University of Geneva

Multi-Period Discrete-Time Model of Financial Markets

1.44
1.2
1 1.2
1
1
t=0 t=1 t=2

1 / 59
University of Geneva

Dynamic Model
One-period model is often inadequate.
E.g., we want to find time−0 price of a call option expiring at time T .
Stock price, hence option payoff (= (ST − K)+ ), takes > 2 values.
⇒ We need > 2 states.
One-period model with bond & stock.
⇒ Market is incomplete (since # assets = 2 < # states).
⇒ Cannot price (most) options.
Yet option price “should” depend on bond & stock only.

Introduce multi-period model.

To develop and study this model, we price a call option.

2 / 59
University of Geneva

Methods of Option Pricing

Continuous-time analytic technique:
For simple instruments, can derive equations describing price.
In some cases, can solve equations in closed form, e.g., Black-Scholes.
Closed-form solutions can be calculated very quickly.
Tree/Lattice pricing technique:
Under standard assumptions, approximates analytic price arbitrarily.
Simple extension of one-period model.
More flexible than analytic: handles discrete dividends & early exercise.
Slower than analytic technique.
Finite difference technique:
Analytic technique ⇒ PDEs describing price. If 6 ∃ closed-form solution,
approximate price obtained by finite difference approximation of PDEs.
Similarly flexible to binomial & trinomial trees.
Monte Carlo technique:
Generate random price paths of underlying asset (so option payoffs).
Discounted expectation of random payoffs is estimate for option price.
More flexible: handles several sources of uncertainty, (e.g., real options).
Very slow.
3 / 59
 University of Geneva

Tree Model - Illustration

The evolution of stock price could be (approximately) represented by trees:
Binomial Model (Stock Price)

5654 Several features and questions:

5370
Binomial versus Trinomial:
5182
5100
Trinomial useful for instruments
depending on 2 underlying assets.
5182
4921 Recombining tree:
4749
Important for calculation speed.
t=0 t=1 t=2
Trinomial Model(Stock Price) 5569 Model Calibration:
5330 5330 Distribution model returns matches
5100 5100 5100 distribution of real returns.
4881 4881
4671

4 / 59
 University of Geneva

Finite-state, Multi-period model

N + 1 dates: date t ∈ {0, 1, . . . , N }.

Note the N + 1 dates define N periods.

Finite number of states of the world

Price of asset at t = 0 is known, but price at t > 0 is uncertain.

Additional Assumptions
Units of assets can be sub-divided for sale/purchase
No transaction costs (e.g., no bid/ask spread).

That is, setup is almost identical to one-period model.

5 / 59
 University of Geneva

Binomial Model of Stock Prices

Distinguish between true return and model return.
True return is the return of the stock we are trying to model. Assume:
IID (independent and identically distributed) ∀ interval lengths,
i.e., are independent and have same distribution function, and
Log-normal, i.e., ln R̃ ∼ N µ, σ 2 .

Model return is the model of the true returns. Assume:

Period returns are IID.
I.e., letting Rt be return from date t − 1 to date t, it is assumed that
Rt and Rs are independent and have the same distribution, ∀t 6= s.
They take two possible values: Ru , Rd ,
P Rt = Ru |Ft−1 = p and P Rt = Rd |Ft−1 = 1 − p.

Note that assumption of 2 possible returns implies that tree recombines.

⇒ N −period model yields N + 1 rather than 2N payoffs at date N .
⇒ 1 + 2 + . . . + N = N (N2+1) rather than 20 + 21 + . . . + 2N −1 = 2N − 1
“operations” needed to calculate option price.
6 / 59
University of Geneva

Binomial Model - Example

Consider the non-recombining representation of the binomial model:
Binomial Model (Stock Prices) State
5954.64 uuu
5654.93
5457.00 uud
5370.30
5457.00 udu
5182.34
5000.96 udd
5100.00
5457.00 duu
5182.34
5000.96 dud
4921.50
5000.96 ddu
4749.25
4583.02 ddd
t=0 t=1 t=2 t=3

