You are on page 1of 48

Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.

The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1. Foundations of Risk Management

Bionic Turtle FRM Practice Questions

Chapter 5. Modern Portfolio Theory (MPT) and the


Capital Asset Pricing Model (CAPM)
By David Harper, CFA FRM CIPM
www.bionicturtle.com
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Key Ideas from authors prior to 2020

KEY IDEAS FROM ELTON, MODERN PORTFOLIO THEORY & INVESTMENT ANALYSIS ...................... 3
KEY IDEAS FROM AMENC, PORTFOLIO THEORY AND PERFORMANCE ANALYSIS, CHAPTER 4 ......... 5

Chapter 5. Modern Portfolio Theory (MPT) and the Capital Asset Pricing Model (CAPM)

P1.T1.20.8. MODERN PORTFOLIO THEORY (MPT) ..................................................................... 6


P1.T1.20.9. PERFORMANCE MEASURES ...................................................................................10
P1.T1.403. TREYNOR, SHARPE, AND ALPHA CALCULATIONS ......................................................13
P1.T1.404. TRACKING ERROR, INFORMATION RATIO, AND SORTINO ...........................................15
P1.T1.61. CAPITAL MARKET LINE .............................................................................................18
P1.T1.31. SHARPE, TREYNOR, AND JENSEN’S ..........................................................................20
P1.T1.32. TRACKING ERROR, SHARPE & SORTINO ...................................................................22
P1.T1.26. CAPITAL MARKET LINE ............................................................................................24
P1.T1.28. CAPITAL ASSET PRICING MODEL (CAPM) ................................................................26
RELATED BIONIC TURTLE YOUTUBE VIDEOS.............................................................................28

APPENDIX (OPTIONAL):

Elton & Gruber, Chapters 5: Delineating Efficient Portfolios

P1.T1.57. PORTFOLIO RETURN AND VOLATILITY........................................................................29


P1.T1.59. MVP AND THE EFFICIENT FRONTIER .........................................................................31

Elton & Gruber, Chapter 14: Nonstandard Forms of Capital Asset Pricing Models

P1.T1.62. NONSTANDARD CAPITAL ASSET PRICING MODEL (NONSTANDARD CAPM) ...................33


P1.T1.63. ARBITRAGE PRICING MODEL (APT)...........................................................................36

Grinold & Kahn, Active Portfolio Management: A Quantitative Approach for Producing
Superior Returns & Controlling Risk, Chapter 7: Expected Returns and the Arbitrage
Pricing Theory

P1.T1.34. ARBITRAGE PRICING THEORY (APT) .........................................................................39


P1.T1.35. APT FORECAST RETURNS .......................................................................................41
P1.T1.36. APT VERSUS CAPM ...............................................................................................43
P1.T1.38. APT FACTOR FORECAST .........................................................................................45
P1.T1.39. STATISTICAL VERSUS STRUCTURAL APT MODELS .....................................................47

2
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Key Ideas from authors prior to 2020


Key Ideas from Elton, Modern Portfolio Theory & Investment Analysis1

 Be prepared to solve for the expected return and variance of a two-asset portfolio; i.e.,
Variance(two-asset portfolio) = X(A)^2*variance(A) + X(B)^2*variance(B) +
2*X(A)*X(B)*covariance(A,B), where X(A) is fraction of portfolio held in A
1
 P   X   X   2 X A X B AB 
2
A
2
A
2
B
2
B
2

 Covariance(A,B) = Standard deviation (A) * Standard deviation (B) *


Correlation(A,B)
 AB
 AB   A B  AB   AB 
 A B
 Be ready especially to apply is case where correlation equals -1.0, zero or 1.0.

 The minimum variance portfolios are the curve of portfolios that represent the minimum
variance conditional on a given return.
o The global minimum variance portfolio (global MVP) is the portfolio of risky
assets with the minimum variance.
o The efficient frontier includes the dominating portfolios on the upper segment,
starting with the global MVP
o The Market Portfolio has the highest Sharpe ratio among the efficient portfolios
on the efficient frontier
 After we introduce (add) the riskfree asset, the (straight-line) CML becomes the
efficient frontier

 The SML expresses the capital asset pricing model (CAPM). The most likely exam
question in this section is an application of the CAPM

 Memorize capital asset pricing model (CAPM): E[R(i)] = Rf + Beta (i,M)*[R(M) - Rf],

Ri  RF  i  RM  RF 

1 Edwin J. Elton, Martin J. Gruber, Stephen J. Brown and William N. Goetzmann, Modern Portfolio Theory and

Investment Analysis, 9th Edition (Hoboken, NJ: John Wiley & Sons, 2014)

3
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Key Ideas from Elton, Modern Portfolio Theory & Investment Analysis2 Continued:

o Beta (i,M) = covariance (i, M)/variance(M)


= correlation(i,M) * StdDev(i)/StdDev(M
iM i
i ,M    i ,M
 M2 M

o CAPM in English
Expected return = Price of time + Price of risk * quantity of risk,
where MRP (market risk premium) is price of risk,
beta is quantity of risk, or

Expected excess return = Price of risk * quantity of risk


 Don’t forget that CAPM can be used to generate a discount rate to compute the present
value!
o For example, you may be asked to use CAPM to compute expected return (R),
then discount a cash flow; e.g., PV = FV/(1+R)^n
 Otherwise, for purposes of the exam, you probably do not need to take a deep dive into
the underlying theory. We advise you focus on:
o Two-asset variance
o Covariance and correlation
o Beta (and it’s direct relationship to covariance/correlation)
o At least a superficial understanding of CML & SML; but including the Sharpe ratio
(e.g., slope of the CML) as it applies elsewhere
o Application of CAPM: the best thing you can do here is practice computing the
expected return, per the CAPM, based on input assumptions.

2
Edwin J. Elton, Martin J. Gruber, Stephen J. Brown and William N. Goetzmann, Modern Portfolio Theory and
Investment Analysis, 9th Edition (Hoboken, NJ: John Wiley & Sons, 2014)

4
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Key ideas from Amenc, Portfolio Theory and Performance Analysis,


Chapter 43
Amenc overlaps with Elton’s CAPM-based chapters. This reflects the FRM’s current emphasis
on the classic introductory finance theory, namely mean-variance framework and single-factor
(i.e., excess market return) risk-adjusted performance metrics (RAPMs). Obviously, you need to
know the same CAPM reviewed by Elton; but Amenc applies it here in the form of Jensen’s
alpha. This alpha is the over- or under-performance not explained by the beta exposure:

Jensen’s alpha is “the excess return equated to alpha plus expected systematic return;” but it is
easier to see that it is simply the difference between the portfolio return, R(p), and the return
that CAPM expects for the portfolio:

E ( RP )  RF   P   P ( E ( RM )  RF )
Among the other metrics, four are most testable. The Sharpe and Treynor have the same
general intent: “excess return per unit of risk.” As such, they are risk-adjusted performance
metrics (RAPMs). The Treynor assumes that beta is risk, so it considers only systematic risk.
The Sharpe assumes volatility is risk, so it considers total risk.

The other two important metrics are tracking error (TE) and information ration (IR).

Sharpe ratio Treynor measure


E( RP )  RF E( RP )  RF
SP  TP 
 ( RP ) P

Tracking Error (TE) Information Ratio


TE   (RP  RB ) E( RP )  E( RB )
IR 
 ( RP  RB )
Two other points:

 The information ratio has three possible definitions, depending on whether we use active
or residual risk/return. GARP should be aware of this issue and you can expect the
question to be specific. More here in a forum post by David
http://www.bionicturtle.com/forum/threads/information-ratio-definition.5554/

You should understand the Sortino ratio conceptually, but as a tedious ex post measure, you
unlikely to need to compute it.

