Chapter 12: An Alternative View of Risk and Return:

The Arbitrage Pricing Theory

1.

Since we have the expected return of the stock, the revised expected return can be determined using
the innovation, or surprise, in the risk factors. So, the revised expected return is:
R = 11% + 1.2(4.2% 3%) 0.8(4.6% 4.5%)
R = 12.36%

2.

a.

If m is the systematic risk portion of return, then:

m = GNP GNP + Inflation Inflation + r Interest rates
m = .000586($5,436 5,396) 1.40(3.80% 3.10%) .67(10.30% 9.50%)
m = 0.83%

b.

The unsystematic return is the return that occurs because of a firm specific factor such as the
bad news about the company. So, the unsystematic return of the stock is 2.6 percent. The
total return is the expected return, plus the two components of unexpected return: the
systematic risk portion of return and the unsystematic portion. So, the total return of the stock
is:
R R m

R = 9.50% + 0.83% 2.6%

R = 7.73%
3.

a.

If m is the systematic risk portion of return, then:

m = GNP GNP + r Interest rates
m = 2.04(4.8% 3.5%) 1.90(7.80% 7.10%)
m = 1.32%

b.

The unsystematic is the return that occurs because of a firm specific factor such as the
increase in market share. If is the unsystematic risk portion of the return, then:
= 0.36(27% 23%)
= 1.44%

c.

The total return is the expected return, plus the two components of unexpected return: the
systematic risk portion of return and the unsystematic portion. So, the total return of the stock
is:
R R m

R = 10.50% + 1.32% + 1.44%

R = 13.26%
4.

The beta for a particular risk factor in a portfolio is the weighted average of the betas of the assets.
This is true whether the betas are from a single factor model or a multifactor model. So, the betas of
the portfolio are:
F1 = .20(1.20) + .20(0.80) + .60(0.95)

Answers to EndofChapter Problems

B175

F1 = 0.97
F2 = .20(0.90) + .20(1.40) + .60(0.05)
F2 = 0.43
F3 = .20(0.20) + .20(0.30) + .60(1.50)
F3 = 0.88
So, the expression for the return of the portfolio is:
Ri = 5% + 0.97 F1 + 0.43 F2 + 0.88 F3
Which means the return of the portfolio is:
Ri = 5% + 0.97(5.50%) + 0.43(4.20%) + 0.88(4.90%)
Ri = 16.45%
5.

We can express the multifactor model for each portfolio as:

E(RP ) = RF + 1 F1 + 2 F2
where F1 and F2 are the respective risk premiums for each factor. Expressing the return equation for
each portfolio, we get:
18% = 6% + 0.75F1 + 1.2F2
14% = 6% + 1.60 F1 0.2 F2
We can solve the system of two equations with two unknowns. Multiplying each equation by the
respective F2 factor for the other equation, we get:
3.6% = 1.2% + .15 F1 + 0.24 F2
16.8% = 7.2% + 1.92 F1 0.24 F2
Summing the equations and solving F1 fr gives us:
20.40% = 8.40% + 2.07 F1
F1 = 5.80%
And now, using the equation for portfolio A, we can solve for F2, which is:
18% = 6% + 0.75(5.80%) + 1.2F2
F2 = 6.38%

6.

a.

The market model is specified by:

R R ( R M RM )

so applying that to each Stock:

Stock A:
R A R A A ( R M RM ) A
Answers to EndofChapter Problems

B176

RA = 10.5% + 1.2(RM 14.2%) + A

Stock B:
R B R B B ( RM RM ) B
RB = 13.0% + 0.98(RM 14.2%) + B

Stock C:
RC RC C ( RM R M ) C

RC = 15.7% + 1.37(RM 14.2%) + C

b.

Since we don't have the actual market return or unsystematic risk, we will get aformula with
those values as unknowns:
RP = .30RA + .45RB + .30RC
RP = .30[10.5% + 1.2(RM 14.2%) + A] + .45[13.0% + 0.98(RM 14.2%) + B]
+ .25[15.7% + 1.37(RM 14.2%) + C
RP = .30(10.5%) + .45(13%) + .25(15.7%) + [.30(1.2) + .45(.98) + .25(1.37)]
(RM 14.2%)+ .30 A + .45 B + .30 C
RP = 12.925% + 1.1435(RM 14.2%) + .30 A + .45 B + .30 C

c.

