P A R T V I Options, Futures, and Other Derivatives

21.3 Binomial Option Pricing

Two-State Option Pricing
A complete understanding of commonly used option-valuation formulas is difficult with
out a substantial mathematics background. Nevertheless, we can develop valuable insight into option valuation by considering a
simple special case. Assume that a stock price can take only two possible values at option expiration: The stock will either
increase to a given higher price or decrease to a given lower price. Although this may seem an extreme sim
plification, it provides a useful introduction to more complicated and realistic models. Moreover, it can be extended to describe far
more reasonable specifications of stock price behavior. In fact, several major financial firms employ variants of this simple model
to value options and securities with option-like features.
Suppose the stock now sells at S0 = $100, and the price will either increase by a factor of u = 1.20 to $120 (u stands for “up”) or
fall by a factor of d = .9 to $90 (d stands for “down”) by year-end. A call option might specify an exercise price of $110 and a time
to expiration of 1 year. The interest rate is 10%. At year-end, the payoff to the call will be either $0, if the stock price falls, or $10,
if the stock price increases to $120.
depends on the stock price at year-end: $ $
These possibilities are illustrated by the following value “trees”:
$ $

Stock price Call option value

Now compare the payoff of the call to that of a portfolio consisting of one share of the stock and borrowing of $81.82 at the
interest rate of 10%. The payoff of this portfolio also

Value of stock at year-end

− Repayment of loan with interest −−
Total

We know the cash outlay to establish the portfolio is $18.18: $100 for the stock, less the
$81.82 proceeds from borrowing. Therefore the portfolio’s value tree is

The payoff of this portfolio is exactly three times that of the call option for either value
of the stock price. Because the portfolio replicates the payoff of the three calls, we call it a
replicating portfolio. Moreover, because their payoffs are the same, the three calls and the
replicating portfolio must have the same value. Therefore,

3C = $18.18

or each call should sell at C = $6.06. Thus, given the stock price, exercise price, interest
rate, and volatility of the stock price (as represented by the spread between the up or
down movements), we can derive the fair value for the call option.
C H A P T E R Option Valuation $ $
$$
This valuation approach relies heavily on the notion of replication. With only two pos sible
end-of-year values of the stock, the payoffs to the levered stock portfolio replicate the payoffs to
three call options and, therefore, command the same market price. Replication is behind most
option-pricing formulas. For more complex price distributions for stocks, the replication technique
is correspondingly more complex, but the principles remain the same.
One way to view the role of replication is to note that, using the numbers assumed for this
example, a portfolio made up of one share of stock and three call options written is perfectly
hedged. Its year-end value is independent of the ultimate stock price:

Stock value
− Obligations from calls written− −
Net payoff

The investor has formed a riskless portfolio, with a payout of $90. Its value must be the
present value of $90, or $90/1.10 = $81.82. The value of the portfolio, which equals $100
from the stock held long, minus 3C from the three calls written, should equal $81.82.
Hence $100 − 3C = $81.82, or C = $6.06.
The ability to create a perfect hedge is the key to this argument. The hedge locks in the
end-of-year payout, which therefore can be discounted using the risk-free interest rate.
To find the value of the option in terms of the value of the stock, we do not need to know
either the option’s or the stock’s beta or expected rate of return. When a perfect hedge
can be established, the final stock price does not affect the investor’s payoff, so the
stock’s risk and return parameters have no bearing.
The hedge ratio of this example is one share of stock to three calls, or one-third. This
ratio has an easy interpretation in this context: It is the ratio of the range of the values of
the option to those of the stock across the two possible outcomes. The stock, which
originally sells for S0 = 100, will be worth either d × $100 = $90 or u × $100 = $120, for a
range of $30. If the stock price increases, the call will be worth Cu = $10, whereas if the
stock price decreases, the call will be worth Cd = 0, for a range of $10. The ratio of
ranges, 10/30, is one-third, which is the hedge ratio we have established.
The hedge ratio equals the ratio of ranges because the option and stock are perfectly
correlated in this two-state example. Because they are perfectly correlated, a perfect
hedge requires that they be held in a fraction determined only by relative volatility. We
can generalize the hedge ratio for other two-state option problems as

H = Cu − Cd
uS0 − d S0
where Cu or Cd refers to the call option’s value when the stock goes up or down, respec
tively, and uS0 and dS0 are the stock prices in the two states. The hedge ratio, H, is the
ratio of the swings in the possible end-of-period values of the option and the stock. If the
investor writes one option and holds H shares of stock, the value of the portfolio will be
unaffected by the stock price. In this case, option pricing is easy: Simply set the value of
the hedged portfolio equal to the present value of the known payoff. Using our example,
the option-pricing technique would proceed as follows:

1. Given the possible end-of-year stock prices, uS0 = 120 and dS0 = 90, and the
exercise price of 110, calculate that Cu = 10 and Cd = 0. The stock price range
is 30, while the option price range is 10.
2. Find that the hedge ratio of 10/30 = 1⁄3.
3. Find that a portfolio made up of 1⁄3 share with one written option would have an
end-of-year value of $30 with certainty.
P A R T V I Options, Futures, and Other Derivatives
4. Show that the present value of $30 with a 1-year interest rate of 10% is $27.27.
 5. Set the value of the hedged position to the present value of the certain payoff:
1
$

