Introduction to Portfolio

Selection and Capital Market

Theory: Static Analysis
BaoheWang
baohewang0592@sina.com
 Introduction
 The investment decision by households as
having two parts:
(a) the “consumption-saving” choice
(b) the “portfolio-selection” choice
 In general the two decisions cannot be
made independently.
 However, the consumption-saving
allocation has little substantive impact on
portfolio theory.
 One-period Portfolio Selection
 The solution to the general problem of
choosing the best investment mix is called
portfolio-selection theory.
 There are n different investment
opportunities called securities.
 The random variable one-period return
per dollar on security j is denoted Z j
 Any linear combination of these securities
which has a positive market value is called
a portfolio.
 U (W ) denote the utility function.
 W is the end-of-period value of the
investor’s wealth measure in dollars.
 U is an increasing strictly concave
function and twice continuously
differentiable.
 So the investor’s decision is relevant to the
subjective joint probability distribution for
( Z1 , Z 2 , . , Z n )
 Assumption 1: Frictionless Markets

 Assumption 2: Price-Taker

 Assumption 3: No-Arbitrage Opportunities

 Assumption 4: No-Institutional Restrictions

 Given these assumptions, the portfolio-
selection problem can be formally stated
as
n
max E{U ( w j Z jW0 )}
{ w1 , w2 ,wn }
1
n
(2.1)
S. T . w
1
j 1

 Where E is the expectation operator for

the subjective joint probability distribution.
 If (w1 , w2 ,, wn ) is a solution (2.1), then it will
satisfy the first-order conditions:

E{U ( Z W0 Z j )} 


W0
Where Z  1 w j Z j is the random variable
 n 


return per dollar on the optimal portfolio.

 With the concavity assumptions on U, if
the variance-covariance matrix of the
return is nonsingular and an interior
solution exists, the the solution is unique.
 Formula (2.1) rules out that any one of th
e securities is a riskless security.
 If a riskless security is added to the menu
of available securities then the portfolio sel
ection problem can be stated as:
n
max E{U ( w j Z jW0  (1  1 w j ) RW0 )}
n

{ w1 , w2 ,wn }
1 (2.4)
n
max E{U ([ w j ( Z j  R )  R ]W0 )}
{ w1 , w2 ,wn }
1
 The first-order conditions can be written
as:
E{U ( Z W0 )( Z j  R )}  0 j  1, 2, , n

can be rewritten as 1 j ( Z j  R)  R
n
 Where Z 
w

 If it is assumed that the variance-

covariance matrix of the returns on the
risky securities is nonsingular and an
interior solution exits, then the solution is
unique.
 But neither (2.1) nor (2.3) reflect that end of
period wealth cannot be negative.
 To rule out bankruptcy, the additional
constraint that, with probability one, Z   0
could be imposed on ( w1 , w2 ,  , wn* ) .
 This constraint is too weak, because the
probability assessments on {Z j }are subjective.
 An alternative treatment is to forbid
borrowing and short-selling securities where,
by law, Z  0 .
j
 The optimal demand functions for risky
securities, {wjW0 } , and the resulting
probability distribution for the optimal
portfolio will depend on
(1) the risk preferences of the investor;
(2) his initial wealth;
(3) the join distribution for the securities’
returns.
 The von Neumann-Morgenstern utility
function can only be determined up to a
positive affine transformation.
 The Pratt-Arrow absolute risk-aversion
function is invariant to any positive affine
transformation of U (W ) .
 The preference orderings of all choices
available to the investor are completely
specified by absolute risk–aversion
function
U (W )
A(W ) 
U (W )
 The change in absolute risk aversion with
respect to a change in wealth is

dA U (W )
 A(W )  A(W )[ A(W )  ]
dW U (W )
 A(W ) ispositive, and such investor are call
risk averse.
 An alternative, measure of risk aversion is
the relative risk-aversion function defined
by
U (W )W
R(W )    A(W )W
U (W )
 Its change with respect to a change in we
alth is given by
R(W )  A(W )W  A(W )
 The certainty-equivalent end-of-period
wealth WC is defined to be such that
U (WC )  E{U (W )}
 WC is the amount of money such that the
investor is indifferent between having this
amount of money for certain or the
portfolio with random variable outcome W .
 We can proof follows directly by Jensen’s
inequality: if U is strictly concave
U (WC )  E{U (W )}  U ( E{W })
 Because U is an increase function, So
WC  E{W }
 The certainty equivalent can be used to
compare the risk aversions of two investor.
 If A is more risk averse than B and they
hold same portfolio, the certainty
equivalent end of period wealth for A is
less than or equal to the certainty
equivalent end of period wealth for B.
 Rothschild and Stiglitz define the meaning
of “increasing risk” for a security so we ca
n compare the riskiness of two securities o
r portfolios.
 If E (W1 )  E (W2 ) , E{U (W1 )}  E{U (W2 )} for a
ll concave Uwith strict inequality holding
for some concave U , we said the first por
tfolio is less risky than the second portfolio
.
 Its equivalence to the two following
definitions:
(1) W2 is equal in distribution to W1 plus some
“noise”.
(2) W2 has more “weight in its tails” than W1 .
 If there exists an increasing strictly
concave function V such that
E{V ( Z )( Z j  R)}  0, j  1, 2, , n., we call this
portfolio is an efficient portfolio.
 All portfolios that are not efficient are

called inefficient portfolios.

