CAPM

9
The Capital Asset Pricing Model

Knowing how risk (market beta) and reward (expected rate of return)
are measured, you are now ready to proceed to the punchline: a formula
that tells you how much reward your investment projects have to offer to
compensate your investors for their risks. If you can judge the risk of new
corporate investment projects, you can then determine the appropriate costs
of capital that you should use in your projects NPV calculations. Alas, like
NPV, the formula may be simple, but the application is hard. The devil is in
the details.
We will first briefly review what you already know. Then you will learn all
about this new modelthe CAPM. Finally, you will get to apply it.
One apology in advance: In this chapter, I do not fully explain where all the
formulas come from. This is because it really takes a full investments course
to derive them. (The appendix goes into more detail, but if you really want
to learn about investments, you need to take a full course on the subject.)
9.1
What You Already Know and What You Want to Know
Lets take stock. First, you already know the right train of thought for capital budgeting
purposes: As a corporate manager, your task is to determine whether you should accept
or reject a project. You make this decision with the NPV formula. To determine the
discount factor in the NPV formula, you need to estimate an appropriate cost of capital
or, more precisely, the opportunity cost of capital for your investors. This means that

you need to judge what a fair expected rate of return, E r , for your project is, given
your projects risk characteristics. If your project offers a lower expected return than
what your investors can earn elsewhere in similarly risky projects, then you should not
put your investors money into your project but instead return their money to them. If
your project offers more expected return, then you should go ahead and invest their
money into your project. Put differently, your goal is to learn what your investors, if
asked, would have wanted you to invest in on their behalves.
Second, the perfect market assumptions are not enough to proceed. We must assume
that investors like overall portfolio reward (expected return) and dislike overall portfolio
risk (variance or standard deviation of return). We also assume that investors are
smart. Presumably, this means that they diversify, hopefully holding many assets and
be reasonably close to the market portfolio. Somewhat less appealing, we also must
219
You are still after an

estimate for your
opportunity cost of
capital.
Assume perfect
markets, that
investors dislike risk
and like reward, and
more.
220
This allows you to

figure out how
theyand
youshould measure
project risk and
reward.
This gives you a

trade-off between risk
and reward in
equilibrium.
assume that investors all have access to exactly the same set of assets. (This means
we are ignoring investments in peoples own houses or education, for example.) And
finally, mostly for convenience, we assume that they want to maximize their wealth in
the market for only one period.
Third, for investors with these preferences and who are therefore already holding
the overall market portfolio, you can follow their trains of thought. You can infer how
they should view the risk and reward of your individual projects. Their reward is their
expected rate of return. Their risk is their overall portfolio risk, not your projects own
standard-deviation risk. Your projects contribution to your investors overall portfolio
risk is the market beta of your projectthink of it as a measure of your projects toxicity.
A project that decreases in value when the market decreases in value, and increases
when the market increases, has a positive market beta. Its toxicinvestors dont like it.
A project that increases in value when the market decreases in value, and vice versa, has
a negative market beta. Its less toxicinvestors like it more. That is, a project with a
low market beta helps an investor who holds a portfolio similar to the market portfolio
to reduce the overall investment risk.
You can also draw some additional conclusions without any math. In our assumed
perfect world, you can guess that investors will have already snatched up the best
projectsthose that have low risk and high expected rates of return. In fact, anyone
selling projects with lower risk contributions can sell them for higher prices, which in
turn immediately drives down their expected rates of return. Consequently, what is
available for purchase in the real world must be subject to some trade-off: Projects
that have more market-risk contribution must offer a higher expected rate of return if
their sellers want to convince investors to purchase them. But what exactly does this
relationship between risk and reward look like? This is the subject of this chapterit is
the domain of the capital asset pricing model, the CAPM.
Q 9.1. What are the assumptions underlying the CAPM? Are the perfect market assumptions among them? Are there more?
9.2
The CAPM gives you
the cost of capital if
you give it the
risk-free rate, the
expected rate of
return on the market,
and your projects
market beta.
Using The Capital Asset Pricing Model (CAPM)
The capital asset pricing model (CAPM) is a model that gives you an appropriate
expected rate of return (cost of capital) for each project if you give it the projects
relevant risk characteristics. The model states that an investments cost of capital is
lower when it offers better diversification benefits for an investor who holds the overall
market portfolioless required reward for less risk contribution. Market beta is its
measure of risk contribution. Projects contributing more risk (market beta) require
a higher expected rate of return for you to want them; projects contributing less risk
require a lower expected rate of return for you to want them. This is the precise
relationship that the CAPM gives you.
9.2. Using The Capital Asset Pricing Model (CAPM)
221
To estimate the required expected rate of return for a project or firmthat is, the cost
of capitalaccording to the CAPM, you need three inputs:
IMPORTANT
1. The risk-free rate of return, rF

2. The expected rate of return on the overall market, E rM
3. A firms or projects beta with respect to the market, i

The CAPM formula is

= rF + E rM rF i
(9.1)

where i is the name of your project and E ri is your projects expected rate of return.
E ri
The difference between the expected rate of return on the risky (stock) market and

the risk-free investment, E rM rF , is called the equity premium or market risk
premium, discussed in more detail later.
You need to memorize the CAPM formula. It is the standard model in the finance.
Lets use the formula. If you believe that the risk-free rate is 3% and the expected
rate of return on the market is 7%, then the CAPM states that
E ri
E ri
A first quick use of

the CAPM formula.
= 3% +
(7% 3%) i
= 3% + 4% i

= rF + E rM rF i
Therefore, a project with a beta of 0.5 should have a cost of capital of 3%+4%0.5 = 5%,
and a project with a beta of 2.0 should have a cost of capital of 3% + 4% 2.0 = 11%.
The CAPM gives an opportunity cost for your investors capital: If the project with the
beta of 2.0 cannot earn an expected rate of return of 11%, you should not take this
project and instead return the money to your investors. Your project would add too
much risk for its reward. Your investors have better opportunities elsewhere.
The CAPM is called an asset-pricing model, even though it is most often expressed
in terms of a required expected rate of return rather than in terms of an appropriate
project price. Fortunately, though messy, the two are equivalentyou can always work
with the CAPM return first, and discount the expected cash flow into an appropriate
price second. A given expected rate of return implies a given price. (If you do not know
the fair price, you will however have to take two aspirins and work with a more difficult
version of the CAPM formula. It is called certainty equivalence and explained in the
chapter appendix.)
The CAPM specifically ignores the standard deviation of individual projects rates
of return. That is, the model posits that investors do not care about it, because they
are smart enough to diversify away any idiosyncratic risk. The CAPM posits that
investors instead care about the project market betas, because these measure the risk
components that investors holding the market portfolio cannot diversify away. (This
It is easier to work in
required returns than
in prices.
Certainty equivalence
CAPM form,
Sect. App.9.A (Companion),
Pg.51.
The CAPM formula

tells you what
investors care about:
comovement with the
market.
222
The CAPM has three

inputs. We will cover
them in detail.
Will history repeat

itself?, Sect. 7.1, Pg.175.
makes a lot of sense for highly-diversified investors, though not for liquidity-constrained
entrepreneurs.)
For the three CAPM inputs, as always, you are really interested in the future: the
future expected rate of return on the market and the future beta of your firm/project
with respect to the market. You really dont care about the past average rates of return
or the past market betas. But, as usual, you often have no choice other than to rely on
estimates that are based at least partly on historical data. In Section 9.4, you will learn
how to estimate each CAPM input. But lets explore the model itself first, assuming that
you know all the inputs.
The Security Market Line (SML)

Examples of CAPM
rates of return that
individual securities
should offer.
The SML is just a

graphical
representation of the
CAPM formula.
If you know the

inputs, the SML is a
sharp line; if you
estimate them, it is a
scatterplot.
Lets apply the CAPM in a specific example. Assume that the risk-free rate is 3% per year
and that the market offers an expected rate of return of 8% per year. The CAPM formula
then states that a stock with a beta of 1 should offer an expected rate of return of
3% + (8% 3%) 1 = 8% per year; that a stock with a beta of 0 should offer an expected
rate of return of 3% + (8% 3%) 0 = 3% per year; that a stock with a beta of 1/2 should
offer an expected rate of return of 3% + (8% 3%) 0.5 = 5.5% per year; that a stock
with a beta of 2 should offer an expected rate of return of 3% + (8% 3%) 2 = 13% per
year; and so on.
The CAPM formula is often graphed as the security market line (SML), which
shows the relationship between the expected rate of return of a project and its beta.
Exhibit 9.1 draws a first security market line for seven assets. Each investment asset
(such as a stock or a project) is a point in this coordinate system. Because all assets
properly follow the CAPM formula in our example, they must lie on a straight line. In
other words, the SML is just a graphical representation of the CAPM formula. The slope

of this line is the equity premium, E rM rF , and the intercept is the risk-free rate, rF .
Alas, in the real world, even if the CAPM holds, you would not have the data to
draw Exhibit 9.1. The reason is that you do not know true expected returns and true
market betas. Exhibit 9.2 plots two graphs in a perfect CAPM world. The top graph
repeats Exhibit 9.1 and falsely presumes that you know CAPM inputsthe true market
betas and true expected rates of return. This line is perfectly straight. In the bottom
graph, you have to rely only on observablesestimates of expected returns and betas,
presumably based mostly on historical data averages. Now you can only fit an estimated
security market line, not the true security market line. Of course, you hope that your
historical data provides good, unbiased estimates of true market beta and true expected
rates of return (and this is a big assumption), so that your fitted line will look at least
approximately straight. A workable version of the CAPM thus can only state that there
should roughly be a linear relationship between the data-estimated market betas and
the data-estimated expected rates of return, just as drawn here.
Q 9.2. The risk-free rate is 4%. The expected rate of return on the market is 7%. What
is the appropriate cost of capital for a project that has a beta of 3?
is the cost of capital for a project that has a beta of 3?
9.2. Using The Capital Asset Pricing Model (CAPM)
223
15
Exp RoR E(ri), in %
10
The intercept is the

3% riskfree interest rate.
s the
ope i emium.
l
s
e
pr
Th
quity
5% e
5
1.0
0.5
0.0
0.5
1.0
1.5
2.0
Project beta with respect to M (i)
Investment Asset
Market Beta
Expected Rate of Return
i
E ri
1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.0%
0.5%
3.0%
5.5%
8.0%
10.5%
13.0%

Exhibit 9.1: The Security Market Line. This graph plots the CAPM relation E ri = rF + [E rM rF ] i =
3% + (8% 3%) i , where i is the beta of an individual asset with respect to the market. In this graph, we
assume that the risk-free rate is 3% and the equity premium is 5%. Each point is one asset (such as a stock, a
project, or a mutual fund). The point M in this graph could also be any other security with a i = 1. F could be
the risk-free asset or any other security with a i = 0.
is the cost of capital for a project that has a beta of 3? Does this make economic sense?
Q 9.5. Is the real-world SML with historical data a perfect straight line?
Q 9.6. The risk-free rate is 4%. The expected rate of return on the market is 7%. A
corporation intends to issue publicly-traded bonds that promise a rate of return of 6%
and offer an expected rate of return of 5%. What is the implicit beta of the bonds?
Q 9.7. Draw the SML if the risk-free rate is 5% and the equity premium is 9%.
Q 9.8. What is the equity premium, both mathematically and intuitively?
224
15
Known E(ri), in %
10
M
Market M
5
F
Riskfree
Treasury
5
1.0
0.5
0.0
0.5
1.0
1.5
2.0
Known MarketBeta (i)
15
Average RoR, in %
10
5
1.0
0.5
0.0
0.5
1.0
1.5
2.0
Historical MarketBeta (i)
Exhibit 9.2: The Security Market Line in an Ideal CAPM World. The lower panel shows what we are usually
confronted with: Historical average returns and historical betas are just estimates from the data. We hope that
they are representative of the true underlying mean returns and true betas, which in turn would mean that they
will also be representative of the future means and betas.
9.3. The CAPM Cost of Capital in the Present Value Formula
9.3
225
The CAPM Cost of Capital in the Present Value Formula
For a corporate manager, the CAPM is needed to get the denominator in the NPV formula,

the opportunity cost of capital, E r :

