Organizational Stuffs: Typed by Vinh Nguyen. Date May 18, 2011

McGill Summer 2011 - Econ 765 - Lecture 1 - Date May 3, 2011
Typed by Vinh Nguyen. Date May 18, 2011.

0. Organizational Stuffs
Course website:
http://russell-davidson.arts.mcgill.ca/e765
Texts: Stochastic Calculus for Finance, Vols I and II, by Steven Shreve
Assessment: 5 assignments + 1 final. The assignments can be counted up to 30% of the final
grade if they help the final grade. Otherwise, the final grade is solely determined by the final
exam.
1. Two-period Model
There are 2 markets: the market for a certain stock (risky) and the money market (riskless). In
period 0, the price of the stock is S
0
> 0. Between period 0 and period 1, we toss a coin whose
outcome determines whether the stock price changes by a factor of u or a factor of d, where
0 < d < u. We write S
1
(H) = uS
0
and S
1
(T) = dS
0
. The per-period rate of interest for the money
market account is r. This is the rate of interest applied to both investing and borrowing.
Arbitrage is a trading strategy that begins with zero capital and trades in the stock
and money markets in order to make money with positive probability without any possibility of
losing money.
Proposition. The 2 period binomial model admits no arbitrage if and only if
0 < d < 1 +r < u.
Proof. Suppose we have no arbitrage. If 1 +r d < u, then in period 0, one can borrow some
amount of money M > 0 to buy
M
S
0
shares of the stock and accumulate a debt of the amount M.
In period 1, the debt grows to (1 +r) M, but the stocks are evaluated at minimum
Printed by Mathematica for Students
M
S
0
(dS
0
) = dM (1 +r) M, so there is no possibility of losing money. Moreover, there is a posi-
tive possibility that the stocks in period 1 are worth
M
S
0
(uS
0
) = uM > (1 +r) M. So there is a
positive probability of making money. This violates the no arbitrage assumption.
Similarly, if u 1 +r, then one can sell D shares of the stock in period 0 to receive DS
0
.
In period 1, this amount grows to (1 +r)DS
0
. Because (1 +r) DS
0
DuS
0
, the amount (1 +r) DS
0
is always at least as large as the value of the sold stocks. So there is zero possibility of losing
money. Moreover, there is a positive possibility that the sold stocks are evaluated in period 1 at
DdS
0
< (1 +r) DS
0
, so there is a positive possiblity to make money. Again, this violates the no
arbitrage assumption.
The previous 2 paragraphs imply that no arbitrage requires 0 < d < 1 +r < u. We now
show that 0 < d < 1 +r < u is also sufficient to preclude arbitrage.
Suppose an investor borrows money in period 0 to buy D
0
> 0 units of stocks. Then in
period 1, the total worth of his stock and money accounts are
X
1
(H) = D
0
S
1
(H) -(1 +r) D
0
S
0
= D
0
S
0
(u-1 -r) > 0
if the coin toss turns head (with some probability p (0, 1)), or
X
1
(T) = D
0
S
1
(T) -(1 +r) D
0
S
0
= D
0
S
0
(d-1 -r) < 0
if the coin toss turns tail (with probability q = 1 - p (0, 1)). We see that there is a positive
possibility q that the investor loses money regardless of D
0
. So there is no arbitrage for this
investor. Similarly, we can argue that there is no arbitrage for an investor who is a holder of
stocks. We conclude 0 < d < 1 +r < u is indeed sufficient for no arbitrage.
In the general 2 period model, we define a derivative security to be a security that pays some
amount V
1
(H) at time 1 if the coin toss results in head and pays a possibly diffent amount
V
1
(T) at time 1 if the coin toss results in tail. A European call option (the option but not the
obligation to buy a share of stock at some fixed price (aka the strike price) at some specified
point in the future) is a particular kind of derivative security. For a European call option, we
2 1. 1. 110503 - Typed. nb
have V
1
= max S
1
, K, where K is the strike price.
The fundamental question of asset pricing is how much the derivative security is worth
at time zero before we know whether the coin toss results in head or tail. The arbitrage pric-
ing theory approach to the asset-pricing problem is to replicate the derivative by trading in
the stock and money markets.
Suppose we begin with initial wealth X
0
and buy D
0
shares of stock at time 0. Then the
value of our portfolio of stock and money market accounts at time 1 will be
X
1
= D
0
S
1
+(1 +r) (X
0
-D
0
S
0
).
In order to replicate (aka hedge) our derivative, we will solve the simultaneous equations
D
0
S
1
(H) +(1 +r) (X
0
-D
0
S
0
) = X
1
(H) = V
1
(H)
D
0
S
1
(T) +(1 +r) (X
0
-D
0
S
0
) = X
1
(T) = V
1
(T)
for X
0
and D
0
. Simple algebra gives:
D
0
=
V
1
(H)-V
1
(T)
S
1
(H)-S
1
(T)
and X
0
=
1
1+r
[p
V
1
(H) +q
V
1
(T)]
where p
and q
, called the risk neutral probabilities, are defined as