 It follows immediately that every efficient
portfolio is a possible optimal portfolio, for
each efficient portfolio there exists an
increasing concave U such that the
efficient portfolio is a solution to (2.1) or
(2.3).
 Because all risk-averse investors have
different utility function, so they will be
indifferent between selecting their optimal
portfolios.
 Theorem 2.1: If Zdenotes the random variabl
e return per dollar on any feasible portfolio a
nd if is riskier
Z e  Zthan
e
in the Rothsch
Z Z
ild and Stiglitz sense, then
( Z is an efficient portfolio)
Z Z
e e

Proof: By hypothesis
E{U [( Z  Z )W0 ]}  E{[( Z e  Z e )W0 ]}
Z  Ze E{U ( ZW0 )}  E{U ( Z eW0 )}
If then trivially .
But Z is a feasible portfolio and
Zis
e an ef

ficient portfolio. By contradiction, Ze  Z

 Corollary 2.1: If there exists a riskless sec
urity with return R, then Z e  ,Rwith equ
ality holding only if isZae riskless security.
 Proof: If is riskless , then by Assumptio
n 3, Z.e If is not riskless, by Theore
m 2.1, . Ze  R Ze
Ze  R
 Theorem 2.2: The optimal portfolio for a n
onsatiated risk-averse investor will be the
riskless security if and only if Z j  R for j=1
,2,…..,n.
 Proof:  If Z   R is an optimal solution, t
hen we have U ( RW0 ) E{Z j  R}  0 By the n
onsatiation assumption, U ( RW0 )  0so Z j  R
 If j Z  R j  1, 2  , n then Z 
 R will sati
sfy U ( Z W0 ) E{Z j  R} because
0 the prope
rty of U, so this solution is unique.
 From Corollary 2.1 and Theorem 2.2, if a r
isk-averse investor chooses a risky portfoli
o, then the expected return on the portfoli
o exceeds the riskless rate.
 Theorem 2.3: Let Zdenote p
the return on any
portfolio p that does not contain security s. If
there exists a portfolio p such that, for
security s, , where
Zs  Z p   s
E{ s | Z j , j  1, 2, , n, j  s}  0 then the fraction of
every efficient portfolio allocated to security s
is the same and equal to zero.
Proof: Suppose is the return on an efficient
Ze
portfolio with fraction allocated to
s  0
security s, be the return on a portfolio with
the same fractional holding Z as except that
instead of security s with portfolio P
Ze
 Hence Z e  Z   s ( Z s  Z p )  Z   s s
So Z e  Z
Therefore ,for  s  0 , Z e is riskier than Z
in the Rothschild-Stiglitz. This contradicts t
hat Z e portfolio.
is an efficient
 Corollary 2.3: Let  denote the set of n se
curities and denote  the same set of sec
urities except that isZreplace s with . ZIfs
 s
Z s  Z s and E{ s | Z }, then
0 all risk avers
e investor would prefer to choose . 
 Theorem 2.3 and its corollary demonstrate
that all risk averse investors would prefer
any “unnecessary” and “noise” to be elimi
nated.
 The Rothschild-Stiglitz definition of increas
ing risk is quite useful for studying the pro
perties of optimal portfolios.
 But this rule is not apply to individual secu
rities or inefficient portfolios.
 2.3 Risk Measures for Securities and
Portfolios in The One-Period model
 In this section, a second definition of
increasing risk is introduced.
 Z ek is the random variable return per dollar
on an efficient portfolio K.
 VK ( Z eK ) denote an increasing strictly
 dVK
concave function such that for VK  dZ eK
E{VK ( Z j  R )}  0 j  1, 2, , n W0  1
VK  E{V }
 Random variable YK 
cov(VK , Z eK )
 Definition: The measure of risk bpK of
portfolio P relative to efficient portfolio K
with random variable return Z eK is defined
by
b  cov(YK , Z P )
K
p

and portfolio P is said to be riskier than

portfolio P relative to efficient portfolio K
if bpK  bpK .
 Theorem 2.4: If Z p is the return on a feasible
portfolio P and e is the return on efficient
K
Z
portfolio K , then Z p  R  bpK ( Z eK  R) .
Proof: From the definition
E{VK ( Z j  R )}  0 j  1, 2,  , n
 j be the fraction of portfolio P allocated to
security j, then
n
Z P    j (Z j  R)  R
and 1

