See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/272157556

Modern Portfolio Theory, Capital Market Theory, and Asset Pricing Models

Chapter · November 2001

CITATIONS READS
2 40,985

2 authors, including:

James L Grant
University of Massachusetts Boston
39 PUBLICATIONS   441 CITATIONS   

SEE PROFILE

All content following this page was uploaded by James L Grant on 11 February 2015.

The user has requested enhancement of the downloaded file.

 Equity Portfolio Management, Frank J. Fabozzi and James L. Grant,
November 2001, Wiley:
http://www.wiley.com/WileyCDA/WileyTitle/productCd-1883249406.html

Chapter 2

Modern Portfolio Theory,

Capital Market Theory, and
Asset Pricing Models
In this chapter, we set forth theories that are the underpinnings for the management of
portfolios: modern portfolio theory and capital market theory. Modern portfolio the-
ory deals with the selection of portfolios that maximize expected returns consistent
with individually acceptable levels of risk. Using quantitative models and historical
data, modern portfolio theory defines “expected portfolio returns” and “acceptable
levels of portfolio risk,” and shows how to construct an “optimal portfolio.”
Capital market theory deals with the effects of investor decisions on secu-
rity prices. More specifically, it shows the relationship that should exist between
security returns and risk if investors constructed portfolios as indicated by modern
portfolio theory. Together, modern portfolio theory and capital market theory provide
a framework to specify and measure investment risk and to develop relationships
between expected security return and risk (and hence between risk and required
return on an investment). These relationships are called asset pricing models.
Modern portfolio theory and capital market theory have revolutionized the
world of investment management by allowing managers to quantify the investment
risk and expected return of a portfolio. Moreover, these theories tell us that the focus
of portfolio management should be the risk of the entire portfolio, not the risk of the
individual assets. That is, it is possible to combine risky assets and produce a portfo-
lio whose expected return reflects its components, but with considerably lower risk.
Our purpose in this chapter is to present the basic principles and implica-
tions. We do not attempt to provide a rigorous derivation of the relationships pre-
sented. In Chapter 4, we discuss the implications of these theories for selecting an
equity management strategy.

MODERN PORTFOLIO THEORY

Prior to modern portfolio theory, practitioners would often speak of risk and
return, but the failure to quantify these important measures made the goal of con-
structing an optimal portfolio highly subjective and provided no insight about the
return investors should expect. Moreover, portfolio managers would focus on the
risks of individual assets without understanding how combining them into a port-
folio can affect the portfolio’s risk.

11
12 Portfolio Theory and Asset Pricing

Some Basic Concepts

The theories presented in this chapter draw on concepts from two fields: financial
economic theory and probability and statistical theory. Chapter 3, provides a
review of some of the basic concepts of probability and statistics that we use in
this chapter. In this section we will describe several concepts from financial eco-
nomic theory that will be used. Many of the concepts have a more technical or
rigorous definition. Our goal here is to keep the explanation simple enough to
appreciate the importance and applicability of these concepts to the development
of modern portfolio theory and capital market theory.

Efficient Portfolios and an Optimal Portfolio

In constructing a portfolio, investors seek to maximize the expected return from
their investment given some level of risk they are willing to accept.1 Portfolios
that satisfy this requirement are called efficient portfolios . The concepts of
expected return and risk will be defined more specifically as we proceed in the
development of modern portfolio theory.
To construct an efficient portfolio, it is necessary to make some assump-
tion about how investors behave in making investment decisions. A reasonable
assumption is that investors are risk averse. A risk-averse investor is one who when
faced with two investments with the same expected return but two different risks
will prefer the one with the lower risk. Given a choice of efficient portfolios from
which an investor can select, an optimal portfolio is the one that is most preferred.

Utility Function and Indifference Curves

In economic theory, there are many situations where individuals and firms face
choices. The “theory of choice” describes the decision-making process with the
help of a concept called the utility function. A utility function is a mathematical
expression that assigns a value to all possible choices. The higher the value, the
greater the utility. Simply put, in portfolio theory the utility function expresses the
preferences of economic entities with respect to perceived risk and expected return.
A utility function can be expressed in graphical form by an indifference
curve. Exhibit 1 shows indifference curves labeled u1, u2, and u3. The horizontal
axis measures risk, and the vertical axis measures expected return. Each curve rep-
resents a set of portfolios with different combinations of risk and return. All the
points on a given indifference curve indicate combinations of risk and expected
return that will give the same level of utility to a given investor. For example, on
utility curve u 1 , there are two points u and u ', with u having a higher expected
return than u', but also having a higher risk. An investor has an equal preference
for (or is indifferent to) any point on the curve, because the curve reflects the
investor’s level of risk aversion. The slope of an indifference curve reflects the fact
that investors require a higher expected return in order to accept higher risk.
1 Alternatively stated, investors seek to minimize the risk that they are exposed to given some target
expected return.
 Chapter 2 13

Exhibit 1: Indifference Curves

For the three indifference curves shown in Exhibit 1, the utility the inves-
tor receives is greater the further the indifference curve is from the horizontal
axis, because that curve represents a higher level of return at every level of risk.
Thus, for the three indifference curves shown in the figure, u3 has the highest util-
ity and u1 the lowest.

Risky Assets versus Risk-Free Assets

A risky asset is one for which the return that will be realized in the future is
uncertain. For example, suppose an investor purchases the stock of Merck today
and plans to hold the stock for one year. At the time the investor purchases the
stock, he or she does not know what return will be realized. The return will
depend on the price of Merck stock one year from now and the dividends that the
company pays during the year. Thus, the pharmaceutical stock, and indeed the
stock of all companies, is a risky asset.
There are assets, however, for which the return that will be realized in the
future is known with certainty today. Such assets are referred to as risk-free or
riskless assets. The risk-free asset is commonly defined as short-term obligations
of the U.S. government. For example, if an investor buys a U.S. government secu-
rity that matures in one year and plans to hold that security for one year, then
there is no uncertainty about the return that will be realized. The investor knows
that in one year, the maturity date of the security, the government will pay a spe-
cific amount to retire the debt.
14 Portfolio Theory and Asset Pricing

Measuring a Portfolio’s Expected Return

Investors are most often faced with choices among risky assets. Here we will look
at how to measure the expected return of a risky asset and the expected return of a
portfolio of risky assets.

Measuring Single-Period Portfolio Return

The actual return on a portfolio of assets over some specific time period is
straightforward to calculate, as shown below:
Rp = w1R1 + w2R2 + ... + wG RG (1)
where
Rp = rate of return on the portfolio over the period
Rg = rate of return on asset g over the period
wg = weight of asset g in the portfolio (i.e., asset g as a proportion of
the market value of the total portfolio)
G = number of assets in the portfolio
Equation (1) states that the return on a portfolio of G assets (Rp) is equal to the
sum of the individual asset weights in the portfolio times its return, for each asset g. The
portfolio return Rp is sometimes called the holding period return, or the ex post return.
For example, consider the following portfolio consisting of three assets:
Asset Market value Rate of return
1 $6 million 12%
2 $8 million 10%
3 $11 million 5%

The portfolio’s total market value is $25 million. Therefore,

R1 = 12% and w1 = $6 million/$25 million = 0.24 or 24%
R2 = 10% and w2 = $8 million/$25 million = 0.32 or 32%
R3 = 5% and w3 = $11 million/$25 million = 0.44 or 44%
and the portfolio’s return is then
Rp = 0.24 (12%) + 0.32 (10%) + 0.44 (5%) = 8.28%

The Expected Return of a Portfolio of Risky Assets

Equation (1) shows how to calculate the actual return of a portfolio over some spe-
cific time period. In portfolio management, the investor also wants to know the
expected (or anticipated) return from a portfolio of risky assets. The expected port-
folio return is the weighted average of the expected return of each asset in the port-
folio. The weight assigned to the expected return of each asset is the percentage of
the market value of the asset to the total market value of the portfolio. That is,
E(Rp) = w1 E(R1) + w2 E(R2) + ... + wG E(RG) (2)
 Chapter 2 15

