Fin 2008 & Econ 3060 Investment Chapter 7. Optimal Risky Portfolios Bokyung Park Department of Finance National Taiwan University Spring, 2023 NTU. Dept of Fin. Investment 1 Last Class Capital Allocation Line (CAL) E(rc) = yE (rp ) + (1 − y)rf Capital Allocation Line (CAL) P E(rp) • E (rc ) = rf + E (rp ) − rf p c F rf σp σc = y p slope = Sharpe Ratio 2 Recall Personal decision: based on investor’s risk aversion Capital Allocation Safe Asset Last Class Risky Asset(s) 1 Technical decision: based on assets’/securities' covariance structure with other securities “Portfolio Optimization” N Today Asset Allocation/ Security Selection 3 Overview ▪ Asset allocation with risky assets □ Agenda 1. Expected return, variance, and co-movement of Risky Portfolio □ Agenda 2. Portfolios of Two risky assets □ Agenda 3. Portfolios of Three risky assets □ Agenda 4. Portfolios of N risky assets • Markowitz Portfolio Optimization model NTU. Dept of Fin. Investment 4 Agenda 1 Power of Diversification: A Simple Example Measures of Co-movement Expected Return & Variance of Risky Portfolios 5 Power of Diversification Example: Ice Cream Co. vs. Straw Sandal Co. Pr rain 1/2 sunshine 1/2 mean s.d . Ice Cream Co Straw Sandal Co -2% 4% -1% 2% 1% 3% 0.5% 1.5% Portfolio :1/2 in each stock -1.5% 3% 0.75% 2.25% cannot reduce risk at all by combining 2 stocks Case of Perfect Positive Correlation (=1) 6 Power of Diversification Example: Umbrella Co. vs. Straw Sandal Co. Pr rain 1/2 sunshine 1/2 mean s.d . 2% -1% -1% 2% Portfolio :1/2 in each stock 0.5% 0.5% 0.5% 1.5% 0.5% 1.5% 0.5% 0 Umbrella Co Straw Sandal Co a complete elimination of risk by combining 2 stocks Case of Perfect Negative Correlation ( = -1) 7 Power of Diversification Definition of Diversification ▪ investment strategy designed to reduce risk by spreading the portfolio across many securities / stocks. always useful unless stocks perfectly move together (corr =1) NTU. Dept of Fin. Investment 8 Measures of Co-Movement Covariance ▪ expected value of the product of the deviations from mean for two variables Covariance (ri , rj ) = Cov (ri , ri ) = E [ri − E (ri )][ rj − E (rj )] = ij n = p( s) [ri ( s ) − E (ri )][rj ( s ) − E (rj )] s =1 Correlation ▪ covariance divided by s.d. of the two variables (coefficient) ▪ always lies between -1 and 1 Correlatio n (ri , ri ) = Corr (ri , r j ) = Cov(ri , r j ) V (ri ) V (r j ) = ij i j = 9 Measures of Co-Movement Calculating Covariance / Correlation Example: Perfect Positive Correlation Pr rain 1/2 sunshine 1/2 mean s.d . Ice Cream Co Straw Sandal Co (rIC − rIC )(rSS − rss ) -2% 4% -1% 2% (-3)(-1.5) = 4.5 (3) (1.5) = 4.5 1% 3% 0.5% 1.5% Cov = 4.5 4.5 =1 Corr = (3)(1.5) 10 Measures of Co-Movement Calculating Covariance / Correlation Example: Perfect Negative Correlation Pr rain 1/2 sunshine 1/2 mean s.d . Umbrella Co Straw Sandal Co (rIC − rIC )(rSS − rss ) 4% -2% -1% 2% (3) (-1.5) = -4.5 (-3)(1.5) = -4.5 1% 3% 0.5% 1.5% Cov = -4.5 − 4. 5 Corr = = −1 (3)(1.5) 11 Measures of Co-Movement Calculating Covariance / Correlation Example: No Correlation Pr rain hot sunshine 1/3 1/3 cold sunshine 1/3 mean s.d . Ice Cream Co Straw Sandal Co (rIC − rIC )(rSS − rss ) -2% 4% -1% 0.33% 2.625% -1% 2% 7.25% 2.75% 3.41% (-7/3)(-3.75) = 26.25/3 (11/3)(-0.75) = -8.25/3 (-4/3)(4.5) = -18/3 Cov = 0 Corr = 0 =0 (2.625 )(3.41) 12 Expected Return & Variance of a Risky Portfolio A Portfolio of 2 Risky Assets stock 1 stock 2 Return r1 r2 mean E (r1 ) E (r2 ) variance Var (r1 ) Var (r2 ) = 12 = 22 a “portfolio” of stock 1 and 2 rp= w1r1 + w2 r2 E(rp)= ? V(rp)= ? 