P) martingale if and only if the condition
(10) is satis…ed.
Daniel Revuz and Marc Yor, page 129, proposition 1.23.
EDS

Quadratic variation I
Introduction

Ito’s lemma
Fundamental th.
Recall that
of calculus
Itô (version 1) Z t Z t
Itô (version 2)
Example Xt X0 + Ks ds + Hs dWs (11)
Ito process 0 0
Quadratic
variation
Itô (version 3)
Example
Itô (version 4)
Taylor
De…nition
Solution to an
SDE Let X be the Ito process de…ned by equation (11), then the
Quadratic
covariation quadratic variation procesof X ,
Multiplication
rule
Example
Multidimensional
Ito
hX i = fhX it : 0 t Tg,
Ito process
Quadratic
covariation
Ito’s lemma
is de…ned to be Z t
Example
hX it = Hs2 ds.
Solutions to 0
SDEs
EDS

Quadratic variation II
Introduction

Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Example. The Brownian motion is an Ito process since
Ito process
Quadratic
variation
Z t
Itô (version 3)
Example
Wt = 1dWs .
Itô (version 4) 0
Taylor
Solution to an
SDE As a consequence,
Quadratic
covariation
Multiplication Z t
rule
Example hW it = 12 ds = t.
Multidimensional
Ito 0
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Third version I
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Theorem
Example
Ito process Ito’s lemma (third version). Let X be an Ito process and let
f : R ! R be a function, the …rst two derivatives of which
Quadratic
variation
Itô (version 3)
Example
Itô (version 4)
exist and are continuous. Then 80 t T ,
Taylor
Solution to an
SDE f (Xt ) f (X0 ) (12)
Quadratic
Z t Z
1 t d 2f
covariation
Multiplication P a.s. df
rule = (Xs ) dXs + (Xs ) d hX is
Example
0 dx 2 0 dx 2
Multidimensional Z t
Ito
P a.s. df
Ito process
Quadratic
= (Xs ) Hs dWs
covariation 0 dx
Ito’s lemma Z t
Example df 1 d 2f
+ (Xs ) Ks + (Xs ) Hs2 ds,
Solutions to
SDEs 0 dx 2 dx 2
EDS

Third version II
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example In its di¤erential form, we …nd
Itô (version 4)
Taylor
df df 1 d 2f
Solution to an
SDE df (Xt ) = (Xt ) Ht dWt + (X t ) K t + (Xt ) Ht2 dt.
Quadratic dx dx 2 dx 2
covariation
Multiplication
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Third version III


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic Justi…cation attempt!
variation
Itô (version 3) Itô’s lemma for Ito processes is
Example
Itô (version 4)
Taylor Z t Z t 2
Solution to an P a.s. df 1 d f
SDE f (Xt ) f (X0 ) = (Xs ) dXs + (Xs ) d hX is .
Quadratic
covariation
0 dx 2 0 dx 2
Multiplication
rule
Example We recognize equation (8) as a particular case of the equality
Multidimensional
Ito
Ito process
above since hW it = t.
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Third version IV
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2) But
Example
Ito process
Quadratic f (X t ) f (X 0 )
variation
Z t Z t 2
Itô (version 3)
df 1 d f
Example = (Xs ) dXs + (X s ) d hX is
Itô (version 4) 0 dx 2 0 dx 2
Taylor Z t Z t 2
Solution to an df 1 d f
SDE = (Xs ) (Ks ds + Hs dWs ) + 2 ( s)
X Hs2 ds
Quadratic
covariation
0 dx 0 dx 2
Z t Z t Z
Multiplication
df df 1 t d 2f
rule
= (Xs ) Ks ds + (Xs ) Hs dWs + (Xs ) Hs2 ds
Example
Multidimensional
0 dx 0 dx 2 0 dx 2
Z t Z t
Ito
df df 1 d 2f
Ito process
= (Xs ) Hs dWs + (X s ) K s + (Xs ) Hs2 ds.
Quadratic
covariation 0 dx 0 dx 2 dx 2
Ito’s lemma
Example

Solutions to
SDEs
EDS

Third version V
Introduction Ito’s lemma
Ito’s lemma The preceding calculation is valid provided the following
Fundamental th.
of calculus exercise has been veri…ed:
Itô (version 1)
Itô (version 2)
Example
Exercise. If
Ito process
Quadratic
Z t Z t
variation
Itô (version 3) Yt = Ks ds + Hs dWs
Example 0 0
Itô (version 4)
Taylor
Solution to an
SDE
where W is a (Ω, F , fFt : t 0g , P) Brownian motion,
Quadratic
covariation then
Multiplication
rule Z t Z t Z t
Example
Multidimensional
Ito
Xs dYs = Xs Ks ds + Xs Hs dWs
Ito process 0 0 0
Quadratic
covariation
Ito’s lemma and Z t Z t
Example

Solutions to Xs d hY is = Xs Hs2 ds
SDEs 0 0
where X is a predictable process.
EDS

