MPT – Modern Portfolio Theory

Business 2039
K. Hartviksen 1
Key Terms
 Harry Markowitz
 MPT
 Expected return
 required return
 portfolio
 systematic risk
 unsystematic risk
 diversification
 beta coefficient
 security market line
K. Hartviksen
 market premium for
risk
 capital asset pricing
model
 cost of capital
 mean, variance,
standard deviation
 correlation
 capital market line
2
Harry Markowitz
 Modern portfolio theory was initiated by
University of Chicago graduate student,
Harry Markowitz in 1952.
 Markowitz showed how the risk of a portfolio
is NOT just the weighted average sum of the
risks of the individual securities…but rather,
also a function of the degree of comovement
of the returns of those individual assets.
Risk and Return - MPT

 Prior to the establishment of Modern Portfolio

Theory, most people only focused upon investment
returns…they ignored risk.

 With MPT, investors had a tool that they could use

to dramatically reduce the risk of the portfolio
without a significant reduction in the expected return
of the portfolio.

3
 Correlation

 The degree to which the returns of two

stocks co-move is measured by the
correlation coefficient.
 The correlation coefficient between the
returns on two securities will lie in the
range of +1 through - 1.
 +1 is perfect positive correlation.
 -1 is perfect negative correlation.
10
 Perfect Negatively Correlated Returns
over Time

Returns
%
A two-asset portfolio
made up of equal parts
of Stock A and B would
be riskless. There
would be no variability
of the portfolios returns
10% over time.

Returns on Stock A
Returns on Stock B
Returns on Portfolio
1994 1995 1996 Time 11
Ex Post Portfolio Returns
Simply the Weighted Average of Past Returns

n
R p   xi Ri
i 1

Where :
xi  relative weight of asset i
Ri  return on asset i

K. Hartviksen 145
Ex Ante Portfolio Returns
Simply the Weighted Average of Expected Returns

Relative Expected Weighted

Weight Return Return
Stock X 0.400 8.0% 0.03
Stock Y 0.350 15.0% 0.05
Stock Z 0.250 25.0% 0.06
Expected Portfolio Return = 14.70%

K. Hartviksen 145
Grouping Individual Assets
into Portfolios
 The riskiness of a portfolio that is made of different
risky assets is a function of three different factors:
 the riskiness of the individual assets that make up the
portfolio
 the relative weights of the assets in the portfolio
 the degree of comovement of returns of the assets making
up the portfolio
 The standard deviation of a two-asset portfolio may
be measured using the Markowitz model:

 p   w   w  2 w A wB  A, B A B
2
A
2
A
2
B
2
B
 Risk of a Three-asset Portfolio
The data requirements for a three-asset portfolio grows
dramatically if we are using Markowitz Portfolio selection formulae.

We need 3 (three) correlation coefficients between A and B; A and

C; and B and C.
A
ρa,b ρa,c
B C
ρb,c

 p   A2 wA2   B2 wB2   C2 wC2  2wA wB  A, B A B  2wB wC  B ,C B C  2wA wC  A,C A C

Risk of a Four-asset Portfolio

The data requirements for a four-asset portfolio grows dramatically

if we are using Markowitz Portfolio selection formulae.

We need 6 correlation coefficients between A and B; A and C; A

and D; B and C; C and D; and B and D.

A
ρa,b ρa,d
ρa,c
B D
ρb,d
ρb,c ρc,d
C
Diversification Potential

 The potential of an asset to diversify a portfolio is

dependent upon the degree of co-movement of
returns of the asset with those other assets that
make up the portfolio.
 In a simple, two-asset case, if the returns of the two
assets are perfectly negatively correlated it is
possible (depending on the relative weighting) to
eliminate all portfolio risk.
 This is demonstrated through the following chart.
Example of Portfolio
Combinations and Correlation
Perfect
Expected Standard Correlation Positive
Asset Return Deviation Coefficient Correlation –
A 5.0% 15.0% 1 no
B 14.0% 40.0% diversification