The E( ) signifies expectations, and E(RP) is sometimes called ex ante return, or

the expected portfolio return over some specific time period.
The expected return on a risky asset is calculated as follows. First, a
probability distribution for the possible rates of return that can be realized must
be specified. A probability distribution is a function that associates the probability
of occurrence to a possible outcome for a random variable. Given the probability
distribution, the expected value of a random variable is simply the weighted aver-
age of the possible outcomes, where the weight is the probability associated with
the possible outcome. Rather than use the term “expected value of the return of an
asset,” we simply use the term expected return . Mathematically, the expected
return of asset i is expressed as

E(Ri) = p1 r1 + p2 r2 + .... + pN rN (3)

where
rn = the n-th possible rate of return for asset i
pn = probability of attaining rate of return n for asset i
N = number of possible outcomes for the rate of return
In practice, the probability distribution is based on historical returns.
Assume that an individual is considering an investment, stock XYZ,
which has a probability distribution for the rate of return for some time period as
given below:
n Rate of return (%) Probability of occurrence
1 15 0.50
2 10 0.30
3 5 0.13
4 0 0.05
5 −5 0.02

Substituting into equation (3), we get

E(RXYZ) = 0.50 (15%) + 0.30(10%) + 0.13 (5%) + 0.05 (0%) + 0.02 (−5%)
= 11%

Thus, 11% is the expected value or mean of the probability distribution for the
rate of return on stock XYZ.

Measuring Portfolio Risk

Investors have used a variety of definitions of risk. Professor Harry Markowitz
changed how the investment community thought about risk by quantifying the
concept of risk. 2 He defined risk in terms of a well-known statistical measure

2Harry M. Markowitz, “Portfolio Selection,” Journal of Finance (March 1952), pp. 77-91, and Portfolio
Selection, Cowles Foundation Monograph 16 (New York: John Wiley & Sons, 1959).
16 Portfolio Theory and Asset Pricing

known as the variance. Specifically, Markowitz quantified risk as the variance

about an asset’s expected return.

Variance as a Measure of Risk

The variance of a random variable is a measure of the dispersion of the possible
outcomes around the expected value. In the case of an asset’s return, the variance
is a measure of the dispersion of the possible outcomes for the rate of return
around the expected return.
The equation for the variance of the return for asset i, denoted var(Ri), is

var(Ri) = p1[r1 − E(Ri)]2 + p2[r2 − E(Ri)]2 + ... + pN[rN − E(Ri)]2 (4)

Using the probability distribution of the return for stock XYZ we can
illustrate the calculation of the variance:

var(RXYZ) = 0.50[15% − 11%]2 + 0.30[10% − 11%]2 + 0.13[5% − 11%]2

+ 0.05[0% − 11%]2 + 0.02[−5% − 11%]2 = 24%

The variance associated with a distribution of returns measures the tightness

with which the distribution is clustered around the mean or expected return. Markowitz
argued that this tightness or variance is equivalent to the uncertainty or riskiness of the
investment. If an asset is riskless, it has a dispersion of zero around the expected return.
Since the variance is squared units, it is common to see the variance con-
verted to the standard deviation or square root of the variance:

SD ( R i ) = var ( R i )

For stock XYZ, then, the standard deviation is:

SD ( R XYZ ) = 24% = 4.9%

The two are conceptually equivalent; that is, the larger the variance or standard
deviation, the greater the investment risk.
There are two criticisms of the use of the variance as a measure of risk.
The first criticism is that since the variance measures the dispersion of an asset’s
return around its expected return, it considers the possibility of returns above the
expected return and below the expected return. Investors, however, do not view
possible returns above the expected return as an unfavorable outcome. In fact,
such outcomes are quite favorable. Because of this, some have argued that mea-
sures of risk should not consider the possible returns above the expected return.
Markowitz recognized this limitation and, in fact, suggested a measure of
downside risk — the risk of realizing an outcome below the expected return —
called the semi-variance. The semi-variance is similar to the variance except that
in the calculation no consideration is given to returns above the expected return.
However, because of the computational problems with using the semi-variance
 Chapter 2 17

and the limited resources available to him at the time, he compromised and used
the variance in developing portfolio theory.
Today, various measures of downside risk are currently being used by
practitioners. However, regardless of the measure used, the basic principles of
portfolio theory developed by Markowitz are applicable. That is, the choice of the
measure of risk may affect the calculation but doesn’t invalidate the theory.
The second criticism is that the variance is only one measure of how the
returns vary around the expected return. When a probability distribution is not
symmetrical around its expected return, then a statistical measure of the skewness
of a distribution should be used in addition to the variance. The variance can be
justified based on empirical evidence which suggests that the historical distribu-
tion of the returns on stocks is approximately symmetrical. 3 Because expected
return and variance are the only two parameters that investors are assumed to con-
sider in making investment decisions, the Markowitz formulation of portfolio the-
ory is often referred to as a two-parameter model.

Measuring the Portfolio Risk of a Two-Asset Portfolio

Equation (4) gives the variance for an individual asset’s return. The variance of a
portfolio consisting of two assets is a little more difficult to calculate. It depends
not only on the variance of the two assets, but also upon how closely one asset
tracks the other asset. The formula is
2 2
var(Rp) = w i var(Ri) + w j var(Rj) + 2 wi wj cov(Ri,Rj) (5)
where
cov(Ri,Rj) = covariance between the return for assets i and j
In words, equation (5) states that the variance of the portfolio return is the sum of
the weighted variances of the two assets plus the weighted covariance between
the two assets.
The covariance is a new term in this discussion and has a precise mathe-
matical translation. However, its practical meaning is the degree to which the
returns on two assets vary or change together. The covariance is not expressed in a
particular unit, such as dollars or percent. A positive covariance means the returns
on two assets tend to move or change in the same direction, while a negative cova-
riance means the returns move in opposite directions.
The covariance is analogous to the correlation between the returns for
two assets. Specifically, the correlation between the returns for assets i and j is
defined as the covariance of the two assets divided by the product of their stan-
dard deviations:
cov ( R i, R j )
cor ( R i, R j ) = -------------------------------------
SD ( R i )SD ( R j )

3 See Chapters 1 and 2 in Eugene Fama, Foundations of Finance (New York: Basic Books, 1976).
18 Portfolio Theory and Asset Pricing

The correlation and the covariance are conceptually equivalent terms.

Dividing the covariance by the product of the standard deviations simply (but
importantly) makes the correlation a number that is comparable across different
assets. The correlation coefficient can have values ranging from +1.0, denoting
perfect comovement in the same direction, to −1.0, denoting perfect comovement
in opposite directions.
In our illustration of how to calculate the expected return, variance, and
standard deviation, we used the probability distributions for the stock. As such,
they are truly “expected values,” since they were derived probabilistically. In
practice, the estimation of these statistical measures is typically obtained from
historical observations on the rate of returns. We show how this is done in the
next chapter.
Now let’s look at the implications of the variance of the portfolio as
given by equation (5). The equation tells us that since the variance of a portfolio
depends on the covariances of its constituent securities, a portfolio’s risk can be
low despite the fact that the risk of individual assets making up the portfolio can
be quite high. This principle has important implications for managing portfolios,
as we shall see below.