3
Noel Amenc and Veronique Le Sourd, Portfolio Theory and Performance Analysis (West Sussex, England: John
Wiley & Sons, 2003)

5
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Chapter 5. Modern Portfolio Theory (MPT) and the Capital


Asset Pricing Model (CAPM)
P1.T1.20.8. Modern portfolio theory (MPT)
P1.T1.20.9. Performance measures
P1.T1.403. Treynor, Sharpe, and Alpha Calculations
P1.T1.404. Tracking Error, Information Ratio, and Sortino
P1.T1.61. Capital Market Line
P1.T1.31. Sharpe, Treynor, and Jensen’s
P1.T1.32. Tracking Error, Sharpe & Sortino
P1.T1.26. Capital Market Line
P1.T1.28. Capital Asset Pricing Model (CAPM)

P1.T1.20.8. Modern portfolio theory (MPT)


Learning objectives: Explain modern portfolio theory and interpret the Markowitz
efficient frontier. Understand the derivation and components of the CAPM. Describe the
assumptions underlying the CAPM. Interpret the capital market line. Apply the CAPM in
calculating the expected return on an asset. Interpret beta and calculate the beta of a
single asset or portfolio.

20.8.1. The plot below illustrates an actual portfolio possibilities curve (PPC, dashed line). Two
prominent portfolios (minimum variance and market portfolio) are also plotted.

The green point is the minimum variance


portfolio, MVP, that has an expected return of
7.70% and standard deviation, σ, of 10.0%.
The red triangle is the Market Portfolio that has
an expected return of 11.0% and standard
deviation, σ, of 13.7%. If the riskfree rate is
4.0%, what is the formula for the capital
market line (CML)?

a) E(R) = 0.040 + 0.511 × σ(P)


b) E(R) = 0.137 + 0.511 × σ(P)
c) E(R) = 0.040 + 0.375 × β(P,M)
d) E(R) = 0.077 + 0.511 × β(P,M)

6
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

20.8.2. Consider the market portfolio plus three other portfolios (A, B, C) with the following
features while the riskfree rate is 3.0%:

Please note the returns are expected gross returns; e.g., the market's expected return is 9.0%
so that its expected excess return is 6.0%. If we assume the capital asset pricing model (CAPM)
is valid then which of the following statements is TRUE?

a) Portfolio A is mispriced or it cannot exist in equilibrium due to both its beta and standard
deviation
b) Portfolio B is mispriced or it cannot exist in equilibrium due to its standard deviation
c) Portfolio C is mispriced or it cannot exist in equilibrium due to its beta
d) All three portfolios can exist in equilibrium under the CAPM

20.8.3. Assume the riskfree rate is 3.0% while the market portfolio has a return of 13.0% (i.e.,
excess return is 10.0%) with volatility of 15.0%. Further, the following four portfolios are
observed:
 Portfolio (A) has an expected return of 11.0% with volatility of 8.0%
 Portfolio (B) has an expected return of 7.0% with volatility of 20.0% and its correlation to
the market is 0.30
 Portfolio (C) has a beta with respect to the market β(C, M) = 1.50 and its realized return
of +20.0% implies an alpha of +2.0% while its correlation to the market is 0.50
 Portfolio (D) has a beta with respect to the market β(C, M) = 0.40 and its realized return
of +9.0% implies an alpha of +2.0% while its correlation to the market is 0.50
Under the conditions of modern portfolio theory (MPT; mean-variance framework) and the
capital asset pricing model (CAPM), each of the above portfolios is valid and plausible EXCEPT
which is not possible?

a) Portfolio A
b) Portfolio B
c) Portfolio C
d) Portfolio D

7
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

20.8.1. A. True: the CML is here given by E(R) = 0.040 + 0.511 × σ(P)

The capital market line (CML) must have the y-intercept intercept at the riskfree rate, which was
given as 4.0%. Its tangent is the market portfolio. The slope of the CML is the market portfolio's
Sharpe ratio, which is (11.0% - 4.0%)/13.7%. Therefore, the CML is given by E(R) = 0.040 +
(0.110 - 0.040)/0.1370 × σ(P) = 0.040 + 0.511× σ(P).

20.8.2. C. TRUE: Portfolio C is mispriced or it cannot exist in equilibrium due to its beta

The market's Treynor ratio is (9.0% - 3.0%) / 1.0 = 0.060. But Portfolio C has a Treynor ratio of
(13.0% - 3.0%)/1.4 = 0.0714; it is above the SML, with a Jensen's alpha of α = 13.0% - 3.0% -
1.4 × (9.0% - 3.0%) = +0.40%. As these are expected returns, this is not possible under CAPM
equilibrium. In regard to (A), (B) and (D), each is FALSE:
 Portfolio A has a Treynor of 0.060 (i.e., on the SML) and Sharpe ratio of (7.8% - 3.0%) /
30.0% = 0.160
 Portfolio B has a Treynor of 0.060 (i.e., on the SML) and Sharpe ratio of (10.2% - 3.0%) /
26.0% = 0.277. It is true that Portfolio B, when compared to Portfolio A, has higher return
and lower standard deviation (as evidenced by it higher Sharpe ratio). Portfolio B is
clearly more efficient than Portfolio A. This is not a violation of CAPM unless a portfolio
has a higher Sharpe than the market portfolio: the market portfolio, being the most
efficient, must have the highest Sharpe ratio. Indeed, the market portfolio's Sharpe ratio
of (9.0% - 3.0%) / 20.0% = 0.30 is the highest.

8
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

20.8.3. A. Portfolio A is unrealistic because it is "more efficient" than the Market portfolio:
it has a Sharpe ratio of 1.0 versus the market portfolio's Sharpe ratio of (13% - 3%)/15% =
0.667.

In regard to (B), (C) and (D), each is plausible


 Portfolio B has β(B, M) = 0.30 * 20.0% / 15.0% = 0.40 with a CAPM expected return,
E(B) = 3.0% + 0.40 * 10.0% = 7.0%
 Portfolio C's Jensen's alpha, α(J, C) = 20.0% - 3.0% - 1.50 * 10.0% = +2.0%
 Portfolio D's Jensen's alpha, α(J, D) = 9.0% - 3.0% - 0.40 * 10.0% = +2.0%

Discuss here in the forum: https://www.bionicturtle.com/forum/threads/p1-t1-20-8-modern-


portfolio-theory-mpt.23118/

9
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.20.9. Performance measures


Learning objectives: Calculate, compare and interpret the following performance
measures: the Sharpe performance index, the Treynor performance index, the Jensen
performance index, the tracking error, information ratio, and Sortino ratio.

20.9.1. The riskfree rate is 3.0% and the market portfolio's expected return is 10.0% (put
another way, the market's excess expected return is 7.0%). Consider two portfolios:
 Portfolio A has a high volatility, σ(A) = 50.0% per annum, but its correlation to the market
portfolio is only, ρ(A, M) = 0.30
 Portfolio B has a moderate volatility, σ(B) = 30.0% per annum, and its correlation to the
market portfolio is, ρ(B, M) = 0.70
If both portfolios lie on the security market line (SML), and if Portfolio A has an expected return,
ER(A) = 7.20%, then what is the expected return of Portfolio B?

a) 5.100%
b) 7.900%
c) 8.880%.
d) 9.540%

20.9.2. Let Rm(P) denote the monthly return of the portfolio and Rm(B) denote the monthly
return of its benchmark. Over a three-year measurement period, the following statistics are
calculated:
 The average monthly return for, respectively, the portfolio and the benchmark was
Rm(P) = 8.50% and Rm(B) = 6.90%; therefore, on average, the portfolio outperformed
its benchmark by +1.60%.
 The monthly standard deviation of the difference between the portfolio's and
benchmark's return, σ[Rm(P) - Rm(B)] = 11.80%
 A regression of the portfolio's excess return against the benchmark's excess return
produced the sample regression function, ERm(P) = -0.0140 + 1.35 * ERm(B); therefore,
the regression intercept (aka, alpha) is -1.40%
 The standard error of the regression (SER), which approximates the volatility of alpha,
σ(α), is 11.0%

Although the periodicity is monthly, an information ratio is generally annualized. In regard to


annualized information ratios, which of the following statements is accurate?
a) The (both annualized) active information ratio is +0.136 and the residual information ratio
is +1.426
b) The (both annualized) active information ratio is +0.470 and the residual information ratio
is -0.441
c) The (both annualized) active information ratio is +0.636 and the residual information ratio
is -0.127
d) The (both annualized) active information ratio is -0.250 and the residual information ratio
is +0.889