3S0 − C0 = $27.27
$33.33 − C0 = $27.27
6. Solve for the call’s value, C0 = $6.06.
What if the option is overpriced, perhaps selling for $6.50? Then you can make arbitrage profits.
Here is how:

Total $ $ . $ .
. Write options. $ $−. Purchase share −
. Borrow $. at % interest;

repay in year

Although the net initial investment is zero, the payoff in one year is positive and
riskless. If the option were underpriced, one would simply reverse this arbitrage strategy:
Buy the option, and sell the stock short to eliminate price risk. Note, by the way, that the
present value of the profit to the arbitrage strategy above exactly equals three times the
amount by which the option is overpriced. The present value of the risk-free profit of
$1.45 at a 10% interest rate is $1.318. With three options written in the strategy above,
this translates to a profit of $.44 per option, exactly the amount by which the option was
overpriced: $6.50 versus the “fair value” of $6.06.

Concept Check .
Suppose the call
option had been
underpriced,
selling at $.. Formulate the arbitrage strategy to exploit the mispricing, and show that it provides a riskless cash flow in
one year of $. per option purchased. Compare the present value of this cash flow to the option mispricing.

Generalizing the Two-State Approach

Although the two-state stock price model seems simplistic, we can generalize it to incor
porate more realistic assumptions. To start, suppose we were to break up the year into
two 6-month segments, and then assert that over each half-year segment the stock price
could take on two values. In this example, we will say it can increase 10% (i.e., u = 1.10)
or decrease 5% (i.e., d = .95). A stock initially selling at 100 could follow these possible
paths over the course of the year:
 .

.
C H A P T E R Option Valuation
The midrange value of 104.50 can be attained by two paths: an increase of 10% followed by a
decrease of 5%, or a decrease of 5% followed by a 10% increase. There are now three possible
end-of-year values for the stock and three for the option:

Cuu Cu
C Cud = Cdu Cd
Cdd

Using methods similar to those we followed above, we could value Cu from knowledge of Cuu and
Cud, then value Cd from knowledge of Cdu and Cdd, and finally value C from knowledge of Cu and
Cd. And there is no reason to stop at 6-month intervals. We could next break the year into four
3-month units, or twelve 1-month units, or 365 1-day units, each of which would be posited to have
a two-state process. Although the calculations become quite numerous and correspondingly
tedious, they are easy to program into a computer, and such computer programs are used widely
by participants in the options market.

Example 21.1 Binomial Option Pricing

Suppose that the risk-free interest rate is % per -month period and we wish to value a call option with exercise price
$ on the stock described in the two-period price tree just above. We start by finding the value of Cu. From this point,
= $ (because at this point the stock price is u × u × S
the call can rise to an expiration date value of Cuu = $) or fall to
$
= (because at this point, the stock price is u × d × S
a final value of Cud = $., which is less than the $ exercise price).
Therefore the hedge ratio at this point is

H = Cuu − Cud
_____________
uuS − udS = $−

$−. =

Thus, the following portfolio will be worth $ at option expiration regardless of the ultimate stock price:

udS

Buy shares at price uS

Write calls at price Cu − Total $

 The portfolio must have a current market value equal to the present value of $:

× − Cu = $/. = $.

Solve to find that Cu = $..

Next we find the value of Cd. It is easy to see that this value must be zero. If we reach this point
(corresponding to a stock price of $), the stock price at option expiration will be either $. or $.; in
either case, the option will expire out of the money. (More formally, we could note that with Cud =
Cdd = , the hedge ratio is zero, and a portfolio of zero shares will replicate the payoff of the call!)
Finally, we solve for C using the values of Cu and Cd. Concept Check . leads you through the
calculations that show the option value to be $..
P A R T V I Options, Futures, and Other Derivatives
Concept Check .

Show that the initial value of the call option in Example . is $.. a. Confirm that the spread in option values is Cu
− Cd = $..
b. Confirm that the spread in stock = $.
values is uS
− dS
c. Confirm that the hedge
ratio is . shares purchased
for each call written.
d.
Demonstrate that the value in one period of a portfolio comprised of . shares and one call written is riskless. e.
Calculate the present value of this payoff.
f. Solve for the option value.

Making the Valuation Model Practical

As we break the year into progressively finer subintervals, the range of possible year-end
stock prices expands. For example, when we increase the number of subperiods to
three, the number of possible stock prices increases to four, as demonstrated in the
following stock price tree:

