Model-Free Option Prices

Kuo-Ping Chang*

First version
Second version

June 2014
August 2014

http://ssrn.com/abstract=2459464

* Department of Quantitative Finance, National Tsing Hua University, Kuang Fu Rd., Hsinchu 300,
Taiwan, E-mail: kpchang@mx.nthu.edu.tw.

Electronic copy available at: http://ssrn.com/abstract=2459464

Model-Free Option Prices

ABSTRACT

In this paper, I have used simple arbitrage argument to derive a dozen of model-free option price
properties. In addition to deriving the Greeks under model-free framework, the results show that first, in
contrast to the traditional view, a European call (put) option for a non-dividend-paying asset can also be a
European call (put) option for any other non-dividend-paying asset, and every non-dividend-paying asset
is also both a European call option and a European put option for any other non-dividend-paying asset.
Second, in some cases the time value of the European put option can be negative, and adjust the exercise
price of an option can decrease or even erase the time value of the option.

I have also used the Arbitrage

Theorem under the binomial option pricing model to examine these properties.

Key words: the Greeks, the time value of option, the Arbitrage Theorem.

JEL Classification: G13, G32.

Electronic copy available at: http://ssrn.com/abstract=2459464

1. Introduction

Start from the seminal work of Black and Scholes (1973), various option price properties have been
derived under the Black-Scholes-Merton and binomial option pricing models. Up to now, there are only
a few model-free option price properties developed in the literature.

In this paper, I use simple arbitrage

argument to derive a dozen of model-free option price properties.

In addition to deriving the Greeks

under model-free framework, the results show that first, in contrast to the traditional view, a European
call (put) option for a non-dividend-paying asset can also be a European call (put) option for any other
non-dividend-paying asset, and every non-dividend-paying asset is also both a European call option and a
European put option for any other non-dividend-paying asset.

Second, in some cases the time value of

the European put option can be negative, and adjust the exercise price of an option can decrease or even
erase the time value of the option.

In Section 3, I also use the Arbitrage Theorem under the binomial

option pricing model to examine these properties.

2. Model-Free Option Prices

A call option gives its owner the right to purchase an asset (the underlying asset) for a given price
(the exercise or strike price: K ) on or before a given date (the expiration date: T ).

A put option gives its

owner the right to sell an asset for a given price on or before the expiration date.

European options can

only be exercised on the expiration date.

American options can be exercised on or before the expiration

date. Because American options have more choices, they are more valuable than European ones.
The following are some important properties of options.1

Property 2.1

Put-Call Parity for European Call and Put Options.

Consider two portfolios at t 0 , where the underlying asset is one share of a non-dividend-paying
stock, and the simple risk-free interest rate between t 0 and t T is r :
Portfolio A: one European call option c with exercise price K , and cash

K
deposited in a bank;
1 r

Properties 2.1-2.3 and the call option part of Property 2.5.1 are the same as in Merton (1973).

Portfolio B: one European put option p with exercise price K , and a share of the stock S 0 .
On the expiration date t T , both portfolios give exactly the same payoff: Max[ST , K ] . Thus, the
costs of the two portfolios at t 0 must be the same:

K
S0 p .
1 r

(1)

Equation (1) is called the put-call parity. Suppose Equation (1) doesnt hold:
Case 1: At t 0 , c

K
S 0 p . Then an investor can immediately get net profit by adopting the
1 r

following strategy: (i) sell one call and borrow

K
from a bank; and (ii) buy one share of the stock
1 r

and one put option. At t T , if ST K , the investor obtains zero profit by letting the put option
expire, selling the share of the stock to the holder of the call option at the price K , and giving K to the
bank. If ST K , the investor obtains zero profit by exercising the put option and giving K to the
bank.
Case 2: At t 0 , c

K
S 0 p . Then an investor can immediately get net profit by adopting the
1 r

following strategy: (i) short-sell one share of the stock and sell one put option; and (ii) buy one call and
deposit

K
in a bank. At t T , if ST K , the investor obtains zero profit by using cash K from
1 r

the bank to exchange for one share of the stock from the holder of the put option, and then give back one
share of the stock. If ST K , the investor obtains zero profit by using cash K from the bank to
exercise the call option, and then give back one share of the stock.

From equation (1) it is easy to find that since call and put options are rights, i.e., their values cannot
be negative, we have the following lower bounds for European call and put options:
Property 2.2

c S0

K
K
S0 .
and p
1 r
1 r

If at t 0 , c S 0

K
, then an investor can immediately obtain net profit by short-selling one
1 r

share of the stock, buying one call option, and depositing cash

K
in a bank. At t T , if ST K ,
1 r

the investor obtains zero profit by using cash K from the bank to exercise the call option, and then give
back one share of the stock. If ST K , the investor can obtain net profit: K ST 0 by getting K
from the bank and spending ST to buy one share of the stock from the market, and then giving back that
share of the stock.
If at t 0 , p

K
S 0 , then an investor can immediately obtain net profit by borrowing cash
1 r

K
from a bank and buying one share of the stock and one put option. At t T , if ST K , the
1 r
investor can obtain net profit: ST K 0 by selling that share of the stock for ST and giving K to
the bank.