10
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

20.9.3. During a 36-month period during which the risk-free rate was 2.0%, consider a
comparison between the market portfolio and two fund managers, Betty and Peter:
 The market portfolio's excess return was +6.0% (its gross return was +8.0%) with a
volatility of 24.0% per annum
 Peter's portfolio's excess return was +7.0% (his gross return was + 9.0%) with a volatility
of 36.0% per annum and beta, β(P,M) = 0.750
 Betty's portfolio's excess return was +11.0% (her gross return was + 13.0%) with a
volatility of 44.0% per annum and beta, β(B,M) = 1.650
If the capital asset pricing model (CAPM) is valid, then each of the following statements is true
EXCEPT which is false?

a) If the investor is diversified (or the portfolio represents only a fraction of the investor's
entire wealth), the Treynor (TPI) is a good measure and Peter's TPI is better than Betty's
b) If the investor is not diversified (or the portfolio represents the investor's entire wealth),
the Sharpe (SPI) is a good measure and Betty's SPI is better than Peter's
c) Betty's portfolio already lies on the efficient frontier such that we do not expect her to
sustain further improvement in the reward-to-variability ratio
d) If Betty's and Peter's portfolio belong to different peer groups, Jensen's alpha is a good
measure for ranking them and Betty's (Jensen's) alpha is better than Peter's

11
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

20.9.1. C. True: 8.880%.

If the both portfolios lie on the SML, then ER(A) = Rf + β(A,M) *[ER(M) - Rf] and ER(B) = Rf +
β(B,M) *[ER(M) - Rf]. We can infer Portfolio A's beta: since 7.20% = 3.0% + β(A, M) * 7.0%, β(A,
M) = 0.60. And since β(A, M) = ρ(A, M)*σ(A)/σ(M), we can infer that σ(M) = [ρ(A, M)*σ(A)]/β(A,
M) = [0.30*0.50]/0.60 = 0.25. Then we can compute the beta of portfolio B: β(B, M) = ρ(B,
M)*σ(B)/σ(M) = 0.70*0.30/0.25 = 0.840. The expected return of of portfolio B is given by ER(B)
= 3.0% + 0.840 * 7.0% = 8.880%.

20.9.2. B. True: The (both annualized) active information ratio is +0.470 and the residual
information ratio is -0.441

The active IR = (active return) / (active risk; aka, tracking error) = +1.60% / 11.80% per month,
when annualized is (+1.60% * 12) / [11.80% * sqrt (12)] = +1.60% / 11.80% * sqrt (12) = +0.470
The residual IR = alpha / σ(alpha) = -1.40% / 11.0%, which annualized is given by -1.40% /
11.0% * sqrt(12) = -0.441

20.9.3. D. False, doubly: if the portfolios belong in SIMILAR peer groups (i.e., similar risk
levels), Jensen's alpha is good for ranking and Peter's is better than Betty's

In regard to (A), (B) and (C), each is TRUE.


 If the investor is diversified (or the portfolio represents only a fraction of the investor's
entire wealth), the Treynor (TPI) is a good measure. Peter's TPI is 9.33% and Betty's
TPI is 6.67%.
 If the investor is not diversified (or the portfolio represents the investor's entire wealth),
the Sharpe (SPI) is a good measure. Peter's SPI is 0.194 and Betty's is 0.250.
 Betty's portfolio already lies on the efficient frontier because her Sharpe ratio equals the
Market's, which is also 6.0% / 24% = 0.250; we do not expect portfolios to be located
beyond the efficient portfolio.

Discuss in the forum here: https://www.bionicturtle.com/forum/threads/p1-t1-20-9-


performance-measures.23129/

12
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.403. Treynor, Sharpe, and alpha calculations


Learning Objectives: Calculate, compare, and evaluate the Treynor measure, the Sharpe
measure, and Jensen's alpha.

403.1. Assume the slope of the security market line (SML) is 0.060 while the riskfree rate is
2.0%. What is the Treynor measure of a security with an alpha of 2.40% and beta of 0.30?

a) 0.140
b) 0.280
c) 0.560
d) 1.120

403.2. A portfolio with a volatility of 30.0% has a Treynor measure of 0.080. The portfolio has a
correlation of 0.50 with the market index which itself has a volatility of 20.0%. What is the
portfolio's Sharpe measure?

a) 0.095
b) 0.200
c) 0.330
d) 0.475

403.3. Assume the market index return is 8.0% while the risk-free rate is 3.0%. A portfolio with a
volatility of 12.0% has a Sharpe measure of 0.50 and a Treynor measure of 0.20. What is the
portfolio's alpha?

a) -2.79%
b) 1.16%
c) 3.83%
d) 4.50%

13
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

403.1. A. 0.140

The slope of the SML is the market's excess return such that the security's excess return is
4.20% = 0.060*0.30 + 2.40%

The security's Treynor measure is therefore 0.140 = 4.20%/0.30 = 0.140.

403.2. B. 0.200

The beta(P,M) = correlation(P,M)*volatility(P)/volatility(M) = 0.50*30%/20% = 0.750.


Because Treynor = (Portfolio's excess return)/beta, the portfolio's excess return = 0.080*0.750 =
6.0% and its Sharpe measure = 6.0%/30.0% = 0.20.

403.3. D. 4.50%

The portfolio's excess return = Sharpe*volatility = 0.50*12% = 6.0%, such that its beta = (excess
return)/Treynor = 6.0%/0.20 = 0.30.
Portfolio alpha = (portfolio's excess return) - beta*(market premium) = 6.0% - 0.30*5.0% =
4.50%.

Discuss in forum here: https://www.bionicturtle.com/forum/threads/p1-t1-403-treynor-sharpe-


and-alpha-calculations.7506/

14
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.404. Tracking error, Information ratio, and Sortino


Learning Objective: Compute and interpret tracking error, the information ratio, and the
Sortino ratio

404.1. Over the previous twelve (12) months, Analyst Robert regressed Portfolio (P) excess
returns against the Benchmark (M) excess returns:

As shown above, the regression equation is given by P = 0.0171 + 0.8195*M. Further, the
standard error of the regression (SER) is 0.0180; this SER can be considered the ex ante
tracking error because it is the square root of the variance of the regression's residual
(prediction error). Which is nearest to the residual-based information ratio (IR)?

a) 0.655
b) 0.728
c) 0.833
d) 0.950

15
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

404.2. Instead of residual-based information ratio (IR), it is also acceptable to compute


information ratio (IR) based on active returns. The following table displays twelve (12) months of
returns comparing a portfolio (P) to its benchmark (B); the final column shows the difference
each month:

The final two rows show the average and sample standard deviation of the monthly return
statistics. Which is nearest to the annualized ex-post (active-based) information ratio (IR)?

a) 0.404
b) 0.651
c) 0.950
d) 1.237

404.3. Consider the following already-annualized statistics for portfolio (P):


 Riskfree rate = 2.00%
 Realized portfolio (P) return (average) = 9.50%
 Portfolio (P) excess return = 9.50% - 2.00% = 7.50%
 Standard deviation of portfolio (P) returns = 14.70%
 Minimum acceptable return (MAR) = 6.00%
 Downside deviation of portfolio (P) returns = 5.60%
Which are nearest, respectively, to the Sharpe measure and Sortino ratio?
a) 0.280 (Sharpe) and 0.100 (Sortino)
b) 0.350 (Sharpe) and 0.433 (Sortino)
c) 0.510 (Sharpe) and 0.625 (Sortino)
d) 0.740 (Sharpe) and 1.290 (Sortino)

16
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

404.1. D. 0.950. Information ratio (IR) = alpha/[tracking error] = 0.0171/0.0180 = 0.950, as the
regression intercept is alpha.

404.2. B. 0.651. Annualized ex post IR = (0.0044*12)/[0.0234*SQRT(12)] = 0.65137.

404.3. C. 0.510 (Sharpe) and 0.625 (Sortino)


Sharpe = excess return/volatility = 0.0750/0.1470 = 0.510;
Sortino = [return above MAR]/downside deviation = (9.50% - 6.00%)/5.60% = 0.6250

Discuss in forum here: https://www.bionicturtle.com/forum/threads/p1-t1-404-tracking-error-


information-ratio-and-sortino.7519/

17
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.61. Capital market line


Learning Objectives: Describe the capital market line. Use the CAPM to calculate the
expected return on an asset.