uS
uS
dS
S
uSdS

udS
udS
dS

udS

Thus, by allowing for an ever-greater number of subperiods, we can overcome one of the
apparent limitations of the valuation model: that the number of possible end-of-period
stock prices is small.
Notice that extreme events such as u3S0 or d3S0 are relatively rare, as they require
either three consecutive increases or decreases in the three subintervals. More
moderate, or mid range, results such as u2dS0 can be arrived at by more than one
path—any combination of two price increases and one decrease will result in stock price
u2dS0. There are three of these paths: uud, udu, duu. In contrast, only one path, uuu,
 results in a stock price of u3S0. Thus midrange values are more likely. As we make the
model more realistic and break up the option maturity into more and more subperiods,
the probability distribution for the final stock price begins to resemble the familiar
bell-shaped curve, with highly unlikely extreme outcomes and far more likely midrange
outcomes. The exact probability of each outcome is given by the binomial probability
distribution, and this multiperiod approach to option pricing is therefore called the
binomial model.
But we still need to answer an important practical question. Before the binomial model can
be used to value actual options, we need a way to choose reasonable values for u and d. The
spread between up and down movements in the price of the stock reflects the volatility of its
rate of return, so the choice for u and d should depend on that volatility. Call σ your estimate of
the standard deviation of the stock’s continuously compounded annualized rate of return, and Δt
the length of each subperiod. To make the standard deviation of the stock in the binomial model
___
match your estimate of σ, it turns out that you can set u = exp(σ Δt )
___ 3
and d = exp(−σ Δt ). You can see that the proportional difference between u and d increases
with both annualized volatility as well as the length of the subperiod. This makes sense, as both
higher σ and longer holding periods make future stock prices more uncer tain. The following
example illustrates how to use this calibration.
3
Notice that d = 1/u. This is the most common, but not the only, way to calibrate the model to empirical volatility.
For alternative methods, see Robert L. McDonald, Derivatives Markets, 3rd ed. (Boston: Pearson/Addison
Wesley, 2013), Ch. 10.
CHAPTER Option Valuation Probability Final Stock Price
Example 21.2 Calibrating u and d to Stock Volatility

Suppose you are using a -period model to value a -year option on a stock with volatil ity (i.e., annualized standard
deviation) of σ = .. With a time to expiration of T = year, and three subperiods, you would calculate Δt = T/n = /, u =
exp (σ Δt ) = exp (. / ) =
. and d = exp (−σ Δt) = exp (−. / ) = .. Given the probability of an up move ment, you could then work out the
probability of any final stock price. For example, suppose the probability that the stock price increases is . and the
probability that it decreases is .. Then the probability of stock prices at the end of the year would be as follows:

Event Possible
Paths

down ddd .=.

movement ddu, × .×.=.
s dud,
down udd
and up uud,
udu, × .×.=.
down
duu
and up
uuu
up
.=.
movement
s

We plot this probability distribution in Figure ., Panel A. Notice that the two middle end-of
period stock prices are, in fact, more likely than either extreme.
 and subperiods. u Δ
Now we can extend Example 21.2 by breaking up the option maturity into ever-shorter
subintervals. As we do, the stock price distribution becomes increasingly plausible, as we
demonstrate in Example 21.3.

Example 21.3 Increasing the Number of Subperiods In Example ., we broke up

the year into three subperiods. Let’s now look at the cases of

Subperiods, n Δt = T/n = exp (σ Δt ) d=exp (−σt )

. exp(.) = . exp(−.) =.
. exp(.) = . exp(−.) =.
. exp(.) = . exp(−.) =.

We plot the resulting probability distributions in Panels B and C of Figure ..

4
Using this probability, the continuously compounded expected rate of return on the stock is .10. In general, the
formula relating the probability of an upward movement to the annual expected rate of return, r, is p = exp(rΔt
)−d
5
u − d. We adjust the probabilities of up versus down movements using the formula in footnote 4 to make the
distribu tions in Figure 21.5 comparable. In each panel, p is chosen so that the stock’s expected annualized,
continuously compounded rate of return is 10%.
P A R T V I Options, Futures, and Other Derivatives
Notice that the right tail of the distribution in Panel C is noticeably longer than the left
tail. In fact, as the number of intervals increases, the distribution progressively approaches
the skewed log-normal (rather than the symmetric normal) distribution. Even if the stock
price were to decline in each subinterval, it can never drop below zero. But there is no
corresponding upper bound on its potential performance. This asymmetry gives rise to the
skewness of the distribution.

5 5 75 5 5 75 5 5
5 5 75 5 5 75 5 5
.45 .4 .35
A .35 .3 .5
B

.3 Probability

.5 . .5 . .5
 5 5 75 5 5 75 5 5

Probability
.
.5
.
.5

Final Stock Price Final Stock Price

C

.5 Probability
. .5

Final Stock Price

Figure 21.5
Probability
distributions for final stock price. Possible outcomes and associated probabilities. In each panel, the stock’s
annualized, continuously compounded expected rate of return is % and its standard deviation is %. Panel A. Three
subintervals. In each subinterval, the stock can increase by .% or fall by
.%. Panel B. Six subintervals. In each subinterval, the stock can increase by .% or fall by .%.
 Panel C. Twenty subintervals. In each subinterval, the stock can increase by .% or fall by .%.

Eventually, as we divide the option maturity into an ever-greater number of subin

tervals, each node of the event tree corresponds to a smaller and smaller time interval.
The possible stock price movement within each time interval is correspondingly small. As
those many intervals pass, the end-of-period stock price more and more closely
WORDS FROM THE STREE We pointed out earlier in the
A Risk-Neutral Shortcut T
chapter that the binomial model

This implies that p = + rf ________

−d
to distinguish it from the true, or “objective,” probability. To
valuation approach is arbitrage-based. We can value the
illustrate, in our two-state example at the beginning of
u − d . We call p a risk-neutral probability
option by replicating it with shares of stock plus borrow
ing. The ability to replicate the option means that its price
Section ., we had u = ., d = ., and rf = .. Given these
relative to the stock and the interest rate must be based

values, p = + . − .

