MScFE xxx [Course Name] - Module X: Collaborative Review Task

Video Transcripts
Module 3
MScFE 622
Continuous-time Stochastic
Processes

© 2019 - WorldQuant University – All rights reserved.

Revised: 07/07/2020
1
MScFE 622 Continuous-time Stochastic Processes – Video Transcripts Module 3

Table of Contents

Unit 1: Square Integrable Martingales .......................................... 3

Unit 2: Localization............................................................................ 5
Unit 3: Semimartingales................................................................... 7

© 2019 - WorldQuant University – All rights reserved.

2
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 1) Module 3: Unit 1

Unit 1: Square Integrable Martingales

Hi, in this video we extend the theory of stochastic integration to the case where the integrator is
a square integrable martingale.

Let 𝑀 be a square integrable martingale. We want to define the stochastic integral of an

appropriate stochastic process, 𝜑 with respect to 𝑀, which is similiar to what we did when 𝑀 was a
Brownian motion:
"
# 𝜑! 𝑑𝑀!
#

We are going to start with what is called a simple process, which is a stochastic process that can be
written like this:
&

𝜑$ = ' 𝐻% 𝐼(* % , (, * %] (")

%'(

where, the τ% 𝑠 are stopping times (i.e. when 0 = τ# < τ( < ⋯ τ& = 𝑇).

𝜑 is a simple process like this:

" &

# 𝜑! 𝑑𝑀! = ' 𝐻% (𝑀" ∧ τ% − 𝑀" ∧ τ%,( )

# %'(

What we see above is the definition of the stochastic integral, which is similar to the case of a
Brownian motion. This means that it also satisfies similar properties to the case of a Brownian
motion.

"
1 𝐸 5∫# 𝜑! 𝑑𝑀! 7 = 0

"
2 This stochastic process, {∫# 𝜑! 𝑑𝑀! : 0 ≤ 𝑡 ≤ 𝑇} is a square integrable martingale with the
following quadratic variation:
" "
=# 𝜑! 𝑑𝑀𝑠> = # 𝜑!0 𝑑⟨𝑀⟩!
# #
"

© 2019 - WorldQuant University – All rights reserved.

3
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 1) Module 3: Unit 1

" 0
3 𝐸 A5∫# 𝜑! 𝑑𝑀! 7 B is equal to the expected value of the integral of the square of 𝜑! with

respect to the quadratic variation of 𝑀 at 𝑠. Written in full:

" 0 "
𝐸 CD# 𝜑! 𝑑𝑀! E F = 𝐸 D# 𝜑!0 𝑑⟨𝑀⟩! E
# #

This is similar to what we had in Brownian motion, because the quadratic variation of
Brownian motion at time 𝑠 is equal to 𝑠 itself.

Once we have defined this for simple processes, we again do an extension.

We begin by defining 𝐿0 (𝑀) to be a set of predictable processes – meaning that we now restrict
this to predictable processes. This restriction was unnecessary in the Brownian case as we just
defined it for progressive processes, which are predictable with respect to the Brownian filtration.
Here, however, we define it for predictable processes, 𝜑, such that 𝜑 is predictable and the norm
of 𝜑 is finite, where the norm of 𝜑 is again defined in a similar way to the Brownian case:
(
2 0
‖𝜑‖1 = D𝐸 # 𝜑!0 𝑑⟨𝑀⟩! E
#

Once we have the above, we can show that for every element of 𝐿0 (𝑀), there exists a sequence of
simple processes that we are going to denote by (𝜑& ), such that 𝜑& converges to 𝜑 in the following
sense:
‖𝜑& − 𝜑‖1 → 0 𝑎𝑠 𝑛 → ∞

In that case, we define the stochastic integral of 𝜑 to be equal to the limit as 𝑛 tends to infinity of
the stochastic integrals of 𝜑& , where this limit is taken in 𝐿0 . Written in full:

" "
# 𝜑𝑑𝑀! : = lim # 𝜑& 𝑑𝑀!
# &→4 #

This limit is independent of the sequence 𝜑& that we have chosen and that completely defines the
stochastic integral with respect to a square integrable martingale. Now that we have defined a
stochastic integral with respect to a square integrable martingale, in the next video we are going
to look at localization.

© 2019 - WorldQuant University – All rights reserved.

4
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 2) Module 3: Unit 2

Unit 2: Localization
Hi, in this video we introduce the notion of localization and we use it to extend the theory of
stochastic integration.

Let ℰ be a class of stochastic processes, which is essentially a set of stochastic processes. For an
example, we can take ℰ to be a set of martingales on the same probability space:

ℰ=𝑀

We say that the stochastic process 𝑋 is locally in ℰ, which we write as 𝑋 ∈ ℰ567 , if it satisfies the
following conditions:

1 An increasing sequence of stopping times must exist, which we will denote by τ& , such
that τ& → ∞ as 𝑛 → ∞.
2 The stopped process, 𝑋 8! ∈ ℰ∀𝑛. This applies even though 𝑋 itself need not belong to ℰ.

As a famous example, if we take ℰ to be a set of all martingales, then ℰ567 , which is the set of all
stochastic processes that are locally in ℰ, in this case will simply coincide with the set of all
stochastic processes that are local martingales:

ℰ = 𝑀, ℰ567 = 𝑀567

A popular example of a local martingale that we will see many times in this course, is when 𝑊 is a
"
Brownian motion and 𝜑 is progressive, with the additional condition that ∫# 𝜑!0 𝑑𝑠 < ∞∀𝑡 and this

is true almost surely. So, if this condition is satisfied, then we know that the stochastic integral
"
∫# 𝜑! 𝑑𝑊! of the stochastic process 𝜑! is well-defined.

