Black Merton Scholes Option

Pricing Formula

Dilip Madan
Department of Finance
Robert H. Smith School of Business
 Black Merton Scholes
Model
Consider an economy with two assets, a money mar-
ket account earning a continuously compounded in-
terest rate of r with account value at time t of

A(t) = ert;

and a risky stock with asset price process S (t) sat-

isfying the stochastic di¤erential equation
dS
= dt + dW
S

where is the mean rate of return on the stock,

is the volatility of the stock, and W (t) is a standard
Brownian motion with increments W (t + h) W (t)
that are normally distributed with mean zero, vari-
ance h that are independent across nonoverlapping
intervals.
It can be shown that the stock price relative is log
2
normally distributed with mean = 2 and
! !
S (t) 2
ln = t + W (t)
S (0) 2

Equivalently we may write the solution to the sto-

chastic di¤erential equation for the stock or that
! !
2
S (t) = S (0) exp t + W (t) :
2

These are the assumptions on the true or physical

probability law of the stock price.
 On Pricing Options
The price of a European call option with strike K and
maturity T under these assumptions is anticipated to
be some function of the stock price and the time to
maturity or calendar time say
c(S (t); t; K; T )

with the property that

c(S (T ); T ; K; T ) = (S (T ) K )+ :

Black, Merton and Scholes develop a partial di¤er-

ential equation (P DE ) that was then solved with
the above boundary condition for the option price.

We shall present the derivation of the PDE later

though not the details for its solution. Our approach
to the option pricing formula will instead build on
what we have already learned from the binomial ap-
proximation model.
The Risk Neutral Law for
the Stock
We have learned that options are not priced under
the original or true probability but that we have to
change probability to the risk neutral probability.

From the risk neutral law we learn that the rate of

return on the stock is the interest rate on the money
market account.
From the risk neutral trees we learn that the in limit
2
the risk neutral log price relative has a drift of r 2
and
! !
S (t) 2
ln = r t + W (t)
S (0) 2

Risk neutrally we have that

! !
2
S (t) = S (0) exp r t + W (t)
2

and
dS
= rdt + dW:
S
 Risk Neutral Option
Pricing
We know from our risk neutral valuation principle
that
h i
Q +
c(S (t); t; K; T ) = Et e r(T t) (S (T ) K)

or the discounted risk neutral expected value of the

payo¤ or …nal cash ‡ow.

We may therefore determine the option price by di-

rectly evaluating this discounted expectation given
the risk neutral law.
Since all that matters is the time to maturity it suf-
…ces to work with t = 0 as

c(S; t; K; T ) = c(S; 0; K; T t):

Hence we evaluate
h i
+
c(S (0); 0; K; T ) = Q
E e rT (S (T ) K) :
The Black Merton Scholes
Option Pricing Formula
The formula asserts that

c(S (0); 0; K; T ) = S (0)N (d1) Ke rT N (d2)

ln(S (0)=K ) r p
d1 = p + + T
T 2
p
d2 = d1 T:

Recall that N (x) is the standard normal distribution

function or the area under the bell curve to the left
of x:
To derive this formula it is helpful to work in terms
of a standard normal density and to write
p
W (T ) = T Z

where Z is a standard normal variate.

We therefore write the risk neutral stock price as

! !
2 p
S (T ) = S (0) exp r T+ TZ :
2

We wish to evaluate
h i
+
E e rT (S (T ) K)
2 0 1+3
2 p
6 S (0) exp r T+ TZ A 7
= E 4e rT @ 2 5
K
0 p 1+
Z 1 2
= e rT @ S (0) exp r 2 T+ Tz A
1 K
n(z )dz
where n(z ) is the standard normal density
!
1 z2
n(z ) = p exp :
2 2
Eliminating the positive
part
The …rst step is to eliminate the positive part by not-
ing that the we only need to integrate in the region

S (T ) > K

as otherwise the integrand is zero.