  E{V  (Z
1
j K j  R )}  E{VK ( Z P  R )}  0
By a similar argument, E{VK ( Z eK  R )}  0
Hence,
 K

cov(VK , Z e )  E[VK ( Z e  Z e )]
K K

 E[VK ( Z eK  R  R  Z eK )]
 
 E[VK ( Z e  R )]  E[VK ( R  Z e )]
K K

 ( R  Z ) E[VK ]
e
K

and
cov(VK , Z P )  ( R  Z P ) E{VK }
By Corollary 2.1 , e  R
K
Z . Therefore
Z p  R  b (Z  R)
K
p e
K
 Hence, the expected excess return on
portfolio P, Z P  R is in direct proportion
to its risk and the larger is its risk , the
larger is its expected return.
 Consider an investor with utility function U
and initial wealth W0 who solves the
portfolio-selection problem:
max E{U ([ wZ j  (1  w) Z ]W0 )}
w
 The first order condition:
E{U ([ w* Z j  (1  w* ) Z ]W0 )( Z j  Z )}
 If Z  Z * then the solution is W *  0 .
 However , an optimal portfolio is an efficie
nt portfolio. By Theorem 2.4
Z j  R  b (Z  R)
*
j
*

 So w*W is similar to an excess demand fun

*
ction . Measures
b j the contribution of se
curity j to the Rothsechild-Stiglitz risk of th
e optimal portfolio.
 By the implicit function theorem, we have:

w *
w W0 E{U ( Z  Z j )}  E{U }
*


Z j 
W0 E{U ( Z  Z j ) } 2

 Therefore , if Z j lies above the risk-return

line in the ( Z , b ) plane, then the investor
would prefer to increase his holding in
security j.
 bpK is a natural measure of risk for
individual securities.
 The ordering of securities by their
systematic risk relative to a given efficient
portfolio will be identical with their
ordering relative to any other efficient
portfolio.
Lemma 2.1:

(i) E{Z P | VK }  E{Z P | Z e } for efficient portf
K

olio K.
cov( Z p ,VK )  0
e }  Z p then
K
(ii) If E{ Z P | Z
(iii) cov( Z p ,VK )  0 for efficient portfolio K if
and only if cov( Z PVL )  0 for every efficient
portfolio L.
Proof: (i) VK is a continuous monotonic fun
ction of and
Z K
e hence VK

and are
Z K
e in o
ne to one correspondence.
 (ii) cov(Z p ,VK )  E{VK ( Z p  Z P )}  E{VK E{Z p  Z P | ZeK }}  0
(iii)Because bpK  0  cov( Z p ,VK )  0
 if p  0 , then Z p  R .
b K

Property I: If L and K are efficient portfolios,

then for any portfolio p, bpK  bLK bpL .
Proof : From Theorem 2.4

Z R L
Zp  R Zp  R
b 
K
L
e
b 
K
p b 
L
p
Z R e
K
Z R
e
K
Z R
e
L
 Property 2: If L and K are efficient
portfolios, then bK  1 and bKL  0 .
K

 Hence, all efficient portfolios have positive

systematic risk, relative to any efficient
portfolio.
 Property 3: Z p  R if and only if bpK  0 for
every efficient portfolio K.
 Property 4: Let p and q denote any two
feasible portfolios and let K and L denote
K  K
any two efficient portfolios. bp bq if
 L 
and only if bp bq
L


 Proof: From Property 1, we have
b b b
K
p
K
L
L
p b b b
K
q
K L
L q

K
 Thus the b measure provides the same
p
orderings of risk for any reference efficient
portfolio.
 Property 5: For each efficient portfolio K
and any feasible portfolio p, Z p  R  bpK ( ZeK  R)   p
where E{ p }  0 and E{ V
p L
 ( Z e )}  0
L
for
every efficient portfolio L.
 Proof: From Theorem 2.4 E{ p }  0 . If por
tfolio q is constructed by holding one dolla
r p, dollars
b K
p riskless security, short sellin
g bp portfolio K, then
dollars
K
Zq  R   p
so bqL  0 for every efficient portfolio L.
But q  0 implies 0  cov( Z q ,VL )  E{ p ,VL}
L
b

for every efficient portfolio L.

