2 The binomial pricing model

2.1 Options and other derivatives

A derivative security is a financial contract whose value depends on some underlying asset
like stock, commodity (gold, oil) or currency. The underlying asset has a market value
St which changes randomly with time t. Most of the derivatives are fixed-term contracts,
which expire at certain time T called maturity. Next are some basic examples.

• A forward contract is an agreement between two parties whereby one contracts to

buy from the other a given asset (stock) at specified price (forward price) K at
delivery date T .

• A (European) call option is a contract which confers the holder the right, but not
the obligation, to buy a given asset (stock) at prescribed price, called strike price K
at time T . The party issuing the contract (‘writer’ of the option, typically a bank)
is obliged to sell the asset should the holder decide to exercise her right.

• A (European) put option is a contract which confers the holder the right, but not
the obligation, to sell a given asset at prescribed price, called strike price K at time
T . The party issuing the contract (‘writer’ of the option) is obliged to buy the asset
should the holder decide to exercise her right.

Speaking of European type of option we mean that the option can be exercised only at
the maturity, but not earlier.
For instance, on 25/09/13 you possess a call option for buying a share of ABZ for 220p
on 25/09/2014. Suppose the market price of ABZ share on 25/09/2014 will be 250p. In
this case you can exercise the option, that is buy a share for 220p, then sell the share for
250p, thus gaining a profit of 30p. In the options exchange, the contract will be settled
by just paying you out 30p. If on 25/09/2014 the market price of a share is 200p, you
will not exercise, as it makes not financial sense and you are not obliged to exercise. On
25/09/13 the value of stock could be higher (the option is ‘in the money’) or lower (‘out
of the money’) than the strike 220p; but in any case the option has some positive time
value due to the chance that at the maturity the option will be in the money.
The payoff at time T for holder of each of the above three types of contracts depends
on the stock price S = ST at time T in a simple way: it is S − K for forward, (S − K)+
for call option, and (K − S)+ for put option. Throughout we shall use the notation x+
for max(x, 0), called positive part of x (e.g. 7+ = 7, (−3)+ = 0).
Exercise 2.1. Graph the payoffs as functions of S.
Exercise 2.2. You buy a call option with strike K1 = 2 and write a call option with
strike K2 = 3. Both options are issued on the same stock, and have the same maturity
time T . Graph the payoff of your portfolio as a function of the stock price ST at maturity.
At any time from the start of the contract to maturity T the contract has some value,
which may be negative for forwards. The holder of the security (buyer) has a ‘long’
position, while the counter-party (seller, writer) is said to have a ‘short’ position.

1
 Puts and calls are known as ‘plain vanilla’ options. More complex options may have
payoff also depending on the stock price before the maturity (exotic options), and can be
exercised before they expire (American options),

2.2 Risk-free investments

Most of the market model include a risk-free investment like bonds or saving account in a
bank. Normally we shall assume that the rate of returns for such investment is the same
for both borrowing and investing, and in simple case that the rate is a constant r > 0.
For instance, let r be the annual interest rate. The interest can be compounded
simply (one time a year), multiply over discrete time periods (e.g. monthly, quarterly) or
continuously. If the interest is compounded k times per year, then investing amount P
(the ‘principal’) returns in a year P (1 + r/k)k . If the interest is compounded continuously,
then investing amount P returns P ert over time t.

Example 2.3. On your credit card you borrow 90 pounds at interest rate 18% com-
pounded monthly. After one year you owe pounds
 12
0.18
90 1 + = 107.6
12

Example 2.4. On your credit card you borrow 90 pounds at interest rate 18% com-
pounded continuously. After five months you owe pounds

90 e0.18 · 5/12 = 97

Exercise 2.5. Suppose the annual interest rate on saving account is r. Let Rk be the
return from investing P pounds for a year with k-times compounded interest, and let R
be the analogous return from investing P pounds with continuously compounded interest.
Argue that Rk converges to R as k → ∞. What is the return on such investment for T
years (in multiply compounded, respectively continuously compounded ineterest)?

It is useful to go back and forth in time with interest rate computations. The present
(time 0) value of capital P at time t is the riskless investment at time 0 which returns
P at time t. The present value of a pound at time 1 is e−r for continuous compounding,
and (1 + r)−1 for simple compounding. More generally, an asset worth a pound at time t
has present value e−rt (for t ≥ 0) for continuous compounding, respectively (1 + r)−t (for
t = 0, 1, 2, . . . ) for simple compounding.