61.1. Which of the following is a DIFFERENCE between the capital asset pricing model (CAPM)
and the capital market line (CML)

a) The CML does not include the riskfree asset, but CAPM does
b) CAPM is a special case of the CML, where the portfolio is diversified and efficient
c) In CAPM, risk is systematic (beta) since it can apply to inefficient portfolios; but in CML,
risk is total (volatility) since it only includes efficient portfolios
d) CAPM assumes the portfolio is diversified and efficient, but CML allows for un-diversified
and/or inefficient portfolios

61.2. A security will produce only two cash flows: $100 at the end of the first year, and $100 at
the end of the second year. The riskfree rate is 3.0% and the Market's expected return is 8.0%.
The security's volatility is 24.0% and the Market's volatility is 15.0%; the correlation (rho)
between the security and the Market is 0.70. Under the capital asset pricing model (CAPM) with
annual discounting, what is the present value of the security?

a) $169.01
b) $176.87
c) $185.95
d) $191.35

61.3. During the most recent period, a Portfolio returned 10.3% when the Market return was
only 8.0%. The riskfree rate was 2.0%. The Market's return was 8.0% with volatility of 29.0%.
Finally, the covariance between the portfolio and Market was 0.134560. Under the CAPM, did
the portfolio outperform?

a) No, Jensen's alpha is -1.30%


b) No, Jensen's alpha is -0.50%
c) Yes, Jensen's alpha is +0.50%
d) Yes, Jensen's alphs is +1.30%

18
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

61.1. C. In CAPM, risk is systematic (beta) since it can apply to inefficient portfolios; but
in CML, risk is total (volatility) since it only includes efficient portfolios
 In regard to (B), the inverse is true: CML is a special case of SML (CAPM) where the
portfolio is efficient and perfectly correlated to the market portfolio; i.e., ρ(i,M) = 1.0.

61.2. B. $176.87
 Beta (i,M) = covariance(i, M)/variance(M) = 24%*15%*0.70/15%^2 = 1.12
CAPM: E[R(i)] = Rf + Beta (i,M)*[R(M) - Rf] = 3% + 1.12*(8%-3%) = 8.60%
PV (annual compounding) = $100/(1.086) + $100/(1.086)^2 = $176.87

61.3. A
 The market risk premium = 8% - 2% = 6%.
Beta (i,M) = covariance(i, M)/variance(M) = 0.134560 / 29%^2 = 1.60.
Under CAPM, the E[portfolio return] = Rf + Beta (i,M)*[R(M) - Rf] = 2% + 1.60*6% =
11.60%.
The Jensen's alpha is 10.3% - 11.6% = -1.30%.

Discuss in the forum here: http://www.bionicturtle.com/forum/threads/p1-t1-61-capital-market-


line-elton-gruber.5273/

19
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.31. Sharpe, Treynor, and Jensen’s


Learning Objective: Calculate, compare, & evaluate Treynor measure, Sharpe measure,
and Jensen’s alpha.

31.1 A portfolio has a correlation of 0.40 with the overall market and produces a Sharpe ratio of
0.2. If the market’s volatility is 20%, what is the portfolio’s Treynor ratio? (take your time...)
a) 4.0%
b) 6.0%
c) 10.0%
d) Not enough information

31.2 Assume the riskfree rate is 3.0% and the price of risk (excess market return) is 5.0%. A
manager’s portfolio produces a return of 9.0% with 30% volatility and a CAPM beta of 0.8 (i.e.,
quantity of risk = 0.8). What is the (ex post or realized) Sharpe ratio?
a) 0.0
b) 0.2
c) 0.4
d) 0.6

31.3 For the same portfolio, what is the Treynor ratio?


a) 5.0%
b) 6.0%
c) 7.5%
d) 9.0%

31.4 For the same portfolio, what is the Jensen’s alpha?


a) -1.0%
b) 0.0%
c) +1.0%
d) +2.0%

31.5 Which is best for RANKING portfolios with the same beta (within peer groups)?
a) Treynor
b) Sharpe
c) Jensen’s
d) None

31.6 Which is BEST for evaluating portfolios that are NOT very well-diversified?
a) Treynor
b) Sharpe
c) Jensen’s
d) None

20
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

31.1. C.
Sharpe = Excess return / volatility and Treynor = Excess return / beta.
Since beta = correlation(M,P)*volatility(P)/volatility(M),
Treynor = Excess return / [correlation(M,P)*volatility(P)/volatility(M)], and
Treynor = Excess return * 1/correlation (M,P) * volatility(M)/volatility(P), and
Treynor = Excess return/volatility (P) * volatility(M)/correlation (M,P), such that
Treynor = Sharpe ratio * volatility(M)/correlation (M,P)

In this case, Treynor = 0.2 * 20%/0.4 = 10%

31.2 B. (9% - 3%)/30% = 0.2

31.3. C (9% - 3%)/0.8 = 7.5%

31.4 D. 9% - (0.8 *5%) - 3% = 2%

31.5. C (Jensen’s)
“The Jensen alpha can be used to rank portfolios within peer groups. Peer groups were
presented in Chapter 2. They group together portfolios that are managed in a similar manner,
and that therefore have comparable levels of risk.”4

31.6. B. If they are not well diversified, then the idiosyncratic risk will not be eliminated to zero;
therefore, volatility (total risk including idiosyncratic) rather than beta (only systemic risk) should
be used.

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-31-sharpe-treynor-


jensens.3460/

4
Noel Amenc and Veronique Le Sourd, Portfolio Theory and Performance Analysis (West Sussex, England: John
Wiley & Sons, 2003)

21
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.32. Tracking Error, Sharpe & Sortino


Learning Objective: Compute/interpret tracking error, information ratio, & Sortino ratio.

32.1 Make the following assumptions:


 Riskfree rate is 3%
 The benchmark is the market (i.e., CAPM) and the benchmark return was 8%
 Portfolio beta is 1.2
 Portfolio return was 10%
 Tracking error was 10%
 Minimum acceptable return (MAR) was 2%
 Downside deviation was 5%
What is (was) the information ratio?
a) 0.10
b) 0.20
c) 0.30
d) 0.40

32.2 IF we would like to be 95% confident that alpha is significant, how many years do we need
to observe?
a) 10
b) 64
c) 125
d) 384

32.3 What was the Sortino ratio?


a) 0.8
b) 1.2
c) 1.4
d) 1.6

32.4 If a portfolio’s volatility (i.e., annualized standard deviation) was 24%, what is the tracking
error (TE) if the benchmark is cash with a constant return of 2% and no volatility?
a) Less than 24%
b) 24%
c) Greater than 24%
d) Need more information

32.5 Which is the information ratio?


a) active return / active risk
b) active return / residual risk
c) residual return / active risk
d) residual return / residual risk

22
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

32.1. A. 0.10
Alpha = 10% - (1.2 beta * 5% ERP) - 3% riskfree rate = 1%.
IR = alpha/TE = 1%/10% = 0.10
... please note that (B) is tempting because alpha of 10% - 8% is tempting. However, that is
active return not residual return (alpha).

32.2 D. 384 years


T = (t-stat / IR)^2. In this case, two-tailed 95% = 1.96 and:
T = (1.96 / 0.10)^2 =~ 384 years (!)

32.3. D. 1.6
(10%-2%)/5% = 1.6

32.4 B. 24%
VAR(A - B) = VAR(A) + VAR(B) - 2*COV(A,B), and if VAR(B) = 0, then
VAR (A-B) = VAR(A) + 0 - 0 = VAR(A)

32.5. Either (A) or (D) is acceptable

Discuss in forum here: https://www.bionicturtle.com/forum/threads/l1-t1-32-tracking-error-


sharpe-sortino.3465/

23
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.26. Capital Market Line


Learning Objective: Describe the capital market line and the construction of the efficient
frontier both with and without a risk-free asset.

