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 Zj dollar on security j is denoted 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 portfolioselection problem can be formally stated as n max E{U ( w j Z jW0 )} { w1 , w2 , wn } 1 n S. T . w j 1 (2.1) 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 1n wj Z j is the random variable 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 the securities is a riskless security. If a riskless security is added to the menu of available securities then the portfolio selection problem can be stated as: n max E{U ( w j Z jW0 (1 1 w j ) RW0 )} { w1 , w2 , wn } n 1 n max E{U ([ w j ( Z j R) R]W0 )} { w1 , w2 , wn } 1 (2.4) The first-order conditions can be written as: E{U (Z W0 )(Z j R)} 0 j 1, 2, ,n Where Z can be rewritten as 1 wj ( Z j R) R If it is assumed that the variancecovariance matrix of the returns on the risky securities is nonsingular and an interior solution exits, then the solution is unique. n 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 j 0 . The optimal demand functions for risky securities, {wjW0} , 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 ) is positive, 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 wealth 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 can compare the riskiness of two securities or portfolios. If E (W1 ) E (W2 ) , E{U (W1 )} E{U (W2 )} for all concave U with strict inequality holding for some concave U , we said the first portfolio 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 Z denotes the random variable return per dollar on any feasible portfolio and if Z e Z e is riskier than Z Z in the Rothschild and Stiglitz sense, then Z e Z ( Z e is an efficient portfolio) Proof: By hypothesis E{U [( Z Z )W0 ]} E{[( Z e Z e )W0 ]} If Z Z e then trivially E{U (ZW0 )} E{U (ZeW0 )} . But Z is a feasible portfolio and Z e is an efficient portfolio. By contradiction, Z e Z Corollary 2.1: If there exists a riskless security with return R, then Z e R , with equality holding only if Z e is a riskless security. Proof: If Z e is riskless , then by Assumption 3, Z e R . If Z e is not riskless, by Theorem 2.1, Z e R . Theorem 2.2: The optimal portfolio for a nonsatiated 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, then we have U ( RW0 ) E{Z j R} 0 By the nonsatiation assumption, U ( RW0 ) 0 so Z j R Z R j 1, 2 , n If then Z R will j satisfy U (Z W0 )E{Z j R} 0 because the property of U, so this solution is unique. From Corollary 2.1 and Theorem 2.2, if a risk-averse investor chooses a risky portfolio, then the expected return on the portfolio exceeds the riskless rate. Theorem 2.3: Let Z p denote the return on any portfolio p that does not contain security s. If there exists a portfolio p such that, for security s, Z s Z p s , where 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 Z e is the return on an efficient portfolio with fraction s 0 allocated to security s, Z be the return on a portfolio with the same fractional holding as Z e except that instead of security s with portfolio P 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 Ze that is an efficient portfolio. Corollary 2.3: Let denote the set of n securities and denote the same set of securities except that Z s is replace with Z s. If Z s Z s s and E{ s | Z} 0 , then all risk averse 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 eliminated. The Rothschild-Stiglitz definition of increasing risk is quite useful for studying the properties of optimal portfolios. But this rule is not apply to individual securities 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 V concave function such that for K dZ K E{VK ( Z j R)} 0 Random variable j 1, 2, ,n W0 1 VK E{V } YK cov(VK , Z eK ) e 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 ) and portfolio P is said to be riskier than portfolio P relative to efficient portfolio K K p if bpK bpK . Theorem 2.4: If Z p is the return on a feasible K Z portfolio P and e is the return on efficient portfolio K , then Z p R bpK (ZeK 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 n E{V (Z j 1 K j R)} E{VK ( Z P R)} 0 By a similar argument, E{VK ( Z eK R )} 0 Hence, K K K cov(VK , Z e ) E[VK ( Z e Z e )] E[VK ( Z eK R R Z eK )] K K E[VK ( Z e R )] E[VK ( R Z e )] and ( R Z ) E[VK ] K e cov(VK , Z P ) ( R Z P ) E{VK } K Z By Corollary 2.1 , e R . Therefore Z p R b ( Z R) K p K e 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 efficient portfolio. By Theorem 2.4 Z j R b (Z R) * j * So w*W is similar to an excess demand * b function . j Measures the contribution of security j to the Rothsechild-Stiglitz risk of the optimal portfolio. By the implicit