Example I
Introduction Return
Ito’s lemma
Fundamental th. Let’s assume that the evolution S = fSt : 0 t T g of
of calculus
Itô (version 1) a risky asset satis…es the stochastic di¤erential equation
Itô (version 2)
Example
Ito process
Quadratic dSt = µSt dt + σSt dWt .
variation
Itô (version 3)
Example
Itô (version 4) The return Rt on such a risky asset, when accumulation is
Taylor
Solution to an continuous, is de…ned as
SDE
Quadratic
covariation St
Multiplication
rule Rt = ln .
Example S0
Multidimensional
Ito
Ito process
Quadratic
Using Ito’s lemma, we can show that R satis…es the
covariation
Ito’s lemma
stochastic di¤erential equation
Example

Solutions to σ2
SDEs dRt = µ dt + σdWt . (13)
2
EDS

Example II
Introduction Return
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) Intermission: Fubini’s theorem. What follows is a
Itô (version 2)
Example particular case of Fubini’s theorem.
Ito process
Quadratic
variation
Itô (version 3)
Example
Theorem
Itô (version 4)
Taylor If Y = fYt : 0 t T g is an adapted stochastic process such
Solution to an
SDE
Quadratic
that, for all 0 t T , Yt 0 then
covariation
Multiplication Z t Z t
EP EP [Ys ] ds,
rule
Example
Multidimensional
Ys ds =
Ito 0 0
Ito process
Quadratic
covariation
Ito’s lemma
i.e. one can exchange integral and expectation.
Example

Solutions to Donald L.Cohn, page 159, proposition 5.2.1.


SDEs
EDS

Example III
Introduction Return
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) Recall that
Itô (version 2)
Example dSt = µSt dt + σSt dWt . (14)
Ito process
Quadratic
variation
Itô (version 3)
First, let’s verify that S truly is an Ito process such that
Example
Itô (version 4)
Ht = σSt and Kt = µSt .
Taylor
Solution to an Indeed, since we have already shown that
SDE
Quadratic
covariation
Multiplication σ2
rule St = S0 exp µ t + σWt (Ft measurable)
Example
Multidimensional
2
Ito
Ito process
Quadratic
covariation
is a solution to the equation (14), the processes
Ito’s lemma
Example
K = fKt = µSt : 0 t T g and
Solutions to H = fHt = σSt : 0 t T g are fFt g adapted.
SDEs
EDS

Example IV
Introduction Return
Ito’s lemma Moreover, since St is a non-negative random variable,
R we
can use Fubini’s theorem and exchangethe EP and
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
operators, which yields
Ito process
Quadratic Z T Z T
EP = j µ j EP
variation
Itô (version 3)
Example
jKu j du Su du
Itô (version 4)
0 0
Taylor Z T
Solution to an
SDE = jµj EP [Su ] du
Quadratic 0
covariation
Multiplication
Z T
EP [S0 ] exp [µu ] du
rule
Example = jµj
Multidimensional 0
Ito
Ito process
Z T
P
Quadratic
covariation = j µ j E [ S0 ] exp [µu ] du
Ito’s lemma 0
Example
eT µ 1
Solutions to
SDEs
= j µ j EP [ S 0 ] <∞
µ
EDS

Example V
Introduction Return
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1) Z T Z T
EP Hu2 du = σ2 EP Su2 du
Itô (version 2)
Example
Ito process 0 0
Quadratic
variation Z T
Itô (version 3)
Example = σ 2
EP Su2 du
Itô (version 4) 0
Taylor Z T
EP S02 exp 2µu + σ2 u du
Solution to an
SDE
Quadratic
= σ2
covariation 0
Multiplication Z T
= σ2 EP S02
rule
Example exp 2µu + σ2 u du
Multidimensional 0
Ito
Ito process T (2µ+σ2 )
e 1
= σ2 EP S02
Quadratic
covariation
Ito’s lemma
Example
2µ + σ2
Solutions to < ∞.
SDEs
EDS

Example VI
Introduction Return
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3) Hence, we can deduce that
Example
Itô (version 4)
Taylor Z T Z T
Solution to an
SDE jKs j ds < ∞ and Hs2 ds < ∞ P a.s.
Quadratic
covariation
0 0
Multiplication
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Example VII
Introduction Return
Ito’s lemma
Now, let’s apply Ito’s lemma. If f (x ) = ln xx0 then
Fundamental th.
of calculus
Itô (version 1)
df 1 d 2f 1
Itô (version 2)
(x ) = and 2 (x ) = .
Example
Ito process dx x dx x2
Quadratic
variation
Itô (version 3) Thus,
Example
Itô (version 4)
Taylor
Solution to an dRt
SDE
Quadratic
covariation = df (St )
Multiplication
rule df df 1 d 2f
Example
= (St ) Ht dWt + ( S t ) Kt + (St ) Ht2 dt
Multidimensional
Ito dx dx 2 dx 2
Ito process
Quadratic 1 1 1 1 2 2
covariation
= σSt dWt + µSt σ St dt
Ito’s lemma
Example
St St 2 St2
Solutions to σ2
SDEs = σ dWt + µ dt.
2
EDS