E C1
E C2
+
+
NPV = C0 +
1 + E r1
1 + E r2
We usually use the

CAPM output, the
expected rate of
return, as our
discount rate.
Together, the CAPM and the NPV formulas tell you again that cash flows that correlate
more with the overall market are of less value to your investors and therefore require

higher expected rates of return (E r ) in order to pass muster (well, the hurdle rate,
which is determined by the alternative opportunities that your model presumes your
investors have).
Deconstructing Quoted Rates of Return Risk Premiums

Let me return to the subject of Section 6.2. You learned that in a perfect and risk-neutral
world, stated rates of return consist of a time premium and a default premium. On
average, the default premium is zero, so the expected rate of return is just the time
premium.
The CAPM extends the expected rate of return to a world in which investors are risk
averse. It gives you an expected rate of return that adds a risk premium (as a reward
for your willingness to absorb risk) to the time premium.
Promised Rate of Return = Time Premium + Default Premium
Actual Earned Rate
+ Risk Premium
Reminder: Stated
bond yields contain
time and default
premiums.
Time and default
premiums, Sect. 6.2,
Pg.129.
The CAPM gives you

the time and risk
premiums.
= Time Premium + Default Realization + Risk Premium
Expected Rate of Return = Time Premium + Expected Risk Premium

|
{z
}
provided by the CAPM
In the risk-neutral perfect world, there were no differences in expected rates of return
across assets. There were only differences in stated rates of return. The CAPM changes
all thisdifferent assets can now also have different expected rates of return.
However, the CAPM does not take default risk into account, much less give you an
appropriate stated rate of return. You should therefore wonder: How do you find the
appropriate quoted rate of return in the real world? After all, it is this stated rate of
return that is usually publicly posted, not the expected rate of return. Put differently,
how do you put the default risk and CAPM risk into one valuation?
Here is an example. Say you want to determine the PV of a corporate zero-bond that
has a beta of 0.25 and promises to deliver $200 next year. This bond pays off 95% of
the time, and 5% of the time it totally defaults. Assume that the risk-free rate of return
is 6% per annum and that the expected rate of return on the market is 10%. Therefore,
the CAPM states that the expected rate of return on your bond must be

E rBond
= 6% +
4%
0.25 = 7%

= rF + [E rM rF ] Bond
This takes care of the time and risk premiums. To take the bonds default risk into
Important: The
CAPM ignores default
risk and, thus, does
not provide a default
premium. You must
take care of it
yourself!
A specific bond
example: First
compute the price
necessary to make you
even relative to the
Treasury if you are
risk-neutral. This
price is based on the
time premium and the
default premium.
226
account, you must still find the numerator. You cannot use the promised payment. You
must adjust it for the probability of default. You expect to receive not $200, but

E CBond
=
95%
$200 +
5%
0
= $190
= Prob(No Default) Promise + Prob(Default) Nothing
Therefore, the present value formula states that the value of the bond is

E CBond
$190
=
$177.57
PVBond =
1 + 7%
1 + E rBond
Given this price, you can now compute the promised (or quoted) rate of return on this
bond:
$200 $177.57
12.6%
$177.57
Promised Cash Flow PV
PV
The risk premium is
above and beyond the
time and default
premiums. On
average, risky
investments earn more
than risk-free
investments now.
= Promised Rate of Return
You can now quantify the three components in this example. For this bond, the time
premium of money is 6% per annumit is the rate of return that an equivalent-term
Treasury offers. The time premium plus the risk premium is provided by the CAPM,
and it is 7% per annum. Therefore, 1% per annum is your average compensation for
your willingness to hold this risky bond instead of the risk-free Treasury. The remaining
12.6% 7% = 5.6% per annum is the default premium: You do not expect to earn money
from this default premium on average. You only earn it if the bond does not default.
12.6%
6%
5.6%
1%
Promised Interest Rate = Time Premium + Default Premium + Risk Premium

In the real world, most bonds have fairly small market betas (often much smaller
than 0.25) and thus fairly low risk premiums. Instead, most of the premium that
ordinary corporate bonds quote above equivalent risk-free Treasury rates is not due to
the risk premium, but due to the default premium. They simply wont pay as much
as they promise, on average. However, for corporate projects and equity shares, the
risk premium can be quite large. (Watch outthere are also some important imperfect
market premiums that you will only learn in the next chapter.)
IMPORTANT
Never forget:
The CAPM provides an expected rate of return.
This return is not a stated (promised, quoted) rate of return, because it does not
include a default premium.
The probability of default must be handled in the NPV numerator (through the
expected cash flow), and not in the NPV denominator (through the expected rate
of return).
9.4. Estimating the CAPM Inputs
227
Q 9.9. A corporate bond with a beta of 0.2 will pay off next year with 99% probability.
The risk-free rate is 3% per annum, and the equity premium is 5% per annum.
1. What is the price of this bond?
2. What is its promised rate of return?
3. Decompose the bonds quoted rate of return into its components.
Q 9.10. Going to your school has total additional and opportunity costs of $30,000 this
year and up-front. With 90% probability, you are likely to graduate from your school.
If you do not graduate, you have lost the entire sum. Graduating from the school will
increase your 40-year lifetime annual salary by roughly $5,000 per year, but more so
when the market rate of return is high than when it is low. For arguments sake, assume
that your extra-income beta is 1.5. Assume the risk-free rate is 3%, and the equity
premium is 5%. What is the value of your education?
9.4
Estimating the CAPM Inputs
How can you obtain reasonable estimates of the three inputs into the CAPM formula

E r i = r F + E r M rF i ?
The Risk-Free Rate and Multi-Year Term-Structure Considerations

The first input into the CAPM formula is the risk-free rate of return (rF ). First, dont
forget to use nominal rates to discount nominal expected cash flows. Now, this nominal
risk-free rate is relatively easy to obtain from U.S. Treasuries. There is one small issue,
thoughwhich Treasury? What if the yield curve is upward sloping (as it usually does)
and Treasuries yield 1% per year over one year, 3% per year over ten years, and 5% per
year over thirty years? Which risk-free interest rate should go into the CAPM?
Unfortunately, the CAPM offers no guidance, because it has no concept of more than
one single time period and thus no concept of a yield curve. However, from a practical
perspective, it makes sense to match projects to similar risk-free bond benchmarks. That
is, pick the risk-free zero-bond yield that is closest to each of your projects specific
expected cash flows at the same time. For example, to value a machine that operates for
three years, use the 1-year T-bond yield to discount the expected cash flow in the first
years NPV term, the 2-year T-bond yield for the second years NPV term, and the 3-year
T-bond yield for the third years NPV term. If you had to use just one risk-free rate for
multiple cash flows (because your pointy-haired boss says so), choose an average of the
three rates or simply the 2-year bond. (There are better ways to do this, but the extra
precision is rarely worth it.)
Which risk-free rate?
US Treasuries,
Sect. 5.3, Pg.97.
Advice: Pick the

interest rate for a
Treasury that is most
similar to your
project.
Yield Curve,
Chapter 5, Pg.85.
228
But dont we need

formal guidance? Isnt
this violating the
letter of the law?
You may think this is a pretty loose method to handle an important question, and you
would be right. However, it is also a reasonable method. Think about the opportunity
cost of capital for a small investment with a market-beta of 0. If your corporations
investors are willing to commit their money for ten years, they could earn the yield on a
ten-year risk-free Treasury bond instead. It is this ten-year rate that would then be the
opportunity cost of capital on your own project cash flow that will materialize in ten
years. If your projects cash flow will occur in three months, your investors could only
earn the rate of return on a three-month T-bill instead. Indeed, there is almost universal
agreement that companies should use a risk-free rate lined up with the project cash flow
timing in the first part of the CAPM formula (where rF appears by itself).
Q 9.11. What is todays risk-free rate for a 1-year project? For a 10-year project?
Q 9.12. If you can use only one Treasury, which risk-free rate should you use for a
project that will yield $5 million each year for 10 years?
The Equity Premium

You want to know the
equity premium,
regardless of the
CAPM
You must provide the

CAPM with the equity
premium. Good luck!
Do not use a
short-term-Treasury
based equity premium
for benchmarking your
far-into-the-future
cash flows.

Your second CAPM input, the equity premium (E rM rF ), is much more difficult to
estimate. It is the extra expected rate of return that risky equity projects have to offer
above and beyond what risk-free bond projects are offering. (It is a difference, so you
can use either two nominal or two real rates.) By the way, regardless of whether the
CAPM holds or not, this is a number of first-order importance to youit helps you decide
whether you should invest your own money in risky equities or in safer bonds.
The theoretical CAPM model assumes that you already know the expected rate of
return on the market perfectly, not that you have to estimate it. But in real life, the
equity premium is not posted anywhere, and no one really knows the correct number.
Worse: Not only is it difficult to estimate, but your estimate often has a large influence
over the CAPMs estimated cost of capital. Cest la vis.
Many other finance text books quote just one equity-premium estimate, and it is
often the expected rate of return on stocks relative to the short-term Treasury yield. This
choice can be reasonable if your own cash flows (that you want to discount) are also
very short-horizon. Stock market investors, who can buy one day and sell the next, can
defend this practice. It also means that an investment in a project with a beta of 1 has
an expected rate of return equal to that in the stock market, because the risk-free rates
in the intercept and slope cancel. Unfortunately, corporate-finance executives can rarely
move in and out of projects on a moments notice. They usually need to use the CAPM
to decide on investments that have cash flows expected to materialize only many years
into the future. In this case, everyone agrees that your CAPM equity premium should
not be expected stock returns above short-term Treasuries. Instead, you should use the
same equivalent-term-to-your-project-cash-flows Treasury rate in your estimate of the
equity premium that you used as your risk-free Treasury in the constant term in the
CAPM formula. (In fact, there is even a second argument to use long-term risk-free rates
in the equity premium: equities are long-term investments, so you should always net
229
out the long-term Treasury rate from expected stock returns, regardless of your own
cash flows horizons.)
There are a number of methods to guesstimate the equity premium. Unfortunately,
for many decades, these methods have not tended to agree with one another. It should
thus not come as a surprise that practitioners, instructors, finance textbook authors
have also been confused and confusing. Exhibit 9.3 shows that each text book seems
to have had its own estimate. (Fortunately, both the disagreement and the average
recommended estimate seem to be slowly declining.)
Should I just give it to

