Financial Analysis, Planning and Forecasting Theory and Application Chapter 7 Risk Estimation and Diversification By Alice C. Lee San Francisco State University John C. Lee J.P. Morgan Chase Cheng F. Lee Rutgers University 1 Outline 7.1 Introduction 7.2 Risk classification 7.3 Portfolio analysis and applications 7.4 The market rate of return and market risk premium 7.5 Determination of commercial lending rates 7.6 The dominance principle and performance evaluation 7.7 Summary Appendix 7A. Estimation of market risk premium Appendix 7B. Normal distribution and Value at Risk Appendix 7C. Derivation of Minimum-Variance Portfolio Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight 2 7.1 Risk classification Total risk=Business risk + Financial risk 3 7.1 Risk classification 7.1A : Method n Expected ROI X X i Pi (7.1) i 1 ROI: return on investment n Variance of ROI ( Xi X ) Pi 2 2 i 1 (7.2) where Xi is ROI and Piis the probability in the ith state of economy Coefficient of variation (CV) X (7.3) 4 7.1 Risk classification 7.1B : Example Table 7-1 State of Economy State Occurring (pi) ROI (A1) of GM(XAI) ROI (Bi) of Ford (XBi) Boom .20 15% 14% Normal .60 12% 11% Recession .20 4% 3% GM Ford Interval Chance of Occurrence .11 – .037 < X < .11 + .037 68.27% .11 – .074 < X < .11 + .074 95.45% .11 – .111 < X < .11 + .111 99.73% .10 – .037 < X < .10 + .037 68.27 % .11 – .074 < X < .11 + .074 95.45% .11 – .111 < X < .11 + .111 99.73% 5 7.2 Portfolio analysis and applications Expected rate of return on a portfolio Variance and standard deviations of a portfolio The efficient portfolios Corporate application of diversification 6 7.3 Portfolio Analysis and Application R P Wa R a Wb Rb Wc Rc (.4) (.1) (.3) (.05) (.3) (.12) .091 R P : expected return rate for portfolio p Wa ,b,c : the propositions invested in securities a,b, and c. n R p Wi Ri where Wi 1 i 1 i 1 1 N COV (W1 R1 ,W2 R2 ) (W1 R1t W1 R1 )(W2 R2t W2 R 2 ) N t 1 W1W2 N ( R1t R 1 )( R2t R 2 ) N t 1 W1W2Cov( R1 , R2 ) n (7.4) (7.5) 7 7.3 Portfolio Analysis and Applications 2 1 N VAR (W1 R1t W2 R2t ) W1R1t W2 R2t W1 R1 W2 R 2 N t 1 W12 12 W22 22 2W1W2COV ( R1 , R2 ) 2 P (7.6) P 1 N 2 ( R R ) P pt N t 1 (7.7) R P : expected return rate for portfolio p Rpt : the portfolio's rate of return in period t Wi : the propositions invested in security i 8 7.3 Portfolio analysis and applications Figure 7-2 Two Portfolios with same mean and different variance 9 7.3 Portfolio analysis and applications P W12VAR( R1 ) W22VAR( R2 ) 2W1W2 COV ( R1 , R2 ) 1 2 COV( R1 , R2 ) 121 2 (7.8) (7.9) Find W1to minimize p by Appendix 7.C 2 ( 2 12 1 ) W1 2 2 1 2 212 1 2 p n 