Group Project - Module 5: Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira

Group Project - Module 5
The Trinomial Model
Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira
Abril 6th, 2019
MScFE 620 Discrete-time Stochastic Processes
WorldQuant University
Let (Ω, ℱ, 𝔽𝔽, ℙ) be the filtered probability space with T = 2, d = 1 risky stock. The evolution of
the stock price is modelled by a trinomial model, where price X is:
𝑇𝑇
𝑋𝑋𝑛𝑛+1 = 𝑋𝑋𝑛𝑛 𝑍𝑍𝑛𝑛+1 , 𝑋𝑋0 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 where 𝑍𝑍𝑍𝑍𝑛𝑛=1 is a sequence of independent random variables,
each taking three distinct values.
Here we define three random values: u is an upwards movement, m is no movement at all, and d
is a downwards movement. Following T = 2, the probability space is defined as such: Ω = {(u,
u), (u, m), (u, d), (m, u), (m, m), (m, d), (d, u), (d, m), (d, d),}
The filtrations are defined as follows:
ℱ0𝑋𝑋 = {Ø, Ω},
ℱ1𝑋𝑋 = {Ω, Ø, {(𝑢𝑢, 𝑢𝑢), (𝑢𝑢, 𝑚𝑚), (𝑢𝑢, 𝑑𝑑)}, {(𝑚𝑚, 𝑢𝑢), (𝑚𝑚, 𝑚𝑚), (𝑚𝑚, 𝑑𝑑)}, {(𝑑𝑑, 𝑢𝑢), (𝑑𝑑, 𝑚𝑚), (𝑑𝑑, 𝑑𝑑)}}
ℱ2𝑋𝑋 = 𝜎𝜎�{𝑋𝑋0, 𝑋𝑋1, 𝑋𝑋2 }� = 2Ω .
The following trinomial tree of a call option has been created using the following criteria:
u = 1.25, d = 1/u = 0.8, m = 1, T = 2, X0 = 100 and strike price K0 = 100.
156.25
125 125
100 100 100
80 80
64
The payoff values are tabulated below:
ω X0(ω) X1(ω) X2(ω) H(ω) = max{X2(ω) – 100, 0}

(u, u) 100 125 156.25 156.25
(u, m) 100 125 125 125
(u, d) 100 125 100 0
(m, u) 100 100 125 125
(m, m) 100 100 100 0
(m, d) 100 100 80 0
(d, u) 100 80 100 0
(d, m) 100 80 80 0
(d, d) 100 80 64 0
In order for the model to be arbitrage-free, we must find an equivalent martingale measure
(EMM) for X. If we take ℙ* to be a probability measure on (Ω, ℱ), then ℙ* must be equal to ℙ for
all filtrations. Looking at ℱ1𝑋𝑋 , the probability of u = δu, m = δm, and d = δd.
Setting up the following equation, we find the set of EMMs for this above model:
ℙ∗ = 𝑎𝑎 ∗ 𝛿𝛿𝑢𝑢 + 𝑏𝑏 ∗ δ𝑚𝑚 + 𝑐𝑐 ∗ δ𝑑𝑑 , where 𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 = 1
𝔼𝔼∗ (𝑋𝑋1 |ℱ0𝑋𝑋 ) = 𝑎𝑎 ∗ 𝛿𝛿𝑢𝑢 + 𝑏𝑏 ∗ δ𝑚𝑚 + 𝑐𝑐 ∗ δ𝑑𝑑 = 𝑋𝑋0 = 100
Completeness of a market is characterized by The Fundamental Theorem of Asset Pricing II

(FTAP II):
For a financial market �(𝛺𝛺, ℱ, 𝔽𝔽, ℙ), 𝑋𝑋� with no arbitrage opportunities, the following are
equivalent: 1. The market is complete; 2. P has exactly one element; 3. X has the predictable
representation property (PRP) with respect to every ℙ* ∈ P: every (𝔽𝔽, ℙ∗ )-martingale M
with M0 can be written as a martingale transform with respect to X.
We want:
(i)
and r < u, but
(ii)
These are our conditions on r, u and qo, where r determines the growth of value for the risk-free
asset of our portfolio and u the growth of the stock value.
Under condition (i) the self-financing portfolio with value
at time t<N for all q∈ℝ,
where V(N) is the value of a self-financing portfolio {HS(t)HB(t)}t∈𝕀𝕀 at time N will be arbitrage-
free.
In particular,
(iii)
We recall that:
Self-financing portfolio {HS(t)HB(t)}t∈𝕀𝕀 is an arbitrage portfolio if it's value satisfies the

following conditions: V(0)=0, V(N,x)≥0 for all x ∈ {-1,0.1}, and there exists some y ∈ {-1,0.1}
such that V(N,y)>0.
For the first period case,
HS(0)=HS(1)=HS
HB(0)=HB(1)=HB
Thus, the portfolio position in the 1-period model is constant over the interval [0,1]. The value of
the portfolio at time t=0 is:
V(0)=HSS(0)+HBB(0)
At time t=1, it is one of the following:
For an arbitrage portfolio, V(0)=0 must hold.