Therefore we identify the region in z when S (T ) >

K:

Now this condition requires that

! !
2 p
S (0) exp r T+ Tz >K
2
or that
! !
S (0) 2 p
ln + r T+ Tz > 0
K 2

or equivalently that
!
ln(S (0)=K ) r p
z > p + T
T 2
= d

with
ln(S (0)=K ) r p
d= p + T:
T 2
 Rewrite the call price
We may now write the call price as
0 0 1 1
Z 1 2
r T A
c = e rT @S (0) exp @
p 2 KA
d + Tz
n(z )dz
= I1 I2

where
Z 1 !
p 2T
I1 = S (0) exp Tz n(z )dz
d 2
Z 1
I2 = Ke rT n(z )dz:
d
Evaluating the Second
Integral
The second integral is just the area under the bell
curve to the right of d and by symmetry of the
bell curve this is the area to the left of d which is
N (d):

Hence we have that

I2 = Ke rT N (d):
 Evaluating the First
Integral
The …rst integral is
Z 1 !
p 2T
I1 = S (0) exp Tz n(z )dz
d 2
Z 1 !
1 z2 p 2T
= S (0) p exp + Tz dz
d 2 2 2
Z 1 p !
1 z2 2 Tz + T 2
= S (0) p exp dz
d 2 2
0 p 21
Z 1 z T C
1
= S (0) p exp B
@ A dz
d 2 2

Now change variables to

p
u = z T
du = dz
and write noting that when z = d we have u =
p
d T
Z !
1 u2
I1 = S (0) p p exp du
d T 2 2
Z
= S (0) p n(u)du
d T

The integral here is the area under the bell curve to

p
the right of d T or by symmetry the area to
p p
the left of d + T or N (d + T ):
We thus have that
p
c = S (0)N d + T Ke rT N (d):

We now note that

p ln(S (0)=K ) r p p
d+ T = p + T+ T
T 2
ln(S (0)=K ) r p
= p + + T
T 2
= d1

and
p
d = d2 = d1 T

Hence

c = S (0)N (d1) Ke rT N (d2):

Black Merton Scholes Put
Option Formula
To derive the put option formula we employ put call
parity and write

p(S (0); 0; K; T )
= c(S (0); 0; K; T ) + Ke rT S (0)
= S (0)N (d1) Ke rT N (d2) + Ke rT S (0)
= Ke rT (1 N (d2)) S (0) (1 N (d1))
= Ke rT N ( d2) S (0)N ( d1):
Pricing an Option with the
Black Merton Scholes
formula
Consider a 3 month put option for S (0) = 100 with
strike 90 and a continuously compounded interest
rate of 6% for a volatility of 20%:

The Black Merton Scoles price is a function

w(p; k; r; ; t; u) = w(100; 90; :06; :2; :25; 0)

where the last argument is iscall.

We have that
ln (100=90) :06 :2 p
d1 = p + + :25
:2 :25 :2 2
= 1:253605156
also
p
d2 = 1:2536 :2 :25
= 1:153605156

The option price is

w = 90 exp( :06 :25) N ( 1:153605156)

100 N ( 1:253605156)
= 88:66007456 N ( d2) 100 N ( d1 )
= 88:66007456 :12433104 100 :10499278
= 11:023199 10:499278
= :5239:

If we want four decimal accuracy at the end it is

imperative that intermediate values are stored at a
higher level of accuracy.

Of course when working with tables we lose this level

of accuracy.
Comparing Black Merton
Scholes and Binomial
We now compare the Black Merton Scholes option
price with the binomial model for a one year 90 put
and a one year 110 call that we studied with our
trees for an interest rate of 6% and a 20% volatility.

