Slides 3 Elementary Market Model

3: ELEMENTARY MARKET MODEL
Marek Rutkowski
School of Mathematics and Statistics
University of Sydney
Semester 2, 2016
M. Rutkowski (USydney) Slides 3: Elementary Market Model 1 / 36

Single Period Market Model
Only one period is considered.
The beginning of the period is usually set as t = 0.
The end of the period is usually set as t = 1 (or T = 1).
At t = 0, the prices of all assets are known and the investor can
choose the investment.
At t = 1, the prices of all assets are observed and the investor obtains
a payoff corresponding to the current portfolio value.
A single period market model is not realistic, but it allows us to
illustrate important economic principles without suffering from (or
enjoying) sophisticated mathematics.
A full understanding of the features (also drawbacks) of a single
period market model is thus necessary for further developments.

Elementary Market Model M = (B, S)
The investor has initial wealth x at t = 0 and is allowed to invest in

the risk-free asset B (bank account) and the risky asset S (stock).
He purchases φ shares of the stock and invests the remaining funds
in his bank account (or borrows cash from the bank).
Notation:
Sample space: Ω = {ω1 , ω2 }.
Probability measure: P(ω1 ) = p > 0 and P(ω2 ) = 1 − p > 0.
A deterministic interest rate r > −1.
The price of a risky asset at time t is denoted by St .
We assume that S0 > 0 and we set u = S1S(ω0 1 ) and d = S1S(ω0 2 ) .
We suppose that 0 < d < u, so that there are two distinct values of
the future stock price: S1 (ω1 ) = uS0 > S1 (ω2 ) = dS0 .
We do not assume that d < 1 or u > 1 (although d stands for ‘down’
and u stands for ‘up’).

Elementary Market Model
1 1 + r
p S1(ω1) = uS0
S0
1 − p
S1(ω2) = dS0
t = 0 t = 1

Why the Elementary Market Model?
In real life, the stock price mouvements are more complicated than
indicated by the elementary market model.
Hence we do not pretend that the elementary model furnishes a
realistic picture of the stock price fluctuations.
Nevertheless, even in this simplistic framework the random character
of the stock price is already visible and thus the problem of options
pricing is non-trivial.
There are two important reasons why we consider this model:
First, the concept of arbitrage-free pricing of derivative securities can
be clearly explained.
Second, the binomial asset pricing model can be seen as a
concatenation of elementary market models.

Outline
We will examine the following issues:

1 Trading Strategies and Arbitrage-Free Models
2 Replication of Contingent Claims
3 Risk-Neutral Probability Measure
4 Put-Call Parity Relationship
5 Summary of the Elementary Market Model
6 Generalisation of the Elementary Market Model

PART 1
TRADING STRATEGIES AND ARBITRAGE-FREE MODELS

Trading Strategy and Wealth Process
For arbitrary real numbers x and φ, where x is the initial wealth and
φ is the number of shares of stock purchased or sold at time 0, the
pair (x, φ) is called the trading strategy.
The initial wealth (or value) equals
V0 (x, φ) := x = (x − φS0 ) B0 + φS0 .
For any trading strategy (x, φ), the investor obtains at time 1 the
random payoff V1 (x, φ), which satisfies for i = 1, 2
V1 (x, φ)(ωi ) = (x − φS0 ) (1 + r) + φS1 (ωi )
Definition (Value Process)

The wealth process (or value process) of the trading strategy (x, φ)
is given by (V0 (x, φ), V1 (x, φ)) where V0 (x, φ) = x and V1 (x, φ) is
the random variable V1 (x, φ) = (x − φS0 ) (1 + r) + φS1 .
Arbitrage
An essential feature of an efficient market is that for any trading

strategy can turn nothing into something, the investor who adopts
it must also face the risk of loss.
The following definition is thus a crucial step in the arbitrage pricing
methodology.
Definition (Arbitrage)
A trading strategy (x, φ) in the single-period market model is called an
arbitrage opportunity if:
A.1. x = 0, that is, no initial investment is required,
A.2. V1 (x, φ) ≥ 0, that is, no risk of losing money,
A.3. There exists an ωi such that V1 (x, φ)(ωi ) > 0.