13 Expected Return & Variance of a Risky Portfolio Additional Properties of Expected Returns If a and b are constants, then; n E (ar1 + br2 ) = p(s)[ar (s) + br (s)] s =1 1 2 n n s =1 s =1 = a p( s )r1 ( s ) + b p( s )r2 ( s ) = aE (r1 ) + bE (r2 ) 14 Expected Return & Variance of a Risky Portfolio Additional Properties of Variance If a and b are constants, then; V (ar1 + br2 ) = E[( ar1 + br2 ) − (aE (r1 ) + bE (r2 ) )]2 = E[a(r1 − E (r1 ) ) + b(r2 − E (r2 ) )] 2 = E[a (r1 − E (r1 ) ) + b (r2 − E (r2 ) ) + 2ab(r1 − E (r1 ) )(r2 − E (r2 ) )] 2 2 2 2 = a E (r1 − E (r1 )) + b E (r2 − E (r2 )) + 2abE[(r1 − E (r1 ) )(r2 − E (r2 ))] = a 2V (r1 ) + b 2V (r2 ) + 2abCov (r1 , r2 ) 2 2 2 2 15 Expected Return & Variance of a Risky Portfolio Alternative Ways of Calculating Variance and Covariance Variance (r ) = E[[ r − E (r )]2 ] = E[r 2 − 2rE (r ) + (E (r ) )2 ] = E ( r 2 ) − 2 E ( r ) E ( r ) + (E ( r ) ) 2 = E ( r ) − (E ( r ) ) 2 2 Covariance(ri , rj ) = E [ri − E (ri )][rj − E (rj )] = E ri rj − E (ri )rj − E (rj )ri + E (ri ) E (rj )] = E (ri rj ) − E (ri ) E (rj ) − E (rj ) E (ri ) + E (ri ) E (rj )] = E (ri rj ) − E (ri ) E (rj ) 16 Expected Return & Variance of a Risky Portfolio A Portfolio of 2 Risky Assets stock 1 stock 2 Return r1 r2 mean E (r1 ) E (r2 ) variance Var (r1 ) Var (r2 ) = 12 = 22 a “portfolio” of stock 1 and 2 rp= w1r1 + w2 r2 E(rp)= w1 E (r1 ) + w2 E (r2 ) V(rp)= w12 12 + w22 22 + 2w1w2 12 = w12 12 + w22 22 + 2w1w2 1 2 17 Expected Return & Variance of a Risky Portfolio A Portfolio of 2 Risky Assets Variance of a Portfolio of 2 Risky Assets: In Detail 2 2 2 2 w + w V(rp)= 1 1 2 2 + 2 w1 w2 1 2 ( w1 1 + w2 2 ) 2= w12 12 + w22 22 + 2w1w2 1 2 s.d.(rp) ≤ ( w1 1 + w2 2 ) s.d. of a portfolio is smaller than weighted average of individual stock’s s.d. => Power of Diversification!! 18 Expected Return & Variance of a Risky Portfolio A Portfolio of 2 Risky Assets Example: Calculating E(rp)& V(rp) of a portfolio using MS Excel Scenario Probability 1 0.14 2 0.36 3 0.30 4 0.20 rates of return r1 r2 -0.10 -0.35 0.00 0.20 0.10 0.45 0.32 -0.19 Form a portfolio that invests 0.4 in stock 1 and 0.6 in stock 2 What is the expected return and variance of this portfolio? 2 methods (1) Formula Approach (2) Scenario Approach 19 Expected Return & Variance of a Risky Portfolio A Portfolio of 2 Risky Assets Example: Calculating E(rp)& V(rp) of a portfolio using MS Excel w1= w2= 0.4 0.6 Scenario Probability 1 2 3 4 0.14 0.36 0.30 0.20 Mean SD Covariance Correlation What happens when you change these weights? rates of retu rn r1 r2 -0.10 0.00 0.10 0.32 -0.35 0.20 0.45 -0.19 0.08 0.1359 0.12 0.2918 -0.0034 -0.0847 Portfolio retu rn w 1*r1+w 2*r2 -0.2500 0.1200 0.3100 0.0140 0.104 0.1788 0.104 0.1788 ↑ ↑ Formula Approach Scenario Approach utilizing mean, sd, & covariances only utilizing full probability distribution 20 Agenda 2 Effect of Changes in Weights on Expected Return Variance of risky portfolios Derivation of Portfolio Opportunity Set Construction of Optimal Risky Portfolio Construction of Optimal Complete Portfolio 21 Effect of Changes in Weights What Happens when you change the weights? Example: continued rates of return r1 r2 Mean SD 0.08 0.1359 Covariance Correlation 0.12 0.2918 -0.0034 -0.0847 1.00 0.99 Portfolio E(rp) s.d.