However, we mentioned in the last module that this is not generally a martingale if 𝜑 does not
"
belong to 𝐿0 (𝑊). If 𝜑 ∉ 𝐿0 (𝑊), then the stochastic integral ∫# 𝜑! 𝑑𝑊!9 need not be a martingale.
This is something important to remember.

"
However, we will show now that ∫# 𝜑! 𝑑𝑊! is always a local martingale.

© 2019 - WorldQuant University – All rights reserved.

5
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 2) Module 3: Unit 2

As a sequence, τ& , which is called a localizing sequence, can be taken as being the first time the
"
infimum, which is ∫# 𝜑!0 𝑑𝑠, is greater than or equal to 𝑛. Written in full:

"
τ& : = 𝑖𝑛𝑓{𝑡 ≥ 0: # 𝜑!0 𝑑𝑠 ≥ 𝑛}
#

8
Now, if we look at the stopped process 𝜑&! , it is always bounded above by 𝑛 and therefore belongs
8
to 𝐿0 (𝑊). This means that 𝜑&! will be a square integrable martingale for all 𝑛. So, the stochastic
" 8
integral ∫# 𝜑! ! 𝑑𝑊! will be a square integrable martingale.

" $ 8"
8
∫# 𝜑! ! 𝑑𝑊! is also equal to the stopped process 5∫# 𝜑! dW: 7 . We can check that this is equivalent

to stopping the integral process itself, and therefore, the integral process is a local martingale
because when we stop it we get a martingale.

Finally, this allows us to extend the theory of stochastic integration to the case where we have
"
∫# 𝜑! 𝑑𝑀 > and this here no longer belongs to 𝐿0 of 𝑀, and the martingale 𝑀 is not necessarily
square integrable – in fact, we can extend it even further to the case where it is just a local
martingale via localization. Further details are covered in the notes.

Now that we've covered localization, in the next video we are going to look at semimartingales.

© 2019 - WorldQuant University – All rights reserved.

6
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 3) Module 3: Unit 3

Unit 3: Semimartingales
Hi, in this video we look at semimartingales.

A stochastic process, 𝑋, is called a continuous semimartingale if 𝑋 can be written in the following

form:
𝑋 = 𝑋# + 𝑀 + 𝐴

where 𝑋# is of course the starting point of the stochastic process. We are going to insist that 𝑀 is
continuous, but it is also a local martingale. 𝐴 is a continuous finite variation process, meaning that
the sample paths of 𝐴 have finite variation.

Now, if 𝑋 is a continuous semimartingale with the above decomposition, we will define the
stochastic integral of the stochastic process 𝜑 with respect to 𝑋 in two steps. The first integral will
be with respect to 𝑀, which is a local martingale as we defined in the previous section, plus the
second integral with respect to 𝐴, which is the Stieltjes integral. To be clear, the first integral is the
stochastic integral, while the second is the ordinary Stieltjes integral. Written in full:

" " "

# 𝜑! 𝑑𝑋! : = # 𝜑! 𝑑𝑀! + # 𝜑! 𝑑𝐴!
# # #

Now 𝜑! will not exist for any arbitrary 𝜑, so, necessarily, we need to make sure that the stochastic
" "
integral exists for when 𝜑 is locally bounded and predictable so that ∫# 𝜑! 𝑑𝑀! and ∫# 𝜑! 𝑑𝐴! both
exist. Those conditions will both be satisfied if we assume that 𝜑 is locally bounded and
predictable. However, there are cases where the stochastic integral exists even though 𝜑 is not
locally bounded and integrable.

We are going to denote by 𝐿(𝑋) the set of all processes 𝜑, such that 𝜑 is 𝑋-integrable in the sense
that the stochastic integral exists.

Now that we have introduced the stochastic integral, we can move on to Ito's Lemma for
continuous semimartingales,

Let 𝑋 be a 𝑑-dimensional semimartingale, 𝑋 = a𝑋( , … , 𝑋 ; c, which means that each component is a

semimartingale, then take ℱ to be a function, where ℱ: ℝ; → ℝ, and assume that ℱ is 𝐶 − 2,
meaning that it is twice continuously differentiable.

© 2019 - WorldQuant University – All rights reserved.

7
MScFE 622 Continuous-time Stochastic Processes – Video Transcript (Lecture 3) Module 3: Unit 3

We will define 𝑌" to be ℱ(𝑋" ), so this is a new stochastic process and a function of 𝑋" .

Ito's Lemma says that 𝑌" is also a semimartingale and it gives us the stochastic differential for
𝑌" with respect to the partial derivatives of the function ℱ as follows:

;
δℱ 1 δ0 ℱ
dY$ = ' 𝑑𝑋" % + ' ' 𝑑m𝑋 % , 𝑋< n"
δ𝑥% 2 δ𝑥%, δ𝑥<
%'( % <

Finally, it is important to mention that the notion of a semimartingale can be extended to a case
where 𝑋 is no longer continuous. In general, a stochastic process 𝑋 is a semimartingale if it can be
decomposed into the three parts discussed at the beginning of the video (𝑋 = 𝑋# + 𝑀 + 𝐴), where
neither 𝑀 nor 𝐴 are necessarily continuous, but 𝑀 is a local martingale and 𝐴 is a finite variation.
Furthermore, we can still define the stochastic integral in the same way for locally bounded and
predictable processes when the process 𝑋 is a just a semimartingale without being continuous.
Further details are included in the notes.

That brings us to the end of this module. In the next module, we are going to look at continuous
trading.

© 2019 - WorldQuant University – All rights reserved.

8