Chapter 5 Risk and Return: Portfolio Theory and Asset Pricing Models

ANSWERS T EN!" #"C$APTER %&EST' NS

5-1

a. A portfolio is made up of a group of individual assets held in combination. An asset that would be relatively risky if held in isolation may have little, or even no risk if held in a well-diversified portfolio. The feasible, or attainable, set represents all portfolios that can be constructed from a given set of stocks. This set is only efficient for part of its combinations. An efficient portfolio is that portfolio which provides the highest expected return for any degree of risk. Alternatively, the efficient portfolio is that which provides the lowest degree of risk for any expected return. The efficient frontier is the set of efficient portfolios out of the full set of potential portfolios. n a graph, the efficient frontier constitutes the boundary line of the set of potential portfolios. b. An indifference curve is the risk!return trade-off function for a particular investor and reflects that investor"s attitude toward risk. The indifference curve specifies an investor"s re#uired rate of return for a given level of risk. The greater the slope of the indifference curve, the greater is the investor"s risk aversion. The optimal portfolio for an investor is the point at which the efficient set of portfolios--the efficient frontier--is $ust tangent to the investor"s indifference curve. This point marks the highest level of satisfaction an investor can attain given the set of potential portfolios. c. The %apital Asset &ricing 'odel (%A&') is a general e#uilibrium market model developed to analy*e the relationship between risk and re#uired rates of return on assets when they are held in well-diversified portfolios. The +', is part of the %A&'. The %apital 'arket ,ine (%',) specifies the efficient set of portfolios an investor can attain by combining a risk-free asset and the risky market portfolio '. The %', states that the expected return on any efficient portfolio is e#ual to the riskless rate plus a risk premium, and thus describes a linear relationship between expected return and risk. d. The characteristic line for a particular stock is obtained by regressing the historical Answers and Solutions: 5 " (

returns on that stock against the historical returns on the general stock market. The slope of the characteristic line is the stock"s beta, which measures the amount by which the stock"s expected return increases for a given increase in the expected return on the market. The beta coefficient (b) is a measure of a stock"s market risk. -t measures the stock"s volatility relative to an average stock, which has a beta of 1... e. Arbitrage &ricing Theory (A&T) is an approach to measuring the e#uilibrium risk!return relationship for a given stock as a function of multiple factors, rather than the single factor (the market return) used by the %A&'. The A&T is based on complex mathematical and statistical theory, but can account for several factors (such as /0& and the level of inflation) in determining the re#uired return for a particular stock. The 1ama-1rench 2-factor model has one factor for the excess market return (the market return minus the risk free rate), a second factor for si*e (defined as the return on a portfolio of small firms minus the return on a portfolio of big firms), and a third factor for the book-to-market effect (defined as the return on a portfolio of firms with a high book-to-market ratio minus the return on a portfolio of firms with a low bookto-market ratio). 'ost people don3t behave rationally in all aspects of their personal lives, and behavioral finance assume that investors have the same types of psychological behaviors in their financial lives as in their personal lives. 5-4 +ecurity A is less risky if held in a diversified portfolio because of its lower beta and negative correlation with other stocks. -n a single-asset portfolio, +ecurity A would be more risky because 5A 6 57 and %8A 6 %87.

Answers and Solutions: 5 " )

S +&T' NS T

EN!" #"C$APTER PR ,+EMS

5-1

a. A plot of the approximate regression line is shown in the following figure9

2. 45 4. 15 1. 5 . -2. -4. -1. -5 -1. -15 -4. . 1. 4. 2. E.

rB (C)

rF
5.

:sing ;xcel, the regression e#uation estimates are9 7eta < ..5=> -ntercept < ...2?> @ 4 < ..A=. b. The arithmetic average return for +tock B is calculated as follows9
r Avg = ( 1E.. + 42.. + ... +1D.4) = 1..=C. ?

