Chapter 6

Geometric Brownian Motions

Normal Distributions

We begin by recalling the normal distribution briefly. Let X be a normally distributed

random variable with mean and standard deviation , i.e., X N (, 2 ). The
probability density function of X is given by

1
(x )2
pX (x) =
exp
, < x < .
(1)
2 2
2 2
This is the
R famous bell-curve of normal distributions. Since probability adds up to 1,
we have pX (x)dx = 1. Clearly E[X], the mean or expected value of X, is equal
to , and Var(X), the variance of X, is equal to 2 .
Next we want to derive the nth moment E[X n ] of X. For this we need the moment
generating function E[etX ]. If it exists for some t > 0, then one can justify that
E[e

tX

]=

X
E[X k ]
k=0

k!

tk .

(2)

Take note that the right hand side of (2) is the power series expansion in t of the
function E[etX ]. Its coefficients are precisely the moments of the random variable X.
Observe that the kth order derivative of the function E[etX ] at t = 0 is given by
dk
E[etX ]|t=0 = E[X k ],
(3)
dtk
which is precisely the kth moment of X.
However, it is not easy to compute the function E[etX ], not to mention its derivatives. Instead we consider the standard normal random variable
X
N (0, 1).
(4)
Z

and use it to compute the moments of Z and hence of X. Putting = 0 and = 1 in

(1), we get
Z
2
1
etz ez /2 dz
E[etZ ] =
2
Z
Z
2
2
2
1
1
1
=
etzz /2 dz =
e 2 [(zt) t ] dz
2
2
Z
2
2
1
t
t2
1
e 2 (zt) dz = e 2 ,
=e2
2

58

MAT4210 Notes by R. Chan

R
where the last equality follows from pZ (z)dz = 1.
From that and (3), we get

0,
when n is odd,
E[Z n ] = (2k)!

, when n = 2k.
2k k!

(5)

From (4), since X = Z + , the moment generating function of X is given by

E[etX ] = E[etZ+t ] = et E[etZ ] = et e

2 t2
2

= et+

2 t2
2

From this we can compute the kth moments of X. In fact, since X = Z, we have
by (5),

0,
when n is odd,
E[(X )n ] =
(2k)!
n
, when n = 2k.
2k k!
From that, one can compute the moments of X by recursion.
Recall that for 0 < < 1/2, the 100 percentage point z of the standard normal
distribution is defined as the number such that
Z
2
1

= Prob{Z > z } =
eu /2 du.
2 z
Referring to a statistical table, we have for examples, z0.050 = 1.6449 and z0.025 =
1.960. More precisely, if Z N (0, 1), then within 95% confidence,
1.96 Z 1.96.
If X N (, 2 ), then by (4),
1.96 X + 1.96.

(6)

If X signifies the daily return of a stock, then we say the stock will have a 2.5% value
at risk (VAR) of $( 1.96). That implies that on average the stock holder expects
a daily return of no more than $( 1.96) in 1 day out of 40 trading days.

A Simple Model for Asset Prices

The asset price and stock price were previously denoted by S(t). Here we are going to
take the volatility into account. The standard, generic symbol for such a randomness is . Thus we denote the asset price at time t by S(t, ) or simply by St (). It is
a stochastic process, i.e. for each fixed t, St () is a random variable depending on .
For those who are not familiar with stochastic processes, you may consider each as a
realization of the process S(t). More precisely, if we are allowed to move back in time
(say a year ago) and start the process S(t) for a year, then because of the randomness
in S(t), we should not be getting the same set of S(t) as we already had. If we can
repeat this experiment, i.e. move back in time a year and start S(t) again, we should
be getting another new set of S(t). The parameter can be considered as the index
of these experiments that you have taken, e.g. St (1) is the first experiment, and St (2)
as the second experiment. However, need not be an integer. In the following, most

Geometric Brownian Motions

59

of the time, we will simply write St () as S(t) or St , with the understanding that it
is a stochastic process depending on .
We are going to build a model for a non-dividend-paying stock S(t). But in general,
what can we say about the stock price other than the fact that it is stochastic? Let
us look at the price charts of HSBC and Yahoo from February 01 to February 06 in
Figure 1. What do you see? We see fluctuation of the prices, that is for sure. But we
also see one important thing. Both stock prices have an upward trend. In the short
term, the price may rises and falls, but in the long term, the price is going to be
higher. It is like putting money in the bank, our return will be positive and increasing
according to the interest rate. For stocks, the long term returns seem to the positive
although the return rate may be different for different stocks. From the charts, the
return rate of Yahoo surely is much higher than that of HSBC.