2.3 The ‘no arbitrage’ principle

A fundamental problem of the financial mathematics is the pricing problem for derivatives.
The question is how to determine a ‘reasonable’ price of a derivative any time before the
maturity. The pricing methodology based on the ‘no arbitrage’ principle has become a
cornestone of the theory.

2
Definition 2.6. An arbitrage opportunity is a trading strategy which requires zero in-
vestment, incurs no losses but may yield a positive profit (with some probability). The
arbitrage pricing paradigm sticks with the assumption that the market has no arbitrage
opportunities.
Example 2.7. (Price of a forward) Let the risk-free investment rate be r, with interest
compounded continuously. Consider a forward contract on a share of a stock, with delivery
price K and maturity T . The payoff of this contract is FT = ST − K at time T . Consider
also a portfolio comprised of one share of the stock, which is priced S0 at time 0, and
a cash position −Ke−rT (negative amount means money borrowed from bank). So the
price of the portfolio at time 0 is S0 − Ke−rT . At time T the value of the portfolio will be
ST − (Ke−rT )erT = ST − K.
We see that at time T the portfolio pays the same as the forward contract.
The arbitrage pricing requires that the price F0 of forward at time 0 must be
F0 = S0 − Ke−rT ,
for otherwise there could be an arbitrage opportunity. The investor who sells a forward
contract for F0 can hedge the risks by investing the sale proceeds into one share of the
stock and −Ke−rT in bonds.
Suppose the contrary, that the forward is traded at time 0 on the market at some price
F0 > S0 − Ke−rT . In that case you can write (or borrow) a forward contract (so assume
obligations to be fulfilled in time T ), receive F0 in cash and invest S0 −Ke−rT in the stock-
and-bond portfolio. At time 0 you pocket positive amount A = F0 − (S0 − Ke−rT ) > 0,
while at time T you will have ST − K to pay back your obligations, keeping positive
amount AerT in your pocket whichever the stock price ST occurs to be.
Example 2.8. (The put-call parity) Observe the identity
(ST − K)+ + K = (K − ST )+ + ST .
Let St , Pt , Ct be the price at time t (for 0 ≤ t ≤ T ) of stock, put option and call option,
respectively. We know that
PT = (K − ST )+ , CT = (ST − K)+
are the payoffs of put and call options, so rearranging CT + K = PT + ST . The portfolio
‘one call option, Ke−r(T −t) pounds in bonds’ is worth some Ct + Ke−r(T −t) pounds at time
t, and will be valued at CT + K = PT + ST at time T , like the portfolio ‘one put option,
one share of stock’. The arbitrage pricing implies that the portfolios should be equally
valued at time t, hence the put-call parity formula
Ct + Ke−r(T −t) = Pt + St
for all t ∈ [0, T ]. Whichever, the put and call prices Pt , Ct happen to be, they must satisfy
the formula.
Exercise 2.9. Suppose the prices of stock, call and put satisfy St − Ke−r(T −t) < Ct − Pt
for some time t < T . Find an arbitrage opportunity.

3
2.4 One-period binomial model
We have been successful with solving promptly the pricing problem for forwards because
these are relatively simple derivatives. For calls, puts and more complex instruments we
need a model for the price development of the underlying (stock). The simplest model of
the kind is the market with one stock and one bond (money), and the one-period evolution
with two possible ‘moves’ of the stock.
Suppose the market evolves over one period, from time 0 to time 1, with no interme-
diate valuations. At time zero we have a stock with some known price S0 > 0, while at
time 1 the price depends on chance. We model the chance as a binary choice from sample
space Ω = {H, T } (here: T stands for ‘tail’, not for ‘maturity’), thus S1 = S1 (ω) is a
random variable depending on ω ∈ Ω. We may think of tossing (perhaps, biased) coin,
so that if the coin lands heads the price is S1 (H) = uS0 , and if tails S1 (T ) = dS0 . The
‘up factor’ u and ‘down factor’ d are some given positive numbers with d < u.
The stock might move up or down with certain ‘true’ or ‘market’ probabilities p and
q. However, these are of secondary interest for us, provided both events H, T are possible.
For the riskless money market (bonds) we assume (simply compounded) interest rate
r, so 1 pound invested at time 0 will yield 1 + r pounds at time 1.
To exclude the arbitrage we must assume

d < r + 1 < u.