7 / 59
University of Geneva

Interest Rate and Compounding

Want to hedge call using bond & stock ⇒ Must talk about interest rates.
Compound interest arises when interest is added to the principal,
i.e., interest already earned also earns interest.
E.g., interest 5% over 6 months compounded every 6 months means:
6−month loan yields 1 + 0.05 of principal.
2
12−month loan yields (1 + 0.05) of principal, because
at end of first 6 months interest of 0.05 is added to principal, and
at end of second 6 months, loan yields 1 + 0.05 of new principal.
3
18−month loan yields (1 + 0.05) of principal.
Let n = # compounding periods/year, t = loan length in years, and
r = n×interest earned in a compounding period (nominal annual rate).

nt
The loan yields 1 + nr of principal.
Keeping r constant and letting n → ∞, get continuous compounding.

nt
The loan yields lim 1 + nr = ert of principal.
n→∞
8 / 59
University of Geneva

Interest Rate and Compounding (continued)

Assume interest R (t) over t years, and continuous compounding.
Nominal annual rate (instantaneous rate of return) r is given by
1
ert = R (t), hence r = t ln R (t).
1 t0
0 t0
Interest over t0 years is R (t0 ) = ert = e t ln R(t) = R (t) t .
t0
If t is integer, then we could also think as follows:
t0
R (t0 ) = R (t) · R (t) · . . . · R (t), hence R (t0 ) = R (t) t .
| {z }
t0
t times

1


Example: Let Rf (1) = 1.04 be risk-free return over a year. What is Rf 12 ?
1/12
1
= Rf (1) 1 = 1.041/12 = 1.0033.


Rf 12

9 / 59
 University of Geneva

Calibration of Model Returns

Must calibrate model stock return distribution, so it matches true one.
Two important observations for true returns:
True returns over ∆t are IID ⇒ log true returns over ∆t are IID, b/c:
FX,Y (x, y) = P (X ≤ x, Y ≤ y) = P (X ≤ x) P (Y ≤ y)
Gln X,ln Y (x, y) = P (ln X ≤ x, ln Y ≤ y) = P (X ≤ ex , Y ≤ ey )
= P (ln X ≤ x) P (ln Y ≤ y) .

So, E, V ar of log true returns increase linearly with time. For V ar:
 
1/∆t
h i X 1 h i
V ar lnR̃ (1) = V ar  lnR̃ ((∆t)i ) = V ar lnR̃(∆t) .
i=1
∆t

Example: Let annual true return be ln R̃ (1) ∼ N 0.096, 0.1522 .

Then monthly return is ln R̃ (1/12) ∼ N 0.008, 0.0442 .

(
p ln Ru + (1 − p) ln Rd = 0.008
Let p = 0.5. Solve 2 2
p (ln Ru ) + (1 − p) (ln Rd ) = 0.0442 + 0.0082
to find Ru = 1.053, Rd = 0.965.
10 / 59
 University of Geneva

True versus Model Return Distribution

Matlab can generate following histograms (pdf estimates) of distributions:
1200 2500

1000
2000

800
1500

600

1000
400

500
200

0 0
0.8 1 1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6 1.8 2

draws = 10000;
R12True = exp(0.096+0.152*randn(draws,1));
hist(R12True,30);
R1=[1.053; 0.965];
for i=1:draws, y=mnrnd(1,[0.5 0.5],12); R12(i) = prod(y*R1); end;
hist(R12,30);

11 / 59
 University of Geneva

True versus Model Return Distribution

Also compare with the result of a different calibration:
The left figure is with calibration p = 0.5, Ru = 1.053, Rd = 0.965, and
the right figure is with calibration p = 0.25, Ru = 1.088, Rd = 0.983.

2500 3000

2500
2000

2000
1500

1500

1000
1000

500
500

0 0
0.8 1 1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6 1.8 2

12 / 59
University of Geneva

Hedging in Dynamically Complete Markets

Stock Price Model (FTSE 100) 5954.64

×1.053
5654.93 ×0.965
×1.053
5370.30 ×0.965
5457.00
×1.053 ×1.053
5100.00 ×0.965
5182.34 ×0.965
×1.053
4921.50 ×0.965
5000.96
×1.053
4749.25 ×0.965

4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th

13 / 59
 University of Geneva

Representation of stock prices

Consider stock with:
Initial price S0 = 5100, and
Monthly return is ln R̃ (1/12) ∼ N 0.008, 0.0442 .