2 2 W i i i 1 n (7.10) WW j 1,i j (7.11) n i 1 i j ij i j 10 7.3 Portfolio analysis and applications Figure 7-3 The Correlation Coefficients 11 7.3 Portfolio analysis and applications Efficient Portfolios W12 12 2 2 W 2 2 W20W1 20,1 W1W20 1,20 2 2 W20 20 Under the mean-variance framework, a security or portfolio is efficient if E(A) > E(B) and var(A) = var(B) Or E(A) = E(B) and var(A) < var (B). 12 7.3 Portfolio analysis and applications Figure 7-4 Efficient Frontier in Portfolio Analysis 13 7.3 Portfolio analysis and applications WMRK .001645 .001069 .1007 by Equation (7.10) .006206 .001645 2(.001069) 2 ( 2 12 1 ) W1 2 1 22 212 1 2 WJNJ 1 WMRK .8993 Table 7-2 Variance-Covariance Matrix MRK JNJ 2 MRK MRK , JNJ .006206 .001069 MRK JNJ ,MRK 2 JNJ JNJ .001069 .001645 14 7.3 Portfolio analysis and applications R MRK 0.002101, R JNJ 0.004233 E ( RP ) (.1007)(.002101) (.8993)(.004233) .004018 p2 (.1007) 2 (.006206) (.8993) 2 (.001645) 2(.1007)(.8993)(.001069) .001587 MRK , JNJ 12 JNJ MRK .001069 1 2 (.006206) (.001645) .001069 .001069 (.078781)(.040554) .003195 .334695 1 2 15 7.4 The market rate of return and market risk premium Table 7-3 Market Returns and T-bill Rates by Quarters I t I t 1 I t 1 (A) Market Return Annualized 3month T-bill Rates (B) Quarterly 3month T-bill Rates (A-B) Quarterly Rates Premium Year Quarter S & P 500 Index 03 4 164.92 04 1 159.17 -3.49 9.52 2.38 -5.87 2 153.17 -3.77 9.87 2.47 -6.24 3 166.09 8.44 10.37 2.59 5.84 4 167.23 .69 8.06 2.02 -1.33 1 180.65 8.02 8.52 2.13 5.89 2 191.84 6.19 6.95 1.74 4.46 05 16 7.4 The market rate of return and market risk premium Table 7-3 Market Returns and T-bill Rates by Quarters (Cont’d) Year 06 Mean Annualized 3month T-bill Rates (B) Quarterly 3month T-bill Rates (A-B) Quarterly Rates Premium Quarter S & P 500 Index (A) Market Return 3 182.07 -5.09 7.10 1.78 -6.87 4 211.27 16.04 7.10 1.78 14.26 1 238.90 13.08 6.56 1.64 11.44 2 250.84 5.00 6.21 1.55 3.45 3 231.32 -7.78 5.21 1.30 -9.08 4 242.17 4.69 5.53 1.38 3.31 195.36 3.50 7.58 1.90 1.61 17 7.5 Determination of commercial lending rates Table 7.4 Economic condition Rf (A) Probability (B) EBIT (C) Probability (D) Rp Boom 10% .25 $2.5m 1.5 .5 .40 .30 .30 2% 3 5 Normal 9 .50 $2.5m 1.5 .5 .40 .30 .30 2 3 5 Poor 8 .25 $2.5m 1.5 .5 .40 .30 .30 2 3 5 18 7.5 Determination of commercial lending rates Table 7-5 Economic condition (A) Rf (B) Probability (C) Rp (D) Probability (B X D) Probability of Occurrence (A + C) Lending Rate Boom 10% .25 2% 3 5 .40 .30 .30 .100 .075 .075 12% 13 15 Normal 9 .50 2 3 5 .40 .30 .30 .200 .150 .150 11 12 14 Poor 8 .25 2 3 5 .40 .30 .30 .100 .075 .075 1.000 10 