If ST K , the investor obtains zero profit by exercising the put option and giving K to the

bank.

Property 2.3

An American call option on a non-dividend-paying stock will never be exercised before

the expiration date.

With the same exercise price K and expiration date T , an American call option C is at least as
valuable as a European call option

c .

At

t T / 2 , equation (1) is written as:

K
K
S p , and if S K 0 , we will have C c S
S K , i.e., the
1 ( r / 2)
1 ( r / 2)

market value of the American call option must be greater than the gain of exercising the call option. For
example, suppose that at t 0 , an issuer issues an American call option with K $102 .

At

t T / 2 , if the stock price is S $105 , then the American call options value C must be greater
than $3 ( S K ) . If C cannot be sold at a price more than $3, it will mean that no one (including
the seller) in the market believes that the stock price at t [T / 2, T ] will be greater than $105. This is,
of course, contradicts the previous assumption that at t T / 2 , S $105 .2

Property 2.4

In the put-call parity, if c p , then K S 0 (1 r ) , i.e., the exercise price must equal the

If all investors believe that at t [T / 2, T ] , St $105 , then ST / 2 $105 cannot sustain.

underlying assets forward price.3

Because at t T , the value of an European call option is: Max[0, ST K ] and the value of an
European put option is: Max[0, K ST ] , at t 0 an issuer of European options can increase K to
decrease c and increase p , i.e.,
Property 2.5.1

c
c
p
p
0 (or
0) and
0 or (
0) . Also,
K
K
K
K

The difference in the values of two otherwise identical European options cannot be

greater than the risk-free-interest-rate-discounted difference in their strike prices, i.e., c

K
and
1 r

K
, where K K 'K , c c'c and p p' p .
1 r

From equation (1), c

Suppose K K 'K 0 .
and p

K
K'
K ' K
S 0 p and c'
S 0 p' , we obtain: c'c
p' p .
1 r
1 r
1 r

Then, c c'c 0 and p p' p 0 which also imply: c

K
1 r

K
.
1 r

Note that equation (1), the put-call parity, is also a restatement of the Modigliani-Miller capital
structure irrelevancy proposition (i.e., the market value of the firm is independent of its capital structure).

At t 0 , equation (1) can be written as S0 c

p where S 0 can be defined as the market
1 r

value of the levered firm, c as the equity of the firm, and

p as the risky debt of the firm.
1 r

the case of riskless debt, p 0 and S 0 c

In

K
K
where
is the riskless debt. At t T , if the
1 r
1 r

For an underlying asset, the relationship between the forward price F0 and the spot price S 0 is: F0 S 0 (1 r ) . For

example, suppose at t 0 , S0 $100 and r 2% . Then the seller of the forward contract will ask for at least $102, i.e.,
she will want: F0 S 0 (1 r ) . However, the seller cannot sell at a price more than $102. This is because the buyer of the
forward contract can always turn to the spot market to buy one unit of the underlying asset at S 0 $100 . Without default,
spend $100 at t 0 is equivalent to spend $102 at t T .

equityholders pay K to the debtholders, then the equityholders can have the firm, ST .4
cannot be negative, we must have:

Because debt

K
p and hence, 0 c S 0 . If at t T the equityholders
1 r

need to pay more to the debtholders to have the firm (i.e., K K 'K 0 ), then at t 0 , c 0 ,

p 0 and 0 S0 S0 S0 c
p imply: lower equity value, higher debt value, no
1 r

change in the market value of the firm, and p

Property 2.5.2

K
.
1 r

c
c
p
p
0 (or
0) and
0 (or
0) .
r
r
r
r

(i) At t 0 , from S0 c
p , if only r increases ( r r'r 0 ) and other factors
1 r

(including S 0 and K ) remain constant, then the firms debt value will decrease, i.e.,

K
K
K
K

p'
p , and 0 S0 S0 c'c

( p' p) implies c c'c 0

1 r'
1 r
1 r' 1 r

(i.e., the firms equity value will increase). Thus,

(ii) At t 0 , for given K and r , riskless debt is:
where 0 is the risk premium.

c
0.
r
K
K
K
p
, and risky debt is:
1 r
1 r
1 r

When only r increases ( r r'r 0 ) and other factors

p
K K
K
K
0 .
remain constant, we will have: 0

p' p . Thus,
r
1 r' 1 r 1 r' 1 r

K
K
Because

( p' p) ()0 , p' p ()0 and

1 r' 1 r
have: p

K
K
K
K
and c
.