Arbitrage-Free Model
Assume that A.2 holds. Then condition A.3 is equivalent to
A.3′ . EP {V1 (x, φ)} > 0, that is, a strictly positive expected payoff.
Definition (Arbitrage-Free Model)

A single-period market model is said to be arbitrage-free if no arbitrage
opportunity exists in the model.
Real markets sometimes exhibit arbitrage, but the forces of supply and
demand take actions to remove it as soon as someone discovers it.
Market model that admit arbitrage are not suitable for our purposes.
Proposition (3.1)
The model M = (B, S) is arbitrage-free if and only if d < 1 + r < u.

Proof of Proposition 3.1
Proof of Proposition 3.1 (⇒).

To prove the ‘only if’ part, we argue by contradiction:
Assume first that d ≥ 1 + r:
At t = 0, the investor borrows the amount S0 of cash on the money
market and buys one share of the stock.
At t = 1, the investor sells the stock and thus receives either uS0 or
dS0 . He also pays back (1 + r)S0 .
Let us now assume that u ≤ 1 + r:
At t = 0, the investor borrows one share of the stock from a broker and
sells it immediately (short-selling). He then invests S0 in the money
market.
At t = 1, the investor obtains (1 + r)S0 from the bank account. He
spends either uS0 or dS0 to buy one share of the stock and returns it
to the broker (hence to the original owner).
In both cases, the investor’s strategy is an arbitrage opportunity.
Hence the proof of the ‘only if’ part is complete.
Proof of Proposition 3.1 (⇐).

The ‘if’ part is also easy to establish:
We start by noting that for x = 0 the equality
V1 (x, φ) = (x − φS0 ) (1 + r) + φS1
becomes
V1 (0, φ) = φ(S1 − (1 + r)S0 ).
More explicitly,
V1 (0, φ)(ω1 ) = φ(S1 (ω1 ) − (1 + r)S0 ) = φS0 (u − (1 + r))

V1 (0, φ)(ω2 ) = φ(S1 (ω2 ) − (1 + r)S0 ) = φS0 (d − (1 + r))
It is thus clear that if d < 1 + r < u, then an arbitrage opportunity

does not exist.
PART 2
REPLICATION OF CONTINGENT CLAIMS

European Options
Definition (European Call and Put Options)

A European call (put) option is a contract which gives the buyer the
right to buy (sell) an asset at a future date T for a predetermined price K.
The underlying asset (for instance, a stock or a stock index), the

maturity date T and the strike price K are given in the contract.
Payoff at T of a European call option:
If ST > K then the holder gets the net payoff ST − K > 0 by
exercising the option, buying the stock at K and selling it at ST .
If ST ≤ K then the holder does not exercise the option and this leads
to the null payoff at T .
Hence the payoff of the call option at time T is
CT = h(ST ) = max (ST − K, 0) = (ST − K)+ .

European Options
Payoff of a European put option:
If ST > K then the holder does not exercise the option and thus the
payoff equals 0.
If ST ≤ K then the holder exercises the option and obtains the net
payoff K − ST > 0 by buying the stock at ST and selling it at K.
Hence the payoff of the put option at time T equals
PT = h(SY ) = max (K − ST , 0) = (K − ST )+ .
European calls and puts are examples of contingent claims. Their

payoffs CT and PT at the expiration date T are random, but they
only depend on the stock price ST and strike K.
We will now address the following general question:
How to select an initial investment x and a trading strategy (x, φ) in
order to obtain the same wealth V1 (x, φ) at time 1 as the payoff
of a given contingent claim X = h(S1 )?
Replication of a Contingent Claim
Definition (Replicating Strategy)

A replicating strategy (or a hedge) (x, φ) for the contingent claim
X = h(S1 ) in the elementary market model M = (B, S) is a trading
strategy which satisfies V1 (x, φ) = h(S1 ), that is,
(x − φS0 ) (1 + r) + φS1 (ω1 ) = h(S1 (ω1 )), (1)

(x − φS0 ) (1 + r) + φS1 (ω2 ) = h(S1 (ω2 )). (2)
Definition (Arbitrage Price)

Assume that the elementary market model M = (B, S) is arbitrage-free.
If (x, φ) is a replicating strategy of a contingent claim then x is called the
arbitrage price (or price) for the claim at t = 0. We write x = π0 (X).