So for p = 0.5, calibration yields Ru = 1.053, Rd = 0.965.

Stock Price Model (FTSE 100) 5954.64

×1.053
5654.93 ×0.965
×1.053
5370.30 ×0.965
5457.00
×1.053 ×1.053
5100.00 ×0.965
5182.34 ×0.965
×1.053
4921.50 ×0.965
5000.96
×1.053
4749.25 ×0.965

4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th

14 / 59
 University of Geneva

Intrinsic Value of a Call Option

Consider 5% out-of-the-money, 3-month call option on this stock.
5% out-of-the-money ⇒ K = 5100 × 1.05 = 5355.
At maturity, C = max {S − K, 0}.
Stock Price 5954.64
5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Call Option Value 599.64
?
? 102.00
? ?
? 0.00
?
0.00

15 / 59
 University of Geneva

Static Hedging Strategy

Let annual risk-free return be Rf (1) = 1.04. Then:
1-month risk-free return is Rf (1/12) = 1.041/12 = 1.0033.
3-month risk-free return is Rf (1/4) = 1.041/4 = 1.0099.
Static hedge (time-0 portfolio that uses bond & stock to “hedge” option):
   
1.0099 5955 600
1.0099 5457 102
We want to solve Ax = b, where A =  , b =  .
1.0099 5001 0
1.0099 4583 0
Note that A·1 is return of bond while A·2 is payoff of stock.
⇒ x1 is money invested in bond, while x2 is units of stock bought.
Can also assume that unit price of bond = 1, hence payoff = Rf .
Incomplete market, so candidate perfect hedge is x̂ = (A0 A)−1 A0 b.
 
  472
−2022  266 
We find x̂ = , but Ax̂ =  , so x̂ isn’t perfect hedge.
0.422 70 
−107
16 / 59
 University of Geneva

Static Hedging Strategy

     
599.64 1.0099 5954.64
102.00 1.0099 5457.00
Graphical illustration of hedging 
 0.00 
 using 
1.0099
 and 5000.96:
 

0.00 1.0099 4583.02

| {z } | {z } | {z }
option payoff safe return stock price
Stock replicates multiples of stock price ⇒ Straight line through origin.
Bond replicates multiples of safe payoff ⇒ Straight horizontal line.
Stock & bond replicate linear combination of two ⇒ Any straight line.
stock portfolio
Payoff

bond portfolio

option

best hedge

K Stock Price

17 / 59
University of Geneva

Dynamic Hedging Strategy

Look at one-period sub-models:
∃ two states with two securities ⇒ Complete market.
At each node, find portfolio replicating value of option in next period.
Complete market ⇒ ∃ Replicating portfolio ⇒ Dynamic hedging works.
Work backwards from the end.
Stock Price 5954.64
5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
?
? 102.00
? ?
? 0.00
?
0.00
18 / 59
University of Geneva

Computing the One-period Hedge at t = 2

     
1.0033 5954.64 599.64 −5337.39
Solve Ax = b, A = , b= , to get x = .
1.0033 5457.00 102.00 1

That is, borrow 5337.39 from the bank, and buy 1 unit of the stock.
So, cost of replicating portfolio is −5337.39 + 1 · 5654.93 = 317.54.
LOOP ⇒ Cost of hedging portfolio = “Price” of call at that node.
Stock Price 5954.64
5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
? 102.00
? ?
? 0.00
?
0.00

19 / 59
University of Geneva

Computing the One-period Hedge at t = 2

     
1.0033 5457.00 102.00 −1115
Solve Ax = b, A = , b= , to get x = .
1.0033 5000.96 0 0.224

That is, borrow 1115 from the bank, and buy 0.224 units of the stock.
So, cost of replicating portfolio is −1115 + 0.224 · 5182.34 = 44.25.
Stock Price 5954.64
5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
? 102.00
? ?
? 0.00
?
0.00

20 / 59
University of Geneva

Computing the One-period Hedge at t = 2

Stock Price 5954.64

5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
? 102.00
? 44.25
? 0.00
?
0.00

Nothing to hedge ⇒ Don’t buy stock or borrow from bank.