1 r' 1 r
1 r' 1 r

Property 2.5.3

c'c ()0 when r'r ()0 , we will

c
c
0 (or
0) .
S0
S0

At t T , if S T K , then the equityholders will not pay K , and the debtholders will have the firm S T .

At t 0 , from S0 c
p , suppose only S 0 increases ( S 0 S ' S 0 0 ) and other factors
1 r

(including r and K ) remain constant.

Case 1: c 0 (i.e.,

Then S 0 c p , and:

c
p
0 ) and p 0 (i.e.,
0 ). This is impossible since S 0 0 . If
S0
S0

this case holds, it will mean that increasing the market value of the firm will make both equityholders and
debtholders suffered.
Case 2: c 0 (i.e.,

c
p
0 ) and p 0 (i.e.,
0 ).
S0
S0

This means that when S 0 0 (the

market value of the firm increases), the equity value will decrease, the debt value will increase, and there
will be some wealth transferring from the equity to the debt.
the debt is riskless, i.e., S 0 c
no effect on the debt.

This result is unlikely. Note that when

K
, S 0 0 only increases the equity value: c S 0 0 and has
1 r

It is implausible that when changing riskless debt to risky debt, S 0 0 will

only benefit debtholders and make equityholders suffered.

Case 3: c 0 (i.e.,

c
p
0 ) and p 0 (i.e.,
0 ).
S0
S0

This means that S 0 0 will only

benefit equityholders and make debtholders suffered.

Case 4: c 0 (i.e.,

c
p
0 ) and p 0 (i.e.,
0 ). This means that S 0 0 will benefit
S0
S0

both equityholders and debtholders.

Property 2.5.4

c p
c p

(or
) for any x S 0 , r, K .
x x
x x

For example, in the Black-Scholes-Merton option pricing model,

c p

where is the volatility.

In the binomial option pricing model where S 0 could go up to S 0 u ( u 1 ) or go down to S 0 d

( 0 d 1 ), we have:

c p
c p

and
.
u u
d d

In the finance literature, the intrinsic value of an option is defined as the maximum of zero and the

value the option would have if it were exercised immediately. For a call option, its intrinsic value is

Max[ S K , 0] . For a put option, its intrinsic value is Max[ K S , 0] . The time value of an option
which arises from the time left to maturity is defined as the difference between option value and intrinsic
value.

At t 0 the time value of the European call option is: TV c c Max[ S0 K , 0] .

Property 2.2, c S 0

K
and hence, TV c 0 .
1 r

From

The time value of the European put option is:

TV p p Max[ K S0 , 0] . We have the following result:

Property 2.6

Even without changing the expiration dates, issuers of European options can adjust

exercise price K to change the time value of European options.

(i)

For S 0 K , if the exercise price increases from K to K ' , i.e., K K 'K 0 , and S 0 K ' ,

TV c
0
K

TV c TV c 'TV c c'( S0 K ' ) [c ( S0 K )] (c'c) ( K ' K ) c K 0 and

TV p TV p 'TV p p' p p 0 and

because by Property 2.5.1, c

TV p
0
K

K
, and K 0 implies c 0 and p 0 .
1 r

(ii) For S 0 K , if the exercise price drops from K to K ' , i.e., K K 'K 0 , and S 0 K ' ,

TV p TV p 'TV p p'( K ' S0 ) [ p ( K S0 )] ( p' p) ( K ' K ) p K 0 and

TV c TV c 'TV c c'c c 0 and

because by Property 2.5.1, p

TV p
0
K

TV c
0
K

K
, and K 0 implies p 0 and c 0 .
1 r

(iii) For S 0 K ,
if K K 'K 0 , TV c c'( S0 K ' ) c (c'c) ( K ' K ) c K 0 and
and

TV p TV p 'TV p p' p p 0 and

TV c
0
K

TV p
0;
K

if K K 'K 0 , TV p p'( K ' S0 ) p ( p' p) ( K ' K ) p K 0 and

TV p
0,
K

TV c TV c 'TV c c'c c 0 and

and

Property 2.7.1

TV c
0.
K

If in equation (1), K is very small relative to S 0 and T is short so that at t T

almost surely ST K , then p 0 , and the European call option price is: c S 0
From Property 2.2, we know c cannot be less than S 0

K 5
.
1 r

K
. For example, a firm provides its
1 r

employees a stock option (a European call option) with K $10.2 , S 0 $200 , r 2% , and T is one
month. It is very unlikely that after one month, the firms stock price will be less than $10.2. Thus,
this call options price at t 0 should be: c S 0
an investor can sell one call option, borrow
$1 ( 191

K
$190 . If not, say, c $191, then at t 0
1 r

$10.2
from a bank, and buy one share of the stock to earn
1 2%

10.2
200) . At t T , ST K , say, ST $100 , the holder of the call option will spend
1 2%

$10.2 to exercise the option, and the investor will obtain zero profit by giving the share of the stock to the
holder and paying $10.2 to the bank.6

