State Prices

Wang Wei Mun

Lee Kong Chian School of Business

Singapore Management University

October 20, 2024

Wang Wei Mun State Prices 1 / 12

Economic Environment

Financial market consists of n risky assets, with initial price of

Pi and (random) final payoff of X̃i for one share of asset i
Financial market has k ≥ 2 “states of nature”, where each
state corresponds to unique set of outcomes for final payoffs
Let Xsi be final payoff for one share of asset i in state s, and
let X be k × n matrix that shows all possible outcomes:
 
X11 · · · X1n
X =  ... .. .. 

. . 
Xk1 · · · Xkn

Notice that each column of X represents different asset, while

each row of X represents different state of nature

Wang Wei Mun State Prices 2 / 12

Complete Market

Financial market is complete if n ≥ k and X has k linearly

independent columns and rows =⇒ X has rank k
If n > k, then can form k portfolios with linearly independent
payoffs: assume that n = k =⇒ X is invertible
Let Y = [Y1 , . . . , Yk ]′ be any k × 1 vector of desired final
payoffs in each state of nature
Let N = [N1 , . . . , Nk ]′ be k × 1 vector of required shares in
each asset, in order to create portfolio with payoffs of Y:

Y = XN =⇒ N = X−1 Y

Hence can create (unique) portfolio to deliver any set of

desired payoffs, or replicate payoffs for any existing investment

Wang Wei Mun State Prices 3 / 12

State Prices

Let P = [P1 , . . . , Pk ]′ be k × 1 vector of initial prices for one

share of each asset: to avoid arbitrage, portfolio with final
payoffs of Y must have initial price of PY = P′ N = P′ X−1 Y
Let es be k × 1 vector of final payoffs for elementary security
(or primitive security or Arrow–Debreu security) that delivers
final payoff of one in state s, and zero in every other state
Let ps be initial price of elementary security for state s:

ps = P′ X−1 es ∀ s = 1, . . . , k =⇒
p1 · · · pk = P′ X−1 e1 · · · ek = P′ X−1
   

Then ps is known as state price for state s, which represents

initial value of receiving final payoff of one in state s

Wang Wei Mun State Prices 4 / 12

Pricing Kernel

There exists unique set of state prices in complete market

Investors who are non-satiated will always be willing to pay for
more consumption, so state prices must be strictly positive
Let πs > 0 be probability for state s, where ks=1 πs = 1
P

Initial price of portfolio with payoffs of Y can be expressed in

terms of state prices, which is related to pricing kernel:

k k  
X X ps h i
PY = ps Ys = πs Ys = E M̃ Ỹ
πs
s=1 s=1

Hence there exists unique pricing kernel in complete market,

which must have value of Ms = ps /πs > 0 in state s

Wang Wei Mun State Prices 5 / 12

Risk-Neutral Probabilities

Initial price of riskless asset with payoff of one in every state:

k k
X X h i 1
Pf = ps = πs Ms = E M̃ =
Rf
s=1 s=1

Let πbs = Rf ps > 0 for s = 1, . . . , k, andPinterpret as set of

(risk-adjusted) state probabilities since ks=1 π bs = 1
Initial price of portfolio that delivers payoffs of Y:

k k
X 1 X 1 bh i
PY = ps Ys = π
bs Ys = E Ỹ
Rf Rf
s=1 s=1

Wang Wei Mun State Prices 6 / 12

Risk-Neutral Probabilities

Here E
b [·] is expectation under probability distribution of π
b
All portfolios will have same expected return under probability
distribution of π
b, which is equal to risk-free rate:

1 bh i
RY = E Ỹ = Rf
PY

Interpret π
b as risk-neutral probability distribution, for which
pricing kernel is non-random: M b s = R −1 for all s
f
Then π represents physical probability distribution, for which
(random) pricing kernel is given by investor’s IMRS
Hence risk-neutral pricing formula (using risk-neutral
probability distribution) is equivalent to using state prices

Wang Wei Mun State Prices 7 / 12

Risk-Neutral Probabilities

Risk-neutral probability distribution puts more (or less) weight

on states where pricing kernel is above (or below) average:
 
Ms
bs = Rf ps = Rf Ms πs =  h i  πs
π
E M̃

Hence “bad” states (where consumption is low and marginal

utility is high) are more likely to occur and “good” states are
less likely to occur, under risk-neutral probability distribution
Then π b is risk-adjusted probability distribution that eliminates
risk premium and induces risk-neutral behaviour (where
expected return is equal to risk-free rate for all assets)

Wang Wei Mun State Prices 8 / 12

Binomial Model

Consider “binomial” model with two states of nature

Risky stock has initial price of S, which will subsequently rise
to uS or drop to dS, where u > d
Riskless bond has initial price of Pf = Rf−1 , where u > Rf > d
Vector of initial prices and matrix of final payoffs:
   
S uS 1
P= , X=
Pf dS 1

Vector of state prices:

 
  ′ −1 1 − dPf uPf − 1
pu pd =PX =
u−d u−d

Wang Wei Mun State Prices 9 / 12

Binomial Model

Vector of risk-neutral probabilities:

 
    Rf − d u − Rf
π̂u π̂d = Rf pu pd =
u−d u−d

Initial price of portfolio that delivers final payoff of Yu or Yd :

1
PY = pu Yu + pd Yd = (π̂u Yu + π̂d Yd )
Rf

Binomial model is often used to price options: initial price will

not be accurate with just one time period, but converges to
true value as model is extended to more time periods

Wang Wei Mun State Prices 10 / 12

Example: Binomial Model

Stock has initial price of 6 and final payoff of 10 or 5

Riskless bond has risk-free rate of 1.05
Vector of initial prices and matrix of final payoffs:
   
6 10 1
P= 1 , X=
1.05 5 1

Vector of state prices:

 
  1
 1
 1 −1  
pu pd = 5 6 1.05 = 0.2476 0.7048
−5 10

Wang Wei Mun State Prices 11 / 12

Example: Binomial Model

Vector of risk-neutral probabilities:

     
π̂u π̂d = 1.05 × 0.2476 0.7048 = 0.26 0.74

Call option gives option (but not obligation) to buy one share
of stock (at end of time period) for “strike price” of 6
Call option will have terminal value of 4 when stock has final
payoff of 10, or 0 when stock has final payoff of 5
Hence initial price of call option:

0.26 × 4
0.2476 × 4 = = 0.99
1.05

Wang Wei Mun State Prices 12 / 12