Investment Science

Chapter 3
Dr. James A. Tzitzouris
<jimt2@ams.jhu.edu>

3.1

Use
rP
A= 1
1− (1+r)n

with r = 7/12 = 0.58%, P = $25, 000, and n = 7 × 12 = 84, to obtain A = $377.32.

3.2

Observe that since the net present value of X is P , the cash flow stream arrived at by cycling X is
equivalent to one obtained by receiving payment of P every n + 1 periods (since k = 0, . . . , n). Let
d = 1/(1 + r). Then
X∞
P∞ = P (dn+1 )k .
k=0

Solving explicitly for the geometric series, we have that

P
P∞ = .
1 − dn+1
Denoting the annual worth by A, we must have
rP
A= ,
1 − dn
so that solving for P as a function of P∞ and substituting the result into the equation for A, we
arrive at
1 − dn+1
 
A=r P∞ .
1 − dn

1
 Investment Science
Chapter 4
Solutions to Suggested Problems
Dr. James A. Tzitzouris
<jimt2@ams.jhu.edu>

4.1

(One forward rate)

(1 + s2 )2 1.0692
f1,2 = −1= − 1 = 7.5%
(1 + s1 ) 1.063

4.2

(Spot Update)
Use 1/(k−1)
(1 + sk )k

f1,k = −1
1 + s1
.
Hence, for example,
1/5
(1.061)6

f1,k = − 1 = 6.32%
1.05
.
All values are
f1,2 f1,3 f1,4 f1,5 f1,6
5.60 5.90 6.07 6.25 6.32

1
4.3

(Construction of a zero)
Use a combination of the two bonds: let x be the number of 9% bonds, and y teh number of 7%
bonds. Select x and y to satisfy

9x + 7y = 0,
x + y = 1.

The first equation makes the net coupon zero. The second makes the face value equal to 100. These
equations give x = −3.5, and y = 4.5, respectively. The price is P = −3.5 × 101.00 + 4.5 × 93.20 =
65.90.
4.5

(Instantaneous rates)
s(t2 )t2 −s(t1 )t1
(a) es(t2 )t2 = es(t1 )t1 eft1 ,t2 (t2 −t1 ) =⇒ ft1 ,t2 = t2 −t1

s(t)t−s(t1 )t1 d[s(t)t]

(b) r(t) = limt→t1 t−t1 = dt = s(t) + s0 (t)

(c) We have

d(ln x(t)) = r(t)dt,

= s(t)dt + s0 (t)dt,
= d[s(t)t].

Hence,

ln x(t) = ln x(0) + s(t)t,

and finally that

x(t) = x(0)es(t)t .

This is in agreement with the invariance property of expectation dynamics. Investing continuously
give the same result as investing in a bond that matures at time t.

4.6

(Discount conversion)

2
The discount factors are found by successive multiplication. For example,

d0,2 = d0,1 d1,2 = 0.950 × 0.940 = 0.893.

The complete set is 0.950, 0.893, 0.770, 0.707, 0.646.

4.7

(Bond taxes)
Let t be the tax rate, xi be the number of bond i purchased, ci be the coupon of bond i, pi be the
price of bond i. To create a zero coupon bond, we require, first, that the after tax coupons match.
Hence

x1 (1 − t)c1 + x2 (1 − t)c2 = 0,

which reduces to

x1 c1 + x2 c2 = 0.

Next, we require that the after tax final cash flows match. Hence

p0 = x1 p1 + x2 p2 .

Using this last relation in the equationfor final cash flow, we find

x1 + x2 = 1.

Combining these equations, we find that

c2 p1 − c1 p2
p0 = .
c2 − c1
After plugging in the given values, we find that

p0 = 37.64.

4.8

(Real zeros)
We assume that with coupon bonds there is a capital gains tax at maturity. We replicate the zero-
coupon bond’s after-tax cash flows using bonds 1 and 2. Let xi be the amount of bond i required

3
4.11
(Running PV examples)

d0,1 d0,2 d0,3 d0,4 d0,5 d0,6

(a) =⇒ N P V = 9.497
0.9524 0.9018 0.8492 0.7981 0.7472 0.7010

Year 0 1 2 34 5 6
Cash Flow −40 10 10 10 10 10 10
(b)
Discount 0.9524 0.9469 0.9416 0.9399 0.9362 9381
PV(n) 9.497 51.970 44.324 36.453 28.144 19.381 10.000

4.12

(Pure duration)

n
X n
X  −k
−k
P (λ) = xk (1 + sk /m) = xk (1 + s0k /m)eλ/m ,
k=0 k=0
n  
dP (λ) X −k −k−1
= xk (1 + s0k /m)eλ/m (1 + s0k /m)eλ/m ,
dλ m
k=0
n  
X −k
= xk (1 + sk /m)−k ,
m
k=0
Pn k

−k
1 dP (λ) k=0 xk m (1 + sk /m)
− = Pn −k
≡ D.
P (λ) dλ k=0 xk (1 + sk /m)

This D exactly corresponds to the original definition of duration as a cash flow weighted average
of the times of cash payments. No modification factor is needed even though we are working in
discrete time.

4.14

(Mortgage division)
(a)

(1 + r)k−1 − 1
   
k−1
P (k) = B − rM (k − 1) = B − r (1 + r) M (0) − B ,
r
= (1 + r)k−1 (B − rM ),

6
 Solution of HW3
Problem 6.1
X0 = outlay (in this case it is equal to the deposit.)
X1 = amount received, equal to the returned deposit plus the profit from shorting.

Thus,

X1= (X0 – X1) + X0

The total return, R = X1/X0 = ((X0 – X1) + X0)/X0

Thus,
R = 2X0 – X1
X0

Problem 6.2
Let a and b be the outcomes of two die rolls. Then Z=ab. By independence, we know:
2
⎛ 1+ 2 + 3 + 4 + 5 + 6 ⎞
E [ ab ] = E[a ]E[b] = ⎜ ⎟ = 3.5 = 12.25
2

⎝ 6 ⎠
var [ Z ] = E ⎡( ab ) ⎤ − ( E [ ab ]) = E ⎡( a ) ( b ) ⎤ − ( E [ a ] E [b ]) =
2 2 2 2 2

⎣ ⎦ ⎣ ⎦
2
⎛ 1 + 4 + 9 + 16 + 25 + 36 ⎞
= E ⎡( a ) ⎤ E ⎡( b ) ⎤ − ( E [ a ] ) ( E [ b ] ) = ⎜ ⎟ − ( 3.5 ) ≈ 79.97
2 2 2 2 4

⎣ ⎦ ⎣ ⎦ ⎝ 6 ⎠

Problem 6.3
Using the answer of the Problem 6.4:

a) α = 19 / 23

α 2σ 12 + (1 − α ) σ 22 = (19 / 23) ( 0.15) + (1 − 19 / 23) ( 0.3)

2 2 2 2 2
b) = 13.4%

c) r = α r1 + (1 − α ) r2 ≈ 11.4

Problem 6.4

Let α equal the percent of investment in stock 1, then the percent in stock 2 is (1-α). Now,
problem becomes Minimize σ2
 2
σ2 = ∑ω ω σ
i , j =1
i j ij

= ω12σ 12 + 2ω1ω 2σ 12 + ω 22σ 22

= α 2σ 12 + 2σ 12 * α * (1 − α ) + σ 22 (1 − α ) 2
= σ 12α 2 + 2σ 12α − 2σ 12α 2 + σ 22 − 2σ 22α + σ 22α 2 : take derivatives respect to α
dσ 2 / dα = 2σ 12α + 2σ 12 − 4σ 12α − 2σ 22 + 2σ 22α = 0
dσ 2 / dα = (2σ 12 − 4σ 12 + 2σ 22 )α + 2σ 12 − 2σ 22 = 0
⎛ σ 2 − σ 12 ⎞
α = ⎜⎜ 2 2 ⎟
2 ⎟
⎝ σ 1 − 2σ 12 + σ 2 ⎠

⎛ σ 2 − σ 12 ⎞
Thus percentage of asset 2 = 1-α = ⎜⎜ 2 1 ⎟⎟
⎝ σ 1 − 2σ 12 + σ 2
2
⎠

The mean rate of return of this portfolio is r = r1 * α + r2 * (1 − α )

r = r 1 [(σ 22 − σ 12 ) /(σ 12 − 2σ 12 + σ 22 )] + r2 [(σ 12 − σ 12 ) /(σ 12 − 2σ 12 + σ 22 )]

= [r 1 (σ 22 − σ 12 ) + r2 (σ 12 − σ 12 )] /(σ 12 − 2σ 12 + σ 22 )
⎛ r 1σ 22 + r 2σ 12 − (r 1 + r 2 )σ 12 ⎞
= ⎜⎜ ⎟
⎟
⎝ σ 2
1 − 2 σ 12 + σ 2
2
⎠

Problem 6.5
Money invested for concert 1 year from now= 1 million
Revenue expected unless it rains= 3 million
Chances of rain=50%
Rain Insurance = $ 0.5 per unit
Money obtained from insurance if it rains= $ 1 per unit
Let no. of units of insurance bought= u

a.)Now total money invested, X0 = money invested + no. of units of insurance bought
= (1*10^6+ 0.5*u)
Revenue obtained X0 = 3*10^6+ 1*u

Now since there are 50% chances of rain we have total revenue:-

X1= (3*10^ 6+u) *0.5

= (1.5*10^6+0.5u)

E (Total Return) = X1/X0=(0.5*3,000,000+0.5*u)/ (1,000,000+0.5u)

Now rate of return E(rate of return)= X1/X0-1=

= ((1.5*10^6+0.5u) - (1*10^6+ 0.5*u))/ (1*10^6+ 0.5*u)
= (500,000)/ (1,000,000+0.5*u)

b.) To get the minimum variance we will have to buy all the 3 million units
which will give us a variance of 0.
We have to Minimize Var(r)

Hence u= 3*10^ 6 units.