For the put option, we have:

Property 2.7.2

If in equation (1), K is very big relative to S 0 and T is short so that at t T

almost surely ST K , then c 0 , and the European put option price is: p

K
S0 .
1 r

Here c is also the value of a forward contract on the non-dividend-paying underlying asset. Suppose there already exists
a forward contract with K as its delivery price. Then, the value of this forward contract f at the current time, i.e., at
t 0 , is f ( F0 K ) /(1 r ) , where F0 is the forward price if both parties negotiated at the current time. Combine this
equation with F0 S 0 (1 r ) , the relationship between the forward price and the spot price, we get f S0 K /(1 r ) .
6

In the Black-Scholes option pricing model, the European call option price is:

c S0 (d1 ) K erT (d2 )

where

d1

ln(

S0
2
) (r
)T
K
2
, d 2 d1 T .
T

If K $10.2 , then S 0 needs to be infinite to make c S0 K erT .

From Property 2.2, we know p cannot be less than

K
S 0 . For example, issuing a European
1 r

put option with K $204, S 0 $10 , r 2% , and T is one month.

It is very unlikely that after one

month, the stock price will be more than $204. Thus, the put options price at t 0 should be:

K
S 0 $190 .
1 r

If not, say, p $191 , then at t 0 an investor can sell one put option,

short-sell one share of the stock, and deposit

$204
204
) . At
in a bank to earn $1 ( 191 10
1 2%
1 2%

t T , ST K , say, ST $100 , the holder of the put option will exercise the option, and the investor
will obtain zero profit by transferring $204 from the bank to the holder and giving back one share of the
stock.

Property 2.7.3

From Properties 2.7.1 and 2.7.2, the Greeks for the European call and put options are:

c
1
p
1
c
K
p
K
p
c

0,

0,

0;

0.
1 0,
1 0 ,
2
K 1 r
K 1 r
r (1 r )
r (1 r ) 2
S0
S0

Property 2.7.4
From Property 2.7.1, c S 0

K
, the time value of the call option is:
1 r

TV c c ( S0 K ) S0
From Property 2.7.2, p

TV c
r
K
r

0.
( S0 K ) K
0 , and
K
1 r
1 r
1 r

K
S 0 , the time value of the put option is:
1 r

TV p p ( K S0 )

TV p
r
K
r

0.
S0 ( K S0 ) K
0 , and
K
1 r
1 r
1 r

Interestingly, the time value of the European put option is negative.7

When K approaches zero,

the time value of the call option will disappear, and C c S 0 , i.e., the call is like a transferrable
restricted stock. These results and Property 2.6 show that, in addition to time, other factors can also

Since American call and put options can be exercised on or before the expiration date, their time values must be
non-negative.

affect the so-called time value of options.8

Property 2.8
(i) For any two non-dividend-paying stocks with the same current share price S 0 , and the same
European call options exercise price K and expiration date T , suppose K is very small relative
to S 0 and T is short so that at t T almost surely ST K .
option prices at the current time must be the same: C c S 0

Then, these two stocks call

K
;
1 r

(ii) For any two non-dividend-paying stocks with the same current share prices S 0 , and the same
European put options exercise price K and expiration date T , suppose K is very big relative to

S 0 and T is short so that at t T almost surely ST K . Then, these two stocks put option
prices at the current time must be the same: p

K
S0 .
1 r

The above results show that an assets call or put option can directly be used as another assets call
or put option. This finding challenges the conventional view that an option must be derived from or
dependent on some specific underlying asset.

Note also that the price behaviours (probability

distributions) of the two underlying assets are irrelevant in pricing these options. To price these options
all we need are: K , r and S 0 , where K is determined by the issuer of options, r is determined by
the economy, and S 0 is influenced by both the economy and the individual company. This indicates
that an assets current price S 0 may reflect not only the assets past information but also the markets
expectations about the assets future performance.

Property 2.9

If c S 0

K
is a European call option for a non-dividend-paying asset, then it can
1 r

K
S0 p where t 0, 1, 2,..., T , and m is the risk-free interest rate in
(1 m)T
each period. If K is very small relative to S 0 and T is short so that at t T almost surely ST K , then p 0 , and
8

For the multi-period put-call parity, c

c
K
0 . If K is very big relative to S 0 and T is short so that
, and
T
T
(1 m)
p
K
0.
at t T almost surely ST K , then c 0 , and the European put option price is: p
S0 , and
T

T
(1 m)
the European call option price is: c S0

also be a European call option for any other non-dividend-paying asset. If p

K
S 0 is a European
1 r

put option for a non-dividend-paying asset, then it can also be a European put option for any other
non-dividend-paying asset.
For example, suppose asset As European call option has K $10.2 where S 0 $200 , r 2% ,
and T is one month. It is very unlikely that after one month, asset As unit price will be less than
$10.2. Thus, the call options price at t 0 is: c S 0

K
$190 . If another asset Bs unit price
1 r

is $50, then we can let the underlying asset be four units of asset B, and this c $190 can also be asset
Bs call option.
Suppose asset Es European put option has K $204 where S 0 $10 , r 2% , and T is one
month. It is very unlikely that after one month, asset Es unit price will be more than $204. Thus, the
put options price at t 0 is: p

K
S 0 $190 . If another asset Fs unit price is $50, then we can
1 r

let the underlying asset be 0.2 units of asset F, and this p $190 can also be asset Fs put option.

Property 2.10

Every non-dividend-paying asset can be written as a European call option and a

European put options for any other non-dividend-paying asset (including itself).
For example, suppose that for two assets, the first assets current unit prices is S 01 $200 , and the
second assets current unit price S02 $198 . Then S 02 can be written as:
(i) a European call option for one unit of the first asset: S 02 S 01

K
, where K $2.04 , r 2% ,
1 r

and T is one month;

(ii) a European put option for one-tenth unit of the first asset: S 02

K
(0.1) S 01 , where K $222.36 ,
1 r

r 2% , and T is one month;

(iii) a European call option for 1.01 units of the second asset: S 02 (1.01) S 02

K
, where K $2.0196,
1 r

r 2% , and T is one month;

(iv) a European put option for one-tenth unit of the second asset: S 02

K
(0.1) S 02 , where
1 r

K $222.156, r 2% , and T is one month.

3. The Arbitrage Theorem and Properties of the Binomial Option Pricing Model

Chang (2012) has introduced the Gordan Theorem/the Arbitrage Theorem (see also Bazaraa et al.,
1993, p. 47).

Gordan Theorem (The Arbitrage Theorem):

Let A be an m n matrix. Then, exactly one of the following systems has a solution:
System 1:

System 2:

Ax 0

for some x R n

Atp 0

1
1

for some p R m , p 0 , e t p 1 where e .

In System 2 of the Arbitrage Theorem, the vector p (which is not the same as the probability
measure in the real world) is usually termed as the risk neutral probability measure, and pi , i 1, ..., m ,
can be interpreted as the current price of one dollar received at the end of period if state i occurs.

If

the matrix A has rank m (i.e., the matrix has m independent rows), the risk neutral probability
measure p will be unique.

m independent rows of A also means a complete market, i.e., every

asset can be replicated by other m assets.9

I now use the following two examples to exemplify these

results.

Example 3.1

Arbitrage and Risk-Neutral Probabilities.

Assume a one-period, two states of nature world with no transaction costs. There are a money
market (Security 1) which provides 1 0.25 dollars at time one if one dollar is invested at time 0 (i.e.,
the risk-free interest rate is r 0.25 ), and two other securities (Security 2 and Security 3) with current
prices $4 and $500, respectively, which at time 1 provide:

In incomplete markets, assets may not be replicated, but with no arbitrage (i.e., System 2 of the Arbitrage Theorem has a
solution), they are still priced by the same (which may not be unique) risk neutral probability measure.

750

S 02 4

S 03 500
2

250

Security 2

Security 3

That is, at time one, when Security 2 provides $8, Security 3 will provide $750; and when Security 2
provides $2, Security 3 will provide $250. In this case, the two securities are not governed by the same
risk neutral probability measure (i.e., System 2 of the Arbitrage Theorem has no solution):

Security 1 :

Security 2 :

Security 3 :

1
1.25 (1 ) 1.25;
1 0.25
' 1 / 2
1 1
1
S 02 4
8 2 ; p'

1 0.25 2
2
1 ' 1 / 2
" 3 / 4
1 3
1

S 03 500
750 250 ; p"

1 0.25 4
4

1 " 1 / 4
S 01 1

i.e., we cannot find a vector p
, 0 1 , such that System 2 holds:
1

1.25 1(1 0.25)

8 4(1 0.25)

750 500(1 0.25)

1.25 1(1 0.25)

2 4(1 0.25)
1
250 500(1 0.25)

0
0 .

0

By System 1 of the Arbitrage Theorem, there must exist arbitrage strategies: e.g., at time 0, we can
short sell one share of Security 3, buy 60 shares of Security 2 and invest $260 ( 500 4 60) in the
money market, and at time 1 we can get net profit:

x1
1.25 1(1 0.25) 8 4(1 0.25) 750 500(1 0.25) 55 0
1.25 1(1 0.25) 2 4(1 0.25) 250 500(1 0.25) 60 = 195 > 0 .