Using the market model, if the return on the market is 15 percent and the systematic risk is
zero, the return for each individual stock is:
RA = 10.5% + 1.20(15% 14.2%)
RA = 11.46%
RB = 13% + 0.98(15% 14.2%)
RB = 13.78%
RC = 15.70% + 1.37(15% 14.2%)
RC = 16.80%
To calculate the return on the portfolio, we can use the equation from part b, so:
RP = 12.925% + 1.1435(15% 14.2%)
RP = 13.84%
Alternatively, to find the portfolio return, we can use the return of each asset and its portfolio
weight, or:
RP = X1R1 + X2R2 + X3R3
RP= .30(11.46%) + .45(13.78%) + .25(16.80%)
RP = 13.84%

7.

a.

Since five stocks have the same expected returns and the same betas, the portfolio also has
the same expected return and beta. However, the unsystematic risks might be different, so the
expected return of the portfolio is:
RP = 11% + 0.84F1 + 1.69F2 + (1/5)( 1 + 2 + 3 + 4 + 5)

Answers to EndofChapter Problems

B177

b.

Consider the expected return equation of a portfolio of five assets we calculated in part a.
Since we now have a very large number of stocks in the portfolio, as:
N , (1/N) 0
But, the Js are infinite, so:
(1/N)( 1 + 2 + 3 + 4+..+ N) 0
Thus:
R P 11% 0.84 F1 1.69 F2

8.

To determine which investment an investor would prefer, you must compute the variance of
portfolios created by many stocks from either market. Because you know that diversification is good,
it is reasonable to assume that once an investor has chosen the market in which she will invest, she
will buy many stocks in that market.
Known:
EF = 0 and = 0.10
E = 0 and S i = 0.20 for all i
If we assume the stocks in the portfolio are equallyweighted, the weight of each stock is
Xi =

1
, that is:
N

1
for all i
N

If a portfolio is composed of N stocks each forming 1/N proportion of the portfolio, the return on the
portfolio is 1/N times the sum of the returns on the N stocks. To find the variance of the respective
portfolios in the 2 markets, we need to use the definition of variance from Statistics:
Var(x) = E[x E(x)]2
In our case:
Var(RP) = E[RP E(RP)]2
Note however, to use this, first we must find RP and E(RP). So, using the assumption about equal
weights and then substituting in the known equation for Ri:
1
Ri
N
1
RP =
(0.10 + F + i)
N
1
RP = 0.10 + F +
i
N
RP =

Also, recall from Statistics a property of expected value, that is:

~ =a ~+ ~
If: Z
X Y
Answers to EndofChapter Problems

B178

~ , ~ , and ~ are random variables, then:

where a is a constant, and Z
X
Y
~
~
~
E (Z ) = E(a) E ( X ) + E (Y )

and
E(a) = a
Now use the above to find E(RP):
1

E ( RP ) E 0.10 F i
N

1
E ( RP ) 0.10 E ( F ) E ( i )
N
1
E ( R P ) 0.10 (0) (0)
N
E(RP) = 0.10
Next, substitute both of these results into the original equation for variance:
Var(RP) = E[RP E(RP)]2
Var(RP) = E 0.10 E ( F )
Var(RP) = E E ( F )

1
i
N

2
Var(RP) = E 2 F 2 F
2
2
Var(RP) = E F

1
i 0.10
N

1
1
2 ( ) 2
N
N

1 2
1
(1 )Cov( i , j )
N
N

Finally, since we can have as many stocks in each market as we want, in the limit, as N ,
1
0, so we get:
N
2
2
Var(RP) = Cov( i , j )

and, since:
Cov ( i , j ) i j ( i , j )

and the problem states that 1 = 2 = 0.10, so:

2
2
Var(RP) = i j ( i , j )
2
Var(RP) = (0.01) 0.04 ( i , j )

So now, summarize what we have so far:

Answers to EndofChapter Problems

B179

R1i = 0.10 + 1.5F + 1i

R2i = 0.10 + 0.5F + 2i
E(R1P) = E(R2P) = 0.10
Var(R1P) = 0.0225 0.04 ( 1 i , 2 j )
Var(R2P) = 0.0025 0.04 ( 2i , 2 j )
Finally we can begin answering the questions a, b, & c for various values of the correlations:
a.

Substitute

( 1i , 1 j ) =

( 2 i , 2 j ) = 0 into the respective variance formulas:

Var(R1P) = 0.0225
Var(R2P) = 0.0025
Since Var(R1P) > Var(R2P), and expected returns are equal, a risk averse investor will prefer to
invest in the second market.
b.