(rp) 0.1200 0.2918 0.1196 0.2887 0.40 0.60 0.1040 0.1788 0.99 0.01 0.0804 0.1344 1.00 0.00 0.0800 0.1359 w1 0.00 0.01 w2 22 Effect of Changes in Weights Relationship between w1 and E(rp) 23 Effect of Changes in Weights Relationship between w1 and s.d.(rp) 24 Effect of Changes in Weights Relationship between w1 and E(rp) Short Sales Allowed When short-sale is allowed some weights can be negative and some can be greater than one. The securities shorted have negative weights and the securities purchased with the proceeds of short-sale have weights greater than one. → 𝑤1 < 0, 𝑤2 > 1 but the sum of all weights still adds up to one. 𝑤𝑖 = 1 25 Effect of Changes in Weights Relationship between w1 and s.d.(rp) Short Sales Allowed 26 Effect of Changes in Weights What happens when ρ changes? Effect of changes in ρ on [w1 - E(rp)] and [w1 - s.d.(rp)] relationship [w1 - E(rp)] relationship: Not affected!! [w1 – s.d.(rp)] relationship: ? Let’s Check with Excel As ρ gets closer to +1 : s.d.(rp) becomes more linear As ρ gets closer to -1 : minimum s.d.(rp) becomes closer to zero 27 Derivation of Portfolio Opportunity Set Combine [w1 - E(rp)] & [w1 - s.d.(rp)] plot => Create [s.d.(rp) - E(rp)] plot Investment Opportunity Set is the set of available portfolio risk-return combinations E(rp)=10.4% • stock 1 stock 2 A risky portfolio of 40% in stock 1, 60% in stock 2 σp=17.88% 28 Derivation of Portfolio Opportunity Set Short Sales Allowed Combine [w1 - E(rp)] & [w1 - s.d.(rp)] plot => Create [s.d.(rp) - E(rp)] plot What if short sales are allowed? • stock 2 •stock 1 29 Derivation of Portfolio Opportunity Set What happens when ρ changes? Effect of changes in ρ on [s.d.(rp) - E(rp)] relationship Let’s Check with Excel As ρ gets closer to +1 : No benefit from diversification As ρ gets closer to -1 : Possible to eliminate risk completely 30 Expected Return & Variance of a Risky Portfolio ▪ Minimum-variance portfolio □ A standard deviation smaller than that of either of the individual components assets V (ar1 + br2 ) = 𝑤12 𝜎12 + 𝑤22 𝜎22 + 2𝑤1 𝑤2 𝐶𝑂𝑉 𝑟1 , 𝑟2 = w12 12 + w22 22 + 2w1w2 1 2 □ −1 < 𝜌 < 1 𝑤𝑀𝑖𝑛 □ 𝜌 = −1 𝜎22 − 𝐶𝑂𝑉 𝑟1 , 𝑟2 𝑟1 = 2 𝜎1 + 𝜎22 − 2𝐶𝑂𝑉 𝑟1 , 𝑟2 𝜎2 (𝑤1 𝜎1 − 𝑤2 𝜎2 )2 = 0 𝑤𝑀𝑖𝑛 𝑟1 = 𝜎 + 𝜎 1 2 As ρ gets closer to +1 : No benefit from diversification As ρ gets closer to -1 : Possible to eliminate risk completely NTU. Dept of Fin. Investment 32 Construction of Optimal Risky Portfolio Recall the Previous 2 Risky Assets Example Mean SD Covariance Correlation rates of return r1 r2 0.08 0.12 0.1359 0.2918 -0.0034 -0.0847 From above information, we derived portfolio opportunity set stock 2 stock 1 Q: Out of all available portfolios, which one is the optimal? Let’s combine these risky portfolios with risk-free asset 33 Construction of Optimal Risky Portfolio Recall from last class: Portfolio of Risk-free Asset and One Risky Asset ▪ P: risky asset (or a portfolio of risky assets) ▪ rp: (uncertain) rate of return of P => E(rp): expected return, σp: standard deviation of rp Previous class, P was arbitrarily given, so E(rp) and s.d.