Portfolio Components Portfolio Characteristics

Expected Standard
Weight of A Weight of B Return Deviation
100.00% 0.00% 5.00% 15.0%
90.00% 10.00% 5.90% 17.5%
80.00% 20.00% 6.80% 20.0%
70.00% 30.00% 7.70% 22.5%
60.00% 40.00% 8.60% 25.0%
50.00% 50.00% 9.50% 27.5%
40.00% 60.00% 10.40% 30.0%
30.00% 70.00% 11.30% 32.5%
20.00% 80.00% 12.20% 35.0%
10.00% 90.00% 13.10% 37.5%
0.00% 100.00% 14.00% 40.0%
Example of Portfolio
Combinations and Correlation
Positive
Expected Standard Correlation Correlation –
Asset Return Deviation Coefficient weak
A 5.0% 15.0% 0.5 diversification
B 14.0% 40.0% potential

Portfolio Components Portfolio Characteristics

Expected Standard
Weight of A Weight of B Return Deviation
100.00% 0.00% 5.00% 15.0%
90.00% 10.00% 5.90% 15.9%
80.00% 20.00% 6.80% 17.4%
70.00% 30.00% 7.70% 19.5%
60.00% 40.00% 8.60% 21.9%
50.00% 50.00% 9.50% 24.6%
40.00% 60.00% 10.40% 27.5%
30.00% 70.00% 11.30% 30.5%
20.00% 80.00% 12.20% 33.6%
10.00% 90.00% 13.10% 36.8%
0.00% 100.00% 14.00% 40.0%
Example of Portfolio
Combinations and Correlation
No
Expected Standard Correlation Correlation –
Asset Return Deviation Coefficient some
A 5.0% 15.0% 0 diversification
B 14.0% 40.0% potential

Portfolio Components Portfolio Characteristics

Expected Standard
Weight of A Weight of B Return Deviation
100.00% 0.00% 5.00% 15.0% Lower
90.00% 10.00% 5.90% 14.1% risk than
80.00% 20.00% 6.80% 14.4% asset A
70.00% 30.00% 7.70% 15.9%
60.00% 40.00% 8.60% 18.4%
50.00% 50.00% 9.50% 21.4%
40.00% 60.00% 10.40% 24.7%
30.00% 70.00% 11.30% 28.4%
20.00% 80.00% 12.20% 32.1%
10.00% 90.00% 13.10% 36.0%
0.00% 100.00% 14.00% 40.0%
Example of Portfolio
Combinations and Correlation
Negative
Expected Standard Correlation Correlation –
Asset Return Deviation Coefficient greater
A 5.0% 15.0% -0.5 diversification
B 14.0% 40.0% potential

Portfolio Components Portfolio Characteristics

Expected Standard
Weight of A Weight of B Return Deviation
100.00% 0.00% 5.00% 15.0%
90.00% 10.00% 5.90% 12.0%
80.00% 20.00% 6.80% 10.6%
70.00% 30.00% 7.70% 11.3%
60.00% 40.00% 8.60% 13.9%
50.00% 50.00% 9.50% 17.5%
40.00% 60.00% 10.40% 21.6%
30.00% 70.00% 11.30% 26.0%
20.00% 80.00% 12.20% 30.6%
10.00% 90.00% 13.10% 35.3%
0.00% 100.00% 14.00% 40.0%
Example of Portfolio
Combinations and Correlation Perfect
Negative
Expected Standard Correlation Correlation –
Asset Return Deviation Coefficient greatest
A 5.0% 15.0% -1 diversification
B 14.0% 40.0% potential

Portfolio Components Portfolio Characteristics

Expected Standard
Weight of A Weight of B Return Deviation
100.00% 0.00% 5.00% 15.0%
90.00% 10.00% 5.90% 9.5% Risk of the
80.00% 20.00% 6.80% 4.0% portfolio is
70.00% 30.00% 7.70% 1.5% almost
eliminated at
60.00% 40.00% 8.60% 7.0%
70% asset A
50.00% 50.00% 9.50% 12.5%
40.00% 60.00% 10.40% 18.0%
30.00% 70.00% 11.30% 23.5%
20.00% 80.00% 12.20% 29.0%
10.00% 90.00% 13.10% 34.5%
0.00% 100.00% 14.00% 40.0%
 Diversification of a Two Asset Portfolio Demonstrated Graphically

The Effect of Correlation on Portfolio Risk:

The Two-Asset Case

Expected Return B

AB = -0.5
12%
AB = -1

8%
AB = 0

AB= +1

A
4%

0%

0% 10% 20% 30% 40%

Standard Deviation
An Exercise using T-bills, Stocks and Bonds
Base Data: Stocks T-bills Bonds
Expected Return 12.73383 6.151702 7.007872
Standard Deviation 0.168 0.042 0.102