.−. =
only on the technology of replication and not on risk ences. It cannot depend on risk aversion or the
capital asset
prefer its expected cash flow at the risk-free rate: C
.
= “E”(CF)
Now let’s see what happens if we use the discounted cash
+ rf . We
pricing model or any other model of equilibrium risk-return
relationships. continue to use the two-state example from Section ..
flow formula to value the option in the risk-neutral economy. We
This insight—that the pricing model must be independent of
We find the present value of the option payoff using the risk
risk aversion—leads to a very useful shortcut to valuing options.
neutral probability and discount at the risk-free interest rate:
Imagine a risk-neutral economy, that is, an economy in which
_______ u
C = “E” (CF) + rf = pC ______________
+ ( − p)Cd
all investors are risk-neutral. This hypothetical economy must cannot affect the valuation formula.
value options the same as our real one because risk aversion
This answer exactly matches the value we found using our no
/× +/×
+ rf = . =.
In a risk-neutral economy, investors would not demand risk
arbitrage approach!
premiums and would therefore value all assets by discounting
We repeat: This is not truly an expected discounted value.
expected payoffs at the risk-free rate of interest. Therefore, a
• The numerator is not the true expected cash flow from the
security such as a call option would be valued by discounting
option because we use the risk-neutral probability, p, rather
than the true probability.
put the expectation operator E in quotation marks to signify
• The denominator is not the proper discount rate for option
that this is not the true expectation, but the expectation that
cash flows because we do not account for the risk.
would prevail in the hypothetical risk-neutral economy. To be
• In a sense, these two “errors” cancel out. But this is not just
consistent, we must calculate this expected cash flow using the
luck: We are assured to get the correct result because the
rate of return the stock would have in the risk-neutral economy,
 no-arbitrage approach implies that risk preferences can
not using its true rate of return. But if we successfully maintain
not affect the option value. Therefore, the value computed
consistency, the value derived for the hypothetical economy
should match the one in our own. our economy.
for the risk-neutral economy must equal the value that we obtain in
How do we compute the expected cash flow from the
When we move to the more realistic multiperiod model, the
option in the risk-neutral economy? Because there are no
calculations are more cumbersome, but the idea is the same.
risk premiums, the stock’s expected rate of return must equal
Footnote shows how to relate p to any expected rate of
the risk-free rate. Call p the probability that the stock price
return and volatility estimate. Simply set the expected rate of
increases. Then p must be chosen to equate the expected rate
return on the stock equal to the risk-free rate, use the resulting
of increase of the stock price to the risk-free rate (we ignore
dividends here): at the risk-free rate, and you will find the option value.
probability to work out the expected payoff from the option, discount
“E”(S) = p(uS) + ( − p)dS = ( + rf)S
These calculations are actually fairly easy to program in Excel.

resembles a lognormal distribution.6 Thus the apparent oversimplification of the two-state

model can be overcome by progressively subdividing any period into many subperiods. At
any node, one still can set up a portfolio that is perfectly hedged over the next tiny time
interval. Then, at the end of that interval, on reaching the next node, a new hedge ratio
can be computed and the portfolio composition could be revised to remain hedged

6
Actually, more complex considerations enter here. The limit of this process is lognormal only if we assume also that
stock prices move continuously, by which we mean that over small time intervals only small price move ments can
occur. This rules out rare events such as sudden, extreme price moves in response to dramatic infor mation (like a
takeover attempt). For a treatment of this type of “jump process,” see John C. Cox and Stephen A. Ross, “The
Valuation of Options for Alternative Stochastic Processes,” Journal of Financial Economics 3 (January–March 1976),
pp. 145–66; or Robert C. Merton, “Option Pricing When Underlying Stock Returns Are Discontinuous,” Journal of
Financial Economics 3 (January–March 1976), pp. 125–44.
P A R T V I Options, Futures, and Other Derivatives
over the coming small interval. By continuously revising the hedge position, the port
folio remains hedged and earns a riskless rate of return over each interval. This is called dynamic
hedging, the continued updating of the hedge ratio as time passes. As the dynamic hedge
becomes ever finer, the resulting option-valuation procedure becomes more precise. The nearby
box offers further refinements on the use of the binomial model.

Concept Check .

In the table in Example ., u and d both get closer to (u is smaller and d is larger) as the time interval Δt
shrinks. Why does this make sense? Does the fact that u and d are each closer to mean that the total
volatility of the stock over the remaining life of the option is lower?

21.4 Black-Scholes Option Valuation

While the binomial model is extremely flexible, a computer is needed for it to be useful in actual
trading. An option-pricing formula would be far easier to use than the tedious algorithm involved in
the binomial model. It turns out that such a formula can be derived if one is willing to make just two
more assumptions: that both the risk-free interest rate and stock price volatility are constant over
the life of the option. In this case, as the time to expiration is divided into ever-more subperiods,
the distribution of the stock price at expi
ration progressively approaches the lognormal distribution, as suggested by Figure 21.5. When the
stock price distribution is actually lognormal, we can derive an exact option
pricing formula.