 Property 6: If a feasible portfolio p has por

1 j j
n
tfolio weight ( 1 ,  , 
,then
n ) b K
p   b K
 Hence , the systematic risk of a portfolio is
the weighted sum of the systematic risks o
f its component securities.
 The Rothschild Stiglitz measure provides o
nly for a partial ordering.
 bp measure provides a complete ordering
K

.
 They can give different rankings.
 The Rothschild Stiglitz definition measure t
he “total risk” of a security. It is appropriat
e definition for identifying optimal portfolio
s and determining the efficient portfolio se
t.
 The b measure the “ systematic risk” of a
K
j
security.
K
 To determine the b j , the efficient
portfolio set must be determined.
 The manifest behavioral characteristic
shared by all risk averse utility
maximization is to diversify.
 The greatest benefits in risk reduction com
e from adding a security to the portfolio w
hose realized return tends to be higher wh
en the return on the rest of the portfolio is
lower.
 Next to such “ countercyclical” investment
s in terms of benefit are the noncyclic sec
urities whose returns are orthogonal to th
e return on the portfolio.
 Theorem 2.5 : If Z p and Z q denote the
returns on portfolio p and q respectively
and if, for each possible value of Z e ,
dG p ( Z e ) dGq ( Z e )
dZ e with strict inequality

dZ e
holding over some finite probability
measure of Z e ,then portfolio p is riskier
than portfolio q and Z p  Z q .
Where G p (Z e )  E{Z p | Z e } , Z e is the
realized return on an efficient portfolio.
 Proof:

bp  bq  cov[Y ( Z e ), Z p  Z q ]  E[Y ( Z e )( Z p  Z q )]
 E[Y ( Z e )( E{Z p | Z e }  E{Z q | Z e })]
 E[Y ( Z e )(Ge ( Z p )  Ge ( Z q ))
 cov[Y ( Z e ), Ge ( Z p )  Ge ( Z q )]
Y (Ze )is a strictly increasing function, Ge ( Z p )  Ge ( Z q )
is a nondecreasing function, so
bp  bq  cov[Y ( Z e ), Ge ( Z p )  Ge ( Z q )]  0
From Theorem 2.4 Z p  Zq
  Theorem 2.6: If Z p and Z q denote the
returns on portfolio p and q respectively
and if, for each possible value of Z e ,
dG p ( Z e ) dGq ( Z e )
  a pq , a constant, then
dZ e dZ e
bp  bq  a pq and Z  Z  a ( Z  R) .
p q pq e
Proof: By hypothesis
Ge ( Z p )  Ge ( Z q )  a pq  h
bp  bq  cov[Y ( Z e ), Ge ( Z p )  Ge ( Z q )]
 cov[Y ( Z e ), a pq Z e  h]  a pq
Z p  R  bp ( Z e  R )  R  bq (Z e  R )  a pq (Z e  R )  Z q  a pq (Z e  R )
 Theorem 2.7: If, for all possible values of Z e
(i)dG (Z ) dZ  1 , then Z p  Z e
p e
e

dG p ( Z e )
(II) 0  dZ e
1 , then R  Z p  Z e

dG p ( Z e )
(III) dZ e
0 , then R  Zp

dG p ( Z e )
(IV) dZ e
 ap , a constant, then
Z p  R  a p (Z e  R)
 Theorems 2.5, 2.6 and 2.7 demonstrate,
the conditional expected return function
provides considerable information about a
security’s risk and equilibrium expected
return.
 2.4 Spanning, Separation, and
Mutual-Fund Theorems
 Definition: A set of M feasible portfolios
with random variable returns ( X 1 , X M )
is said to span the space of portfolios
contained in the set  if and only if for
any portfolio in  with return denoted by Z p
there exist numbers (1 , M ) , 1  i  1
M

such that Z p  1  j X j
M
 A mutual fund is a financial intermediary
that holds as its assets a portfolio of
securities and issues as liabilities shares
against this collection of assets.
 Theorem 2.8 If there exist M mutual funds
whose portfolio span the portfolio set  ,
then all investors will be indifferent
between selecting their optimal portfolios
from  and selecting from portfolio
combination of just the M mutual funds.
 Therefore the smallest number of such

funds M is a particularly important
spanning set.
 When such spanning obtain, the investor’s
portfolio-selection problem can be
separated into two steps.
 However, if the smallest funds can be
constructed only if the fund managers
know the preferences, endowments, and
probability beliefs of each investor.
Theorem 2.9: Necessary conditions for the
M feasible portfolios with return ( X 1 , , X M )
to span the portfolio set  are (a) that
f

the rank of   M and (b) that there exist

numbers 1 1  j  1 such that the
M
( ,  ,  M ),
random variable 1  j X j has zero variance.
M