Figure 1. HSBC and Yahoo prices for the last 5 years.

We will build our model of the stock prices based on the above observation. We
first recall the return is defined to be the change in the price divided by the original
value, i.e.
change in price
return =
.
original value
Consider a small subsequent time interval (t, t + t), during which S(t) becomes
S(t + t) = S(t) + S(t), where S(t) = S(t + t) S(t). The return now can be
expressed as
S(t + t) S(t)
S(t)
=
.
(7)
S(t)
S(t)
Let us adopt the differential notation used in calculus. Namely, we use the notation
dt for the small change in any quantity over this time interval when we intend to
consider it as an infinitesimal change. Then (7) can be written as dS/S. How might
we model this return dS/S on the asset S(t)? According to the observation in Figure
1, the commonest model decomposes this return into two parts.
1. One is a predictable, deterministic and anticipated return akin to the return on
money invested in a risk-free bank. It gives a contribution dt to the return dS/S,
where is a measure of the average rate of growth of the asset price. Such a
parameter is also known as the drift. In simple models, is taken to be a
constant. In more complicated models, can be a function of S and t.

60

MAT4210 Notes by R. Chan

2. The second contribution to dS/S models the random change in the asset price
in response to external effects, such as unexpected news. It is represented by
a random sample dX(t, ) drawn from a normal distribution with mean 0 and
adds a term dX(t, ) to dS/S. Here is a number called the volatility, which
measures the standard deviation of the returns, and emphasizes that dX(t, )
is a random process. For simplicity, we will sometimes simplify the writing of
dX(t, ) by dXt (), dX(t) or dXt .
Putting these two contributions together, we obtain the stochastic differential equation
dS(t)
= dt + dX(t),
(8)
S(t)
which is the mathematical representation of our simple recipe for the price of a nondividend-paying stock. Equivalently, (8) can also be expressed as
dS(t) = S(t)dt + S(t)dX(t),
and

Z
St () = St0 () +

Su ()du +
t0

(9)

Su ()dXu ().

(10)

t0

Later in Definition 5, we will give more precise definitions of the drift parameter
and the volatility . The only symbol in (8) whose role is not yet clear is dX(t).
Suppose we were to cross out this term by taking = 0. We then would be left with
the ordinary differential equation
dS
= dt or
S

dS
= S.
dt

When is a constant, this can be solved exactly to give exponential growth in the
value of the asset, i.e.,
S(t) = S(t0 )e(tt0 ) ,
where S(t0 ) is the value of the asset at time t0 . Thus, if = 0, the asset price is totally
deterministic, and we can predict the future price of the asset with certainty. In fact,
it increases exponentially (or geometrically with common factor e ). Its like putting
money in the bank with interest rate .
The model (8)(10) for stock prices seems to have been first mentioned by Paul
Samuelson in 1965 in his paper Rational Theory of Warrant Prices, Industrial Management Review, 6 (1965), pp. 1331. For reasons that will be clear in Chapter 7, the
model is sometimes also referred as geometric Brownian motion.
Those who are not familiar with stochastic differential equations may tempt to
integrate (8) directly and get
S(t) = S(t0 )et+X(t) .
We will see in the next chapter that this is not the correct solution to (8), and in fact
dS/S 6= d ln S. Thus there are difference between stochastic differential equations and
ordinary differential equations. The difference can fortunately be handled easily by
Itos Lemma discussed in the next chapter.