The Black-Scholes Formula

Financial economists searched for years for a workable option-pricing model before Black and
Scholes7 and Merton8 derived a formula for the value of a call option. Scholes and Merton shared
the 1997 Nobel Prize in Economics for their accomplishment.9 Now widely used by options market
participants, the Black-Scholes pricing formula for a call option is
C0 = S0N(d1) − Xe−rTN(d2) (21.1) where

d1 = ln(S0 /X ) + (r + σ2/2)T
σT
d2 = d1 − σ T

7
Fischer Black and Myron Scholes, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy 81
(May–June 1973).
8
Robert C. Merton, “Theory of Rational Option Pricing,” Bell Journal of Economics and Management Science 4 (Spring
1973).
9
Fischer Black passed away in 1995.
C H A P T E R Option Valuation
and

C0 = Current call option value.

S0 = Current stock price.
N(d) = The probability that a random draw from a standard normal distribution will be less than d.
This equals the area under the normal curve up to d, as in the shaded area of Figure 21.6. In
Excel, this function is called NORMSDIST(d) or NORM.S.DIST(d, TRUE).
X = Exercise price.
e = The base of the natural log function, approximately 2.71828. In Excel, ex can be evaluated
using the function EXP(x).
r = Risk-free interest rate (the annualized continuously compounded rate on a safe asset with the
same maturity as the expiration date of the option, which is to be distinguished from rf , the discrete
period interest rate).
T = Time to expiration of option, in years.
ln = Natural logarithm function. In Excel, ln(x) can be calculated using the built-in function LN(x).
σ = Standard deviation of the annualized continuously compounded rate of return of the stock.
Notice a surprising feature of Equation 21.1: The option value does not depend on the expected
rate of return on the stock. In a sense, this information is already built into the for mula with the
inclusion of the stock price, which itself reflects the stock’s risk and return characteristics. This
version of the Black-Scholes formula is predicated on the assumption that the stock pays no
dividends.
Although you may find the Black-Scholes formula intimidating, we can explain it at a somewhat
intuitive level. The trick is to view the N(d) terms (loosely) as risk-adjusted probabilities that the call
option will expire in the money. First, look at Equation 21.1 assuming both N(d) terms are close to
1.0, that is, when there is a very high probability the option will be exercised. Then the call
option value is equal to S0 − Xe−rT, which
is what we called earlier the adjusted intrin sic
value, S0 − PV(X). This makes sense; if
exercise is certain, we have a claim on a stock
, and an obliga
with current value S0 tion with
present value PV(X), or, with con tinuous N(d) =
Shaded area
compounding, Xe−rT.
Now look at Equation 21.1 assuming the
N(d) terms are close to zero, meaning the
option almost certainly will not be exercised.
Then the equation confirms that the call is
worth nothing. For middle-range values of N(d) d
0
between 0 and 1, Equation 21.1 tells us that
the call value can be viewed as the present
value of the call’s potential payoff adjusting for How do the N(d) terms serve as risk
the probability of in-the-money expiration.

Figure 21.6 A standard normal curve

adjusted probabilities? This question quickly

P A R T V I Options, Futures, and Other Derivatives
leads us into advanced statistics. Notice, however, that ln(S0/X), which appears in the numerator of
d1 and d2, is approximately the percentage amount by which the option is currently in or out of the
money. For example, if S0 = 105 and X = 100, the option is 5% in the money, and ln(105/100) =
.049. Similarly, if S0 = 95, the option is 5% out of the money, and ln(95/100) = −.051. The
denominator, σ T , adjusts the amount by which the option is in or out of the money for the volatility
of the stock price over the remaining life of the option. An option in the money by a given percent
is more likely to stay in the money if both stock price volatility and time to expiration are low.
Therefore, N(d1) and N(d2) increase with the probability that the option will expire in the money.

Example 21.4 Black-Scholes Valuation

You can use the Black-Scholes formula fairly easily. Suppose you want to value a call option under the following
circumstances:
=
Stock price: S
Exercise price: X =
Interest rate: r = . (% per year)
Time to expiration: T = . ( months or one-quarter of a year)
Standard deviation: σ = . (% per year)

First calculate

= ln( /) + (. + .
/ ).
d
=.
..

d
=.−.. =.
 Next find N(d) and N(d). The values of the normal distribution are tabulated and may be found in
many statistics textbooks. A table of N(d) is provided here as Table .. The nor
mal distribution function, N(d), is also provided in any spreadsheet program. In Microsoft Excel, for
example, the function name is NORMSDIST or NORM.S.DIST. Using either Excel or Table . we
find that

N(.) = .
N(.) = .

Thus the value of the call option is

C = × . − e−.× . × .
= . − . = $.

Concept Check .

Recalculate the value of the call option in Example . using a standard deviation of . instead of .. Confirm that
the option is worth more using the higher stock-return volatility.