Var(r) = 0

Therefore r = (10^6(1.5-1))/ (1*10^6+0.5*3*10^6)

= (0.5)/(1+1.5)
= 0.2
= 20%

Problem 6.6
(A)

The efficient set (at minimum-variance point)

assets

(B)
Since assets are uncorrelated,
n
Var = ∑ ωi2σ i2
i =1
n
⎛ n ⎞
L = ∑ ωi2σ i2 − µ ⎜ ∑ ωi − 1⎟
i =1 ⎝ i =1 ⎠
∂L
= 2ωiσ i2 − µ = 0
∂ωi
µ = 2ωiσ i2
µ
ωi = 2
2σ i
Then,
n n
µ
∑ ω = ∑ 2σ
i =1
i
i =1
2
=1
i
n
1 2
∑σ
i =1
2
=
µ
i

2
µ= n
= 2σ 2
1
∑σ
i =1
2
i

Therefore,
σ2
ωi =
σ i2
n
σ4 n
1 n
Var = ∑ ω σ = ∑ 2 = σ 4 ∑ 2 2 2

i =1 σ i i =1 σ i
i i
i =1

Var = σ 2

Problem 6.7
Covariance matrix
2 1 0
1 2 1 V=
0 1 2
Expected rates of return

E(r1)= 0.4
E(r2)= 0.8
E(r3)= 0.8
a)
Minimum variance portfolio (consider that w1=w3 by symmetry)

Markowitz formulation:
n

∑σ w
j =1
ij j
− λr i − µ = 0

∑w r
i =1
i i
=r
(1.1)
n

∑w =1
i =1
i

w =w 1 3

For solving the system of equations (1.1) we omit the last normalizing constraint. Then,
we normalize the obtained solution {v1 , v2 , v3 } to the solution {w1 , w2 , w3 } that satisfies
the constraint.

System with 6 variables and 6 equations:

w1 w2 w3 λ µ E(r) =
2 1 0 -0.4 -1 0 0
1 2 1 -0.8 -1 0 0
0 1 2 -0.8 -1 0 0
0.4 0.8 0.8 0 0 -1 0
1 1 1 0 0 0 1
1 0 -1 0 0 0 0

By using calculator:
w1 w2 w3 λ µ E(r)
0.5 0 0.5 0 1 0.6

b)
λ=1, µ=0

System with 3 variables and 3 equations:

v1 v2 v3 =
2 1 0 0.4
1 2 1 0.8
0 1 2 0.8

By using calculator:
v1 v2 v3 w1 w2 w3 E(r)
0.1 0.2 0.3 0.167 0.333 0.5 0.733
c)
Given that there is a risky-free part, there is a single fund of risky assets such that any
efficient portfolio can be constructed as a combination of the fund and the risk-free
instrument. By using equation 6.10, we get a system of 3 equations with 3 variables.

∑
n
i =1
σ ki vi = rk − r f , k = 1,.., n
where r f = 0.2

v1 v2 v3 =
2 1 0 0.2
1 2 1 0.6
0 1 2 0.6

v1 v2 v3 w1 w2 w3
0 0.2 0.2 0 0.5 0.5

Problem 6.8

a)
n
Min Var (∑ α i ri − rM )
i =1
n
S .T . ∑α
i
i =1
n n n n
L = ∑∑ α iα jσ ij − 2∑ α iσ iM + σ M2 + µ (1 − ∑ α i )
i =1 j =1 i =1 i

∂L n
= 2∑ α jσ ij − 2σ iM − µ = 0 for each i
∂α i j =1

∂L n
= 1 − ∑αi = 0
∂µ i =1

So,
n
µ
∑α σ
j =1
j ij =
2
+ σ iM for each i

∑α
i =1
i =1
b)

n
Min Var (∑ α i ri − rM )
i =1
n
S .T . ∑α i
i =1
n ~
∑ αi ri = r
i
n n n n ~ n
L = ∑∑ α iα jσ ij − 2∑ α iσ iM + σ M2 + µ (1 − ∑ α i ) + λ (r − ∑ α i ri )
i =1 j =1 i =1 i i

∂L n
= 2∑ α jσ ij − 2σ iM − µ − λ ri = 0 for each i
∂α i j =1

∂L n
= 1 − ∑αi = 0
∂µ i =1

∂L ~ n
= r − ∑ α i ri = 0
∂λ i

So,
n
µ
∑α σ
j =1
j ij =
2
+ σ iM + λ r for each i

∑α
i =1
i =1
n ~
∑α r = r
i
i i

Problem 6.9
(Beeting Wheel) Consider a general betting wheel with n segments. The payoff for a $1
bet on a segment i is Ai. Suppose you bet an amount Bi = 1/Ai on segment i for each i.
Show that the amount you win is independent of the outcome of the wheel. What is the
risk-free rate of return for the wheel? Apply this to the wheel in Example 6.7

Sol:
Data:
n segments (from segment i =1 to segment n)
Amount received from segment i = Ai for each $1 invested
Fraction of segment i invested = Bi = 1/Ai (Wi)

a)
Expected return for segment i, Ri = Ai / $1

Amount to win if the outcome equal to segment i:

Expected return for segment i times the fraction invested in segment i

Ai 1
Ri * Wi ⇒ Ri * Bi ⇒ * = $1
1 Ai

For any outcome on the wheel the amount we will win will be $1.

Risk-free rate of return for the wheel:

For any outcome of the wheel, the amount that we will win will be = $1., X0 = $1
(amount received)
1
The amount invested will be X1= ∑
Ai

X0 $1
R= =
X1 1
∑ Ai
−1
⎛ 1 ⎞
r = R −1 = ⎜∑ ⎟ −1
⎝ Ai ⎠

Problem 6.10
n
Derive ∑σ
i =1
ki λwi = rk − rf .
Solution
n

∑ w (r − r
i i f )
tan θ = i =1
1/ 2
⎛ n ⎞
⎜ ∑ σ ij wi w j ⎟
⎜ ⎟
⎝ i , j =1 ⎠
 n

⎛ n ⎞
1/ 2
⎛ n ⎞
∑ σ kj w j
( rk − rf )⎜⎜ ∑ σ ij wi w j ⎟⎟ − ⎜ ∑ wi ( ri − rf ) ⎟
j =1
1/ 2
⎝ i , j =1 ⎠ ⎝ i =1 ⎠⎛ n ⎞
⎜ ∑ σ ij wi w j ⎟
∂ tan θ ⎜ ⎟
= ⎝ i , j =1 ⎠ = 0 , k = 1, n
∂wk n

∑ σ ij wi w j
i , j =1
n
Since ∑σ
i , j =1
ij wi w j = σ 2 > 0 , we have:

⎛ n ⎞ ⎛ n ⎞n
( rk − rf )⎜⎜ ∑ σ ij wi w j ⎟⎟ − ⎜ ∑ wi ( ri − rf ) ⎟∑ σ kj w j = 0
⎝ i , j =1 ⎠ ⎝ i =1 ⎠ j =1
⎛ n
⎞
⎜ ∑ wi ( ri − rf ) ⎟ n
⎝ i =1 ⎠ σ w = (r − r )
⎛ n
∑ kj j k f
⎞ j =1
⎜ ∑ σ ij wi w j ⎟
⎜ ⎟
⎝ i , j =1 ⎠
⎛ n ⎞
⎜ ∑ wi ( ri − rf ) ⎟
lets denote λ = ⎝ i =1 ⎠
⎛ n
⎞
⎜ ∑ σ ij wi w j ⎟
⎜ ⎟
⎝ i , j =1 ⎠
n

∑σ
i =1
ki λwi = rk − rf
Problem 7.1

7.1
Assume that the expected rate of return on the market portfolio is 23% and the rate of
return on T-bills(the risk-free rate) is 7%.The standard deviation of the market is
32%.Assume that the market portfolio is efficient
a) What is the equation of the capital market line?

formulation

expected rate of return =[ [ risk-free rate + (expected rate of return from market-risk free
rate)]/ standard deviation of market ] * standard deviation

expected rate of return from market = 0.23

risk-free rate = 0.07
standard deviation of market = 0.32

expected rate of return = 0.07+ (0.23-0.07)/0.32 *(standard deviation)

= 0.07+0.5 (standard deviation)
b) i) if expected return of 39% is desired, what is the standard deviation of this position?

expected return = 0.39

standard deviation position = 0.64

ii) If you have $1,000 to invest, how should you allocate it to achieve the above
position?
Expected return is 0.39 we can get 1000(1+0.39) = 1390
1390 = x (1+0.07) +(1000-x)(1+0.23)
x= -1000
risk free asset -1000
maket portfolio 2000

c) If you invest $300 in the risk-free asset and $700 in the market portfolio, how much money
should you expect to have at the end of the year?