Example VIII
Introduction Return
Ito’s lemma
Fundamental th.
But the integral form of equation (13)
of calculus
Itô (version 1)
Itô (version 2)
σ2
Example
Ito process
dRt = µ dt + σdWt
Quadratic 2
variation
Itô (version 3)
Example
Itô (version 4)
is
Taylor
Z t Z t
Solution to an
SDE σ2
Quadratic Rt = R0 + σ dWs + µ ds
covariation
Multiplication
0 2 0
rule
Example σ2 S0
Multidimensional = σWt + µ t car R0 = ln = ln 1 = 0
Ito
Ito process
2 S0
σ2
Quadratic
covariation
Ito’s lemma N µ t; σ2 t .
Example 2
Solutions to
SDEs
In the Black-Scholes world, returns are Gaussian.
EDS

Fourth version I
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Theorem
Ito process
Quadratic
variation
Ito’s lemma (fourth version). Let X be an Ito process, i.e.
Itô (version 3)
Example
dXt = Kt dt + Ht dWt , and let f : R [0, ∞) ! R be a
Itô (version 4)
Taylor
function, the …rst and second partial derivatives of which exist
Solution to an
SDE
and are continuous. Then 80 t T ,
Quadratic
covariation
Multiplication
rule f (Xt , t ) f (X0 , 0)
Example Z t Z t
Multidimensional P a.s. ∂f ∂f
Ito
Ito process
= (Xs , s ) dXs + (Xs , s ) ds
Quadratic 0 ∂x 0 ∂t
covariation Z t 2
Ito’s lemma 1 ∂ f
Example + (Xs , s ) d hX is
Solutions to
2 0 ∂x 2
SDEs
EDS

Fourth version II
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
We can also write
variation
Itô (version 3)
Example f (Xt , t ) f (X0 , 0)
Itô (version 4)
Z t
Taylor
P a.s. ∂f
Solution to an
SDE
= (Xs , s ) Hs dWs
Quadratic 0 ∂x
covariation Z t
Multiplication ∂f ∂f 1 ∂2 f
rule + (Xs , s ) Ks + (Xs , s ) + (Xs , s ) Hs2 ds.
Example
Multidimensional
0 ∂x ∂t 2 ∂x 2
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Fourth version III


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
In the di¤erential form, we obtain
variation
Itô (version 3)
Example
Itô (version 4)
df (Xt , t )
Taylor
Solution to an ∂f
SDE = (Xt , t ) Ht dWt
Quadratic
covariation
∂X
Multiplication
∂f ∂f 1 ∂2 f
rule
Example + (Xt , t ) Kt + (Xt , t ) + (Xt , t ) Ht2 dt.
Multidimensional
Ito
∂X ∂t 2 ∂X 2
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Relation with Taylor series I


Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
The Taylor series expansion of the continuous function
Itô (version 2)
Example
f : R ! R (the …rst derivatives of which exist and are
Ito process
Quadratic
continuous) about point x0 is
variation
Itô (version 3)
Example
(x x0 )2
Itô (version 4)
Taylor f (x ) = f (x0 ) + f 0 (x0 ) (x x0 ) + f 00 (x0 )
Solution to an
SDE
2
Quadratic
covariation
Multiplication
where = means that it is an approximation.
rule
Example
Multidimensional
If the continuous function f : R ! R is twice continuously
Ito
Ito process
di¤erentiable, then Ito’s lemma is stated as follows:
Quadratic
covariation Z t Z t
Ito’s lemma
0 1
Example f (Xt ) = f (X0 ) + f (Xs ) dXs + f 00 (Xs ) d hX is
Solutions to 0 2 0
SDEs
where X is an Ito process.
EDS

Relation with Taylor series II


Introduction Ito’s lemma
Ito’s lemma If f 0 (Xs ) and f 00 (Xs ) could be respectively replaced with
Fundamental th.
of calculus f 0 (X0 ) and f 00 (X0 ) (that would be a good approximation
Itô (version 1)
Itô (version 2) if t is small), then we could write
Example
Ito process
Quadratic
variation f (Xt )
Itô (version 3) Z t Z
Example 1 t 00
Itô (version 4) = f (X0 ) + f 0 (Xs ) dXs + f (Xs ) d hX is
Taylor
0 2 0
Solution to an
Z t Z
SDE
1 t 00
Quadratic
covariation = f (X0 ) + f 0 (X0 ) dXs + f (X0 ) d hX is
Multiplication 0 2 0
rule Z t Z t
Example 1
Multidimensional
Ito
= f (X0 ) + f 0 (X0 ) dXs + f 00 (X0 ) d hX is
Ito process 0 2 0
Quadratic
1
= f (X0 ) + f 0 (X0 ) (Xt X0 ) + f 00 (X0 ) (hX it hX i0 ) .
covariation
Ito’s lemma
Example 2
Solutions to
SDEs which looks more like the Taylor series expansion. But the
latter approximation is not Ito’s lemma.
EDS