⇒ Value of option = 0.

21 / 59
University of Geneva

Hedging at t < 2

At t < 2 we don’t directly replicate the option payoff at t = 3.

We replicate the amount of money needed at each node at t + 1, so

we can buy the replicating portfolio that at t = 3 yields option payoff.
At each node at t = 1, replicate amount of money needed at following
nodes at t = 2, so we can buy replicating portfolios found above.
⇒ At each node at t = 1 find portfolio eventually yielding option payoff.
This is the value of the option at t = 1, at each node.
At t = 0, replicate amount of money needed at nodes at t = 1, so we
can buy portfolios found above.
⇒ At t = 0 find portfolio eventually yielding option payoff.
This is the value of the option at t = 0.

22 / 59
 University of Geneva

General Solution of one-period Hedging

   
R St Ru C
At each node of t solve Ax = b, with A = f , b = t,u , i.e.:
Rf St Rd Ct,d

bankt Rf + deltat St Ru = Ct,u

bankt Rf + deltat St Rd = Ct,d .
The solution is:
Ct,u − Ct,d
deltat =
St Ru − St Rd
Ct,d St Ru − Ct,u St Rd Ct,d Ru − Ct,u Rd
bankt = = .
(St Ru − St Rd ) Rf (Ru − Rd ) Rf
LOOP: No-arbitrage value of C is cost of portfolio, bankt + deltat × St :
Ru − Rf 1 Rf − Rd 1
Ct,d · + Ct,u · .
Ru − Rd Rf Ru − Rd Rf
| {z } | {z }
state price of d state price of u
Check these state prices are same you get from S = A0 ψ (or 1 = R0 ψ).
23 / 59
University of Geneva

Hedging at t = 1
Stock Price 5954.64
5654.93
5370.30 5457.00
5100.00 5182.34    
4921.50 5000.96 St Ru 5654.93
4749.25 =
4583.02 St Rd 5182.34
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
? 102.00
? 44.25    
? 0.00 Ct,u 317.54
0.00 =
0.00 Ct,d 44.25

317.54−44.25
delta = 5654.93−5182.34 = 0.578 bank = −2943

option value = 162.66

24 / 59
University of Geneva

Hedging at t = 1

Stock Price 5954.64

5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
162.66 102.00
? 44.25
? 0.00
0.00
0.00

option value = 19.19

25 / 59
University of Geneva

Hedging at t = 0

Stock Price 5954.64

5654.93
5370.30 5457.00
5100.00 5182.34
4921.50 5000.96
4749.25
4583.02
t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th
Value of Replicating Portfolio 599.64
317.54
162.66 102.00
? 44.25
19.19 0.00
0.00
0.00

option value = 81.36

26 / 59
 University of Geneva

Self-financing trading strategy

A strategy such that you don’t need to inject money, except at t = 0.

Current portfolio value equals initial investment plus trading gains.

Assuming no dividends, a trading strategy {xt } is self-financing if:

t
S0t xt = S00 x0 + (Ss − Ss−1 )0 xs−1 , ∀t,
P

s=1
or equivalently
S0t xt = S0t xt−1 , ∀t.

S and x are stochastic processes, so values at t depend on state.

⇒ The equalities need to hold for all states.

27 / 59
 University of Geneva

Self-financing trading strategy

Stock Price 5954.64 Value of the Replicating Portfolio 599.64
5654.93 317.54
5370.30 5457.00 162.66 102.00
5100.00 5182.34 81.36 44.25
4921.50 5000.96 19.19 0.00
4749.25 0.00
4583.02 0.00
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 599.64 Number of Shares 0.000
-5337.4 1.000
-2942.9 102.00 0.578 0.000
-1548.9 -1114.9 0.320 0.224
-483.6 0.00 0.102 0.000
0.00 0.000
0.00 0.000