(rp) was fixed ▪ F: risk-free asset ▪ rf: (certain) rate of return of F ▪ C: portfolio of P and F ▪ rc: rate of return of C => rc = yrp + (1 − y )r f E(rc) = rf + y[ E (rp ) − rf ] c = y p Today, we choose P from available portfolios that provides best result E (rc ) = r f + Optimal Risky Portfolio [ E (rp ) − r f ] p c 34 Construction of Optimal Risky Portfolio A Portfolio of 2 Risky Assets a risky asset P Key Criteria in mixing possible P’s with F: + Risk-free Asset rf = 7% risk-free asset F Optimal Risky Portfolio What is w1 at this point? choose w1 (and w2) such that; Sharpe Ratio = E (rp ) − rf will be maximized CAL 3 CAL 2 • • s.d .(rp ) F rf CAL 1 • • 35 Construction of Optimal Risky Portfolio The Optimal CAL ▪ The optimal CAL is the one with the highest slope (or highest Sharpe ratio). □ The optimal CAL is the tangency line between the risk-free rate and the risky investment opportunity curve □ The risky portfolio that the optimal CAL is tangent with is called the tangency portfolio. 𝑬(𝒓 )−𝒓 𝑬(𝒓 )−𝒓 𝐌𝐚𝐱 𝑺𝒍𝒐𝒑𝒆𝒑 = 𝑾𝒊 𝒑 𝒇 𝝈𝒑 −𝟎 = 𝒑 𝝈𝒑 𝒇 (Sharpe Ratio of the portfolio) s.t 𝑤𝑎 + 𝑤𝑏 =1 𝑤𝑏 = 𝐸 𝑟𝑏 − 𝑟𝑓 𝜎𝑎2 − 𝐸 𝑟𝑎 − 𝑟𝑓 𝐶𝑂𝑉(𝑟𝑎 , 𝑟𝑏 ) 𝐸 𝑟𝑏 − 𝑟𝑓 𝜎𝑎2 + 𝐸 𝑟𝑎 − 𝑟𝑓 𝜎𝑏2 − 𝐸 𝑟𝑏 − 𝑟𝑓 + 𝐸 𝑟𝑎 − 𝑟𝑓 𝐶𝑂𝑉(𝑟𝑎 , 𝑟𝑏 ) NTU. Dept of Fin. Investment 36 Construction of Optimal Risky Portfolio A Portfolio of 2 Risky Assets + Risk-free (7%) Optimal Risky Portfolio P (using a formula) w1 = 52% w2 = 48% E(rp) = 9.92% s.d.(rp) = 15.14% Sharpe Ratio = E (rp ) − rf 9.92 − 7 = 15.14 s.d .(rp ) Q: What about asset allocation between optimal risky portfolio & risk-free asset? = 19.3% need information about investor’s risk aversion 37 Construction of Optimal Complete Portfolio A Portfolio of 2 Risky Assets + Risk-free (7%) CAL + Risk Aversion (ex. A = 4) y = * Optimal Risky Portfolio E ( rp ) − r f A p2 0.0992 − 0.07 = 4 * 0.1514 2 indifference curve = 31.85% Optimal Complete Portfolio What is y* (proportion invested in risky asset) at this point? NTU. Dept of Fin. Investment 38 Construction of Optimal Complete Portfolio Optimal Complete Portfolio (2 risky assets): Summary Assumptions r1 Mean SD Covariance 0.08 0.1359 -0.0034 r2 0.12 0.2918 rf = 7% A=4 w1 = 52% w2 = 48% E(rp) = 9.92% s.d.(rp) = 15.14% y * = 31.85% Final Asset Allocation Proportion invested in stock 1 = 52%*31.85% = 16.56% Proportion invested in stock 2 = 48%*31.85% = 15.29% Proportion invested in risk-free asset = 1-31.85% = 68.15% E(rc) = 31.85%*9.92% + 68.15%*7% = 7.93% your advice to your client!!! s.d.(rC) = 31.85%*15.14% = 4.82% 39 Agenda 3 Variance of a Portfolio: 3 Risky Assets Derivation of Portfolio Opportunity Set Construction of Optimal Risky Portfolio Construction of Optimal Complete Portfolio 42 Variance of a Risky Portfolio A Portfolio of 3 Risky Assets Recall variance of a portfolio of 2 risky assets 2 2 2 2 V(rp)= w1 1 + w2 2 + 2w1w2 1 2 = w + w + 2w1w2 12 2 2 1 1 2 2 2 2 = w1 w1 11 + w2 w2 22 + w1w2 12 + w1w2 12 w1 w2 w1 w1*w1*Cov(r1,r1) w2*w1*Cov(r2,r1) w2 w1*w2*Cov(r1,r2) w2*w2*Cov(r2,r2) 43 Variance of a Risky Portfolio A Portfolio of 3 Risky Assets Variance of a portfolio of 3 risky assets w1 w2 w3 w1 w2 w3 w1*w1*Cov(r1,r1) w2*w1*Cov(r2,r1) w3*w1*Cov(r3,r1) w1*w2*Cov(r1,r2) w2*w2*Cov(r2,r2) w3*w2*Cov(r3,r2) w1*w3*Cov(r1,r3) w2*w3*Cov(r2,r3) w3*w3*Cov(r3,r3) V(rp)= w12 12 + w22 22 + w32 32 + 2w1w2 12 + 2 w2 w3 23 + 2w1w3 13 44 Variance of a Risky Portfolio A Portfolio of 3 Risky Assets Recall Previous 2 Risky Assets Example Scenario Probability 1 2 3 4 0.14 0.36 0.30 0.20 Mean SD between Covariance Correlation rates of retu rn r1 r2 => Let’s add a 3rd stock r3 -0.10 0.00 0.10 0.32 -0.35 0.20 0.45 -0.19 0.14 -0.10 0.17 0.05 0.08 0.1359 0.12 0.2918 0.0446 0.1163 Form a portfolio of 3 stocks 1-2 2-3 1-3 -0.0034 -0.0847 0.0016 0.0483 0.0028 0.1753 45 Variance of a Risky Portfolio A Portfolio of 3 Risky Assets Example: continued rates of return r1 r2 Mean SD between Covariance Correlation w1 0.3 r3 0.08 0.1359 0.12 0.2918 0.0446 0.1163 1-2 -0.0034 -0.0847 2-3 0.0016 0.0483 1-3 0.0028 0.1753 w2 Portfolio w3 E(rp) s.d.