What if the option price in Example 21.4 were $15 rather than $13.70? Is the option
mispriced? Maybe, but before betting your fortune on that, you may want to reconsider
the valuation analysis. First, like all models, the Black-Scholes formula is based on some
simplifying abstractions that make the formula only approximately valid.
C H A P T E R Option Valuation N N N N N N

d(d) d(d) d(d) d(d) d(d) d(d)

 −. . −. . .. .. . . . . . . . . . . . .
−.. −.. ..
.............
−. . −. . ..
−.. −.. .. .............
−. . −. . ..
.............
−.. −.. ..
−. . −. . −. . .. .....
−.. −.. −.. ..
..
−. . −. . −. . ..
−.. −.. −.. .. ..
−. . −. . −. . .. ..
−.. −.. −.. .. ..
−. . −. . −. . .. ..
−.. −.. −.. .. ..
−. . −. . −. . .. ..
−.. −.. −.. .. ..
−. . −. . −. . .. ..
−.. −.. −.. ..
..
−. . −. . −. . ..
..
−.. −.. −.. ..
−. . −. . −. . ..
−.. −.. −.. ..
−. . −. . −. . ..
−.. −.. −.. ..
−. . −. . −. . ..
−.. −. . −. . ..
−. . −. . −. . ..
−.. −. . −. . ..
−. . −. . −. . ..
−.. −. . −. . ..
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ..
−. .
−. . −. . ......
−. .
−. . ..
−. .
−. . ..
−. .
−. . .. ................
−. .
−. .
........
−. .
−. .
..
−. .
−. . ..
−. . ..
..
..
..
..
..
..
..
..
 Table 21.2
Cumulative normal distribution
P A R T V I Options, Futures, and Other Derivatives
Some of the important assumptions underlying the formula are the following:
1. The stock will pay no dividends until after the option expiration date.
2. Both the interest rate, r, and variance rate, σ2, of the stock are constant (or in
slightly more general versions of the formula, both are known functions of time—
any changes are perfectly predictable).
3. Stock prices are continuous, meaning that sudden extreme jumps such as those in the
aftermath of an announcement of a takeover attempt are ruled out.
Variants of the Black-Scholes formula have been developed to deal with many of these limitations.
Second, even within the context of the Black-Scholes model, you must be sure of the accuracy of
the parameters used in the formula. Four of these—S0, X, T, and r—are straightforward. The stock
price, exercise price, and time to expiration are readily deter mined. The interest rate used is the
money market rate for a maturity equal to that of the option, and the dividend payout is reasonably
predictable, at least over short horizons.
The last input, though, the standard deviation of the stock return, is not directly observable. It must
be estimated from historical data, from scenario analysis, or from the prices of other options, as we
will describe momentarily.
We saw in Chapter 5 that the historical variance of stock market returns can be calcu
lated from n observations as follows:
n 2
n − 1 ∑t = 1 (rt − r )
σ2 = n _____

n
where r is the average return over the sample period. The rate of return on day t is defined to be
consistent with continuous compounding as rt = ln(St /St − 1). [We note again that the natural
logarithm of a ratio is approximately the percentage difference between the numer ator and
denominator so that ln(St /St − 1) is a measure of the rate of return of the stock from time t − 1 to
time t.] Historical variance commonly is computed using daily returns over periods of several
months. Because the volatility of stock returns must be estimated, however, it is always possible
that discrepancies between an option price and its Black Scholes value are simply artifacts of error
in the estimation of the stock’s volatility.
In fact, market participants often give the option-valuation problem a different twist. Rather than
calculating a Black-Scholes option value for a given stock’s standard devia tion, they ask instead:
What standard deviation would be necessary for the option price that I observe to be consistent
with the Black-Scholes formula? This is called the implied volatility of the option, the volatility level
for the stock implied by the option price. Inves tors can then judge whether they think the actual
stock standard deviation exceeds the implied volatility. If it does, the option is considered a good
buy; if actual volatility seems greater than the implied volatility, its fair price would exceed the
observed price.
Another variation is to compare two options on the same stock with equal expiration dates but
different exercise prices. The option with the higher implied volatility would be considered relatively
expensive, because a higher standard deviation is required to justify its price. The analyst might
consider buying the option with the lower implied volatility and writing the option with the higher
implied volatility.
The Black-Scholes valuation formula, as well as the implied volatility, is easily calcu
lated using an Excel spreadsheet like Spreadsheet 21.1. The model inputs are provided in column
B, and the outputs are given in column E. The formulas for d1 and d2 are provided in the
spreadsheet, and the Excel formula NORMSDIST(d1) or NORM.S.DIST(d1, TRUE) is used to
calculate N(d1). Cell E6 contains the Black-Scholes formula. (The formula in the spreadsheet
actually includes an adjustment for dividends, as described in the next section.)
C H A P T E R Option Valuation A C D E
FGHIJ
Stock price 100 N(d2) 0.4231 N

Exercise price 105 B/S call value 7.000 B5

B 0

1
2
3
4

eXc el 5
6
INPUTS OUTPUTS

Standard deviation 0.2783 d1 0.002

(annual) 9

Maturity (in years) 0.5 d2 –0.193

9

0.06 N(d1) 0.501

2
7 Dividend yield (annual)
0
B/S put value
8.8968
Ris B6*EXP(-B4*B3)*(1-E5)-B5*EXP(-B7*B3)*(1-E4)