Hedging of a Contingent Claim
Computation of the hedge and (unique) arbitrage price x of a contingent

claim X = h(S1 ):
By subtracting (2) from (1), we find the hedge ratio
h(S1 (ω1 )) − h(S1 (ω2 )) h(uS0 ) − h(dS0 )

φ= = . (3)
S1 (ω1 ) − S1 (ω2 ) (u − d)S0
Equality (3) is called the delta hedging formula.

One can substitute (3) into (1) or (2) in order to compute x.
To derive a convenient representation for x, we introduce the notation
1+r−d
pe := ∈ (0, 1) (4)
u−d
e by: P(ω
and we define the probability P e 1 ) = pe and P(ω
e 2 ) = 1 − pe.

Pricing of a Contingent Claim
e satisfies
One can check that P

S1 1
EPe = [e
pS1 (ω1 ) + (1 − pe) S1 (ω2 )] = S0 . (5)
1+r 1+r
We rewrite (1) and (2) in the following way:

S1 (ω1 ) 1
x+ − S0 φ = h(S1 (ω1 )) (6)
1+r 1+r

S1 (ω2 ) 1
x+ − S0 φ = h(S1 (ω2 )) (7)
1+r 1+r
If we multiply (6) and (7) with pe and 1 − pe, respectively, and add them,
then we obtain

1
x+ [e
pS1 (ω1 ) + (1 − pe) S1 (ω2 )] − S0 φ
1+r
(8)
1
= [e
ph(S1 (ω1 )) + (1 − pe) h(S1 (ω2 ))] .
1+r
Model Completeness
In view of (5), the term with φ vanishes. Hence (8) yields the
following convenient representation for the price x
1
x= [e
ph(S1 (ω1 )) + (1 − pe) h(S1 (ω2 ))] . (9)
1+r
It can be seen from (9) that the price x depends on pe and 1 − pe, but
it is independent of the real-world probabilities p and 1 − p.
Note that equalities (3) and (9) hold for an arbitrary payoff function
h(S1 ). Hence for any contingent claim X we have a unique
replicating strategy and a unique arbitrage price.
Definition (Completeness)
Since all contingent claims in the elementary market model have
replicating strategies, the model is called complete.

PART 3
RISK-NEUTRAL PROBABILITY MEASURE

Risk-Neutral Probability Measure
Definition (Risk-Neutral Probability Measure)

A probability measure Q on the sample space Ω = {ω1 , ω2 } is called a
risk-neutral probability measure (equivalent martingale measure) for
the market model M = (B, S) if Q is equivalent to P and the following
equality holds
S1
EQ = S0 .
1+r
Proposition (3.2)
The risk-neutral probability measure for the model M = (B, S) is
unique and it satisfies Q = Pe if and only if d < 1 + r < u.
If 1 + r ≤ d or u ≤ 1 + r, then no risk-neutral probability exists.

Proof of Proposition 3.2.

It suffices to observe that the equality

S1
EQ = S0
1+r
means that Q(ω1 ) ≥ 0, Q(ω2 ) ≥ 0, Q(ω1 ) + Q(ω2 ) = 1 and
Q(ω1 )S1 (ω1 ) + Q(ω2 )S1 (ω2 ) = (1 + r)S0 .
The last equality above yields

1+r−d e
Q(ω1 ) = = P(ω1 ). (10)
u−d
If d = 1 + r or u = 1 + r then Q given by (10) is well defined, but it
is not equivalent to P since either Q = (1, 0) or Q = (0, 1).
Expected Rates of Return on Traded Assets
Assume that u < 1 + r < d. Then the risk-neutral probability

measure Q = Pe exists and is unique.
The expected rate of return on the savings account B equals

B1 − B0 B1 − B0
EPe (rB ) = EPe = = r.
B0 B0
The expected rate of return on the stock under P e equals r. This

follows from the equality

S1 − S0 (1 + r)S0 − S0
EPe (rS ) = EPe = = r.
S0 S0
The probability measure Pe is the unique probability measure Q under

which the equality EQ (rB ) = EQ (rS ) holds.

Risk-Neutral Valuation Formula
Proposition (3.3)
For any claim X = h(S1 ), the arbitrage price of X at time 0 in the
arbitrage-free elementary market model M = (B, S) satisfies

1 X
π0 (X) = Ee (h(S1 )) = EPe . (11)
1+r P 1+r

We know (see (9)) that the price x satisfies
1
x= [e
ph(S1 (ω1 )) + (1 − pe) h(S1 (ω2 ))] .
1+r
Formula (11) now follows immediately.
We will now give a shorter proof for (11).

It is assumed that u < 1 + r < d. Let (x, φ) be any trading strategy.
From the equality

V1 (x, φ) S1
=x+φ − S0
1+r 1+r
e
and formula (5), we deduce that investing is a ‘fair game’ under P
meaning that for any strategy (x, φ) we have

V1 (x, φ)
EPe = x.
1+r
In particular, if (x, φ) replicates X then V1 (x, φ) = X and thus we

obtain the risk-neutral valuation formula (11) since x = π0 (X).

PART 4
PUT-CALL PARITY RELATIONSHIP

Example: Call and Put Options
Example (3.1)
Consider the elementary market model M = (B, S) with parameters
r = 13 , S0 = 1, u = 2, d = 21 , p = 35 and T = 1.
Recall that the risk-neutral probability measure P e is given as
e 1 ) = pe and P(ω
P(ω e 2 ) = 1 − pe where
1+r−d
pe := .
u−d
e equals
Hence the risk-neutral probability measure P
1 1
e 1 ) = pe = 1 + 3 −
P(ω 2
=
5
, e 2 ) = 1 − pe = 4 .
P(ω
2 − 12 9 9

Example (3.1)
S1 (ω1 ) = 2 C1 (ω1 ) = 1 P1 (ω1 ) = 0

??

p=5/9
e

S0 = ?1 K=1 C1 = (S1 − K)+ P1 = (K − S1 )+
??
??
??
??1−p=4/9
e
??
??
??
?
S1 (ω2 ) = 0.5 C1 (ω2 ) = 0 P1 (ω2 ) = 0.5

Example (3.1 Continued)
The price of the European call with strike price K = 1 equals

1 1
C0 = EPe (CT ) = peCT (ω1 ) + (1 − pe)CT (ω2 )
1+r 1+r
1 3 5 5
= pe(uS0 − K) = × × 1 = .
1+r 4 9 12
The price of the European put with strike price K = 1 equals

1 1
P0 = EPe (PT ) = pePT (ω1 ) + (1 − pe)PT (ω2 )
1+r 1+r
1 3 4 1 1
= (1 − pe)(K − dS0 ) = × × = .
1+r 4 9 2 6

Put-Call Parity
The arbitrage prices at time 0 computed in Example (3.1) satisfy

1 3 1
C0 − P0 = = 1 − = S0 − K. (12)
4 4 1+r
Equality (12) is a special case of the put-call parity.
Proposition (3.4)
The put-call parity can be represented as follows, for t = 0, 1,
Ct − Pt = St − B(t, T )K (13)
where the zero-coupon bond equals B(t, T ) = (1 + r)−(T −t) , so that

B(0, 1) = (1 + r)−1 and B(1, 1) = 1.
It is clear that CT − PT = (ST − K)+ − (K − ST )+ = ST − K.