Correlation Coefficient Matrix:

Stocks 1 -0.216 0.048
T-bills -0.216 1.000 0.380
Bonds 0.048 0.380 1.000

Portfolio Combinations:

Weights Portfolio
Expected Standard
Combination Stocks T-bills Bonds Return Variance Deviation
1 100.0% 0.0% 0.0% 12.7 0.0283 16.8%
2 90.0% 10.0% 0.0% 12.1 0.0226 15.0%
3 80.0% 20.0% 0.0% 11.4 0.0177 13.3%
4 70.0% 30.0% 0.0% 10.8 0.0134 11.6%
5 60.0% 40.0% 0.0% 10.1 0.0097 9.9%
6 50.0% 50.0% 0.0% 9.4 0.0067 8.2%
7 40.0% 60.0% 0.0% 8.8 0.0044 6.6%
8 30.0% 70.0% 0.0% 8.1 0.0028 5.3%
9 20.0% 80.0% 0.0% 7.5 0.0018 4.2%
10 10.0% 90.0% 0.0% 6.8 0.0014 3.8%
11 0.0% 100.0% 0.0% 6.2 0.0017 4.2%
Results Using only Three Asset Classes

Attainable Portfolio Combinations

and Efficient Set of Portfolio Combinations

14.0
Efficient Set
Portfolio Expected Return (%)

12.0
Minimum
Variance
10.0
Portfolio

8.0
6.0
4.0
2.0
0.0
0.0 5.0 10.0 15.0 20.0
Standard Deviation of the Portfolio (%)
 Plotting Achievable Portfolio Combinations

Expected Return on
the Portfolio

12%

8%

4%

0%

0% 10% 20% 30% 40%

Standard Deviation of the Portfolio

 The Efficient Frontier

Expected Return on
the Portfolio

12%

8%

4%

0%

0% 10% 20% 30% 40%

Standard Deviation of the Portfolio

 The Capital Market Line
Capital
Market Line

Expected Return on
the Portfolio

12%

8%

4%

Risk-free
rate
0%

0% 10% 20% 30% 40%

Standard Deviation of the Portfolio

 The Capital Market Line and Iso Utility Curves

Highly
A risk-
Risk
taker
Expected Return on Averse
the Portfolio
Investor

12%

Capital
8%
Market Line

4%

Risk-free
rate
0%

0% 10% 20% 30% 40%

Standard Deviation of the Portfolio

 The Capital Market Line and Iso Utility Curves

The risk- A risk-taker’s

taker’s utility curve
Expected Return on
the Portfolio optimal
portfolio
12%
combination

Capital
8%
Market Line

4%

Risk-free
rate
0%

0% 10% 20% 30% 40%

Standard Deviation of the Portfolio

CML versus SML
 Please notice that the CML is used to
illustrate all of the efficient portfolio
combinations available to investors.
 It differs significantly from the SML that
is used to predict the required return
that investors should demand given
the riskiness (beta) of the investment.
Data Limitations
 Because of the need for so much data,
MPT was a theoretical idea for many
years.
 Later, a student of Markowitz, named
William Sharpe worked out a way
around that…creating the Beta
Coefficient as a measure of volatility
and then later developing the CAPM.
CAPM

 The Capital Asset Pricing Model was

the work of William Sharpe, a student
of Harry Markowitz at the University of
Chicago.
 CAPM is an hypothesis …
Capital Asset Pricing Model

Return
Required return = Rf + s [kM - Rf]
%

km
Market Security Market
Premium Line
for risk
Rf
Real Return
Premium for expected inflation

BM=1.0 Beta Coefficient

6
CAPM

 This model is an equilibrium based model.