you?
12
10
Equity Premium
0
1980
1985
1990
1995
2000
2005
2010
Year
Exhibit 9.3: Equity Premia from Different Textbooks. Source: Pablo Fernandez, SSRN, 2013..
Ultimately, we finance-textbook authors have two choices: The first is to throw you
one estimate, pretend it is the correct one, and hope you forget to ask hard questions.
If you like a formulaic painting-by-numbers approach, this would leave you (wrongly)
satisfied. The second is to tell you about the different methods that lead to different
estimates. This is the route I will takeexplaining different reasoning behind different
estimatesif only because the first would eventually leave you startled to discover
that your boss is using some other equity-premium and therefore has come up with
a different cost-of-capital estimate. I will both explain the intuition behind the mostcommon methods and describe the magnitude that each suggests nowadays. You can
make up your own mind what you deem to be the best estimate. (I will tell you my own
personal estimate only at the end.)
Lets show you how

people are reasoning.
230
Historical Averages I
Here are the historical
numbers.
The first and most common guesstimation method is to assume that whatever the
equity premium was in the past will also be the case in the future. Lets look at the
historical performance of stocks vs. bonds in two different time samples, 1926-2012
and 1970-2012:
Ari
Value-Weighed Stock Market
net of 1-Year Treasuries
net of 30-Year Treasuries
net of Long-Term Corporates
PS: 30-Year Treasuries had

market betas of < 0.1.
Morningstar Ibbotson
Averages, Exhibit 7.5,
Pg.168.
1926-2012
Geo
Sdv
1970-2012
Ari
Geo
Sdv
11.6%
9.7%
19.8%
11.3%
9.8%
17.3%
7.9%
5.5%
5.2%
6.1%
4.0%
3.6%
20%
22%
20%
5.8%
1.7%
1.7%
4.4%
0.8%
0.7%
17%
20%
18%
Stocks returned about 11.5% in arithmetic terms with a standard deviation of about
17-20% per year. (The value-weighted stock market is actually the correct portfolio from
a CAPM perspective, but it wouldnt be much different if you used the S&P 500 instead.)
The geometric return of about 9.5% was in line with the rule-of-thumb formula on
Page 162. Although the stock market rate of return was pretty much the same in both
samples, the equity premium was not: bond returns were higher after 1970, especially
the long-term Treasuries. Thus, the historical equity premium you would want to use
depends on the (matched) duration of your own project cash flow, not only for the

aforementioned rF , but also for the E rM rF term.
We can roughly reconcile the difference between the highest equity-premium figure
of 7.9% and the lowest figure of 0.7% in the table as follows:
Arithmetic Equity Premium 1926 to 2012 vs. Short-Term Bonds
Minus Later Sample Period 1970 to 2012
Minus Long-Term T-Bonds Instead of Short-Term T-Bills
Minus Use of Geometric Return
Minus Cross-Product of Above Three
8%
2%
2%
2%
1%
Geometric Equity Premium 1970-2012 vs. Long-Term Bonds
1%
Earlier textbooks touted the equivalent of the 7.9% figure, which thus etched itself
into the minds of generations of students, practitioners, and finance professors. (In fact,
many other finance textbooks still etch it, without a second thought!) But 7.9% is not
necessarily the right one to use. Lets go through the three differences one by one:
1. Sample Period?: You have to judge what historical sample is appropriate. You
probably want to end the sample recently (say 2012). But it is not clear whether
you should start, say, in 1926 (when most of our data series become available)
or in 1970 (about half-way). Although your estimate can seem statistically more
reliable if you use more years, using the long sample means that you are then
leaning more heavily on the (heroic) assumption that the world has not changed.
Is the world really still the same in 2013 as it was in 1926? (And is the United
States really the right country to consider alone? Maybe it just had an unusually
lucky streak during (first half of) the American Century, which is unlikely to
231
repeat. In this case, the average countrys experience may be a better forecast for
todays U.S., too.) No one knows the correct choice. I prefer the latter sample,
and more so not because (noisier) stocks have performed differently, but because
(less noisy) Treasuries have performed betterand continue to perform better.
2. Long-Term or Short-Term Bonds?: You have to judge whether short-term or longterm bonds are the appropriate benchmark. As already mentioned, the CAPM
theory itself does not understand the concept of a term structure (Chapter 5).
Thus, it does not understand yield differentials for cash flows over different
horizons. And thus, it offers you no easy guidance which one you should use. As
with our choice for the risk-free rate in the first term of the CAPM, we have no
theory guidance. We need a reasonable approach here, too.
Again, from the perspective of an investor who can make monthly decisions and
shift effortlessly between risk-free bonds and stocks, using short bonds as your
benchmark makes sense. From the perspective of a manager who needs to decide
on a short-term project, using short T-bills as your benchmark can also make sense.
However, from the perspective of a manager who needs to commit funds to a
long-term project with cash flows over decades, it does not. If all investors can
earn a higher yield in Treasuries if they commit their money for 20 years, and if
your own project requires them to commit their money for 20 years, too, then
your project should also be benchmarked to this long-term expected rate of return.
Conveniently, we already know a reasonable approximation of the term premium
that your firm has to offer for your own longer-term projects vs. your shorter-term
projects: the prevailing yield differential that similar-horizon long-term Treasuries
are offering over short-term Treasuries. And, better yet, you can use the yield
curve to (simultaneously) reduce your equity premium estimate and raise your
risk-free rate. And, more better yet, for projects with betas around 1, this means
that risk-free rates cancel and you would expect a rate of return similar to that
of the overall stock market. Just dont commit the mistake of using a (high)
long-term risk-free rate in the first CAPM term, and a (high) equity premium over
the short-term T-bill rate in the second CAPM term.
3. Geometric or Arithmetic?: Should you use geometric or arithmetic rates of return
in your benchmark cost of capital in the NPV formula? The answer is not clear, as
you can may recall from Section 7.1. There was a convention of assuming that
past returns represent equally likely future outcomes, many CAPM users compound
the annual arithmetic average stock return or equity premium. However, doing so
means that you expect the future multi-year stock performance relative to bonds
to be better than it was in the past.
You should probably compound an equity premium estimate somewhere in between the arithmetic and geometric averages. (The correct value depends on
your own cash flows duration. Besides, your own expected future cash flows are
normally geometric, too. If you think in terms of arithmetic expected cash flows
compounded over many periodsi.e., if you consider the expected cash flow on a
project that first earns +200% and then 100% [for a complete overall loss] to be
a positive, then you should use the arithmetic average. Hardly anyone thinks this
way.)
CAPM Term Structure,

Sect. 9.4, Pg.227.
Geometric vs.
Arithmetic Returns and
Extrapolation, Sect. 7.1,
Pg.161.
232
My recommendation.
Yet another problem:

your margin of error.
My own preference is to use the later 40 years, to use bonds with similar maturity
as the cash flow that are discounted, and to use an average between the arithmetic
and geometric historical average stock returns. Thus, to discount expected cash flows
that will occur in about 10 years and beyond, my own equity-premium estimate is
around 1.5%which is much lower than the 3-5% that would be touted in other books.
Conveniently, my way of estimating means that I can also use the same risk-free rate in
both the first and the second term of the CAPM. It also means that my equity premium
estimate is lower for longer-term cash flows, but my cost of capital estimate is
usually not. I still assign higher costs of capital to Longer-term cash flows, but this just
manifests itself more through the first term (the risk-free rate) than the second term
(the equity premium).
We are not done with all the problems. Small (and often seemingly innocuous)
variations in how you estimate the CAPM inputs can lead to very different cost-of-capital
estimatesthink 3% vs 5%. Even if the CAPM were correct under one definition, neither
you nor I nor anyone else know exactly which one it is. And besides the problem of
assessing the expected equity premium point estimate, there is also the problem of the
fairly large margin of error. The standard deviation
of annual returns of 20%ptranslates
p
into a standard error of error of about 0.2/ 86 2% over 86 years and 0.2/ 43 3%
over 43 years. If you are willing to assume that nothing has changed over the sample,
then you can use some additional statistical artillery: You are then about 95% sure (a
confidence range popular in statistics) that the mean geometric stock return over long
bonds was between 0% and 8% from 1926 to 2012. From 1970 to 2012, you are about
95% sure that the same number was between 2% and +7%. Frankly, this large a range
doesnt tell you much. We already knew, or at least believed, that the equity premium
should not have been negative.
Peso Problems
To make matters even more complex, some economists believe that the historical
data are not telling the full story. There are tiny probability of desasters that just
happened not to happen. (This is sometimes called a Peso problem, based on a similar
unobserved crash situation first described in an otherwise obscure academic paper about
the Mexican Peso.) If you might have lost all your money, its no wonder that you
would have earned more in the scenario in which this big disaster did not occur. We
just happened to have lived in this world, and so we now see superior returns when we
look back. There is some empirical evidence that investors behave exactly as if they fear
such a crashbut we do not know whether such a fear is (or was) rational and we are
not sure how much of the historical or future equity premium such fear can explain. A
reasonable order of magnitude is that extra compensation for crash risk could account
for no more than a 1% equity premium per annum and perhaps for nothing (given that
stock investors lost more than a third of their investments from 2000-2002 and in 2008
alone).
A sarcastic view: It
aint great!
If your estimate of the forward-looking equity premium is based on the historical

averages I method, then you can defend a choice of 1% (for long-term cash flows). If
you are aggressive, you can defend even a choice of 8% (for short-term cash flows),
and ranges from 0% to beyond 10% if need be (or, more cynically, if you are an expert
witness paid to opine so). Are you in awe (or disgust) of the wide possible range here?
233
Historical averages II
The second method is to look at historical equity premiums in the opposite light. If
stocks have become more desirable, perhaps this is because investors have become less
risk averse, because more investors thus competed to own stocks, drove up the prices,
and thereby lowered their future expected rates of return. High historical rates of return
would then be indicative of low future expected rates of return.
An even more extreme version of this argument suggests that high past equity
returns could have been not just due to high ex-ante equity premiums, but due to
historical bubbles in the stock market. The proponents of the bubble view usually
cannot quantify the appropriate equity premium, but they do argue that it is lower
after recent market run-upsexactly the opposite of what proponents of the historical
averages I method argue.
However, you should be aware that not everyone believes that there were bubbles in
the stock-market.
Method 2: Inverse
historical averages.
Current predictive ratios