Assuming it does hold, i.e.
HSS(0)+HBB(0)=0 (iv)
Also, V(1)≥0 must hold, i.e.
HSS(0)eu+HBB(0)er ≥0
HSS(0)e-u+HBB(0)er ≥0
HSS(0)+HBB(0)er ≥0
And also, one of the inequalities must be strict.
From (iv), HSS(0)= -HBB(0) (iv)
⇒ HSS(0)(eu-er )≥0
HSS(0)(e-u-er) ≥0
HSS(0)(1-er) ≥0
Since we must have at least one strict inequality, HS≠0
We then must analyze two cases:
1) HS>0
⇒ (eu-er )≥0
(e-u-er) ≥0
(1-er) ≥0
(1-er) ≥0 ⇒ r ≤0,
but r ≥0
⸫r = 0.
But u >0, therefore (e-u-er) ≥0 does not hold.
⇒⇐ (contradiction)
2) HS < 0
⇒ (eu-er )≤0
(e-u-er) ≤0
(1 - er) ≤0
(1-er) ≥0 ⇒ r ≤0,
but r ≥0
⸫r = 0.
But u >0, therefore (e-u-er) ≥0 does not hold.
⇒⇐ (contradiction)
We recall that:
For the multi-period model, the value at t=0 of a self-financing portfolio satisfies equation (iii):
As q-1, q0, q+1 are positive, V(0)=0 can only hold if V(N, x)≡0 along all paths.
It follows that the market cannot be an arbitrage, under these conditions.
It also follows that when our risk-neutral measure (EMM), the triple q-1, q0, q+1, constitutes a
probability, it implies an absence of arbitrage in the market.
A market is said to be complete if the arbitrage-free price of a derivative is uniquely defined. In
such a market, the price will coincide with the value of a hedging (also called replicating)
strategy.
The standard trinomial model is an example of an incomplete market since the price will depend
on q0 and hence not be uniquely defined.
The trinomial model can be made complete by adding a second risky asset. More generally, a
market model with m number of states can only be made complete if there exist at least m-1 risky
assets – in the trinomial model case of three states, at least 2 risky assets.
Here, we will assume that:
1) the risky assets are independent,

2) there are no constraints regarding change of states, and
3) there is no arbitrage opportunity.
We consider risky asset Si which moves according to
In order to determine the martingale measure Q, for the completed trinomial market, we use the
system of equations:
Under the assumption of no arbitrage, the system has the solution

q-1, q0, q+1 are positive numbers belonging to (0,1) if we choose the model parameters in a
suitable way. Thus q-1, q0, q+1 constitute a probability measure Q (EMM) under this assumption.
References:
1. A Pascucci and W J Runggaldier (2012), Financial Mathematics: Theory and Problems

for Multi-period Models; Springer; Milan.
2. Johan Bjorefeldt et al. (2016), The Trinomial Asset Pricing Model; Chalmers University
of Technology; Gothenburg.

Group Project - Module 5: Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira

Uploaded by

Copyright:

Available Formats

Group Project - Module 5: Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira

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

Group Project - Module 5: Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira

Uploaded by

Copyright:

Available Formats

Group Project - Module 5

The Trinomial Model

Eric Cooper, Chrispin Phiri, Francis Mtambo, and Chamunorwa Karimanzira

Abril 6th, 2019

MScFE 620 Discrete-time Stochastic Processes

The filtrations are defined as follows:

ℱ0𝑋𝑋 = {Ø, Ω},

ℱ2𝑋𝑋 = 𝜎𝜎�{𝑋𝑋0, 𝑋𝑋1, 𝑋𝑋2 }� = 2Ω .

u = 1.25, d = 1/u = 0.8, m = 1, T = 2, X0 = 100 and strike price K0 = 100.

100 100 100

The payoff values are tabulated below:

ω X0(ω) X1(ω) X2(ω) H(ω) = max{X2(ω) – 100, 0}

ℙ∗ = 𝑎𝑎 ∗ 𝛿𝛿𝑢𝑢 + 𝑏𝑏 ∗ δ𝑚𝑚 + 𝑐𝑐 ∗ δ𝑑𝑑 , where 𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 = 1

𝔼𝔼∗ (𝑋𝑋1 |ℱ0𝑋𝑋 ) = 𝑎𝑎 ∗ 𝛿𝛿𝑢𝑢 + 𝑏𝑏 ∗ δ𝑚𝑚 + 𝑐𝑐 ∗ δ𝑑𝑑 = 𝑋𝑋0 = 100

Completeness of a market is characterized by The Fundamental Theorem of Asset Pricing II

and r < u, but

at time t<N for all q∈ℝ,

Self-financing portfolio {HS(t)HB(t)}t∈𝕀𝕀 is an arbitrage portfolio if it's value satisfies the

For the first period case,

At time t=1, it is one of the following:

For an arbitrage portfolio, V(0)=0 must hold.

Also, V(1)≥0 must hold, i.e.

And also, one of the inequalities must be strict.

From (iv), HSS(0)= -HBB(0) (iv)

Since we must have at least one strict inequality, HS≠0

We then must analyze two cases:

But u >0, therefore (e-u-er) ≥0 does not hold.

It follows that the market cannot be an arbitrage, under these conditions.

Here, we will assume that:

1) the risky assets are independent,

We consider risky asset Si which moves according to

Under the assumption of no arbitrage, the system has the solution

1. A Pascucci and W J Runggaldier (2012), Financial Mathematics: Theory and Problems

You might also like