Equality (13) is thus an easy consequence of Proposition 3.3.
PART 5
SUMMARY OF THE ELEMENTARY MARKET MODEL

Summary: Properties
Let us summarise the properties of the elementary market model:

1 The two-state single-period market model M = (B, S) is
arbitrage-free if and only if d < 1 + r < u.
2 The arbitrage-free property of the model M = (B, S) does not
depend on the real-world probability measure P.
3 An arbitrary contingent claim X can be replicated by means of a
unique trading strategy. Hence the model M = (B, S) is complete.
4 The initial value of a replicating strategy for X is called the arbitrage
price for X and it is denoted as π0 (X).
5 The risk-neutral probability Pe exists and is unique if and only if
d < 1 + r < u (that is, whenever the model M is arbitrage-free). By
e is equivalent to P.
definition, P
6 The arbitrage price π0 (X) of any claim X can be computed from
the risk-neutral valuation formula.
Summary: Theorem
Theorem (3.1 Elementary Market Model)

The elementary market model M = (B, S) is arbitrage-free
if and only if d < 1 + r < u.
Any contingent claim X can be replicated so the model is complete.
Formally, X = V1 (x, φ) for some (x, φ) ∈ R2 .
If d < 1 + r < u then any contingent claim X admits the unique
arbitrage price π0 (X) := x where X = V1 (x, φ).
e for the model M exists and
The risk-neutral probability P
is unique if and only if d < 1 + r < u.
If 1 + r ≤ d or u ≤ 1 + r, then no risk-neutral probability Q exists.
If d < 1 + r < u, then the arbitrage price π0 (X) of any contingent
claim X satisfies
X
π0 (X) = EPe .
1+r
PART 6
GENERALISATION OF THE ELEMENTARY MARKET MODEL

Generalisation of the Elementary Market Model
We generalise the elementary market model by postulating that:

1 We still deal with two primary traded assets: B and S.
2 The sample space Ω = {ω , ω , . . . , ω } where k ≥ 3.

1 2 k
3 Hence S = (S (ω ), . . . , S (ω )) where we assume that
1 1 1 1 k
S1 (ωk ) < S1 (ωk−1 ) < · · · < S1 (ω2 ) < S1 (ω1 ).

4 The model is arbitrage-free if and only if
S1 (ωk ) < S0 (1 + r) < S1 (ω1 ).
5 The risk-neutral probability measure Q exists if and only if
S1 (ωk ) < S0 (1 + r) < S1 (ω1 ), but it is not unique.
6 The model is incomplete: no replicating strategy exists for some
contingent claims X = (X(ω1 ), . . . , X(ωk )).
7 We will not examine this model in detail, since it can be seen
as a special case of a general single-period market model.
Generalisation of the Elementary Market Model
Theorem (3.2 Generalised Elementary Market Model)

The generalised elementary market model M = (B, S) with k ≥ 3
is arbitrage-free if and only if S1 (ωk ) < S0 (1 + r) < S1 (ω1 ).
Only some (but not all) contingent claims X can be replicated,
that is, are attainable. Hence M = (B, S) is incomplete.
If S1 (ωk ) < S0 (1 + r) < S1 (ω1 ), then any attainable claim X has
the unique arbitrage price π0 (X).
The risk-neutral probability Q for the model M exists if and only if
S1 (ωk ) < S0 (1 + r) < S1 (ω1 ). It is not unique.
If S1 (ωk ) < S0 (1 + r) < S1 (ω1 ), then the arbitrage price π0 (X) of
any attainable claim X satisfies for any risk-neutral probability Q

X
π0 (X) = EQ .
1+r

Slides 3 Elementary Market Model

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Slides 3 Elementary Market Model

Загружено:

Авторское право:

Доступные форматы

3: ELEMENTARY MARKET MODEL

M. Rutkowski (USydney) Slides 3: Elementary Market Model 1 / 36

M. Rutkowski (USydney) Slides 3: Elementary Market Model 2 / 36

The investor has initial wealth x at t = 0 and is allowed to invest in

M. Rutkowski (USydney) Slides 3: Elementary Market Model 3 / 36

M. Rutkowski (USydney) Slides 3: Elementary Market Model 4 / 36

M. Rutkowski (USydney) Slides 3: Elementary Market Model 5 / 36

We will examine the following issues:

M. Rutkowski (USydney) Slides 3: Elementary Market Model 6 / 36

TRADING STRATEGIES AND ARBITRAGE-FREE MODELS

M. Rutkowski (USydney) Slides 3: Elementary Market Model 7 / 36

V0 (x, φ) := x = (x − φS0 ) B0 + φS0 .