Spreadsheet 21.1
Please visit us at
Spreadsheet to calculate
Black-Scholes call option values

www.mhhe.com/Bodiee

To compute an implied volatility, we can use the Goal Seek command from the What
BCDEFGHIJ
If Analysis menu (which can be found under the Data tab) in Excel. See Figure 21.7 for an
illustration. Goal Seek asks us to change the value of one cell to make the value of
another cell (called the target cell) equal to a specific value. For example, if we observe a
call option selling for $7 with other inputs as given in the spreadsheet, we can use Goal
Seek to change the value in cell B2 (the standard deviation of the stock) to set the option
value in cell E6 equal to $7. The target cell, E6, is the call price, and the spreadsheet
manipu
lates cell B2. When you click OK, the spreadsheet finds that a standard deviation equal to
.2783 is consistent with a call price of $7; this would be the option’s implied volatility if it
were selling at $7.

A
 INPUTS OUTPUTS FORMULA FOR OUTPUT IN COLUMN E
1

.783 d1 .9 (LN(B5/B ^)*B3)/(B

6)+ (B4-B7+.5
*B

.5 d2 -.939 E-B*SQRT(B3)

.6 N(d1) .5 NORMSDIST(E)

N(d2) .43 NORMSDIST(E3)

5 B/S call value 7. B5*EXP( B7*B3)*E -B4*B3)*E5

- 4-B6*EXP
(

B/S put value 8.8968 B6*EXP( B4*B3)*(-E5) B5*EXP(

- -

c el

Standard deviation (annual)

2
Maturity (in years)
3
Risk-free rate (annual)
4
Figure 21.7 Using Goal Seek to find implied volatility
Stock price
5
Exercise price
6
*SQRT(B3))
Dividend yield (annual)
7
8
9
10
11
B7*B3)*(-E4)
12
13
14
15
16
17
 Please visit us at
www.mhhe.com/Bodiee

eX
P A R T V I Options, Futures, and Other Derivatives
The Chicago Board Options Exchange
7 Financial Crisis
Implied Volatility (%) 6

regularly computes the implied volatil ity of major

stock indexes. Figure 21.8 is
U.S. Debt
a graph of
the implied
(30-day)
volatil
5
LTCM
Downgrade
Iraq 9/
4 Gulf War War), August 1998 (collapse of
3 ity of the S&P 500 since 1990. Long-Term Capital Management),
During periods of turmoil, implied September 11, 2001 (terrorist
volatility can spike quickly. Notice attacks in
the peaks in January 1991 (Gulf
99 5 8 4 7 993

996

999
Iraq), and, most dramatically, during 2008
(the financial crisis). Because implied
the U.S.), 2002 (build-up to invasion of
Chicago Board Options Exchange, www.cboe.com.

volatility correlates with crisis, it is some times

called an “investor fear gauge.” A futures
contract on the 30-day implied volatility of the
S&P 500 has traded on the CBOE Futures
Exchange

Figure 21.8 Implied volatility of the S&P (VIX index) Source:

since 2004. The payoff of the contract depends on market implied volatility at the expira
tion of the contract. The ticker symbol of the contract is VIX.
Figure 21.8 reveals an awkward empirical fact. While the Black-Scholes formula is
derived assuming that stock volatility is constant, the time series of implied volatilities
derived from that formula is in fact far from constant. This contradiction reminds us that
the Black-Scholes model (like all models) is a simplification that does not capture all
aspects of real markets. In this particular context, extensions of the pricing model that
allow stock volatility to evolve randomly over time would be desirable, and, in fact, many
extensions of the model along these lines have been developed.10
The fact that volatility changes unpredictably means that it can be difficult to choose
 the proper volatility input to use in any option-pricing model. A considerable amount of
recent research has been devoted to techniques to predict changes in volatility. These
tech
niques, known as ARCH and stochastic volatility models, posit that changes in volatility
are partially predictable and that by analyzing recent levels and trends in volatility, one
can improve predictions of future volatility.11

Concept Check .
Suppose the call
option in
Spreadsheet .
actually is selling for $. Is its implied volatility more or less than .%? Use the spreadsheet (available in Connect) and
Goal Seek to find its implied volatility at this price.