E.g., assume true ω = udu, i.e., price goes Up, Down, and Up. Then:
S01 x1 = S01 x0 : −2942.9+0.578 · 5370.3 ≈ −1548.9 · 1.0033+0.320 · 5370.3
S02 x2 = S02 x1 : −1114.9+0.224 · 5182.3 ≈ −2942.9 · 1.0033+0.578 · 5182.3
S03 x3 = S03 x2 : 102 ≈ −1114.9 · 1.0033+0.224 · 5457.0
28 / 59
University of Geneva

Dynamic Option Hedging - Example

We want to price a call option using a 3−period model. Assume:
1


∃ bond with semi-annual return Rf 2 = 1.03,


∃ stock with IID annual log return ln R̃ (1) ∼ N 0.12, 0.242 , and
initial price S0 = 10,
Option on stock has strike price K = 10.1, and expires in 3 months.

Solution:
1 Calibrate model (find Ru , Rd assuming, e.g., p = 1/2; also find Rf ).
2 Construct stock price tree.
3 Write down option payoff at expiration.
4 Hedge the payoff going backwards.

Write down:
Amount of money deposited in the bank.
Number of shares (option delta).
Cost of hedging (value of the replicating portfolio).
29 / 59
University of Geneva

Dynamic Option Hedging - Example (continued)

Stock Price Value of the Replicating Portfolio

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

30 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

1. Calibrate model:
Remember, IID true returns ⇒ E and V ar increase linearly with time:
h i t0 h i h i t0 h i
E ln R̃ (t0 ) = E ln R̃ (t) , V ar ln R̃ (t0 ) = V ar ln R̃ (t)
t t
1



So, ln R̃ 12 ∼ N 0.01, 0.072 .
1
Let p = 2 and solve:
 h
i
p ln Ru + (1 − p) ln Rd 1
= E ln R̃ 12
h i  h i2
p (ln R )2 + (1 − p) (ln R )2 = V ar ln R̃ 1
+ E ln R̃ 1

u d 12 12

h r
1

i h
1

i
to find ln Ru = E ln R̃ 12 + V ar ln R̃ 12 .
So Ru = 1.083, Rd = 0.942.
1

1

1/12
1/2
1
Remember, we calculate Rf 12 = Rf 2 = 1.03 6 = 1.005.
31 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

2. Construct the stock-price tree by starting from S0 = 10 and using

St = St−1 × Ru for “up” nodes and St = St−1 × Rd for “down” nodes.

3. Write down option pay-off at expiration. It is max {ST − K, 0}.

4. Hedge the payoff going backwards, using the formulae:

Ct,u − Ct,d
delta =
St Ru − St Rd
Ct,d St Ru − Ct,u St Rd
bank =
(St Ru − St Rd ) Rf
value of option at t = bankt + deltat × St .

32 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

33 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

10

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

34 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

10.83
10

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

35 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

10.83
10
9.42

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

36 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

11.73
10.83
10 10.20
9.42

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

37 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price Value of the Replicating Portfolio

11.73
10.83
10 10.20
9.42
8.87

t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3

Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

38 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price 12.70 Value of the Replicating Portfolio

11.73
10.83 11.05
10 10.20
9.42 9.61
8.87
8.36
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

39 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73
10.83 11.05 0.95
10 10.20
9.42 9.61 0
8.87
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account Number of Shares

40 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73
10.83 11.05 0.95
10 10.20
9.42 9.61 0
8.87
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
0.95 0
0 0
0 0

41 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 12.70 2.60 −10.05
Solve Ax = b with A = ,b= , to get x = .
1.005 11.05 0.95 1
Cost of replicating portfolio is −10.05 + 1 · 11.73 = 1.68.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.95
10 10.20
9.42 9.61 0
8.87
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
0.95 0
0 0
0 0

42 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 11.05 0.95 −6.31
Solve Ax = b with A = ,b= , to get x = .
1.005 9.61 0 0.66
Cost of replicating portfolio is −6.31 + 0.66 · 10.20 = 0.42.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.95
10 10.20 0.42
9.42 9.61 0
8.87
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
0.95 0
-6.31 0.66
0 0
0 0