Solution to an SDE I
Introduction Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2) The solution to a stochastic di¤erential equation (SDE)
Example
Ito process consists in …nding the stochastic process(es) satisfying
Quadratic
variation
Itô (version 3)
thegiven equation.
Example
Itô (version 4) But Ito’s lemma allows to do the opposite, i.e. it helps
Taylor
Solution to an
SDE
…nd the SDE satis…ed by a given stochastic process. If it
Quadratic
covariation
turns out that the stochastic process chosen satis…es the
Multiplication
rule given SDE, then we will have found a solution to that
Example
Multidimensional equation. In that sense, Ito’s lemma provides us with
Ito
Ito process
Quadratic
solutions to SDEs.
covariation
Ito’s lemma However, it is not a systematic way to obtain solutions. It
Example

Solutions to
requires educated guesses!
SDEs
EDS

Quadratic covariation
Introduction

Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example De…nition
Ito process
Quadratic
variation
Let X and Y be two Ito process such that
Itô (version 3)
Example
Itô (version 4)
et dt + H
dXt = Kt dt + Ht dWt and dYt = K e t dWt
Taylor
Solution to an
SDE
Quadratic then the quadratic covariation process hX , Y i is de…ned for
covariation
Multiplication
rule
all t 2 [0, T ] as
Example
Multidimensional Z t
Ito
Ito process hX , Y it = e s ds.
Hs H
Quadratic
covariation 0
Ito’s lemma
Example

Solutions to
SDEs
EDS

Multiplication rule I
Introduction

Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example The multiplication rule is useful when we want to study,
Itô (version 4)
Taylor for example, the behavior of the present value of an asset
Solution to an
SDE
Quadratic
while knowing the processes for both the asset price
covariation
Multiplication evolution and the discount factor.
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Multiplication rule II
Introduction

Ito’s lemma
Fundamental th.
Theorem
of calculus
Itô (version 1)
Itô (version 2)
The multiplication rule. Let X and Y be two Ito processes
Example
Ito process
such that
Quadratic
variation
Itô (version 3) et dt + H
dXt = Kt dt + Ht dWt and dYt = K e t dWt .
Example
Itô (version 4)
Taylor
Solution to an The multiplication rule is
SDE
Quadratic
covariation
Multiplication
rule
dXt Yt
Example
Multidimensional = Xt dYt + Yt dXt + d hX , Y it
Ito
Ito process
Quadratic et dt + H
= Xt K e t dWt + Yt (Kt dt + Ht dWt ) + Ht H
e t dt
covariation
Ito’s lemma
Example
= et + Yt Kt + Ht H
Xt K e t dt + Xt H
e t + Yt Ht dWt .
Solutions to
SDEs
EDS

Example I
Introduction Present value of an asset
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Let’s assume that the stochastic process X , satisfying the
Example
Ito process
equation
Quadratic
variation
dXt = µXt dt + σXt dWt ,
Itô (version 3)
Example
Itô (version 4) is the price evolution of a risky asset.
Taylor
Solution to an
SDE
The process f βt = e rt : t 0g is the discount factor.
Quadratic
covariation Note that
Multiplication d
rule
β = re rt = r βt ,
Example
Multidimensional
dt t
Ito
Ito process
Quadratic
i.e. the process β is an Ito process satisfying the equation
covariation
Ito’s lemma
di¤erential
Example
d βt = r βt dt.
Solutions to
SDEs
EDS

Example II
Introduction Present value of an asset
Ito’s lemma
Fundamental th.
The process Y = βX represents the evolution of the
of calculus
Itô (version 1) present value of the asset. The multiplication rule yields
Itô (version 2)
Example that
Ito process
Quadratic
variation
Itô (version 3) dYt = d βt Xt
Example
Itô (version 4)
Taylor
= βt dXt + Xt d βt + d h β, X it
Solution to an
SDE = βt (µXt dt + σXt dWt ) + Xt ( r βt dt )
Quadratic
covariation
Multiplication = (µ r ) βt Xt dt + σβt Xt dWt
rule
Example
Multidimensional
= (µ r ) Yt dt + σYt dWt .
Ito
Ito process
Quadratic
covariation In its integral form, the latter equation is written
Ito’s lemma
Example Z t Z t
Solutions to
SDEs
Yt = Y0 + (µ r) Ys ds + σ Ys dWs . (15)
0 0
EDS