11 13 19 7.5 Determination of commercial lending rates R (.100)(12%) (.75)(13) (.075)15 (.200)11 (.150)(12) (.150)(14) (.100)10 (.074)(13%) 12.2% (.100)(12 12.2) 2 (.075)(13 12.2) 2 (.075)(15 12) 2 (.200)(11 12.2) 2 (.150)(12 12.2) 2 (.150)(14 12.2) 2 (.100)(10 12.2) (.075)(11 12.2) (.075)(13 12.2) .004 .048 .588 .288 .006 .436 .484 .108 .048 1.43% 2 2 1 2 2 20 7.6 The dominance principle and performance evaluation Figure 7-5 Distribution of Leading Rate(R) 21 7.6 The dominance principle and performance evaluation Figure 7-6 The Dominance Principle in Portfolio Analysis 22 7.6 The dominance principle and performance evaluation Table 7-6 The example of Sharpe Performance Measure Smith Fund Jones Fund Average return ( R ) 18% 16% Standard deviation (σ) 20% 15% Risk-free rate = Rf = 9.5% RS Rf .18 .095 SPS .425 S .2 R J R f .16 .095 SPJ .433 J .15 23 7.7 Summary In Chapter 7, we defined the basic concepts of risk and risk measurement. Based on the relationship of risk and return, we demonstrated the efficient portfolio concept and its implementation, as well as the dominance principle and performance measures. Interest rates and market rates of return were used as measurements to show how the commercial lending rate and the market risk premium are calculated. 24 Appendix 7A. Estimation of market risk premium Table 7A-1 Summary Statistics of Annual Returns (1926-2006) Geometric Mean Arithmetic Mean Standard Deviation S&P 500 Index 6.02% 7.90% 19.39% U.S. Treasury Bills (3 Month) 3.79% 3.83% 3.45% Long-Term Corporate Bonds (20 Year) 6.44% 7.04% 3.05% Long-Term Government Bonds (20 Year) 5.24% 5.60% 9.04% Series Sources: (1) The Center for Research in Security Prices, Wharton School of Business, The University of Pennsylvania. (2) Federal Reserve Economic Data, The Federal Reserve Bank of St. Louis. 25 Appendix 7A. Estimation of market risk premium Exhibit 7A-1: Derived Series: Summary Statistics of Annual Component Returns (1926-2006) Series Geometric Mean Arithmetic Mean Standard Deviation Equity risk premia (stockbills) 2.13% 4.06% 19.48% Default premia (LT corps-LT govts.) 1.1 1.44 8.13 Horizon premia (LT govts. - bills) 1.34 1.76 9.4 Real interest rates (bills – inflation) .63 .71 4.03 Distribution 26 Appendix 7A. Estimation of market risk premium Exhibit 7A-2: Simulated Total Return Distributions of Common Stock (1977-2000) by Geometric Average Annual Rates 27 Appendix 7B. Normal distribution and Value at Risk f x ( x) 1 2 ( x )2 e 2 2 Figure 7B-1 Probability Density Function for a Normal Distribution, Showing the Probability That a Normal Random Variable Lies between a and b (Shaded