function theorem, we have: w w W0 E{U (Z Z j )} E{U } 2 Z j W0 E{U (Z Z j ) } * * 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: K (i) E{Z P | VK } E{Z P | Z e } for efficient portfolio K. K ) 0 cov( Z , V E { Z | Z } Z p K (ii) If P e p then (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 K K V Z Z function of e and hence K and e are in one to one correspondence. (ii) cov(Z p ,VK ) E{VK (Z p ZP )} E{VK E{Z p Z P | ZeK }} 0 (iii)Because bpK 0 cov(Z p ,VK ) 0 K b if p 0 , then Z p R . Property I: If L and K are efficient portfolios, then for any portfolio p, bpK bLK bpL . Proof : From Theorem 2.4 L Z K e R bL K Ze R b K p Zp R Z R K e b L p Zp R Z R L e Property 2: If L and K are efficient K portfolios, then bK 1 and bKL 0 . 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. b p bq if and only if bpL bqL Proof: From Property 1, we have b b b K p K L L p b b b K q K L L q Thus the b measure provides the same 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 L E { V ( Z E { } 0 where and for p L e )} 0 p every efficient portfolio L. K p Proof: From Theorem 2.4 E{ p } 0 . If portfolio q is constructed by holding one K b dollar p, p dollars riskless security, short selling bpK dollars portfolio K, then Z q R p so bqL 0 for every efficient portfolio L. L b But q 0 implies 0 cov( Z q ,VL ) E{ p , VL} for every efficient portfolio L. Property 6: If a feasible portfolio p has n K portfolio weight (1 , , n ) ,then bp 1 j b Kj Hence , the systematic risk of a portfolio is the weighted sum of the systematic risks of its component securities. The Rothschild Stiglitz measure provides only for a partial ordering. K bp measure provides a complete ordering. They can give different rankings. The Rothschild Stiglitz definition measure the “total risk” of a security. It is appropriate definition for identifying optimal portfolios and determining the efficient portfolio set. The b measure the “ systematic risk” of a 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. K j The greatest benefits in risk reduction come from adding a security to the portfolio whose realized return tends to be higher when the return on the rest of the portfolio is lower. Next to such “ countercyclical” investments in terms of benefit are the noncyclic securities whose returns are orthogonal to the 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 ) dG ( Z ) q e dZ e dZ e with strict inequality 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 )] 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 Z q Y (Ze ) 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 (Ze R) R bq (Ze R) a pq (Ze R) Zq a pq (Ze R) Theorem 2.7: If, for all possible values of Z e (i)dG (Z ) dZ 1 , then Z p Ze p e e (II) 0 dG p ( Z e ) dZ e (III) dG p ( Z e ) (IV) dG p ( Z e ) dZ e dZ e 1 0 ap , then R Z p Ze , then R Zp , 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 ( X1 , 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 M there exist numbers (1 , M ) , 1 i 1 such that Z p 1M j X j 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 ( X1 , , X M ) f to span the portfolio set are (a) that the rank of M and (b) that there exist M ( , , ), numbers 1 1 j 1 such that the M M random variable 1 j X j has zero variance. n Proposition 2.1: If Z p 1 a j Z j b is the 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, p allocated to the security with return Z p; n and 1 p 1 j allocated to the riskless security with return R, if j is chosen such n a that j p j ,then Z R p [b R (1 1 a j )] is riskless security and therefore Z R but can be chosen arbitrarily. So we get the result. , n; Hence, as long as there are no arbitrage opportunities, it can be assumed without loss of generality that one of the portfolios in any candidate spanning set is the riskless security. Theorem 2.10: A necessary and sufficient f condition for ( X1 , , X m , R) to span is that there exist number {aij } such that Z j R 1 aij ( X i R ) j 1, 2, m , n. Proof: If ( X1 , M M 1 ij 1 such that Z j 1 ij X i . Because m X M R and substituting Mj 1 1 ij , we m have Z j R 1 aij ( X i R) j 1, 2, , n. ij aij we pick the portfolio weights m i 1, , m for and Mj 1 1 ij , from M which it follows that Z j 1 ij X i .But every f portfolio in can be written as a portfolio combination of ( Z1 , , Z n ) and R. f , X m , R) span , then Corollary 2.10: A necessary and sufficient condition for ( X1 , , X m , R) to be the smallest number of feasible portfolio that span is that 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 m exist number {aij }such that Z j Z j 1 aij ( X i X i ) m j 1, , n for . Therefore Z j b j 1 aij X i m where b j Z j 1 aij X i by Theorem 