V1 (x, φ)(ωi ) = (x − φS0 ) (1 + r) + φS1 (ωi )

Definition (Value Process)

An essential feature of an efficient market is that for any trading

M. Rutkowski (USydney) Slides 3: Elementary Market Model 9 / 36

Assume that A.2 holds. Then condition A.3 is equivalent to

Definition (Arbitrage-Free Model)

M. Rutkowski (USydney) Slides 3: Elementary Market Model 10 / 36

Proof of Proposition 3.1 (⇒).

Proof of Proposition 3.1 (⇐).

V1 (x, φ) = (x − φS0 ) (1 + r) + φS1

V1 (0, φ)(ω1 ) = φ(S1 (ω1 ) − (1 + r)S0 ) = φS0 (u − (1 + r))

It is thus clear that if d < 1 + r < u, then an arbitrage opportunity

REPLICATION OF CONTINGENT CLAIMS

M. Rutkowski (USydney) Slides 3: Elementary Market Model 13 / 36

Definition (European Call and Put Options)

The underlying asset (for instance, a stock or a stock index), the

CT = h(ST ) = max (ST − K, 0) = (ST − K)+ .

M. Rutkowski (USydney) Slides 3: Elementary Market Model 14 / 36

European calls and puts are examples of contingent claims. Their

Definition (Replicating Strategy)

(x − φS0 ) (1 + r) + φS1 (ω1 ) = h(S1 (ω1 )), (1)

Definition (Arbitrage Price)

M. Rutkowski (USydney) Slides 3: Elementary Market Model 16 / 36

Computation of the hedge and (unique) arbitrage price x of a contingent

h(S1 (ω1 )) − h(S1 (ω2 )) h(uS0 ) − h(dS0 )

Equality (3) is called the delta hedging formula.

M. Rutkowski (USydney) Slides 3: Elementary Market Model 17 / 36

M. Rutkowski (USydney) Slides 3: Elementary Market Model 19 / 36

RISK-NEUTRAL PROBABILITY MEASURE

M. Rutkowski (USydney) Slides 3: Elementary Market Model 20 / 36

Definition (Risk-Neutral Probability Measure)

M. Rutkowski (USydney) Slides 3: Elementary Market Model 21 / 36

Proof of Proposition 3.2.

means that Q(ω1 ) ≥ 0, Q(ω2 ) ≥ 0, Q(ω1 ) + Q(ω2 ) = 1 and

Q(ω1 )S1 (ω1 ) + Q(ω2 )S1 (ω2 ) = (1 + r)S0 .

The last equality above yields

Assume that u < 1 + r < d. Then the risk-neutral probability

The expected rate of return on the stock under P e equals r. This

The probability measure Pe is the unique probability measure Q under

M. Rutkowski (USydney) Slides 3: Elementary Market Model 23 / 36

Proof of Proposition 3.3.

Proof of Proposition 3.3.

In particular, if (x, φ) replicates X then V1 (x, φ) = X and thus we

M. Rutkowski (USydney) Slides 3: Elementary Market Model 25 / 36

PUT-CALL PARITY RELATIONSHIP

M. Rutkowski (USydney) Slides 3: Elementary Market Model 26 / 36

M. Rutkowski (USydney) Slides 3: Elementary Market Model 27 / 36

S1 (ω1 ) = 2 C1 (ω1 ) = 1 P1 (ω1 ) = 0

M. Rutkowski (USydney) Slides 3: Elementary Market Model 28 / 36

Example (3.1 Continued)

The price of the European call with strike price K = 1 equals