The third method is to try to predict the stock market rate of return actively with
historical dividend yields (i.e., the dividend payments received by stockholders). Higher
dividend yields should make stocks more attractive and therefore predict higher future
equity premiums. The equity premium estimation is usually done in two steps: First, you
must estimate a statistical regression that predicts next years equity premium with this
years dividend yield; then, you substitute the currently prevailing dividend yield into
your estimated regression to get a prediction. Sometimes, as in 2008, current dividend
yields were so low that the predicted equity premium was negativewhich would make
no sense. Variations of this method have used interest rates or earnings yields, typically
with similar results. In any case, the empirical evidence suggests that this method does
not yield great predictionsfor example, it predicted low equity premiums in the 1990s,
which was a period of superb stock market performance.
Method 3: Dividend
or earnings yields.
Philosophical prediction
The fourth method is to wonder how much rate of return is required to entice reasonable
investors to switch from bonds into stocks. Even with an equity premium as low as 3%,
over 25 years, an equity investor would end up with more than twice the money of a
bond investor. Naturally, in a perfect market, nothing should come for free, and the
reward for risk-taking should be just about fair. Therefore, equity premiums of 6-8% just
seem too high for the amount of risk observed in the stock market. This philosophical
method generally suggests equity premiums of about 1% to 3%.
Sidenote: A bubble is a runaway market, in which rationality has temporarily disappeared.
There is a lot of debate as to whether bubbles in the stock market ever occurred. A strong case
can be made that technology stocks experienced a bubble from around 1998 to 2000. It is
often called the dot-com bubble, the internet bubble, or simply the tech bubble. There is no
convincing explanation based on fundamentals that can explain both why the NASDAQ Index
climbed from 2,280 in March 1999 to 5,000 by March 2000, and why it then dropped back to
1,640 by April 2001.
Method 4:
Introspection and
philosophy.
234
Consensus survey
Method 5: Just ask!
What to choose? Welcome to the club! No one knows the true equity premium. So, the
fifth method is to ask the expertsor anyone else who may or may not know. Its the
blind leading the blind. The ranges of estimates have varied widely (and they are often
also conveniently tilted in the interest of those giving them):
The Social Security Administration uses an estimate of around 4%.
The consulting firm McKinsey uses a standard of around 5%.
Analysts estimates
are all over the map,
too. Estimates
between 2% and 6%
per annum seem
reasonable.
Around the turn of the millenium, the most common equity premium estimates
recommended by professors of finance were 5% for a 1-year horizon and 6% for a
30-year horizon, both with a range from 3% to 8%. The estimates were generally
similar in the U.S., Spain, Germany, and the UK.
On Monday, February 28, 2005, the Wall Street Journal reported the following
average after-inflation forecasts from then to 2050 (per annum):
Name
Organization
William Dudley
Jeremy Siegel
David Rosenberg
Ethan Harris
Robert Shiller
Robert LaVorgna
Parul Jain
John Lonski
David Malpass
Jim Glassman
Goldman Sachs
Wharton
Merrill Lynch
Lehman Brothers
Yale
Deutsche Bank
Nomura
Moodys
Bear Stearns
JP Morgan
Government
Stocks Bonds
5.0%
6.0%
4.0%
4.0%
4.6%
6.5%
4.5%
4.0%
5.5%
4.0%
2.0%
1.8%
3.0%
3.5%
2.2%
4.0%
3.5%
2.0%
3.5%
2.5%
Corp.
Bonds
Equity Premium
Rel Gov Rel Corp
2.5%
2.3%
4.0%
2.5%
2.7%
5.0%
4.0%
3.0%
4.3%
3.5%
3.0%
4.2%
1.0%
0.5%
2.4%
2.5%
1.0%
2.0%
2.0%
1.5%
2.5%
3.7%
0.0%
1.5%
1.9%
1.5%
0.5%
1.0%
1.2%
0.5%
Arithmetic Average (Difference):

Volatility-Adjusted Geometric Average 1% :
2.0%
1.0%
1.4%
0.4%
As you already know, it matters (a) whether you quote geometric or arithmetic
averages; and (b) whether you quote the equity premium with respect to a shortterm or a long-term interest rate. If you want to use the short rate, then you need
to add another 1-2% to the equity-premium estimates in this table. (Unrelated,
for the equity premium, it does not matter whether equity premium numbers are
inflation adjusted. Inflation cancels out, because the equity premium is itself a
difference in nominal rates.)
In 2005, a poll by Graham and Harvey (from Duke) and CFO Magazine reported
an average equity premium estimate of CFOs of around 3%.
In mid-2008, Merrill Lynchs survey of 300 institutional investors reported 3%.
In 2012, Fernandez reported that analysts and companies in the U.S., Spain,
Germany and the U.K. all used average estimates between 5% and 6%just like
finance professors, and with the same typical range from about 3% to 8%.
235
Of course, these estimates are themselves based on the first four methods, they do
not take your own cash flow duration into account, and they occur in echo chambers
they are what analysts, companies, consultants, students, and professors have been
reading in corporate finance textbooks (like this one) for many years now.
One aspect that does not make sense is that these estimates seem to correlate too
strongly with very recent stock market returns. For example, in late 2000, right after
a huge run-up in the stock market, surveys by Fortune or Gallup/Paine Webber had
investors expecting equity premiums as high as 15% per year. (They were acutely
disappointed: The stock market dropped by as much as 30% over the following two
years. Maybe they just got the sign wrong?!)
Internal Cost of Capital (ICC)

A hybrid method combining survey methods and analysis is the Internal Cost of Capital.
Basically, this uses analysts consensus projections about S&P 500 earnings over the
next few years, and then uses a perpetuity model to back out the cost of capital that
makes the price equal to the analysts discounted future earnings. These estimates vary
over the business cycle, which is why one usually uses an average ICC over many years.
The estimates that come out of these models are about 2.5%-3% per annum relative to
10-year bonds in arithmetic terms, and about 1.5% in geometric terms. (And, as with
historical estimates, different variants can give estimates with a much larger range, say
from 0% all the way to 7%.)
Method 6: Ask and

Use!
Conclusion
You now know that no one can tell you the authoritative number for the equity premium.
Such authority does not exist. Everyone is guessing, but there is no way around ityou
have to take a stance on the equity premium. I could not shield you from this problem.
I could only give you the arguments that you should contemplate when you are picking
your number. My own take is this: First, I have my doubts that equity premiums will be
8% in the future. (The twentieth century was the American Century for a good reason:
There were a lot of positive surprises for American investors.) I personally prefer equity
premium estimates around 2%, and this is actually in line with the majority of methods
mentioned above. But realize that reasonable expert witnesses can cherry-pick equity
premium estimates as low as 1% or as high as 8%. Of course, I personally find their
estimates less believable the farther they are from my own personal estimate. And I
find anything outside this 1% to 8% range just too tough to swallow. Second, whatever
equity premium you do choose, be consistent. Do not use 3% for investing in one project
and 8% for investing in another similarly-timed project. And do not use a risk-free rate
based on long-term bonds as your risk-free rate in the CAPM and an equity premium
estimate based on short-term bills. Being consistent can sometimes reduce your relative
mistakes in choosing one project over another.
Yes, the equity premium is difficult to estimate, but there is really no way around your
taking a stance. Even if you had never heard of the CAPM, you would still consider the
equity premium to be one of the two most important numbers in finance (together with
the risk-free rate, the other CAPM input). If you believe that the equity premium is
Remain consistent:
Dont use different
equity premium
estimates for different
projects.
The equity premium is

an extremely
important number,
even without the
CAPM.
236
ANECDOTE
Was the 20th Century Really the American Century?
The compound rate of return in the United States was about 8% per year from 1920 to 1995. Adjusted for
inflation, it was about 6%. In contrast, an investor who had invested in Romania in 1937 experienced not
only the German invasion and Soviet domination, but also a real annual capital appreciation of about 27%
per annum over its 4 years of stock market existence (19371941). Similar fates befell many other Eastern
European countries, but even countries not experiencing political disasters often proved to be less than stellar
investments. For example, Argentina had a stock market from 1947 to 1965, even though its only function
seems to have been to wipe out its investors. Peru tried three times: From 1941 to 1953 and from 1957 to
1977, its stock market investors lost all their money. But the third time was the charm: From 1988 to 1995,
its investors earned a whopping 63% real rate of return. Indias stock market started in 1940 and offered its
investors a real rate of return of just about 1% per annum. Pakistan started in 1960 and offered about 0.1%
per annum.
Even European countries with long stock market histories and no political trouble did not perform as well as
the United States. For example, Switzerland and Denmark earned nominal rates of return of about 5% per
annum from 1920 to 1995, while the United States earned about 8% per annum. A book by Dimson, Marsh,
and Staunton looks at 101 years of global investment returns and argue that measurement and hindsight biases
can account for much of this superior return.
Nevertheless, the United States stock market was an unusual above-average performer in most of the twentieth
century. Will the twenty-first century be the Chinese century? And do Chinese asset prices already reflect this?
Or already reflect too much of this?
Goetzmann and Jorion (1999)
The CAPM is about

relative pricing, not
absolute pricing.
high, you would want to allocate a lot of your personal assets to stocks. Otherwise,
you would allocate more to bonds. You really do need to know the equity premium even
for basic investing purposes, toono escape possible.
In a corporate context, like every other corporate manager, you cannot let your
limited knowledge of the equity premium stop you from making investment decisions.
In order to use the CAPM, you do need to judge the appropriate reward for risky projects
relative to risk-free projects. Indeed, you can think of the CAPM as telling you the relative
expected rate of return for projects, not the absolute expected rate of return. Given your
estimate of how much risky average stock market projects should earn relative to safe
projects, the CAPM can tell you the costs of capital for projects of a specific beta. But
the basic judgment of the appropriate spread between high-beta and low-beta projects
is left up to you.
Q 9.13. What are appropriate equity premium estimates? What are not? What kind of
reasoning are you relying on?
237
Investment Projects Market Betas

Your third CAPM input is your projects market beta (i ). It measures how the rate of
return of your project fluctuates with that of the overall market. Unlike the previous
two inputs, which are the same for every project in the economy, the beta input depends
on your specific project characteristics: Different investments have different betas.
Unlike the risk-free

rate and the equity
premium, beta is
specific to each
project.
The Implications of Beta for a Projects Risk and Reward

You already understand the role of market beta in determining the expected rate of
return for an asset. This is the security market linethat is, the CAPM formula itself is
an upward-sloping line when the expected rate of return is plotted against beta. But
market beta also has implications for the standard deviation of assets. First, note that
assets with a low beta are not very exposed to market risk. Thus, assets that have either
a very high or a very low market beta tend to have higher standard deviation. Second,
note that you can only learn much about an assets market-beta in months in which the
market does not turn in the same performance as the risk-free security. If the market
and the risk-free asset turn in the same performance in a given month, then any assets
expected rate of return is just the risk-free rate, regardless of its market-beta.
Projects with higher