Measuring the Risk of a Portfolio with

More Than Two Assets
Thus far we have given the portfolio risk for a portfolio consisting of two assets.
The extension to three assets — i, j, and k — is as follows:
2 2 2
var(Rp) = w i var(Ri) + w j var(Rj) + w k var(Rk) + 2 wi wj cov(Ri,Rj)
+ 2 wi wk cov(Ri,Rk) + 2 wj wk cov(Rj,Rk) (6)

In words, equation (6) states that the variance of the portfolio return is
the sum of the weighted variances of the individual assets plus the sum of the
weighted covariances of the assets. Hence, the variance of the portfolio return is
the weighted sum of the individual variances of the assets in the portfolio plus the
weighted sum of the degree to which the assets vary together.
In general, for a portfolio with G assets, the portfolio variance is
G G G

∑ ∑ ∑ wg wh cov ( Rg, Rh )
2
var ( R p ) = w g var ( R g ) + (7)
g=1 g = 1 h=1
for h ≠ g

Portfolio Diversification
Often, one hears investors talking about “diversifying” their portfolio. By this an
investor means constructing a portfolio in such a way as to reduce portfolio risk
without sacrificing return. This is certainly a goal that investors should seek.
However, the question is how does one do this in practice.
 Chapter 2 19

Some investors would say that a portfolio can be diversified by including

assets across all asset classes. For example, one investor might argue that a port-
folio should be diversified by investing in stocks, bonds, and real estate. While
that might be reasonable, two questions must be addressed in order to construct a
diversified portfolio. First, how much should be invested in each asset class?
Should 40% of the portfolio be in stocks, 50% in bonds, and 10% in real estate, or
is some other allocation more appropriate? Second, given the allocation, which
specific stocks, bonds, and real estate should the investor select?
Some investors who focus only on one asset class such as common stock
argue that such portfolios should also be diversified. By this they mean that an
investor should not place all funds in the stock of one corporation, but rather
should include stocks of many corporations. Here, too, several questions must be
answered in order to construct a diversified portfolio. First, which corporations
should be represented in the portfolio? Second, how much of the portfolio should
be allocated to the stocks of each corporation?
Prior to the development of portfolio theory, while investors often talked
about diversification in these general terms, they never provided the analytical
tools by which to answer the questions posed above. A major contribution of port-
folio theory is that using the concepts discussed above, a quantitative measure of
the diversification of a portfolio is possible, and it is this measure that can be used
to achieve the maximum diversification benefits.

Markowitz Diversification
The Markowitz diversification strategy is primarily concerned with the degree of
covariance between asset returns in a portfolio. Indeed a key contribution of
Markowitz diversification is the formulation of an asset’s risk in terms of a portfo-
lio of assets, rather than in isolation. Markowitz diversification seeks to combine
assets in a portfolio with returns that are less than perfectly positively correlated,
in an effort to lower portfolio risk (variance) without sacrificing return. It is the
concern for maintaining return, while lowering risk through an analysis of the
covariance between asset returns, that separates Markowitz diversification from
the naive approaches mentioned above and makes it more effective.
Markowitz diversification and the importance of asset correlations can be
illustrated with a simple two-asset portfolio example. To do this, we will first
show the general relationship between the risk of a two-asset portfolio and the
correlation of returns of the component assets. Then we will look at the effects on
portfolio risk of combining assets with different correlations.

Portfolio Risk and Correlation In our two-asset portfolio, assume that com-
mon stock C and common stock D are available with expected returns and stan-
dard deviations as shown:
E(R) SD(R)
Stock C 10% 30%
Stock D 25% 60%
20 Portfolio Theory and Asset Pricing

If an equal weighting (50%) is assigned to both stocks C and D, the expected

portfolio return can be calculated as:
E(Rp) = 0.50(10%) + 0.50(25%) = 17.5%
The variance of the return on the two-stock portfolio is
2 2
var(Rp) = w C var(RC) + w D var(RD) + 2 wC wD cov(RC,RD)
= (0.5)2 (30%)2 + (0.5)2 (60%)2 + 2(0.5)(0.5)cov(RC,RD)

From the relationship between correlation and covariance we have

cov ( R C, R D )
cor ( R C, R D ) = ----------------------------------------
SD ( R C )SD ( R D )
so
cov(RC,RD) = SD(RC) SD(RD) cor(RC,RD)
Since SD(RC) = 30% and SD(RD) = 60%, then
cov(RC,RD) = (30%)(60%) cor(RC,RD)
Substituting into the expression for var(Rp), we get
var(Rp) = (0.5)2(30%)2 + (0.5)2(60%)2 + 2(0.5)(0.5)(30%)(60%) cor(RC,RD)
Taking the square root of the variance:
2 2 2 2
SD ( R p ) = ( 0.5 ) ( 30% ) + ( 0.5 ) ( 60% ) + 2 ( 0.5 )0.5 ( 30% ) ( 60% )cor ( R C, R D )

= 0.1125 + ( 0.09 )cor ( R C, R D )

The Effect of the Correlation of Asset Returns on Portfolio Risk How

would the risk change for our two-asset portfolio with different correlations between
the returns of the component stocks? Let’s consider the following three cases for
cor(RC,RD): −1.0, 0, and +1.0. Substituting into the equation for the standard devia-
tion of the portfolio as given above for these three cases of cor(RC,RD), we get
cor(RC,RD) E(Rp) SD(Rp)
+1.0 17.5% 45.0%
0 17.5% 33.5%
−1.0 17.5% 15.0%

As the correlation between the returns on stocks C and D decreases from

+1.0 to 0.0 to −1.0, the standard deviation of the portfolio return also decreases from
45% to 15%. However, the expected portfolio return remains 17.5% for each case.
This example clearly illustrates the effect of Markowitz diversification
which states that as the correlation (covariance) between the returns for assets
that are combined in a portfolio decreases, so does the variance (hence the stan-
 Chapter 2 21

dard deviation) of the return for that portfolio. This is due to the degree of corre-
lation between the asset returns. The good news is that investors can maintain
expected portfolio return and lower portfolio risk by combining assets with lower
(and preferably negative) correlations. However, the bad news is that very few
assets have small to negative correlations with other assets! The problem, then,
becomes one of searching among large numbers of assets in an effort to discover
the portfolio with the minimum risk at a given level of expected return or, equiva-
lently, the highest expected return at a given level of risk. The stage is now set for
a discussion of Markowitz efficient portfolios and their construction.

Choosing a Portfolio of Risky Assets

Diversification in the manner suggested by Professor Markowitz leads to the con-
struction of portfolios that have the highest expected return at a given level of
risk. Such portfolios are called Markowitz efficient portfolios or, simply, efficient
portfolios. In order to construct Markowitz efficient portfolios, the theory makes
some basic assumptions about asset selection behavior.
First, it assumes that the only two parameters that affect an investor’s
decision are the expected return and the variance. That is, investors make deci-
sions using the two-parameter model formulated by Markowitz. Second, it
assumes that investors are risk averse (i.e., when faced with two investments with
the same expected return but different risk levels, investors will prefer the one
with the lower risk). Third, it assumes that all investors seek to achieve the high-
est expected return at a given level of risk. Fourth, it assumes that all investors
have the same expectations regarding expected return, variance, and covariances
for all risky assets. This assumption is referred to as the homogeneous expecta-
tions assumption. Finally, it assumes that all investors have a common one-period
investment horizon.

Constructing Markowitz Efficient Portfolios

The technique of constructing Markowitz efficient portfolios from large groups of
stocks requires a massive number of calculations. In a portfolio of G securities,
there are (G2 − G)/2 unique covariances to calculate. Hence, for a portfolio of just
50 securities, there are 1,225 covariances that must be calculated. For 100 securi-
ties, there are 4,950. Furthermore, in order to solve for the portfolio that mini-
mizes risk for each level of return, a mathematical technique called quadratic
programming (and a computer) must be used.4 A discussion of this technique is
4 Alternative methods that require less computer time can be used by investors who want a useful approxi-
mation to the Markowitz efficient frontier. The basic idea behind these alternative methods is that an esti-
mated relationship between the return on a stock and some common stock market index or indexes can be
used in lieu of the variance and covariance of returns. One of the approaches suggested by Sharpe uses one
common stock index, and the approach is referred to as the single-index market model. (See William F.
Sharpe, “A Simplified Model for Portfolio Analysis,” Management Science (January 1963), pp. 277-293.)
Multiple indexes have been suggested in Kalman J. Cohen and Jerry A. Pogue, “An Empirical Evaluation of
Some Alternative Portfolio Selection Models,” Journal of Business (April 1967), pp. 166-193.
22 Portfolio Theory and Asset Pricing

beyond the scope of this chapter. However, it is possible to illustrate the general
idea of the construction of Markowitz efficient portfolios by referring again to the
simple two-asset portfolio consisting of stocks C and D.
Recall that for two assets, common stocks C and D, E(RC) = 10%,
SD(R C) = 30%, E(R D) = 25%, and SD(R D) = 60%. We now further assume that
cor(RC,RD) = −0.5. The expected portfolio return and standard deviation are cal-
culated for five different proportions of C and D in the portfolio in Exhibit 2.
Given these available combinations of stocks C and D, it is now possible to intro-
duce the notion of a feasible portfolio and a Markowitz efficient portfolio.