10
Influential articles on this topic are J. Hull and A. White, “The Pricing of Options on Assets with Stochastic
Volatilities,” Journal of Finance (June 1987), pp. 281–300; J. Wiggins, “Option Values under Stochastic Vola
tility,” Journal of Financial Economics (December 1987), pp. 351–72; and S. Heston, “A Closed-Form Solution
for Options with Stochastic Volatility with Applications to Bonds and Currency Options,” Review of Financial
Studies 6 (1993), pp. 327–43. For a review, see E. Ghysels, A. Harvey, and E. Renault, “Stochastic Volatility,” in
Handbook of Statistics, Vol. 14: Statistical Methods in Finance, ed. G. S. Maddala (Amsterdam: North Holland,
1996).
11
For an introduction to these models see Carol Alexander, Market Risk Analysis, Vol. 4 (England: Wiley, 2009).
C H A P T E R Option Valuation
Dividends and Call Option Valuation
We noted earlier that the Black-Scholes call option formula applies to stocks that do not pay
dividends. When dividends are to be paid before the option expires, we need to adjust the formula.
The payment of dividends raises the possibility of early exercise, and for most realistic dividend
payout schemes the valuation formula becomes significantly more complex than the
Black-Scholes equation.
When stocks pay quarterly dividends, their share prices decline by a discrete amount, roughly
equal to the amount of the dividend. In some cases, it will be rational for the call holder to exercise
just before the stock goes ex dividend. This introduces uncertainty into “maturity” of the call. Will it
be exercised at the ex-dividend date or held until the expira
tion date?12 Variations on the Black-Scholes formula have been developed that can accom modate
dividends, but the resulting valuation formulas are more complex and become rapidly more difficult
as the number of possible dividend payments increase.13
In one special case, the dividend adjustment takes a simple form and allows us to use a slight
variant of the Black-Scholes formula. Suppose the underlying asset pays a con tinuous flow of
income. This might be a reasonable assumption for options on a stock index, where different
stocks in the index pay dividends on different days, so that dividend income arrives in a more or
less continuous flow. If the dividend yield, denoted δ, is con stant, one can show that the present
value of that dividend flow accruing until the option expiration date is S0 (1 − e−δT).14 In this case,
S0 − PV(Div) = S0e−δT, and we can use the Black-Scholes call option formula on the
dividend-paying asset simply by substituting S0e−δT for S0 in the original formula. This approach is
used in Spreadsheet 21.1.
One warning about this practice, however. Even with continuous dividends, it may be rational to
exercise the call option early, so strictly speaking, the modified Black-Scholes formula would apply
only to European options. As a general rule, American calls on divi dend paying stocks will be
worth more than European ones even if dividends are continuous.

Put Option Valuation

We have concentrated so far on call option valuation. We can derive Black-Scholes European put
option values from call option values using the put-call parity theorem. To value the put option, we
simply calculate the value of the corresponding call option in Equation 21.1 from the Black-Scholes
formula, and solve for the put option value (on a non-dividend paying stock) as
P = C + PV(X ) − S0
= C + Xe−rT − S0(21.2)
We calculate the present value of the exercise price using continuous compounding to be
consistent with the Black-Scholes formula.

12
While the stock price falls by a discrete amount on the ex-dividend date, the option price does not. The dividend is
announced in advance and is anticipated by the market. The option price will adjust smoothly over time to reflect the
approaching dividend payment.
13
An exact formula for American call valuation on dividend-paying stocks has been developed in Richard Roll, “An Analytic
Valuation Formula for Unprotected American Call Options on Stocks with Known Dividends,” Journal of Financial
Economics 5 (November 1977). The technique has been discussed and revised in Robert Geske, “A Note on an Analytical
Formula for Unprotected American Call Options on Stocks with Known Dividends,” Journal of Financial Economics 7
(December 1979); and Robert E. Whaley, “On the Valuation of American Call Options on Stocks with Known Dividends,”
Journal of Financial Economics 9 (June 1981).
14
For intuition about this formula, notice that e−δT approximately equals 1 − δT, so the value of the dividend is approximately
δTS0.
P A R T V I Options, Futures, and Other Derivatives
Sometimes, it is easier to work with a put option valuation formula directly. If we sub
stitute the Black-Scholes formula for a call in Equation 21.2, we obtain the value of a European put
option as

P = Xe−rT 1 − N(d2) − S01 − N(d1) (21.3)

Example 21.5 Black-Scholes Put Valuation

Using data from Example . (C = $., X = $, S

= $, r = ., σ = ., and

T = .), Equation . implies that a European put option on that stock with identical
exercise price and time to expiration is worth

$e−.× .( − .) − $( − .) = $.

Notice that this value is consistent with put-call parity:

P = C + PV(X) − S = . + e−.× . − = .

As we noted traders can do, we might then compare this formula value to the actual put price
as one step in formulating a trading strategy.

Dividends and Put Option Valuation

Equation 21.2 and Equation 21.3 are valid for European puts on non-dividend-paying stocks. As
we did for call options, if the underlying asset pays a dividend, we can find European put values by
substituting S0 − PV(Div) for S0. Cell E7 in Spreadsheet 21.1 allows for a continuous dividend flow
with a dividend yield of δ. In that case S0 −
PV(Div) = S0e−δT.
However, listed put options on stocks are American options that offer the opportunity of early
exercise, and we have seen that the right to exercise puts early can turn out to be valuable. This
means that an American put option must be worth more than the cor
responding European option. Therefore, Equation 21.2 and Equation 21.3 describe only the lower
bound on the true value of the American put. However, in many applications the approximation is
very accurate.

21.5 Using the Black-Scholes Formula

Hedge Ratios and the Black-Scholes Formula
In Chapter 20, we considered two investments in FinCorp stock: 100 shares or 1,000 call options.
We saw that the call option position was more sensitive to swings in the stock price than was the
all-stock position. To analyze the overall exposure to a stock price more precisely, however, it is
necessary to quantify these relative sensitivities. We can summa
rize the overall exposure of portfolios of options with various exercise prices and times to
expiration using the hedge ratio, the change in option price for a $1 increase in the stock price. A
call option, therefore, has a positive hedge ratio and a put option a negative hedge ratio. The
hedge ratio is commonly called the option’s delta.