300(1+0.07)+700*(1+0.23)=1182

Problem 7.2
A small world
Consider a world in which there are only two risky assets, A and B, and a risk free asset
F. Te two risky assets are in equal supply in the market; that is, M= ½ (A+B). The
following information is known:

rF = .10
2
A = .04
AB = .01
2
B = .02

M = .18

(a) Find a general expression for M2, A, B  

Since the market has only two risky assets A and B, then the expected rate of return and
the variance of the market depend solely on the expected rate of return and the variance
of the assets A and B.
2 2 2
M = 

A +2 (1- ) 

AB + (1- )2 

Since A and B are in equal amounts in the market  = 1- ( =0.5)  

2


A = [cov (rA,rM)]/ M cov(rA,rM) = cov (rA, rA + (1- )rB)  

= A2 + (1- ) AB  

2 2 2


A =[ 

A + (1- ) 

AB]/[ 

A +2 (1- )  

AB + (1- )2 

B ]

= [ A2 + AB]/ ( A2 +2 AB + 
2
B )=
=2 [ A2 + AB]/( A2 +2 AB + B
2

2


B = [cov (rB,rM)]/ M cov(rB,rM) = cov (rB, rA + (1- )rB)  

= B2 + (1- ) AB  

2 2 2


B =[ 

B + (1- ) 

AB]/[ 

A +2 (1- )
 

AB + (1- )2 

B ]

= [ B2 + AB]/ ( A2 + 2 AB + 
2
B )=
= 2[ B2 + AB]/( A2 + 2 AB + B
2

(b)According to CAPM, what are the numerical values of

A and B
i – rf = i ( M-rf) 

= A ( M- rf) + rf A 

=[ [ A + AB]/ ( A2 2 AB + B2)] ( M- rf) + rf

2

= [(.04+.01)/(.5(.04+2(.01)+.02))][.18-.10] + .10
= .2

B= 

B( M- rf) + rf
2 2 2

=[ [ B + AB]/  ( A 2 AB + B )] ( M- rf) + rf
 = [(.02+.01)/(.5(.04+2(.01)+.02))][.18-.10] + .10
= .16

Problem 7.3
7.3 (Bounds on returns) Consider a universe of just three securities. They have expected
rates of return of 10%, 20%, and 10%, respectively. Two portfolios are known to lie on
the minimum-variance set. They are defined by the portfolio weights w = [.60, .20, .20], v
= [.80, -.20, .40]. It’s also known that the market portfolio is efficient.
(a) Given this information, what are the minimum and maximum possible values for the
expected rate of return on the market portfolio?
(b) Now suppose you are told that w represents the minimum-variance portfolio. Does
this changer your answers to part (a)?
(a)
Market portfolio can’t contain negative amount of security.




 



    





       



!#"$&% ' 




( )*  ,+ -#.0/ #- .01 -#.01%-#.   2 - #34
( )*   ,+! -#.  %#- .  %-#. -#. 4 2 - # 2 -#.05
( )*  ,+6 -#.  -#. 4 -#. 4 %-#.   2 - # 3 
78 9     #- .05 3#3 

Expected rate of return on the market portfolio:

:;=<)  >  
:?   *  9
:
  :
 



   








   
 * 9    "  



!"$!%
' 



 '
( )*  ,+ ! -#@ -#.0/ #- .01 -#.0-1%-#. - 
( )*  ,+! -#@ #- .  % -#.  %-#.0- 4 -#.0-1 
( )*  ,+6 !-#@ #- .  -#. 4 -#.0- 4 %-#. - 
7

-#.0-1-#.0- 4

From rM = 0.08+0.04 and 0.5

3 3 2, we know 0.13 r 3 0.16.
 

M
(b)
If given portfolio1 is the minimum-variance portfolio, rate of return of portfolio1 is the
minimum rate of return of market portfolio.
Rate of return of portfolio1=0.06+0.04+0.02=0.12.
3
Therefore, the expected rate of return of market portfolio became 0.12 rM 0.16.
3

Problem 7.4
_
(Quick CAPM derivation) Derive the CAPM formula for rk − r f by using Equation (6.9)
in Chapter 6. [Hint: Note that
n
σ ik wi = cov(rk , rM ).]
i=n
Apply (6.9) both to asset k and to the market itself.

Solution

Equation 6-9 from the textbook

n _
σ ki λwi = rk − r f

k = 1,2,….n
i=n

We apply the above equation both to asset k and to market M. So we get

λ [σ i2 wi + σ iM wM ] = ri − rf and
_

λ [σ iM wi + σ M2 wM ] = rM − r f
_

From the hint:

[σ i
2
]
wi + σ iM wM = cov(ri , rM ) = σ iM and

[σ iM ]
wi + σ M2 wM = cov(rM , rM ) = σ M2

Substituting in above equations we get:

_
λσ iM = ri − r f
 _
λσ M2 = rM − r f

Eliminating λ from the above two equations and solving, we get:

_
rM − rf
λ= . Therefore,
σ M2
_
rM − r f _
σ iM
σ iM = ri − r f . From the textbook, = β i . Substituting, we obtain
σ 2
M σ M2

_ _
(rM − r f ) β i = ri − rf Which is the required capital asset pricing (CAPM) formula.

Problem 7.5
β j = σ jM / σ M2
σ M2 = σ i2 xi2 cov = 0
n
i =1

σ jM = x jσ 2j i contributes according to its weight

Therefore β j = x jσ 2j / σ i2 xi2
n
i =1

Problem 7.6
Problem 7.6 – Simpleland

In Simpleland, there are only two risky stocks, A and B, whose details are listed below:
- Number of shares outstanding:
o A: 100
o B: 150
- Price per share
o A: $1.50
o B: $2.00
- Expected rate of return
o A: 15%
o B: 12%
- Standard deviation of return
o A: 15%
o B: 9%
Furthermore, the correlation coefficient between the returns of stocks A and B is ab=1/3.

There is also a risk free asset, and Simpleland satisfies the CAPM exactly.
 a. Expected rate of return of the market portfolio

Market Capitalization
- Stock A: 100*1.50=150
- Stock B: 150*2.00=300
- Total: 450
We can deduce the respective weights in the market portfolio:
- Stock A: 1/3 (=150/450)
- Stock B: 2/3

So the expected rate of return of the market portfolio is:

E(rm)=0.15*1/3+0.12*2/3
E(rm)=13%

b. Standard deviation
2 2 2 2
Var(rm)= a * a + b * b +2* a* b* ab* a* b

Var(rm)=(1/3)2*0.152+(2/3) 2*0.092+2*1/3*2/3*1/3*0.15*0.09
Var(rm)=0.0081
(rm)=0.09

c. Beta of stock A
2
 a=Cov(ra,rm)/ m

And we know that: rm=1/3*ra+2/3*rb

Hence: Cov(ra,rm)= Cov(ra, 1/3*ra+2/3*rb)
Cov(ra,rm)= 1/3*Cov(ra,ra)+2/3*Cov(ra,rb)
Cov(ra,rm)= 1/3* a2+2/3* ab* a* b

Cov(ra,rm)= 1/3*0.152+2/3*1/3*0.15*0.09
Cov(ra,rm)=0.0105
So the Beta of stock A is:
a=0.0105/(0.09)2


a =1.296


d. Risk Free Asset in Simpleland

Relation (7.2) gives us: E(ra)-rf = a*(E(rm)- rf) 

So: rf =(E(ra)- a*E(rm)) / (1- a)  

rf =(0.15-1.296*0.13) / (1-1.296)
rf =6.25%

Problem 7.7
 (Zero-beta assets) Let w0 be the portfolio (weights) of risky assets corresponding
the minimum-variance point in the feasible region. Let w1 be any other portfolio on
the efficient frontier. Define r0 and r1 to be the corresponding returns.

a) There is a formula of the form 01= A 0². Find A. [Hint: Consider the portfolios
(1- )w0 + w1, and consider small variations of the variance of such portfolios near

=0.