Feasible and Efficient Portfolios

A feasible portfolio is a portfolio that an investor can construct given the assets
available. The collection of all feasible portfolios is called the feasible set of port-
folios. With only two assets, the feasible set of portfolios is graphed as a curve
that represents those combinations of risk and expected return that are attainable
by constructing portfolios from the available combinations of the two assets. In
Exhibit 3, the feasible set of portfolios is defined by the combinations of stocks C
and D producing the E(Rp) and SD(Rp) given in Exhibit 2, and is represented by
the curve 1–5. If combinations of more than two assets were being considered, the
feasible set is no longer the curved line. It would be approximated by the shaded
area in Exhibit 4.
In contrast to a feasible portfolio, a Markowitz efficient portfolio is one
that gives the highest expected return of all feasible portfolios with the same risk.
A Markowitz efficient portfolio is also said to be a mean-variance efficient portfo-
lio. Thus, for each level of risk there is a Markowitz efficient portfolio. The col-
lection of all efficient portfolios is called the Markowitz efficient set of portfolios.

Exhibit 2: Portfolio Expected Returns and Standard

Deviations for Varying Proportions of Stocks C and D
Data for stock C: E(RC) = 10% SD(RC) = 30%
Data for stock D: E(RD) = 25% SD(RC) = 60%
Correlation between stock C and stock D = −0.5

Proportion of Proportion of
Portfolio stock C (wC) stock D (wD) E(Rp) SD(Rp)
1 100% 0% 10.0% 30.0%
2 75 25 13.8 19.8
3 50 50 17.5 26.0
4 25 75 21.3 41.8
5 0 100 25.0 60.0
 Chapter 2 23

Exhibit 3: Feasible and Efficient Sets of Portfolios for

Stocks C and D

Exhibit 4: Feasible and Efficient Sets of Portfolios When

There Are More Than Two Assets

This can be seen graphically in Exhibit 3. Efficient combinations of stocks

C and D lie on the curve section 2–5 in the exhibit. These Markowitz efficient com-
binations of stocks C and D offer the highest expected return at a given level of
risk. Notice that portfolio 1 [with E(Rp) = 10% and SD(Rp) = 30%] is not included
in the Markowitz efficient set, since there are two portfolios, for example, in the
24 Portfolio Theory and Asset Pricing

Markowitz efficient set (portfolios 2 and 3) that have higher expected returns and
lower risk levels. Portfolios to the left of section 2-3-4-5 are not attainable from
combinations of stocks C and D and are, therefore, not candidates for the Markow-
itz efficient set. Portfolios to the right of section 2-3-4-5 are not included in the
Markowitz efficient set, since there exists some other portfolio that would provide
a higher expected return at the same level of risk or, alternatively, a lower level of
risk at the same expected return. To see this, consider portfolio 6 in Exhibit 3. Port-
folios 4 and 6 have the same level of risk, but portfolio 4 has a higher expected
return. Likewise, portfolios 2 and 6 have the same expected returns, but portfolio 2
has a lower level of risk. Thus, portfolios 4 and 2 are said to dominate portfolio 6.
Exhibit 4 also shows the Markowitz efficient set. All the portfolios on the
Markowitz efficient set dominate the portfolios in the shaded area.
The Markowitz efficient set of portfolios is sometimes called the
Markowitz efficient frontier, because graphically all the Markowitz efficient port-
folios lie on the boundary of the set of feasible portfolios that have the maximum
return for a given level of risk. Any portfolios above the Markowitz efficient fron-
tier cannot be achieved. Any portfolios below the Markowitz efficient frontier are
dominated by portfolios on the Markowitz efficient frontier.

Choosing a Portfolio in the Markowitz Efficient Set

Now that we have constructed the Markowitz efficient set of portfolios, the next
step is to determine the optimal portfolio.
An investor will want to hold one of the portfolios on the Markowitz effi-
cient frontier. Notice that the portfolios on the Markowitz efficient frontier repre-
sent trade-offs in terms of risk and return. Moving from left to right on the
Markowitz efficient frontier, the expected risk increases, but so does the expected
return. The question is, which is the best portfolio to hold? The best portfolio to
hold of all those on the Markowitz efficient frontier is the optimal portfolio.
Intuitively, the optimal portfolio should depend on the investor’s prefer-
ence or utility as to the trade-off between risk and return. As explained at the begin-
ning of this chapter, this preference can be expressed in terms of a utility function.
In Exhibit 5, three indifference curves and the efficient frontier are drawn
on the same diagram. In our application, the indifference curve indicates the com-
binations of risk and expected return that give the same level of utility. Moreover,
the farther from the horizontal axis the indifference curve, the higher the utility.
From Exhibit 5, it is possible to determine the optimal portfolio for the
investor with the indifference curves shown. Remember that the investor wants to
get to the highest indifference curve achievable given the Markowitz efficient fron-
tier. Given that requirement, the optimal portfolio is represented by the point where
an indifference curve is tangent to the Markowitz efficient frontier. In Exhibit 5,
* *
that is the portfolio P MEF . For example, suppose that P MEF corresponds to portfo-
lio 3 in Exhibit 3. From Exhibit 2 we see that this portfolio is a combination of
50% of stock C and 50% of stock D, with E(Rp) = 17.5% and SD(Rp) = 26%.
 Chapter 2 25

Exhibit 5: Selection of the Optimal Portfolio

Consequently, for the investor’s preferences for risk and return as deter-
mined by the shape of the indifference curves, and his or her expectations for
returns and covariance of stocks C and D, portfolio 3 maximizes utility. If this
investor had a different preference for expected risk and return, there would have
been a different optimal portfolio.
At this point in our discussion, a natural question is how to estimate an
investor’s utility function so that the indifference curves can be determined.
Unfortunately, there is little guidance about how to construct one. In general,
economists have not been successful in measuring utility functions.
The inability to measure utility functions does not mean that the theory is
flawed. What it does mean is that once an investor constructs the Markowitz effi-
cient frontier, the investor will subjectively determine which Markowitz efficient
portfolio is appropriate given his or her tolerance to risk.

CAPITAL MARKET THEORY AND ASSET PRICING MODELS

Having introduced the principles of modern portfolio theory, we will now

describe capital market theory and the implications of both that theory and mod-
ern portfolio theory for the pricing of financial assets. We first focus on one well-
known asset pricing model called the capital asset pricing model (CAPM) and
then another popular asset pricing model. The asset pricing models we describe in
this chapter are equilibrium models. That is, given assumptions about the behav-
26 Portfolio Theory and Asset Pricing

ior and expectations of investors, and assumptions about capital markets, these
models predict the theoretical equilibrium price of an asset.

Assumptions
Just like the model for the selection of Markowitz efficient portfolios, capital mar-
ket theory and asset pricing models are abstractions of the real world and, as such,
are based upon some simplifying assumptions. Some of these assumptions may
even seem unrealistic. However, these assumptions make the theory more tracta-
ble from a mathematical standpoint. The assumptions to derive the CAPM are that
(1) investors rely on two factors in making their decisions: expected return and
variance; (2) investors are rational and risk averse and subscribe to Markowitz
methods of portfolio diversification; (3) investors all invest for the same period of
time; (4) investors share all expectations about assets (the “homogeneous expec-
tations assumption”); (5) there is a risk-free investment and investors can borrow
and lend any amount at the risk-free rate; and (6) capital markets are completely
competitive and frictionless.
The theory assumes all investors make investment decisions over some
single-period investment horizon. How long that period is (i.e., six months, one
year, two years, etc.) is not specified. In reality, the investment decision process is
more complex than that, with many investors having more than one investment
horizon — such as short-term and long-term planning horizons. Nonetheless, the
assumption of a one-period investment horizon is necessary to simplify the math-
ematics of the theory.
The first four assumptions deal with the behavior of investors in making
investment decisions. It is also necessary to make assumptions about the character-
istics of the capital market in which investors transact. These are covered by the
last two assumptions.