1

When investors adopt this strategy, the time-0 price of Security 2 will go up and that of Security 3 will go
down. In equilibrium (with no arbitrage), the time-0 prices of the two securities will adjust to the point

2 / 3
that they all are priced by the same risk neutral probability measure, say, p
,
1 / 3

Money Mark et (Security 1) :

Security 2 :

Security 3 :

S 01 1

1 2
1

1.25 1.25
1 0.25 3
3

S 02 4.8

1 2
1
8 2
1 0.25 3
3

2
1 2
1

S 03 466
750 250
3 1 0.25 3
3

or

1.25 1(1 0.25)

8 4.8(1 0.25)

2
750 466 (1 0.25)
3

1.25 1(1 0.25)

2 4.8(1 0.25)

2
250 466 (1 0.25)
3

0
2 / 3
1 / 3 0 .

Since, in the above equation, the rank of the matrix is 2 (which equals the number of the states of the

2 / 3
nature), the market is complete, and the risk-neutral probability measure
must be unique. Also,
1 / 3
any security can be replicated by other two assets:

(1) Replicate Security 2s time-1 payment: At time 0, spend $4.8 (or short-sell one share of Security 2)

2
to buy n 0.012 shares of Security 3 and borrow $0.8 ( 466 0.012 4.8) from the money
3
market, and at time 1, we can get:

$8

750

(
1

0
.
25
)(
4
.
8

466
n) if the time - 1 price of Security 3 is $750

$2 250 n (1 0.25)(4.8 466 2 n) if the time - 1 price of Security 3 is $250

3
(2) Replicate Security 3s time-1 payment: At time 0, spend $466
3) to buy n 250/ 3 shares of Security 2 and deposit $66
market, and at time 1, we can get:

2
(or short-sell one share of Security
3

2
2
250
( 466 4.8
) in the money
3
3
3

$750 8 n (1 0.25)(466 3 4.8 n) if the time - 1 price of Security 3 is $8

$250 2 n (1 0.25)(466 2 4.8 n) if the time - 1 price of Security 3 is $2

Example 3.2

Contracts are European Options for One Another.

Assume a one-period, two states of nature world with no transaction costs. There already exists
two assets: a money market with the risk-free interest rate r 0.25 , and a contract (Contract A) with
time-0 price $10, which at time-1 is worth $16 if the economy growth rate is more than 3% and $6 if the
growth rate is less or equal to 3%. Now, suppose that another contract (Contract B) promises at time-1
to pay $60 if the growth rate is more than 3%, and $20 if the growth rate is less or equal to 3%. Then,
what will the time-0 price of Contract B be?

16

60

S 0B ?

S 0A 10
8

20

Contract A

Contract B

1.25
16
The money markets time-1 payment vector
and Contract As time-1 payment vector

1.25
8
are linearly independent, and the number of the linearly independent vectors equals the number of the
states of the nature. Hence, by System 2 of the Arbitrage Theorem, the market is complete, and with no

9 / 16
arbitrage, there exists a unique risk-neutral probability measure p
:
7 / 16

Money Mark et :

Contract A :

or

S 01 1

1 9
7

1.25 1.25
1 0.25 16
16

S02 10

1 9
7

16 8
1 0.25 16
16

1.25 1(1 0.25)

16 10(1 0.25)

1.25 1(1 0.25)

8 10(1 0.25)

9 / 16
7 / 16

0
0 .

Contract Bs time-0 price is S 0B 34 , i.e.,

Money Mark et :

Contract A :

Contract B :

S 01 1

1 9
7

1.25 1.25
1 0.25 16
16

S 0A 10

1 9
7

16 8
1 0.25 16
16

S 0B 34

1 9
7

60 20
1 0.25 16
16

(2)

or

1.25 1(1 0.25)

16 10(1 0.25)

60 34(1 0.25)

1.25 1(1 0.25)

8 10(1 0.25)

20 34(1 0.25)

0
9 / 16
7 / 16 0 .

Equation (2) can also be written as:

Money Mark et :

Contract A :

Contract B :

S 01 1

1 9
7

1.25 1.25
1 0.25 16
16

S 0A 10

1 9
7

(60 44) (20 12)

1 0.25 16
16

S 0B 34

1 9
7

(76 16) (28 8)

1 0.25 16
16

(2)

That is, Contract A can be shown as a European call option c $10 for Contract B, where the exercise
price K is $44 if the time-1 price of Contract B is $60, and $12 if the time-1 price of Contract B is $20.
Contract B, on the other hand, can be shown as a European put option p $34 for Contract A, where
the exercise price K is $76 if the time-1 price of Contract A is $16, and $28 if the time-1 price of
Contract A is $8.

In the following, I will use the Arbitrage Theorem to examine the properties of the binomial option
pricing model.
Assume a one-period, two states of nature world with no transaction costs. There are a money
market with the risk-free interest rate r , and an asset (a stock) with time-0 price S 0 which at time-1
provides S 0 u (u 1) if the economy is good, and S 0 d (0 d 1) if the economy is bad.