26.1 Assume the expected return on the market portfolio is 15.0% with volatility of 12.8%. The
risk-free rate is 4%.
a) What is the market portfolio's Sharpe ratio?
b) What is the slope of the CML?
c) What is the Sharpe ratio of a portfolio equally allocated between the riskfree asset (50%)
and the market portfolio (50%)?
d) What is the beta of a portfolio equally allocated between the risk-free asset (50%) and
the market portfolio (50%)?
e) What is the function that characterizes the CML here, in slope-intercept form?

26.2 Is each of the following TRUE or FALSE?


a) According to Equilibrium Theory, only portfolios on the CML are efficient
b) According to Equilibrium Theory, different allocations WITHIN the universe of risky
assets cannot improve
c) According to Equilibrium Theory, all investors will seek the same allocation between the
market portfolio and the riskless asset
d) The CML requires that all investors have the same beliefs in regard to asset returns,
variances and covariance values

24
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

26.1
a) Sharpe = excess return / volatility = (15% - 4%) / 12.8% = ~0.86
b) 0.86, the same as the Sharpe. The slope of CML = rise/run = E(market return) - riskfree
/volatility = ~0.86.
c) 0.86, no calculation is necessary: all the portfolios on the CML have the same Sharpe
ratio. In this case, the 50/50 portfolio has expected return of ~9.5% and volatility of
~6.4% (i.e., 50% * 12.8%) which gives a Sharpe ratio of (9.5-4%)/6.4% = 0.86
d) As beta is a linear combination and the beta of the riskfree asset is zero, the beta is 0.5
(50% * 1). Or, equivalently: E(portfolio return) = riskfree + beta * E(excess market
return), such that, beta = E(excess portfolio return)/E(excess market return), in this case:
beta = (9.5% - 4%) / (15% - 4%) = 0.5
e) CML can be given by: E(portfolio return) = 4% + 0.86*volatility(portfolio)
On the other hand, SML can be given by: E(portfolio return) = 4% + (% invested in
market portfolio)*11%, if the only choice is between market portfolio (beta = 1.0) and
risk-free asset (beta=0) such that (% invested in market portfolio) = beta, since 11% is
the market's excess return.

26.2
a) True. After the addition of the riskfree asset, the (linear) CML is the efficient frontier
b) True! The allocation decision is only between the market portfolio of risky assets and the
riskless asset; under the assumptions, it can be shown that alternative mixes of risky
assets are suboptimal
c) False. All points on the CML have the same Sharpe ratio and are equally efficient, this
mix choice is a matter of individual risk aversion
True! This is the key (unrealistic) assumption required to validate the findings; this is the
condition that justifies a single consensus market portfolio. From Amenc: “Up until now
we have only considered the case of an isolated investor. By now assuming that all
investors have the same expectations concerning assets, they all then have the same
return, variance and covariance values and construct the same efficient frontier of risky
assets. In the presence of a risk-free asset, the reasoning employed for one investor is
applied to all investors. The latter therefore all choose to divide their investment between
the risk-free asset and the same risky asset portfolio M.5”

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-26-cml.3449/

5
Noel Amenc and Veronique Le Sourd, Portfolio Theory and Performance Analysis (West Sussex, England: John
Wiley & Sons, 2003)

25
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.28. Capital Asset Pricing Model (CAPM)


Learning Objective: Describe the Capital Asset Pricing Model (CAPM), list its underlying
assumptions, and explain its implications, contributions and limitations.

28.1 Assume the riskfree rate is 4% and the expected (overall) market return is 12% with 20%
volatility. Our portfolio (P) has volatility of 30% and a correlation with the market of 0.4.
According to CAPM, what is the portfolio’s expected return?
a) 6.0%
b) 8.8%
c) 11.2%
d) 12.0%

28.2 CAPM assumptions: Which of the following is NOT an underlying assumption of the CAPM
model?
a) Investors only consider the first two moments of return distribution: the expected return
and the variance
b) All investors have the same forecast return, variance and covariance expectations for all
assets
c) All investors prefer to be fully invested in the market portfolio
d) Markets are perfect: there are no taxes and no transaction costs. All assets are traded
and are infinitely divisible

28.3 CAPM contributions: According to CAPM, what is the expected return of diversifiable
(idiosyncratic) risk?
a) Zero
b) Beta
c) Beta * Excess Market Return
d) Riskless rate + (Beta * Excess Market Return)

28.4 CAPM limitations: In order for CAPM to hold, markets must exhibit the STRONG form of
efficiency; is it NOT enough that markets are weak- or semi-strong efficient. Which assumption
is the key to achieving the strong form market efficiency that supports the CAPM conclusion?
a) Investors are risk averse
b) Investors only care about first two moments
c) All investors make the same forecasts concerning the assets
d) Markets are frictionless, without taxes and transaction costs, with infinitely divisible
assets

26
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

28.1. B. 8.8%
beta = cov(P,M)/variance(M) = (20%*30%*0.4)/20%^2 = 0.6
CAPM says E(P return) = 4% + 0.6*(12%-4%) = 8.8%

28.2 C. All investors prefer to be fully invested in the market portfolio

The market portfolio is the MOST EFFICIENT mix of risky assets but the CAPM does not restrict
all investors to prefer a beta of 1.0.

28.3 A. Zero

The CAPM is a SINGLE-FACTOR model that says expected return is a function of SYSTEMIC
RISK (beta); in the diversified portfolio, unsystematic risk is eliminated and receives no
compensation. Amenc: “The unsystematic risk, which is also called the diversifiable risk, is not
rewarded by the market. In fact, it can be eliminated by constructing diversified portfolios. The
correct measure of risk for an individual asset is therefore the beta, and its reward is called the
risk premium. The asset betas can be aggregated: the beta of a portfolio is obtained as a linear
combination of the betas of the assets that make up the portfolio. According to the CAPM, the
diversifiable risk component of each security is zero at equilibrium…6”

28.4 C. All investors make the same forecasts concerning the assets
Investor homogeneity is arguably the most important, and unrealistic (!), assumption.

Here is the very crux of CAPM and its Achilles’ heel: “The demonstration of the CAPM is
based on the efficiency of the market portfolio at equilibrium. This efficiency is a consequence of
the assumption that all investors make the same forecasts concerning the assets. They all
construct the same efficient frontier of risky assets and choose to invest only in the efficient
portfolios on this frontier. Since the market is the aggregation of the individual investors’
portfolios, i.e. a set of efficient portfolios, the market portfolio is efficient. In the absence of this
assumption of homogeneous investor forecasts, we are no longer assured of the efficiency of
the market portfolio, and consequently of the validity of the equilibrium model. The theory of
market efficiency is therefore closely linked to that of the CAPM. It is not possible to test the
validity of one without the other. This problem constitutes an important point in Roll’s criticism of
the model. We will come back to this in more detail at the end of the chapter.6”

Discuss in forum here: https://www.bionicturtle.com/forum/threads/l1-t1-28-capital-asset-


pricing-model-capm.3453/

6
Noel Amenc and Veronique Le Sourd, Portfolio Theory and Performance Analysis (West Sussex, England: John
Wiley & Sons, 2003)

27
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Related Bionic Turtle YouTube Videos


 Capital market line (CML) versus Security market line (SML), 12-min video:
https://youtu.be/m3kBczlC84c
 RAPMs: Treynor, Jensen's, Sharpe, 13-min video: https://youtu.be/V1RoJMoaHw4
 Capital asset pricing model, 15-min video: https://youtu.be/2QtAa2nmS0g
 Information ratio, 12-min video: https://youtu.be/fZmuJ2A9TC8
 Downside risk measures: semi-deviation, downside deviation, and Sortino ratio, 13-min
video: https://youtu.be/wEz9Zfzx9Bs

28
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

APPENDIX (OPTIONAL):
Please note: Elton & Gruber’s Chapter 5 (“Delineating Efficient Portfolios”) was
previously assigned in the FRM but removed in 2014. However, we have retained these
practice questions as the concepts continue to be highly relevant. Elton & Gruber’s
Chapter 14 (Nonstandard CAPM), on the other hand, we have relegated to the Appendix:
it was also previously assigned but removed in 2014; however, we do not consider them
highly relevant.

