Chapter 7 Portfolio Theory: 1 Introduction and Overview
Chapter 7 Portfolio Theory: 1 Introduction and Overview
7
Portfolio Theory
Road Map
Part A Introduction to finance.
Part B Valuation of assets, given discount rates.
Part C Determination of risk-adjusted discount rates.
Part D Introduction to derivative securities.
Main Issues
• Returns of Portfolios
• Diversification
• Diversifiable vs. Non-diversifiable risks
• Optimal Portfolio Selection
Chapter 7 Portfolio Theory 7-1
1 Introduction and Overview
In order to understand risk-return trade-off, we observe:
2 Portfolio Returns
Asset 1 2 · · · n
Mean Return r¯ r¯ · · · r¯n
r1 r2 ··· r
r1 σ12 σ12 · · · σ1
r2 σ21 σ22 · · · σ2
... n .
... ... . . ...
rn 2
σn1 σn2 · · · σn
r¯
1 r¯
2 r˜ r˜
1 2
0.0149 0.0100 r˜ 0.007770 0.002095
1
r2̃ 0.002095 0.003587
A portfolio is simply a collections of assets, characterized by the
mean, variances/covariances of their returns, r˜1 , r˜2 , . . . , r˜n :
• Mean returns:
Mean returns Covariance matrix
A portfolio of these two assets is characterized by the value
invested in each asset.
V = V1 + V2 .
Then, w1 + w2 = 1 .
Example. You have $1,000 to invest in IBM and Merck stocks.
If you invest $500 in IBM and $500 in Merck, then wIBM =
wMerck = 500 /1000 = 50% . This is an equally weighted
portfolio of the two stocks.
Example. If you invest $1,500 in IBM and −$500 in Merck (short
sell $500 worth of Merck shares), then wIBM = 1500 /1000 =
150% and wMerck = −500 /1000 = −50% .
Fall 2006 c J. Wang 15.401 Lecture Notes
7-4 Portfolio Theory Chapter 7
Return on a portfolio with two assets
The portfolio return is a weighted average of the individual returns:
Example. Suppose you invest $600 in IBM and $400 in Merck
for a month. If the realized return is 2.5% on IBM and 1.5% on
Merck over the month, what is the return on your total portfolio?
Expected return on a portfolio with two assets
r¯ p = w 1 r¯ 1 + w 2 r¯ 2 .
Variance of return on a portfolio with two assets
2 2
= w1 σ1 + w σ + 2 w1 w2 σ
w1 r˜1 w2 r˜ 2 2
12 . 2 2
w1 r˜1 w1 σ1 w1 w2 σ12
r¯
1 r¯
2 r˜ r˜
1 2
0.0149 0.0100 r˜ 0.007770 0.002095
1
Consider the equally weighted portfolio:r2̃ w0.002095 0.003587
1 = w 2 = 0 . 5 .
= 0.003888
σp = 6.23%.
2.2 Portfolio of Three Assets
2 1 2
σp = × (Sum of all elements of covariance matrix)
3
= 0.003130
σp = 5.59%.
σIBM = 8 .81%
wir˜i.
n
r¯p = E [ rp] = w1 r¯1 + w2 r¯2 + · · · + wn r¯n wir¯i.
=
i=1
σ p = Var[ r˜p] = σ p2 .
Fall 2006 c J. Wang 15.401 Lecture Notes
The variance of portfolio return can be computed by summing up
all the ent ries to the foll
owing tab w 1 r1 w 2 r 2 ··· w n le:
2 2
w 1 r1 w1 σ1 w1 w2 σ12 · · · w1 wn σ1
n
w 2 r2 2 2
w2 w1 σ21 w2 σ2 · · · w2 wn σ
··· n
... ... ... ...
w n rn
2 2
The variance of a sum is not just the sum of variances! We also
need to account for the covariances.
(a) portfolio weights
(b) individual variances
(c) all covariances.
15.401 Lecture Notes c J. Wang Fall 2006
1. Two assets:
Chapter 7 Portfolio Theory
Diversification wit h t wo a sse t s 7-11
1.5
3 1
corre la tion = 0 . 1
0.5
excessreturn
Diversification 0
−1
t im e
2. Multiple assets:
t im e
1.5
0.5
− 0 .5
−1
0 10 20 30 40 50 60
7-12 Portfolio Theory Chapter 7
Example. Given two assets with the same annual return StD,
σ1 = σ2 = 35% , consider a portfolio p with weight w in asset 1
and 1−w in asset 2.