Let w 0 and w 1 be two portfolios on the the efficient frontier. A third portfolio can be constructed based
on w 0 and w 1 .
σ 2 2 = (1 − α ) 2 σ 0 2 + 2α (1 − α )σ 01 + α 2σ 1 2
∂ (σ 2 )
2

∂α α = 0

∂ (σ 2 )
2

= −2(1 − α )σ 0 + (2 − 4α )σ 01 + 2ασ 1
2 2

∂α α = 0

∂ (σ 2 )
2

= −2(1 − 0)σ 0 + (2 − 4(0))σ 01 + 2(0)σ 1

2 2

∂α α = 0

∂ (σ 2 )
2

= −2σ 0 + 2σ 01
2

∂α α = 0

σ 0 2 = σ 01
σ 01 = Aσ 0 2
A =1

b) Corresponding to the portfolio w1 there is a portfolio wz on the minimum-

variance set that has zero beta with respect to w1; that is, 1,z=0. This portfolio can
be expressed as wz= (1- )w0 + w1. Find the proper value of .




Let w 0 be the minimum variance point portfolio. If σ 1z is zero, then the weighted combinatio n of w 0 and
σ z 2 = (1 − α ) 2 σ 0 2 + 2α (1 − α )σ 01 + α 2σ 1 2
σ 0 2 = (1 − β ) 2 σ z 2 + 2β (1 − β )σ z1 + β 2σ 1 2
By linear combinatio n of the three portfolios, it can be concluded that :
1 α
=1− β β =−
1-α 1− α
Since w 0 is the MVP, then :
∂σ 0
2

= 0 0 = −2(1 − β )σ z + (2 − 4 β )σ z1 + 2 βσ 1
2 2
Set :
∂β
Because σ z1 = 0
0 = −2(1 − β )σ z + 2 βσ 1
2 2

σ z2
β= 2
σ z + σ 12
2 2
σ 2


σ z2





σ0 2
= 1− 2 z 2


σz 2
+ 

σ 12
σ z + σ1 σ z 2 + σ 12
 

2 2
σ 12


σ z2





σ0 2
= 

σz 2
+ 

σ 12
σ z 2 + σ 12 σ z 2 + σ 12
 

σ 1 4σ z 2 + σ 1 2σ z 4
σ0 2
= σ 0 2 (σ z 2 + σ 1 2 ) 2 = σ 1 2σ z 2 (σ 12 + σ z 2 )
(σ z + σ 1 ) 2
2 2

σ 0 2σ 1 2
σz 2
= 2
σ1 − σ 02
σ 0 2σ 1 2 1 σ 0 2σ 1 2 σ 12 − σ 0 2
β= * = *
σ 12 − σ 0 2 σ 0 2σ 1 2 σ 1 2 − σ 0 2 σ 0 2σ 1 2 + σ 1 2 (σ 1 2 − σ 0 2 )
+ σ 2

σ 12 − σ 0 2
1

σ 02 α σ 02
β= 2 − = −ασ 1 = (1 − α )σ 0
2 2

σ1 1 − α σ 12
σ 02
α=
σ 0 2 − σ 12

c) Show the relation of the three portfolios on a diagram that includes the feasible
region.
r
w1
r 1

r 0
w0

wz
r z

σ σz σ σ
0 1

d) If there is no risk-free asset, it can be shown that other assets can be priced
according to the formula

r i − r z = β iM (r M − r z )

where the subscript M denotes the market portfolio and r z is the expected rate of
return on the portfolio that has zero beta with the market portfolio. Suppose that
the expected returns on the market and the zero-beta portfolio are 15% and 9%
respectively. Suppose that a stock i has a correlation coefficient with the market of
.5. Assume also that the standard deviation of the returns of the market and stock i
are 15% and 5% respectively. Find the expected return of stock i.

r M = 15%
r z = 9%
σ M = 15%
σ i = 5%
ρ iM = 0.5 

σ iM = 0.5 * 0.15 * 0.05 = 0.00375

r i − r z = β iM (r M − r z )


0.00375
r i = 0.09 + (0.15 − 0.09)


0.15 2

r i = 0.1 = 10%
Problem 7.8
Electron Wizards, Inc. has a new idea for producing TV sets, and it is planning to
enter the development stage. Once the product is developed (which will be at the end of 1
year), the company expects to sell its new process for a price p, with expected
value p = $24M . However, this sale price will depend on the market of TV sets at the
time. By examining the stick histories of various TV companies, it is determined that the
final sales price p is correlated wth the market return us E p − p rM − rM = $20Mσ Μ2 . [( )( )]
To develop the process, EWI must invest in a research and development project. The
cost c of the project will be known shortly after th project is begun. The current estimate
is that the cost will be either c=$20M or c=$16M, and each of these is equally likely.
(This uncertainty is uncorrelated with the final price and is also uncorrelated with the
market). Assume that the risk free rate is rf=9% and the expected return of the market
rM = 33% .
(a) What is the expected rate of return of this project?
(b) What is the beta of this project?
(p − p ) (r ) [( )( )]


1


M − rM =E E p − p rM − rM


Hint: E


c c

c) Is this an acceptable project based on a CAPM criterion? In particular what is the

excess rate of return (+ or -) above the predicted by the CAPM?

Solution:

p−c p


= − 1 E (r ) = E
p
a) r = − 1 . Due to the fact that p,c are independent we

 

c c c


have E (r ) = E E ( p ) − 1 = 0.05625 * 24 − 1 E (r ) = 0.35

1  

b)

[( )( )] ( p − c ) − ( p − c ) (r ) (p − p ) (r )
 

   

E r − r rM − rM = E − rM =E − rM
 

   

M M
c c c
 

[( )( )]


1
=E E p − p rM − rM = 1.125σ Μ2
 

σ rM 1.125σ M2
Then: β = 2 = = 1.125
σM σ M2

c) Based on the CAPM the expected return is:

 (
r − rf = β r M − rf ) ( )
r = β r M − r f + r f = 1.125(0.33 − 0.09) + 0.09 r = 0.36
The expected excess rate of return is:
r − r f = +0.27

Thus we conclude that, based on the CAPM model, the project is not acceptable since it
gives smaller return rate than CAPM. Nevertheless, the difference is only 0.01 therefore
the final decision should not be based only on the CAPM criterion.

Problem 7.9
Formulation:

(Gavin’s problem)

Prove to Gavin Jones that the results he obtained in egs. 7.5 and 7.7 were not accidents.
Specifically, for a fund with return rf + (1- ) rm show that both CAPM models give

the price of $ 100 worth of fund assets as $100.

We have to prove that the formulas, “certainty equivalent form of the CAPM” &
the “CAPM as a pricing formula” both will give the same results for pricing an
asset P by appropriately discounting the final return Q. we have to do this for the
case mentioned for Gavin Jones in egs. 7.5 & 7.7 of the book, i.e. for an asset
combination of two securities with weights and (1- ).  

The return for an asset mixture (Q) with above weights is given by

(please note that Q or rM with a bar on top is represented as Q or rM i.e., boldface)

Q=P(
rf + (1- ) rM + 1)………………………………………………………(1)

Also for the covariance we can write,

cov (Q, rM ) = cov ( P ( rf + (1- ) rM + 1), rM ) …………………………….…(2)

= P(1- ) 2 M



from certainty equivalent form of the CAPM, we have,

2
P = [1/ (1+ rf )] * [ Q- {cov (Q, rM )*( rM - rf ) / 

M}

Substituting (1) and (2) from above,

2 2
P = [1/ (1+ rf )]*[ P ( rf + (1- ) rM + 1) – {P(1- )





M )*( rM - rf ) / 

M }]

= [1/ (1+ rf )]* P [ ( rf + (1- ) rM + 1) – { (1- ) *( rM - rf )}]

expanding and canceling out common terms,

P = [1/ (1+ rf )]* (1+ rf )

= P.

hence the proof.

Problem 8.1
(A simple portfolio) Someone who believes that the collection of all stocks
satisfies a single-factor model with the market portfolio serving as the factor gives
you information on three stocks which makes up a portfolio. In addition, you know
that the market portfolio has an expected rate of return of 12% and a standard
deviation of 18%. The risk-free rate is 5%.
(a) What is the portfolio’s expected rate of return?
(b) Assuming the factor model is accurate, what is the standard deviation of this
rate of return?

Market
Rate of Standar
return deviation
12% 18% Risk free rate 5%

Stock Beta Standar Weight

deviation of in
random error porfolio
term
A 1.1 7.00% 20%
B 0.8 2.30% 50%
C 1 1.00% 30%

a) The equation for a single-factor model for stock returns is:

ri − rf = β i × (rm − r f )
So, solving we have:
R1 = 1.1*(0.12- 0.05)+0.05 = 0.127 = 12.7%
R2 =0.8*(0.12- 0.05)+0.05 = 0.160= 10.6%
R3 =1.0*(0.12- 0.05)+0.05 = 0.120 = 12.0%
Since Rportfolio =
n
rportfolio = ri × wi
i =1
Rportfolio = 0.127*.20+0.160*0.50+0.120*0.30 = 0.1144 = 11.44%

b) Standard Deviations:

We know that:
3 3
σ 2 = b 2 × σ 2f + σ e2 b=
wi × bi σ e2 =  wi2 × σ ei2
i =1 i =1
Solving, we have:
B2 = (0.20*1.1+0.50*0.8+0.30*1)^2 = 0.846
 2
M = 0.18^2 = 0.324 = 3.24%
2
e = 0.2^2*0.07^2+0.5^2*0.023^2+0.3^2*0.01^2 = 0.03%

Portfolio's standard deviation = (84.6%*3.24%+0.03%)^0.5 = 16.66%

Problem 8.2
Two stocks are believed to satisfy the two-factor model

r1 = a1 + 2f1 + f2

r2 = a1 + 3f1 + 4f2

In addition there is a risk-free asset with a rate of return on 10%. It is known that r-bar1 =

15% and r-bar2 = 20%. What are the values of 0, 1, and 2 for this model?