betas have more
market risk, so their
own idiosyncratic
variances tend to be
higher, too.
Beta Estimation
How do you find good forward-looking market-beta estimates for your own project? As
usual, when we do not know the input, we rely on statistical analysis of past data. The
mechanics of finding the beta for a stock are easy. You run a market-model regression
on historical stock returns. The independent variable is the rate of return on the stockmarket (the S&P500 percent change, even without dividends, is usually good enough).
The dependent variable is the rate of return on your project. Usually, you should run
such regressions with daily rather than with monthly returns and you should use about
3-5 years of data. Any statistical package (and common computer spreadsheet programs)
readily give you the regression coefficients. The slope is the historical market-beta.
Unfortunately, although estimates of future betas are better than estimates of the
future equity premium, they are still not great. The reason is that stock returns are very,
very noisy. (And projects are rarely the same as stock, and project and stocks both often
change their characteristics over time, too, but lets ignore this for the moment.) Thus,
statisticians recommend that you should shrink your beta estimates further. Shrinking
comes in two forms:
Instead of using your own historical rates of returns, use the historical rates of
return on a broader portfolio. For example, if you want to estimate the future
market-beta of AMD, do not use the historical rates of return of AMD in your
market-model, but those of the computer hardware sector instead. In other
words, assume that all computer hardware makers have about the same stock
market beta, and that AMDs own future beta will look more like that of its sector
in the past than like that of its own past.
Instead of using the coefficient estimate from the regression, use an average
between the regression estimate and the number 1 (which is the average of
Ways to estimate
beta.
238
Chems
BldMt
1.4
BldMt
1.4
Rubbr
Rubbr
ElcEq
ElcEq
Paper
1.2
PerSv
Guns
1.0
Banks
Softw
Soda
Ships
Util
Meals
Fun
Mines
Drugs
Beer
Oil
Clths
Smoke
Other
Food
Insur
Autos
Mach
Boxes
Steel
Telcm
Gold
BusSv
RlEst
Cnstr
Whlsl
Coal
LabEqHardw Txtls
0.8
MedEq
Books
Toys
Agric
Retail
FabPr
Hshld
Fin
0.6
Aero
Chips
0.4
Correlation: 0.61
Beta 2010 = 0.35 + 0.63*Beta 2006
0.2
49 Industries' Equity MarketBeta 2010
MedEq
Paper
1.2
Cnstr
Hlth
PerSv
Autos
Guns
Banks Mach
1.0
Util
LabEq
etail
0.8
Whlsl
Fun
Telcm
Soda
Softw
Boxes
Coal
Hardw
Txtls
Oil
Beer
Gold
BusSv
Clths
RlEst
Steel
Mines
Ships
Drugs
Toys
Agric
Books
Smoke
Other
FabPr
Food
Insur
Meals
Hshld
Fin
0.6
Aero
Chips
0.4
Correlation: 0.05
Beta 2010 = 0.96 + 0.05*Beta 2002
0.2
0.0
0.0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Exhibit 9.4: Betas For 49 Industries Far Into The Future. These figures plot industry market betas at the end of
2010 against their own value a few years earlier. Industries that had high market-betas in 2006 still tended to
have high market-betas in 2010although you should have not have used your exact estimates but shrunk
them towards 1 to reflect their tendency to mean-revert. In contrast, industries that had high market-betas in
2002 unfortunately did not have high market-betas in 2010. If you had to guess market-betas in 2002 for 2010,
you may as well have guessed the same value for every industry, ignoring the prevailing 2002 market-betas. The
0.05 coefficient is unusually low. In other eight-year samples, it was more like 0.3. Data Sources: 49 industries
from Fama-French. Betas from 3 years of daily data.
all stocks market-betas). For example, if your market-model coefficient estimate

based on past data is 2.6, use 1/2 2.6 + 1/2 1.0 = 1.8 for your estimate of
the future. Many studies have confirmed that such shrunk market-betas perform
better in predicting subsequent market-betas than the unshrunk coefficient estimates themselves. The market-betas that are posted on many websites, such as
FINANCE, are also shrunk.
Unfortunately, while these two shrinkages combined work reasonably well for predicting
stock market-betas over the next quarter, they do not work so well for predicting stock
market betas for cash flows that will occur in many years. Figure 9.4 shows how the
stock market-betas for 49 different industries and then shrunk again. These industry
239
betas typically range from about 0.3 to about 1.5, but change over time. The left panel
shows that 2006 market-betas were still similar to those in 2010. The right panel shows
that 2002 market-betas were not. (The left panel was better than usual, the right panel
was worse than usual.) Based on a more detailed statistical study, my advice is to shrink
the market-betas for cash flows in more than 2-5 years a second time. In our example of
an industry market-beta of 2.6, shrunk once to 1.8 for cash flows that occur within the
next year, if you had to assess the market betas of cash flows in about 5 to 15 years, you
would shrink your beta a second time, say to 1/2 1.8 + 1/2 1.0 = 1.4.
Unfortunately, as a corporate manager, you are rarely interested in the market-beta
of an industry or even a stock. Usually, you are interested in the market-beta of a new
project that you are considering. Sometimes, your firm is not even publicly traded, so
you would not even have historical data if you wanted to. (And, if not publicly traded,
then it is quite possible that your investors would not have been fully diversified, which
is an essential assumption in the CAPM. If your main investor is undiversified, you
may care about idiosyncratic standard deviation more than about the market-beta.) In
this case, corporate CAPM users must thus rely more on economic intuition than pure
statistics. You can rearrange the CAPM formula to obtain a beta estimate. Now, do
you think your project cash flows and its future project value (which is influenced by
changes in the economy) is likely to move more or less with the overall stock market
(and, possibly, the overall economy)?

E ri rF

E ri = rF + E rM rF i i =
E rM rF
The right side of this formula helps translate your intuition into a beta estimate. What
rate of return (above the risk-free rate) will your project have if the market were to
have +10% or 10% rate of return (above the risk-free rate)? Clearly, such guesswork is
difficult and error-pronebut it can provide a beta estimate when no other is available.
Or, perhaps you can start with an industry market-beta and shrink it appropriately,
perhaps adjusting for the fact that some (smaller) firms typically have higher betas?
Equity and Asset Betas Revisited

No matter how good your estimates of your stock betas are, it is important that you
always distinguish between asset betas and equity betas. Let me remind you with an
example. Assume that the risk-free rate is 4% and the equity premium is 5%. You own
a $100 million project with an asset beta of 2.0 that you can finance with $20 million
of risk-free debt. By definition, risk-free debt has a beta of 0. To find your equity beta,
write down the formula for your asset beta (firm beta):
20%
(0) +
80%
(Equity ) = 2.0

Debt value
Equity value
Firm =
Debt +
Equity
Firm value
Firm value
Solve this to find that your equity beta is 2.5. This is what you would find on
FINANCE. You would not want to base your hurdle rate for your firms typical
average project on the equity beta: Such a mistake would recommend you use a hurdle

rate of E ri = rF + E rM rF i = 4% + 5% 2.5 = 16.5%. This would be too high.

Instead, you should require your average projects to return E ri = 4%+5%2.0 = 14%.
Dont use the equity

beta to estimate your
projects hurdle rate.
Use the asset beta
instead.
Asset and equity betas,

Formula 8.7, Pg.212.
Typical, average, and

marginal betas,
Sect. 12.3, Pg.343.
240
If you use
comparables, first
unlever them.
Credit ratings,
Sect. 6.2, Pg.130.
Conversely, if your project is private but the potential future owners are welldiversified, you may have to find its hurdle rate by looking at public comparables. Lets
FINANCE lists
presume you find a similarly-sized firm with a similar business that
with a beta of 4, or perhaps better yet, the firms industry. Remember that financial
websites always list only the equity beta. The CAPM tells you that the expected rate of
return on the equity is 4% + 5% 4 = 24%. However, this is not necessarily the hurdle
FINANCE, you may notice that
rate for your project. When you look further on
your comparable is financed with 90% debt and 10% equity. (If the comparable had very
little debt, a debt beta of 0 might have been a good assumption, but, unfortunately, in
this case it is not.) Corporate debt rarely has good historical return data that would allow
you to estimate a debt beta. Consequently, practitioners often estimate the expected rate
of return on debt via debt comparables based on the credit rating. Say your comparables
debt is rated BB and say that BB bonds have offered expected rates of return of 100 basis
points above the Treasury. (This might be 200 basis points quoted above the Treasury).
With the Treasury standing at 4%, you would estimate the comparables cost of capital
on debt to be 5%. The rest is easy. The expected rate of return on your project should
be
E rProject
90% 5%
= wDebt E rDebt
= 6.9%
10% 24%
+ wEquity E rEquity
This would make a good hurdle rate estimate for your project.
Q 9.14. According to the CAPM formula, a zero-beta asset should have the same
expected rate of return as the risk-free rate. Can a zero-beta asset still have a positive
standard deviation? Does it make sense that such a risky asset would not offer a higher
rate of return than a risk-free asset in a world in which investors are risk averse?
Q 9.15.A comparable firm (with comparable size and in a comparable business) has a
FINANCElisted equity beta of 2.5 and a debt/asset ratio of 2/3. Assume that the
debt is risk free.
1. Estimate the equity beta for your firm if your projects have similar betas, but your
firm will carry a debt/asset ratio of 1/3.
2. If the risk-free rate is 3% and the equity premium is 2%, then what should you
use as your firms hurdle rate?
3. What do investors demand as the expected rate of return on the comparable firms
equity and on your own equity?
Q 9.16. You own a stock portfolio that has a market beta of 2.4, but you are getting
married to someone who has a portfolio with a market beta of 0.4. You are three times
as wealthy as your future significant other. What is the beta of your joint portfolio?
9.5. Is the CAPM the Right Model?
9.5
241
Is the CAPM the Right Model?
Now you know how securities should be priced in a perfect CAPM world, in which
investors have good knowledge of the parameters. What would happen if a stock offered
more than its appropriate expected rate of return? Investors in the economy would
want to buy more of the stock than would be available: Its price would be too low. It
would be too good a deal. Investors would immediately flock to it, and because there
would not be enough of this stock, investors would bid up its price and thereby lower
its expected rate of return. The price of the stock would settle at the correct CAPM
expected rate of return. Conversely, what would happen if a stock offered less than its
due expected rate of return? Investors would not be willing to hold enough of the stock:
The stocks price would be too high, and its price would fall. Neither situation should
happen in the real world.
Is this an arbitragea free money situation? No. When stocks do not to follow the
CAPM formula, buying them is still risky. Yes, some stocks would offer a higher or lower
expected rate of return and thus seem to be too good or too bad a deal, attracting too
many or too few investors chasing a limited amount of value in this stockbut these
stocks would still remain risky investments. No investor could earn risk-free profits.
There is no arbitrage here. The market forces working on correcting the (CAPM) mispricing are modest. And remember that there are good reasons why the CAPM may not
hold in the first place, too. For example, it relies on many perfect-market assumptions.
If investors are taxed or liquidity-constrained (that is, they cannot easily diversify, e.g.,
because the firm is a startup or family firm) or do not agree on the inputs, then it is
quite plausible that some firms or even sectors (such as value firms or growth firms)
would offer higher or lower expected rates of return than the CAPM suggests.
Q: What happens if a
stock offers too much
or too little expected
rate of return? A:
Investor stampedes.
Assets not priced

according to the
CAPM do not allow
you to make money
for nothing. However,
it could imply good
deals.
What is The Scientific Evidence?

Unfortunately, in real life, despite its wide use, the evidence in favor of practical use and
application of the CAPM is either weak or non-existent. If you use the CAPM, you do so
based primarily on a belief that it should work, not based on empirical evidence. Say
again: the evidence suggests that, even if the CAPM held, input estimates for corporate
cash flows that will occur far in the future are usually so imprecise that they render the
CAPM practically useless.
Why use the CAPM?
Huh? Did you really read me right?