2.10 span f It follows from Corollary 2.10 that a necessary and sufficient condition for f nontrivial spanning of is that some of the risky securities are redundant securities. By Theorem 2.10, if investors agree on a set of portfolios ( X1 , , X m , R) such that m Z j R 1 aij ( X i R ) j 1, 2, , n. and if they agree on the number {a } ,then ( X1 , , X m , R) span f even if investors do not agree on the joint distribution of ( X1 , , X m , R) ij Proposition 2.2: If Z e is the return on a portfolio contained in e , then any portfolio that combines positive amount of Z e with the riskless security is also contained e in , where e is the set of all efficient f portfolios contained in . Proof: Let Z Ze (1 ) R , because Z e is an efficient portfolio, so E{V ( Z e )( Z j R)} 0 Define U (W ) V (aW b) where a 1 and , Hence E{U ( Z )( Z j R)} 0 , b ( 1)R thus 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 ( X1 , , X m ) denote the return on m feasible portfolios. If, for security j, there exist number {aij } such that m Z j Z j 1 aij ( X i X i ) j where E{ jVK (ZeK )} 0 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 construction , E{ j } 0 and hence cov( Z ,V ) 0 Therefore the systematic risk of portfolio p, K is zero. From Theorem 2.4 bp Zp R therefore Z j R m aij ( X i R) m p 1 K Hence, if the return on a security can be written in this linear form relative to the portfolios ( X1, , 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 ( X1 , , X m , R) e span the set of efficient portfolios . Proof: Z 1 w j Z j 1 w j [ R 1 aij ( X i R ) j ] n K e n K m K 1 w j R 1 1 w j aij ( X i R ) 1 wKj j n n K m m K R 1 iK ( X i R) K m K w Where 1 j aij K i n K 1 wK j m j Construct portfolio Z 1m iK X i (1 1m iK ) R K K K Z Z Thus e where E{ | Z} 0 K K Z Hence, for 0 , e is riskier than Z, K which contradicts that Z e is and efficient K portfolio. So 0 . We get the result. K w Theorem 2.13: Let j denote the fraction of efficient portfolio K allocation to e security j, j 1, , n. ( X1 , , X m , R) span if and only if there exist number {aij } for every m security j such that Z j R 1 aij ( X i R) j m K n K K E { | X } 0, w where 1 j aij for j 1 i i i 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 every efficient portfolio K. If X R , then K from Corollary 2.1 0 for every e efficient portfolio K and R span . Otherwise, from Theorem 2.2, K 0 for every efficient portfolio. By Theorem 2.13, E{ j | K X } 0 so E{ j | X } 0 e f Since is contained in , any properties proved for portfolios that span e must be properties of portfolio that span f . Z eK K ( X R) R 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 ( X1 , , X m ) linearly independent with finite variances and if the return on security j, Z j has a finite variance, then the {aij } i 1, , m, in Theorems 2.12 and 2.13 are given by m aij 1 vik cov( X K , Z j ) where vik is the ikth 1 element 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 riskaverse investor’s optimal portfolio allocated to security j, j 1, , n, then for every such risk-averse investor j w k j * k w j , k 1, 2, ,n Because every optimal portfolio is an efficient portfolio and the holding of risky securities in every efficient portfolio are proportional to the holding in X. If there exist numbers j where , j, k 1, n * and 1 j ,then the portfolio with * * ( , proportions 1 n ) is called the Optimal * j j * k k Combination of Risky Assets. e e ( X , R ) Proposition 2.5: If span , then is a convex set. ,n Z e1 1 ( X R) R Z e2 2 ( X R ) R Proof: Let 1 2 Z Z (1 ) Z and 1 2 , e e . By substitution, the expression for Z can be 1 Z ( Z rewritten as e R ) R , where ( 2 )(1 ) .Therefore by Proposition 1 2.2, Z is an efficient portfolio. It follow by induction that for any integer k and number i such that 0 i 1, i 1, , k and k k k i 1, Z Z 1 i 1 i e is the return on an e efficient portfolio. Hence , 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 security j and VR denote the value of the M riskless security, then j is the fraction of security j held in a market portfolio. M j Vj V n 1 j VR e Theorem 2.14: If is a convex set, and if the securities’ market is in equilibrium, then a market portfolio is an efficient portfolio. Proof: Let there be K risk averse investor n k K Z R in the economy.Define 1 w j ( Z j R ) to be the return on investor k’s optimal K portfolio. In equilibrium, 1 wkj W0k V j , k W where 0 is the initial wealth of investor K n K K, and 1 W0 W0 1 V j VR . Define W W k 1, K . By definition of a market K portfolio 1 wkj k jM j 1, , n. Multiplying by Z j R and summing over j, it follows that K n k K K k k 0 0 w ( Z R) ( Z R) Z R k 1 n 1 j 1 M i j j 1 M K ( Z R) because 1 . Hence, Z M is a convex combination of the returns on K e efficient portfolios. Therefore , if is convex, then the market portfolio is e contained in . 