Since there is a risk-free asset with rate-of-return of 10%: 0 = .10

Using the relationship: if ri = ai +  bij fj then r-bari = 0 +  bij j

Yields:

.15 = .10 + 2 1 + 2

.20 = .10 + 3 1 +4 2

1 = .02 2 = .01

Problem 8.3
Suppose there are n random variables x1, x2, … xn and let V be the corresponding
covariance matrix. An eigen vector of V is a vector v = (v1, v2, … vn) such that Vv = λv
for some λ (called an eigenvalue of V). The random variable v1 x1 + v2 x2 + … + vnxn is a
principle component. The first principle component is the one corresponding to the
largest eigenvalue of V, the second to the second largest, and so forth.

A good candidate for the factor in a one-factor model of n asset returns, is the first
principle component extracted from the n returns themselves: that is, by using the
principle eigenvector of the covariance matrix of the return. Find the first principle
component for the data of example 8.2. Does this factor (when normalized) resemble the
return on the market portfolio? [Note: For this part, you need an eigenvector calculator as
available in most matrix operation packages.]
 From the example 8.2 we have the following data

Year Stock 1 Stock 2 Stock 3 Stock 4 Market Riskless

1 11.91 29.59 23.27 27.24 23 6.2

2 18.37 15.25 19.47 17.05 17.54 6.7
3 3.64 3.53 -6.58 10.2 2.7 6.4
4 24.37 17.67 15.08 20.26 19.34 5.8
5 30.42 12.74 16.24 19.84 19.81 5.9
6 -1.45 -2.56 -15.05 1.51 -4.39 5.2
7 20.11 25.46 17.8 12.24 18.9 4.9
8 9.28 6.92 18.82 16.12 12.78 5.5
9 17.63 9.73 3.05 22.93 13.34 6.1
10 15.71 25.09 16.94 3.49 15.31 5.84

Average 15.00 14.34 10.90 15.09 13.83 5.854

Var 90.26 107.23 162.20 68.25 72.12
Cov 65.09 73.62 100.78 48.99 72.12
β 0.9 1.02 1.4 0.68 1
α 1.95 0.34 -6.11 3.82 0
e-var 31.54 32.09 21.37 34.99

From the above first we calculate the averages, variance and finally the
covariance for the above data

In the covariance we have to divide the value that is obtained from the excel
solver by 9*10 to adjust the bias.

Using the above Covariance values the Covariance matrix is constructed

From the excel we get it as

1 2 3 4
1 90.26 50.89 79.02 40.18
2 50.89 107.23 105.38 30.99
3 79.02 105.38 162.20 56.54
4 40.18 30.99 56.54 68.25

Covariance matrix

From the above matrix we have to calculate the eigen values

 To find the eigen values we solve using Excel the following equation:
det (V − λ E ) = 0 .
This equation has several roots.
We are interested in the largest eigen value.
We get the largest eigen value = 311.16

The corresponding eigen vector is found to be

V = [0.217 0.263 0.360 0.153]

Since this is the normalized form we have the eigen vector as

V = 0.217r1 + 0.263r2 + 0.360r3 + 0.153r4

Problem 8.4
Let ri, for i = 1,2,…,n, be independent samples of a return r of mean µ and variance σ 2 .
Define the estimates
1 n
µ=
ˆ ri
n i =1

1 n
s2 = (ri − µˆ ) 2 .
n − 1 i =1
Show that E (s 2 ) = σ 2 .

Solution.

1 n
E (s 2 ) = E (ri − µˆ ) 2

 

n − 1 i =1


n
1
= (ri − µˆ ) 2

 

E (by properties of the expectation)

n − 1 i =1


n
1
= (ri 2 − 2ri µˆ + µˆ 2 )

E


n − 1 i =1


n
1
= ri 2 − 2nµˆ 2 + nµˆ 2

E


n − 1 i =1


n
1
= E (ri 2 ) − nE ( µˆ 2 )

(by the linearity of the expectation)



n − 1 i =1


Since E (Y 2 ) = V (Y ) + E (Y ) 2 , then
 σ2


n
1


E (s ) = (σ + µ ) − n +µ



2 2 2 2




n −1


n

i =1

σ2


1


= n(σ 2 + µ 2 ) − n + µ2



 

n −1


n

=σ .2

q.e.d.

Problem 8.5

a) Show that σ (rˆ ) is independent of n

r
rn =
n rˆ = nrˆn
and
σ2 σ 2 = nσ n2
σ n2 =
n

var(rˆ ) = var(nrˆn ) = n 2 * var(rˆn ) ……………………….(1)

σ n2 σ2
since var(rn ) = and σˆ = 2
n
n n

σ2 σ2
var(rn ) = =
n*n n2

σ2
From (1), var(rn ) = n 2 * 2
=σ2
n

It shows that σ (rˆ ) is independent of n

b) Show that how σ (σˆ 2 ) depends on “n”

var(σˆ 2 ) = var(n * σˆ n2 ) = n 2 * var(σˆ n2 )................................(2)

Assume that the return are normally distributed,

 2σ n4 2σ 4
var(σ n2 ) = = 2
n − 1 n * (n − 1)

2σ 4
Therefore, var(σˆ 2 ) =
n −1

2σ 2
Then, σ (σˆ ) =
2

(n − 1)

More data does not help to estimate the mean more precisely but it improves the
estimation of the volatility.

Problem 8.6
A record of annual percentage rates of return of the stock S is shown in the following table.

Record of Rates of Return:

Month Percent rateof return Month Percent rateof return
1 1.0 13 4.2
2 0.5 14 4.5
3 4.2 15 -2.5
4 -2.7 16 2.1
5 -2.0 17 -1.7
6 3.5 18 3.7
7 -3.1 19 3.2
8 4.1 20 -2.4
9 1.7 21 2.7
10 0.1 22 2.9
11 -2.4 23 -1.9
12 3.2 24 1.1

a) Estimate the arithmetic mean rate of return, expressed in percent per year.
b) Estimate the ariithmetic standard deviation of these returns, again as percent per year.
c) Estimate the accuracy of the estimates found in parts (a) and (b).
d) How do you think the answers to © would change is you had 2 years of weekly data
instead of monthly data? (See exercise 5.)

n = number of periods = 24
Period i
(month) ri
(r − r̂ )
i
2

1 1.0 4.93038E-32
2 0.5 0.25
3 4.2 10.24
4 -2.7 13.69
5 -2.0 9
6 3.5 6.25
7 -3.1 16.81
 8 4.1 9.61
9 1.7 0.49
10 0.1 0.81
11 -2.4 11.56
12 3.2 4.84
13 4.2 10.24
14 4.5 12.25
15 -2.5 12.25
16 2.1 1.21
17 -1.7 7.29
18 3.7 7.29
19 3.2 4.84
20 -2.4 11.56
21 2.7 2.89
22 2.9 3.61
23 -1.9 8.41
24 1.1 0.01
a) n
1 n = 24
rˆ = ri = =

( ri − rˆ ) 2
n i =1
1.0 165.4 i =1

rˆy = 12rˆ =
12.0%

b)
1 n = 24
s2 = ( ri − rˆ ) 2 =
n − 1 i =1


7.191304
sy = 9.289545%

c)
The accuracy of the estimators is determined by taking the following standard deviations:

Accuracy of the mean estimator

σ
σ rˆ = =
6.5687% n
where n = 2 since we are dealing with yearly standard deviations

Accuracy of the variance estimator

2σ 2
stdev( s ) monthly =
2
=
n −1 2.1206= 0.021206%
stdev( s ) yearly = 12 * stdev( s ) monthly =
2 2
25.4472=0.254472%
where n = 24 and σ is the monthly σ.

Since we are quaring the standard deviation with units as percent, then we need to multiply by 100 to get
the answer as a percent and not percent squared
d)
Exercise 8-5 proves that the accuracy of the mean estimator is independent of the number of periods, n. As
a result, the accuracy of the mean estimator would not change by using 2 years of weekly data instead of 2
years of monthly data. The accuracy of the standard deviation estimator however, is dependent on the
number of periods. The accuracy would increase with more periods to estimate the standard deviation.

