This preview shows pages 1–11. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: MeanVariance Optimization Outline Types of Securities Return Statistics and their Estimation Formulation of the MeanVariance Problem Solution of the MeanVariance Problem Inclusion of RiskFree Borrowing/Lending The Mutual Fund Principle (Separation Theorem) and the Capital Market Line Types of Securities Treasury Bills (Tbills) Bonds Government Bonds Corporate Bonds Stocks: large, medium, small caps, index (S&P 500, TSX 300) RiskFree vs. Risky assets RiskFree rate: r f Rate of return on TBills Returns on other assets are often expressed in terms of their spread above the riskfree rate: ) ( asset asset f f r r r r + = Return Statistics: Expected Return (Reward) For a discrete random variable with P[R=x i ]=p i For a continuous distribution with probability density function p(x): i i i R x p R E ∑ = = μ ] [ ∫ = = dx x xp R E R ) ( ] [ μ Return Statistics: Variance and Standard Deviation (Risk) ∑ = = i R i i R R x p R E 2 2 2 ) ( ] ) [( μ μ σ For a discrete random variable with P[R=x i ]=p i For a continuous distribution with probability density function p(x): The standard deviation σ R is the square root of the variance. It measures the variability of the returns. dx x p x R E R R R ) ( ) ( ] ) [( 2 2 2 ∫ = = μ μ σ Historical Estimates {r i , i=1,…,N} observed (historical) rate of return in period i. Empirical Estimates: ∑ ∑ = = = = N i R i R N i i R r N r N 1 2 2 1 ) ˆ ( 1 1 ˆ 1 ˆ μ σ μ Covariance For two assets A and B with discrete returns distributions, with scenarios indexed by i, P[R A =x i and R B =y i ]=p i : If the returns are continuous random variables, with joint distribution f(x,y) ∑ = ⋅ = = i B i A i i B B A A B A AB y x p R R E R R ) )( ( )] ( ) [( ) , cov( μ μ μ μ σ ∫ ∫ = ⋅ = = dxdy y x f y x R R E R R B A B B A A B A AB ) , ( ) )( ( )] ( ) [( ) , cov( μ μ μ μ σ Correlation The correlation between R A and R B is: The correlation must satisfy: ρ AB measures the degree of linear dependence. ρ=1 implies that R A and R B satisfy a linear relationship. ρ=0 (uncorrelated returns) does not imply independence. b A AB AB σ σ σ ρ = 1 1 ≤ ≤ AB ρ Historical Estimation {r Ai , r Bi ,i=1,…,N} observed (historical) rates of return in period i. ) ˆ )( ˆ ( 1 1 ˆ 1 B Bi N i A Ai AB r r N μ μ σ = ∑ = Return Statistics for Portfolios Now consider S securities with returns given by the random variables R j . Suppose that we invest the fraction w j of wealth in security j....
View
Full
Document
This note was uploaded on 09/18/2011 for the course ACTSC 372 taught by Professor Maryhardy during the Winter '09 term at Waterloo.
 Winter '09
 MARYHARDY

Click to edit the document details