Proposition 2.1: If Z p  1 a j Z j  b is the

n

return on some security or portfolio and if

there are no “ arbitrage opportunities”
then
(a ) b  (1  1 a j ) R and (b) Z p  R  1 a j ( Z j  R )
n n
 Proof: Let Z  be the return on a portfolio
with fraction  j allocated to security j, j  1,, n;


p Zp
1   p  1 to
allocated the security with return
n 
j

; and allocated  jto the riskless s



 
  a
ecurity j with return
p j p [b  R (1  
R, Zif  Ris chosen

such
n
a )] t
1 j

hat ,then Z   R is
riskless security and therefore but
can be chosen arbitrarily. So we get the
result.
 Hence, as long as there are no arbitrage o
pportunities, it can be assumed without lo
ss of generality that one of the portfolios i
n any candidate spanning set is the riskles
s security.
 Theorem 2.10: A necessary and sufficient
condition for ( X 1 , , X m , R) to span is that
f

there exist number {aij } such that

Z j  R  1 aij ( X i  R ) j  1, 2, , n.
m
 Proof:  If ( X 1 , , X m , R) span  f
 , then
 such that Z j  1  ij X i . Because
M M
1
 ij  1

X M  R and substituting  Mj  1  1  ij , we ha
m

ve Z j  R  1 aij ( X i  R ) j  1, 2, , n.
m

 we pick the portfolio weights  ij  fo

aij
r i  1,  , mand  Mj  1  1 ,ij from which
m

it follows that Zj   .But

 ij Xevery portfo
M
1 i

lio in canbe written as a portfolio comb

f

ination of ( Z1 , ,andZ n ) R.
 Corollary 2.10: A necessary and sufficient
condition for ( X 1 , , X m , R) to be the smallest
number of feasible portfolio that span is th
at the rank of equals the rank of  X  m
 Proof:  If the rank of  X  m , then X
are linearly independent. Moreover
hence, if the rank of   m then there e
that Z j  Z j  1 aij ( X i  X i )
m
xist number {aijsuch }
for . Therefore Z j  b j  1 aij X i wher
m
j  1,  , n
b j  Z j  1 by
m
e aij XTheorem
i 2.10
span
f
 It follows from Corollary 2.10 that a
necessary and sufficient condition for
nontrivial spanning of  is f
that some of the
risky securities are redundant securities.
 By Theorem 2.10, if investors agree on a set
of portfolios such that
( X 1 , , X m , R ) and if they
agree
Z j  R  1 aij (number
on the
m
X i  R ) j  1,,then2, , n. span
even if investors do {not a } agree (on
ij
the
X 1 , , X m , R)
joint distribution
 f of
( X 1 , , X m , R )
 Proposition 2.2: If Z e is the return on a por
tfolio contained in , then
 e
any portfolio t
hat combines positive amount of with the Z e
riskless security is also contained in ,

where
e
is theset
e
of all efficient portfolio
s contained in . f
 Proof: Let Z   Z e  (1   ) R , because Z eis
an efficient portfolio, so E{V ( Z e )( Z j  R)}  0
Define U (W )  V (aW  b) where a  1 and
b  (  1)R , Hence E{U ( Z )( Z j  R)}  0 , th

us Z is an efficient portfolio.
 It follows from Proposition 2.2 that, for
every number Z such that Z  R , there
exists at least one efficient portfolio with
expected return equal to Z .
 Theorem 2.11: Let ( X 1 , , X m ) denote the
return on m feasible portfolios. If, for
security j, there exist number {aij } such that
Z j  Z j  1 aij ( X i  X i )   j where E{ jVK ( Z eK )}  0
m

for some efficient portfolio K, then

Z j  R  1 aij ( X i  R )
m
 Proof: Let Z p   Z j  1  i X i  (1    1  i )R
m m


because Z j  Z j  1 aij ( X i  X i )   j , thus

m

Z p  R   [ Z j  R  1 aij ( X i  R )]   j by
m

construction , E{ j }  0 and hence cov(Z p ,VK )  0

Therefore the systematic risk of portfolio
p, bp is zero. From Theorem 2.4 Z p  R
K

therefore Z j  R  1m aij ( X i  R)

 Hence, if the return on a security can be
written in this linear form relative to the
portfolios ( X 1 , , X m ) , then its expected
excess return is completely determined by
the expected excess returns on these
portfolios and the weights {aij } .
 Theorem 1.12: If, for every security j,
there exist numbers {aij } such that
Z j  R  1 aij ( X i  R )   j
m

where E{ j | X 1 , , X m }  0 , then ( X 1 , , X m , R)

span the set of efficient portfolios  e .
 Proof:
Z  1 w j Z j  1 w j [ R  1 aij ( X i  R)   j ]
K n K n K m
e

 1 w j R  1 1 w j aij ( X i  R)  1 wKj  j
n K n m K m

 R  1  iK ( X i  R)   K
m

1 j aij
n
   K  1 w Kj  j
K K
Where
m
w i

Construct portfolio Z  1  X i  (1  1  iK ) R
m K m
i

Thus Z e  Z   where E{ | Z }  0

K K K

K
Z
Hence, for   0 , e is riskier than Z,
K

K
which contradicts that Z e is and efficient
portfolio. So  K
 0 . We get the result.
 K
 Theorem 2.13: Let j denote the fraction
w
of efficient portfolio K allocation to security
j, j  1, , n. ( X 1 ,span 
, X m , R) if and only
e

if there exist number for every {aij } security j

such that Z j  R  1 aij ( X i  R )   j
m

where E{ j | 1  i X i }  0,  i  1 wKj aij for

m K K n

every efficient portfolio K.