Exercise 2.10. Suppose r + 1 ∈

/ (d, u). Find an arbitrage opportunity.
To apply the ‘no arbitrage’ principle for pricing an option, we need to find a hedging
strategy in the stock+money market.
Example 2.11. Suppose S0 = 4, u = 2, d = 1/2, r = 1/4. Then S1 (H) = 8 and S1 (T ) =
2. Consider a European call with maturity time 1 and strike K = 5. To hedge the option
we may start with capital X0 = 1.2, buying ∆0 = 21 shares at time 0, and investing
X0 − ∆0 S0 = −0.8 in bonds (cash position).
At time 1 the cash position will be (1 + r)(X0 − ∆0 S0 ) = −1, and the stock position
will be either 12 S0 (H) = 4 or 21 S0 (T ) = 1, leaving us with the wealth either
1
X1 (H) = S1 (H) + (1 + r)(X0 − ∆0 S0 ) = 3
2
or
1
X1 (T ) = S1 (T ) + (1 + r)(X0 − ∆0 S0 ) = 0.
2
On the other hand, the payoff of call is (S1 (H) − K)+ = 8 − 5 = 3 if ω = H; and
(S1 (T ) − K)+ = (2 − 5)+ = 0 if ω = T .
We see that a stock-bond portfolio of worth X0 = 1.2 hedges (replicates) the call. By
the ‘no arbitrage’ principle the time-0 value of the call option must be X0 = 1.2.

The example suggests a general way of pricing derivatives in the one-period model.
Suppose an option has payoff V1 (ω) depending on ω ∈ {H, T }. We are looking for a
portfolio of some worth X0 , with ∆0 shares of stock and X0 − ∆0 S0 invested in bonds,

4
such that at time 1 the portfolio worth is V1 (ω) whichever ω. The value of portfolio at
time 1 is

X1 = ∆0 S1 + (1 + r)(X0 − ∆0 S0 ) = (1 + r)X0 + ∆0 (S1 − (1 + r)S0 ).

We want to choose X0 , ∆0 to have X1 (ω) = V1 (ω) for ω = H, T , that is

 
1 1
X0 + ∆0 S1 (H) − S0 = V1 (H) (1)
r+1 r+1
 
1 1
X0 + ∆0 S1 (T ) − S0 = V1 (T ). (2)
r+1 r+1
A smart way to solve (1),(2) is to multiply them by yet indefinite factors p̃, respectively,
q̃ = 1 − p̃, then add them up to get
 
1 1
X0 + ∆0 [p̃S1 (H) + q̃S1 (T )] − S0 = [p̃V1 (H) + q̃V1 (T )].
r+1 r+1
Suppose we have chosen p̃ to achieve
1
S0 = [p̃S1 (H) + q̃S1 (T )], (3)
r+1
then
1
X0 = [p̃X1 (H) + q̃V1 (T )]. (4)
r+1
From (3) easily
1+r−d u−1−r
p̃ = , q̃ = , (5)
u−d u−d
and subtracting (2) from (1) we get

V1 (H) − V1 (T )
∆0 = . (6)
S1 (H) − S1 (T )
Formula (6) is called the delta hedging formula. A portfolio worth (4) with ∆0 shares will
hedge short position in the option. The ‘no arbitrage’ principle prescribes that the time-0
value of the option be
1
V0 = [p̃V1 (H) + q̃V1 (T )].
r+1
The numbers p̃, q̃ are positive and add to 1. We can therefore interpret them as prob-
abilities for H and T (upward/downward moves). These are the risk-neutral probabilities,
which are characterised by the condition (3), which says that the the discounted expected
stock price at time 1 is equal to S0 . In other words, under the risk-neutral probability
the stock moves, on the average, exactly like the riskless bond.
Warning: the risk-neutral probabilities should not be confused with the market prob-
abilities p, q. Under the market probabilities the stock should perform, on the average,
better than a bond, for otherwise the investors had no incentive to invest in stocks.