Capital Market Theory

Earlier in this chapter we distinguished between a risky asset and a risk-free asset. We
described how to create Markowitz efficient portfolios from risky assets. We did not
consider the possibility of constructing Markowitz efficient portfolios in the presence
of a risk-free asset, that is, an asset where the return is known with certainty.
In the absence of a risk-free rate, portfolio theory tells us that Markowitz
efficient portfolios can be constructed based on expected return and variance and
that the optimal portfolio is the one that is tangent to the investor’s indifference
curve. Once a risk-free asset is introduced and assuming that investors can borrow
and lend at the risk-free rate (assumption 5, above), the conclusion of Markowitz
portfolio theory can be qualified as illustrated in Exhibit 6. Every combination of
the risk-free asset and the Markowitz efficient portfolio M is shown on the capital
market line (CML). The line is drawn from the vertical axis at the risk-free rate
tangent to the Markowitz efficient frontier. The point of tangency is denoted by
M. All the portfolios on the capital market line are feasible for the investor to con-
 Chapter 2 27

struct. Portfolios to the left of M represent combinations of risky assets and the
risk-free asset. Portfolios to the right of M include purchases of risky assets made
with funds borrowed at the risk-free rate. Such a portfolio is called a levered port-
folio since it involves the use of borrowed funds.
Now compare a portfolio on the capital market line to the portfolio on the
Markowitz efficient frontier with the same risk. For example, compare portfolio
PA, which is on the Markowitz efficient frontier, with portfolio PB, which is on the
capital market line and therefore some combination of the risk-free asset and the
Markowitz efficient portfolio M. Notice that for the same risk the expected return
is greater for PB than for PA. A risk-averse investor will prefer PB to PA. That is,
PB will dominate PA. In fact, this is true for all but one portfolio on the line: port-
folio M, which is on the Markowitz efficient frontier.
Recognizing this, we must modify the conclusion from portfolio theory
that an investor will select a portfolio on the Markowitz efficient frontier, the par-
ticular portfolio depending on the investor’s risk preference. With the introduction
of the risk-free asset, we can now say that an investor will select a portfolio on the
line representing a combination of borrowing or lending at the risk-free rate and
purchasing the Markowitz efficient portfolio M.
The particular efficient portfolio that the investor will select on the line
will depend on the investor’s risk preference. The investor will select the portfolio
on the line that is tangent to the highest indifference curve. In the absence of a
risk-free asset, it would not be possible to construct such a portfolio.

Exhibit 6: The Capital Market Line

28 Portfolio Theory and Asset Pricing

Several researchers demonstrated that the opportunity to borrow or lend at

the risk-free rate implies a capital market where risk-averse investors will prefer to
hold portfolios consisting of combinations of the risk-free asset and some portfolio
M on the Markowitz efficient frontier.5 William Sharpe, one of the developers of the
theory, called the line from the risk-free rate to portfolio M on the efficient frontier
the capital market line, and this is the name that has been adopted in the industry.
One more key question remains: How does an investor construct portfo-
lio M? Eugene Fama answered this question by demonstrating that M must consist
of all assets available to investors, and each asset must be held in proportion to its
market value relative to the total market value of all assets.6 So, for example, if
the total market value of some asset is $200 million and the total market value of
all assets is $X, then the percentage of the portfolio that should be allocated to
that asset is $200 million divided by $X. Because portfolio M consists of all
assets, it is referred to as the market portfolio.
Now we can restate how a risk-averse investor who makes investment
decisions as suggested by Markowitz and who can borrow and lend at the risk-
free rate should construct efficient portfolios. This should be done by combining
an investment in the risk-free asset and the market portfolio. While all investors
will select a portfolio on the capital market line, the optimal portfolio for a spe-
cific investor is the one that will maximize that investor’s utility function.
The capital market line can be derived algebraically. We will dispense with
the proof here and simply indicate the formula for the capital market line below:
E ( R M ) – RF
E ( R p ) = R F + ------------------------------ SD ( R p ) (8)
SD ( R M )
where
E(RM) = expected return for the market
RF = risk-free return
SD(RM) = standard deviation of the market return
SD(RP) = standard deviation of the portfolio
Under the assumptions stated earlier, equation (8) is a straight line repre-
senting the efficient set for all risk-averse investors.
As we mentioned, both capital market theory and the derivation of the
Markowitz efficient frontier assume that all investors have the same expectations
for the inputs into the model. With homogeneous expectations, SD(R M ) and

5 William F. Sharpe, “Capital Asset Prices,” Journal of Finance (September 1964), pp. 425- 442; John Lint-

ner, “The Valuation of Risk Assets and the Selection of Risky Investments in Stock Portfolio and Capital
Budgets,” Review of Economics and Statistics (February 1965), pp. 13-37; Jack L. Treynor, “Toward a The-
ory of Market Value of Risky Asset,” Unpublished Paper, Arthur D. Little, 1961; and, Jan Mossin, “Equilib-
rium in a Capital Asset Market,” Econometrica (October 1966), pp. 768- 783.
6 Eugene F. Fama, “Efficient Capital Markets: A Review of Theory and Empirical Work,” Journal of

Finance (May 1970), pp. 383- 417.

 Chapter 2 29

SD(Rp) are the market’s consensus for the return distributions for the market port-
folio and portfolio p. The slope of the CML is
E ( R M ) – RF
------------------------------
SD ( R M )
Let’s examine the economic meaning of the slope. The numerator is the
expected return of the market in excess of the risk-free return. It is a measure of
the risk premium, or the reward for holding the risky market portfolio rather than
the risk-free asset. The denominator is the risk of the market portfolio. Thus, the
slope measures the reward per unit of market risk. Since the CML represents the
return offered to compensate for a perceived level of risk, each point on the line is
a balanced market condition, or equilibrium. The slope of the line determines the
additional return needed to compensate for a unit change in risk. That is why the
slope of the CML is also referred to as the equilibrium market price of risk.
The CML says that the expected return on a portfolio is equal to the risk-
free rate plus a risk premium equal to the price of risk (as measured by the differ-
ence between the expected return on the market and the risk-free rate divided by
the standard deviation of the market return) times the quantity of market risk for
the portfolio (as measured by the standard deviation of the portfolio). That is,
E(Rp) = RF + Market price of risk × Quantity of market risk

Capital Asset Pricing Model

Up to this point, we know how a risk-averse investor who makes decisions based
on two parameters (expected return and variance) should construct an efficient
portfolio: using a combination of the market portfolio and the risk-free rate.
Based on this result, a model can be derived that shows how a risky asset should
be priced. In the process of doing so, we can fine-tune our thinking about the risk
associated with an asset. Specifically, we can show that the appropriate risk that
investors should be compensated for accepting is not the variance of an asset’s
return but some other quantity. In order to do this, let’s take a closer look at the risk.

Systematic and Unsystematic Risk

In the development of portfolio theory, Professor Markowitz defined the variance of
the portfolio rate of return as the appropriate measure of risk. This risk measure can
be divided into two general types of risk: systematic risk and unsystematic risk.
Professor William Sharpe defined systematic risk as the portion of an
asset’s variability that can be attributed to a common factor.7 It is also sometimes
called undiversifiable risk or market risk. Systematic risk is the minimum level of
risk that can be obtained for a portfolio by means of diversification across a large
number of randomly chosen assets. As such, systematic risk is that which results
from general market and economic conditions that cannot be diversified away.
7 William F. Sharpe, “A Simplified Model for Portfolio Analysis,” Management Science (January 1963), pp.

277-293.
30 Portfolio Theory and Asset Pricing

Exhibit 7: Systematic and Unsystematic Portfolio Risk

Sharpe defined the portion of an asset’s variability that can be diversified

away as unsystematic risk. It is also sometimes called diversifiable risk, unique
risk, residual risk, or company-specific risk. This is the risk that is unique to a
company, such as a strike, the outcome of unfavorable litigation, or a natural catas-
trophe. As examples of this type of risk, one need only recall the case of product
tampering involving Tylenol capsules (manufactured by Johnson & Johnson, Inc.)
in October 1982 or the chemical accident at the Union Carbide plant in Bhopal,
India, in December 1984. Both of these unforecastable and hence unexpected trag-
edies had negative impacts on the stock prices of the two companies involved.
How diversification reduces unsystematic risk for portfolios can be illus-
trated with a graph. Exhibit 7 shows that at a portfolio size of about 20 randomly
selected stocks, the level of unsystematic risk is almost completely diversified
away.8 Essentially, all that is left is systematic, or market, risk.
Therefore the total risk of an asset can be measured by its variance. However,
the total risk can be divided into its systematic and unsystematic risk components.
Next we will show how this can be done so as to be able to quantify both components.
Market Model
The market model or single-index market model developed by Sharpe can be used
to isolate the systematic and unsystematic components of an asset’s return.9 The
linear relationship can be expressed as follows:
8 Empirical evidence for Exhibit 7 is provided in Wayne H. Wagner and Sheila Lau, “The Effect of Diversi-

fication on Risks,” Financial Analysts Journal (November-December 1971), p. 50.