43 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 9.61 0 0
Solve Ax = b with A = ,b= , to get x = .
1.005 8.36 0 0
Cost of replicating portfolio is 0.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.95
10 10.20 0.42
9.42 9.61 0
8.87 0
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
0.95 0
-6.31 0.66
0 0
0 0
0 0

44 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 11.73 1.68 −7.94
Solve Ax = b with A = ,b= , to get x = .
1.005 10.20 0.42 0.82
Cost of replicating portfolio is −7.94 + 0.82 · 10.83 = 0.94.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.94 0.95
10 10.20 0.42
9.42 9.61 0
8.87 0
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
-7.94 0.95 0.82 0
-6.31 0.66
0 0
0 0
0 0

45 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 10.20 0.42 −2.79
Solve Ax = b with A = ,b= , to get x = .
1.005 8.87 0 0.32
Cost of replicating portfolio is −2.79 + 0.32 · 9.42 = 0.22.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.94 0.95
10 10.20 0.42
9.42 9.61 0.22 0
8.87 0
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
-7.94 0.95 0.82 0
-6.31 0.66
-2.79 0 0.32 0
0 0
0 0

46 / 59
University of Geneva

Dynamic Option Hedging - Example - Solution

     
1.005 10.83 0.94 −4.57
Solve Ax = b with A = ,b= , to get x = .
1.005 9.42 0.22 0.51
Cost of replicating portfolio is −4.57 + 0.51 · 10 = 0.53.

Stock Price 12.70 Value of the Replicating Portfolio 2.60

11.73 1.68
10.83 11.05 0.94 0.95
10 10.20 0.53 0.42
9.42 9.61 0.22 0
8.87 0
8.36 0
t=0 t=1 t=2 t=3 t=0 t=1 t=2 t=3
Jan 28th Feb 28th Mar 28th Apr 28th Jan 28th Feb 28th Mar 28th Apr 28th
Bank Account 2.60 Number of Shares 0
-10.05 1
-7.94 0.95 0.82 0
-4.57 -6.31 0.51 0.66
-2.79 0 0.32 0
0 0
0 0

47 / 59
University of Geneva

Examples

48 / 59
University of Geneva

Example - Asian Option

3-period binomial model, each period corresponding to half a year.
Bond: continuously compounded interest; instantaneous rate r = 0.446.
Stock: price S0 = 4 at t = 0, return Ru = 2 or Rd = 21 each period.
Consider Asian call option that expires at t = 3 and has strike K = 4.
It’s like European call option, except payoff at t = 3 is
max {Y3 − K, 0} rather than max {S3 − K, 0}, where
k
1 P
for 0 ≤ k ≤ 3, Yk = k+1 Si is average stock price from t = 0 to k.
i=0
We are asked the following:
a. For each one-period sub-model, calculate the state prices.
b. Draw tree containing stock prices and payoffs of option at maturity.
c. Using the state prices, find option’s value at each t = 2 node.
d. Find option’s price at t = 0.
e. Set up at t = 0 self-financing strategy that replicates options’ payoff.
49 / 59
 University of Geneva

Solution - (a)

a. Find 6-month risk-free return: Rf = ert = 1.25, with r = 0.446, t = 0.5.

Letting S, A, R, and ψ denote as usual in each 1−period sub-model,
−1
LOOP ⇒ S = A0 ψ, or equivalently 1 = R0 ψ, so ψ = (R0 ) 1.
0 −1
Solve ψ = (R ) 1 to find state prices (for up and down states):
1 Rf − Rd
ψu = ·
Rf Ru − Rd
1 Ru − Rf
ψd = · ,
Rf Ru − Rd

which yields ψu = 0.4 and ψd = 0.4.

Note Rf , Ru , Rd are the same in all sub-models, so ψu , ψd are same.

50 / 59
 University of Geneva

Solution - (b)
b. To draw price tree: Start with S0 = 4 at unique node at t = 0.
Multiply by 2 to get price after “up” branch.
Divide by 2 to get price after “down” branch.
Payoffs at maturity are: C3 (uuu) = max {60/4 − 4, 0} = 11.0,
C3 (uud) = max {36/4 − 4, 0} = 5.0, etc.
In particular, we have
Stock Prices Payoff
32 11.0
16
8 5.0
8
8 2.0
4
2 0.5
4
8 0.5
4
2 0.0
2
2 0.0
1
0.5 0.0
t=0 t=1 t=2 t=3

51 / 59
 University of Geneva

Solution - (c, d)
c. Price of option at each node of t = 2 is given by

C2 = ψu C3u + ψd C3d .

d. In general, price of option at each node at t is given by

u d
Ct = ψu Ct+1 + ψd Ct+1 .