The arithmetic average rate of return on the market portfolio, determined similarly, is 14.1C. 1or +tock B, the estimated standard deviation is 12.1 percent9
B = ( 1E.. 1..=) 4 + (42.. 1..=) 4 +... + (1D.4 1..=) 4 = 12.1C. ? 1

Answers and Solutions: 5 " *

The standard deviation of returns for the market portfolio is similarly determined to be 44.= percent. The results are summari*ed below9 Average return, r Avg +tandard deviation, 5 +tock B 'arket &ortfolio 1..=C 14.1C 12.1 44.=

+everal points should be noted9 (1) 5 ' over this particular period is higher than the historic average 5' of about 15 percent, indicating that the stock market was relatively volatile during this period> (4) +tock B, with B < 12.1C, has much less total risk than an average stock, with Avg < 44.=C> and (2) this example demonstrates that it is possible for a very low-risk single stock to have less risk than a portfolio of average stocks, since 5B G 5'. c. +ince +tock B is in e#uilibrium and plots on the +ecurity 'arket ,ine (+',), and given the further assumption that r B = r B and r ' = r ' --and this assumption often does not hold--then this e#uation must hold9
r B = r@1 + ( r r@1 )b B .

This e#uation can be solved for the risk-free rate, r@1, which is the only unknown9
1..= = r@1 + (14.1 r@1 ) ..5= 1..= = r@1 + =.D ..5= r@1 ..EE r@1 = 1..= =.D r@1 = 2.D ! ..EE = D.=C.

Answers and Solutions: 5 " -

d. The +', is plotted below. Hata on the risk-free security (b@1 < ., r@1 < D.=C) and +ecurity B (bB < ..5=, r X < 1..=C) provide the two points through which the +', can be drawn. r' provides a third point.
r(%) k(%)

20

rXkX == 10.6% 10.6%

kM = 12.1%

rRF = 8.6%10
kRF = 8.6

1.0

2.0

Beta

e. -n theory, you would be indifferent between the two stocks. +ince they have the same beta, their relevant risks are identical, and in e#uilibrium they should provide the same returns. The two stocks would be represented by a single point on the +',. +tock F, with the higher standard deviation, has more diversifiable risk, but this risk will be eliminated in a well-diversified portfolio, so the market will compensate the investor only for bearing market or relevant risk. -n practice, it is possible that +tock F would have a slightly higher re#uired return, but this premium for diversifiable risk would be small.

Answers and Solutions: 5 " 5

5-4 a. The regression graph is shown above. :sing a speadsheet, we find b < ..=4.
rk (%) Sy (%)
45

30

15

-30

-15

15

30

45

rM(%)

kM (%)

-15

b. 7ecause b < ..=4, +tock F is about =4 percent as volatile as the market> thus, its relative risk is about =4 percent of that of an average firm. c. 1. Total risk (4 F ) would be greater because the second term of the firm"s risk 4 4 4 e#uation, F = b 4 F ' + eF , would be greater. 4. %A&' assumes that company-specific risk will be eliminated in a portfolio, so the risk premium under the %A&' would not be affected. d. 1. The stock"s variance would not change, but the risk of the stock to an investor holding a diversified portfolio would be greatly reduced. 4. -t would now have a negative correlation with r'. 2. 7ecause of a relative scarcity of such stocks and the beneficial net effect on portfolios that include it, its Irisk premiumI is likely to be very low or even negative. Theoretically, it should be negative.

Answers and Solutions: 5 " .

5-2

a. ri = r@1 + (r' r@1 )b i = r@1 + (r' r@1 )

i' i . ' r' r@1 ' ri' i .

b. %',9

r p = r@1

r ' r@1 + '

p .