9 Sharpe, “A Simplified Model for Portfolio Analysis.”
 Chapter 2 31

Rit = αi + βi RMt + eit (9)

where
Rit = return on asset i over the period t
RMt = return on the market portfolio over the period t
αi = a term that represents the nonmarket component of the return on
asset i
βi = a term that relates the change in asset i’s return to the change in
the market portfolio
eit = random error term that reflects the unique risks associated with
investing in an asset over period t
The market model given by equation (9) says that the return on an asset
depends on the return on the market portfolio and the extent of the asset’s respon-
siveness as measured by beta (βi). In addition, the return will also depend on con-
ditions that are unique to the firm as measured by eit.
The term β, or beta, is the slope of the market model for the asset, and
measures the degree to which the historical returns on the asset change systemati-
cally with changes in the market portfolio’s return. Hence, beta is referred to as an
index of that systematic risk due to general market conditions that cannot be
diversified away. For example, if a stock has a beta of 1.5, it means that, on aver-
age, on the basis of historical data, the stock had a return equal to 1.5 times that of
the market portfolio’s return. The beta for the market portfolio is 1.0.
The term in the market model, popularly referred to as alpha is equal to
the average value over time of the unsystematic returns for the stock. For most
stocks, alpha tends to be small and unstable.
Recall from our earlier discussion, the total risk of an asset can be
decomposed into market or systematic risk and unique or unsystematic risk. We
can use the market model to quantify these two risks. To see how, let’s look at the
total risk of the return of stock i as measured by the variance of its return. This is
done by determining the variance of equation (9). We show without proof that the
variance would be
2
var(Ri) = β i var(RM) + var(ei) (10)

Equation (10) says that the total risk as measured by var(Ri) is equal to the sum of
2
(1) the market or systematic risk as measured by β i var(RM), and (2) the unique
risk as measured by var(ei)
In Chapter 3 we explain how the market model is estimated by applying
statistical techniques to historical data on returns. Another product of the statisti-
cal technique used to estimate beta is the percentage of systematic risk to total
risk. In statistical terms, it is measured by the coefficient of determination (R 2)
from the regression, which indicates the percentage of the variation in the return
of the asset explained by the market portfolio return. The value of the coefficient
32 Portfolio Theory and Asset Pricing

ranges from 0 to 1. For example, a coefficient of determination of 0.3 for an asset

means that 30% of the variation in the return of that asset is explained by the
return of the market portfolio. Unsystematic or unique risk is then the amount not
explained by the market portfolio’s return. That is, it is 1 minus the coefficient of
determination (1 − R2).
Studies have shown that for the average New York Stock Exchange
(NYSE) common stock, systematic risk is about 30% of return variance, while
unsystematic risk is about 70%. In contrast, the coefficient of determination for a
well-diversified portfolio of stocks will typically exceed 90%, indicating that
unsystematic risk is less than 10% of total portfolio return variance. This supports
the point made in Exhibit 7 that with a well-diversified portfolio, most of the port-
folio risk is systematic risk.

The Security Market Line

The capital market line represents an equilibrium condition in which the expected
return on a portfolio of assets is a linear function of the expected return on the
market portfolio. A directly analogous relationship holds for individual security
expected returns:
[ E ( R M ) – RF ]
E ( R i ) = R F + ----------------------------------- SD ( R i ) (11)
SD ( R M )

Equation (11) simply uses risk and return variables for an individual security in
place of the portfolio values in the formula for the CML in equation (8). This ver-
sion of the risk- return relationship for individual securities is called the security
market line (SML). As in the case of the CML, the expected return for an asset is
equal to the risk-free rate plus the product of the market price of risk and the
quantity of risk associated with the security.
Another more common version of the SML relationship uses the beta of a
security. To see how this relationship is developed, look back at equation (10). In
a well-diversified portfolio (i.e., Markowitz diversified), the unique risk is elimi-
2
nated. Consequently, equation (10) can be rewritten as var(Ri) = β i var(RM) and
the standard deviation as SD(Ri) = βi SD(RM).
If βi SD(RM) is substituted into equation (11) for SD(Ri), we have the beta
version of the SML or the capital asset pricing model as shown in equation (12):

E(Ri) = RF + βi [E(RM) − RF] (12)

Equation (12) states that, given the assumptions of the CAPM, the expected
(or required) return on an individual asset is a positive linear function of its index of
systematic risk as measured by beta. The higher the beta, the higher the expected
return. Notice that it is only an asset’s beta that determines its (differential) expected
return.
 Chapter 2 33

Exhibit 8: The Security Market Line

Let’s look at the prediction of the CAPM for several values of beta. The
beta of a risk-free asset is zero, because the variability of the return for a risk-free
asset is zero and therefore it does not covary with the market portfolio. So if we
want to know the expected return for a risk-free asset, we would substitute zero for
βi in equation (12). Thus, the return on a risk-free asset is simply the risk-free
return. Of course, this is what we expect. The beta of the market portfolio is 1. If
asset i has the same beta as the market portfolio, then substituting 1 for βi in equa-
tion (12) gives a value equal to E(RM). In this case, the expected return for the asset
is the same as the expected return for the market portfolio. If an asset has a beta
greater than the market portfolio (i.e., greater than 1), then the expected return will
be higher than for the market portfolio. The reverse is true if an asset has a beta less
than the market portfolio. A graph of the SML is presented in Exhibit 8.
In equilibrium, the expected return of individual securities will lie on the SML
and not on the CML. This is true because of the high degree of unsystematic risk that
remains in individual securities that can be diversified out of portfolios of securities.
It follows that the only risk that investors will pay a premium to avoid is
market risk. Hence, two assets with the same amount of systematic risk will have
the same expected return. In equilibrium, only efficient portfolios will lie on both
the CML and the SML. This underscores the fact that the systematic risk measure,
beta, is most correctly considered as an index of the contribution of an individual
security to the systematic risk of a well-diversified portfolio of securities.
There is one more version of the SML that is worthwhile to discuss. It
can be demonstrated that SML can be expressed as:
Cov ( R i, R M )
E ( R i ) = R F + ------------------------------- [ E ( R M ) – R F ] (13)
var ( R M )
34 Portfolio Theory and Asset Pricing

This version of the SML emphasizes that it is not the variance or stan-
dard deviation of an asset that affects its return. It is the covariance of the asset’s
return with the market’s return that affects its return. An asset that has a positive
covariance will have a higher expected return than the risk-free asset; an asset
with a negative covariance will have a lower expected return than the risk-free
asset. The reason has to do with the benefits of diversification we discussed earlier
in this chapter. If the covariance is positive, this increases the risk of an asset in a
portfolio and therefore investors will only purchase that asset if they expect to
earn a return higher than the risk-free asset. If an asset has a negative covariance,
recall from our discussion earlier in this chapter that this will reduce the portfolio
risk and investors would be willing to accept a return less than the risk-free asset.
Finally, it is important to point out the difference between the market
model and the CML and SML. The CML and the SML represent an ex ante, or
predictive, model for expected returns. The market model is an ex post, or
descriptive, model used to describe historical data. Hence, the market model
makes no prediction of what expected returns should be.