The binomial model prices are from a risk neutral

tree of various sizes.
n 90P 110C
50 2.1264 6.4022
100 2.1141 6.4537
200 2.1031 6.4301
300 2.1065 6.4429
500 2.1037 6.4357
BMS 2.1044 6.4373
 SPX 201 10419 to 201 2041 8
145 0

140 0

135 0

130 0
SPX level

125 0

120 0

115 0

110 0

105 0
0 50 100 150 200 250 300
day s fro m 20110 419

Figure 1: SPX level 20110419 to 20120418

Black Merton Scholes in

Practice
Consider as underlier the S&P 500 index.

We present a graph of the index over the one year

period 20110419 to 20120418.

The annualized standard deviation of returns over

this period was
v0
u
u 0 12 1
p uB 1 X 1 X A C
0:2312 = 252 u@ rt2 @ rt A
t
252 t 252 t
p
= 252 0:0146

There were 256 OTM (out-of-the-money) options

trading with moneyness levels within 30% for 10 ma-
turities between a month and a year on this day.

We present a graph of the option prices by maturity

in two sets of 5 graphs.
 maturities 1 month to s ix months
70

60

50

40

30

20

10

0
900 1000 1100 1200 1300 1400 1500 1600

maturities s ix months to one y ear

120

100

80

60

40

20

0
900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Implied Volatility
The BMS option price is a function of the form

w = bmsp(p; k; r; q; ; t; u)

where u = iscall:

All the inputs are known except the volatility.

We will adjust for dividend yields q later.

The implied volatility is de…ned as the volatility im-
plied in the Black Merton Scholes formula by the
observed market price.

Given market price of option w(k; t) we solve for

implied volatility IV (k; t) such that

w(k; t) = bmsp(p; k; r; q; IV (k; t); t; u):

We now graph the implied volatilities for our options

on 20120418:
 maturities 1 month to s ix months
0.45

0.4

0.35

0.3

0.25

0.2

0.15

0.1
900 1000 1100 1200 1300 1400 1500 1600

maturities s ix months to one y ear

0.35

0.3

0.25

0.2

0.15

0.1
900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Figure 2: SPX Implied volatilities 20120418

Historical Volatility
Mispricing
We may observe quite an extensive mispricing at his-
torical volatility.
strike maturity iscall BMSP Price IV
1390 .16 1 142.17 32.30 .17
1075 .25 0 0.83 4.85 .32
1500 .70 1 90.07 28.5 .16
1025 .94 0 13.12 30.05 .30
We certainly cannot price options using the Black
Merton Scholes formula using historical volatility or
any constant volatility for that matter.

The real world is far more complicated and the ideal-

ized model assumptions leading to the BMS formula,
though a signi…cant advance are too simple for real
world option pricing.
 Option Quotations
Nonetheless the formula is extensively used …rst and
foremost to quote option prices.

Options are quoted in BMS implied volatilities and

it is often asserted that the implied volatility is the
wrong number to put in the wrong formula to get
the right price.

This is useful for comparing across strike, maturities

and assets much like bond prices are quoted in terms
of implied yields or interest rates.

Just as you need to know how to get the bond price

from a rate quotation, you need the Black Merton
Scholes formula to get the option price from the
quoted volatility.
 Option Hedging
We have seen that the Black Merton Scholes hedge
position in the stock is given by
cu cd @c
m= u
S Sd @S

Recognizing that the current price is only correct if

one uses the implied volatility the hedge position for
a derivatives book on a single underlier is worked out
by evaluating the BMS delta at the implied volatility.

For a position of xk;t options long or short in strike

k and maturity t the hedge for the book is
X @BM SP (S; k; r; q; IV (k; t); t; u)
m= xk;t :
k;t
@S
 The Option Delta
The BM S option delta is
@BM SP (S; k; r; q; ; t; u)
@S

is evaluated for each option at its implied volatility.

For a call option the delta is

N (d1)

while for a put it is

N ( d2):
 110 one y ear c all delta
1

0.8

0.6

delta
0.4

0.2

0
70 80 90 100 110 120 130
s toc k pric e

90 one y ear put delta

0

-0.2

-0.4
delta

-0.6

-0.8

-1
70 80 90 100 110 120 130
s toc k pric e

We present a graph of the call and put deltas for our

110 one year call and the 90 put as a function of the
stock price.