Introduction
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example
Itô (version 4) The phenomenon we would like to model may contain several
Taylor
Solution to an
SDE
sources of uncertainty. That is why we consider Ito-type
Quadratic
covariation processes which involve several Brownian motions.
Multiplication
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Ito process I
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Let W (1 ) , ..., W (n ) be independent Brownian motions built
Quadratic
variation on the …ltered probability space
Itô (version 3)
Example (Ω, F , fFt : 0 t T g , P).
Itô (version 4)
Taylor
Solution to an
Let K (i ) , i 2 f1, 2, ..., m g and H (i ,k ) , where
SDE
Quadratic i 2 f1, 2, ..., m g and k 2 f1, 2, ..., ng, be fFt g adapted
covariation
Multiplication processes.
rule
Example
h R T (i ) i
Multidimensional
Ito
P ∑m i =1 0 K s ds < ∞ = 1,
Ito process hR i
T (i ,k ) (j ,k )
Quadratic
covariation P 0 Hs Hs ds < ∞ = 1, i, j 2 f1, 2, ..., m g and
Ito’s lemma
Example k, k 2 f1, 2, ..., ng .
Solutions to
SDEs
EDS

Ito process II
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic De…nition
variation
Itô (version 3)
Example For all i 2 f1, 2, ..., m g, we de…ne the process X (i ) as
Itô (version 4)
Taylor
Solution to an
Z t n Z t

SDE (i ) (i ) (i ) (i ,k ) (k )
Quadratic Xt = X0 + Ks ds + Hs dWs (16)
covariation 0 k =1 0
Multiplication
rule
Example
Multidimensional (1 ) (m )
Ito where X0 , ..., X0 are F0 measurable random variables.
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Quadratic covariation I
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
De…nition
Example
Ito process The
D quadratic E covariation
nD process
E o
Quadratic
variation (
X ,Xi ) ( j ) = ( i
X ,X ) ( j ) :0 t T de…ned as
Itô (version 3)
Example
t
Itô (version 4)
Taylor D E n Z t

Solution to an (i ) (j ) (i ,k ) (j ,k )
SDE X ,X = Hs Hs ds.
Quadratic t
covariation k =1 0
Multiplication
rule
Example
Multidimensional
Note that D E D E
Ito
Ito process X (i ) , X (j ) = X (j ) , X (i )
Quadratic
covariation
Ito’s lemma
Example
and D E D E
Solutions to X (i ) , X (i ) = X (i ) .
SDEs
EDS

Quadratic covariation II
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example. We want to calculate the quadratic covariation
Example
Ito process
of two correlated Brownian motions.
Quadratic
variation
Itô (version 3)
Let’s set B1 and B2 , two Brownian motions such that
Example
Itô (version 4)
Corr[B1 (t ) , B2 (t )] = ρ for all t > 0.
Taylor
Solution to an it is possible to writepB1 (t ) = W1 (t ) and
SDE
Quadratic
covariation
B2 (t ) = ρW1 (t ) + 1 ρ2 W2 (t ) where W1 and W2
Multiplication
rule are independent Brownian motions.
Example
Multidimensional
Ito
Since dB1 (t ) = 1dW1 (p
t ) + 0dW2 (t ) and
Ito process
Quadratic
dB2 (t ) = ρdW1 (t ) + 1 ρ2 dW2 (t ) .
covariation Rt Rt p
Ito’s lemma
Example
Then hB1 , B2 it = 0 1ρds + 0 0 1 ρ2 ds = ρt
Solutions to
SDEs
EDS

Ito’s lemma I
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
Example
Itô (version 4) Let W (1 ) , ..., W (n ) be independent Brownian motions built
Taylor
Solution to an
SDE
on the …ltered probability space
Quadratic
covariation
(Ω, F , fFt : 0 t T g , P) in the Ito processes de…ned
Multiplication
rule in equation (16).
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Ito’s lemma II
Introduction Multidimensional Ito’s lemma
Ito’s lemma
Fundamental th.
of calculus
Theorem
Itô (version 1)
Itô (version 2) Let f : Rm [0, ∞) ! R be a continuous function such that
Example 2 2f
Ito process
Quadratic
the partial derivatives ∂x , , ∂ f and ∂x∂ i ∂s
∂f ∂f
i ∂s ∂x i ∂x j
, i, j 2 f1, ..., m g
variation
Itô (version 3) exist and are continuous. Then 8t 0,
Example
Itô (version 4)
Taylor (1 ) (m ) (1 ) (m )
Solution to an f Xt , ..., Xt ,t f X0 , ..., X0 ,0
SDE
Quadratic
m Z t
covariation
P a.s. ∂f

(1 ) (m ) (i )
Multiplication
rule = Xs , ..., Xs , s dXs
Example
i =1 0 ∂xi
Multidimensional
Ito Z t
Ito process
∂f (1 ) (m )
Quadratic + Xs , ..., Xs , s ds
covariation 0 ∂t
Ito’s lemma Z t D E
Example 1 m m ∂2 f
2 i∑ ∑
(1 ) (m )
Solutions to + Xs , ..., Xs , s d X (i ) , X (j ) .
SDEs =1 j =1 0 ∂xi ∂xj s
EDS