+',9 ri = r@1 +

Jith some arranging, the similarities between the %', and +', are obvious. Jhen in this form, both have the same market price of risk, or slope,(r ' - r@1)!5'. The measure of risk in the %', is 5 p. +ince the %', applies only to efficient portfolios, 5p not only represents the portfolio"s total risk, but also its market risk. Kowever, the +', applies to all portfolios and individual securities. Thus, the appropriate risk measure is not 5i, the total risk, but the market risk, which in this form of the +', is ri'5i, and is less than for all assets except those which are perfectly positively correlated with the market, and hence have ri' < L1... 5-E a. :sing the %A&'9 ri < r@1 L (r' - r@1)bi < ?C L (1.1)(=.5C) < 1E.15C b. :sing the 2-factor model9 ri < r@1 L (r' M rrf)bi L (r+'7)ci L (rK',)di < ?C L (1.1)(=.5C) L (5C)(..?) L (EC)(-..2) < 1=.E5C

Answers and Solutions: 5 " /

M'N' CASE

To 0egin1 0riefly re2ie3 the Chapter - Mini Case4 Then1 e5tend your kno3ledge of risk and return 0y ans3ering the follo3ing 6uestions4 a4 Suppose asset A has an e5pected return of (7 percent and a standard de2iation of )7 percent4 Asset , has an e5pected return of (. percent and a standard de2iation of -7 percent4 'f the correlation 0et3een A and , is 74-1 3hat are the e5pected return and standard de2iation for a portfolio co8prised of *7 percent asset A and /7 percent asset ,9
N& =w A r NA + (1 w A ) r N7 r =..2(..1) +..?(..1=) =..1E4 = 1E.4C.
4 4 4 4 p = JA A +(1 JA ) 7 +4 JA (1 JA ) A7 A 7

Ans3er:

= ..2 4 (..4 4 ) +..? 4 (..E 4 ) +4(..2)(..?)(..E)(..4)(..E) =..2.A

04 Ans3er:

Plot the attaina0le portfolios for a correlation of 74-4 No3 plot the attaina0le portfolios for correlations of :(47 and "(474

AB = +0.4: Attainable Set of Risk/Return Combinations

20% #$"e%te& return 15% 10% 5% 0% 0%

10%

20% Risk! "

0%

40%

Mini Case: 5 " ;

 AB = +1.0: Attainable Set of Risk/Return Combinations 20% #$"e%te& return 15% 10% 5% 0% 0% 10% 20% Risk! " 0% 40%

AB = '1.0: Attainable Set of Risk/Return Combinations 20% #$"e%te& return 15% 10% 5% 0% 0% 10% 20% Risk! " 0% 40%

Mini Case: 5 " <

c4

Suppose a risk"free asset has an e5pected return of 5 percent4 ,y definition1 its standard de2iation is =ero1 and its correlation 3ith any other asset is also =ero4 &sing only asset A and the risk"free asset1 plot the attaina0le portfolios4

Ans3er:
Attainable Set of Risk/Return Combinations (it) Risk'*ree Asset
15% #$"e%te& return

10%

5%

0% 0% 5% 10% Risk! " 15% 20%

Mini Case: 5 " (7

d4

Construct a reasona0le1 0ut hypothetical1 graph 3hich sho3s risk1 as 8easured 0y portfolio standard de2iation1 on the 5 a5is and e5pected rate of return on the y a5is4 No3 add an illustrati2e feasi0le >or attaina0le? set of portfolios1 and sho3 3hat portion of the feasi0le set is efficient4 What 8akes a particular portfolio efficient9 !on@t 3orry a0out specific 2alues 3hen constructing the graphA 8erely illustrate ho3 things look 3ith Breasona0leB data4

Ans3er:
Expecte !"rt#"$%" Ret+r, Expecte !"rt#"$%"
Ret+r,' k"
+

r,
E##%c%e,t Set (('B)

B )