We observe that as the call gets in the money the

delta goes to unity and zero at the other end.

For the put it goes to negative unity as the stock

drops and to zero at the other end.

The call option hedger buys as the market rises while

the put option hedger sells as the market drops.
 Theta and Gamma
The next two theoretically important option Greeks
are the option theta and the option gamma.

The theta is the derivative of option value with re-

spect to calendar time which is the negative of the
derivative with respect to maturity
@BM SP (S; k; r; q; ; t; u)
>0
@t

As calendar time increases, time to maturity de-

creases, and if nothing else changes, the option price
will fall to eventually coincide with the payo¤.

The call option theta is then negative and the put

option theta is also negative for zero rates but can
be positive for positive rates as the present value of
the strike rises as calendar time increases and time
to maturity decreases.
 11 0 o ne y ea r c all the ta
0

-2

th eta
-4

-6

-8
70 80 90 10 0 11 0 12 0 13 0
s toc k pric e

90 on e y ear put th eta

3

1
th eta

-1

-2
70 80 90 10 0 11 0 12 0 13 0
s toc k pric e

We present graphs of the theta for our two options

at the 6% rate.
 The Option Gamma
The option gamma is a measure of the degree of
convexity of the option price and is given by
@ 2BM SP (S; k; r; q; ; t; u)
@S 2

The greater the gamma the poorer the delta as a

hedge and the greater is hedge loss for a …xed hedge
rebalancing interval.

We present a graph of the option prices as a function

of the stock price and the respective gammas.

We observe that the gammas are greatest when the

price convexity is at its highest.
 C all an d Put
30

25

11 0 C all

20

90 put
Optio n Pr ice

15

10

0
70 80 90 10 0 11 0 12 0 13 0
s toc k pr ic e

Figure 3: Call and Put Prices as a function of the spot.

C all an d Put Gammas

0.025

0.02

0.015
Gamma

0.01

0.005

0
70 80 90 10 0 11 0 12 0 13 0
Stoc k Pric e

Figure 4: Call and Put Gammas

Black Merton Scholes PDE
Recall the original BMS assumptions of a money
market account satisfying

dA = rAdt

and a risky stock satisfying

dS
= dt + dW
S

that we write in discrete time approximations as

A = rA t

and

S = S t+ S W !
p W
= S t+ S t p
t
We recognize now that by construction
W
p =Z
t
a standard normal variate.
Consider
p
( S )2 = S t+ S tZ
= 2S 2 tZ 2 + o( t)

Where o( t) denotes terms that go to zero faster

than t:

We may note that since ( S )2 involves a term of or-

der t it was recognized that in building approxima-
tions of the e¤ects of random variables on functions
of these random variables one should take expansions
to second order with respect to such random e¤ects.

In fact to …rst order in t we may write

( S )2 2S 2 t

given that E [Z 2] = 1:
Approximating a change in
call values
Suppose we have a call price function say c(S; t) and
we wish to approximate the change in call value over
a small interval of length t:

A second order expansion in S and …rst order is called

for and we write
@c @c 1 @ 2c 2
c t+ S+ ( S )
@t @S 2 @S 2
 The BMS PDE
The PDE for the call price asserts that
@c @c 2S 2 @ 2c
+ rS + 2
= rc
@t @S 2 @S
To etsablish the argument proceeds by considering
a portfolio that is short one call option and long m
shares as a hedge.