Problem 8.7
Gavin Jones figured out a clever way to get 24 samples of monthly returns in just over
one year instead of only samples; he takes overlapping samples; that is, the first sample
covers Jan. 1 to Feb. 1 and the second sample covers Jan. 15 to Feb. 15, and so forth. He
figures that error in his estimate of r , the mean monthly return, will be reduces by this
method. Analyze Gavin’s idea. How does the variance of his estimate compare with that
of the usual method of using 12 nonoverlapping monthly returns?

Solution

1 n
E ( Head r ) = E ( ri ) = r
n i =1
Monthly rate of return=r
Half monthly rate of return =r/2
E (ri ) = r = 1 / 12ry
E ( ρ i ) = r / 2 = 1 / 24ry

V (ri ) = σ 2 = 1 / 12σ y2
V ( ρ i ) = 1 / 2σ 2 = 1 / 24σ y2

From the hint: ri = ρ i + ρ i +1

Cov (ri , ri +! ) = E (ri ri +1 ) − E (ri ) E (ri +1 )

= E[( ρ i + ρ i +1 )( ρ i +1 + ρ i + 2 )] − r * r
= E[ ρ i ρ i +1 ] + E[ ρ i ρ i + 2 ] + E[ ρ i +1 ]2 + E[( ρ i +1 )( ρ i + 2 )] − r * r

From the hint:

ρi are uncorrelated-
independent.
E ( ρ i ) = V ( ρ1 ) + E ( ρ i ) 2 = 1 / 2σ 2 + (r / 2) 2
2

From the formula above

= E[ ρ i ]E[ ρ i +1 ] + E[ ρ i ]E[ ρ i + 2 ] + E[ ρ i +1 ]2 + E[ ρ i +1 ]E[ ρ i + 2 ] − r 2
= r / 2 * r / 2 + r / 2 * r / 2 + σ 2 / 2 + ( r / 2) 2 + r / 2 * r / 2 − r 2
= 3* r 2 / 4 + r2 / 4 + σ 2 / 2 − r2
=σ2 /2

From the page 215 and the hint:

1 24 1 24
V ( Head r ) = V ( ri ) = 2 V ( ri )
24 i =1 24 i =1

1
= 2 [V (r1) + V (r 2) + .... + V (r 24) + 2 cov(r1, r 2) + 2 cov(r 2 + r 3) + ... + 2 cov(r 23 + r 24)]
24
1
= 2 [24σ 2 + 2 * 24(σ 2 / 2)] T
24
1
= σ2
12

he result is equal to page 215 (8.11)

Problem 8.8

We use the general model with p=Pr+e where P is an mxn matrix, r is an n-dimensional
vector, and P and e are m-dimensional vectors. The vector p is a set of observation
values and e is a vector of errors having mean 0. The error vector has covariance matrix
Q. Then the estimate r’ of r is

r’=inv(tr(P)inv(Q)P)tr(P)inv(Q)p

where inv is the inverse of a matrix and tr is the transpose of a matrix.

a) If there is a single asset an just one measurement of the form p=r+e, we must show
that r’=p.
Solution: Suppose that p=r+e, then P and Q are scalars with tr(P)=P=1. Thus

r’=inv(Pinv(Q)P)Pinv(Q)p=Qinv(Q)p=p.
b) Suppose we have two uncorrelated measurements with values p1 and p2 having
variances 1 and 2 respectively. We need to show that

r’=[1/(1/ 1^2 +1/ 2^2)](p1/ 1^2 + p2/ 2^2).

Solution: Here we have

P=tr(1 1)

p=tr(p1 p2)

r=(r)

and Q is the matrix with entries Q11= 1^2, Q12=0, Q21=0, Q22= 2^2 with the
inverse of Q, inv(Q), given by the entries

inv(Q)11=1/ 1^2
inv(Q)12=0
inv(Q)21=0
inv(Q)22=1/ 2^2

Putting these into the above formula we get

r’=[1/(1/ 1^2 +1/ 2^2)](p1/ 1^2 + p2/ 2^2).

c) We consider example 8.5. There are measurements of the form

r1=p1+e1
r2=p2+e2
r3=p3+e3
r4=p4+e4

r1=rf+ 1*fM

r2=rf+ 2*fM

r3=rf+ 3*fM

r4=rf+ 4*fM

where the ei’s are uncorrelated and where cov(ei,fM)=0.25 i^2. Assuming the i’s are
known exactly, find the best estimates of the ri’s.

Solution: First we note that the inverse of a 2x2 matrix A is given with entries A11=a,
A12=b,
A21=c, A22=d.

Inv(A)=[1/(ad-bc)]M
Where M is the matrix with entries M11=d, M12=-b, M21=-c, and M22=a.

For part (c) we use the above formula stated at the beginning with

P=tr(1 1).

We also use the formula

r_ei=r_f + i(r_M-r_f)

where r_ei and r_M are the expected values of the return for security i (using the
equilibrium model) and the market respectively, r_f is the risk free rate, and

_ei= i _M/( 10)

for the variance of the error in the measurement of r_ei. We have as well that

cov(e_i, f_M)=0.25 _i^2

where the errors e_i (for the measurements r_h) are uncorrelated.

For stock 1 we have r_h=15.00 and r_e=13.05 as computed in example 8.5. So

p=tr(15.00 13.05)

and Q is the matrix with entries Q11=3.00^2, Q12=Q21=0.25(3.00)^2, and Q22=2.42^2.

Inserting these into the above formula we get

r’=13.72%.

Note that the first entry for Q is /( 10)

For stock 2 we have r_h=14.34% and r_e=13.99% using 2=1.02 and we also compute

p=tr(14.34 13.99)

and Q is the matrix with entries Q11=3.275^2, Q12=Q21=0.25(3.275)^2, and

Q22=2.74^2 where

2.74=1.02 _M/( 10).

Inserting these into the above formula we get

r’=14.12%.


For stock 3 we have r_h=10.90% and r_e=17.03% using 3=1.40 and we also compute

p=tr(10.90 17.03)

and Q is the matrix with entries Q11=4.027^2, Q12=Q21=0.25(4.027)^2, and

Q22=3.76^2 where

3.76=1.40 _M/( 10).

Inserting these into the above formula we get

r’=14.25%.

For stock 4 we have r_h=15.09% and r_e=11.27% using 3=0.68 and we also compute

p=tr(15.09 11.27)

and Q is the matrix with entries Q11=2.613^2, Q12=Q21=0.25(2.613)^2, and

Q22=1.826^2 where

1.826=0.68 _M/( 10).

Inserting these into the above formula we get

r’=12.19%.
Problem 1

An investor has a utility function U(x) = x^1/4 for salary. He has a new job offer which
pays $80,000 with a bonus. The bonus will be $0, $10,000, $20,000, $30,000, $40,000,
$50,000, or $60,000, each with equal probability. What is the certainty equivalent of this
job offer?

Certainty Equivalent
U(x) = x^1/4
Salary = $80,000 with bonus
Bonus = $0, $10,000, $20,000, $30,000, $40,000, $50,000, $60,000
p of each = 1/7
Find the certainty equivalent

Salary = $80,000, $90,000, $100,000, $110,000, $120,000, $130,000, $140,000

E(U(x)) = 1/7 ($16.82 + $17.32 + $17.78 + $18.21 + $18.61 + $18.99 + $19.34)
E(U(x)) = 1/7($127.077)
E(U(x)) = $18.154

Need to find A value C such that U(C) = $18.15

Cx^1/4 = $18.15
C = $18.15^4
C = $108,610.04

The value C above is the Certainty Equivalent.

Problem 2
Suppose an investor has exponential utility function U(x) = -e-ax and an initial
wealth level of W. The investor is faced with an opportunity to invest an amount
w W and obtain a random payoff x. Show that his evaluation of this incremental
investment is independent of W.

To evaluate the incremental investment, we will compare the investment versus

not making the investment.

E [U (W − w + x )] is the expected value of the utility function if the investment is

made, and E [U (W )] is the expected value of the utility function of wealth if the
investment is not made.

E[U (W − w + x)] > E[U (W )]

[ ] [ ]
E − e − a (W − w+ x ) > E − e −aW
E [− e e
− aW ( )
] > E[− e ]
−a x−w − aW

E [− e ]⋅ E [e ( ) ] > E [− e ]
− aW −a x − w − aW

E [e ( ) ] < 1
−a x−w

The initial wealth, W, is no longer part of the equation; only the investment w and
the payoff x are. This shows that the evaluation is independent of W.
Problem 3
3) Suppose U(x) is utility function with Arrow-Pratt risk aversion coefficient a(x)
Let V(x) = C + bU(x). What is the risk aversion coefficient of V?
a(x) = -U”(x)/U’(x)
V’(x) = bU’(x)
V”(x) = bU”(x)
a(V) = bU”(x)/bU’(x)
a(V) = U”(x)/U’(x)
Therefore the risk aversion coefficient = -a(x)
Problem 4
The Arrow-Pratt relative risk aversion coefficient is
(x) = x * U”(x) / U’(x)
Show that the risk aversion coefficients are constant for U(x)1 = ln x and U(x)2 = * x



U’(x)1 = 1 / x
U”(x)1 = -1 / x2
So, (x)1 = x * (-1 / x2) / (1 / x) = -(1 / x) / (1 / x) = -1 = constant
U’(x)2 = 2 * x( -1)



U”(x)2 = ( 3 – 2) * x( -2) = 2 ( -1)* x( -2)





So, (x)2 = -1= constant

Problem 5
A young woman uses the first procedure described in Section 9.4 to deduce her utility
function U(x) over the range A<=x<=B. She uses the normalization U(A)=A, U(B)=B.
To check her result, she repeats the whole procedure over the range A’<=x<=B’, where
A<A’<B’<B. The result is a utility function V(x), with V(A’)=A’, V(B’)=B’. If the
results are consistent, U and V should be equivalent; that is, V(x)=aU(x)+b for some a>0
and b. Find A and B.