*
Fea&%-$e' "r (tta%,a-$e' Set

E
R%&k' "

r%&k' P

The figure above shows the feasible set of portfolios. The points 7, %, H, and ; represent single securities (or portfolios containing only one security). All the other points in the shaded area, including its boundaries, represent portfolios of two or more securities. The shaded area is called the feasible, or attainable, set. The boundary A7 defines the efficient set of portfolios, which is also called the efficient frontier. &ortfolios to the left of the efficient set are not possible because they lie outside the attainable set. &ortfolios to the right of the boundary line (interior portfolios) are inefficient because some other portfolio would provide either a higher return with the same degree of risk or a lower level of risk for the same rate of return.

e4

No3 add a set of indifference cur2es to the graph created for part ,4 What do these cur2es represent9 What is the opti8al portfolio for this in2estor9 #inally1 add a second set of indifference cur2es 3hich leads to the selection of a different opti8al portfolio4 Why do the t3o in2estors choose different portfolios9 Mini Case: 5 " ((

Expecte !"rt#"$%" Ret+r,' Expecte !"rt#"$%" 2 2 Ret+r,' k p rp

B ) /pt%0a$ !"rt#"$%" .,1e&t"r B * /pt%0a$ !"rt#"$%" .,1e&t"r (

. . . . .

B2 B1

(3 (2 (1

R%&k' p r%&k' P

Ans3er: The figure above shows the indifference curves for two hypothetical investors, A and 7. To determine the optimal portfolio for a particular investor, we must know the investor"s attitude towards risk as reflected in his or her risk!return tradeoff function, or indifference curve. %urves -a1, -a4, and -a2 represent the indifference curves for individual A, with the higher curve (-a2) denoting a greater level of satisfaction (or utility). Thus, -a2 is better than -a4 for any level of risk. The optimal portfolio is found at the tangency point between the efficient set of portfolios and one of the investor"s indifference curves. This tangency point marks the highest level of satisfaction the investor can attain. The arrows point toward the optimal portfolios for both investors A and 7. The investors choose different optimal portfolios because their risk aversion is different. -nvestor A chooses the portfolio with the lower expected return, but the riskiness of that portfolio is also lower than investor"s 7 optimal portfolio, because investor a is more risk averse.

f4

What is the capital asset pricing 8odel >CAPM?9 What are the assu8ptions that underlie the 8odel9

Ans3er: The Capital Asset Pricing Model (CAPM) is an e#uilibrium model which specifies the relationship between risk and re#uired rates of return on assets when they are held in Mini Case: 5 " ()

well-diversified portfolios. The %A&' re#uires an extensive set of assumptions9 All investors are single-period expected utility of terminal wealth maximi*ers, who choose among alternative portfolios on the basis of each portfolio"s expected return and standard deviation. All investors can borrow or lend an unlimited amount at a given risk-free rate of interest. -nvestors have homogeneous expectations (that is, investors have identical estimates of the expected values, variances, and covariances of returns among all assets). All assets are perfectly divisible and perfectly marketable at the going price, and there are no transactions costs. There are no taxes. All investors are price takers (that is, all investors assume that their own buying and selling activity will not affect stock prices). The #uantities of all assets are given and fixed. What i8pact does this ha2e on the efficient

g4

No3 add the risk"free asset4 frontier9

Ans3er: The risk-free asset by definition has *ero risk, and hence 5 < .C, so it is plotted on the vertical axis. 0ow, given the possibility of investing in the risk-free asset, investors can create new portfolios that combine the risk free asset with a portfolio of risk! assets. This enables them to achieve any combination of risk and return that lies along any straight line connecting r@1 with any portfolio in the feasible set of risky portfolios. Kowever, the straight line connecting r@1 with m, the point of tangency between the line and the portfolio"s efficient set curve, is the one that all investors would choose. +ince all portfolios on the line r@1m* are preferred to the other risky portfolio opportunities on the efficient frontier A7, the points on the line r @1m* now represent the best attainable combinations of risk and return. Any combination under the r @1m* line offers less return for the same amount of risk, or offers more risk for the same amount of return. Thus, everybody wants to hold portfolios which are located on the r@1m* line.