Elton & Gruber, Chapters 5: Delineating Efficient Portfolios


P1.T1.57. PORTFOLIO RETURN AND VOLATILITY
P1.T1.59. MVP AND THE EFFICIENT FRONTIER

P1.T1.57. Portfolio return and volatility


AIM: Calculate the expected return and volatility of a portfolio of risky assets

57.1. A portfolio is invested equally in two asset classes: bonds with expected return of 3.0%
per annum and volatility of 18.0%; equities with expected return of 7.0% per annum and
volatility of 26.0%. Their correlation is 0.40. If the portfolio re-allocates from equally weighting to
60% equities and 40% bonds, what is the net change to the portfolio's expected return?
a) No change to portfolio's expected return
b) Increase of 0.4%
c) Increase of 0.8%
d) Increase of 1.2%

57.2. A portfolio is invested equally in two asset classes: 50% in bonds with expected return of
4.0% per annum and volatility of 20.0%; equities with expected return of 9.0% per annum and
volatility of 32.0%. If the portfolio's variance is 0.04520, what is the implied correlation (of
returns) between bonds and equities?
a) zero
b) 0.019
c) 0.300
d) 0.467

57.3. A portfolio is 40% invested in Colonel Motors stock which has a volatility of 18.0% per
annum and 60% invested in Separated Edison which has a volatility of 38.0% per annum. Their
correlation is 0.24. Assuming the mean daily return is zero, what is the 1-day 99% confident
delta-normal value at risk of the portfolio? i.e., the delta-normal VaR with a holding period of one
day and a confidence level of 99%. Assume 250 trading days per year.
a) 1.61%
b) 1.90%
c) 2.65%
d) 3.75%

29
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

57.1. B. Increase of 0.4%


Before: 50%*3%+50%*7% = 5.0%
After: 40%*3%+60%*7% = 5.4%.
Expected return increases by 0.4%.
Note that neither volatilities nor correlation impacts expected return.

57.2. C. 0.300
Since variance(portfolio) = X(A)^2*variance(A) + X(B)^2*variance(B) +
2*X(A)*X(B)*covariance(A,B), where X(A) is fraction of portfolio held in A, per Elton & Gruber:
2*X(A)*X(B)*covariance(A,B) = variance(portfolio) - X(A)^2*variance(A) - X(B)^2*variance(B);
covariance(A,B) = [variance(portfolio) - X(A)^2*variance(A) - X(B)^2*variance(B)] /
[2*X(A)*X(B)];
StdDev(A)*StdDev(B)*correlation(A,B) = [variance(portfolio) - X(A)^2*variance(A) -
X(B)^2*variance(B)] / [2*X(A)*X(B)];
correlation(A,B) = [variance(portfolio) - X(A)^2*variance(A) - X(B)^2*variance(B)] /
[2*X(A)*X(B)*StdDev(A)*StdDev(B)]; in this case,
correlation(A,B) = [0.04520 - 50%^2*20%^2 - 50%^2*32%^2] / [2*50%*50%*20%*32%] = 0.30

57.3. D. 3.75%
The annual portfolio volatility = SQRT(40%^2*18%^2 + 60%^2*38%^2 +
2*40%*60%*18%*38%*0.24) = 25.5044%;
The daily portfolio volatility = 25.5044% / SQRT(250) = 1.6130%.
99% VaR = 1.6130% * 2.33 = 3.752%

Discuss in forum here: http://www.bionicturtle.com/forum/threads/p1-t1-57-portfolio-return-


and-volatility.5213/

30
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.59. MVP and the efficient frontier


AIMs: Define the minimum variance portfolio. Define the efficient frontier and describe
the impact on it of various assumptions concerning short sales and borrowing (Elton &
Gruber)

59.1. A portfolio contains only two assets. The first asset (a) has volatility of 9.0% and the
second asset (b) has volatility of 16.0%. If the assets are uncorrelated, what is nearest to the
weight in the first asset (a) in the minimum variance portfolio?

a) 43.75%
b) 56.25%
c) 75.96%
d) 100.0%

59.2. The market portfolio (M) contains the optimal allocation of only risky assets and no risky
assets. Let the (S1) be the Sharpe ratio of this market portfolio. There exists a risk-free asset.
Initially, an investor is fully (100%) invested in (M) with a portfolio Sharpe ratio of (S1).
Subsequently, the investor borrows 30% at the risk-free rate, such that she is 130% invested in
the market portfolio (M) where this leverage portfolio has a Sharpe ratio of (S2). After the
leverage (i.e., borrowing at the riskfree rate to invest +30% in M), is the investor still on the
efficient frontier and how do the Sharpe ratios?

a) No (no longer efficient), and S2 < S1


b) No, but S1 = S2
c) Yes (still efficient), but S2 < S1
d) Yes and S2 = S1

59.3. According to Elton & Gruber, which of the following statements is true?

a) The nominal returns of U.S. Treasury bills are risk-free returns


b) Predicted variance is always greater than historical variance
c) When using historical data to determine expected return inputs into a mean-variance
portfolio optimization model, the longest possible time frame is best
d) Treasury bill returns tend to be positively autocorrelated and this implies that T-bills are
effectively a decreasingly risky asset as the investment time horizon grows

31
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

59.1. C. 75.96%

 In this case of zero correlation, the weight allocated to asset (A) in the minimum variance
portfolio = 16%^2/(16%^2 + 9%^2) = 75.9644%.
 We can derive this by taking the first partial derivative:
variance(portfolio) = w(a)^2*sigma(a)^2 + w(b)^2*sigma(b)^2 + 0;
variance(portfolio) = w(a)^2*sigma(a)^2 + [1 - w(a)]^2*sigma(b)^2 + 0;
d variance(p) / d w(a) = 2*[w(a)*sigma(a)^2 - sigma(b)^2 + w(a)*sigma(b)^2], setting
equal to zero and solving for w(a):
0 = 2*[w(a)*sigma(a)^2 - sigma(b)^2 + w(a)*sigma(b)^2];
0 = w(a)*sigma(a)^2 - sigma(b)^2 + w(a)*sigma(b)^2;
w(a) = sigma(b)^2/[sigma(a)^2 + sigma(b)^2].

59.2. D. Yes and S2 = S1

 The ability to borrowing or lend morphs the concave/convex efficient frontier into the
linear CML; i.e., the leveraged portfolio is efficient with higher risk and higher return.
 All portfolios on the CML have the same Sharpe ratio: the slope of the CML.

59.3. B. "Predicted variance is always greater than historical variance because of


uncertainty as to the future mean.7"
 In regard to (A), this is false due to inflation risk.
 In regard to (C), this is false
 In regard to (D), the opposite: positive autocorrelation increases volatility (above the
square root of time)

Discuss in forum here: http://www.bionicturtle.com/forum/threads/p1-t1-59-mvp-and-the-


efficient-frontier.5241/

7
Edwin J. Elton, Martin J. Gruber, Stephen J. Brown, William N. Goetzmann, Modern Portfolio Theory and
Investment Analysis (John Wiley & Sons, Nov 16, 2009)

32
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Elton & Gruber, Chapter 14: Nonstandard Forms of Capital


Asset Pricing Models
P1.T1.62. NONSTANDARD CAPITAL ASSET PRICING MODEL (NONSTANDARD CAPM)
P1.T1.63. ARBITRAGE PRICING MODEL (APT)

P1.T1.62. Nonstandard capital asset pricing model (nonstandard


CAPM)
AIMs: Describe the impact on the CAPM of the following: Short sales disallowed;
Riskless lending and borrowing; Personal taxes; Nonmarketable assets; Heterogeneous
expectations; Non-price-taking behavior. Describe the following multi-period versions of
CAPM: Consumption-oriented CAPM; CAPM including inflation; Multi-beta CAPM

62.1. The standard CAPM makes several unrealistic assumptions (ten, according to Elton!) in
order to prove a general equilibrium relationship. The financial literature includes many analyses
of an equilibrium relationship in the event that some of these restrictive assumptions are relaxed
or violated, including: disallowing short sales; inability to borrow at the riskless rate; the
introduction of capital gains and income (divided) tax rates; the presence of nonmarketable
assets; and heterogeneous expectations. Which of the following statements most nearly
summarizes the viability of relaxing the assumptions of the standard CAPM?