Given:
U(A)=A, U(B)=B from A to B
V(A’)=A’, U(B’)=B’ from A’ to B’
V(x)=aU(x)+b a>0
Find:
a and b
Solution:
V ( A ' ) = A ' = aU ( A ' ) + b
V ( B ' ) = B ' = aU ( B ' ) + b
A '− b
a=
U ( A' )
A '− b A ' U ( B ' ) bU ( B ' )
B' = U (B') + b = − +b
U ( A' ) U ( A' ) U ( A' )
A 'U ( B ' ) 1 − U (B') U ( A' ) − U ( B ' )
B '− =b =b
U ( A' ) U ( A' ) U ( A' )
B 'U ( A ' ) U ( A ' ) A 'U ( B ' )
b= −
U ( A ' ) − U ( B ' ) U ( A ' ) (U ( A ' ) − U ( B ' ) )
B 'U ( A ' ) − A 'U ( B ' )
b=
U ( A' ) − U ( B ' )

V ( A' ) = A' = aU ( A' ) + b

V ( B' ) = B' = aU ( B' ) + b
b = B'− aU ( B' )
A' = aU ( A' ) + B'− aU ( B' )
A'− B' = a(U ( A' ) − U ( B' ) )
A'− B'
a=
U ( A' ) − U ( B' )
Problem 6
The HARA ( for hyperbolic absolute risk aversion) class of utility functions is defined by
γ
1− γ


ax
U ( x) = + b , b>0.


γ 1−γ

Show how the parameters γ , a and b can b chosen to obtain the following special cases
(or an equivalent from).
(a) Linear or risk neutral: U(x) =x
Let λ =1 so we have ax=x then a=1, b=0.

1
(b) Quadratic: U ( x ) = x − cx 2
2
Let λ =2 so we will have 0.5ax^2+0.5b^2+abx/(1-b)
So a=b=1

(c) Exponential: U ( x ) = −e − ax
λ =- ∞
γ
1−γ
1


ax ax
+ b) = (lim1 + x) x +b =x

 

(lim


x →∞ 1 − γ γ 1− γ


x →∞

then b=1 a=1

(d) Power: U ( x) = cx γ
γ
1− γ


ax
U ( x) = +b



γ 1−γ


U ( x) = cx γ then a=1 b=1 γ = γ

(e) Logarithmic: U ( x ) = ln x
γ
1− γ


ax
U ( x) = +b



γ 1−γ


1−γ
*
1
(ax + b(1 − γ ))λ =
(1 − γ )
1−γ
(ax + b(1 − γ )) γ =
(1 − γ ) x γ
γ (1 − γ ) γ
γ γ
∂U ( x) (1 − γ )
1−γ
= γx γ −1 = (1 − γ )1−γ x γ −1
∂x γ
γ = 0, a = 1, b = 0 thenU ( x) = ln x

The Arrow-Pratt risk aversion coefficient is,

 

γ −1
ax
U ′( x) = a +b


1− γ

γ −2
ax
U ′′( x) = − a +b


2 

1− γ

γ −2
ax
+b


2
a


U ′′( x) 1− γ

a 1
a( x ) = − = γ −1 = =
U ′( x) a 1 b


ax
+b x+b x+


a 1− γ 1− γ


1− γ a


Problem 7
(The venture capitalist) A venture capitalist with a utility function U(x)=Sqrt(x)
carried out the procedure of Example 9.3. Find an analytical expression for C as a
function of e, and for e as a function of C. Do the values in Table 9.1 of the example
agree with these expressions?

Venture capitalist U(x)=Sqrt(x)

Lottery outcomes, either $1M or $9M
p(receiving $1M) varies
For p(either two outcomes)=0.5, E[U(x)]=$5M.
Certainty equivalent, C=$4M
U(C)=e

p 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

e 9 8.2 7.4 6.6 5.8 5 4.2 3.4 2.6 1.8 1
C 9 7.84 6.76 5.76 4.84 4 3.24 2.56 1.96 1.44 1

Analytical expression for C as a function of e is found:

For this problem, we can solve U(C) for C. From this, we will have an equation with an
unknown, variable probability p. Since we know the probabilities in the table, we can
calculate the expected values in terms of probability and vice versa. Substituting
probabilities as a function of expected values into the certainty equivalent equation yields
the desired certainty equivalent as a function of expected values. To find the second part
of the problem, we simply solve the equation from the first part for expected values to get
expected value as a function of certainty equivalent. Finally, we compare these two
equations to the table by substituting in values from the table for C and e to determine
that the equations remain true statements. If they were not true, the table and equations
would not agree.

U(C)=Sqrt(C)=e=E[U(x)]

For an undetermined probability value for the $1M lottery outcome, U(C)=p*U(1)+(1-
p)*U(9)

From the table

e=p*1+(1-p)*9
e=p+9-9*p=9-8*p

Since we are looking for an expression for C as a function of e, we can solve the last
equation for p and substitute this into the equation for C to find the answer.

e=9-8*p
e-9=-8*p
p=(9-e)/8

an equation for C
U(C)=p*U(1)+(1-p)*U(9)=Sqrt(C)
C=[p*U(1)+(1-p)*U(9)]^2
U(1)=Sqrt(1)=1
U(9)=Sqrt(9)=3
C=[p*1+(1-p)*3]^2=[p+3*(1-p)]^2=(p+3-3*p)^2=(3-2*p)^2=4*p^2-12*p+9

Substituting for p
C=4*[(9-e)/8]^2-12*[(9-e)/8]+9
=4*[(9-e)^2]/64-12*(9-e)/8+9
=(4/64)*[e^2-18*e+81]-(108+12*e)/8+9
=(1/16)*(e^2-18*e+81)-(216/16)+(24*e)/16+(144/16)
=(1/16)*e^2-(18/16)*e+(81/16)-(216/16)+(24*e)/16+(144/16)
=(1/16)*e^2+(6/16)*e+(9/16)=(1/16)*(e^2+6*e+9)=(1/16)*(e+3)^2

So, C(e)=[(e+3)^2]/16

Analytical expression for e as a function of C is found:

U(C)=Sqrt(C)=e=E[U(x)]
U(C)=p*U(1)+(1-p)*U(9)=Sqrt(C)
U(1)=Sqrt(1)=1
U(9)=Sqrt(9)=3
p=(9-e)/8
U(C)=Sqrt(C)=(9-e)/8+[1-(9-e)/8]*3=(9-e)/8+3-[3*(9-e)]/8
=(9-e)/8+3-(27/8)+3*e/8=(9-e+24-27+3*e)/8
=(2*e+6)/8=(e+3)/4
Sqrt(C)=(e+3)/4
4*Sqrt(C)-3=e

So, e(C)=4*Sqrt(C)-3

Do the values of Table 9.1 in Example 9.3 agree with these expressions?

Substituting values from table into these equations

C(6.6)=[(6.6+3)^2]/16=5.76, the same value as the table
e(5.76)=4*Sqrt(5.76)-3=6.6, the same value as the table

From these observations, the values in the table agree with these equations.
Problem 8
There is a useful approximation to certainty equivalent that is easy to derive. A second-
_
order expansion near x =E(x) gives

_ _ _ _ _
1
U ( x ) ≈ U ( x) + U ' ( x)( x − X ) + U ' ' ( x)( x − X ) 2
2
_ _
1
E[U ( x )] ≈ U ( x) + U ' ' ( x )Var ( x )
2
_
On the other hand, if we let c donate the certainty equivalent and assume it is close to x ,
we can use the first-order expansion
_ _ _
U (C ) = U ( X ) + U ' ( X )(C − X )

Using these approximations, show that

_
_
U (C ) − U ( X )
C = X+ _
U '( X )

As the definition
U (C ) = E[U ( X )]
_ _
1
U (C ) ≈ U ( X ) + U " ( X )Var ( X )
2
_ _ _
1
_ [U ( X ) + U " ( X )Var ( X )] − U ( X )
C = X+ 2
_
U '( X )
_
_
1 U " ( X )Var ( X )
C = X+ _
2
U '( X )
Problem 9
An investor with unit wealth maximizes the expected value of the utility function U(x) =
ax – bx2/2 and obtains a mean-variance efficient portfolio. A friend of his with wealth W
and the same utility function does the same calculation, but gets a different portfolio
return. However, changing b to b’ does yield the same result. What is the value of b’?