(rp) 0.2 0.0929 0.1521 0.5 What happens when you change these weights? 46 Derivation of Portfolio Opportunity Set What Happens when you change the weights? Example: continued rates of return r1 r2 Mean SD between Covariance Correlation w1 0.08 0.1359 0.12 0.2918 0.0446 0.1163 1-2 -0.0034 -0.0847 2-3 0.0016 0.0483 1-3 0.0028 0.1753 w2 0.0 0.0 0.9 1.0 r3 0.0 0.1 0.1 0.0 Portfolio w3 E(rp) s.d.(rp) 1.0 0.0446 0.1163 0.1163 0.0000 0.9 0.1338 0.0932 0.0 0.0 0.0840 0.1233 0.0800 0.1359 47 Derivation of Portfolio Opportunity Set Portfolios of 3 Risky Assets stock 2 stock 1 stock 3 48 Derivation of Portfolio Opportunity Set Portfolios of 3 Risky Assets stock 2 stock 1 Previous 2 stocks case: Portfolio opportunity set with stocks 1 and 2 only stock 3 49 Derivation of Portfolio Opportunity Set Portfolios of 3 Risky Assets → If we add more assets, the curve extends more to the Northwest corner of the graph (diversification) 50 Derivation of Portfolio Opportunity Set Portfolios of 3 Risky Assets “Efficient Frontier” global minimumvariance portfolio • stock 2 Efficient Frontier – graph representing the set of portfolios that maximizes expected return at each level of portfolio risk. stock 1 stock 3 minimum - variance frontier The efficient frontier shows us the best possible returns we can get for any risk (standard deviation) 51 Derivation of Portfolio Opportunity Set Q: How do we obtain weights of Efficient Portfolios? => Let’s use Excel Solver [Efficient Weights based on Excel results] w1 w2 w3 E(rp) s.d.(rp) 0.00 0.25 0.50 0.68 0.56 0.43 0.31 0.15 1.00 0.75 0.50 0.28 0.21 0.13 0.06 0.00 0.00 0.00 0.00 0.04 0.23 0.43 0.63 0.85 0.12 0.11 0.10 0.09 0.08 0.07 0.06 0.05 0.2918 0.2186 0.1556 0.1194 0.1014 0.0915 0.0925 0.1042 52 Derivation of Portfolio Opportunity Set Efficient Portfolios with 3 Risky Assets 53 Construction of Optimal Risky Portfolio Mean SD between Covariance Correlation rates of return r1 r2 0.08 0.12 0.1359 0.2918 r3 0.0446 0.1163 1-2 -0.0034 -0.0847 1-3 0.0028 0.1753 2-3 0.0016 0.0483 From above information, we derived efficient frontier Q: Out of all available portfolios, which one is the optimal? Let’s combine these risky portfolios with risk-free asset NTU. Dept of Fin. Investment 54 Construction of Optimal Risky Portfolio + Risk-free Asset rf = 2% a risky asset P risk-free asset F A Portfolio of 3 Risky Assets Key Criteria in mixing possible P’s with F: CAL 2 Optimal Risky Portfolio What is w1 at this point? choose w1 (and w2) such that; Sharpe Ratio = E ( rp ) − r f • • s.d .(rp ) will be maximized F rf CAL 1 • NTU. Dept of Fin. Investment 55 Construction of Optimal Risky Portfolio A Portfolio of 3 Risky Assets + Risk-free (2%) [Optimal Risky Portfolio based on Excel Solver] Optimal Risky Portfolio 56 Construction of Optimal Risky Portfolio A Portfolio of 3 Risky Assets + Risk-free (2%) Optimal Risky Portfolio (based on Excel Solver) w1 = 60% w2 = 23% w3 = 17% E(rp) = 8.3% s.d.