 Corollary 2.13: (X,R) span  e if and only if
there exist a number a j for each security j,
j  1, , n, such that Z j  R  a j ( X  R )   j
where E{ j | X }  0
 Proof: By hypothesis, for e
Z eK   K ( X  R)  R
very efficient portfolio K. If X  R, then fro
m Corollary 2.1  for
K
0 every efficient p
ortfolio K and R span  e
. Otherwise, from
Theorem 2.2, for everyefficient
K
 0 po
rtfolio. By Theorem 2.13,
E{ j |  K X }  0 so E{ j | X }  0
 is contained in  , any properties
e f
 Since
proved for portfolios that span  e must be
properties of portfolio that span  f .
 From Theorem 2.10, 2.12, 2.13, the
essential difference is that to span the
efficient portfolio set it is not necessary
that linear combinations of the spanning
portfolios exactly replicate the return on
each available security.
 All the models that do not restrict the
class of admissible utility function, the
distribution of individual security returns
must be such that
Z j  R  1 aij ( X i  R )   j
m
 Proposition 2.3: If, for every security j,
E{ j | X 1 , , X m }  0 with ( X 1 , , X m ) linearly i
ndependent with finite variances and if the
return on security j, has
Z j a finite varianc
e, then the {aij } i in Theorems
1, , m, 2.12
and 2.13 are given by
aij  1 vik cov( X K , Z j ) where vik is the ikth ele
m

1
ment of  .X
 Hence given some knowledge of the joint e
distribution of a set of portfolio that span 
with Z j  Z j , we can determining the aijand Z j
 Proposition 2.4: If ( Z1 , , Z n ) contain no
redundant securities,  j denotes the
fraction of portfolio X allocated to security

j, and w j denotes the fraction of any risk-
averse investor’s optimal portfolio
allocated to security j, j  1, , n, then for
every such risk-averse investor
w 
jj
 j , k  1, 2, , n
w k
*
k
 Because every optimal portfolio is an effici
ent portfolio and the holding of risky securi
ties in every efficient portfolio are proporti
onal to the holding in X.
 
If there exist numbers  j where   , j, k  1,, n
 *
 j
 *
j

k k

1 j
n
 *

and ( ,the
,then
*
1  )portfolio with proporti
*
n
ons is called the Optimal Combina
tion of Risky Assets.( X , R) e e
 Proposition 2.5: If span , then i
s a convex set.
 Proof: Let Z e1  1 ( X  R )  R Z e2   2 ( X  R )  R

and 1 2 ,   Z   Z 1
e  (1   ) Z e . By
2

substitution, the expression for Z can be

rewritten as   e  R )  R , where
1
Z ( Z
    ( 2  )(1   ) .Therefore by Proposition
2.2, Z is an efficient portfolio. It follow by
1

induction that for any integer k and

number i such that 0  i  1, i  1, , k and
1 i 1 i e is the return on an
k k
  1, Z k
  Z i

efficient portfolio. Hence ,  e

is a convex
set.
 Definition: A market portfolio is defined as
a portfolio that holds all available
securities in proportion to their market
values.
 The equilibrium market value of a security
for this purpose is defined to be the
equilibrium value of the aggregate
demand by individuals for the security.
 The market value of a security equals the
equilibrium value of the aggregate amount
of that security issued by business firms.
 We use V j denote the market value of sec
urity j and denote
VR the value of the riskl

ess security, then is the
M
j fraction of secur
ity j held in a market portfolio.
Vj
 M