In general;

E[U(x)] = E[ax – 1/2bx2]

= aE[x] –1/2bE[x2]
= aE[x] – 1/2b(var[x] + E[x]2)

In this situation, if the random payoff of the portfolio of the investor with unit wealth is
R, it would maximize:

E[U(R)] = aE[R] – 1/2b(var[R] + E[R]2)

Similarly, if the investor with wealth W purchases the same portfolio, the payoff will be
WR and R should maximize:

E[U(x2)] = aE[RW] – 1/2b2(var[RW] + E[RW]2)

= aWE[R] – 1/2b2(W2var[R] + W2E[R]2)
= W[aE[R] – 1/2b2W(var[R] + E[R]2)]

If b2 = b’= b/W is substituted in the final equation for the second investor, the same
R will solve the expected value of the utility function as the R using unit wealth.
Problem 10
Suppose an investor has utility function U. There are n risky assets with rate of return
ri=1, 2, …, n, and one risk-free asset with rate of return rf. The investor has initial wealth
W0. Suppose that the optimal portfolio for this investor has random) payoff x*. Show that
E[U (x*)(ri-rf)]=0
for i = 1, 2, …, n.

1. From (9.4) p. 243 we know that E[U (x*)di ]= Pi. If there is a risk-free asset with rate
of return R, then di = R and Pi = 1. Thus,

= E[U (x*)] R => = E[U (x*)] (1+ rf )

2. If there is a asset i with total return Ri, then di = Ri and Pi = 1. Thus,

= E[U (x*) Ri] => = E[U (x*)(1+ ri )]
=> = E[U (x*)(1+ ri )] = E[U (x*)] (1+ rf )
E[U (x*)(1+ ri)] - E[U (x*)(1+ rf)] = 0
E[U (x*)(1+ ri )-U (x*)(1+ rf)]=0
E[U (x*)( ri - rf)]=0
Problem 11
0.9 0.4
+ =λ
3θ1 + 1.2θ 2 + 6θ 3 + 3000 θ1 + 1.2θ 2 + θ 4
0.36 0.48 0.36
+ + =λ
3θ1 + 1.2θ 2 + 6θ 3 + 3000 θ1 + 1.2θ 2 + θ 4 1.2θ 2 + θ 4
1.8
=λ
3θ1 + 1.2θ 2 + 6θ 3 + 3000
0.4 0.3
+ =λ
θ1 + 1.2θ 2 + θ 4 1.2θ 2 + θ 4
θ1 + θ 2 + θ 3 + θ 4 = W
1
λ=
W
1.8W = 3θ 1 + 1.2θ 2 + 6θ 3 + 3000
0.8W = θ 1 + 1.2θ 2 + θ 4
0.6W = 1.2θ 2 + θ 4
θ1 + θ 2 + θ 3 + θ 4 = W
thus
θ1 = 0.2W
θ 2 = −1250
θ 3 = 0.2W − 250
θ 4 = 0.6W + 1500

The price of money back guaranteed investment

P = $1,500
Problem 12
Formulation:
The following is a general result from matrix theory: Let A be mxn matrix.
Suppose that the equation Ax = p can achieve no p ≥ 0 except p = 0 . Then
there is a vector y > 0 with AT y = 0 . Use this result to show that if there is no
arbitrage, there are positive state prices; that is, prove the positive state price
theorem in Section 9.9. [Hint: If there are S states and N securities, let A be
an appropriate (S + 1)xN matrix]

Solution:
 

d11 d12 d13 ........ d1 N




d 21 d 22 d 23 d2N
Let construct a matrix A = ...... 

...... ...... ........ ....... . Here d ij is




d S1 dS2 dS3 ........ d SN
− p1 − p2 − p3 ........ − p N


dividend of the security j in the state i . Let take vector θ = (θ 1θ 2θ 3 .......θ N ) as

vector of weights of the securities in the portfolio.
θ1
     

d11 d 12 d 13 ........ d1 N  

D1 

d 2N θ 2
  

d 21 d 22 d 23  
D2 

Aθ = ......

...... ...... ........ ....... θ 3 =  

D3 

=T
  

d S1 dS2 d S3 ........ d SN 
..... 
..... 

− p1 − p2 − p3 ........ − p N θN − PN





(*)
Here Di is dividend in state i and PN is price of the portfolio. Now lets
assume that PN = 0 or PN < 0 and there is some k s.t. Dk > 0 . In this case we
have arbitrage because with 0 or negative price there is a possibility to get
dividend in one of the states. In order to avoid the arbitrage we need to
conclude that if PN ≤ 0 then for all i Di = 0 . It means that system (*) can
achieve with T = 0 . According to the algebra we have that ∃y i > 0 s.t.
 

y1
− p1
 

 

d 11 d 21 ....... d S1 

y2
− p2
 

d 12 d 22 ....... dS2
 

A y=
T
y3 =0 (**)
 

........ ........ ........ ........ ........ 

.....
 

− pN
 

d1 N d 2N ........ d SN 

y S +1
Because y S +1 ≠ 0 we can divide all y i by y S +1 and define them as state price.
Expression (**) will looks like:
ψ1
 

d 21 ....... d S1 − p1


 

d 11
ψ2


d 22 ....... d S 2 − p 2


d 12


A Tψ = ψ 3 = 0 where ψ i > 0


........ ........ ........ ........ ........



.....


d1 N d 2 N ........ d SN − p N





1
S
By solving this we have for each state that pi =  d jiψ j which means that for
j =1
S
each state we constructed a positive state price ψ i s.t. pi =  d jiψ j .
j =1
Problem 13
[ ]
From the above exercise, we have (rk − r )E U ' (x * ) = 0 , where r is the risk free rate. In
the4 quadratic case, we have U’(x)=1-cx. We denote by the RM the return of the
portfolio, and using the fact that initial capital is W we get (rk − r )E [U ' (WRM )] = 0 ,
equivalently E [(1 − cWR M )(Ri − R )] = 0 , so
Ri − R = cW [E (R M Ri ) − R M R ] = cW [cov(RM , Ri ) + R M ( Ri − R )], so,
Ri − R − cWR M ( Ri − R ) = cW [cov(R M , Ri )] , and equivalently Ri − R = γCov (R M , Ri ) ,
where γ = 1 − CWRM . If we apply this relation to the portfolio, we obtain
Cov (RM , Ri )
RM − R = γCov (RM , RM ) = γVar ( RM ) , so Ri − R = ( RM − R ) = β i (RM − R ) ,
Var ( RM )
and so, the problem is solved.
Problem 9.14. (At the track) At the horse race one Saturday afternoon Gavin Jones
studies the racing form and concludes that the horse No Arbitrage has a 25% chance to
win and is posted at 4 to 1 odds. (For every dollar Gavin bets, he receives $5 if the horse
wins and nothing if it loses.) He can either bet on this horse or keep money in his pocket
Gavin decides that he has a square-root utility for money.
(a) What fraction of his money should Gavin bet on No Arbitrage?
(b) What is the implied winning payoff of a $1 bet against No Arbitrage?

Solution:

(a) If we denote by α fraction of his money m G. J. should bet we need

1 3
max E[U ] = m + 4α m + (1 − α )m
4 4
So we need the 1st derivative to be equal to zero, that is,
dE 1 m 3 m
= − =0
dα 2 m + 4α m 8 (1 − α )m
7
Solving this equation we obtain α = = 0.1346
52
5
(b) Implied wining payoff of a $1 bet against No Arbitrage is = 1.25
4
Problem 15
(General risk-neutral pricing) We can transform the log-optimal pricing formula into a
risk-neutral pricing equation. From the log-optimal pricing equation we have
d
P = E( )
R*
Where R* is the return on the log-optimal portfolio. We can then define a new
expectation operation E by

Rx
E( x ) = E( ).
R*
This can be regarded as the expectation of an artificial probability. Note that the usual
rules of expectation hold. Namely:


1 1
a) If x is certain, then E ( x ) = x . This is because E ( )= .

R* R  

b) For any random variables x and y, there holds E (ax + by ) = aE ( x ) + bE ( y ) .



c) For any nonnegative random variable x, there holds E ( x ) ≥ 0 .

Using this new expectation operation, with the implied artificial probabilities, show that
the price of any security d is


E (d )
P= .
R
This is risk neutral pricing.

1  

d
From the rules b), we know * E (d ) = E ( ) (1)
R R



Rx
According to the definition of the operation E , for any variable x, E ( x ) = E ( ) , assume
R*
d
R*

d R ) = E( d )
d/R as a variable, we can get E( ) = E( *
(2)
R R R*
d
From the log-optimal pricing equation we have: P = E ( ) (3)
R*
So,
d
P = E( ) By (3)
R*
d
R*
d
R ♦

= E( ) = E( ) By (2)
R* R
1

= * E (d ) By (1)
R