Geometric Brownian Motions

61

Wiener Process and its Generalizations

The term dX(t) in (8) which gives the randomness to S(t) is certainly the main feature
of the geometric Brownian motion. In fact, X(t) is a Wiener process and is known to
follow Brownian motions. Historically speaking, such a random process was observed
by Robert Brown, an English botanist, in the summer of 1827, that pollen grains
suspended in water performed a continual swarming motion. Hence it was named
after Robert Brown, called Brownian motion.
Brownian motion is a process of tremendous practical and theoretical significance.
It was used as an model to explain the ceaseless irregular motions of tiny particles
suspended in a fluid. It had also been used as a model of the stock market in Louis
Bacheliers (1900) work. His paper was at first largely ignored by academics for many
decades, but now his work stands as the innovative first step in a mathematical theory
of stock markets that has greatly altered the financial world today.
In 1905, Albert Einstein gave a satisfactory explanation and asserted that the
Brownian motion originates in the continued bombardment of the pollen grains by
the molecules of the surrounding water, with successive molecular impacts coming
from different directions and contributing different impulses to the particles. (Incidentally, 1905 is the same year in which Einstein set forth his theory of relativity and his
quantum explanation for the photoelectric effect.) But, Brownian motion is complicated and it is not surprising that it took more than another decade to get a clear
picture of the Brownian motion stochastic process.
A rigorous mathematical foundation upon which Brownian motion could be built
had to wait until 1920s. In 1923, Norbert Wiener (18941964) laid a rigorous mathematical foundation and gave a proof of its existence. Hence, it explains why it is now
also called a Wiener process. In the sequel, we will use both Brownian motion and
Wiener process interchangeably.
Definition 1. We say a stochastic process {Xt (), t 0} is a standard Wiener process
if it satisfies the following conditions:
(a)
(b)
(c)
(d)

X0 () = 0 for all ;
for all , the map t 7 Xt () is a continuous function for t 0 ;
for every t and h 0, the change [Xt+h () Xt ()] N (0, h); and
{Xu () Xv () : 0 v u t} are independent.

Condition (a) says that the starting point of a standard Wiener process is the
origin. We in fact frequently speak of { + Xt () : t R+ } as a Wiener process
started at . Note that this starting point can be a fixed, real number, or a random
variable independent of Xt ().
The conditions (b)(c) are the really essential ones. For each fixed , the function
t 7 Xt () is called sample path (realization, trajectory) of the Wiener process associated with . By (b), Wiener paths are continuous. Again for simplicity, we will simply
write Xt () as Xt or X(t). By (c), the increments are normally distributed with mean
0 and variance equal to the time difference. Thus by (1), for any event A,
Z
2
1
ex /2h dx.
Prob{Xt+h Xt A} =
2h A
Example 1. Say, a Wiener process Y (t) is initially equals 25 and the time t is measured
in years. Let us compute the probability distribution of Y (t) at the end of one year.

62

MAT4210 Notes by R. Chan

Let {X(t), t 0} be a standard Wiener process starting at 0. Then, Y (t) is given by

Y (t) = X(t) + 25. By Condition (c), X(1) = X(1) X(0) N (0, 1). Therefore, the
probability distribution of Y (1) is normal with mean 25 and variance 1.
From (d) and (a), one can deduce that, for 0 t0 < t1 < . . . < tn , the random
variables
Xt0 , (Xt1 Xt0 ), . . . , (Xtn Xtn1 )
are independent, i.e., Xt has independent, normally distributed, increments. In particular, we have Xt+h Xt is independent on Xt . Hence a Wiener process has what
the so-called Markov property, which means that only the present value of the process
is relevant for predicting the future, while the past history of the process and the
way that the present has emerged from the past are irrelevant. A stochastic process
which satisfies the Markov property is called a Markov process. Generally speaking,
the following expression says that a Wiener process is Markov: for t 0 and h > 0,
Prob{Xt+h A | Xs , 0 s t} = Prob{Xt+h A | Xt }.

(11)

Take note that stock prices are usually assumed to follow a Markov process. Suppose that the price of XYZ stock is $50 now. If the stock price follows a Markov
process, our prediction for the future should be unaffected by the price one week ago,
one month ago, or one year ago. The only relevant piece of information is the price
now which is $50. Obviously, predictions for the future are uncertain and must be
expressed in terms of probability distributions. The Markov property implies that the
probability distribution of the price at any particular future time is not dependent on
the particular path followed by the price in the past.
Definition 2. Let {X(t), t 0} be a standard Wiener process. The process
Y (t) = X(t) + t + ,

(12)

where is any real parameter and > 0, is called a generalized Wiener process,
starting at , with a drift parameter and a variance rate 2 .
Using the properties of Xt , we see that
Y (t + h) Y (t) N (h, 2 h).
In differential notations, (12) becomes

Y (0) = ,
dY (t) = dt + dX(t),

for t 0.