The

S0u
1 r
money markets time-1 payment vector
and the stocks time-1 payment vector
are

S0 d
1 r
linearly independent, and the number of the linearly independent vectors equals the number of the states
of the nature. Hence, by System 2 of the Arbitrage Theorem, the market is complete, and with no

arbitrage, there exists a unique risk-neutral probability measure p
. Suppose also that there are
1
one European call option and one European put option based on the same stock with exercise price K :

cu M a [xS 0 u K ,0]

pu M a [xK S 0 u,0]

S0 u

S0 u

S0

c?

S0

p?

S0 d
cd M a [xS 0 d K ,0]

S0 d
pd M a [xK S 0 d ,0]

By System 2 of the Arbitrage Theorem, the prices of these assets are:

Money Mark et (Security 1) :

Asset :

Security 3 :

Security 4 :

1
(1 r ) (1 )(1 r )
1 r
1
S0u (1 ) S0 d
S0
1 r
. (3)
1
Max( S0u K , 0) (1 ) Max( S0 d K , 0)
c
1 r
1
Max( K S0u, 0) (1 ) Max( K S0 d , 0)
p
1 r
1

From equation (3), we can derive the following properties for S 0 u K S 0 d .10

Property 3.1

S0 p

Put-Call Parity.

1
S0u (1 ) S0 d 1 (1 ) ( K S0 d ) 1 ( S0u K ) K
1 r
1 r
1 r
1 r

K
.
1 r

Property 3.2
(i) c

1
( S0u K ) 1 S0u (1 ) S0 d 1 (1 ) S0 d 1 K
1 r
1 r
1 r
1 r

S0

1
K (1 ) S0 d S0 1 K (1 ) K S0 K S0 K .
1 r
1 r
1 r

This indicates that for any American call option C , even if S 0 K 0 , it wont be exercised.
(ii)

1
(1 )( K S0 d ) 1 S0u (1 ) S0 d 1 (1 ) K 1 S0u
1 r
1 r
1 r
1 r

S0

1
S0u (1 ) K S0 1 K (1 ) K K S0 .
1 r
1 r
1 r

Property 3.3
Suppose c p , i.e.,

1
( S0u K ) 1 (1 )( K S0 d ) .
1 r
1 r

Then K S 0 (1 r ) , i.e., the

exercise price must equal the underlying assets forward price.

An example could be:

Money Mark et (Security 1) :

Security 2 :

Security 3 :

Security 4 :

Security 5 :

S 01 1

1 3
1

1.25 1.25
1 0.25 4
4

S 02 48

1 3
1

75 15
1 0.25 4
4

S 03 68

1 3
1

100 40
1 0.25 4
4

c9

1 3
1 3

(75 60)
(100 85)
1 0.25 4
1 0.25 4

p9

1 1
1 1

(60 15)
(85 40)
1 0.25 4
1 0.25 4

where the same call and put option price: c p $9 can be dependent on Security 2 ( S02 ) with

10

For S0u S0d K , c S0

K
K
and p 0 . For K S0u S0d , p
S0 and c 0 .
1 r
1 r

K S 02 (1 r) 48(1 0.25) $60 ,

or

dependent

on

Security

( S03 )

with

K S 30 (1 r )

68(1 0.25) $85 .

Let K K 'K 0 and other factors other factors (i.e., u, d , S 0 and r ) remain

Property 3.4
constant.

Then

c c'c
i.e.,

1
( S0u K ' ) 1 ( S0u K ) 1 ( K ' K ) 0 ,
1 r
1 r
1 r

c
K
0 , and c
because r 0 and 0 1;
K
1 r

p p' p
i.e.,

Property 3.5

1
(1 ) ( K ' S0 d ) 1 (1 ) ( K S0 d ) 1 (1 ) ( K ' K ) 0 ,
1 r
1 r
1 r

p
K
0 , and p
because r 0 and 0 1.
K
1 r
c p
c p

0 and

0.
u u
d d

(i)

Let u u'u 0 and other factors (i.e., K , d , S 0 and r ) remain constant.

S0

1
(1 r ) d (1 r ) d
[ ( S 0 u ) (1 )( S 0 d )] ,

' . with
1 r
ud
u'd

and

p'

Then from

1
[(1 )( K S 0 d )]
1 r

1
p' p p
p' 1 '
[(1 ' )( K S 0 d )] , we have:

0.

1 and
1 r
u'u u
p 1

c p

0 can be derived from:

u u

(1 )
'
(1 ' )

(c'c) ( p' p)
( S 0 u' K )
( S 0 u K )
( K S0 d )
( K S 0 d )
1 r
1 r
1 r
1 r

(ii) Let

1
' ( S0u' ) (1 ' )(S0 d ) 1 ( S0u) (1 )(S0 d ) S0 S 0 0 .
1 r
1 r

d d 'd 0

and other factors (i.e.,

K , u, S 0 and r ) remain constant.