Tests of the CAPM

The number of articles found under the general heading “tests of the CAPM” is
impressive. One bibliographic compilation lists almost 1,000 papers on the topic.
The major implication that is tested is that beta should be the only factor
that is priced by the market. That is, other factors such as the variance or standard
deviation of the returns, and variables that we will discuss in later chapters such as
the price-earnings ratio, dividend yield, and firm size, should not be significant in
explaining the return on stocks. However, the empirical evidence finds that beta is
not the only factor priced by the market. Several studies have discovered other fac-
tors that explain stock returns. These include the other factors that we just cited.
One of the most controversial papers written on the CAPM is Richard
Roll’s “A Critique of the Asset Pricing Theory’s Tests.”10 The CAPM is a general
equilibrium model based upon the existence of a market portfolio that is defined
as the value-weighted portfolio of all investment assets. Furthermore, the market
portfolio is defined to be ex ante mean-variance efficient. Roll argues that this
means that the market portfolio lies on the ex ante Markowitz efficient frontier for
all investors and demonstrates that the only true test of the CAPM is whether the
market portfolio is in fact ex ante mean-variance efficient. However, the true mar-
ket portfolio is, in fact, ex ante mean-variance efficient since it includes all invest-
ment assets (e.g., stocks, bonds, real estate, art objects, and human capital). The
consequences of this “non-observability” of the true market portfolio are that the
CAPM is not testable until the exact composition of the true market portfolio is
known, and the only valid test of the CAPM is to observe whether the ex ante true
market portfolio is mean-variance efficient. As a result of his findings, Roll states
10 Richard Roll, “A Critique of the Asset Pricing Theory’s Tests,” Journal of Financial Economics (March

1977), pp. 129-176

 Chapter 2 35

that he does not believe there ever will be an unambiguous test of the CAPM. He
does not say that the CAPM is invalid. Rather, Roll says that there is likely to be
no unambiguous way to test the CAPM and its implications due to the non-
observability of the true market portfolio and its characteristics.
Does this mean that the CAPM is useless to the financial practitioner?
The answer is no, it does not. What it means is that the implications of the CAPM
should be viewed with caution.

Zero-Beta Version of the CAPM

As explained earlier, in a world without a risk-free asset, an investor will select
some portfolio on the Markowitz efficient frontier. When a risk-free asset is
assumed, the CML can be generated and dominates the Markowitz efficient fron-
tier. From the CML, the CAPM is derived.
Not only is the existence of a risk-free asset important in developing the
CAPM, but there are two related assumptions. First, it is assumed that investors
can borrow or lend at the risk-free rate. The risk-free asset is one in which there is
no uncertainty about the return that will be realized over some investment hori-
zon. To realize that return, it is assumed the borrower will not default on its obli-
gation. In the United States the short-term obligations of the federal government
are viewed as default-free and therefore risk-free assets. There is not just one
interest rate in an economy but a structure of interest rates. The U.S. government
pays the lowest interest rate, and individual borrowers pay a higher rate. The
greater the perceived risk that the borrower will default, the higher the interest
rate. Thus, while the U.S. government may be able to borrow at the risk-free rate,
an individual investor must pay a higher rate. Consequently, this assumption does
not reflect the situation facing investors in the real world.
The second related assumption is that investors can borrow and lend at the
same risk-free rate. In real-world markets, investors typically lend and borrow
money at different rates, the former being less than the latter. Again, the assumption
does not reflect the economic situation facing investors in real-world capital markets.
Fischer Black examined the implications for the original CAPM, when
there is no risk-free asset in which the investor can borrow and lend.11 He demon-
strated that neither the existence of a risk-free asset nor the requirement that inves-
tors can borrow and lend at the risk-free rate is necessary for the theory to hold.
However, without the risk-free asset a different form of the CAPM will result.
Black’s argument is as follows. The beta of a risk-free asset is zero. That is,
since there is no variability of the return on a risk-free asset, it cannot covary with
the market. Suppose that a portfolio can be created such that it is uncorrelated with
the market. This portfolio would have a beta of zero. We shall refer to any portfolio
with a beta of zero as a zero-beta portfolio. Black demonstrated how a zero-beta
portfolio can be constructed and that the CAPM would be modified as follows:
11Fischer Black, “Capital Market Equilibrium with Restricted Borrowing,” Journal of Business (July
1972), pp. 444-455.
36 Portfolio Theory and Asset Pricing

E(Rp) = E(RZ) + βp [E(RM) − E(RZ)] (14)

where E(RZ) is the expected return on the zero-beta portfolio and [E(RM) − E(RZ)] is
the market risk premium.
The version of the CAPM as given by equation (14) is the same as equa-
tion (12) except that the expected return for the zero-beta portfolio is substituted
for the risk-free rate. Black’s zero-beta version of the CAPM is called the two-
factor model. Empirical tests of the two-factor model suggest that it does a better
job in explaining historical returns than the pure CAPM.12
The basic principle in developing the zero-beta portfolio is that by means
of short selling, a zero-beta portfolio can be created from a combination of securi-
ties. The reason why short selling is a necessary assumption is that since assets
such as stocks are positively correlated, the only way to get a portfolio that is
uncorrelated with the market portfolio is to create a portfolio in which stocks are
owned and stocks are shorted. Thus, when the price of stocks increases, there will
be a gain on the stocks owned in the portfolio, giving a positive return; however,
there will be a loss on the stocks that have been shorted and therefore a negative
return. The zero-beta portfolio is created such that this combination of stocks
owned and stocks shorted will have a beta of zero.
Unfortunately, not all investors are permitted to sell short. Many institu-
tional investors are prohibited or constrained from selling short. Thus, the two-fac-
tor version of the CAPM avoids relying on the myth of “borrowing and lending at a
risk-free rate.” It still cannot reflect the real world for all investors, however,
because it does require unrestricted short selling, which is not available to everyone.

Arbitrage Pricing Theory Model

Professor Stephen Ross developed an alternative asset pricing model based purely
on arbitrage arguments, and hence called the arbitrage pricing theory (APT)
model.13 While we do not spend as much time in developing the underlying the-
ory for this asset pricing model, in Chapter 9 we discuss in more detail how factor
models are used in practice.
The arbitrage pricing theory model postulates that a security’s expected
return is influenced by a variety of factors, as opposed to just the single market
index of the CAPM. Specifically, look back at equation (9), which states that the
return on a security is dependent on its market sensitivity index and an unsystem-
atic return. The APT in contrast states that the return on a security is linearly
related to H “factors.” The APT does not specify what these factors are, but it is
assumed that the relationship between security returns and the factors is linear.

12 Fischer Black, Michael C. Jensen, and Myron Scholes, “The Capital Asset Pricing Model,” in Michael C.

Jensen (ed.), Studies in the Theory of Capital Markets (New York: Praeger 1972).
13 Stephen A. Ross, “The Arbitrage Theory of Capital Asset Pricing,” Journal of Economic Theory (Decem-

ber 1976), pp. 343-362,

 Chapter 2 37

The following notation will be used:

R̃ i = the random rate of return on security i
E(Ri) = the expected return on security i
F̃ h = the h-th factor that is common to the returns of all assets (h =
1, ..., H)
βi,h = the sensitivity of the i-th security to the h-th factor
H = number of factors
ẽ i = the unsystematic return for security i
The APT model asserts that the random rate of return on security i is
given by the following relationship:

R̃ i = RF + βi,1 F̃ 1 + βi,2 F̃ 2 + ... + βi,H F̃ H + ẽ i (15)

For equilibrium to exist, the following condition to insure no arbitrage

profits must be satisfied: using no additional funds (wealth) and without increas-
ing risk, it should not be possible, on average, to create a portfolio to increase
return. In essence, this condition states that there is no “money machine” avail-
able in the market.
Ross has shown that the following risk and return relationship will result
for each security i in a no-arbitrage world:

E(Ri) = RF + βi,F1[E(RF1) − RF] + βi,F2[E(RF2) − RF]

+ ... + βi,FH[E(RFH) − RF] (16)

where
βi,Fh = the sensitivity of security i to the h-th factor, and
E(RFh) − RF = the expected excess return of the h-th systematic factor
over the risk-free rate, and can be thought of as the
price (or risk premium) for the h-th systematic risk.
Equation (16) is the APT model. It states that investors want to be com-
pensated for all the factors that systematically affect the return of a security. The
compensation is the sum of the products of each factor’s systematic risk (βi,Fh),
and the risk premium assigned to it by the financial market [E(RFh − RF)]. As in the
case of the CAPM, an investor is not compensated for accepting unsystematic risk.
Examining the equations, we can see that the CAPM as given by equation
(12) is actually a special case of the APT model as given by equation (16). If the
only factor in equation (16) is market risk, the APT model reduces to equation (12).
Supporters of the APT model argue that it has several major advantages
over the CAPM. First, it makes less restrictive assumptions about investor prefer-
ences toward risk and return. CAPM theory assumes investors trade off between
risk and return solely on the basis of the expected returns and standard deviations
of prospective investments. The APT, in contrast, simply requires that some rather
38 Portfolio Theory and Asset Pricing

unobtrusive bounds be placed on potential investor utility functions. Second, no

assumptions are made about the distribution of security returns. Finally, since the
APT does not rely on the identification of the true market portfolio, the theory is
potentially testable.
The disadvantage is the identification of the factors. The factors must be
determined statistically. We’ll discuss this further when we look at the APT model
in Chapter 9 where we explain how factor models are used by practitioners.