If we assume

( 1 i , 1 j ) = 0.9, and

Var(R1P) = 0.0225 + 0.04

( 2 i , 2 j ) = 0, the variance of each portfolio is:

( 1 i , 1 j )

Var(R1P) = 0.0225 + 0.04(0.9)

Var(R1P) = 0.0585
Var(R2P) = 0.0025 + 0.04 ( 2i , 2 j )
Var(R2P) = 0.0025 + 0.04(0)
Var(R2P) = 0.0025
Since Var(R1P) > Var(R2P), and expected returns are equal, a risk averse investor will prefer to
invest in the second market.
c.

If we assume

( 1 i , 1 j ) = 0, and

Var(R1P) = 0.0225 + 0.04

( 2 i , 2 j ) = 0.5, the variance of each portfolio is:

( 1 i , 1 j )

Var(R1P) = 0.0225 + 0.04(0)

Var(R1P)= 0.0225
Var(R2P) = 0.0025 + 0.04 ( 2 i , 2 j )
Var(R2P)) = 0.0025 + 0.04 (0.5)
Var(R2P) = 0.0225
Since Var(R1P)= Var(R2P), and expected returns are equal, a risk averse investor will be
indifferent between the two markets.

Answers to EndofChapter Problems

B180

d.

Since the expected returns are equal, indifference implies that the variances of the portfolios
in the two markets are also equal. So, set the variance equations equal, and solve for
thecorrelation of one market in terms of the other:
Var(R1P)= Var(R2P))
0.0225 + 0.04

( 1i , 1 j ) = 0.0025 + 0.04

( 2 i , 2 j ) =

( 1 i , 1 j ) + 0.5

( 2 i , 2 j )

Therefore, for any set of correlations that have this relationship (as found in part c), a risk
adverse investor will be indifferent between the two markets.
12.9

a.

In order to find standard deviation, , you must first find the Variance, since Var .
Recall from Statistics a property of Variance:
If: Z~ a X~ Y~
~ , ~ ,an ~ are random variables, then:
where a is a constant, and Z
X
Y
~
~
~
Var ( Z ) a 2 Var( X ) Var (Y )

The problem states that returngeneration can be described by:

Ri,t = i + i (RM) + i,t
Realize that Ri,t, RM, and i,t are random variables, and i and i are constants. Then, applying
the above properties to this model, we get:
Var(Ri,) = i 2 Var(RM) + Var( i)
and now we can find the standard deviation for each asset:

A2 = 0.72(0.0121) + 0.01 = 0.015929

A

0.015929 = .1262 or 12.62%

B2 = 1.22(0.0121) + 0.0144 = 0.031824

B

0.031824 = .1784 or 17.84%

C2 = 1.52(0.0121) + 0.0225 = 0.049725

C

b.

0.049725 = .2230 or 22.30%

From above formula for variance, note that as N ,

Var ( i )
, so you get:
N

Var(Ri) = i2 Var(RM)
So, the variances for the assets are:

Answers to EndofChapter Problems

B181

 A2 = 0.72(.0121) = 0.005929
B2 = 1.22(.0121) = 0.017424
C2 = 1.52(.0121) = 0.027225
c.

We can use the model:

Ri= RF + i (RM RF)
which is the CAPM (or APT Model when there is one factor and that factor is the Market).
So, the expected return of each asset is:
RA = 3.3% + 0.7(10.6% 3.3%) = 8.41%
RB = 3.3% + 1.2(10.6% 3.3%) = 12.06%
RC = 3.3% + 1.5(10.6% 3.3%) = 14.25%
We can compare these results for expected asset returns as per CAPM or APT with the
expected returns given in the table. This shows that assets A & B are accurately priced, but
asset C is overpriced (the model shows the return should be higher). Thus, rational investors
will not hold asset C.

10.

d.

If short selling is allowed, rational investors will sell short asset C, causing the price of asset
C to decrease until no arbitrage opportunity exists. In other words, the price of asset C should
decrease until the return becomes 14.25 percent.

a.

Let:
X1= the proportion of Security 1 in the portfolio and
X2= the proportion of Security 2 in the portfolio
and note that since the weights must sum to 1.0,
X 1= 1 X 2
Recall that the beta for a portfolio (or in this case the beta for a factor) is the
weighted average of the security betas, so
P1 = X111 + X2 21
P1 = X111 + (1X1) 21
Now, apply the condition given in the hint that the return of the portfolio does not depend on
F1. This means that the portfolio beta for that factor will be 0, so:
P1 = X111 + X2 21
P1 = 0 = X1(1.0) + (1 X1)(0.5)
and solving for X1 and X2:

Answers to EndofChapter Problems

B182

X 1= 1
X 2= 2
Thus, sell short Security 1 and buy Security 2.
To find the expected return on that portfolio, use
RP = X1R1 + X2R2
so applying the above:
E(RP) = 1(20%) + 2(20%)
E(RP) = 20%
P1 = 1(1) + 2(0.5)
P1 = 0
b.