Example I
Introduction Exchange rate
Ito’s lemma
Fundamental th. Assume that the stochastic process X , satisfying equation
of calculus
Itô (version 1)
Itô (version 2)
Example
dXt = µXt dt + σXt dWt ,
Ito process
Quadratic
variation
Itô (version 3)
represents the price evolution of a stock in British pounds.
Example
Itô (version 4) Note that d hX it = σ2 Xt2 dt.
Taylor
Solution to an
SDE Assume also that the Canadian dollar-value evolution of
Quadratic
covariation the British Pound is modeled by a process C where
Multiplication
rule
Example
Multidimensional
ft + γCt dWt .
dCt = αCt dt + βCt d W
Ito
Ito process
Quadratic
covariation If the Brownian motions W and W f are built on the same
Ito’s lemma
Example …ltered probability space and are independent, then
Solutions to
SDEs
d hC it = β2 Ct2 + γ2 Ct2 dt = β2 + γ2 Ct2 dt
d hX , C it = (σXt γCt + 0βCt ) dt = σγXt Ct dt.
EDS

Example II
Introduction Exchange rate
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Example
Ito process
Quadratic
variation
Itô (version 3)
The price evolution Y of the stock in Canadian dollars is
Example
Itô (version 4)
such that, at any time,
Taylor
Solution to an
SDE
Quadratic Yt = Xt Ct .
covariation
Multiplication
rule
Example
Multidimensional
Ito
Ito process
Quadratic
covariation
Ito’s lemma
Example

Solutions to
SDEs
EDS

Example III
Introduction Exchange rate
Ito’s lemma
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
Let a function f : R2 ! R be de…ned as
Example
Ito process
Quadratic
variation f (x, c ) = xc.
Itô (version 3)
Example
Itô (version 4)
Taylor Note that Yt = f (Xt , Ct ) , 8t 0,
Solution to an
SDE
Quadratic
that the …rst derivatives are ∂∂xf (x, c ) = c and
covariation ∂f
Multiplication
rule
∂c (x, c ) = x
Example and that the second derivatives are
Multidimensional
Ito
Ito process ∂2 f ∂2 f ∂2 f
Quadratic ( x, c ) = ( x, c ) = 0 and (x, c ) = 1.
covariation
Ito’s lemma
∂x 2 ∂c 2 ∂x ∂c
Example

Solutions to
SDEs
EDS

Example IV
Introduction Exchange rate
Ito’s lemma Ito’s lemma yields
Fundamental th.
of calculus
Itô (version 1)
Itô (version 2)
dYt = df (Xt , Ct )
Example
Ito process ∂f ∂f
Quadratic = (Xt , Ct ) dXt + (Xt , Ct ) dCt
variation
Itô (version 3)
∂x ∂c !
1 ∂2 f ∂2 f
Example
Itô (version 4) ∂x 2
( Xt , Ct ) d h X i t + ( X t , C t ) d h X , C i t
Taylor + 2f
∂x ∂c
∂2 f
Solution to an 2 + ∂c∂ ∂x (Ct , Xt ) d hC , X it + ∂c 2 ( X t , C t ) d h C i t
SDE
Quadratic
covariation = Ct dXt + Xt dCt + d hX , C it
Multiplication
rule
Example ft + γCt dWt
= Ct (µXt dt + σXt dWt ) + Xt αCt dt + βCt d W
Multidimensional
Ito
Ito process
Quadratic
+σγXt Ct dt
covariation
Ito’s lemma = (µXt Ct + αXt Ct + σγXt Ct ) dt
Example

Solutions to ft
+ (σXt Ct + γCt Xt ) dWt + βXt Ct d W
SDEs
ft .
= (µ + α + σγ) Yt dt + (σ + γ) Yt dWt + βYt d W
EDS

What is a solution? I
Introduction
There exist two types of solutions to a stochastic
Ito’s lemma
di¤erential equation: strong solutions and weak solutions.
Solutions to
SDEs
Gaussian models
Consider the stochastic di¤erential equation
Particular cases
A¢ nes Models
Lipschitz dXt = µ (Xt ) dt + σ (Xt ) dWt
Coe¢ cients

where fWt : t 0g is a
(Ω, F , fFt : t 0g , P) standard Brownian motion.

De…nition
A strong solution consists of …nding a stochastic process X
existing on the same …ltered probability space
(Ω, F , fFt : t 0g , P) as the Brownian motion and satisfying
the equation
Z t Z t
Xt = X0 + µ (Xs ) ds + σ (Xs ) dWs .
0 0
EDS

What is a solution? II
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models e
The strong solution is said to be unique if, when X and X
Lipschitz
Coe¢ cients
represent two strong solutions to the same stochastic
di¤erential equation, we have
h i
P Xt = X et , t 0 = 1.
EDS

What is a solution? III


Introduction
De…nition
Ito’s lemma
We have a weak solution if we can construct:
n o
Solutions to
SDEs 1) e Fe , Fet : t 0 , P
…ltered probability space Ω, e ,
Gaussian models n o
Particular cases
A¢ nes Models 2) a Ω,e Fe , Fet : t 0 , P e standard Brownian
Lipschitz
Coe¢ cients
n o
motion W ft : t 0 and
n o
3) a Ω,e Fe , Fet : t 0 , P e stochastic process
n o
et : t 0 such that
X
Z t Z t
et = X
X e0 + es
µ X ds + es
σ X fs
dW
0 0