2
σp = w2 σ1 + (1 −w)2 σ2 + 2 w(1 −w)σ12 .
From the plot below, the StD of the portfolio return is less than
the StD of each individual asset.
0.45
0.40 corr(1,2)=0.0 corr(1,2)=1.0 corr(1,2)=0.5 corr(1,2)=-0.5
0.35
0.30
PortfolioStD
0.25
0.20
0.15
0.10
0.05
0.00
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Weight in asset 1
15.401 Lecture Notes c J. Wang Fall 2006
Chapter 7 Portfolio Theory 7-13
ρ¯ = 1.0 and ρ¯ = 0.5
Risk eliminated
by diversification
Total risk of typical
security
Undiversifiable
or market risk
Number of securities in portfolio
Example. An e qually-
w1 r1 · · · wnrn
weighted portfol io of n assets:
w r w 2 2
σ ··· w σ
1 1 1 1 1 wn 1n
..
. ... ... ...
2 2
wnrn wn w1σn1 · · · wn σn
σ .
1 2 σ
• A typical variance term: n ii
1 n 2 − n
= (average variance) + (average covariance).
n n2
As n becomes very large:
4 Optimal Portfolio Selection
How to choose a portfolio:
r¯
✻
maximizing
✻return
✛
minimizing
risk
✲
σ
i
=
1
i=
1
n
ir
¯
i
=
r
¯
p.
i=
1
Fall 2006 c J. Wang 15.401 Lecture Notes
4.1 Solving optimal portfolios “graphically”
6.00
5.00
Return(%,permonth)
Portfolio Frontie
Frontier with
six assets
4.00
r fro m Stocks of
3.00 IBM, Merck, Intel, AT&T, JP Morgan and GE
2.00
1.00
0.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Definition: Given an expected return, the portfolio that minimizes
risk (measured by StD) is a mean-StD frontier portfolio.
Definition: The locus of all frontier portfolios in the mean-StD
plane is called portfolio frontier. The upper part of the portfolio
frontier gives efficient frontier portfolios.
Portfolio Frontier with Two Assets
4.2
Then
r¯p −
w = r¯ .
2
r¯1 −
r¯2
Covariances r˜ r˜
IBM Merck
r˜ 0.007770 0.002095
IBM
r˜ 0.002095 0.003587
Merck
Mean (%) 1.49 1.00
StD (%) 8.81 5.99
Fall 2006 c J. Wang 15.401 Lecture Notes
Without Short Sales
r¯p −r¯2 2 2 r¯p −r¯1 2 2
When short sales are not allowed, w1 ≥ 0 , w2 ≥ 0 and
Return(%,permonth)
r¯ 1 ≥ r¯p ≥ r¯
2.
Portfolio Frontier when Short Sales Are Not Allowed
StandardDeviation (%,perm onth)
15.401 Lecture Notes
2
c J. Wang Fall 2006
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
0 2 4 6 8 10 12
Chapter 7 Portfolio Theory 7-19
With Short Sales
Example. (Contin ued.)
Covariances r˜ r˜
IBM Merck
r˜ 0.007770 0.002095
IBM
r˜ 0.002095 0.003587
Merck
Mean 1.49% 1.00%
StD 8.81% 5.99%
Portfolios of IBM and Merck
Weight in IBM (%) -40 -20 0 20 40 60 80 100 120 140
Mean return (%) 0.80 0.90 1.00 1.10 1.20 1.29 1.39 1.49 1.59 1.69
StD (%) 7.70 6.69 5.99 5.72 5.95 6.62 7.61 8.81 10.16 11.60
Return(%,permonth)
Portfolio Frontier when Short Sales Are Allowed
StandardDeviation (%,perm onth)
7-20 1.6
Chapter 7
Return(%,permonth)
1.5 Portfolio Theory
1.4
Several special situations (without short sales)
1.3
1.2
1
Portfolio Frontier with A Risk-Free Asset
0.9
0.8
0 2 4 6 8 10 12
1.8
1.6
1.4
1.2
corr = 0.397 corr =0.0 corr =1.0 corr =-1.0
1
Return(%,permonth)
0.8
0.6
StandardDeviation (%,perm onth)
0.4
0 Portfolio Frontiers with Special Return Correlation
0 2 4 6 8 10 12
StandardDeviation (%,perm onth)
4.3 Portfolio Frontier with Multiple Assets
With more than two assets, we need to solve the constrained
optimization problem (P).