(rp) = 10.61% Sharpe Ratio = E (rp ) − rf s.d .(rp ) 8.3 − 2 = 10.61 What about asset allocation between optimal risky portfolio & risk-free asset? = 59.4% need information about investor’s risk aversion 57 Construction of Optimal Complete Portfolio A Portfolio of 3 Risky Assets + Risk-free (2%) + Risk Aversion (ex. A = 6) indifference curve y = * What is y* (proportion invested in risky asset) at this point? Optimal Complete Portfolio E ( rp ) − r f A p2 0.083 − 0.02 = 6 * 0.10612 = 93.27% NTU. Dept of Fin. Investment 58 58 Construction of Optimal Complete Portfolio Optimal Complete Portfolio (3 risky assets): Summary Assumptions Mean SD between Covariance Correlation rates of return r1 r2 0.08 0.12 0.1359 0.2918 r3 0.0446 0.1163 1-2 -0.0034 -0.0847 1-3 0.0028 0.1753 2-3 0.0016 0.0483 rf = 2% A=6 w1 = 60% w2 = 23% w3 = 17% E(rp) = 8.3% s.d.(rp) = 10.61% y * = 93.27% NTU. Dept of Fin. Investment 59 Construction of Optimal Complete Portfolio Optimal Complete Portfolio (3 risky assets): Summary Final Asset Allocation Proportion invested in stock 1 = 60%*93.27% = 55.96% Proportion invested in stock 2 = 23%*93.27% = 21.45% Proportion invested in stock 3 = 17%*93.27% = 15.86% Proportion invested in risk-free asset = 1-93.27% = 6.73% E(rc) = 93.27%*8.3% + 6.73%*2% = 7.88% s.d.(rC) = 93.27%*10.61% = 9.896% NTU. Dept of Fin. Investment your advice to your client!!! 60 Construction of Optimal Complete Portfolio General Steps to Obtain Optimal Complete Portfolio 1. Specify the return characteristics of all securities (E(ri), s.d.(ri),Cov(ri,rj)) 2. Establish the risky portfolio - Calculate the optimal risky portfolio, P (i.e. the weights, wi’s) - Calculate the properties of P (E(rp), s.d.(rp), Sharpe Ratio) using wi’s 3. Allocate funds between the (optimal) risky portfolio and risk-free asset - Determine the weights between P and risk-free asset - Calculate final weights allocated to each security in P NTU. Dept of Fin. Investment 61 Agenda 4 N risky assets Expected Return & Variance of Risky Portfolios Diversification and Portfolio Risk 62 Expected Return & Variance of a Risky Portfolio Multiple Risky Assets stock 1…stock N Return mean r1 E (r1 ) rN E (rN ) variance Var (r1 ) Var (rN ) a “portfolio” of N stocks N rp = wi ri i =1 portfolio weight on stock i N w = 1 i =1 i N N E (rp ) = E wi ri = wi E (ri ) i =1 i =1 N Var (rp ) = Var wi ri = i =1 ? 63 Expected Return & Variance of a Risky Portfolio Multiple Risky Assets Recall variance of a portfolio of 3 risky assets w1 w2 w3 w1 w2 w2 w1*w1*Cov(r1,r1) w2*w1*Cov(r2,r1) w3*w1*Cov(r3,r1) w1*w2*Cov(r1,r2) w2*w2*Cov(r2,r2) w3*w2*Cov(r3,r2) w1*w3*Cov(r1,r3) w2*w3*Cov(r2,r3) w3*w3*Cov(r3,r3) V(rp)= w12 12 + w22 22 + w32 32 + 2w1w2 12 + 2w2 w3 23 + 2w1w3 13 64 Expected Return & Variance of a Risky Portfolio Multiple Risky Assets Variance of a portfolio of N risky assets w1 | wN w1 ………………..….. wN w1*w1*Cov(r1,r1) | wN*w1*Cov(rN,r1) ………………..….. w1*wN*Cov(r1,rN) | wN*wN*Cov(rN,rN) ………………..….. N V(rp)= N N w Var (r ) + w w Cov(r , r ) i =1 2 i i i j i j i j 65 Expected Return & Variance of a Risky Portfolio Multiple Risky Assets The shaded boxes contain (weighted) variance terms; 1 2 3 the remainder contain (weighted) covariance terms. 