Area) 28 Appendix 7B. Normal distribution and Value at Risk Figure 7B-2 Probability Density Function of Normal Random Variables with Equal Variances: Mean 2 is Greater Than 1. Figure 7B-3 Probability Density Functions of Normal Distributions with Equal Means and Different Variances 29 Appendix 7B. Normal distribution and Value at Risk Table 7B-1 Probability, P, That a Normal Random Variable with Mean Standard Deviation σ lies between K – σ and K – σ. P K/ σ .50 .674 .60 .842 .70 1.036 .80 1.281 .90 1.645 .95 1.960 and Mean =12. If the investor may believe there is a 50% chance that the actual return will be between 10.5% and 13.5%. K=(13.5-10.5)/2=1.5 and K/ σ=0.674 2 Then σ=1.5/0.674=2.2255, =4.95 30 Appendix 7B. Normal distribution and Value at Risk Figure 7B-5 For a Normal Random Variable with Mean 12, Standard Deviation 4.95, the Probability is .5 of a Value between 10.5 and 13.5 31 Appendix 7C. Derivation of Minimum-Variance Portfolio Min . w 2 1 p 2 P w12 12 w22 22 2w1w2Cov( R1 , R2 ) (7.C.2) w12 12 (1 w1 )2 22 2w1 (1 w1 )Cov( R1 , R2 ) w12 12 22 2w1 22 w12 22 2w1Cov( R1 , R2 ) 2w12Cov( R1 , R2 ) By taking partial derivative of p2 with respect to w1, we obtain p2 w1 0 2w1 12 2 22 2w1 22 2Cov( R1 , R2 ) 4 w1 Cov( R1 , R2 ) 0 [ 12 22 2Cov( R1 , R2 )]w1 22 Cov( R1 , R2 ) 22 Cov( R1 , R2 ) w1 2 1 22 2Cov( R1 , R2 ) 32 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight Max S p w1 E ( Rp ) R f p where E(R ) = expected rates of return for portfolio P. R = risk free rates of return S p = Sharpe performance measure p as defined in equation (7.C.1) of Appendix C p f 33 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight S p 0 (7.D.1) E ( R p ) w1 E ( R1 ) w2 E ( R2 ) (7.D.2) w1 p [w w 2w1w2Cov( R1 , R2 )] 2 1 2 1 2 2 2 2 w1 w2 1 1/ 2 (7.D.3) (7.D.4) 34 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight E ( R p ) R f w1 E ( R1 ) w2 E ( R2 ) R f w1E ( R1 ) (1 w1 ) E ( R2 ) R f f ( w1 ) p [ w12 12 w22 22 2w1w2Cov( R1 , R2 )]1/ 2 [ w12 12 (1 w1 ) 2 22 2w1 (1 w1 )Cov( R1 , R2 )]1/ 2 g ( w2 ) f ( w1 ) S p g ( w2 ) f ( w1 ) g ( w2 ) f ( w1 ) g ( w2 ) 0 2 w1 w2 [ g ( w2 )] (7.D.5) (7.D.6) (7.D.7) 35 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight f '( w1 ) g '( w1 ) f ( w1 ) E ( R1 ) E ( R2 ) w1 (7.D.8) g ( w1 ) w1 1 1 1 2 2 2 2 2 [ w1 1 (1 w1 ) 2 2 w1 (1 w1 )Cov( R1 , R2 )] 2 [2 w1 12 2 w1 22 2 22 2Cov( R1 , R2 ) 4 w1Cov( R1, R2 )] [ w1 12 w1 22 22 Cov( R1 , R2 ) 2 w1Cov( R1 , R2 )] [w12 12 (1 w1 ) 2 22 2 w1 (1 w1 )Cov( R1 , R2 )]1/ 2 (7.D.9) 36 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight f '( w1 ) g ( w1 ) f ( w1 ) g '( w1 ) 0 (7.D.10) f '( w1 ) g ( w1 ) f ( w1 ) g '( w1 ) [ E ( R1 ) E ( R2 )] [ w (1 w1 ) 2 w1 (1 w1 )Cov( R1 , R2 )] 2 1 2 1 2 2 2 1 2 [ w1 E ( R1 ) (1 w1 ) E ( R2 ) R f ] [ w1 12 w1 22 22 Cov( R1 , R2 ) 2w1Cov( R1 , R2 )] 1 [ w (1 w1 ) 2w1 (1 w1 )Cov( R1 , R2 )] 2 1 2 1 2 2 2 2 (7.D.11)37 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight [ E ( R1 ) E ( R2 )] [ w12 12 (1 w1 ) 2 22 2w1 (1 w1 )Cov ( R1 , R2 )] [ w1 E ( R1 ) (1 w1 ) E ( R2 ) R f ] [ w1 12 w1 22 22 Cov ( R1 , R2 ) 2w1Cov ( R1 , R2 )] (7.D.12) 38 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight Left hand side of equation (D12): [ E ( R1 ) E ( R2 )] [ w12 12 (1 w1 ) 2 22 2 w1 (1 w1 )Cov( R1 , R2 )] [ E ( R1 ) E ( R2 )] [ w12 12 22 2w1 22 w12 22 2w1Cov( R1 , R2 ) 2w12Cov( R1 , R2 )] [ E ( R1 ) E ( R2 )] w12 [ 12 22 2Cov( R1 , R2 )] 2w1[Cov( R1 , R2 ) 22 ] 22 [ E ( R1 ) E ( R2 )] w12 [ 12 22 2Cov( R1 , R2 )] [ E ( R1 ) E ( R2 )] 2w1[Cov( R1 , R2 ) 22 ] [ E ( R1 ) E ( R2 )] 22 [ E ( R1 ) E ( R2 )] [ 12 22 2Cov( R1 , R2 )]w12 2[ E ( R1 ) E ( R2 )] [Cov( R1 , R2 ) 22 ]w1 [ E ( R1 ) E ( R2 )] 22 39 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight Right hand side of equation (D12) [ w1 E ( R1 ) (1 w1 ) E ( R2 ) rf ] [ w1 12 w1 22 22 Cov( R1 , R2 ) 2w1Cov( R1 , R2 )] {w1[ E ( R1 ) E ( R2 ) w1 E ( R2 ) R f ]} [ w1 12 w1 22 2 w1Cov( R1 , R2 ) 22 Cov( R1 , R2 )] {w1[ E ( R1 ) E ( R2 )] [ E ( R2 ) R f ]} {w1[ 12 22 2Cov( R1 , R2 )] Cov ( R1 , R2 ) 22 } w1[ E ( R1 ) E ( R2 )] w1[ 12 22 2Cov( R1 , R2 )] w1[ E ( R1 ) E ( R2 )] [Cov( R1 , R2 ) 22 ] [ E ( R2 ) R f ] w1[ 12 22 2Cov( R1 , R2 )] [ E ( R2 ) R f ] [Cov( R1 , R2 ) 22 ] [ E ( R1 ) E ( R2 )] [ 12 22 2Cov( R1 , R2 )]w12 [ E ( R1 ) E ( R2 )] [Cov( R1 , R2 ) 22 ]w1 [ E ( R2 ) R f ] [ 12 22 2Cov( R1 , R2 )]w1 [ E ( R2 ) R f ] [Cov( R1 , R2 ) 22 ] 40 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight [ E ( R1 ) E ( R2 )] [Cov( R1 , R2 ) 22 ]w1 [ E ( R1 ) E ( R2 )] 22 [ E( R2 ) R f ] [12 22 2Cov( R1 , R2 )]w1 [ E( R2 ) R f ] [Cov( R1, R2 ) 22 ] (7.D.13) [ E( R1 ) E( R2 )]2 22 2 [ E( R2 ) R f ] [Cov( R1, R2 ) 22 ] [ E( R2 ) R f ] [1 2 2Cov( R1, R2 )]w1 [ E( R1 ) E( R2 )] [Cov( R1, R2 ) 22 ]w1 (7.D.14) 41 Appendix 7D. Sharpe Performance Approach to Derive Optimal Weight [ E ( R1 ) E ( R2 ) E ( R2 ) R f ] 22 [ E ( R2 ) R f ]Cov( R1 , R2 ) {[ E ( R2 ) R f ] 12 [ E ( R2 ) R f ] 22 [ E ( R2 ) R f ][2Cov( R1 , R2 )] [ E ( R1 ) E ( R2 )]Cov( R1 , R2 ) [ E ( R1 ) E ( R2 )] 22 }w1 [ E ( R1 ) R f ] 22 [ E ( R2 ) R f ] 12 [ E ( R1 ) R f E ( R2 ) R f ]Cov( R1 , R2 )] w1 w1 [ E ( R1 ) R f ] 22 [ E ( R2 ) R f ]Cov( R1 , R2 ) [ E ( R1 ) R f ] 22 [ E ( R2 ) R f ]12 [ E ( R1 ) R f E ( R2 ) R f ]Cov( R1 , R2 ) 42