 It is called a single-factor model because the slope of the SML
is caused by a single measure of risk … the beta.
 Although this model is a simplification of reality…it is robust (it
explains much of what we see happening out there) and it
enjoys widespread use in a great variety of applications.
 Although it is called a ‘pricing model’ there are not prices on
that graph….only risk and return.
 It is called a pricing model because it can be used to help us
determine appropriate prices for securities in the market.
Risk

 Risk is the chance of harm or loss;

danger.
 We know that various asset classes
have yielded very different returns in
the past:
Historical Returns and Standard Deviations
1948 - 941

Average Return Standard Deviation

Canadian common stock 12.73% 16.81%
U.S. common stock (Cdn $) 14.09 16.60
Long term bonds 7.01 10.20
Small cap stocks 15.67 24.40
Inflation 4.52 3.54
Treasury bills 6.15 4.17

___________________
1The Alexander Group
Risk and Return

 The foregoing data point out that those asset

classes that have offered the highest rates of
return, have also offered the highest risk
levels as measured by the standard
deviation of returns.
 The CAPM suggests that investors demand
compensation for risks that they are exposed
to…and these returns are built into the
decision-making process to invest or not.
Capital Asset Pricing Model

Return
Required return = Rf + s [kM - Rf]
%

km
Market Security Market
Premium Line
for risk
Rf
Real Return
Premium for expected inflation

BM=1.0 Beta Coefficient

6
CAPM

 The foregoing graph shows that investors:

 demand compensation for expected inflation
 demand a real rate of return over and above expected inflation
 demand compensation over and above the risk-free rate of return for
any additional risk undertaken.
 We will make the case that investors don’t need
compensation for all of the risk of an investment
because some of that risk can be diversified
away.
 Investors require compensation for risk they can’t
diversify away!
 Beta Coefficient
 The beta is a measure of systematic risk of an investment.
 Systematic risk is the only relevant risk to a diversified investor
according to the CAPM since all other risk may be diversified away.
 Total risk of an investment is measured by the securities’ standard
deviation of returns.
 According to the CAPM total risk may be broken into two parts…
systematic (non-diversifiable) and unsystematic (diversifiable)

TOTAL RISK = SYSTEMATIC RISK + UNSYSTEMATIC RISK

 The beta can be determined by regressing the holding period returns

(HPRs) of the security over 30 periods against the returns on the
overall market.

7
Measuring Risk of the
Individual Security
 Risk is the possibility that the actual return that will
be realized, will turn out to be different than what we
expect (or have forecast).
 This can be measured using standard statistical
measures of dispersion for probability distributions.
They include:
 variance
 standard deviation
 coefficient of variation
Standard Deviation

 The formula for the standard deviation

when analyzing population data
(realized returns) is:
n

 (k i  ki ) 2

  i 1

n 1
Standard Deviation

 The formula for the standard deviation when

analyzing forecast data (ex ante returns) is:

n
  (k
i 1
i  k i ) Pi
2

 it is the square root of the sum of the squared

deviations away from the expected value.
Using Forecasts to Estimate
Beta

The formula for the beta coefficient for a stock ‘s’ is:

Cov ( k s k M )
Bs 
Variance (k M )
Obviously, the calculate a beta for a stock, you must
first calculate the variance of the returns on the
market portfolio as well as the covariance of the
returns on the stock with the returns on the market.
Systematic Risk

 The returns on most assets in our economy are influenced by

the health of the ‘system’
 Some companies are more sensitive to systematic changes in
the economy. For example durable goods manufacturers.
 Some companies do better when the economy is doing poorly
(bill collection agencies).
 The beta coefficient measures the systematic risk that the
security possesses.
 Since non-systematic risk can be diversified away, it is
irrelevant to the diversified investor.
Systematic Risk

 We know that the economy goes

through economic cycles of expansion
and contraction as indicated in the
following:
Canada’s Business cycles from 1873-1992
Canada’s Business cycles from 1873-1992