SOME PRINCIPLES TO TAKE AWAY

In this chapter, we have covered the heart of what is popularly called modern
portfolio theory and capital market theory. We have emphasized the assumptions
and their critical role in the development of these theories. While you may under-
stand the topics covered, you may still be uncomfortable as to where we have pro-
gressed in investment management given the lack of theoretical and empirical
support for the CAPM or the difficulty of identifying the factors in the APT
model. You’re not alone. There are a good number of practitioners and academics
who feel uncomfortable with these models, particularly the CAPM.
Nevertheless, what is comforting is that there are several general princi-
ples of investing that are derived from these theories that very few would ques-
tion. All of these principles are used in later chapters.

1. Investing has two dimensions, risk and return. Therefore, focusing only on
the actual return that a manager has achieved without looking at the risk
that had to be accepted to achieve that return is inappropriate.
2. It is inappropriate to look at the risk of an individual asset when deciding
whether it should be included in a portfolio. What is important is how the
inclusion of an asset into a portfolio will affect the risk of the portfolio.
3. Whether investors consider one risk or 1,000 risks, risk can be divided into
two general categories: systematic risks that cannot be eliminated by
diversification, and unsystematic risk which can be diversified.
4. Investors should only be compensated for accepting systematic risks.
Thus, it is critical in formulating an investment strategy to identify the
systematic risks.
 Chapter 2 39

QUESTIONS

1. “A portfolio’s expected return and variance of return are simply the weighted
average of the individual asset’s expected returns and variances.” Explain why
you agree or disagreement with this statement?

2. Professor Harry Markowitz, co-recipient of the 1990 Nobel Prize in Economic

Science, wrote the following: “A portfolio with sixty different railway securi-
ties, for example, would not be as well diversified as the same size portfolio
with some railroad, some public utility, mining, various sort of manufacturing,
etc.” Why is this true?

3 a. What is meant by a Markowitz efficient frontier?

b. Explain why all feasible portfolios are not on the Markowitz efficient frontier.

4. Two portfolio managers are discussing modern portfolio theory. Manager A

states that the objective of Markowitz portfolio analysis is to construct a port-
folio that maximizes expected return for a given level of risk. Manager B dis-
agrees. He believes that the objective is to construct a portfolio that minimizes
risk for a given expected return. Which portfolio manager is correct?

5. Explain the critical role of the correlation between assets in determining the
potential benefits from diversification.

6. Suppose that you are evaluating the global diversification opportunities of a

portfolio consisting of the following four asset classes: U.S. large cap stocks,
U.S. small cap stocks, stocks of developed foreign countries, and stocks of
emerging market countries.

a. Write out the rate of return variance for this 4-asset global portfolio.
b. How many unique covariances are there in this portfolio having four asset
classes? Verify your answer by using the following formula:

Unique covariances = (G2 − G)/2

(G = Number of asset classes)
c. Should a risk averse investor prefer a high or low return correlation among U.S.
equities and, say, the stocks of emerging market countries? Explain your
answer using an efficient frontier.

7. A well-known mutual fund company offers an investment product called “The

30 Percent Solution — A Global Guide For Investors Seeking Better Perfor-
mance with Reduced Portfolio Risk.” The investment product recommends
investing 70% in U.S. equities and 30% in international equities.
40 Portfolio Theory and Asset Pricing

Use the following average return and risk information for U.S. and interna-
tional equities over a recent 20-year period to verify the risk management
accuracy (or inaccuracy) of the proposed “30 Percent Solution.”
U.S. Equities International Equities “70/30 Solution”
Average Return 15% 17% ?
Standard Deviation 14% 22% ?

The figures are based on S&P 500 and Morgan Stanley EAFE (Europe, Australia,
and Far East) total return indexes during 20-year period, 1977-1996.

Assume a correlation of 0.3 between domestic U.S. and international equity

returns in answering this question. Explain your findings.

8. The following excerpt is from Warren Bailey and Rene M. Stulz, “Benefits of
International Diversification: The Case of Pacific Basin Stock Markets,” Jour-
nal of Portfolio Management (Summer 1990):

Recent international diversification literature uses monthly data

from foreign stock markets to make the point that American
investors should hold foreign stock to reduce the variance of a
portfolio of domestic stocks without reducing its expected
return. (p. 57)

a. Why would you expect that the justification of diversifying into foreign stock
markets would depend on empirical evidence regarding the ability to “reduce
the variance of a portfolio of domestic stocks without reducing its expected
return”?
b. Typically in research papers that seek to demonstrate the benefits of interna-
tional diversification by investing in a foreign stock market, two efficient fron-
tiers are compared. One is an efficient frontier constructed using only domestic
stocks; the other is an efficient frontier constructed using both domestic and
foreign stocks. If there are benefits to diversifying into foreign stocks, should
the efficient frontier constructed using both domestic and foreign stocks lie
above or below the efficient frontier constructed using only domestic stocks?
Explain your answer.

9. The following excerpt is from John E. Hunter and T. Daniel Coggin, “An Anal-
ysis of the Diversification from International Equity Investment,” Journal of
Portfolio Management (Fall 1990):

The extent to which investment risk can be diversified depends

upon the degree to which national markets were completely domi-
nated by a single world market factor (i.e., if all cross-national cor-
relations were 1.00, then international diversification would have no
benefit). If all national markets were completely independent (that
 Chapter 2 41

is, if all cross-national correlations were zero), then international

diversification over an infinite number of countries would com-
pletely eliminate the effect of variation in national markets. (p. 33)

a. Why are the “cross-national correlations” critical in justifying the benefits from
international diversification?
b. Why do Hunter and Coggin state that there would be no benefit from interna-
tional diversification if these correlations are all 1.00?

10 a. In the “real world” the investor’s borrowing rate of interest is higher than the
lending rate. Show how this capital market imperfection impacts the set of
efficient portfolios that are available to risk averse investors. (Hint: the
“Capital Market Line” now consists of two linear segments.)
b. In a world where investors can borrow and lend at the same (risk-free) rate of
interest, we learned that there is one tangency portfolio consisting of all risky
assets called the market portfolio, M. Illustrate where the market portfolio, M,
would be positioned in the portfolio opportunity set when there exists a diver-
gence between borrowing and lending rates of interest in the economy.
Explain your answer.

11. What is the single factor that is used to explain the expected return for a secu-
rity or portfolio in the traditional CAPM?

12 a. How does the Black “zero-beta” model differ from the traditional CAPM?
Be sure to discuss the underlying logic of the Black portfolio model.
b. Show where Black’s “zero beta” portfolio lies on the boundary of feasible
investment portfolios. (Hint: Portfolio “Z,” for zero-beta portfolio, plots
directly across from a line drawn tangent to the market portfolio M and inter-
secting the expected return axis (vertical axis) in expected return and portfolio
risk space.)

13 a. What is a multi-factor asset-pricing model?

b. Why were multi-factor models developed, and what are their potential benefits
and limitations?

14. One of the major challenges for corporate pension plan (defined benefit) spon-
sors is to be sure that there are sufficient assets in a portfolio to cover pro-
jected payment of future benefits to retired (and retiring) employees. From a
“surplus risk management” perspective, should plan sponsors prefer positive
or negative return correlation in a portfolio having both assets and liabilities?

Explain your answer in the context of the rate of return variance for a “2-asset”
portfolio—namely, a portfolio consisting of a conventional asset class like com-
mon stocks and a liability (or negative asset) having the payoff characteristics of a
“fixed income” security.
View publication stats