a) The CAPM is "simply not robust" to any relaxation of assumption; even weakly,
equilibrium requires all assumptions
b) The CAPM is "restrictively robust" to the introduction of personal taxes, but equilibrium
"generally fails" if either short sales are disallowed or investors cannot borrow at the
riskfree rate
c) The CAPM is "remarkably robust" (an general equilibrium relationship exists) to the
relaxation (or violation) of any ONE of these assumptions
d) The CAPM is "remarkably robust" (an general equilibrium relationship exists) to the
simultaneous relaxation (and violation) of ALL assumptions

33
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

62.2. One assumption of the standard CAPM is unlimited lending and borrowing at the riskless
rate: "The investor can lend or borrow any amount of funds desired at a rate of interest equal to
the rate for riskless securities.8" If this assumption is eliminated, such that investor can neither
lend nor borrow at the riskfree rate, Elton shows that the zero-beta CAPM follows. Each of the
following is TRUE about the zero-beta CAPM EXCEPT:

a) Zero-beta CAPM still implies a security market line (SML), although the slope and
intercept are different
b) Zero-beta CAPM still has all investors holding the same portfolio in equilibrium
c) In the Zero-beta CAPM the market portfolio is still efficient
d) Although there is no riskless asset in the zero-beta CAPM, theoretically a riskless
portfolio (zero standard deviation) is possible

62.3. The consumption-oriented CAPM is directly analogous to the simple form of the CAPM.
But which of the following most nearly summarizes the key difference?

a) The consumption-oriented CAPM replaces the single common factor, switching the
CAPM's excess return on the market portfolio with the growth rate in per capita
consumption
b) The consumption-oriented CAPM adds an additional factor such that its two common
factors include both the market risk premium (market's price of risk) and per capita
consumption growth
c) The consumption-oriented CAPM exhibits a non-linear relationship between expected
return and per capita consumption growth
d) It is false that the consumption-oriented CAPM is analogous to CAPM

8
Edwin J. Elton, Martin J. Gruber, Stephen J. Brown, William N. Goetzmann, Modern Portfolio Theory and
Investment Analysis (John Wiley & Sons, Nov 16, 2009)

34
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

62.1. C. The CAPM is remarkably robust (an general equilibrium relationship exists) to
the relaxation (violation) of any ONE of these assumptions
 As the assumptions are relaxed/violated, the reading generally shows how an alternative
equilibrium model can be derived. However, this is not the case simultaneously.
 In regard to (A) and (B), please note the reading makes a point to demonstrate CAPM
is robust to disallowing short sales and the absence of riskfree borrowing/lending.

62.2. B. All investors no longer hold the same portfolio in equilibrium.


 In regard to (A), (C) and (D), each is TRUE.
 In regard to (D), the ability to SHORT allows investors to produce riskless portfolios
without the riskfree asset.

62.3. A. The consumption-oriented CAPM replaces the single common factor, switching
the CAPM's excess return on the market portfolio with the growth rate in per capita
consumption

 In regard to (B), CCAPM still assumes single common factor exposure (a limitation, to
be sure).
 In regard to (C), CCAPM posits a linear relationship

Discuss in forum here: http://www.bionicturtle.com/forum/threads/p1-t1-62-nonstandard-


capital-asset-pricing-model-nonstandard-capm.5291/

35
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.63. Arbitrage pricing model (APT)


AIMs: Describe the APT and the assumptions underlying it. Use the APT to calculate the
expected returns on an asset.

63.1. Each of the following is an assumption of the arbitrage pricing model (APM) EXCEPT for:

a) Homogeneous expectations
b) A security (stock) is linearly related to a set of indexes (factors); i.e., linearity assumption
c) Investors utilize a mean-variance framework
d) Error terms are uncorrelated; i.e., E(e(i),e(j)] = 0

63.2. Your colleague Robert uses a two-factor model in order to estimate the volatility of a
Portfolio. He specifies the covariance matrix as follows:

The Portfolio has the following factor sensitivities (i.e., betas): 0.60 to the Global Equity Factor
and 0.25 to the Global Bond Factor. The volatility of the Portfolio is nearest to which value?

a) 16.44%
b) 18.60%
c) 21.15%
d) 25.30%

36
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

63.3. In order to determine the covariance between Markets A and B, you develop the following
factor covariance matrix using a two-factor model:

Suppose that Market A exhibits the following factor sensitivities: 0.80 to the GDP Factor and
0.10 to the Interest Rate factor. Market B exhibits the following factor sensitivities: 0.70 to GDP
and 0.20 to Interest Rates. Which value is nearest to the covariance between Market A and
Market B? (note: this is a variation on GARP's actual sample question in 2011 Practice Exam,
Part 1, Exam 1, Question 6).

a) 0.0028
b) 0.0075
c) 0.0164
d) 0.1259

37
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

63.1. C. In APT, the CAPM assumption that investors utilize a mean-variance framework
is replace by an assumption of the process generating security returns.

63.2. D. 25.30%

 Variance(Portfolio) = 0.60^2*0.090 + 0.25^2*0.160 + 2*0.60*0.25*0.0720 = 0.0640


 Volatility (Portfolio) = SQRT(0.0640) = 25.30%

63.3. C. 0.0164

 Cov(Market A, Market B) = Cov(0.80*GDP + 0.10*I, 0.70*GDP + 0.20*I) =


(0.80)(0.70)*Variance(GDP) + (0.10)(0.20)*Variance(I) + (0.80)(0.20)*Cov(GDP,I) +
(0.10)(0.70)*Cov(I,GDP)
= (0.80)(0.70)*Variance(GDP) + (0.10)(0.20)*Variance(I) + [(0.80)(0.20) +
(0.10)(0.70)]*Cov(GDP,I)
= (0.80)(0.70)*0.03240 + (0.10)(0.20)*0.01440 + [(0.80)(0.20) + (0.10)(0.70)]*-0.00864
= 0.016445

Discuss in forum here: http://www.bionicturtle.com/forum/threads/p1-t1-63-arbitrage-pricing-


model-apt.5306/

38
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Grinold & Kahn, Active Portfolio Management: A Quantitative


Approach for Producing Superior Returns & Controlling Risk,
Chapter 7: Expected Returns and the Arbitrage Pricing
Theory
P1.T1.34. ARBITRAGE PRICING THEORY (APT)
P1.T1.35. APT FORECAST RETURNS
P1.T1.36. APT VERSUS CAPM
P1.T1.38. APT FACTOR FORECAST
P1.T1.39. STATISTICAL VERSUS STRUCTURAL APT MODELS

P1.T1.34. Arbitrage pricing theory (APT)


AIM: Define and describe the components of the Arbitrage Pricing Theory (APT) model

34.1 Which component is NOT in the APT model (assigned Grinold)?

a) Factor exposure
b) Factor return
c) Factor correlations
d) Specific (idiosyncratic) return

34.2 Is each of the following statements about APT true or false?

a) True or false: APT is a model of expected returns


b) True or false: APT is appropriate for consensus expected returns
c) True or false: APT factors can be fundamental (e.g., ROE), technical (e.g., liquidity),
macroeconomic (e.g., GDP), firm-specific (e.g., size) or statistical
d) True or false: The factor exposures must sum to 1.0
e) True or false: APT theory includes specification of the factors
f) True or false: CAPM is a special case of APT

39
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

34.1. C
 APT gives expected return (first moment) as a linear combination of factors in which
correlation does not enter; rather, correlation impacts variance/volatility (second
moment).

34.2 True

a) True
b) False. Grinold: "The flexibility of the APT makes it INAPPROPRIATE as a model for
consensus expected returns, but an APPROPRIATE model for a manager's expected
returns.9"
c) True. The APT is flexible in regard to factors. Says Grinold, "the theory doesn't say what
the factors are, how to calculate a stock's exposure to the factors, or what the weights
should be in the linear combination. This is where science steps out and art steps in.9 "
d) False. Although the factors exposures are "weights" on the factor forecasts, they will not
sum to 1.0 exceed by unlikely coincidence
e) False. Consistent with (c) above
f) True (!). CAPM is APT but with only one factor (factor = excess overall market return).