5.00
Frontier with
3.00
Portfolio Frontier of IBM, Merck, Intel, AT&T, JP Morgan and GE
2.00
1.00
1.8
0.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0
1.7
1.6
1.5
1.4
1.3
1.2
1.1
StandardDeviation (%,perm onth)
0.9
0.8
0 2 4 6 8 10 12
Observation: When more assets are included, the portfolio frontier
improves, i.e., moves toward upper-left: higher mean returns and
lower risk.
Intuition: Since one can choose to ignore the new assets, including
them cannot make one worse off.
5 Portfolio Frontier with A Safe Asset
When there exists a safe (risk-free) asset, each portfolio consists
of the risk-free asset and risky assets.
Observation: A portfolio of risk-free and risky assets can be
viewed as a portfolio of two portfolios:
Example. Consider a portfolio with $40 invested in the risk-free
asset and $30 each in two risky assets, IBM and Merck:
Consider a portfolio p with x invested in a risky portfolio q, and
1−x invested in the risk-free asset. Then,
r¯ p = (1 −x )rF
+ xr¯q
Return(%,permonth)
2 2 2
σp = x σq .
When there is a risk-free asset, the frontier portfolios are
combinations of:
(1) the risk-free asset
Portfolio Frontier with A Risk-Free Asset (Capital Market Line or CML)
StandardDeviation (%,perm onth)
r¯ p −
Frontier portfolios have the highest Sharpe ratio: .
r F
σp
Fall 2006 5.00
c J. Wang 15.401 Lecture Notes
7-24 Portfolio Theory Chapter 7
4.50
6 Summary
4.00
3.50
3.00
2.00
• Diversifiable (non-systematic)
1.00
0.50
• Non-diversifiable (systematic)
0.00
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
4. When there is risk-free asset, frontier portfolios are linear
combinations of
• the risk-free asset, and
• the tangent portfolio.
7 Appendix: Solve Frontier Portfolios
Fall 2006 c J. Wang 15.401 Lecture Notes
7-26 Portfolio Theory Chapter 7
We now outline the steps involved in obtaining the optimal portfolio using the
Solver.
Step 1 Enter the data on
Step 2 Create the cells for the optimization using Solver.
• Specify a required rate of return for the portfolio. (In cell B26, we
specified this to be 0.00%.)
• Enter some initial values for w1 and w2 in cells C26 and D26. We
used the initial values of 0.25 and 0.35. These are the cells that Solver
will change to find the optimal portfolio weights.
• Enter, in E26, the portfolio constraint: w3 = 1 − w1 − w2 .
• Cell F26 simply checks that the weights satisfy the constraint that their
sum is equal to one.
• Enter the portfolio variance in G26 using the weights given in cells C26
and D26 and the information on variances and covariances in Step 1.
Then, use this definition of portfolio variance to specify the portfolio
StD, in cell H26.
• Finally, enter the definition of portfolio return in cell I26, again using
the weights given in cells C26 and D26 and the information on expected
returns that we entered in Step 1.
Step 3 Open Solver (from the “Tools” menu). If Solver is not installed, you
can try and install it from the “Add-ins” choice under “Tools”. If this does
not work, then you will need to get professional help.
• The minimization will be done by changing the cells $C$26:$D$26 —
these are the cells that contain the portfolio weights w1 and w2 .
• Now add the constraint that the portfolio expected return given in cell
I26 must be equal to the return specified in cell B26.
Step 4 Click on “Solve” and the optimal portfolio weights should appear in
cells C26 and D26.
Step 5 You can repeat this process for other values of portfolio expected return
(as in B27, B28, . . . and redoing Steps 2-4 with the appropriate changes).
Step 6 Plotting the points in columns H and I (using x-y plots) will give the
portfolio frontier.
Assignment:
• Problem Set 4.
15.401 Lecture Notes c J. Wang Fall 2006