Trough to ExpansionPeak to Contraction

Trough to ExpansionPeak to Contraction
(months from trough to peak)(months from peak to trough)
(months from trough to peak)(months from peak to trough)
Nov 1873 66
Nov 1873 66
May 1879 38 July 1882 32
May 1879 38 July 1882 32
Mar 1885 23 Feb 1887 12
Mar 1885 23 Feb 1887 12
Feb 1888 29 July 1890 9
Feb 1888 29 July 1890 9
Mar 1891 23 Apr 1893 13
Mar 1891 23 Apr 1893 13
Mar 1894 17 Aug 1895 12
Mar 1894 17 Aug 1895 12
Aug 1896 44 Apr 1900 10
Aug 1896 44 Apr 1900 10
Feb 1901 22 Dec 1902 18
Feb 1901 22 Dec 1902 18
June 1904 30 Dec 1906 19
June 1904 30 Dec 1906 19
July 1908 20 Mar 1910 16
July 1908 20 Mar 1910 16
July 1911 16 Nov 1912 25
July 1911 16 Nov 1912 25
Jan 1915 36(WWI) Jan 1918 15
Jan 1915 36(WWI) Jan 1918 15
Apr 1919 14 June 1920 15
Apr 1919 14 June 1920 15
Sep 1921 21 June 1923 14
Sep 1921 21 June 1923 14
Aug 1924 56 Apr 1929 47 (Depression)
Aug 1924 56 Apr 1929 47 (Depression)
Mar 1933 52 July 1937 15 (Depression)
Mar 1933 52 July 1937 15 (Depression)
Oct 1938 80(WWII) June 1945 8
Oct 1938 80(WWII) June 1945 8
Feb 1946 33 Oct 1948 11
Feb 1946 33 Oct 1948 11
Sep 1949 44(Korean War) May 1953 14² 耀 uly 1954
Sep 1949 44(Korean War) May 1953 14² 耀 uly 1954
31 Feb 1957 12
31 Feb 1957 12
Feb 1958 26 Apr 1960 10
Feb 1958 26 Apr 1960 10
Feb 1961 160 June 1974 10
Feb 1961 160 June 1974 10
Apr 1975 58 Feb 1980 6
Apr 1975 58 Feb 1980 6
July 1980 12 July 1981 6
July 1980 12 July 1981 6
Nov 1982 89 Apr 1990 22
Nov 1982 89 Apr 1990 22
Feb 1992
Feb 1992
Companies and Industries

 Some industries (and by implication the companies

that make up the industry) move in concert with the
expansion and contraction of the economy.
 Some lead the overall economy. (stock market)
 Some lag the overall economy. (ie. automotive
industry)
Amount of Systematic Risk

 Some industries may find that their fortunes are

positively correlated with the ebb and flow of the overall
economy…but that this relationship is very insignificant.
 An example might be Imperial Tobacco. This firm does
have a positive beta coefficient, but very little of the
returns of this company can be explained by the beta.
Instead, most of the variability of returns on this stock is
from diversifiable sources.
 A Characteristic line for Imperial Tobacco would show
a very wide dispersion of points around the line. The
R2 would be very low (.05 = 5% or lower).
Characteristic Line for Imperial
Tobacco
Characteristic
Returns on
Line for Imperial
Imperial
Tobacco
Tobacco %

Returns on the
Market %
(TSE 300)
High R2

 An R2 that approaches 1.00 (or 100%) indicates that

the characteristic (regression) line explains virtually
all of the variability in the dependent variable.
 This means that virtually of the risk of the security is
‘systematic’.
 This also means that the regression model has a
strong predictive ability. … if you can predict what
the market will do…then you can predict the returns
on the stock itself with a great deal of accuracy.
Characteristic Line General
Motors
Characteristic
Returns on
Line for GM
General
Motors % (high R2)

Returns on the
Market %
(TSE 300)
Diversifiable Risk
(non-systematic risk)

 Examples of this type of risk include:

 a single company strike
 a spectacular innovation discovered through the company’s
R&D program
 equipment failure for that one company
 management competence or management incompetence for
that particular firm
 a jet carrying the senior management team of the firm crashes
 the patented formula for a new drug discovered by the firm.
 Obviously, diversifiable risk is that unique factor that
influences only the one firm.
Partitioning Risk under the
CAPM
 Remember that the CAPM assumes that total risk (variability of a
security’s returns) can be separated into two distinct components:

Total risk = systematic risk + unsystematic risk

100% = 40% + 60% (GM)
or
100% = 5% + 95% (Imperial Tobacco)

Obviously, if you were to add Imperial Tobacco to your portfolio, you

could diversify away much of the risk of your portfolio. (Not to
mention the fact that Imperial has realized some very high rates of
return in addition to possessing little systematic risk!)
Using the CAPM to Price
Stock
 The CAPM is a ‘fundamental’ analyst’s tool to
estimate the ‘intrinsic’ value of a stock.
 The analyst needs to measure the beta risk of the firm
by using either historical or forecast risk and returns.
 The analyst will then need a forecast for the risk-free
rate as well as the expected return on the market.
 These three estimates will allow the analyst to
calculate the required return that ‘rational’ investors
should expect on such an investment given the other
benchmark returns available in the economy.
Required Return

 The return that a rational investor should demand is

therefore based on market rates and the beta risk of
the investment.
 To find this, you solve for the required return in the
CAPM:

R(k )  R f   s [k M  R f ]
 This is a formula for the straight line that is the SML.
Security Market Line

 This line can easily be plotted.