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. K k 1, Z M 1 K Z k K Proposition 2.6: In all portfolio models with homogeneous beliefs and risk-averse investors the equilibrium expected return on the market portfolio exceeds the return on the riskless security. Proof: From the proof of Theorem 2.14 K k Z R ( Z and Corollary 2.1. M 1 k R ) , because Z k R , k 0 . Hence Z M R The market portfolio is the only risky portfolio where the sign of its equilibrium expected excess return can always be predicted. Returning to the special case where e is spanned by a single risky portfolio and the riskless security, the market portfolio is efficient. So the risky spanning portfolio can always be chosen to be the market portfolio. e ( Z , R ) Theorem 2.15: If M span , then the equilibrium expected return on security j can 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 , , n, are unique, (b) value j , j 1, 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, aj cov( Z j , X ) var( X ) , j 1, , n. Where min denote the set of portfolios f contained in such that there exists no f other portfolio in with the same expected return and a smaller variance. Proof: Let ij denote the ijth element of 1 v and ij denote the ijth element of . So all portfolios in min with expect return u, we need solutions the problem min 1 1 i j ij n n S .T Z ( ) R R If then Z ( R) R and j 0, j 1, 2 , n Consider the case when R . The n firstorder conditions are 0 1 j ij u ( Zi 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 n j u 1 vij ( Z i R ) j 1, , 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 bkK as a measure of a security’s systematic risk provided that the chosen for X is such that R . e Theorem 2.17: If ( X , R ) span and if X has a finite variance, then e is contained in min . Proof: Let Ze R ae ( X R) . Let Z p be f the return on any portfolio in such that Ze Z p . By Corollary 2.13 Z R a ( X R) where E{ p } E{ p | X } 0 p p p Therefore a p ae Thus var(Z p ) a2p var( X ) var( p ) a p var( X ) var( Ze ) 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 e ( X , R ) span . 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 k is normal distributed , and because cov( X , k ) 0 , so they are independent. Therefore E{ k } E{ k | X } 0 , From Corollary 2.13 e it follows that ( X , R ) span Theorem 2.19: If p(Z1, , Zn ) is a symmetric function with respect to all its arguments, then there exists a portfolio with return X e such that ( X , R )span . Proof: By hypothesis p(Z1 , Zi , Z n ) p(Zi , 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 ( X1 , , X M ) such that X i and Yi are perfectly correlated, i 1, , m, then ( X1 , , 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 usual format is to derive equilibrium V j 0 given the distribution of V j . Theorem 2.20: If ( X1, , X m ) denote a set of linearly independent portfolios that satisfy the hypothesis of Theorem 2.12, and all securities have finite variances, then a necessary condition for equilibrium in the securities’ market is that V j 1 m Vj0 where m 1 vik cov( X k ,V j )( X j R) vik is the ikth R 1 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 ) m 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 ) We can get V j 1 m Vj0 m 1 vik cov( X k ,V j )( X j R) R Hence, from Theorem 2.20, a sufficient set of information to determine the equilibrium value of security j is the first and second moments for the join distribution of ( X 1 , , X m ,V j ) . Corollary 2.20a: If the hypothesized conditions of Theorem 2.20 hold and if the end-of-period value a security is given by n V 1 jV j then in equilibrium V0 1 jV j 0 n This property of formula is called “ value additivity”. 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 | X1 , , X m} u and E{q} E{q | X1 , , 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 1 fk 0 V j 0 (I j ) where f k 0 is the equilibrium initial value of financial claim 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 u ( X , R ) return X such that span if and only if Ai (W ) 1 (ai bW ) 0 , where Ai is the absolute risk-aversion function for investor i in u . Because the b in the statement of Theorem 2.22 does not have a subscript i , u therefore all investors in must have virtually the same utility function. Cass and Stiglitz (1970) conclude: it is requirement that there be any mutual funds, and not the limitation on the number of mutual funds. This is a negative report on the approach to developing spanning theorems. The End thanks