V
j n
1 j  VR
 Theorem 2.14: If  e
is a convex set, and if
the securities’ market is in equilibrium, the
n a market portfolio is an efficient portfolio
.
 Proof: Let there be K risk averse investor i
n the economy.Define Z  R  1 w j ( Z j  R)
K n k

to be the return on investor k’s optimal po

rtfolio. In equilibrium, 1 wkj W0k ,Vwhere
K
j

is the 0initial wealth of investor K, and

k
W
1 W0  W0 .Define
K n
K
V  VR
1 j

  W W k  1, K . By definition of a market po

k
0
k

1 j k j j  1,, nMultiplying
0

rtfolio by
K
w k
   M
.
 R summing over j, it follows that
Z jand

 k 1 w ( Z j  R )  1 K ( Z K  R)
K n k K
1 j


n

1 i
M
( Z j  R)  Z M  R
because 1 k  1, Z M  1 K Z k . Hence, Z M is
K K

a convex combination of the returns on K


efficient portfolios. Therefore , if is
e

convex, then the market portfolio is

contained in  .
e

 The efficiency of the market portfolio

provides a rigorous microeconomic

justification for the use of a “
representative man” to derive equilibrium
prices in aggregated economic models.
 Proposition 2.6: In all portfolio models wit
h homogeneous beliefs and risk-averse inv
estors the equilibrium expected return on t
he market portfolio exceeds the return on
the riskless security.
 Proof: From the proof of Theorem 2.14 an
d Corollary 2.1. 1 k  R), beca
K
Z M  R   ( Z k

use Z ,  R k . Hence
k 0 ZM  R
 The market portfolio is the only risky portf
olio where the sign of its equilibrium expec
ted excess return can always be predicted.
 Returning to the special case where  e is s
panned by a single risky portfolio and the
riskless security, the market portfolio is eff
icient. So the risky spanning portfolio can
always be chosen to be the market portfoli
o.
 Theorem 2.15: If M ( Z , R ) 
span , then the
e

equilibrium expected return on security j c

an be written as Z j  R   j ( Z M  R )
where cov( Z j , Z M )
j 
var( Z M )
 This relation, called the Security Market
Line, was first derived by Sharpe.
 In the special case of Theorem 2.15,  j
measure the systematic risk of security j
relative to the efficient portfolio Z M .
  j can be computed from a simple
covariance between Z j and Z M . But the
k
sign of b j can not be determined by the
sign of the correlation coefficient between
Z j and Z ek
 Theorem 2.16: If ( Z1 , , Z n ) contain no
redundant securities, then (a) for each

value  ,  j , j  1, , n, are unique, (b)
there exists a portfolio contained in
with return X such that ( X , R) span  min ,
and (c) Z j  R  a j ( X j  R) where,
cov( Z j , X )
aj  , j  1, , n.
var( X )
 Where  min denote the set of portfolios
contained in  such that there exists no
f

other portfolio in  with the same

f

expected return and a smaller variance.

 Proof: Let  ij denote the ijth element of  a
nd ijdenote the ijth element of
v  . So al
1

l portfolios in  minwith expect return u, we

need solutions the problem
min 1 1  i j ij
n n

S .T Z (  )  
If   R then Z ( R)  R and j  0, j  1, 2 , n
 R

Consider the case when   R . The n first-

order conditions are
0  1  j ij  u ( Z i  R) i  1, 2, , n
n
Multiplying by and summing, we get
1 1 i j ij  i i  (Zi  R)  0
n n    n
    
u  var[ Z (  )] (   R)

By definition of  min ,  must be the same

for all Z (  ) . Because  is nonsingular, the
linear equation has unique solution
 j  u 1 vij ( Z i  R ) j  1, , n
 n

This prove (a). From this solution we have

 j  k are the same for every value  .
Hence all portfolios in  min are perfectly
correlated. Hence we can pick any
 portfolio in  min with   R and call its
return X. Then we have
Z ( )    ( X  R)  R
Hence ( X , R) span  min which proves (b).
and from Corollary 2.13 and Proposition
2.3 (c) follows directly.
 From Theorem 2.16, ak will be equivalent
to k as a measure of a security’s
b K

systematic risk provided that the

chosen for X is such that   R .
 Theorem 2.17: If ( X , R) span  e
and if X
has a finite variance, then  e is contained
in  min .
 Proof: Let Z e  R  ae ( X  R ) . Let Z p be
the return on any portfolio in  f
such
that Z e  Z p . By Corollary 2.13 Z  R  a ( X  R)  
p p p

where E{ p }  E{ p | X }  0

 Therefore a p  ae
Thus
var( Z p )  a 2p var( X )  var( p )  a p var( X )  var( Z e )

Hence, Z e is contained in  min .

 Theorem 2.18: If ( Z1 , , Z n ) have a joint
normal probability distribution, then there
exists a portfolio with return X such that
( X , R) span  .
e
 Proof: construct a risky portfolio contained
in  min , and call its return X. Define
 k  Z k  R  ak ( X  R ), k  1, , n by
Theorem 2.16 part (c) E{ k }  0 and by
construction cov( k , X )  0 . Because Z1  Z n
are normally distributed, X will be normally
distributed. Hence is normal
k
distributed , and because cov( X ,  k )  0 , so
they are independent. Therefore
E{ k }  E{ k | X }  0 , From Corollary 2.13
it follows that ( X , R ) span  e
 Theorem 2.19: If p(Z1 , , Z n ) is a symmetric
function with respect to all its arguments,
then there exists a portfolio with return X
such that ( X , R )span  .
e