 Draw Cartesian coordinates.
 Plot the yield on 91-day Government of Canada Treasury Bills
as the risk-free rate of return on the vertical axis.
 On the horizontal axis set a scale that includes Beta=1 (this is
the beta of the market)
 Plot the point in risk-return space that represents your
expected return on the market portfolio at beta =1
 Draw a straight line to connect the two points.
 Plot the required and expected returns for the stock at it’s
beta.
 Plot the Risk-Free Rate

Return
%

Rf

1.0 Beta Coefficient

 Plot Expected Return on the
Market Portfolio
Return
%

km =12%

Rf = 4%

1.0 Beta Coefficient

 Draw the Security Market Line

Return SML
%

km =12%

Rf = 4%

1.0 Beta Coefficient

 Plot Required Return
(Determined by the formula = Rf + s[kM - Rf]

Return SML
%
R(k) = 13.6%

km =12% R(k) = 4% + 1.2[8%] = 13.6%

Rf = 4%

1.0 1.2 Beta Coefficient

 Plot Expected Return
E(k) = weighted average of possible returns

Return SML
%
R(k) = 13.6%
R(k) = 4% + 1.2[8%] = 13.6%
km =12%

E(k)

Rf = 4%

1.0 1.2 Beta Coefficient

 If Expected = Required Return
The stock is properly (fairly) priced in the market. It is in
EQUILIBRIUM.

Return SML
%
R(k) = 13.6%
R(k) = 4% + 1.2[8%] = 13.6%
km =12%
E(k)

Rf = 4%

1.0 1.2 Beta Coefficient

 If E(k) < R(k)
The stock is over-priced. The analyst would issue a sell recommendation in anticipation
of the market becoming ‘efficient’ to this fact. Investors may ‘short’ the stock to take
advantage of the anticipated price decline.

Return SML
%
R(k) = 13.6%
R(k) = 4% + 1.2[8%] = 13.6%
km =12%

E(k) = 9% E(k)

Rf = 4%

1.0 1.2 Beta Coefficient

 Let’s Look at the Pricing
Implications
 In this example:
 E(k) = 9%
 R(k) = 13.6%
 If the market expects the company to pay a dividend of $1.00 next
year, and the stock is currently offering an expected return of 9%, then
it should be priced at:
d1
P0 
E (k s )
$1.00
P0   $11 .11
.09
 But, given the other rates in the economy and our judgement about the
riskiness of this investment we think that this stock should be worth:
$1.00
P0   $7.35
.136
Practical Use of the CAPM

 Regulated utilities justify rate increases using the model to

demonstrate that their shareholders require an appropriate return on
their investment.
 Used to price initial public offerings (IPOs)
 Used to identify over and under value securities
 Used to measure the riskiness of securities/companies
 Used to measure the company’s cost of capital. (The cost of capital is
then used to evaluate capital expansion proposals).
 The model helps us understand the variables that can affect stock
prices…and this guides managerial decisions.
Rf rises
SML2
Return
% SML1

ks2

ks1
Rising
Risinginterest
interestrates
rateswill
will
cause
causeall
allrequired
requiredrates
ratesofof
Rf2 return
returntotoincrease
increaseand
andthis
this
Rf1 will
willforce
forcedown
downstock
stockand
and
bond
bondprices.
prices.

Bs=1.2 Beta Coefficient

The Slope of The SML rises
(indicates growing pessimism about the future of the economy)

SML2
Return
% SML1

ks2
Growing
Growingpessimism
pessimism
will
willcause
causeinvestors
investorstoto
ks1
demand
demandgreater
greater
compensation
compensationfor for
taking
takingononrisk…this
risk…this
will
will mean priceson
mean prices on
Rf1
high
highbeta
betastocks
stockswill
will
fall
fallmore
morethan
thanlow
low
beta
betastocks.
stocks.

Bs=1.2 Beta Coefficient