4 weights 5 6 To calculate the portfolio’s variance, add up the boxes N 1 2 3 4 5 6 N weights 66 Expected Return & Variance of a Risky Portfolio Multiple Risky Assets stock 1…stock N Return r1 rN a “portfolio” of N stocks N rp = wi ri i =1 mean E (r1 ) E (rN ) variance Var (r1 ) Var (rN ) portfolio weight on stock i N w = 1 i =1 i N N E (rp ) = E wi ri = wi E (ri ) i =1 i =1 N Var (rp ) = Var wi ri i =1 N N = wi w j Cov(ri , rj ) i =1 j =1 N N i =1 i j N = wi2Var (ri ) + wi w j Cov(ri , rj ) 67 Diversification and Portfolio Risk What happens to portfolio variance when N goes to infinity? N N N x Var (r ) + x x Cov(r , r ) i =1 2 i i i j i j i 2 N N 1 1 1 Var ( r ) + Cov(ri , r j ) i N N i =1 N i j N assume xi = 1/N, then j 1 = N 2 N 1 1 N N Var (ri ) + Cov(ri , rj ) N N i j i =1 N N Var (r ) 1 = N i i =1 N 1 1 + ( N 2 − N ) N N N Cov(r , r ) i i j j N2 − N 1 1 = average variance + 1 − average covariance N N 68 Diversification and Portfolio Risk What happens to portfolio variance when N goes to infinity? N N i =1 i j N 2 x i Var (ri ) + xi x jCov(ri , rj ) 1 1 average variance + 1 − average covariance N N assume xi = 1/N, then = further assume all securities have common standard deviation, σ, and all security pairs have a common correlation coefficient, ρ. Then; 1 2 1 2 = + 1 − N N 69 Diversification and Portfolio Risk What happens to portfolio variance when N goes to infinity? (σ = 0.5) ∞ 0.00 31.62 Source: BKM, Chap 7 70 Diversification and Portfolio Risk What happens to portfolio variance when N goes to infinity? Case 1: Average Covariance = 0 Case 2: Average Covariance > 0 average covariance 71 Diversification and Portfolio Risk What happens to portfolio variance when N goes to infinity? Source: Statman, 1987, Journal of Financial and Quantitative Analysis 72 Derivation of Efficient Frontier Consider stocks A, B, and C We can form portfolios from A and B or from B and C (blue curves) We can also make portfolios out of the portfolios we created (black curve) This provides us even greater diversification benefits and shifts the possibilities to higher returns and lower standard deviations. Derivation of Efficient Frontier E(Rp) The composite of all stock sets constitutes the efficient frontier Global Minimum Variance Portfolio Each half egg shell represents the possible weighted combinations for two stocks (portfolios). σp Derivation of Efficient Frontier Mean-Variance Optimization N = wi E (ri ) i =1 N N = wi w j Cov(ri , rj ) i =1 j =1 75 Markowitz’s Key Result / Implications What did Markowitz’s dissertation committee think about his paper ? …His work on portfolio theory, of which this article was the beginning, won him a share of the Nobel prize in economics in 1990. Apparently, the importance of his work was not widely-recognized at first. Milton Friedman, a member of Markowitz’s doctoral dissertation committee and who also became a Nobel laureate, questioned whether the work met the requirements for an economics Ph.D. See Bernstein (Bernstein 1992). Bernstein, Peter L., 1992, Capital Ideas: The Improbable Origins of Modern Wall Street (Free Press). Requoted from G. Pennacchi, 2002, Asset Pricing 76 Markowitz’s Key Result / Implications ▪ The phrase “don’t put all your eggs in one basket” existed long before modern finance theory. □ It was not until 1952, however, that Harry Markowitz published a formal model of portfolio selection embodying diversification principles □ His model is precisely step one of portfolio management: the identification of the efficient set of portfolios, or the efficient frontier of risky assets. • The principal idea behind the frontier set of risky portfolios is that, for any risk level, we are interested only in that portfolio with the highest expected return. • Alternatively, the frontier is the set of portfolios that minimizes the variance for any target expected return. 