h4

Write out the e6uation for the capital 8arket line >CM+? and dra3 it on the Mini Case: 5 " (*

graph4 'nterpret the CM+4 No3 add a set of indifference cur2es1 and illustrate ho3 an in2estor@s opti8al portfolio is so8e co80ination of the risky portfolio and the risk"free asset4 What is the co8position of the risky portfolio O

Expecte !"rt#"$%"
Expecte !"rt#"$%" Ret+r,' 2 Ret+r,' k p

3 B

rp
M

( k RF

rRF
R%&k' p

r%&k' P

Ans3er: The line r@1m* in the figure above is called the capital market line (CM"). -t has an intercept of r@1 and a slope of ( r ' r@1 ) ! ' . Therefore the e#uation for the capital market line may be expressed as follows9

P r' r @1 %',9 r p = r + @1 p . '

The %', tells us that the expected rate of return on any efficient portfolio (that is, Mini Case: 5 " (-

any portfolio on the %',) is e#ual to the risk-free rate plus a risk premium, and the risk premium is e#ual to ( r ' r@1 ) ! ' multiplied by the portfolio"s standard deviation, p . Thus, the %', specifies a linear relationship between expected return and risk, with the slope of the %', being e#ual to the expected return on the market portfolio of risky stocks, r M , minus the risk-free rate, r@1, which is called the market risk premium, all divided by the standard deviation of returns on the market portfolio, 5m .
;xpected @ate of @eturn,
Expecte Rate rp 2 "# Ret+r,' kp
2

2 .1

)M4

k RF

rRF

/pt%0 a$ !"rt#"$%"

R%&k' p

The figure above shows a set of indifference curves (i1, i4, and i2), with i1 touching the %',. This point of tangency defines the optimal portfolio for this investor, and he or she will buy a combination of the market portfolio and the risk-free asset. The risky portfolio, m, must contain every asset in exact proportion to that asset"s fraction of the total market value of all assets> that is, if security g is x percent of the total market value of all securities, x percent of the market portfolio must consist of security g.

Mini Case: 5 " (5

i4

What is a characteristic line9 $o3 is this line used to esti8ate a stockCs 0eta coefficient9 Write out and e5plain the for8ula that relates total risk1 8arket risk1 and di2ersifia0le risk4

Ans3er: 7etas are calculated as the slope of the characteristic line, which is the regression line formed by plotting returns on a given stock on the y axis against returns on the general stock market on the x axis. -n practice, 5 years of monthly data, with =. observations, would be used, and a computer would be used to obtain a least s#uares regression line. The relationship between stock Q"s total risk, market risk, and diversifiable risk can be expressed as follows9
T TA, @-+R = 8A@-A0%; = 'A@R;T @-+R + H-8;@+-1-A7,; @-+R 4 Q =
4 b4 Q '

4 eQ

2 2 Kere J is the variance or total risk of stock $, M is the variance of the market, b$ 2 is stock Q"s beta coefficient, and eJ is the variance of stock Q"s regression error term. -f stock Q is held in isolation, then the investor must bear its total risk. Kowever, when stock Q is held as part of a well-diversified portfolio, the regression error term, 2 eJ is driven to *ero> hence, only the market risk remains.