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-34-arbitrage-pricing-


theory-apt.3470/

9
Richard Grinold and Ronald Kahn, Active Portfolio Management: A Quantitative Approach for Producing Superior
Returns and Controlling Risk, 2nd Edition (New York: McGraw‐Hill, 1999).

40
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.35. APT forecast returns


AIM: Calculate a security’s expected excess returns using the APT model and interpret
the results.

Assume the following set of stocks (adapted from Grinold10 Tables 7.1 and 7.2).

Please note: Two industry classifications: expected excess return is 8.0% for the chemical
industry and 6.0% for all other industries. The factor forecasts are (top row): 2.0% for Growth,
2.5% for Bond, -1.5% for Size, and 0.0% for ROE. The CAPM excess market return (aka, equity
risk premium) is 6.0%.

35.1 What is the APT forecast for AT&T?


a) 4.80%
b) 5.35%
c) 6.00%
d) 6.25%

35.2 What is the APT forecast for Dow?


a) 3.80%
b) 4.60%
c) 6.60%
d) 7.40%

35.3 What is the CAPM forecast for AT&T?


a) 4.80%
b) 5.35%
c) 6.00%
d) 6.25%

35.4 What is the CAPM forecast for Dow?


a) 3.80%
b) 4.60%
c) 6.60%
d) 7.40%

10
Richard Grinold and Ronald Kahn, Active Portfolio Management: A Quantitative Approach for Producing Superior
Returns and Controlling Risk, 2nd Edition (New York: McGraw‐Hill, 1999).

41
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

35.1 B.
6% industry + (-0.20 * 2.0%) + (0.80 * 2.5%) + (1.5 * -1.5%) + (-0.60 * 0%) = 5.35%

35.2 A
8% industry + (-0.60 * 2.0%) + (-0.90 * 2.5%) + (0.50 * -1.5%) + (0.20 * 0%) = 3.80%

35.3 A
0.8 beta * 6.0% ERP = 4.80% excess return

35.4 C.
1.1 beta * 6.0% ERP = 6.60% excess return

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-35-apt-forecast-


returns.3473/

42
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.36. APT versus CAPM


AIM: Discuss the relationship between APT and the CAPM.

36.1 How is APT different from CAPM?


a) APT is a linear model
b) APT is a factor model
c) APT cannot use a market-related factor (CAPM beta is a market factor)
d) APT can use a small group of securities

36.2 Beta in CAPM is a single-factor analog to the multiple factor exposures (a.k.a., factor
loadings) in APT. What is, respectively, an average beta and an average APT factor exposure
per Grinold’s approach?
a) 0.0 (beta) and 0.0 (factor loading)
b) 0.0 and 1.0
c) 1.0 and 0.0
d) 1.0 and 1.0

43
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

36.1 D.
 Both are linear factor models. In regard, to (C), APT can use market-related, macro,
fundamental, firm-specific, and/or statistical factors. But (D) is a key difference: APT
does not require that a market portfolio of all risky assets.
 ... Recall this requirement is a critical weakness of CAPM Grinold: “The APT is
marvelously flexible. We can concentrate on any desired group of stocks. Among any
group of N stocks, there will be an efficient frontier for portfolios made up of the N risky
stocks11.”

36.2 C. beta = 1.0 is the “perfect” sensitivity to the overall market.


 But the APT factor exposures are STANDARDIZED with mean of zero. Grinold: “It is
easier to make comparisons across factors if they are all expressed in the same fashion.
One way to do this is to standardize the exposure by subtracting the market average
from each attribute and dividing by the standard deviation. In that way, the average
exposure will equal zero, roughly 66% of the stocks will have exposures running from –1
up to +1, and only 5% of the stocks will have exposures above +2 or below –211.”
 ... viewed another way: in CAPM, only the riskfree asset (beta = 0.0) escapes common
factor exposure, but the APT has several factors; in Grinold’s example11, the securities
are all immediately risky due to the industry exposure (itself implicitly a beta of 1.0)

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-36-apt-versus-


capm.3477/

11
Richard Grinold and Ronald Kahn, Active Portfolio Management: A Quantitative Approach for Producing Superior
Returns and Controlling Risk, 2nd Edition (New York: McGraw‐Hill, 1999).

44
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.38. APT factor forecast


AIM: Describe the difficulties involved with factor forecasting.

38.1 Assume a manager can forecast expected returns on 40 stocks but she can only forecast
10 APT factors. How much more skill in forecasting the factors is required, relative to skill in
forecasting the stocks, in order to produce the same value add?

a) 1x (same)
b) 2.0x (double)
c) 3.0x
d) 4.0x

38.2 According to Grinold, which type of APT factor is easiest to develop:

a) fundamental, stationary
b) fundamental, non-stationary
c) statistical, stationary
d) statistical, non-stationary

45
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

38.1 B.
Value add = Function [IC^2 * BR] = Function [skill^2 * breadth]
In this case, equivalent value-add is when:
40 breadth * skill ^2 = 10 breadth * (m * skill) ^ 2, such that
40b * s^2 = 10b * m^2 * s^2, and
40b = 10b * m^2,
m = SQRT(4) = 2.0

38.2 A.
Stationary makes the use of backcast more reliable: "The simplest approach to forecasting mk
[the factor] is to calculate a history of factor returns and take their average. This is the forecast
that would have worked best in the past—i.e., a backcast rather than a forecast. If we hope that
the past average helps in the future, we are implicitly assuming an element of stationarity in the
market.12"...

In regard to fundamental versus statistical, "Factor forecasts are easier if there is some
explicit link between the factors and our intuition. Consider, for example, a "bond market beta"
factor. This factor will show how the stock will react as bond prices (i.e., interest rates) change.
A forecast for this factor is in fact a forecast for the bond market. This doesn't mean that
forecasting future interest rates is easy. It means that the knowledge that you are, in fact,
forecasting interest rates should make the task clearer.12 "

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-38-apt-factor-


forecast.3484/

12
Richard Grinold and Ronald Kahn, Active Portfolio Management: A Quantitative Approach for Producing Superior
Returns and Controlling Risk, 2nd Edition (New York: McGraw‐Hill, 1999).

46
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

P1.T1.39. Statistical versus structural APT models


AIM: Describe some of the methods typically used in factor forecasting. Describe and
compare the attributes of purely statistical and structural APT models.

39.1 Which of the following type of factors are most likely to appear in a principal component
analysis (PCA)?

a) Macroeconomic
b) Fundamental (e.g., EPS growth, ROE)
c) Market-related (e.g., industry, beta)
d) None of the above

39.2 Which of the following is a statistical APT model?

a) MLE
b) BARRA
c) Fama-French three-factor model
d) CAPM

39.3 Assume a BARRA model that starts with the identification of fundamental or market
FACTOR EXPOSURES (e.g., industry, size). What is the most likely next step in the
construction of an APT model?

a) PCA to estimate factor returns


b) MLE to estimate factor returns
c) Linear regression to estimate factor returns
d) Linear regression to estimate alphas

47
Licensed to Christian Rey Magtibay at christianrey_magtibay@yahoo.com. Downloaded August 21, 2021.
The information provided in this document is intended solely for you. Please do not freely distribute.

Answers:

39.1 D. The others are all “structural.” PCA is a statistical model in which the factors have no
interpretation.

39.2 A. (Maximum Likelihood Estimation, along with PCA, is statistical)

39.3 C.

Grinold on Structural Model 113: “Given Exposures, Estimate Factor Returns” The BARRA
model, described in detail in Chap. 3, “Risk,” can function as an APT model. As we saw above,
it qualifies with ease, and it is just as easy (i.e., not very) to forecast returns to the BARRA
factors as to any others. The BARRA model takes the factor exposures as given based on
current characteristics of the stocks, such as their earnings yield and relative size. The factor
returns are estimates.”... Per the rest of Grinold, multivariate linear regression is the most likely
approach to estimating the factor returns13.

Discuss in forum here: http://www.bionicturtle.com/forum/threads/l1-t1-39-statistical-versus-


structural-apt-models.3487/

13
Richard Grinold and Ronald Kahn, Active Portfolio Management: A Quantitative Approach for Producing Superior
Returns and Controlling Risk, 2nd Edition (New York: McGraw‐Hill, 1999).

48

You might also like