p
=
1
u-d
(1 +r -d) > 0 and q
= 1 - p
=
1
u-d
(u-1 -r).
The formula for D
0
above is called the delta-hedging formula. With X
0
and D
0
as given above,
the investor has emulated the short position in the derivative security. The derivative security
that pays V
1
at time 1 should be priced at
V
0
= X
0
=
1
1+r
[p
V
1
(H) +q
V
1
(T)] .
This formula is called the risk-neutral pricing formula. Neither does the risk-neutral pric-
1. 1. 110503 - Typed. nb 3
ing formula nor the delta-hedging formula depend on the actual probabilities of the coin toss
turning head or tail. Note that they don't depend on the initial stock price S
0
either.
2. Multiperiod Model
We can extend the 2-period model above to a general multiperiod model. For example, the path
of possible stock prices in a 4-period binomial model looks like:
(Figure is from Shreve, Vol I, page 8.) Here S
0
= 4, u = 2, and d =
1
2
. We solve the multiperiod
model recursively, reducing it to the 2-period model, which we have solved above. The details
are in the following theorem.
Theorem. (Theorem 1.2.2 in Shreve, Vol I) (Hedging the multiperiod bino-
mial model)
Consider an N-period binomial asset-pricing model, with 0 < d < 1 +r < u, and with
p
=
1
u-d
(1 +r -d) > 0 and q
= 1 - p
=
1
u-d
(u-1 -r).
4 1. 1. 110503 - Typed. nb
Let V
N
be a random variable (a derivative security paying off at time N) depending on the first
N coin tosses w
1
w
2
w
N
(each w
n
= H or T). Note that all possible values of V
N
are completely
specified by the problem. Define recursively backward in time the sequence of random variables
V
N-1
, V
N-2
, , V
0
by
V
n
(w
1
w
n
) =
1
1+r
[p
V
n+1
(w
1
w
n
H) +q
V
n+1
(w
1
w
n
T)],
so that each V
n
depends on the first n coin tosses w
1
w
n
, where n ranges between N -1 and 0.
Next define
D
n
(w
1
w
n
) =
V
n+1
(w
1
w
n
H)-V
n+1
(w
1
w
n
T)
S
n+1
(w
1
w
n
H)-S
n+1
(w
1
w
n
T)
,
where again n ranges between 0 and N -1. If we set X
0
= V
0
and define recursively forward in
time the portfolio values X
1
, , X
N
by the wealth equation:
X
n+1
(w
1
w
n
w
n+1
) = D
n
(w
1
w
n
) S
n+1
(w
1
w
n
w
n+1
)
+(1 +r) (X
n
(w
1
w
n
) -D
n
(w
1
w
n
) S
n
(w
1
w
n
))
then we will have
X
N
(w
1
w
N
) = V
N
(w
1
w
N
) for all w
1
w
N
.
Proof. See Shreve, Vol 1, page 12.
3. Probability Space
Definition. A probability space is a triple (W, F, P). These symbols are explained as fol-
lows.
(1) W is a nonempty set called the set of outcomes. A single element w of W is called an out-
1. 1. 110503 - Typed. nb 5
come, whereas a subset A of W is called an event.
(2) F is a subset of the powerset of W: F 2
W
. Moreover, F is a s-algebra on W, meaning:
(i) F;
(ii) A F implies A
c
W\ A F;
(iii) a sequence of sets A
1
, A
2
, belongs to F implies |
n=1
A
n
F.
(3) P, called a probability measure, is a real-valued function that that assigns every element
A F with a number P(A), called the probability of A, such that
(i) P(A) [0, 1] for all A F;
(ii) P(W) = 1;
(iii) (Countable Additivity) whenever A
1
, A
2
, is a sequence of disjoint sets in F, then
P||
n=1
A
n
] = _
n=1
P(A
n
).
Lemma. (1) A F, B F imply A|B F and AB F.
(2) If A
1
, A
2
, is a sequence in F, then
n=1
A
n
F.
(3) P(A
c
) = 1 -P(A).
Definition. Let (W, F, P) be a probability space. If a set A F satisfies P(A) = 1. we say that
the event A occurs almost surely.
4. Random Variable
The Borel s-algebra on the real line is the smallest s-algebra containing all open (closed)
sets. Openess and closedness are based on the natural topology on . Notation: sometimes, we
use both symbols ] a, b[ and (a, b) to denote the open interval from a to b.
Definition. Let (W, F, P) be a probability space and let be endowed with the Borel s-alge-
bra B. Then, a random variable X is a function X : (W, F) (, B) such that X is measur-
able, meaning: G B implies X
-1
(G) F.
A random variable X on (W, F, P) induces the probability measure m
X
: B [0, 1], called the
distribution measure of X, defined as
6 1. 1. 110503 - Typed. nb
m
X
(B) = P|X
-1
(B)].
We also define the cumulative distribution function (cdf) F
X
:
F
X
(x) = P(X x).
If we know m
X
, then we can retrieve F
X
because F
X
(x) = m
X
(-, x]. On the other hand, suppose
we know F
X
. Then, noting [x, y] =
n=1
|x -
1
n
, y|, we have
m
X
[x, y] = lim
n
m
X
|x -
1
n
, y| = lim
n
|F
X
(y) -F
X
|x -
1
n
]].
And once the distribution measure m
X
[a, b] is known for every interval [a, b] , it is deter-
mined for every Borel subset of . Therefore, in principle, knowing the cdf F
X
for a random
variable X is the same as knowing its distribution measure m
X
.
Next class: integrals and expectations.
1. 1. 110503 - Typed. nb 7

Organizational Stuffs: Typed by Vinh Nguyen. Date May 18, 2011

Uploaded by

Copyright:

Available Formats

Organizational Stuffs: Typed by Vinh Nguyen. Date May 18, 2011

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Organizational Stuffs: Typed by Vinh Nguyen. Date May 18, 2011

Uploaded by

Copyright:

Available Formats

McGill Summer 2011 - Econ 765 - Lecture 1 - Date May 3, 2011

Typed by Vinh Nguyen. Date May 18, 2011.

, called the risk neutral probabilities, are defined as

You might also like