D4

What are t3o potential tests that can 0e conducted to 2erify the CAPM9 What are the results of such tests9 What is rollCs criti6ue of CAPM tests9

Ans3er: +ince the %A&' was developed on the basis of a set of unrealistic assumptions, empirical tests should be used to verify the %A&'. The first test looks for stability in historical betas. -f betas have been stable in the past for a particular stock, then its historical beta would probably be a good proxy for its e# ante, or expected beta. ;mpirical work concludes that the betas of individual securities are not good estimators of their future risk, but that betas of portfolios of ten or more randomly selected stocks are reasonably stable, hence that past portfolio betas are good estimators of future portfolio volatility. The second type of test is based on the slope of the SM". As we have seen, the %A&' states that a linear relationship exists between a security"s re#uired rate of return and its beta. 1urther, when the +', is graphed, the vertical axis intercept Mini Case: 5 " (.

should be r@1, and the re#uired rate of return for a stock (or portfolio) with beta < 1.. should be rm, the re#uired rate of return on the market. 8arious researchers have attempted to test the validity of the %A&' model by calculating betas and reali*ed rates of return, plotting these values in graphs, and then observing whether or not (1) the intercept is e#ual to r @1, (4) the regression line is linear, and (2) the +', passes through the point b < 1.., rm. ;vidence shows a more-or-less linear relationship between reali*ed returns and market risk, but the slope is less than predicted. Tests that attempt to assess the relative importance of market and company-specific risk do not yield definitive results, so the irrelevance of diversifiable risk specified in the %A&' model can be #uestioned. $oll #uestioned whether it is even conceptually possible to test the %A&'. @oll showed that the linear relationship which prior researchers had observed in graphs resulted from the mathematical properties of the models being tested, hence that a finding of linearity proved nothing about the validity of the %A&'. @oll"s work did not disprove the %A&' theory, but he did show that it is virtually impossible to prove that investors behave in accordance with the theory. -n general, evidence seems to support the %A&' model when it is applied to portfolios, but the evidence is less convincing when the %A&' is applied to individual stocks. 0evertheless, the %A&' provides a rational way to think about risk and return as long as one recogni*es the limitations of the %A&' when using it in practice. k4 ,riefly e5plain the difference 0et3een the CAPM and the ar0itrage pricing theory >APT?4

Ans3er: The %A&' is a single-factor model, while the Arbitrage Pricing %heor! (AP%) can include any number of risk factors. -t is likely that the re#uired return is dependent on many fundamental factors such as the /0& growth, expected inflation, and changes in tax laws, and that different groups of stocks are affected differently by these factors. Thus, the apt seems to have a stronger theoretical footing than does the %A&'. Kowever, the apt faces several ma$or hurdles in implementation, the most severe being that the apt does not identify the relevant factors--a complex mathematical procedure called factor analysis must be used to identify the factors. To date, it appears that only three or four factors are re#uired in the apt, but much more research is re#uired before the apt is fully understood and presents a true challenge to the %A&'.

Mini Case: 5 " (/

l4

Suppose you are gi2en the follo3ing infor8ation4 The 0eta of co8pany1 0 i1 is 74<1 the risk free rate1 rR#1 is .4;E1 and the e5pected 8arket pre8iu81 rM"rR#1 is .4*E4 ,ecause your co8pany is larger than a2erage and 8ore successful than a2erage >i4e41 it has a lo3er 0ook"to"8arket ratio?1 you think the #a8a"#rench *" factor 8odel 8ight 0e 8ore appropriate than the CAPM4 Fou esti8ate the additional coefficients fro8 the #a8a"#rench *"factor 8odel: the coefficient for the si=e effect1 Ci1 is "7451 and the coefficient for the 0ook"to"8arket effect1 d i1 is G 74*4 'f the e5pected 2alue of the si=e factor is -E and the e5pected 2alue of the 0ook"to"8arket factor is 5E1 3hat is the re6uired return using the #a8a"#rench *"factor 8odel9 >assu8e that Ai H 7474? What is the re6uired return using CAPMO

Ans3er: The 1ama-1rench model9 ri < r@1 L (rm - r@1)bi L (rsmb)ci L (rhmb)d$ ri < =.DC L (=.2C)(..A) L (EC)(-..5) L (5C)(-..2) < D.A?C The %A&'9 ri < rrf L (rm - rrf)bi ri < =.DC L (=.2C)(..A) < 14.E?C

Mini Case: 5 " (;