The initial portfolio value is

V = c + mS:

Next we develop the expression for the change in

value V as

V = c + m S:
 The Change in Value
We may the write
@c @c 1 @ 2c 2
V = t S ( S )
@t @S 2 @S 2
+m S

We next introduce the expressions for S in terms

of t and W:
The e¤ects of time and
uncertainty
We then obtain the expression
@c @c
V = t ( S t + S W)
@t @S
1 @ 2c 2 2
2
S t
2 @S
+m ( S t + S W )
Collecting terms in t and W we have
0 1
@c @c S
V = @ @t @S A t
2
1 @ c 2S 2 + m S
2 @S 2
@c
+m S W
@S

It is now clear that the choice of delta hedging with

@c
m=
@S

removes the e¤ect of uncertainty on the change in

value
With result that
!
@c 1 @ 2c 2S 2
V = t
@t 2 @S 2

However this is a deterministic change in value and

hence it must equal the interest on value or

V = rV t
@c
= r c+ S t:
@S
Equating the two expressions for V we have
!
@c @c 1 @ 2c 2S 2
r c+ S =
@S @t 2 @S 2

or that
@c @c 2S 2 @ 2c
+ rS + = rc
@t @S 2 @S 2

or the BMS PDE.

Interestingly not only is the e¤ect of uncertainty

eliminated, but the …nal equation makes no reference
to the mean rate of return and hence the solution is
independent of the stock drift.
The original derivation of
the BMS formula
The original paper sought a function c(S; t) satisfy-
ing the PDE with the boundary condition

c(S; T ) = (S K )+ :

The result was obtained by making various changes

of variable that reduce the equation down to the
heat equation in Physics for which the solutions were
known.

The binomial approximation and the associated risk

neutral approach of pricing under a change of mea-
sure was developed later as people tried to under-
stand what in the world did the heat equation have
to do with the price of an option on the stock.
The PDE from a Traders
Perspective
We rewrite the equation as
@c @c 2S 2 @ 2c
=r S c + 2
:
@t @S 2 @S

The left hand side is the negative of the theta and

as we have seen the option theta is negative and
its negative is the income to be made on selling an
option and waiting to maturity as its value drops to
payout.

The …rst term on the right hand side is the cost of

the hedge as we buy delta shares, borrow the value
less the call premium and incur the interest cost of
the loan.

We now turn our attention to the last term which is

half the gamma times the stock variance.
half gamma times variance
When one hedges a call option as a liability with
delta shares one may then evaluate the quality of
the mishedge.

Locally one may write as a good approximation that

@c
c= S + a( S )2
@S

or that
@c
c(S + x) c(S ) = x + ax2
@S

Taking derivatives with respect to x we see that

@c(S + x) @c
= + 2ax:
@S @S
Taking a second derivative and evaluating at x = 0
we see that
@ 2c
2
= 2a
@S

or that
1 @ 2c
a= :
2 @S 2

The gap between the change in the liability and the

hedge return
@c 1 @ 2c 2
c S= ( S )
@S 2 @S 2

or half gamma times stock variance.

The stock hedge is not perfect and the gap on a

market down or up move is positive and equal to
half gamma times stock variance.
 H edg e Gap o n Both Side s
14

12
H edg e Gap U p

10

8 H edg e Gap D o w n

6
Value

-2

-4
80 85 90 95 100 105 110 115 120
Stoc k Pric e

Figure 5: Hedge Gaps on Both Sides as the market rises

or falls

Theta as compensation
The negative theta is earned on an option sale as
compensation for the interest cost of the hedge and
the expected level of mishedging given by half gamma
times variance.

For short time intervals like a day, we may ignore

interest costs and write the pro…t and loss as
@c 1 @ 2c 2
t ( S )
@t 2 @S 2

For small moves in the stock we expect a pro…t while

for large moves there is a loss.

The break even move for a day is given by

v
u @c
u2 t
u @t
j Sj = t
@ 2c
@S 2

High gamma values reduce the break even level of

the stock move.

For a 120 call with interest rate 6% and volatility :2

and a one year maturity of one year we have
price = 3:5095
theta = 5:1163
gamma = 0:0175
The break even for a day is
q
2 5:1163 (1=365)=0:0175 = 1:2675: