6. Binomial Trees
6.1 One-step Binomial Trees
6.1.1 Definition of Binomial Model
Binomial model: after one period, the value of the underlying asset will either go up to
S
u
S_u
Su or go down
S
d
S_d
Sd.
6.1.2 No-arbitrage Argument
Law of one price: Assets that produce identical(完全相同的) future cash flows regardless of future everts should have the same price.
We can use option and stock to construct a portfolio which has certain value after one period. If there is no arbitrage opportunity, the option price can be derived from the cost of this portfolio.
Example: The stock is currently trading at $
50
50
50. The stock price will either go up to $
75
75
75 or go down to $
25
25
25 after one year. The risk-free rate is
2
%
2\%
2%. Please calculate the value of a 1-year European call option with an exercise price of $
60
60
60 using one-step binomial tree.
We create risk-neutral portfolio, Long
Δ
\Delta
Δ stock and short
1
1
1 call option.
The value of the portfolio after 1 year: Δ ∗ S u − C u = Δ ∗ S d − C d → Δ = 0.3 \Delta*S_u-C_u=\Delta*S_d-C_d \to \Delta=0.3 Δ∗Su−Cu=Δ∗Sd−Cd→Δ=0.3
The portfolio value at t 0 t_0 t0: 50 × 0.3 − c 50\times0.3-c 50×0.3−c
The portfolio value at t 1 t_1 t1: 75 × 0.3 − 15 = 7.5 75\times0.3-15=7.5 75×0.3−15=7.5
Thus, ( 50 × 0.3 − c ) × e 0.02 = 7.5 → c = 7.6485 (50\times0.3-c)\times e^{0.02}=7.5 \to c=7.6485 (50×0.3−c)×e0.02=7.5→c=7.6485
Calculata Δ \Delta Δ of a stock option: Δ × S u − C u = Δ × S d − C d \Delta \times S_u-C_u=\Delta \times S_d-C_d Δ×Su−Cu=Δ×Sd−Cd
Δ = C u − C d S u − S d \Delta=\frac{C_u-C_d}{S_u-S_d} Δ=Su−SdCu−Cd
6.1.3 Risk Neutral Valuation
In a risk neutral world, all individuals are indifferent to risk. Investors requires no compensation for risk and expected return on all securities is the risk-free interest rate.
The option should be valued based on risk neutrality.
In a risk-neutral world, the expected return on all securities is the risk-free interest rate, and the discount rate is the risk-free interest rate.
u
=
e
σ
t
,
d
=
1
u
=
e
−
σ
t
u=e^{\sigma\sqrt{t}}\;, d=\frac{1}{u}=e^{-\sigma\sqrt{t}}
u=eσt,d=u1=e−σt
The higher the standard deviation, the greater the dispersion between stock prices in up and down states.
π
u
×
S
0
×
u
+
(
1
−
π
u
)
×
S
0
×
d
=
S
0
e
r
t
→
π
u
=
e
r
t
−
d
u
−
d
\pi_u\times S_0 \times u+(1-\pi_u)\times S_0 \times d=S_0e^{rt}\to \pi_u=\frac{e^{rt}-d}{u-d}
πu×S0×u+(1−πu)×S0×d=S0ert→πu=u−dert−d
6.1.4 Risk neutral valuation-Process
The valuation process of one-step binomial tree as follow:
- Calculate u u u and d d d, then construct the whole binomial tree.
- Calculate the payoff of the option at the maturity node.
- Compute the risk-neutral up and down probability, then calculate the expected value of option in one period.
- Use risk-free rate to discount the expected value to present.
6.2 Two-step Binomial Trees
6.2.1 Two-Step Binomial Model
The basic valuation process of a two-step European option is similar with one-step binomial model, but with more steps.
6.2.2 Two-Step Binomial Model - European
An analyst is using two-step binomial model to calculate the price of a 2-year European put option with strike price of $ 75 75 75. The continuously compounded risk-free rate is 5 % 5\% 5%. The stock pays no dividend and is trading at $ 70 70 70. The volatility of the stock price is 20 % 20\% 20%. What will be the European put option price?
Firstly, calculate
u
u
u and
d
d
d, then construct the whole binomial trees:
u
=
e
σ
T
=
e
0.02
∗
1
=
1.2214
,
d
=
1
u
=
e
−
σ
T
=
0.8187
u=e^{\sigma \sqrt{T}} = e^{0.02 * \sqrt{1}} =1.2214, d=\frac{1}{u}=e^{-\sigma \sqrt{T}}=0.8187
u=eσT=e0.02∗1=1.2214,d=u1=e−σT=0.8187
Secondly, calculate the payoff of the option at the maturity node:
Thirdly, compute the risk-neutral up and down probability, and then calculate the expected value in one period.
π
u
=
e
r
t
−
d
u
−
d
=
e
5
%
×
1
−
0.8187
1.2214
−
0.8187
=
57.7529
%
\pi_u=\frac{e^{rt}-d}{u-d}=\frac{e^{5\%\times1}-0.8187}{1.2214-0.8187}=57.7529\%
πu=u−dert−d=1.2214−0.8187e5%×1−0.8187=57.7529%
π
d
=
1
−
π
u
=
1
−
57.7528
%
=
42.2471
%
\pi_d=1-\pi_u=1-57.7528\%=42.2471\%
πd=1−πu=1−57.7528%=42.2471%
At last, use risk-free rate to discount the expected value to present.
Node
2
=
(
0
×
57.7529
%
+
5
×
42.2471
%
)
e
−
5
%
=
2.0093
\text{Node}2=(0\times 57.7529\%+5\times 42.2471\%)e^{-5\%}=2.0093
Node2=(0×57.7529%+5×42.2471%)e−5%=2.0093
Node 3 = ( 5 × 57.7529 % + 28.0811 × 42.2471 % ) e − 5 % = 14.0317 \text{Node}3=(5\times 57.7529\%+28.0811\times 42.2471\%)e^{-5\%}=14.0317 Node3=(5×57.7529%+28.0811×42.2471%)e−5%=14.0317
Node 1 = ( 2.0093 × 57.7529 % + 14.0317 × 42.2471 % ) e − 5 % = 6.7427 \text{Node}1=(2.0093\times 57.7529\%+14.0317\times 42.2471\%)e^{-5\%}=6.7427 Node1=(2.0093×57.7529%+14.0317×42.2471%)e−5%=6.7427
The European Put Option price is 6.7427 6.7427 6.7427
6.2.3 Two-Step Binomial Model - American
We need to determine if the option will be exercised at each node including Node 1.
Node
1
=
(
2.0093
×
57.7529
%
+
17.691
×
42.2471
%
)
e
−
5
%
=
8.2133
\text{Node}1=(2.0093\times 57.7529\%+17.691\times 42.2471\%)e^{-5\%}=8.2133
Node1=(2.0093×57.7529%+17.691×42.2471%)e−5%=8.2133
Node1 should also be checked. In this case, if there is an early exercise in Node 1, payoff will be 5 which is less than 8.2133. Therefore, the option will not be exercised early in Node 1 and value of this American put option is 8.2133.
6.2.4 As time periods are added
Suppose that a binomial tree with n n n steps in its life T, if n n n approaches infinity, the length of each step approaches to zero. A continuous binomial tree will be achieved and this is one of the ways that derive the Black-Scholes-Merton model.
6.2.5 Options on Other Assets
Options on stock indices with continuous dividend yield q q q
π
u
×
S
0
×
u
+
(
1
−
π
u
)
×
S
0
×
d
=
S
0
e
(
r
−
q
)
t
\pi_u\times S_0 \times u+(1-\pi_u)\times S_0 \times d=S_0e^{(r-q)t}
πu×S0×u+(1−πu)×S0×d=S0e(r−q)t
π
u
=
e
(
r
−
q
)
t
−
d
u
−
d
\pi_u=\frac{e^{(r-q)t}-d}{u-d}
πu=u−de(r−q)t−d
Options on currencies with thedomestic risk-free rate R DC R_{\text{DC}} RDC and foreign risk-free rate R FC R_{\text{FC}} RFC
π u = e ( r DC − r FC ) t − d u − d \pi_u=\frac{e^{(r_{\text{DC}}-r_{\text{FC}})t}-d}{u-d} πu=u−de(rDC−rFC)t−d
Option on futures
π
u
=
1
−
d
u
−
d
\pi_u=\frac{1-d}{u-d}
πu=u−d1−d