So we can calculate price C0 recursively.

Using the values we have calculated, we find

C0 = 1.216.

52 / 59
 University of Geneva

Solution - (e)

e. At t = 0 want strategy paying 2.96 in up and 0.08 in down node at t = 1.

Let    
1.25 2.00 2.96
R= ,b =
1.25 0.50 0.08
be returns matrix for 1−period sub-model at t = 0 and the focus asset.

We are looking for vector x such that Rx = b.

Market is complete, so find replicating portfolio as

 
−0.704
x = R−1 b = .
1.920

That is, at t = 0 borrow 0.704 from bank and buy 1.920 worth of stock.

53 / 59
University of Geneva

Example - American Barrier Put Option

3-period binomial model, each period corresponding to a month.
Bond: continuously compounded interest; instantaneous rate r = 0.05.
Stock: price S0 = 100 at t = 0, and rises/falls 10 each period.
American barrier put option expiring at t = 3 with strike K = 100.
Like European put except
Early exercise is possible
Option is worthless if St ≥ 110 for some t ≤ 3.

We are asked the following:

a. Draw price lattice with payoffs of option at maturity.
b. For each one-period model, calculate the state prices.
c. Find option’s value at each t = 2 node, assuming no early exercise.
d. At which t = 2 nodes would you want to exercise option early?
So what is the option’s value at each t = 2 node?
e. Compute the option’s value recursively, to find its price at t = 0.
54 / 59
 University of Geneva

Solution - (a)

Stock Prices Payoff

t=0 t=1 t=2 t=3

55 / 59
 University of Geneva

Solution - (a)

Stock Prices Payoff

130 0
120
110 0
110
110 0
100
90 0
100
110 0
100
90 10
90
90 10
80
70 30
t=0 t=1 t=2 t=3

56 / 59
 University of Geneva

Solution - (b)
1
b. Find 1-month risk-free return: Rf = ert = 1.004, with r = 0.05, t = 12 .
Letting S, A, R, and ψ denote as usual in each 1−period sub-model,
−1
LOOP ⇒ S = A0 ψ, or equivalently 1 = R0 ψ, so ψ = (R0 ) 1.
0 −1
Solve ψ = (R ) 1 to find state prices (for up and down states):
1 Rf − Rd
ψu = ·
Rf Ru − Rd
1 Ru − Rf
ψd = · .
Rf Ru − Rd
E.g., for the top node at t = 2, we have
1 1.004 − 110
120
ψu = · 130 110 = 0.522
120 − 120
1.004
130
1 120 − 1.004
ψd = · 130 110 = 0.474.
120 − 120
1.004
Note: Ru , Rd are NOT same in all sub-models, so ψu , ψd NOT same.
57 / 59
 University of Geneva

Solution - (c, d)

c. Price of option at each node of t = 2 is given by

P2 = ψu P3u + ψd P3d .

In particular, from top to bottom node, we find prices 0, 0, 4.78, 19.6.

d. Only node at t = 2 in which you might want to exercise is bottom.

There, payoff from exercising it at t = 2 would be 20, i.e., > 19.60

Thus, you prefer to exercise the option.
So, allowing for early exercise, prices at t = 2 are 0, 0, 4.78, and 20.

58 / 59
 University of Geneva

Solution - (e)

e. Thus, the price of the option at each node at t = 1 is given by

P1u = 0
P1d = ψu × 4.78 + ψd × 20 = 0.516 × 4.78 + 0.480 × 20 = 12.07.

Early exercise isn’t profitable at either node.

Thus, price of option at t = 0 is given by

P0 = ψu × 0 + ψd × 12.07 = 0.518 × 0 + 0.478 × 12.07 = 5.77.

59 / 59