It is possible to have several strong solutions to a same


equation.
EDS

What is a solution? IV
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models
Lipschitz
Coe¢ cients
If the strong solution is unique, then there will also be a
unique weak solution. By contrast, if the weak solution is
unique, there may be several strong solutions.
EDS

Gaussian models I
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models
De…nition
Lipschitz
Coe¢ cients Some stochastic processes that give rise to Gaussian models
are of the form

dXt = (µ1 (t ) Xt + µ2 (t )) dt + σ (t ) dWt (17)

where µ1 ( ), µ2 ( ) and σ ( ) deterministic functions of time


and fWt : 0 t T g is a (fFt g , P) standard Brownian
motion.
EDS

Gaussian models II
Introduction

Ito’s lemma
Theorem
Solutions to
SDEs
hR i
T σ 2 (s )
Gaussian models If P 0 Φ2 (s ) ds < ∞ = 1, then the strong solution to the
Particular cases
A¢ nes Models
Lipschitz
stochastic di¤erential equation (17) is
Coe¢ cients
Z t Z t
µ2 (s ) σ (s )
Xt = Φ (t ) X0 + ds + dWs
0 Φ (s ) 0 Φ (s )

where the function Φ ( ) is the solution to the ordinary


di¤erential equation

d Φ (t ) = µ1 (t ) Φ (t ) dt, Φ (0) = 1.

Bisière, page 117.


EDS

Gaussian models III


Introduction

Ito’s lemma

Solutions to
SDEs Proof . Let’s set
Gaussian models
Particular cases Z t
A¢ nes Models σ (s )
Lipschitz Yt dWs , 0 t T. (18)
Coe¢ cients
0 Φ (s )

The process Y is an Ito process if


Z T 2
σ (s )
P ds < ∞ = 1. (19)
0 Φ2 (s )

σ (t ) σ 2 (t )
Note that dYt = Φ (t )
dWt and d hY it = Φ2 (t )
dt.
EDS

Gaussian models IV
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models
Lipschitz h Rt i
Coe¢ cients µ2 (s )
Let’s set g (t, y ) = Φ (t ) X0 + 0 Φ (s )
ds + y .
Then g (t, Yt ) = Xt and g (0, Y0 ) = Φ (0)X0 = X0 .
| {z }
=1
EDS

Gaussian models V
Introduction

Ito’s lemma
Recall that
Solutions to Z t
µ2 (s )
SDEs
g (t, y ) = Φ (t ) X0 + ds + y .
Gaussian models
Particular cases 0 Φ (s )
A¢ nes Models
Lipschitz
Coe¢ cients Moreover,
∂g
(t, y )
∂t
Z t
∂Φ µ2 (s ) µ (t )
= (t ) X0 + ds + y + Φ (t ) 2
|∂t{z } | 0 Φ (s ) Φ (t )
{z }
= µ1 (t ) Φ (t )
= gΦ(t,y
(t )
)

= µ1 (t ) g (t, y ) + µ2 (t ) ,
∂2 g
∂g
∂y (t, y ) = Φ (t ) , and ∂y 2
(t, y ) = 0.
EDS

Gaussian models VI
Introduction

Ito’s lemma

Solutions to Using Ito’s lemma, we obtain


SDEs
Gaussian models
Particular cases
A¢ nes Models dXt
Lipschitz
Coe¢ cients
= dg (t, Yt )
∂g ∂g 1 ∂2 g
= (t, Yt ) dt + (t, Yt ) dYt + (t, Yt ) d hY it
∂t ∂y 2 ∂y 2
= (µ1 (t ) g (t, Yt ) + µ2 (t )) dt + Φ (t ) dYt
|{z}
= Φσ((tt)) dW t

= (µ1 (t ) g (t, Yt ) + µ2 (t )) dt + σ (t ) dWt .


EDS

Gaussian models VII


Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models
Lipschitz
Coe¢ cients
Why Gaussian?
It is possible to show that the joint distribution of
(Xt1 , Xt2 , ..., Xtn ) is the multivariate normal law.
EDS

Ornstein-Uhlenbeck Process I
Introduction Gaussian Models
Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases Vasicek (1977)
A¢ nes Models
Lipschitz
Coe¢ cients
The spot interest rate satis…es

drt = κ (θ rt ) dt + σ dWt . (20)

Referring to the general form,

dXt = (µ1 (t ) Xt + µ2 (t )) dt + σ (t ) dWt ,

we conclude that µ1 (t ) = κ, µ2 (t ) = κθ et σ (t ) = σ.
Bisière, pages 115-117.
EDS