Then

1 '
c

u (1 r ) u (1 r )

1
u d'
ud

or

'

With

c'

1
[ ' ( S 0 u K )]
1 r

and

1
c' '
c'c c
[ ( S 0 u K )] , we obtain:
1 and

0.
1 r
c
d 'd d
c p

0 can be derived from:

d d

(1 )
'
(1 ' )

(c'c) ( p' p)
( S0 u K )
( S 0 u K )
( K S0 d ' )
( K S 0 d )
1 r
1 r
1 r
1 r

Property 3.6

1
' ( S0u) (1 ' )(S0 d ' ) 1 ( S0u) (1 )(S0 d ) S0 S 0 0 .
1 r
1 r

Even without changing the expiration dates, issuers of European options can adjust

exercise price K to change the time value of options.

(i)

If S 0 K , then TV p p Max[ K S0 ,0] p 0 and from Property 3.2, TV c c ( S0 K ) 0 .

Let K K 'K 0 and S 0 K ' , we have:

TV c

TV c TV c 'TV c c'( S0 K ' ) [c ( S0 K )] ( K ' K )1

0
0 and
K
1 r
1
TV p

(1 ) ( K ' K ) 0 and
TV TV 'TV p' p
0.
1 r
K
p

(ii) If S 0 K , then TV c c 0 . Let K K 'K 0 and S 0 K ' ,

TV p
1
TV p TV p 'TV p p'( K ' S0 ) [ p ( K S0 )] ( K ' K )
0
1 0 and
K
1 r
1
TV c
( K ' K ) 0 and
TV TV 'TV c'c
0.
1 r
K
c

(iii) For S 0 K ,
if K K 'K 0 ,

TV c

TV c c'( S0 K ' ) c (c'c) ( K ' K ) ( K ' K )1

0
and

K
1 r

and

TV p TV p 'TV p p' p

p
1
(1 ) ( K ' K ) 0 and TV 0 ;
1 r
K

if K K 'K 0 ,

TV p
1
TV p p'( K ' S0 ) p ( p' p) ( K ' K ) ( K ' K )
0,
1 0 and
K
1 r
and

Property 3.7

TV c TV c 'TV c c'c

c
1
( K ' K ) 0 and TV 0 .
1 r
K

Every non-dividend-paying asset is also a European call option and a European put option

for any other non-dividend-paying asset.

An example could be:

1 3
1

1
Money Mark et (Security 1) : S 0 1 1 0.25 4 1.25 4 1.25

1 3
1

S 02 48
70 30
Security 2 :
1 0.25 4
4

1 3
1

S 03 50
75 25
Security 3 :
1 0.25 4
4

1 3
1
c6
10 0
Call Option :
1 0.25 4
4

1 3
1

p6
Put Option :
0 30
1 0.25 4
4

where
(i) c can be a call option for Security 2 (with K $60 ) or Security 3 (with K $65 ). Also, p can
be a put option for Security 2 (with K $60 ) or Security 3 (with K $55 ).
(ii) One share of Security 2 is:
(a) a European call option for two shares of Security 3 where K $80 if the time-1 share price of
Security 3 is $75 , and K $20 if the time-1 share price of Security 3 is $25 ;
(b) a European put option for one share of Security 3 where K $145 if the time-1 share price of
Security 3 is $75 , and K $55 if the time-1 share price of Security 3 is $25 .
One share of Security 3 is:

(a) a European call option for two shares of Security 2 where K $65 if the time-1 share price of
Security 2 is $70 , and K $35 if the time-1 share price of Security 2 is $30 ;
(b) a European put option for one share of Security 2 where K $145 if the time-1 share price of
Security 2 is $70 , and K $55 if the time-1 share price of Security 2 is $30 .

4. Concluding Remarks

In this paper, I have used simple arbitrage argument to derive a dozen of model-free option price
properties. In addition to deriving the Greeks under model-free framework, the results show that first, in
contrast to the traditional view, a European call (put) option for a non-dividend-paying asset can also be a
European call (put) option for any other non-dividend-paying asset, and every non-dividend-paying asset
is also both a European call option and a European put option for any other non-dividend-paying asset.
Second, in some cases the time value of the European put option can be negative, and adjust the exercise
price of an option can decrease or even erase the time value of the option.

I have also used the Arbitrage

Theorem under the binomial option pricing model to examine these properties.

REFERENCES

Bazaraa, Mokhtar, Hanif Sherali and C. M. Shetty, 1993, Nonlinear Programming: Theory and
Algorithms, John Wiley & Sons, Inc., New York.
Black, Fischer and Myron Scholes, 1973, The Pricing of Options and Corporate Liabilities, Journal of
Political Economy 81, 637-654.
Chang, Kuo-Ping, 2012, Are Securities Also Derivatives?, American Journal of Operations Research 2,
430-441.

doi: 10.4236/ajor.2012.23051.

Also in http://ssrn.com/abstract=987522.

Merton, Robert, 1973, Theory of Rational Option Pricing, The Bell Journal of Economics and
Management Science 4, 141-183.