14-Arbitrage Page 397 Wednesday, February 4, 2004 1:08 PM
Arbitrage Pricing: Finite-State Models 397
The market value Sθθθθof a portfolio θθθθat time 0 is a scalar given by
the scalar product:
N
Sθθθθ= Sθθθθ= ∑ Siθi
i = 1
Its payoff dθθθθat time 1 is the M-vector:
dθθθθ= D′θθθθ
The price of a security and the market value of a portfolio can be negative
numbers. In the previous example of a two-state, three-security market
we obtain
Sθθθθ= Sθθθθ= S 1 θ 1 + S 2 θ 2 + S 3 θ 3
dθθθθ D′θθθθ
d 11 d 21 d 31
d 12 d 22 d 32
θ 1
θ 2
θ 3
d 11 θ 1 + d 21 θ 2 + d 31 θ 3
d 12 θ 1 + d 22 θ 2 + d 32 θ 3
= = =
Let’s introduce the concept of arbitrage in this simple setting. As we
have seen, arbitrage is essentially the possibility of making money by trad-
ing without any risk. Therefore, we define an arbitrage as any portfolio θ
which has a negative market value Sθ = Sθθθθ< 0 and a nonnegative payoff
Dθ = D′θθθθ≥ 0 or, alternatively, a nonpositive market value Sθ = Sθθθθ≤ 0 and
a positive payoff Dθ = D′θθθθ> 0.
State Prices
Next we define state prices. A state-price vector is a strictly positive M-
vector ψψψψsuch that security prices can be written as S = Dψψψψ. In other
words, given a state-price vector, if it exists, security prices can be
recovered as a weighted average of the securities’ payoffs, where the
state-price vector gives the weights. In the previous two-state, three-security
example we can write:
ψ 1
ψψψψ=
ψ 2
S = Dψ