Example 2. Consider the situation where the cash position Y (t) of a company, measured in thousands of dollars, follows a generalized Wiener process with a drift parameter = 20 per year and a variance rate of 900 per year. Initially, the cash position
is 50. Let us compute the cash position of the company at the end of 5 years. Clearly,
Y (t) = X(t) + t + , where = 50, = 20 and = 30. At the end of 5 years, the
probability distribution of Y (5) is normal with mean
+ t = 50 + 20 5 = 150,
and variance

2 t = 900 5 = 4500.

In other words, Y (5) N (150, 4500).

Geometric Brownian Motions

63

Definition 3. Let {X(t), t 0} be a standard Wiener process. The process

dY (t) = a(Y, t)dt + b(Y, t)dX(t),
where both the drift a(Y, t) and the variance rate b(Y, t) are functions of the underlying
process Y (t) and time t is called an It
o process.
The geometric Brownian motion in (9) is an Ito process.

Geometric Brownian Motions

Let us consider the geometric Brownian motion S(t) in 2 again. The process X(t) in
(8)(10) is a standard Wiener process. Note that dX(t) (= dXt ) can be considered as
the limit of X(t + dt) X(t) as dt 0. Hence by Condition (c) in Definition 1, we
have, as dt 0:
(i) dX(t) is a random variable, drawn from a normal distribution;
(ii) the mean of dX(t) is zero, i.e., E[dX(t)] = 0;
(iii) the variance of dX(t) is dt, i.e., Var(dX(t)) = dt.
In short, dXt N (0, dt) and we can write

dXt = dt,

(13)

where N (0, 1) is a standard normal random variable. How about the random
variable (dXt )2 ? First it mean is E[(dXt )2 ] = Var(dXt ) = dt. By (5), we also get
Var((dXt )2 ) = Var(2 (dt)) = E[(2 E[2 ])2 ](dt)2 = 2(dt)2 .
Using this we can estimate the mean and variance of the return on S(t).
Proposition 4. The return dS/S on the asset S(t) satisfies
dS
N (dt, 2 dt).
S

(14)

Proof. Since dS/S = dt + dXt where dXt is normally distributed, dS/S is also
normally distributed. By (8) and (ii) above,

dS
E
= E[dt + dXt ] = dt.
(15)
S
By (15), (8) and (iii), the variance is given by
"
"

2 #
2 #
dS
dS
dS
dS
Var
=E
E
=E
dt
= E[(dXt )2 ] = 2 dt.
S
S
S
S
(16)

The standard deviation of dS/S is thus equal to dt.

Equation (15) says that on average, the return on the asset S(t) is increasing as a
rate of dt. Thus the next value for S is higher than the old one by an amount Sdt.
If we recall the definition of interest rate in (3.4), this indicates that asset on average

64

MAT4210 Notes by R. Chan

13

13

=0.15, =0.30

12.5
12

12
11.5

11

S(t)

S(t)

11.5

10.5
10
9.5

11
10.5
10
9.5

8.5

8.5

=0.05, =0.15

12.5

0.2

0.4

0.6

0.8

0.2

Time

0.4

0.6

0.8

Time

Figure 2. Asset prices for two different stocks with different and .

is earning an interest rate . From (15) and (16), we also see that if we compare two
different asset-price processes as described by (8), the one with the larger value of
usually rises more steeply and the one with the larger value of appears more jagged,
see Figure 2 for two charts simulated according to the geometric Brownian motion
(8). Note that the simulated price charts do have some resemblance to the real price
charts in Figure 1.
Recall in (11) that X(t) is Markov: it does not depend on past history. Moreover
(9) does not refer to the past history of the asset price. Therefore the next asset price
S(t + dt) depends solely on todays price S(t) and not on the past asset price. Thus
the geometric Brownian motion S(t) also has the Markov property.
The discrete version of (14) is
S
N (t, 2 t),
S
where the variable S is the change in the stock price S in a small interval of time
t. By (4), we then have