If there is no empirical evidence that CAPM use is justified, then why do we torture
you with it? This is a much easier question to answer than how stocks are priced in the
real world or what the best estimate of the appropriate hurdle rates for your project is.
Good intuition: The CAPM has impeccable intuition. It is a model that shines through
its simplicity and focuses on what should matter when owners are many
diversification. It gets executives away from the false notion that many small
public investors care about the idiosyncratic risk of projects that the investors can
diversify away. It also helps you understand that corporate diversification into a
conglomerate is not likely to add value. Your investors can diversify themselves.
The CAPM is based

on the important
concept of
diversification.
242
They dont need your firm to diversify you for them. And, it explains nicely why
stocks should have higher rates of returns than bonds and how to lever and
unlever assets. In general, it is a nice conceptual framework that helps you think
about what should matter.
Faith.
Strong Belief: Many instructors and practitioners find the CAPM to be so plausible that
they are willing to live with absence of CAPM evidence. They do not take this
absence to mean evidence of CAPM absence. Thus, they adopt the CAPM based
on their prior belief and faith, not based on evidence. Doing this is acceptable as
long as you are fully aware that this is really what you are doing. (However, even
if you do adopt the CAPM and even if this is not a Rumsfeld-level blunder, you
still have to realize that you should greatly shrink your beta and equity-premium
inputs for long-term cash flows.)
A crutch
Standin for Expected Cash Flow Default: The CAPM often assigns higher costs of
capital to projects that are more likely to fail. If you have not fully adjusted your
expected cash flow estimates downwards to adjust for failure (a common human
error), the CAPM cost of capital often helps to impose a higher hurdle rate on
riskier cash flows.
Important: Everyone
expects you to know
the CAPM!
Everyone uses it: The CAPM is the standard. Exhibit 9.5 shows that 73% of the CFOs
reported that they always or almost always use the CAPM. (And use of the CAPM
was even more common among large firms and among CFOs with an MBA.) No
alternative method was used very often. Consequently, you have no choice but
to understand the CAPM model wellif you will work for a corporation, then the
CAPM is the benchmark model that your future employer will likely use and will
expect you to understand well. Again, the CAPM is simply the standard. The CAPM
is also used as a benchmark by many investors rating their (investment) managers,
by government regulatory commissions, by courts in tort cases, and so on. It is
literally the dominant, if not the only, widely-used model to estimate the cost of
capital. Indeed, there is a whole section on the CFA exam about the CAPM!
There is no
generally-used
alternative to the
CAPM.
Alternativesplease stand up: The famous sociologist Lewin wrote that there is
nothing more practical than a good theory. If not the CAPM, then what else would
you use? There are no commonly-accepted alternatives. (A related justification
for the CAPM has been that we consider the CAPM like linguists consider Latina
good language that prepares you well to learn other languages that descended
from it. The problem is that the CAPM-descendant models dont work well, either.
At best, they are so flexible and slippery that we cannot know whether they work
or not. At worst, they or their use has been rejected by the data, too.)
Do you want a
bedtime story that
the world is ok in
order to be able to go
to sleep?
Be aware that my treatment of the CAPM in an introductory corporate-finance

textbook borders on heresy. Most corporate finance text-books make the CAPM their
centerpiece. They do this not because the authors believe in it, but because it is dogma
that new finance students are too fragile to deserve the hard truth. I am sorryI wish I
could have told you a happy bed-time story about how the world is nice and orderly,
too. But it would have been a lie.
Now, if you still want to use the CAPM, here is my advice. As a corporate executive,
you should always first think hard about when you want to use the CAPM. Think about
Never make the

following errors,
please.
Method
243
Usage Frequency
CAPM
Historical Average Returns
Modified CAPM
Backed Out from Gordon Model
Whatever Investors Tell Us
(73%)
(39%)
(34%)
(16%)
(14%)
Usage Recommendation
With Caution
Rarely
With Caution
Occasionally
Occasionally
Explained in
Chapter 9
Chapter 8
Chapter 9
Chapter 3
Chapter 2
Exhibit 9.5: CFO Valuation Techniques for the Cost of Capital. Rarely means usually no, and often used
incorrectly. Not reported, use of the CAPM is more common among managers with an MBAand in firms who
rely on consultants who in turn use the CAPM. Original Source: John Graham and Campbell Harvey, 2001.
ANECDOTE
Cost of Capital Expert Witnessing
When Congress tried to force the Baby Bells (the split-up parts of the original AT&T) to open up their local
telephone lines to competition, it decreed that the Baby Bells were entitled to a fair return on their infrastructure
investmentwith fair return to be measured by the CAPM. (The CAPM is either the de facto or legislated
standard for measuring the cost of capital in many other regulated industries, too.) The estimated value of the
telecommunication infrastructure in the United States is about $10 to $15 billion. A difference in the estimated
equity premium of 1% may sound small, but even in as small an industry as local telecommunications, it meant
about $100 to $150 million a yearenough to hire hordes of lawyers and valuation consultants opining in court
on the appropriate equity premium. Some of my colleagues bought nice houses with the legal fees.
I did not get the call. I lack the ability to keep a straight face while stating that the equity premium is exactly x
point y percent, which was an important qualification for being such an expert. In an unrelated case in which I
did testify, the opposing expert witness even explicitly criticized my statement that my cost-of-capital estimate
was an imprecise rangeunlike me, he could provide an exact estimate, and it was 11% per year!
Bradford Cornell, UCLA
whether it is useful for your own cost-of-capital estimates, or whether the CAPM errors
seem too large to be useful for your particular needs. Here is what I would definitely
warn about:
Accuracy: The CAPM is a poor model if you want precision. If you believe that CAPM
expected rates of return should be calculated with any digits after the decimal
point, then you are deluded. Please realize that, at best, the CAPM can only offer
expected rates of return that are of the right order of magnitude, plus or minus
a few percentage points perhaps. Actually, if accuracy and precision are important,
you are in trouble. We do not have any models that can offer it. (Fortunately, it is
often less important to be accurate than it is to be better estimating value than
Dont expect accuracy

and dont use it for
financial investing.
244
your competitors. And always remember that valuation is as much an art as it is a

science.)
Avoid using the
CAPM for financial
investment purposes.
Mean-variance
optimization in detail,
Sect. App.8.C (Companion),
Pg.35.
Corporate
Time-Varying Costs of
Capital, Sect. 5.5,
Pg.112.
Investment purposes: If you are not a corporate executive looking to determine your
project hurdle rate, but a financial investor looking for good investments from
the universe of financial instruments, with an ability to shift your money around
every day, then please do not use the CAPM. Although the CAPM offers the correct
intuition that wide diversification needs to be an important part of any good
investment strategy, there are many better investment strategies than just investing
in the market index. Some are explained in Section App.9.C (Companion); more
will be discussed in an advanced investments course.
Please do not confuse the CAPM with the mean-variance framework discussed in
the previous chapter. Mean-variance optimization is an asset-selection technique
for your individual portfolio, and it works, regardless of whether or not the CAPM
holds.
Long-Term Differences: If you are a corporate executive, be cautious. Look at your
cost of capital more holistically. The CAPM has two terms.
The first term is the risk-free rate which applies to all projects, regardless of beta.
Fortunately, there is great evidence what you should use. You should use higher
costs of capital for cash flows that will occur in the more distant future. And
you have a great estimate of the premium that long-term projects need to offer
over short-term projects, based on the Treasury yield-curve. You dont even need
historical estimates: you can use the prevailing Treasury yield curve. Use it! It
works!
It is the second term (the beta multiplied by the risk-premium), i.e., your beta
risk-adjustment, that is dubious. If your cash flows will occur in many years, be
modest. Do not overstate the risk-inputs in the CAPM. Shrink and shrink again.
As a corporate manager, compare the cost of capital on your equity vs. the cost
of capital on your debt for your long-term cash flows. With an equity premium
based on the performance of stocks vs. long-term Treasuries of about 1-2%
from 1970 to today, it may not matter much whether your project A has a
beta of 0.8 and your project B has a beta of 1.2. The implied cost-of-capital
difference between these two projects of under (1.2 0.8) 2% 1%/year is
already small.
Asset betas are often

even closerthey
often give it
time-stability, though.
For long-term cash flows, your best estimate of your equity market-betas
should be tilted much more towards 1 than what you think your market-beta
is today. Thus, if you fit your historical market-beta to be 0.5 for A and 1.5
for B today, you may well want to use a market-beta shrunk to around 0.9
for A and 1.1 for B if those equity cash flows will occur in 10-20 years. Think
about this: A and B would now have a different implied cost of equity capital
of 0.2 2% 0.4%. This is way below your noise-and-uncertainty threshold.
But lets continue. Say your projects are partly debt-financed, too. Now
you need to calculate asset-betas rather than equity betas. Lets say both
projects have 50% debt that is almost risk-free. Then your asset beta would
be 0.5 0.0 + 0.5 0.9 = 0.45 for A and 0.5 0.0 + 0.5 1.1 = 0.55 for B.
245
Now you have a project cost of capital difference (0.55 0.45) 2% 0.2%
between A and B.
How does this expected rate of return difference between A and B compare to
your own uncertainty about your projects relative expected cash flows? Does the
CAPM beta risk-adjustment really matter much in light of your uncertainty?
The estimated CAPM

cost of capital for
long-term cash flows
are fragile.
Alternatives
Let me summarize what I believe the data do tell us that is solid enough a rock to build
a house on it:
What is solid
empirical evidence?
There definitely is a time-value of money.

There definitely is a term structure. Long-term cash flows usually require higher
costs of capital than short-term cash flows. Your investors can earn higher expected
rates of return elsewhere for longer-term commitments, too.
There definitely is a credit component. Assets with higher probabilities of default
have to make up for it with higher promised yields; that is, higher yields when
they succeed.
As a preview to Chapter 10, market imperfections seem to play a role. There seems
to be a liquidity premium. Assets that can be quickly liquidated in a market crash
are more expensive, and different asset classes seem to have different degrees
of liquidity. Because of their collateral, mortgage debt tend to have lower costs
of capital than general bonds. Firms with less access to capital markets, such
as startups, seem to pay higher costs of capital, although adjusting for default
makes this difficult to measure. Investors pay more in personal income tax for
interest receipts than they do for capital gains, which makes equities relatively
more desirable and reduces their after-tax income. And sentiment and agency
considerations seem to play a role in equity trading that is not unimportant. Many
of these market imperfections embody some concept of risk, but it is not the
market-beta.
Market Imperfections,
Chapter 10, Pg.257.
After taking into account the premia just mentioned, the remaining equity premium is probably relatively small (1-2%), although we do not know for sure.
Our uncertainty is much larger than our certainty about its magnitude. And you
need to realize that betas for cash flows far into the future are much closer to 1
than historical regressions would suggest. The CAPM beta impact is relatively
unimportant.
So what would I do if I was not constrained by my boss? My best alternative
cost-of-capital recommendation would start out just like the CAPM: As the first term
in a formula, I would recommend that you use the rate of return on bonds of similar
maturity as the cash flow that you want to value. Usually, this means that you assign
higher costs of capital to cash flows farther in the future. It is only on the second
term, the equity risk-adjustment, that I would tinker. Instead of the (shrunk) CAPM
market-beta multiplied by the historical equity premium (of 2% or less per annum), I
would recommend a more holistic approach.
Use reasonable risk

adjustments.
246
Take into consideration that projects with high volatility and/or with high leverage
are more risky. The equity on these projects probably requires a higher expected
rate of return to keep your investors happy. Projects with higher idiosyncratic risk
are also the same projects where executives are often the most over-optimistic.
(Check again: are you sure your expected cash flows in the NPV numerator are
not over-confident?)
Take into consideration whether you and your owners are well-diversified. If you
are not, then you should require higher rates of return on riskier projects. In this
case, it is not beta risk that matters, but total risk.
Take into consideration that your investors may like growth firms and are often
willing to pay higher prices and thus accept lower average rates of returns for
some such projects.
Long-Run Excess
Profits, Sect. 20.3,
Pg.672.
What would I do if the

boss liked the CAPM?
Income Taxes and Cost

of Capital, Chapter 17,
Pg.545.
Comparables,
Chapter 14, Pg.431.
There is little harm if you calculate a (repeatedly-shrunk) CAPM market-beta with a