 Proof: By hypothesis
p ( Z1 , Z i , Z n )  p ( Z i , Z1 , Z n ) for each set
of given values. Therefore every risk
averse investor will choose 1 i . But this  
  

is true for all i. Hence , all investor will

hold all risky securities in the same
relative proportions. Then ( X , R ) span  e
 The APT model developed by Ross
provides an important class of linear-factor
models that generate spanning without
assuming joint normal probability
distributions.
 If we can construct a set of m portfolios

with returns ( X 1 , , X M ) such that X i and Yi

are perfectly correlated, i  1, , m, then
( X 1 , , X M , R) will span  e
 The APT model is attractive because the
equilibrium structure of expected returns
and risks of securities can be derived
without explicit knowledge of investors’
preferences or endowments.
 For the study of equilibrium pricing, the us
ual format is to derive equilibrium Vgiven j0

the distribution of . Vj
 Theorem 2.20: If ( X 1 , , X m ) denote a set of
linearly independent portfolios that satisfy
the hypothesis of Theorem 2.12, and all se
curities have finite variances, then a neces
sary condition for equilibrium in the securit
ies’ market is that
V j  1 
m m
vik cov( X k , V j )( X j  R)
Vj0  1
R
1
where vik is the ikth 
element of X
 Proof: By linear independence V j  Z jV j 0
by Theorem 2.12 V j  V j 0 [ R   m aij ( X i  R)   j ]
1
where E{ j | X 1 , , X m }  0 . Take
expectations, we have
V j  V j 0 [ R  1 aij ( X i  R)]
m

Noting that cov( X k , V j )  V j 0 cov( X k , Z j )

From Proposition 2.3 aij  1 vik cov( X K , Z j )
m

Thus V j 0 aij  1 vik cov( X K ,V j )

m

We can get
V j  1 
m m
vik cov( X k ,V j )( X j  R )
Vj0  1
R
 Hence, from Theorem 2.20, a sufficient se
t of information to determine the equilibriu
m value of security j is the first and secon
d moments for the join distribution of
. ( X 1 , , X m , V j )
 Corollary 2.20a: If the hypothesized condit
ions of Theorem 2.20 hold and if the end-
of-period value a security is given by
V  1  jV j then in equilibrium
n

V0  1  jV j 0
n

 This property of formula is called “ value a

dditivity”.
 Corollary 2.20b: If the hypothesized
conditions of Theorem 2.20 hold and if the
end-of-period value of a security is given
by V  qV j  u , where E{u}  E{u | X 1 , , X m }  u
and E{q}  E{q | X 1 , , X m }  q then in
equilibrium V0  qV j 0  u R
 Hence, to value two securities whose end
of period values differ only by
multiplicative or additive “noise”, we can
simply substitute the expected values of
the noise terms.
 Theorem 2.20 and its corollaries are
central to the theory of optimal investment
decisions by business firms.
 Although the optimal investment and
financing decisions by a form generally
require simultaneous determination, under
certain conditions the optimal investment
decision can be made independently of
the method of financing.
 Theorem 2.21: If firm j is financed by q
different claims defined by the function
f k (V j ) k  1, , q, and if there exists an
equilibrium such that the return
distribution of the efficient portfolio set
remains unchanged from the equilibrium in
which firm j was all equity financed, then


q
f  V (I )
where is the equilibrium initial value of
k0 j0 j
is the equilibrium initial value of
1

financialf k 0claim k.
 Hence, for a given investment policy, the
way in which the firm finances its
investments changes the return
distribution of the efficient portfolio set.
 Clearly, a sufficient condition for Theorem
2.21 to obtain is that each of the financial
claims issued by the firm are “ redundant
securities”.
 An alternative approach to the
development of nontrivial spanning
theorems is to derive a class of utility
functions for investors .
 Such that even with arbitrary joint
probability distributions for the available
securities,investors within the class can
generate their optimal portfolios from the
spanning portfolios.
 Let  u denote the set of optimal portfolios
selected from  f by investors with strictly
concave von Neumann-Morgenstern utility
functions.
 Theorem 2.22 There exists a portfolio with
span  if and onl
u
return X such that ( X , R )
y if Ai (W )  1 (ai  bW ) , where
0 isAithe abs
olute risk-aversion function for investor
i in  u .
 Because the b in the statement of Theore
m 2.22 does not have a subscript , theref i
ore all investors in 
must have virtually t
u

he same utility function.

 Cass and Stiglitz (1970) conclude: it is req
uirement that there be any mutual funds,
and not the limitation on the number of m
utual funds.
 This is a negative report on the approach t
o developing spanning theorems.
The End

thanks