Following the same logic as in part a, we have

P2 = 0 = P1 = X331 + (1X3) 41
P2 = 0 = X3(1) + (1 X3)(1.5)
and
X 3= 3
X4= 2
Thus, sell short Security 4 and buy Security 3. Then,
E(RP2) = 3(10%) + (2)(10%)
E(RP2) = 10%
P2 = 3(0.5) 2(0.75)
P2 = 0
Note that since both P1 and P2 are 0, this is a risk free portfolio!

c.

The portfolio in part b provides a risk free return of 10%, which is higher than the 5% return
provided by the risk free security. To take advantage of this opportunity, borrow at the risk
free rate of 5% and invest the funds in a portfolio built by selling short security four and
buying security three with weights (3,2) as in part b.

d.

First assume that the risk free security will not change. The price of security four (that
everyone is trying to sell short) will decrease, and the price of security three (that everyone is
trying to buy) will increase. Hence the return of security four will increase and the return of
security three will decrease.
The alternative is that the prices of securities three and four will remain the same, and the
price of the riskfree security drops until its return is 10%.

Answers to EndofChapter Problems

B183

Finally, a combined movement of all security prices is also possible. The prices of security
four and the riskfree security will decrease and the price of security three will increase until
the opportunity disappears.

Answers to EndofChapter Problems

B184

MINI CASE The Fama-French Multifactor Model and Mutual Fund Returns
NOTE: The example below shows the results for returns between January 2001 and December 2005.
The actual answer to the case will change based on current market conditions.
1.

For a largecompany stock fund, we would expect the beta for the market risk premium to be near
one since large company returns account for a large part of the total market return on a marketvalue
basis. We would expect the betas for the SMB and HML risk factors to be low, and possibly
negative, since large company stock returns are not highly related to small company stock returns
and large company stocks tend to be more oriented toward value stocks.

2.

The following shows the results of the regression estimates for the period between January 2001 and
December 2005. The actual answer to the case will change based on current market returns.
Fidelity Magellan:
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations

0.98136
0.96306
0.96108
0.00876
60

ANOVA
Regression
Residual
Total

df
3
56
59

SS
0.111927178
0.004292802
0.116219979

Coefficients
Intercept
MktRF
SMB
HML

0.00080
0.98379
0.20016
0.07497

Standard
Error
0.00123
0.02991
0.03985
0.04202

MS
0.037309
7.67E05

F
486.7002

t Stat
0.64652
32.89282
5.02241
1.78426

Pvalue
0.52059
0.00000
0.00001
0.07980

MS
0.02465
0.00035

F
70.63527

Significance F
4.60021E40

Fidelity LowPriced Stock Fund:

Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations

0.88937
0.79097
0.77977
0.01868
60

ANOVA
Regression
Residual
Total

df
3
56
59

SS
0.07396
0.01954
0.09350

Answers to EndofChapter Problems

Significance F
4.99566E19

B185

Coefficients
Intercept
MktRF
SMB
HML

0.00944
0.73959
0.43843
0.35001

Standard
Error
0.00263
0.06382
0.08504
0.08965

t Stat
3.59433
11.58912
5.15571
3.90414

Pvalue
0.00069
0.00000
0.00000
0.00026

MS
0.03894
0.00069

F
56.31373

t Stat
0.58715
8.96478
5.79182
1.54484

Pvalue
0.55946
0.00000
0.00000
0.12802

Baron Small Cap Fund:

Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations

0.86663
0.75105
0.73771
0.02630
60

ANOVA
Regression
Residual
Total

df
3
56
59

SS
0.11682
0.03872
0.155539609

Coefficients
Intercept
MktRF
SMB
HML

0.00217
0.80529
0.69327
0.19495

Standard
Error
0.00370
0.08983
0.11970
0.12619

Significance F
6.50745E17

3.

This answer will depend on the returns used by the students.

4.

If the market is efficient, all assets should have an alpha of zero. In this case, only one of the three
funds has a positive alpha, so the evidence provided here against market efficiency is limited. We
should note that the alpha for the Fidelity LowPriced Stock Fund is extremely large, 0.94 percent
per month, or just over 11 percent per year. It is still possible that the model misspecification
problem exists and that other risk factors are present that we are not examining in this regression
model.

5.

On a riskadjusted basis, the Fidelity LowPriced Stock Fund performed the best over this period
since it was the only fund with a positive alpha.

Answers to EndofChapter Problems

B186