low equity premium (say 2%) to assess whether any other non-CAPM cost-of-capital
assessments seems reasonably similar to your CAPM assessment. In this sense, the CAPM
can still be a little helpful. Finally, realize that it is in general very difficult to assess
over many years whether corporate projects will offer higher or lower average rates of
return than the average project in the economy. If you make smart decisions, after your
projects initial growth phase is over, would it be reasonable to assume that it will earn
similar rates of returns as most other good projects in the economynot better, not
worse?
And if my boss required an approach like the CAPM, what would I do?
If I ran a large firm with good access to capital markets, I would assume an equity
premium of 1-2% per annum and apply this to the equity components of all my
long-term cash flows. The exception would be projects for which I would have a
strong prior that their market-betas will be very extreme, say, below 1 or greater
than 3 (and I would then shrink those betas further to, say, 0 and 1.5, respectively,
to account for long-term uncertainty about betas). I would consider long-term
corporate debt to have a higher cost of capital than equivalent Treasuries but
a lower cost of capital than my own equitythe latter primarily because debt
provides a corporate income tax shield (as you will learn in Chapter 17) and not
because the equity premium over long-term corporate bonds is high.
Deviating from the CAPM, if I ran a startup firm, I would assume a cost of capital
of 2% to 6% above the expected rate of return on my uncollateralized debt. The
expected rate of return on my debt could be very highit could even be in the
double digits. (This reflects the fact that more volatile cash flows and firms that
struggle with more market imperfections must pay higher costs of capital.) Risk
definitely plays a role, but not in the strict CAPM market-beta sense. Alternatively,
I would abandon NPV-based models altogether and try to estimate what other
similar projects are offering their investors. This is the route we take in Chapter 14.
And I would never use any of my schemes here (or the CAPM) for the pricing of bonds,
derivatives, or other extreme kinds of projects.
Am I the only professor who recommends against using the CAPM? No. Eugene
Fama, perhaps the most famous active finance professor alive and partly responsible for
the original spread of the CAPM, nowadays strongly recommends against the combined
use of NPV models with asset-pricing models like the CAPM, where you use the CAPM
expected rate of return as your cost of capital in an NPV calculation. Such use means
you divide one uncertain number by another. This practice combines your errors and
uncertainty about expected cash flows in the numerator with your errors and uncertainty
about expected returns in the denominator. Yikes!
247
NPV or Comparables?
Eugene Fama thinks
Comparables are
better.
Conclusion
IMPORTANT
The CAPM is the benchmark model in the real world. Most corporations use it.
Everyone will expect you to understand the CAPM. Regardless of whether the
model holds or not, you have to know it.
The empirical evidence suggests that the CAPM is not a great model for predicting
expected rates of return.
The first CAPM term (that long-term projects have to offer higher expected rates
of return) seems to hold better than the second CAPM term (the risk adjustment).
For cash flows many years into the future, you must realize (a) that market-betas
revert back towards 1 and (b) that the equity premium is low.
The CAPM never offers great accuracy.
Mean-variance optimization (Section 8.2) works even if the CAPM does not.
Q 9.17. Does the empirical evidence suggest that the CAPM is correct?
Q 9.18. If the CAPM is wrong, why do you need to learn it?
Q 9.19. Is the CAPM likely to be more accurate for a project where the beta is very high,
one where it is very low, or one where it is zero?
Q 9.20. To value an ordinarily risky project, that is, a project with a beta in the vicinity
of about 1, what is the relative contribution of your personal uncertainty (lack of
knowledge) in (a) the risk-free rate, (b) the equity premium, (c) the beta, and (d) the
expected cash flows? Consider both long-term and short-term investments. Where are
the trouble spots?
248
Summary
This chapter covered the following major points:
The CAPM provides an opportunity cost of capital for investors, which corporations can use as the cost of capital in the NPV formula. The CAPM formula is
E ri

= rF + E rM rF i
Thus, there are three inputs: the risk-free rate of return (rF ), the expected rate of

return on the market (E rM ), and the projects or firms market beta (i ). Only
the latter is project-specific.
The line plotting expected rates of return against market beta is called the security
market line (SML).
The CAPM provides an expected rate of return, consisting of the time premium and
the risk premium. It ignores the default premium. In the NPV formula, the default
risk and default premium work through the expected cash flow in the numerator,
not through the expected rate of return (cost of capital) in the denominator.
For rF , you should use bonds that match the timing of your projects cash flows.
Thus, cash flows farther in the future often require higher opportunity costs of
capital. Even if you do not believe the CAPM, term adjustment is important.
The expected rate of return on the market is a critical CAPM input if market beta
is highbut it is difficult to guess. There are many guesstimation methods, but no
one really knows which one is best. Reasonable estimates for the equity premium

(E rM rF ) can range from about 1% to 8% per annum, although 2% seems
most reasonable to me for cash flows more than a few years into the future.
There are a number of methods to estimate market beta. Many users rely on
industry betas and not on firms own historical betas as estimates of future market
betas, and they shrink them towards 1. When your cash flows are farther in the
future, you have to shrink your beta estimates even more drastically towards 1.
Never believe the CAPM blindly. Its estimates are poor. Use it more like a general
direction estimate than like an accurate guide estimate.
Even though its estimate are poor, understand the CAPM well. Everyone will
expect you to.
The chapter appendix discusses certainty equivalence and CAPM alternatives
(such as the APT and the Fama-French-Momentum model). You must use the
certainty equivalence form of the CAPM when projects are purchased or sold for
prices other than their fair market values. It is also often the only method if only
underlying cash flows rather than value estimates are available.
End of Chapter 9 Material
249
This negative perspective on the CAPM is so uncommon in a textbook (but not

among the experts actually studying the models) that it is important that you dont
misunderstand it. Lets end this chapter with a FAQ:
Q: Should riskier projects not have to promise higher rates of return?
A: Riskier projects have to promise a higher rate of return, i.e., offer
higher default premiums. This is not the same as higher risk premiums
in the CAPM sense. In NPV applications, make sure to reflect the default
risk in the expected cash flow numerator. Riskier projects need to pay
off a lot more when they succeed, just to make up for the fact that they
fail more often.
Q: Should riskier long-term cash flows not require higher expected rates of return?
A: Long-term projects command term premiums. Thus, in NPV applications, you should usually use higher required costs of capital for more
distant cash flows. You can but do not need the CAPM for this. The
U.S. Treasury Yield Curve gives you a working first estimate about how
much extra premium long-term cash flows should require.
Q: Should riskier stocks and cash flows have higher expected discount rates?
A: Maybe, but be careful. First, make it modest. Dont be too overconfident in your ability to judge equity risks. If you can judge the risks well,
make sure your estimates first flow into your expected cash flows in
the NPV numerator. Second, dont be too wedded to the CAPM for an
extra risk-premium kicker. Instead, combine your cost-of-capital estimate with judgment-based and other risk measures, such as volatility
(especially if your owners are not fully diversified).
Preview of the Chapter Appendix in the Companion

The appendix to this chapter explains
the certainty equivalence value (CEV) which allows you to use the CAPM for
projects that you are not buying at the appropriate equilibrium price. For example,
you need the CEV to work out how to value an inheritance that will be higher if
your business fails. (Being free today does not mean that there is no value to such
a promise.)
how to use the CEV formula to estimate the value of a project for which you have
historical cash flows, but no market value information.
how the CAPM is derived from the fact that the optimal portfolio is always the
combination of two portfolios, one of which may be the risk-free asset.
what the CAPM alternatives are and how to use them. The first alternative is
the APT (arbitrage pricing theory) and its relative, the Intertemporal CAPM. The
In the
Appendix
250
second alternative are Fama-French value and momentum models. These seem to
predict better than any alternatives, but are less grounded in theory (or, you may
say, reason) than the former.
Keywords
Asset-pricing model, 221.
Bubble, 233.
CAPM, 220.
Capital asset pricing model, 220.
Certainty
equivalence, 221.
Dot-com bubble, 233.
Dow Jones 30, ??.
Equity premium, 221.
Internet bubble, 233. Market beta, 237. Market risk premium, 221. Market-model, 237. Peso problem, 232. Risk
premium, 225. SML, 222. Security market line, 222. Shrinking, 237. Tech bubble, 233.
Answers
Q 9.1 Yes, the perfect market is an assumption underlying the
CAPM. In addition,
Q 9.7 The security market line is

15
2. Investors care only about overall portfolio mean rate of return

and risk at one given point in time.
10
3. All parameters are known (not discussed until later in the

chapter).
4. All assets are traded. Every investor can purchase every asset.

Q 9.2 With rF = 4% and E rM = 7%, the cost of capital for
a project with a beta of 3 is E r = rF + [E rM rF ] i =
4% + (7% 4%) 3 = 13%.

Q 9.3 With rF = 4% and E rM = 12%, the cost
of capital
for a project with a beta of 3 is E r = rF + [E rM rF ] i =
4% + (12% 4%) 3 = 28%.

Q 9.4 With rF = 4% and E rM =
12%, the cost of capital for
a project with a beta of 3 is E r = rF + [E rM rF ] i =
4% + (12% 4%) (3) = 20%. Yes, it does make sense that a
project can offer a negative expected rate of return. Such projects
are such great investments that you would be willing to expect
losses on them, just because of the great insurance that they are
offering.
Q 9.5 Nothe real-world SML is based on historical data and
not true expectations. It would be a scatterplot of historical risk and
reward points. If the CAPM holds, a straight, upward-sloping line
would fit them best.

Q 9.6
Write down the CAPM formula and solve E ri = rF +
[E rM rF ] i = 4% + (7% 4%) i = 5%. Therefore, i = 1/3.
Note that we are ignoring the promised rate of return.
E(ri), in %
1. Investors are rational utility maximizers.
Market M
Riskfree
Treasury
5
1.0
0.5
0.0
0.5
1.0
1.5
2.0
True MarketBeta (i)

Q 9.8 The equity premium, E rM rF , is the premium that the
market expects to offer on the risky market above and beyond what
it offers on Treasuries.
Q 9.9 It does not matter what you choose as the per-unit payoff
of the bond. If you choose $100, you expect it to return $99.
1. Thus, the price of the bond is PV = $99/(1 + [3% + 5% 0.2])
$95.19.
2. Therefore, the promised rate of return on the bond is
$100/$95.19 1 5.05%.
3. The risk-free rate is 3%, so this is the time premium (which
contains any inflation premium). The (expected) risk premium is 1%. The remaining 1.05% is the default premium.
Q 9.10 The cost needs to be discounted with the current interest
rate. Because payment is up-front, this cost is $30,000 now! The
appropriate expected rate of return for cash flows (of your earnings)
is 3% + 5% 1.5 = 10.5%. You can now use the annuity formula to
determine the PV if you graduate:
$5,000
10.5%