S
(17)
= t + t,
S
where N (0, 1) is a random number drawing from a standard normal distribution.
The left-hand side of (17) is the return provided by the stock in a short period
of
time, t. The term t is the expected value of this return, and the term t
is the stochastic component of the return. The variance of the stochastic component
(and, therefore, of the whole return) is 2 t. If we let t = 1 unit time, we have the
following definition.
Definition 5. The drift of a stock is the expected return of the stock per unit time,
and the volatility of the stock is the standard derivation of the return of the stock
per unit time.
Example 3. Consider a stock that pays no dividends, has a volatility of 30% per annum, and provides an expected return of 15% per annum with continuous compounding. That is, it follows the model (8) with = 0.30 and = 0.15. To be precise,
dS(t)
= 0.15dt + 0.30dXt .
S(t)

Geometric Brownian Motions

65

If S is the stock price at a particular time and S is the increase in the stock price
in the next small interval of time, then

S
= 0.15t + 0.30 t,
S
where N (0, 1). Let us consider a time interval of one week or 0.0192 year and
suppose that the initial stock price is $100. Then, t = 0.0192, S = 100, and hence
S = 100(0.00288 + 0.0416) = 0.288 + 4.16,
showing that the price increase is a random variable following the normal distribution
with mean $0.288 and standard deviation $4.16.
A Monte Carlo simulation of a stochastic process is a procedure for sampling
random outcomes for the process. As an example, we can use (17) to simulate the
stock price S(t) that follows a geometric Brownian motion. We just need to draw
random numbers in N (0, 1).
Example 4. Suppose that the expected return from a stock is 14% per annum and
that the standard deviation of the return (i.e., the volatility) is 20% per annum. This
means that = 0.14 and = 0.20. Suppose that t = 0.01 so that we are considering
changes in the stock price in time intervals of length 0.01 year (or 3.65 days). From
(17),

S = 0.14 0.01S + 0.2 0.01S = 0.0014S + 0.02S.

(18)
A path for the stock price can be simulated by sampling repeatedly for from a
standard normal distribution N (0, 1) and substituting it into (18). For instance, the
following table is one particular set of results from doing this.
Stock Price at
Start of Period
20.000
20.236
20.847
20.518
21.146
20.883
20.603
20.719
20.292
20.617
21.124

Random Sample
for
0.52
1.44
0.86
1.46
0.69
0.74
0.21
1.10
0.73
1.16
2.56

Change in Stock-price
During Period
0.236
0.611
0.329
0.628
0.262
0.280
0.115
0.427
0.325
0.507
1.111

Here the initial stock price is assumed to be $20. For the first period, is sampled as
0.52. From (18), the change during the first time period is
S = 0.0014 20 + 0.02 20 0.52 = 0.236.
At the beginning of the second time period, the stock price is, therefore, $20.236. The
value of sampled for the next period is 1.44. From (18), the change during the second
time period is
S = 0.0014 20.236 + 0.02 20.236 1.44 = 0.611.

66

MAT4210 Notes by R. Chan

At the beginning of the next period the stock price is, therefore, $20.847; and so on.
Note that, because the process we are simulating is Markov, the samples that we
generate should be independent of each other. Also the above table only shows one
possible pattern of stock price movements. Different random samples would lead to
different price movements. In fact, if we start all over again, we should be getting a
completely new path for S(t).

An Important Fact: (dXt )2 = dt

In (13), we have seen that dXt = dt, where N (0, 1). Hence (dXt )2 = 2 dt
where E[2 ] = Var[] = 1. In fact, we can show that the smaller dt becomes, the more
certainly (dXt )2 is equal to dt. To prove this, we need to establish a famous inequality
first.
Proposition 6. (Chebyshevs inequality) For any continuous random variable R and
any positive , we have
1
Prob{|R| } 2 E[R2 ].
(19)

Proof. Let p(r) be the probability density function of R. We have

Z
Z
Z
2
2
2
E[R ]
r p(r)dr
r p(r)dr
2 p(r)dr = 2 Prob{|R| }.

|r|

|r|

Theorem 7. Let {X(t), t 0} be a standard Wiener process starting at 0. Then

Z t

Prob (dXs )2 t = 0
(20)
0

for any > 0. In other words, as dt 0, (dXt )2 dt with probability 1.