5
2.5 Multiperiod binomial model
The one-period model can be readily extended to multiple periods, with two possible
stock price moves in every situation. When the move is ‘up’, the price is multiplied by
factor u, ‘down’ – by factor d.
For N -period model, we can think of N coin tosses, and adopt for the sample space
Ω = {H, T }N , which is the set of 2N sequences ω = ω1 . . . ωN with ωj ∈ {H, T }. The
stock price at time 0 is a fixed value S0 and in time n = 1, . . . , N the stock price Sn
depends on ‘chance’ ω through the first n coin tosses ω1 . . . ωn , so does not depend on
further coin-tosses ωn . . . ωN . We can write therefore Sn (ω1 . . . ωn ) instead of Sn (ω) with
ω = ω1 . . . ωN . For instance S3 (HT H) = u2 dS0 , S4 (HT T T ) = ud3 S0 , etc.
Consider an option (derivative security) which pays VN at maturity N , where VN =
VN (ω1 . . . ωN ) depends on N coin-tosses. This notation includes the possibility that the
option is path dependent (aka ‘exotic’), with a payoff VN depending not only on SN but
also on stock prices before maturity N .
In the multiperiod market, a trading strategy may adjust the portfolio with every new
‘stock price tick’ at times n = 1, . . . , n − 1
Let (Xn , ∆n ) for n = 0, 1, . . . , N be a stock-bond portfolio process with total wealth
Xn and ∆n shares of stock at time n. The portfolio process is self-financing if it satisfies
the recursion

Xn+1 = ∆n Sn+1 + (1 + r)(Xn − ∆n Sn ), n = 0, 1, . . . , N − 1, (7)

meaning that the wealth of portfolio at time n + 1 comes from the investment at time n.
No dividends, no consumption.
Introduce
1+r−d u−1−r
p̃ = , q̃ = ,
u−d u−d
and define recursively backward in time VN −1 , VN −2 , . . . , V0 by
1
Vn (ω1 . . . ωn ) = [p̃Vn+1 (ω1 . . . ωn H) + q̃Vn+1 (ω1 . . . ωn T ), (8)
r+1
so each Vn depends on ω1 . . . ωn only. Further define

Vn+1 (ω1 . . . ωn H) − Vn+1 (ω1 . . . ωn T )

∆n (ω1 . . . ωn ) = . (9)
Sn+1 (ω1 . . . ωn H) − Sn+1 (ω1 . . . ωn T )

Theorem 2.12. The portfolio process defined by equations (7), (8), (9) replicates the
option, that is
XN (ω1 . . . ωN ) = VN (ω1 . . . ωN )
for all ω1 . . . ωN .

To construct a hedge at time N − 1, when the tosses ω1 . . . ωN −1 are fixed, we just

argue like in the 1-period model, finding ∆N −1 and XN −1 . This will give
1
XN −1 (ω1 . . . ωN −1 ) = [p̃VN (ω1 . . . ωN −1 H) + q̃VN (ω1 . . . ωN −1 T )].
r+1

6
A full proof is by induction, showing that

Xn (ω1 . . . ωn ) = Vn (ω1 . . . ωn )

for all ω1 . . . ωn and n = 1, . . . , N .

Following the ‘no arbitrage’ principle we must set the option price at time n equal to
Vn , for all n = 0, . . . , N .

Exercise 2.13. In three-period binomial model, consider a path-dependent option which

pays
V3 = max Sn − S3
0≤n≤3

at the maturity time N = 3. Suppose u = 1.5, d = 0.7, r = 0.1, S0 = 5. Calculate the

price of the option at time 0.

2.6 Pricing calls and puts

Call and put options are not path dependent. Their payoff is determined only by the price
of underlying stock at the maturity time. This feature simplifies formulas, because at any
time n only Sn should be taken into account, and not all previous stock price moves.
Within the framework of N -period model let VN = vN (SN ), where vN is some given
function of the stock price. As can be argued by backward induction (and will be shown
later), the option price Vn depends on Sn , hence can be written as Vn = vn (Sn ) for some
function vn , n ≤ N .
The recursion for the time-n value of the option simplifies as
1
vn (s) = [p̃vn+1 (us) + q̃vn+1 (ds)], n = 0, . . . , N − 1. (10)
r+1
Moreover the time-0 price of the option becomes
N  
1 X N j N −j j N −j


V0 = p̃ q̃ vN S0 u d . (11)
(r + 1)N j=0 j

Exercise 2.14. Write analogous formula for vn (s). Show that this formula agrees with
the recursion (10).

Consider the coin-tossing space Ω = {H, T }N with the risk-neutral probability measure

P(ω)
e = p̃#heads(ω) q̃ #tails(ω)
(e.g. #heads(HHHT HT ) = 6 − #tails(HHHT HT ) = 4). Note that #heads has
Binomial(N, p̃) distribution. We can write (11) as the discounted expected payoff (or
present value of the expected payoff)
1
V0 = Ev(S
e N ).
(1 + r)N