1
1
1 + 10.5%
251
40
$47,619 98.2%
$46,741.46
With 90% probability, you will do so, which means that the appropriate risk-adjusted and discounted cash flow is about $42,067.32.
The NPV of your education is therefore about $12,067.32.
Q 9.11 Use the 1-year Treasury rate for the 1-year project, especially if the 1-year project produces most of its cash flows at the end
of the year. If it produces constant cash flows throughout the year, a
6-month Treasury rate might be more appropriate. Because the 10year project could have a duration of cash flows much shorter than
10 years, depending on use, you might choose a risk-free Treasury
rate that is between 5 and 10 years. Of course, it would be even
better if you match the individual project cash flows with individual
Treasuries.
Q 9.12 The duration of this cash flow is around, or a little under, 5 years. Thus, a 5-year zero-coupon U.S. Treasury would be
a reasonably good guess. You should not be using a 30-day or
30-year Treasury. A 10-year zero-coupon Treasury would be a better
match for a project that yields cash only once at the end of 10 years.
That is, for our project that has cash flows each year for 10 years,
the 10-year Treasury as a benchmark would have too much of its
payments as principal repayment at the end of its 10-year term.
Q 9.13 An estimate between 1% and 8% per year is reasonable.
Anything below 0% and above 10% would seem unreasonable to
me. For reasoning, please see the different methods in the chapter.
Q 9.14 Yes, a zero-beta asset can still have its own idiosyncratic
risk. And, yes, it is perfectly kosher for a zero-beta asset to offer the
same expected rate of return as the risk-free asset. The reason is
that investors hold gazillions of assets, so the idiosyncratic risk of
the zero-beta asset will just diversify away.
Q 9.15 This is an asset beta versus equity beta question. Because
the debt is almost risk free, we can use Debt 0.
1. First, compute an unlevered asset beta for your comparable with its debt-to-asset ratio of 2 to 3. This
is Asset =
wDebt Debt +wEquity Equity = 2/3 0+ 1/3 2.5 0.833.
Next, assume that your project has the same asset beta, but
a smaller debt-to-asset ratio of 1 to 3, and compute your
own equity beta:
Asset =

wDebt Debt + wEquity Equity
0.833 1/3 0 + 2/3 Equity Equity = 1.25.
2. With an asset beta of 0.83, your firms asset hurdle rate
should be E ri = 3% + 2% 0.83 4.7%.
3. Your comparables equity expected rate of return would
be E rComps Equity = 3% + 2% 2.5 = 8%. Your own equitys expected rate of return would be E rYour Equity =
3% + 2% 1.25 = 5.5%
Q 9.16 Your combined happy-marriage
Combined = (3/4) 2.4 + (1/4) 0.4 = 1.9.
beta
would
be
Q 9.17 No, the empirical evidence suggests that the CAPM does
not hold. The most important violation seems to be that value firms
had market betas that were low, yet average returns that were high.
The opposite was the case for growth firms.
Q 9.18 Even though the CAPM is empirically rejected, it remains
the benchmark model that everyone uses in the real world. Moreover, even if you do not trust the CAPM itself, at the very least it
suggests that covariance with the market could be an important
factor.
Q 9.19 The CAPM should work very well if beta is about 0. The
reason is that you do not even need to guess the equity premium if
this is so.
Q 9.20 For short-term investments, the expected cash flows are
most critical to estimate well (see Section 4.1 on Page 64). In this
case, the trouble spot (d) is really all that matters. For long-term
projects, the cost of capital becomes relatively more important to
get right, too. The market betas and risk-free rates are usually
relatively low maintenance (though not trouble free), having only
modest degrees of uncertainty. The equity premium will be the most
important problem factor in the cost-of-capital estimation. Thus,
the trouble spots for long-term projects are (b) and (d).
End of Chapter Problems

Q 9.21. What are the assumptions underlying the
CAPM? Are the perfect market assumptions among
them? Are there more?
Q 9.23. In a perfect world and in the absence of externalities, should you take only the projects with the
highest NPV?
Q 9.22. If the CAPM holds, then what should you do

as the manager if you cannot find projects that meet
the hurdle rate suggested by the CAPM?
Q 9.24. Write down the CAPM formula. Which are

economy-wide inputs, and which are project-specific
inputs?
252
Q 9.25. The risk-free rate is 6%. The expected rate of

return on the stock market is 8%. What is the appropriate cost of capital for a project that has a beta of
2?
Q 9.26. The risk-free rate is 6%. The expected rate
of return on the stock market is 10%. What is the
appropriate cost of capital for a project that has a beta
of 2? Does this make economic sense?
Q 9.27. Draw the SML if the true expected rate of return on the market is 6% per annum and the risk-free
rate is 2% per annum. How would the figure look if
you were not sure about the expected rate of return
on the market?
Q 9.28. A junk bond with a beta of 0.4 will default
with 20% probability. If it does, investors receive only
60% of what is due to them. The risk-free rate is 3%
per annum and the risk premium is 5% per annum.
What is the price of this bond, its promised rate of
return, and its expected rate of return?
Q 9.29. What would it take for a bond to have a
larger risk premium than default premium?
Q 9.30. A corporate zero-bond promises 7% in one
year. Its market beta is 0.3. The equity premium is
4%; the equivalent Treasury rate is 3%. What is the
appropriate bond price today?
Q 9.31. Explain the basic schools of thought when it
comes to equity premium estimation.
Q 9.32. If you do not want to estimate the equity
premium, what are your alternatives to finding a costof-capital estimate?
Q 9.33. Explain in 200 words or less: What are reasonable guesstimates for the market risk premium
and why?
Q 9.34. Should you use the same risk-free rate of
return both as the CAPM formula intercept and in the
equity premium calculation, or should you assume
an equity premium that is independent of investment
horizon?
Q 9.35. Should a negative-beta asset offer a higher
or a lower expected rate of return than the risk-free
asset? Does this make sense?
Q 9.36. An unlevered firm has an asset market beta

of 1.5. The risk-free rate is 3%. The equity premium
is 4%.
1. What is the firms cost of capital?
2. The firm refinances itself. It repurchases half of
its stock with debt that it issues. Assume that
this debt is risk free. What is the equity beta of
the levered firm?
3. According to the CAPM, what rate of return
does the firm have to offer to its creditors?
4. According to the CAPM, what rate of return
does the firm have to offer to its levered equity
holders?
5. Has the firms weighted average cost of capital
changed?
Q 9.37. Consider the following historical rate of return series:
Year
IBM
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
0.175
0.400
0.156
0.322
0.257
0.676
0.393
0.775
0.175
0.208
S&P 500
Year
IBM
0.263
0.045
0.071
0.015
0.341
0.203
0.310
0.267
0.195
0.101
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
0.430
0.355
0.205
0.072
0.158
0.198
0.129
0.208
0.586
0.143
S&P 500
0.130
0.234
0.264
0.090
0.030
0.136
0.035
0.385
0.235
0.128
Assume that IBM had so little debt that it was practically risk-free.
1. What was IBMs equity beta over this sample
period?
2. If IBM had a debt-equity ratio of 70%, what was
its asset beta? (Hint: To determine a D/A ratio,
make up an example in which a firm has a 70%
D/E ratio.)
3. How important is the 1992 observation to your
beta estimate?
4. If HP is similar to IBM in its business but has a

debt-equity ratio of 10%, what would you expect HPs levered equity beta to be? (Hint: Use
the same leverage conversion trick.)
FINANCE today, and
Q 9.38. Look up betas on
compare them to those in Exhibit 8.6 on Page 209.
1. How does the beta of Intel today compare to its

earlier estimate from May 2008? Was its beta
stable (over time)?
2. How does the beta of AMD today compare to its
earlier estimate from May 2008? Was its beta
stable?
3. AMD is a much smaller firm than Intel. How do
their betas compare?
Q 9.39. A comparable firm (in a comparable business) has an equity beta of 2.5 and a debt-equity ratio
of 2. The debt is almost risk free. Estimate the beta
for your equity if projects have constant betas, but
your firm will carry a debt-equity ratio of 1/2. (Hint:
To translate a debt-equity ratio into a debt-asset ratio,
make up an example.)
Q 9.40. A Fortune 100 firm is financed with $15 billion in debt and $5 billion in equity. Its historical
equity beta has been 2. If the firm were to increase its
leverage from $15 billion to $18 billion and use the
cash to repurchase shares, what would you expect its
levered equity beta to be?
Q 9.41. The prevailing risk-free rate is 5% per annum. A competitor to your own firm, though publicly
traded, has been using an overall project cost of capital of 12% per annum. The competitor is financed
by 1/3 debt and 2/3 equity. This firm has had an estimated equity beta of 1.5. What is it using as its equity
premium estimate?
253
Q 9.42. Apply the CAPM. Assume the risk-free rate of

return is the current yield on 5-year bonds. Assume
that the markets expected rate of return is 3% per
year above this. Download 5 years of daily rate of
return data on four funds: NAESX, VLACX, VUVLX,
and VWUSX.
What were the historical average rates of return?
What were the historical market betas?
What were the historical market betas, adjusted
(shrunk) toward 1 by averaging with 1?
How do these estimates compare to the market beta estimates of the financial website from
which you downloaded the data?
Does it appear as if these funds followed a
CAPM-like relationship?
Q 9.43. Draw some possible security markets relations that would not be consistent with the CAPM.
The x axis would be the true market beta, the y axis
would be the true expected rate of return.
Q 9.44. Does the empirical evidence suggest that the
CAPM is correct?
Q 9.45. Why do you need to understand the CAPM?
Q 9.46. Under what circumstances is the CAPM a
good model to use? What are the main arguments in
favor of using it? When is it not a good model?
Q 9.47. If you use the CAPM, explain for what kinds
of projects it is important to get accurate equitypremium estimates.

CAPM

Uploaded by

CAPM

Uploaded by

9

The Capital Asset Pricing Model

What You Already Know and What You Want to Know

You are still after an

This allows you to

This gives you a

The Capital Asset Pricing Model

Using The Capital Asset Pricing Model (CAPM)

9.2. Using The Capital Asset Pricing Model (CAPM)

1. The risk-free rate of return, rF

3. A firms or projects beta with respect to the market, i

A first quick use of

The CAPM formula

The CAPM has three

Will history repeat

The Capital Asset Pricing Model

The Security Market Line (SML)

The SML is just a

If you know the

9.2. Using The Capital Asset Pricing Model (CAPM)

Exp RoR E(ri), in %

The intercept is the

Project beta with respect to M (i)

The Capital Asset Pricing Model

Known MarketBeta (i)

Historical MarketBeta (i)

9.3. The CAPM Cost of Capital in the Present Value Formula

The CAPM Cost of Capital in the Present Value Formula

We usually use the

Deconstructing Quoted Rates of Return Risk Premiums

The CAPM gives you

= Time Premium + Default Realization + Risk Premium

Expected Rate of Return = Time Premium + Expected Risk Premium

The Capital Asset Pricing Model

= Promised Rate of Return

Promised Interest Rate = Time Premium + Default Premium + Risk Premium

9.4. Estimating the CAPM Inputs

Estimating the CAPM Inputs

The Risk-Free Rate and Multi-Year Term-Structure Considerations

Which risk-free rate?

Advice: Pick the

But dont we need

The Capital Asset Pricing Model

The Equity Premium

You must provide the

9.4. Estimating the CAPM Inputs

Should I just give it to

Lets show you how

The Capital Asset Pricing Model

PS: 30-Year Treasuries had

Geometric Equity Premium 1970-2012 vs. Long-Term Bonds

9.4. Estimating the CAPM Inputs

CAPM Term Structure,

The Capital Asset Pricing Model

Yet another problem:

If your estimate of the forward-looking equity premium is based on the historical

9.4. Estimating the CAPM Inputs

Current predictive ratios

The Capital Asset Pricing Model

Arithmetic Average (Difference):

9.4. Estimating the CAPM Inputs

Internal Cost of Capital (ICC)

Method 6: Ask and

The equity premium is

The Capital Asset Pricing Model

Was the 20th Century Really the American Century?

The CAPM is about

9.4. Estimating the CAPM Inputs

Investment Projects Market Betas

Unlike the risk-free

The Implications of Beta for a Projects Risk and Reward

Projects with higher

The Capital Asset Pricing Model

49 Industries' Equity MarketBeta 2010

49 Industries' Equity MarketBeta 2010

49 Industries' Equity MarketBeta 2006