Ornstein-Uhlenbeck Process II
Introduction Gaussian Models
Ito’s lemma

Solutions to
SDEs Since the ordinary di¤erential equation
Gaussian models
Particular cases
A¢ nes Models
Lipschitz d Φ (t ) = µ1 (t ) Φ (t ) dt, Φ (0) = 1
Coe¢ cients

becomes

d Φ (t ) = κΦ (t ) dt, Φ (0) = 1,

the funciton Φ ( ) is

Φ (t ) = exp ( κt ) , 0 t T.
EDS

Ornstein-Uhlenbeck Process III


Introduction Gaussian Models
Ito’s lemma

Solutions to The process Y where


SDEs
Gaussian models
Z t Z t
Particular cases σ (s )
A¢ nes Models
Yt = dWs becomes Yt = σ exp (κs ) dWs
Lipschitz
Coe¢ cients 0 Φ (s ) 0

Which is an Itô’s process because


Z T 2 Z T
σ (s )
P ds < ∞ = P σ 2
exp (2κs ) ds < ∞
0 Φ2 (s ) 0
σ2 2T κ
= P e 1 <∞

= 1.
EDS

Ornstein-Uhlenbeck Process IV
Introduction Gaussian Models
Ito’s lemma

Solutions to
The general solution
SDEs
Z t Z t
Gaussian models
µ2 (s ) σ (s )
Particular cases
Xt = Φ (t ) X0 + ds + dWs
Φ (s ) Φ (s )
A¢ nes Models
Lipschitz 0 0
Coe¢ cients

is
2 3
6 Z t Z t 7
6 7
rt = exp ( κt ) 6r0 + κθ exp (κs ) ds + σ exp (κs ) dWs 7
4 0 0 5
| {z }
=θ (exp (κt ) 1 )
Z t
= r0 exp ( κt ) + θ (1 exp ( κt )) + σ exp ( κ (t s )) dWs
0
Z t
= θ + (r0 θ ) exp ( κt ) + σ exp ( κ (t s )) dWs .
0
EDS

A¢ ne Models I
Introduction

Ito’s lemma

Solutions to
A model is said to belong to the exponential a¢ ne class if
SDEs
Gaussian models
there is determinisic functions a (t, T ) and b (t, T ) for
Particular cases
A¢ nes Models
which the time t value of a zero-coupon bond paying 1
Lipschitz
Coe¢ cients
dollar at maturity T is

P (t, T ) = exp (a (t, T ) r (t ) + b (t, T )) .

In that case, the rishless interest rate as the form


q
drt = (µ1 (t ) rt + µ2 (t )) dt + σ1 (t ) rt + σ2 (t ) dWt

where µ1 ( ), µ2 ( ), σ1 ( ) and σ2 ( ) are deterministic


function of time.
Bisière, pages 116-117 et 128-129.
EDS

A¢ ne Models II
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases The Gaussian models belong the the exponential-a¢ ne
A¢ nes Models
Lipschitz classe of model..
Coe¢ cients

Square-root Processus (Feller’s branching di¤usion)


Cox-Ingersoll-Ross (1985), generalised par Hull-White
(1990)
p
drt = κ (θ rt ) dt + σ rt dWt .
Bisière, pages 116-117.
EDS

Lipschitz Coe¢ cients


Introduction
Theorem
Ito’s lemma

Solutions to
Consider dXt = b (t, Xt ) dt + σ (t, Xt ) dWt . If b ( , ) and
SDEs
Gaussian models
σ ( , ) are continuous functions such that there is a constant
Particular cases
A¢ nes Models
K < ∞, for which
Lipschitz
Coe¢ cients
jb (t, x ) b (t, y )j + jσ (t, x ) σ (t, y )j K jx yj
jb (t, x )j + jσ (t, x )j K (1 + jx j)
E X0 2
< ∞

then, for all T 0, the equation (EDS) admits a unique


solution in the interval [0, T ]. Moreover, this solution veri…es

E sup jXt j2 < ∞.


0 t T

There are su¢ cent but not nessary conditions. (Lamberton et


Lapeyre, p. 59-60.)
EDS

Références I
Introduction

Ito’s lemma

Solutions to
Martin Baxter et Andrew Rennie (1996). Financial
SDEs Calculus, an indroduction to derivative pricing, Cambridge
Gaussian models
Particular cases university press.
A¢ nes Models
Lipschitz
Coe¢ cients Christophe Bisière (1997). La structure par terme des taux
d’intérêt, Presses universitaires de France.
Donald L. Cohn (1980). Measure Theory, Birkhäuser.
Richard Durrett (1996). Stochastic calculus : a practical
introduction, CRC Press.
Richard R. Goldberg (1976). Methods of real analysis,
second edition, Wiley.
Ioannis Karatzas et Steven E. Shreve (1988). Brownian
motion and stochastic calculus, deuxième édition,
Springer-Verlag.
EDS

Références II
Introduction

Ito’s lemma

Solutions to
SDEs
Gaussian models
Particular cases
A¢ nes Models
Lipschitz Damien Lamberton et Bernard Lapeyre (1991).
Coe¢ cients
Introduction au calcul stochastique appliqué à la …nance,
Ellipses.
Daniel Revuz et Marc Yor (1999). Continuous martingales
and brownian motion, third edition, Springer.

You might also like