Proof. Let us take the partition {0, t/n, 2t/n, . . . , t} of the interval [0, t], and denote
i = i t/n. Note that by the definition of integration
Z t
n
X
2
(dXs )2 = lim
(X(i ) X(i1 )) ,
0

i=1

where we will assume here and in the following that all limits exist. By Definition 1(c),
if we define
X(i ) X(i1 )
p
,
Zn,i
t/n
then for each n, the sequence Zn,1 , Zn,2 , . . . is a set of independent, identically distributed N (0, 1) random variables. Moreover, we have
Z t
n
n
t X 2
t X 2
(dXs )2 t = lim
Zn,i t = lim
(Zn,i 1) = lim Rn ,
(21)
n n
n n
n
0
i=1
i=1
where

t X 2
Rn
(Z 1).
n i=1 n,i

Geometric Brownian Motions

67

To prove (20), we will apply the Chebyshev inequality (19) on Rn . Thus we need to
compute E[Rn2 ]. Note that

n
! n
2
X
X
t
2
2
(Zn,i
1) (Zn,j
Rn2 = 2
1)
n
i=1
j=1

n
n
n
X
t2 X X 2 2
2
Zn,i Zn,j 2n
Zn,i
+ n2 .
= 2
n
i=1 j=1
i=1
Note that for all n and i, Zn,i N (0, 1), hence

n X
n
n
2
X
X
t
2
2
2
E[Rn2 ] = 2
E[Zn,i
Zn,j
] 2n
E[Zn,i
] + n2
n
i=1 j=1
i=1

n
n
t2 X X
2
2
E[Zn,i
Zn,j
] n2 .
= 2
n
i=1 j=1
Recall that Zn,i and Zn,j are independent of each other if i 6= j, hence

n
n X
n
X
t2 X
4
2
2
2
2
E[Zn,i ] +
E[Zn,i ]E[Zn,j ] n
E[Rn ] = 2
n
i=1
i=1 j6=i
!
n
t2 X
4
= 2
] + n(n 1) n2 .
E[Zn,i
n
i=1
By (5), we then have,
E[Rn2 ] =

2t2
t2
2
3n
+
n(n

1)

n
=
.
n2
n

Putting this into the Chebyshev inequality (19), we then have

0 Prob{|Rn | }

2t2
,
2 n

for all > 0.

Taking limit as n , we have

2t2
= 0.
n 2 n

0 lim Prob{|Rn | } lim

n

Recall that probability is just an integration and limit and integration can interchange
order under sufficiently nice conditions, which we assume are satisfied here. Hence
Prob{| lim Rn | } = Prob{ lim |Rn | } = lim Prob{|Rn | } = 0.
n

Hence by (21), we have (20). Thus with probability 1,

Z t
(dXs )2 = t,
0

or in differential form (dXt )2 = dt.

(22)

68

MAT4210 Notes by R. Chan

Let us illustrate (22) by a numerical example. We set the time interval ti ti1 =
0.001 for all i = 1,P
, 10000, and simulate X(ti ) X(ti1 ) N (0, 0.001). Then we
n
i1 2
i
) X( 1000
)) for all n = 1, , 10000. The sum for each
compute the sum i (X( 1000
n is shown as the red curve in Figure 3. We see that the red curve is indeed very close
to the function y(t) = t.

Figure 3. A numerical result to show that

Rt
0

(dXs )2 = t

From Theorem 7, we can derive the following estimates:

Proposition 8. As dt 0, we have

(i) dXt = O( dt); i.e. limdt0 {dXt / dt} = constant.

(ii) dXt dt = o(dt); i.e. limdt0 {(dXt dt)/dt} = 0.
(iii) (dXt )2 = dt;
Example 5. Squaring both sides of (9), we get
(dS)2 = 2 S 2 (dt)2 + 2S 2 dtdXt + 2 S 2 (dXt )2 .
By Proposition 8, it becomes
(dS)2 = 2 S 2 dt + o(dt).
In the limit dt 0, we have
(dS)2 = 2 S 2 dt.