77 Derivation of Efficient Frontier Optimum Portfolio Selection (without risk-free asset) E(Rp) indifference curves Global Minimum Variance Portfolio σp NTU. Dept of Fin. Investment 78 Derivation of Optimal Risky Portfolio Introduce Risk Free Asset (J. Tobin, 1958) risky assets risk free asset Return rp rf mean E (rp ) E (rf ) = rf variance Var (rp ) Var (rf ) = p2 = 2f = 0 definition of risk free asset a “combination” of risky assets and risk free asset rc = yrp + (1 − y )rf E (rc ) = E yrp + (1 − y)rf = yE (rp ) + (1 − y)rf Var (rc ) = Var yrp + (1 − y)rf = y 2 p2 E (rc ) = rf + E (rp ) − rf NTU. Dept of Fin. Investment p c => CAL 79 Derivation of Optimal Risky Portfolio Optimal Risky Portfolio CAL3 E(Rc) p3 = Optimal Risky Portfolio Risk-free rate (ex. T-Bill) CAL2 CAL1 E(rp3) slope = Sharpe Ratio E (rc ) = rf + rf σp1σp2 σp3 E (rp ) − rf p c σc 80 Derivation of Optimal Risky Portfolio Optimum Portfolio Selection (with risk-free asset) CAL3 E(Rc) indifference curves E(rp3) p3 = Optimal Risky Portfolio rf σp3 σc 81 Portfolio Selection: A Summary Step 1: Construction of Efficient Frontier Specify the return characteristics of all securities (E(ri), s.d.(ri),Cov(ri,rj)) : inputs to mean variance optimization * number of estimates required for N securities N estimates of expected returns, N estimates of variances (N2 – N)/2 estimates of covariances Total: 0.5N(N+3) number of estimates ex. If N = 50, then total estimates = 1,325 Obtain efficient frontier through mean-variance optimization 82 Portfolio Selection: A Summary Step 2: Construction of Optimal Risky Portfolio Introduce risk-free asset Mix (i.e. form a portfolio) with risky portfolios along the frontier Obtain CAL with the highest slope (Sharpe Ratio) Obtain optimal risky portfolio P P will be the same for all clients/investors!!!! Separation Property 84 Portfolio Selection: A Summary Step 3: Construction of Optimal Complete Portfolio Determine the weights between P and risk-free asset => Depends on individual’s preferences towards risk (risk aversion) 85 Portfolio Selection: A Summary Step 3: Construction of Optimal Complete Portfolio Example: Risk Tolerance Questionnaire 86 Portfolio Selection: A Summary Separation Property ▪ Portfolio choice problem may be separated into two independent tasks - Product Decision (steps 1& 2): determination of the optimal risky portfolio (will generally be the same for all investors) = > technical issue - Consumption Decision (step 3) : allocation between risky portfolio and risk-free asset (can be different across investors) = > personal issue 87 Quiz σP = Absolute value [wAσA − wBσB] 0 = 5 × wA − [10 (1 – wA)] wA = 0.6667 The expected rate of return for this risk-free portfolio is: E(r) = (0.6667 × 10) + (0.3333 × 15) = 11.667% Therefore, the risk-free rate is: 11.667% NTU. Dept of Fin. Investment 89 Exercise Questions (BKM) ▪ Problem set #7-12 ▪ For midterm exam, □ Agenda 1,2, and 4 in these slides □ However, you should understand key concepts in Agenda 3. NTU. Dept of Fin. Investment 90